Citation
A discourse analysis of expository Appalachian English

Material Information

Title:
A discourse analysis of expository Appalachian English
Creator:
Montgomery, Michael, 1950-
Publication Date:
Language:
English
Physical Description:
viii, 182 leaves : map ; 28 cm.

Subjects

Subjects / Keywords:
Anaphora ( jstor )
Counties ( jstor )
Discourse ( jstor )
Field research ( jstor )
Language ( jstor )
Linguistics ( jstor )
Monologues ( jstor )
Narratives ( jstor )
Spoken communication ( jstor )
Towns ( jstor )
English language -- Dialects -- Tennessee -- White Pine ( lcsh )
English language -- Discourse analysis ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Includes bibliographical references (leaves 172-181).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Michael Bryant Montgomery.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
06442834 ( OCLC )
ocm06442834
0023314048 ( ALEPH )

Downloads

This item has the following downloads:

EVBK1YOW4_BJMZCN.xml

AA00012898_00001.pdf

discourseanalysi00mont_0049.txt

AA00012898_00001_0108.txt

discourseanalysi00mont_0179.txt

AA00012898_00001_0027.txt

discourseanalysi00mont_0096.txt

discourseanalysi00mont_0113.txt

AA00012898_00001_0037.txt

discourseanalysi00mont_0168.txt

AA00012898_00001_0014.txt

AA00012898_00001_0063.txt

AA00012898_00001_0018.txt

discourseanalysi00mont_0056.txt

discourseanalysi00mont_0026.txt

discourseanalysi00mont_0157.txt

AA00012898_00001_0184.txt

discourseanalysi00mont_0131.txt

AA00012898_00001_0148.txt

discourseanalysi00mont_0081.txt

discourseanalysi00mont_0011.txt

AA00012898_00001_0021.txt

AA00012898_00001_0010.txt

AA00012898_00001_0022.txt

AA00012898_00001_0073.txt

discourseanalysi00mont_0040.txt

AA00012898_00001_0074.txt

discourseanalysi00mont_0120.txt

AA00012898_00001_0051.txt

discourseanalysi00mont_0189.txt

AA00012898_00001_pdf.txt

AA00012898_00001_0117.txt

AA00012898_00001_0047.txt

discourseanalysi00mont_0084.txt

AA00012898_00001_0150.txt

discourseanalysi00mont_0099.txt

discourseanalysi00mont_0073.txt

discourseanalysi00mont_0055.txt

discourseanalysi00mont_0066.txt

discourseanalysi00mont_0094.txt

discourseanalysi00mont_0063.txt

discourseanalysi00mont_0118.txt

AA00012898_00001_0060.txt

AA00012898_00001_0097.txt

discourseanalysi00mont_0186.txt

AA00012898_00001_0192.txt

AA00012898_00001_0179.txt

AA00012898_00001_0038.txt

AA00012898_00001_0048.txt

AA00012898_00001_0050.txt

discourseanalysi00mont_0030.txt

AA00012898_00001_0118.txt

AA00012898_00001_0158.txt

AA00012898_00001_0008.txt

AA00012898_00001_0067.txt

AA00012898_00001_0167.txt

discourseanalysi00mont_0082.txt

discourseanalysi00mont_0164.txt

discourseanalysi00mont_0108.txt

AA00012898_00001_0094.txt

AA00012898_00001_0114.txt

discourseanalysi00mont_0176.txt

discourseanalysi00mont_0139.txt

AA00012898_00001_0016.txt

AA00012898_00001_0156.txt

AA00012898_00001_0101.txt

AA00012898_00001_0185.txt

discourseanalysi00mont_0141.txt

discourseanalysi00mont_0136.txt

discourseanalysi00mont_0042.txt

AA00012898_00001_0183.txt

AA00012898_00001_0007.txt

discourseanalysi00mont_0098.txt

AA00012898_00001_0011.txt

AA00012898_00001_0189.txt

AA00012898_00001_0081.txt

discourseanalysi00mont_0083.txt

discourseanalysi00mont_0126.txt

discourseanalysi00mont_0057.txt

discourseanalysi00mont_0153.txt

AA00012898_00001_0091.txt

discourseanalysi00mont_0071.txt

AA00012898_00001_0110.txt

discourseanalysi00mont_0079.txt

discourseanalysi00mont_0045.txt

discourseanalysi00mont_0155.txt

AA00012898_00001_0013.txt

AA00012898_00001_0026.txt

discourseanalysi00mont_0087.txt

AA00012898_00001_0129.txt

AA00012898_00001_0049.txt

AA00012898_00001_0154.txt

AA00012898_00001_0070.txt

AA00012898_00001_0134.txt

AA00012898_00001_0036.txt

AA00012898_00001_0121.txt

discourseanalysi00mont_0137.txt

AA00012898_00001_0143.txt

AA00012898_00001_0188.txt

AA00012898_00001_0190.txt

AA00012898_00001_0123.txt

discourseanalysi00mont_0106.txt

AA00012898_00001_0023.txt

discourseanalysi00mont_0188.txt

discourseanalysi00mont_0033.txt

discourseanalysi00mont_0105.txt

AA00012898_00001_0193.txt

discourseanalysi00mont_0111.txt

AA00012898_00001_0145.txt

discourseanalysi00mont_0178.txt

discourseanalysi00mont_0162.txt

AA00012898_00001_0046.txt

discourseanalysi00mont_0167.txt

AA00012898_00001_0100.txt

discourseanalysi00mont_0058.txt

discourseanalysi00mont_0146.txt

discourseanalysi00mont_0032.txt

AA00012898_00001_0165.txt

AA00012898_00001_0163.txt

AA00012898_00001_0115.txt

discourseanalysi00mont_0068.txt

AA00012898_00001_0166.txt

discourseanalysi00mont_0172.txt

discourseanalysi00mont_0159.txt

AA00012898_00001_0019.txt

AA00012898_00001_0059.txt

discourseanalysi00mont_0103.txt

discourseanalysi00mont_0158.txt

discourseanalysi00mont_0130.txt

discourseanalysi00mont_0059.txt

discourseanalysi00mont_0035.txt

AA00012898_00001_0111.txt

discourseanalysi00mont_0017.txt

AA00012898_00001_0039.txt

AA00012898_00001_0015.txt

discourseanalysi00mont_0061.txt

discourseanalysi00mont_0109.txt

AA00012898_00001_0042.txt

discourseanalysi00mont_0190.txt

discourseanalysi00mont_0173.txt

discourseanalysi00mont_0143.txt

AA00012898_00001_0177.txt

discourseanalysi00mont_0124.txt

AA00012898_00001_0144.txt

AA00012898_00001_0171.txt

AA00012898_00001_0012.txt

AA00012898_00001_0041.txt

AA00012898_00001_0125.txt

AA00012898_00001_0137.txt

discourseanalysi00mont_0023.txt

discourseanalysi00mont_0028.txt

discourseanalysi00mont_0088.txt

discourseanalysi00mont_0077.txt

AA00012898_00001_0087.txt

AA00012898_00001_0078.txt

discourseanalysi00mont_0115.txt

AA00012898_00001_0088.txt

discourseanalysi00mont_0142.txt

AA00012898_00001_0002.txt

AA00012898_00001_0113.txt

AA00012898_00001_0116.txt

AA00012898_00001_0109.txt

discourseanalysi00mont_0031.txt

AA00012898_00001_0176.txt

AA00012898_00001_0126.txt

discourseanalysi00mont_0051.txt

AA00012898_00001_0172.txt

discourseanalysi00mont_0132.txt

AA00012898_00001_0061.txt

discourseanalysi00mont_0170.txt

AA00012898_00001_0168.txt

discourseanalysi00mont_0037.txt

discourseanalysi00mont_0138.txt

AA00012898_00001_0182.txt

AA00012898_00001_0093.txt

AA00012898_00001_0140.txt

discourseanalysi00mont_0070.txt

AA00012898_00001_0131.txt

AA00012898_00001_0068.txt

discourseanalysi00mont_0128.txt

discourseanalysi00mont_0163.txt

AA00012898_00001_0139.txt

AA00012898_00001_0159.txt

AA00012898_00001_0005.txt

AA00012898_00001_0092.txt

discourseanalysi00mont_0067.txt

discourseanalysi00mont_0089.txt

AA00012898_00001_0142.txt

discourseanalysi00mont_0180.txt

discourseanalysi00mont_0014.txt

AA00012898_00001_0124.txt

discourseanalysi00mont_0133.txt

AA00012898_00001_0083.txt

AA00012898_00001_0055.txt

discourseanalysi00mont_0065.txt

AA00012898_00001_0095.txt

AA00012898_00001_0096.txt

discourseanalysi00mont_0050.txt

discourseanalysi00mont_0127.txt

AA00012898_00001_0024.txt

discourseanalysi00mont_0043.txt

discourseanalysi00mont_0009.txt

discourseanalysi00mont_0147.txt

discourseanalysi00mont_0034.txt

discourseanalysi00mont_0039.txt

discourseanalysi00mont_0175.txt

discourseanalysi00mont_0075.txt

AA00012898_00001_0130.txt

discourseanalysi00mont_0116.txt

AA00012898_00001_0133.txt

discourseanalysi00mont_0171.txt

AA00012898_00001_0112.txt

AA00012898_00001_0025.txt

AA00012898_00001_0035.txt

discourseanalysi00mont_0184.txt

discourseanalysi00mont_0019.txt

discourseanalysi00mont_0151.txt

discourseanalysi00mont_0097.txt

AA00012898_00001_0076.txt

discourseanalysi00mont_0091.txt

discourseanalysi00mont_0122.txt

discourseanalysi00mont_0015.txt

discourseanalysi00mont_0192.txt

AA00012898_00001_0120.txt

AA00012898_00001_0009.txt

AA00012898_00001_0155.txt

AA00012898_00001_0157.txt

discourseanalysi00mont_0078.txt

discourseanalysi00mont_0117.txt

discourseanalysi00mont_0191.txt

discourseanalysi00mont_0074.txt

discourseanalysi00mont_0054.txt

AA00012898_00001_0099.txt

discourseanalysi00mont_0027.txt

AA00012898_00001_0029.txt

discourseanalysi00mont_0183.txt

AA00012898_00001_0056.txt

AA00012898_00001_0072.txt

AA00012898_00001_0147.txt

discourseanalysi00mont_0007.txt

discourseanalysi00mont_0095.txt

discourseanalysi00mont_0121.txt

discourseanalysi00mont_0125.txt

AA00012898_00001_0057.txt

AA00012898_00001_0187.txt

AA00012898_00001_0065.txt

discourseanalysi00mont_0006.txt

discourseanalysi00mont_0154.txt

discourseanalysi00mont_0010.txt

discourseanalysi00mont_0080.txt

discourseanalysi00mont_0169.txt

discourseanalysi00mont_0053.txt

discourseanalysi00mont_0003.txt

discourseanalysi00mont_0187.txt

AA00012898_00001_0033.txt

AA00012898_00001_0138.txt

discourseanalysi00mont_0076.txt

AA00012898_00001_0119.txt

AA00012898_00001_0181.txt

AA00012898_00001_0054.txt

AA00012898_00001_0146.txt

AA00012898_00001_0062.txt

AA00012898_00001_0175.txt

AA00012898_00001_0089.txt

AA00012898_00001_0174.txt

discourseanalysi00mont_0064.txt

AA00012898_00001_0186.txt

discourseanalysi00mont_0177.txt

discourseanalysi00mont_0062.txt

AA00012898_00001_0030.txt

AA00012898_00001_0173.txt

discourseanalysi00mont_0012.txt

discourseanalysi00mont_0005.txt

discourseanalysi00mont_0165.txt

AA00012898_00001_0086.txt

discourseanalysi00mont_0100.txt

discourseanalysi00mont_0013.txt

discourseanalysi00mont_0119.txt

discourseanalysi00mont_0101.txt

discourseanalysi00mont_0060.txt

AA00012898_00001_0162.txt

discourseanalysi00mont_0152.txt

AA00012898_00001_0082.txt

AA00012898_00001_0003.txt

AA00012898_00001_0044.txt

discourseanalysi00mont_0112.txt

AA00012898_00001_0034.txt

discourseanalysi00mont_0000.txt

AA00012898_00001_0149.txt

discourseanalysi00mont_0104.txt

AA00012898_00001_0103.txt

discourseanalysi00mont_0047.txt

AA00012898_00001_0105.txt

discourseanalysi00mont_0093.txt

AA00012898_00001_0161.txt

AA00012898_00001_0106.txt

AA00012898_00001_0032.txt

AA00012898_00001_0001.txt

discourseanalysi00mont_0182.txt

AA00012898_00001_0136.txt

AA00012898_00001_0191.txt

AA00012898_00001_0164.txt

AA00012898_00001_0069.txt

AA00012898_00001_0127.txt

AA00012898_00001_0066.txt

AA00012898_00001_0102.txt

discourseanalysi00mont_0020.txt

AA00012898_00001_0160.txt

AA00012898_00001_0107.txt

discourseanalysi00mont_0134.txt

AA00012898_00001_0079.txt

discourseanalysi00mont_0018.txt

discourseanalysi00mont_0002.txt

discourseanalysi00mont_0086.txt

discourseanalysi00mont_0022.txt

discourseanalysi00mont_0107.txt

AA00012898_00001_0151.txt

discourseanalysi00mont_0156.txt

AA00012898_00001_0028.txt

AA00012898_00001_0090.txt

AA00012898_00001_0180.txt

AA00012898_00001_0053.txt

AA00012898_00001_0104.txt

discourseanalysi00mont_0150.txt

discourseanalysi00mont_0149.txt

discourseanalysi00mont_0038.txt

discourseanalysi00mont_0148.txt

discourseanalysi00mont_0110.txt

AA00012898_00001_0045.txt

AA00012898_00001_0080.txt

AA00012898_00001_0170.txt

discourseanalysi00mont_0021.txt

AA00012898_00001_0031.txt

discourseanalysi00mont_0048.txt

discourseanalysi00mont_0161.txt

discourseanalysi00mont_0090.txt

AA00012898_00001_0141.txt

discourseanalysi00mont_0001.txt

AA00012898_00001_0084.txt

AA00012898_00001_0152.txt

AA00012898_00001_0004.txt

discourseanalysi00mont_0114.txt

AA00012898_00001_0135.txt

discourseanalysi00mont_0102.txt

AA00012898_00001_0020.txt

AA00012898_00001_0006.txt

AA00012898_00001_0064.txt

discourseanalysi00mont_0072.txt

discourseanalysi00mont_0129.txt

discourseanalysi00mont_0025.txt

discourseanalysi00mont_0024.txt

AA00012898_00001_0052.txt

AA00012898_00001_0071.txt

AA00012898_00001_0132.txt

AA00012898_00001_0169.txt

discourseanalysi00mont_0046.txt

discourseanalysi00mont_0160.txt

discourseanalysi00mont_0085.txt

AA00012898_00001_0122.txt

discourseanalysi00mont_0174.txt

discourseanalysi00mont_0008.txt

discourseanalysi00mont_0052.txt

AA00012898_00001_0040.txt

AA00012898_00001_0085.txt

AA00012898_00001_0128.txt

discourseanalysi00mont_0092.txt

discourseanalysi00mont_0029.txt

AA00012898_00001_0077.txt

discourseanalysi00mont_0185.txt

AA00012898_00001_0075.txt

discourseanalysi00mont_0041.txt

discourseanalysi00mont_0144.txt

discourseanalysi00mont_0145.txt

discourseanalysi00mont_0140.txt

discourseanalysi00mont_0036.txt

discourseanalysi00mont_0166.txt

AA00012898_00001_0058.txt

AA00012898_00001_0017.txt

AA00012898_00001_0153.txt

discourseanalysi00mont_0044.txt

discourseanalysi00mont_0123.txt

discourseanalysi00mont_0135.txt

discourseanalysi00mont_0004.txt

discourseanalysi00mont_0181.txt

AA00012898_00001_0098.txt

EVBK1YOW4_BJMZCN_xml.txt

discourseanalysi00mont_0069.txt

AA00012898_00001_0178.txt

discourseanalysi00mont_0016.txt

AA00012898_00001_0043.txt


Full Text











A DISCOURSE ANALYSIS OF
EXPOSITORY APPALACHIAN ENGLISH





BY
MICHAEL BRYANT MONTGOMERY

















A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY



UNIVERSITY OF FLORIDA























Copyright 1979

by
Michael Bryant Montgomery










ACKNOWLEDGMENTS


For their assistance to me throughout my graduate studies and

especially in undertaking the fieldwork for this dissertation I am

indebted to my parents. In many ways this dissertation would have

been impossible without them, and to them I dedicate it.

The fieldwork for this dissertation could not have been carried

out without the gracious assistance of my contact persons in White

Pine, the Atchleys-Earl, Emogene, Steve, and Philip-who enabled

me to make the necessary contacts with and to establish a rapport

with my informants. To them I will always be grateful.

I also wish to thank the people of White Pine for their hospital-

ity. Their open-hearted welcome and open invitations to return will

long be appreciated.

To the chairman of my committee, Professor William J. Sullivan, I

am thankful for patient direction and many invaluable suggestions.

To the other members of my committee, Professors Jean Casagrande,

Robert Thomson, Roger Thompson, and Norman Markel, I am grateful for

patient tolerance of an erratic and hurried submission of my work.

My colleagues at the Department of English at the University of

Arkansas at Little Rock deserve thanks for their encouragements to

finish this work this past year.

Finally I wish to thank Professor Robert Longacre of the Summer

Institute of Linguistics for his invaluable personal advice to me.









TABLE OF CONTENTS

Page
ACKNOWLEDGMENTS ................ ....... .... iii

ABSTRACT . .. vi

CHAPTER I INTRODUCTION .... 1

General Remarks ................... .... 1
The White Pine Setting ................... 4
The Economic Life of White Pine ... 15
The Religious Life of White Pine .. 17
The Southern Appalachian Region .... 20
The Study of Appalachian Speech .... 33
Fieldwork Procedures ..................... 38
The Sample ................... .... 50
Transcription Procedure ... 53
Fieldwork Journal and Interview Catalog .. 56
Explanation to Informants .. 57
Taping of the Interviews ... 58
The Value of Fieldwork .... 58
Notes .......... .... .............. 60

CHAPTER II BACKGROUND AND METHODOLOGY ... 61

Preface . .. 61
Discourse Studies and the Approach of This Study. ...... 62
The Study of Monologue Expositions .. 72
Oral Paragraph and a Review of Paragraph Studies. .. 74
Longacre's Typology of Discourse Genres .. 84
Notes . 88

CHAPTER III THEMATIC DEVICES .... 89

Introduction ......................... 89
Left Dislocation ................... .... 96
The Variety of Forms .................. 97
Specifications ................... .. 104
Functions . .. 109
Relationship to Subsequent Discourse .. 117
Existential Sentences as Thematic Devices .. 124

CHAPTER IV DEICTICS AND SUMMARY DEVICES. .. .... 131

General Remarks ....................... 131
Left Dislocation with Deictics ... 136
Extrapositions with Deictics. ... 137









Page
Identifying Clauses with Deictics. .. 139
Peak Sentences ................... .. 143

CHAPTER V CONJUNCTIONS AND INTRODUCERS. .. 154

CHAPTER VI SUMMARY AND CONCLUSIONS .... 161

APPENDIX A LIST OF INFORMANTS. 163

APPENDIX B SAMPLE INTERVIEW. .. 164

BIBLIOGRAPHY ....... ............ ....... 172

BIOGRAPHICAL SKETCH .. .. 182









Abstract of Dissertation Presented to the Graduate Council of
the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



A DISCOURSE ANALYSIS OF
EXPOSITORY APPALACHIAN ENGLISH

By

Michael Bryant Montgomery

August 1979

Chairman: William J. Sullivan III
Major Department: Linguistics

This dissertation contributes to the fields of discourse anal-

ysis and Southern Appalachian studies by describing the discourse

functions of several linguistic structures from a large sample of

spontaneous expository (explanatory and descriptive) discourses for

a cross-section of informants from a small East Tennessee hill commu-

nity. The community, White Pine, is a typical Southern Appalachian

one.

Five different contexts are taken into consideration for the

examined linguistic structures: 1) Speaker Type (a representative sam-

ple of forty informants was interviewed); 2) Frequency of Occurrence

(data are quantified when possible); 3) Social Situation (researcher

assumed one role with his informants); 4) Task (a standard interview

was controlled to elicit only expository discourse); and 5) Linguis-

tic (the sentence and discourse position of all data was considered).

This study is based on the premise that speech is hierarchical









as well as linear in structure, that it has, in other words, struc-

tural units larger than the sentence. It hypothesizes in particular

that speech has paragraph structure. If this first hypothesis is

true, oral paragraphs should be describable to some degree of pre-

cision and generality.

Paragraphs may be viewed as having phonological, lexemic, and

grammatical structure. This study examines their grammatical struc-

ture and makes a second hypothesis: that the grammatical structure

of oral expository paragraphs is organized by certain surface struc-

ture cohesive devices. It investigates this hypothesis by considering

the functions) and distribution of the following devices:

1) Introducers of themes into the stream of discourse. Such

introducers include left dislocation (as in "My mother, she's a great

cook"), existentials (as "There's a man here to see you"), it-clefts

(as "It was my brother who called") and WH-clefts (as "What I did was

to completely strike out").

2) Deictics (pointing words such as "that" and "this"), both

cataphoric (forward-pointing) and anaphoric (backward-pointing).

Structures examined are that-extraposition (as "That is good that John

finally left"), identifying sentences (as "That's what I meant"), and

peak sentences (as "That's the best thing is to leave early").

3) Introducers and conjunctions ("and," "but," "now," "then,"

"like," and others) when they accompany left dislocation and assist

in introducing themes.










All structures are examined with the view of identifying what is

typical about them, which is why they are quantified numerically when-

ever possible. This is to insure that the data are not chance findings

of random observation.

It was found that spontaneous expository monologues contain super-

sentential units that are thematically unified and that various surface

structure cohesive devices play significant roles in creating such

unity for the benefit of a hearer. These units are shown to be oral

paragraphs.













CHAPTER 1

INTRODUCTION


General Comments

This dissertation describes the discourse functions of several

linguistic structures from a large sample of spontaneous expository

discourses for a cross-section of informants from a small East Ten-

nessee hill community. The community, White Pine, is viewed as ty-

pically Southern Appalachian. This dissertation thus contributes to

both discourse analysis and Southern Appalachian studies.

This study is based on a set of forty interviews conducted by the

researcher between December 1977 and February 1978 and averaging

slightly more than an hour each in length (forty-five hours total).

The interviews consisted of a small core of questions and were de-

signed to elicit large amounts of expository (explanatory and descrip-

tive) discourse from the informants in the form of spontaneous mono-

logues. Subsequently the interviews were transcribed orthographically

and constitute a 1,070 page typed corpus from which data in this dis-

sertation are taken.

Language, as a social instrument, consists of a set of systematiz-

able patterns of social behavior. These patterns, although presumed

to be repetitive and systematizable on all levels, have been largely

ignored for English at levels above that of the sentence. The








essential claim of this dissertation is that oral discourse structure

must be considered primary to written discourse, a principle which

has long been axiomatic for other levels of linguistic description

(Bloomfield 1933:21ff.). In studying oral discourse linguists have

heretofore dealt with highly formularized types (folktales and other

narratives) in English and not with spontaneous expository discourse.

This study is based on the premise that speech is hierarchical

as well as linear in structure, that it has, in other words, struc-

tural units larger than the sentence and that it has structures whose

domains extend to groups of sentences. Further it leads to the

hypothesis that speech (expository in this case) has describable

paragraph structure.

Oral paragraphs may be viewed as having three different kinds

of structure: phonological, grammatical, and lexemic (Klammer 1971).

This study examines their grammatical structure and its second hypoth-

esis is that the grammatical structure of oral expository paragraphs

is organized by certain surface structure cohesive devices (as des-

cribed by Longacre and Levinsohn 1978). It investigates this hypoth-

esis by examining the functions) and distribution of the following

devices:

1) Introducers of themes into the stream of discourse. Such

introducers include left dislocation (as in "My mother, she's a

great cook"), existentials (as "There's a man here to see you"), it-

clefts (as "It was my brother who called") and WH-clefts (as "What I

did was to completely strike out").

2) Deictics (pointing words such as "that" and "this"), both








cataphoric (forward-pointing) and anaphoric (backward-pointing).

Structures examined are that-exptraposition (as "That is good that

John finally left"), identifying sentences (as "That's what I meant"),

and peak sentences (as "That's the best thing is to leave early").

3) Introducers and conjunctions ("and," "but," "now," "then,"
"like," and others) when they accompany left dislocation and assist

in introducing themes.

All devices are examined with the view of identifying their

typical functions) and distribution within greater discourse block

structures. Hence they are quantified numerically whenever possible.

This is to insure that the data are not chance findings of random

observation. Also to insure the nature of the data gathered, five

different parameters are controlled in this study:

1) The speaker context (a representative sample of forty infor-

mants was interviewed);

2) The frequency context (data are quantified whenever possible);

3) Social context (a matter of controlling the speech style

elicited; the fieldworker-researcher assumed one role with all infor-

mants);

4) Task control (interview was controlled to elicit only explana-

tions and descriptions from informants);

5) Linguistic context (the sentence and discourse position of
all data was taken into account).

For the most part (except the fourth and, in part, the fifth),
these have been controlled in the studies of Labov. Labov









(1966, 1972) has controlled the first two by adopting random sampling

techniques and quantitative analysis of his data. He has controlled

the third by deliberately eliciting several different speech styles

(four in his New York City investigations). And he has taken into ac-

count the sentence context of data by devising the variable rule (Labov

1969). The present study, however, controls both the task context and

the discourse position context. Recent studies in the ethnography of

communication approach take the task and discourse position into ac-

count but tend to ignore the others.

This study, in investigating a selection of surface structure co-

hesive devices, found that they were to a significant extent used by

speakers to communicate thematically unified blocks of information to

hearers and that they reflected the grammatical structure of para-

graphs. This confirms the hypothesis that oral expository speech

(Southern Appalachian in particular) has describable paragraph and

discourse structure.



The White Pine Setting
The small East Tennessee hill town of White Pine (White PIne to

its inhabitants), in which the interviews for this dissertation were

conducted, is no longer the rural village it was only a little more

than a generation ago, and no longer is farming the principal occupa-

tion of its residents. Today the town of White Pine is primarily res-

idential, a "bedroom" community (in 1970, eight times as many of its

people derived their livelihoods from manufacturing as from farming),









and a majority of its citizens are supported by employment in facto-

ries which located in the surrounding area after the Second World War

because of the labor surplus and the low taxes. The factories re-

tarded the emigration so disastrously prevalent throughout East Ten-

nessee as well as other areas of Southern Appalachia in the past two

generations and attracted workers from the foothill and mountain areas

farther east, so that White-Pine's population has continued to grow

steadily since the war (from 1,035 in 1950 to 1,530 in 1970 to 1,622

in 1975 to-unofficially-1,830 in 1977). It remains, however, a

cozy, closeknit (many would say "clannish") community despite the in-

flux of new residents in recent years. Never has it needed more than

one stoplight, and its mayor still dutifully accepts his annual one

hundred fifty dollar check for his thirty hour a week job.

In this study, "White Pine" designates the White Pine community,

consisting of the town and adjacent neighborhoods and having a popula-

tion of approximately 3,000 persons. It is the White Pine community

wherein the forty interviews for this study were conducted.

White Pine (refer to the map on page 6) lies at the extreme

northeastern corner of Jefferson County, forty miles east northeast

of Knoxville and within view of the Great Smoky Mountains. Seven

miles north on Highway 25E is Morristown (population 30,000+) and

five miles south is Cocke County (infamous for its rumrunning in the

20's and sometimes called the "Moonshining Capital of the World" dur-

ing that era). East of White Pine are its sister and erstwhile rival,

but incorporated, community Leadvale and the Nolichucky River; south

and west is Douglas Lake, a Tennessee Valley Authority reservoir.


























>0

-




LL LL

IL
0 6



Ytl


Figure 1.1 Jefferson County, Tennessee
(Source: East Tennessee Development District)









Jefferson County (1960 population 21,493, 1970 population 24,940)

lies in the ridge and valley section of the Great Valley of East Ten-

nessee between the Cumberland Plateau and the Blue Ridge. Today it is

sandwiched between the TVA reservoirs, Cherokee Lake (the dammed Hol-

ston River) and Douglas Lake (the dammed French Broad River) built in

the 1930's. The county is one of the earliest settled areas of what

was to become the state of Tennessee. The Holston, which formed the

county's northern border, brought settlers in from the northeast,

from upper East Tennessee and from Virginia (through the Shenandoah

Valley), while the French Broad, which largely determined Jefferson

County's southern extent and which flows west from Asheville after

flowing north, routed newcomers from the east into the area. Many of

these had received dispensations of land for Revolutionary services.

The county's first white residents arrived in 1783, thirteen years

before Tennessee became a state. The county seat, Dandridge, was

incorporated in 1793.

Settlement in the White Pine vicinity goes back nearly two hun-

dred years, as shown by the fact that two of its present-day churches,

Westminister Presbyterian and Beth Car Methodist, were first organized

in 1787. Although the town of White Pine was not incorporated and

named until 1893, it achieved the status of a community and the char-

acter of a town (and was called "Dandridge Crossing") when a railroad

stop and a post office were established there shortly after the Civil War.

A number of today's White Pine residents can trace back their ancestries









many generations in White Pine itself; some are descendants of the

original settlers in the area.

The rolling terrain of the White Pine area is better than average

farm and pasture land, with corn, tobacco, and wheat being the staple

crops. Although not in the mountains itself, White Pine lies in the

center of the region referred to as Southern Appalachia and is in many

respects a typical small town of the region. Although it is sometimes

supposed that all Appalachian communities are socially and culturally

homogeneous, only the extremely isolated areas actually are. There-

fore White Pine is perhaps more typical of Southern Appalachia because

it is neither homogeneous nor isolated.

To the outsider, White Pine seems a slow-paced and unassuming

small town, all of whose residents know one another. It seems to be a

town with few pretentions. On a not atypical day one can get a good

flavor of the town's personality by visiting the town's center of ac-

tivity and one of its chief public forums, the drugstore at Main and

Maple, where the mayor, the chief of police, one or two city aldermen,

the town's physician, and a local minister hobnob over coffee.

Such an appearance would be rather deceptive, for in recent years

White Pine has entered a period of great transition socially and eco-

nomically. It has also considerably changed in the way it views it-

self (the near doubling of the town's population since 1960 through

annexation and immigration would suggest this anyway). Life has be-

come more complex for its residents as the changes have come. For

such a long time White Pine resisted change of any kind that many of









its citizens feel that the current pace of change in housing, in the

economic pressures on the town, and in the arrival of new people de-

manding more from the town than it can offer is accelerating out of

control. Not too many years ago White Pine residents did pretty much

know everyone else in town, but with new residents recently moving in

faster than ever, they feel uneasy and sense a lack of cohesiveness

they once took for granted. Many of the longtime residents have a

strong sense of the symbiosis of the community, of how everyone plays

a part in supporting the general welfare. A typical expression of

this was made by a thirty nine year old female:


It's a feeling of closeness and connentuity [sic]. If I could
draw you a picture, I would draw you a big ball of people all
wrapped up together, like worms. You know, like you're fishing
with worms, and all those worms are intertwined. That's the
way this town is.

For many, though, this sense of closeness and symbiosis is inseparable

from the "traditional" way of doing things, so that change is threat-

ening.

Probably the most important event in the past fifty years for

White Pine and East Tennessee is the arrival of TVA. For White Pine,

the coming of TVA has in a sense been responsible for most of the

changes during the past forty-five years (TVA began in 1933, and the

construction of Douglas Dam near White Pine began shortly thereafter).

TVA's initial effect was to dislocate many farmers in Jefferson County

from good bottom land and to resettle them, which was the cause of some

resentment. Today few people in the county regret the changes TVA made









possible. As one sixty-year-old farmer who witnessed the relocation

of the thirties put it,


...the landowners were against it. They didn't want uh, they
didn't want to move out....Of course, we realize now that TVA
is our, our main uh source of energy and power and so, I don't
know what we'd do without it.


TVA, through its water management and its provision of electricity, set

the stage for the ensuing economic development of the area, for the

profitability of agriculture, and especially for the coming of indus-

try. Since the war several dozen major factories have relocated to

East Tennessee.

Until after the Second World War and the locating of the facto-

ries White Pine was very much an isolated farming community. Its

sphere of interest hardly extended beyond its boundaries. A half dozen

stores provided the necessities and there was little pressure to de-

velop the town residentially, commercially, or industrially. The sen-

timent was to keep the town as quiet as possible and to keep any con-

scious development out. This is how the state of affairs in White Pine

remained, too, until only a few years ago. Eventually the preoccupa-

tion with prohibiting development was buffeted by demands for more

housing, for recreational facilities, and for planning in general, so

the town has had to undergo considerable growing pains in meeting these

demands. These pains will persist for a long time because the pres-

sures, primarily from the younger generation, will continue.

The most profound development in White Pine over the past thirty

years has been its evolution into a "bedroom" or almost strictly









residential community. Until a generation ago, most people in White

Pine grew up on the farm, but with the coming of industry came rising

economic expectations and consequently a revolution of rising demands

for private residential housing. These demands have partially been met

by the parceling up of farmland into lots and a mobile home park, but

as another generation has grown up, it has become acutely obvious that

the town especially lacks social activities.

In White Pine the principal social institutions are the churches

(ten in the community), and more than likely any social get-together or

meeting will be church-sponsored. Although there are several organized

civic groups (Ruritans, Lions, Beta Sigma Phi) in town, there is little

regular social activity. The occasional special event (church revival,

political candidate speech) may attract a group of citizens, and the

annual Pine Festival is an event that meets some of the social needs of

the town.

The Pine Festival, sponsored by the local Beta Sigma Phi and begun

in 1967, is a one-day collection of shows, contests, and displays held

on July 4 each year. The festival has become the focal event of the

community's life because it is the one occasion when a majority of its

people take part. The festival integrates the community and gives it

a greater sense of self-identity than it has ever had because it is hea-

vily patronized by people from surrounding counties and communities.

Although there is tremendous socioeconomic disparity between the

very rich and the very poor in the community, one finds little overt

residential segregation by economic status. As the thirty-nine-year-









old female quoted earlier put it,


...there are very, very wealthy people in this town, and there
are very, very poor people in this town. But to save my life,
I can't tell you what the difference is. I mean I can't point
to one person and say "Okay, this is a poor person and this is
a wealthy person." They, they blend.

This sentiment evidences the symbiosis referred to earlier.

One major division in the community is that between the middle

and older generations on the one hand and a large part of the younger

generation on the other. This division is a product of the rising ex-

pectations of the town referred to above. In general the community is

more optimistic about its future than a generation ago because of im-

proved local education. A comprehensive high school has been added and

a junior college has been opened five miles up the road. A more diver-

sified job market has also developed. But most youths with professional

and career aspirations still see inadequate prospects in the nearby

area. Many of them, once they leave White Pine, are embarrassed to be

from the town. Often White Pine is for them a place to escape, a place

that can and does trap its ambitious youth. One twenty-nine-year-old

male characterized the town's subtle pressures not to aspire:


...most of the people prior to uh the sixties, when I got out,
they, they were settled. The idea was to, when you got out of
that high school was to get married, cause marriage was stabi-
lization. You got married, you had children right then, as
quick as you could, because that made stay at home. You didn't
carouse. You didn't go on. You weren't apt to, to leave the
area and get involved in things that way.

This desire to escape is usually tempered by an attachment to the

familiar surroundings and personality of the town, though, so that some









of its college-educated citizens settle in town despite its lack of

opportunities. One twenty-five-year-old male college senior puts it

this way:

It's hard to say, you know, when the settling down comes, you
know, when, when I might settle down. I might settle down next
year, you know, or else it might be five more years from now,
maybe longer. But I like to think that I'd come back to this
town, you know. I, I think right now that it's still a fairly
good town, want to plant your roots.

Others choose to remain in the town because of a feeling of belonging

to the community, because they value the cohesiveness (although not as

tight as it once was) of White Pine and live without many of the ameni-

ties of larger communities.

As indicated earlier, White Pine is an underdeveloped community.

Thw town itself, in contrast with the surrounding area, has virtually

no industry and relatively few commercial outlets. It has no motel, no

entertainment outlet, and no restaurant. It has none of these things

by design and promises to resist them in the near future, content to

maintain its "bedroom" status. Although it is not accurate to call the

town entirely unprogressive, one finds that even the young and college-

educated members of the community sometimes have reservations about the

town's growth and expansion. The college senior quoted above says,

regarding "progress,"

progress is on its way. It's on its way, you know, but for
some reason I can't but feel a kind of remorse, you know, ..
it's a reaction that I really do feel sad sometimes and regret and,









and feel that progress is not, is not as good maybe as a lot of
people think it is. It has its disadvantages as well as its ad-
vantages, you know.

The effect of White Pine's resistance to growth and its sense of

community is that it is very slow to accept new residents, which causes

the outsider to view its people as clannish. With the accelerating im-

migration into White Pine, however, of people with a diversity of back-

grounds, the residents have had to learn to assimilate new neighbors

more quickly. The minister of the town's largest Baptist church moved

to White Pine from Kentucky in 1969. He describes the community's re-

ception of new people as follows:


I think the people in White Pine are very slow to accept you
within the community. They're very friendly, but a distant
friendliness. Uh they keep you at a distance, not only when
you move, but throughout your stay in White Pine to some de-
gree. Uh I find the people in White Pine very, very friendly
outwardly, but uh not across the board as in some communities.
They tend to stay within family groups or little cliques of
three or four families. There seems to be no broad base of,
of acceptance,...

This reluctance to accept newcomers manifests two salient traits-tra-

ditionalism and individualism-long considered characteristic of South-

ern Appalachian culture, as pointed out by Ford (1962:11). Later in

this chapter we examine four basic traits discussed by Ford and see how

they aptly characterize White Pine as an Appalachian community. But

first we survey the economic and religious sides of White Pine's life

and then put this study into the perspective of others on Appalachian

English.









The Economic Life of White Pine

White Pine has made significant economic strides in the past two

decades and has undergone a thorough economic transformation in the

past generation, but it remains in many important ways undeveloped.

The percentage of its population below the poverty line (19.4) is well

above the national average (approximately 12). Most of its citizens

must look beyond the city limits for their livelihoods.

Fortunately the surrounding ten mile radius of White Pine has been

able to provide considerable employment opportunities for the last

three decades. This employment has not only stemmed but also largely

reversed the serious problem of emigration of the labor force which

plagued East Tennessee and White Pine in the 30's, 40's, and 50's, al-

though it was not quite as severe as elsewhere in Appalachia, espe-

cially in the coal mining regions of Kentucky and West Virginia. This

reversal has largely been accomplished through the benefits accruing

from the development of TVA. In the 1940's, Jefferson County's emigra-

tion rate surpassed ten percent of its eligible adult work force, and

in the 1950's it was almost as high. But in the 1960's, immigrants

began to outnumber emigrants and the migration patterns underwent a

dramatic turnaround. In the 1970's, the county and especially White

Pine have seen a significant infusion of new residents and a decrease

in emigration. Employment prospects for incoming members of the job

market, although still limited, are better than twenty or thirty years

ago.

The dislocations brought about by the emigration over the years

have left their marks, though. White Pine has lost much of its Black








population (only 1.7 percent of its 1970 population was Black), and

many families have been split because the children have had to mi-

grate north to find jobs. This need to leave home for employment

was aggravated for many natives of East Tennessee and Southern Appa-

lachia by their inability to adjust to the urban North and their fond

attachment to home in the hills. Instances of their choosing to return

to certain unemployment or underemployment in the hills have been

frequently documented, as by Hicks (1976:29-30). This is perhaps not

so common in White Pine as in other areas in the region. A similar

attachment to native ground continues today; relatively few graduating

high school students and other new members of the labor force wish to

settle down elsewhere.

Almost entirely responsible for the improved economic profile for

Jefferson County and White Pine has been the tremendous growth of

local factories, especially those of the American Enka Corporation and

several major furniture firms. Attracting such industry to the area

have been the labor surplus, various tax incentives, and the availa-

bility of cheap electricity. By 1970, more than half (51 percent) of

White Pine's employed population was in manufacturing, eight times as

many as in agriculture, forestry, and fisheries combined (6.3 percent).

In the same year, trade, service, and finance supported another 23.1

percent of White Pine. Construction employed 9.1 percent, mining 0.4

percent, and all other types of employment, 10.0 percent. The move-

ment of workers into manufacturing has entailed the dissolution of the









tenant farming system and the development of more specialized agri-

culture in the area.

This is not to say that farming is no longer a major concern to

most of the people of White Pine. It is. It must be remembered that

most of its citizens are a generation or less off the farm- many of

them still religiously tend kitchen gardens of an acre or two. But

few citizens must still depend on it as their primary source of income.

The new sources of income, the factory payroll primarily, have

brought in their wake profound changes in White Pine (above p. 10).

The better income is more dependable than one might derive from farm-

ing. It has caused White Pine to become a bedroom community. At the

same time, the improved standard of living has raised the optimism of

White Pine's citizens for their future and especially for the future of

their children. For those wishing to break the cycle of having to go

to work immediately after leaving public school, there are more oppor-

tunities than ever before. Within the past decade a junior college has

opened only five miles away, and many of White Pine's youth have de-

cided to undertake some college work rather than enter the job market

right away. The area's economy, although still not providing suffi-

cient skilled positions, continues healthy at the same time that White

Pine, a self-sufficient community a generation ago, finds itself more

and more dependent on the surrounding area.

The Religious Life of White Pine

The community of White Pine lies in the region often referred to

as the "Bible Belt," so called because much of its population still









adheres to fundamentalist beliefs in the Bible and because of the role

of its clergy as both "preachers of the Word" and as community spokes-

men commanding political and social influence. The community of White

Pine is typical of the region. Not only are the churches by far the

most prominent institutions in the community, but also religious values

permeate so thoroughly every aspect of its life that it is simply quite

impossible to deal meaningfully with most aspects of the community's

existence without taking them into consideration.

Churches have long been numerous in Appalachia. A 1935 report of

the United States Department of Agriculture reported there were in 1926

as many as 3.25 churches for each 1,000 inhabitants. It might be

thought that such a high ratio would be accounted for by the fact that

few people could motor to church fifty years ago. But the present-day

ratio in White Pine is the same: ten churches for its approximately

3,000 inhabitants (3.3 per 1,000 people). Church membership, if not

necessarily participation in church activities, is almost required for

a White Pine citizen. Perhaps as much as eighty percent of the popula-

tion in the community are church members.

The two largest churches in White Pine, one a Southern Baptist and

the other a Missionary Baptist, have approximately one thousand members

between them, and the other eight are by no means all small congrega-

tions. Besides five Baptist churches (of several different affilia-

tions), there are two Methodist, two Pentecostal, and one Presbyterian

church. There is not a single non-Protestant church in the entire

county.









Throughout all of the churches one finds a firm fundamentalism

(Puritanical in kind for most churches). One finds also a strong com-

mitment to evangelistic activity and to the need for the individual to

undergo a conversion experience and for the congregation to support

missionary activities and frequent week-long revivals.

All the White Pine inhabitants refer to each one of the ten min-

isters in the community as "Preacher" regardless of an inhabitant's own

church affiliation. Undoubtedly this reflects the fact that until a

generation ago, local churches usually did not have their own ministers.

Rather, they shared a minister with one or more churches outside the

community (the days of the "circuit riders"). It also seems to reflect

the present-day notion that preaching the Word of God is the most impor-

tant duty a minister has. The epithet thus continues even though cler-

gymen have adopted all the everyday duties of the profession (counsel-

ing, visitation, etc.) in recent years. Preachers are also invested

with a great deal of moral authority in the community. They can wield

considerable political influence as well, if they choose, on many is-

sues of local import such as the granting of beer permits or of permits

for businesses to open on Sunday. These are quite important issues in

White Pine.

One finds little evidence of sectarianism in White Pine. There

seems little difference between the Baptists, the Methodists, and the

Presbyterians in either their theology or in their way of conducting

business. Indeed, they coexist quite harmoniously. One's church af-

filiation is viewed generally as dependent on the affiliation of one's

parents or one's spouse, not on one's personal convictions.










The foregoing does not strictly hold for the Pentecostals, who do

consider themselves to an extent to be a people apart. The extremes

of Pentecostalism in Southern Appalachia have long been pointed out

(as in La Barre's study (1964) on the snake-handling cult). One can

find congregations of "snake handlers" not more than twenty miles east

of White Pine in Cocke County. But people in the area overwhelmingly

view such practices as quite objectionable and offensive.

The primary functions of the church itself are to provide relig-

ious services and training for its members, yet in a community where

such a large percentage of the inhabitants are members, churches inevi-

tably offer more than the fulfillment of spiritual needs. Churches in

White Pine provide most of the organized social activities. Scheduled

weekend activities are relatively infrequent in White Pine, but any

given activity is more than likely a church-related one. Since 1975,

when the county's high schools were consolidated, the only organized

youth activities in White Pine are those that are church-sponsored.

In short, then, the influence of the organized religion in White

Pine is pervasive. The churches are viewed as having an indispensable

and rightful place in the life of the community and in the life of the

individual.

The Southern Appalachian Region
Alghough the Appalachian Mountains extend thirteen hundred miles

from Vermont to Alabama, the region usually called "Appalachia" and









viewed as sharing certain distinctive cultural characteristics and

social and economic conditions encompasses the southern half of the

mountain range and includes as its core a sliver of North Georgia, the

eastern thirds of Tennessee and Kentucky, the western thirds of

Virginia and North Carolina, and the entire state of West Virginia.

This is in addition to a variety of fringe sections, depending upon

which demarcation of the region one chooses. Perhaps the most exten-

sive demarcation is the political one made by the Appalachian Regional

Commission. The Commission defines Appalachia by county, based on

economic conditions. Their "Appalachia" includes parts of Alabama,

Mississippi, South Carolina, Ohio, Maryland, Pennsylvania, and New

York, in addition to the core area mentioned above.

Gastil, whose regions are determined by sociocultural features,

considers the core area outlined above and a small corner of South

Carolina to be the "Mountain" area of the South (Gastil 1975:174ff).

Vance (1962) views the "Southern Appalachians" as a distinctive region

sharing agricultural and economic conditions, which covers parts of

West Virginia, Kentucky, Tennessee, North Carolina, South Carolina,

Georgia, and Alabama. The demarcation of the "Southern Highland" re-

gion by John C. Campbell (see map on the following page), a lifelong

student of the region, comprises the core area of portions of six states

and also the fringes of three others, a 112,000 square mile territory.

For Campbell, this area exhibits similar economic, educational, and re4-

igious characteristics. What is important for us is that, whatever the

delimitation of Southern Appalachia, the community of White Pine has always



















































Figure 1.2 Campbell's Southern Highland Region
(Source: Campbell 1921:xxxviii)









been considered in its heart. Further reason for designating White

Pine as an "Appalachian" community is presented below.

The linguistic analogue of Southern Appalachia is the South Mid-

land area (Kurath 1949, Kurath and McDavid 1961), which encompasses

the mountain districts and also most of Tennessee and considerable por-

tions of Mississippi, Alabama, and Georgia. This area, determined by a

combination of lexical, phonological, and morphological isoglosses, has

been traditionally viewed as more closely related linguistically to the

North Midland area (Pennsylvania and parts of Maryland, Delaware, and

New Jersey) than to the South.

The question of the relative affinity of linguistic and other

traits in Southern Appalachia (or the South Midland) to regions north

and south is still very much unanswered, in fact, but it will not overly

concern us here. In a provocative reassessment of the linguistic dis-

tinctiveness of the area, Bailey (1968) concludes there is no "Midland"

area at all which has a core of distinctive features and that North Mid-

land should be retermed 'Lower Northern' while South Midland be redesig-

nated 'Outer Southern.' Gastil views the cultural affinities as

lying closer to the South: "Culturally the mountains are an extreme

version of the Upland [South]" (1975:194).

All this is not to say, of course, that even the most restricted
definition of Southern Appalachia implies that it is socially homogene-

ous or that its people have a common background because they now may

share certain cultural or linguistic features or even a common outlook

on life. Yet commentators on the area, both native and non-native,










infer that its inhabitants are alike in ancestry and cultural back-

ground, especially on the basis of a number of linguistic relics

apparently distinctive of its people. The most extreme forms of such

inferences issue an assertions that Appalachian people are "Elizabe-

than" or that they preserve the "purest Anglo-Saxon blood in the United

States" or that they are "almost pure Scotch-Irish" (note the contra-

diction here). These inferences have formed a part of the chronic

misconceptions about the inhabitants of Southern Appalachia.

Even such a serious commentator as Campbell has abetted such

inferences by his statement that Southern Appalachian people form "a

definite racial group" (1921:xiv), although he conceived of

their having developed into such a distinct group of people by virtue

of "their common interests, hardships, and struggles" (p. 71) and not

by virtue of their lineal descent from one stock of Old World people.

His lengthy study of the surnames of early mountain settlers revealed

almost equal proportions of Scotch-Irish and English names, and a

large number of Germans as well:

Without a doubt the Scotch-Irish and English elements are the
strongest in the mountain population, though the Highland
people are not different from the Lowland Southerners in this
respect. The Scotch-Irish strain is strongest in some mountain
sections, the English in others; and in some communities may be
surmised an influence of German ancestry. (p. 71)

Several other groups, most especially the French Huguenots, were numer-

ous among the early settlers in Appalachia too. Thus the region's

present-day descendants are a hybrid, whencever came their language (a







handful of its features have indeed been shown to be carryovers from

Northern England and Scotland) or its musical or other traditions.

For a variety of unscientific and often impressionistic reasons,

however, the people of Southern Appalachia are reputed to be quite dif-

ferent from all other Americans. If they do not descend "directly from

the Scotch-Irish," there is a quaintness in their approach to life, it

is said, harking back to an earlier era, or their speech has a quality

and character indisputably unique. Not only does Appalachian speech

preserve various relics, that is, but it also has a distinctive quality

in a more general sense. The literature on the subject abounds in de-

scriptions of how mountain folk talk, more often than not written by

college-educated children remembering fondly earlier days in the hills.

Here are three representative descriptions:


There are certain peculiarities of enunciation which it might be
well to speak of here. For the Smoky Mountaineer the nose is as
much an organ of speech as the larynx. This is particularly no-
ticeable in old people, and may be because of the catarrh
brought on by constant exposure and diseased tonsils. The moun-
taineer drawls to an alarming extent, even more so than his
neighbors of the plantations. His voice is utterly without ca-
dence, almost a droning monotone. On the other hand, it has a
deep, resonant quality which catarrh and embarrassment never
wholly obliterate. (Walker 1939:3-4)

The mountaineer has clung to the Shakespearean word these many
years because he has not learned the modern word. Indeed, he
has had little contact with those who speak the modern word; and
since his surroundings and habits and thoughts have been largely
the same as they were in Shakespeare's day, he has had no need
for new or changed speech. (Coleman 1936:30)

As a race, the people of our southern mountains speak softly.
Even their most commonplace remarks somehow are made to sound
secret, and of greatest importance. The women especially speak
softly, their voices often plaintively sweet, pitched in a
minor key and lifting upward, so that the last word of every
sentence is almost sung. (Hannum 1969:29)









The contention in these excerpts is that the speech of Southern Appa-

lachian people has a marked flavor and character all its own, beyond

the presence of archaic expressions and words therein. Commentators

have long been at pains to account for this on the basis of one extra-

linguistic factor or another (isolation, pace of life, etc.). Cratis

Williams, lifelong resident of the mountains, attributes in part the

distinctiveness of the way people in the region speak to "the highland-

er's habit of thrusting his chin forward rigidly when he speaks."

(1961:10) Brewer and Brandes try to explain why Southern Appalachian

English impresses outsiders as so different by describing the kinesics

of its speakers, detailing the posture and the eye and body contact of

the typical speaker in the region.

Almost all the studies and comments on the speech of Southern Ap-

palachia have had the unfortunate effect, however, of lumping the en-

tire region together linguistically and assuming that it is homogeneous

within itself as well as distinct from other regions of the country.

To this date, very little attention has been devoted to regional lin-

guistic variation within Appalachia. Only the two linguistic atlas

projects, the Linguistic Atlas of the Middle and South Atlantic States

and the Linguistic Atlas of the Gulf States, have given significant

consideration to such variation. This presumption of underlying uni-

formity has been espoused frequently, as in Carpenter (1933). He con-

tends that the only "variation in the vernacular of the southern and

central Appalachians has been due to the difference in the degree of

isolation of the various sections" (p. 22). Most writers ignore the









subject of geographical variation entirely. This creates the impres-

sion that there is none. No attention was given until very recently

to variation by age, socioeconomic background, or any other variable.

This presumption of little, if any, variation is directly at odds

with the view of Southern Appalachian speakers themselves. They be-

lieve that Southern Appalachia has much subregional variation. Speak-

ers in the region as a rule (and White Pine exemplifies this well) con-

tend that they speak differently from neighboring communities only a

few miles down the road or one or two hills over, although it is diffi-

cult to tell what differences they have in mind. These subjective

views have occasionally been supported by statements in the linguistic

literature, as by Berrey: "the dialect may vary slightly with the lo-

cality, and even from family to family" (1940:46) Or by Walker: "It

must be realized that the speech of the mountaineer varies from one

district to another" (1939:6). Wolfram and Christian, in commenting

on vowel variation in southern West Virginia, state that "there are

apparent vowel differences from region to region within Appalachia.

Even within the restricted locale we have studied here, there is evi-

dence that several different vowel systems must be recognized"

(1976:64).

To what extent the term "Appalachian English" refers to a deter-
minable linguistic entity is then unsettled. On the one hand, some

commentators and Appalachian people recognize an almost hollow-to-hol-

low and community-to-community variation. On the other, such long-

time students of American English as Mencken (1963) and Wise (1957)









consider the dialect of the Southern Appalachian area distinct and

identifiable. One reasonable approach to the issue of whether the

term "Appalachian English" should be used is that of Wolfram, who

adopts the term and applies it to a "unique combination of linguistic

characteristics rather than structures as such" (Wolfram 1977:

97). For him, Appalachian English refers to a common denominator of

non-unique features which are still subject to variation for any given

district within Appalachia or for subgroups within such a district.

Since this study is designated as one on "Appalachian" speech,

this must now be justified and explained. This study does not inves-

tigate linguistic features dealt with systematically in earlier studies

on the speech of the Southern Appalachian region or of the speech of

any other variety of American English. It thus lacks the comparability

of evidence required to distinguish the speech of White Pine as "Appa-

lachian" solely on the basis of linguistic criteria, although nearly

all the features-phonological, lexical, and grammatical-usually

termed "Appalachian," as by Brewer and Brandes (1977), are found in

the speech of White Pine inhabitants. Nor does it adopt the design-

ation solely on the basis of White Pine's geographical location. If

it did, "East Tennessee" would have been used. Rather, the speech of

the inhabitants of White Pine is called "Appalachian" because the

community is, ethnographically speaking, an Appalachian community, on

the basis of several of those salient characteristics which have long

been associated with Southern Appalachia-individualism and self-

reliance, traditionalism, fatalism, and fundamentalist religion

(Ford 1962:llff.).









Throughout the literature on Southern Appalachia runs the view

that its people and culture are distinctive in at least a broad sense.

Most commentators on the region point out that a combination of such

traits as "fierce" individualism and mistrustfulness of all strangers,

among others, characterize the Southern Appalachian native and make him/

herdifferent from other Americans. White Pine is more of a valley than

a mountain community, and such characterizations have been more accu-

rate of the isolated communities in the region than of the urban and

valley sections of Southern Appalachia. This has sometimes been recog-

nized, as by John C. Campbell, who lived and taught in the region for a

quarter century. Campbell was concerned that some of his statements in

his The Southern Highlander and His Homeland would not be understood as

"applicable to the remote rural folk who were the particular object of

his study [and] were not true of their urban and valley kinsfolk"

(1921:xiv). But such characterizations, while sometimes extreme,

rarely claim that Southern Appalachia is culturally or socially homoge-

neous. White Pine itself is certainly neither. Yet several broad cul-

tural traits have for so long been pointed out as typical of Southern

Appalachian people that it is worthwhile to determine if they are ap-

plicable to White Pine and can enable us better to understand it as a

Southern Appalachian community. We will ocnsider the four traits or

themes mentioned above that Ford chose to examine by conducting a

region-wide survey of Southern Appalachia. Ford found these traits to

be significantly apposite to both urban and metropolitan communities,

although less so than to rural communities.









Ford surveyed nearly 1,500 homes in a region closely coterminous

with Campbell's Southern Highland region (page 22 above) in three dif-

ferent kinds of community-rural, urban, and metropolitan-with a

questionnaire devised to indicate the strength of four prominent

traits-individualism and self-reliance, traditionalism, fatalism, and

fundamentalist religion. It was hypothesized that, because of the eco-

nomic and social changes which had taken place after the Second World

War, a disparity between rural and non-rural manifestation of these

traits would be found. Although he did find some disparity, his results

show that the four traits he chose form a common denominator for Southern

Appalachia in general. Ford outlines his choice of traits in this man-

ner:


In examining the web of mountain life, one finds these themes
intertwined and generally, though not always, mutually support-
ing. Most so-called "mountain traits" are to be found in one
form or another throughout the nation, particularly in rural
areas. At the same time, each of them has its antithesis in
contemporary industrial society. The self-reliant individual-
ist, at least as an "ideal type," stands at the far end of the
scale from the much berated "organization man." Traditionalism,
not only in the sense of clinging to an earlier heritage but
also in the exaltation of resistance to social change, is viewed
as both anachronistic and vaguely immoral by a larger society
that values progress through rational, scientific endeavor.
Even more reprehensible to a culture that stresses achievement,
self-betterment, and mastery over nature is a passive resigna-
tion to one's situation in life, particularly if it is a situa-
tion viewed as both undesirable and remediable. Less subject to
censure by the larger society, perhaps, but contrasting as
sharply with its dominant values-and not immune from ridicule-
is the rigid, pervasive religious ethos of the [Southern Appala-
chian] Region. (p. 11)

Commentators on Southern Appalachia have long characterized its

people as individualistic, self-reliant, and non-social. Campbell









devotes an entire chapter (1921:90-122) to analyzing their individualism

and views their self-reliance as a result of their extreme isolation.

Horace Kephart, who lived in western North Carolina the first two dec-

ades of this century, characterizes the people as having "fiery indi-

vidualism" and as being "non-social" (1922:382-3) and goes on to say

Except as kinsmen or partisans they cannot pull together. Speak
to them of community of interests, try to show them the advan-
tages of cooperation, and you might as well be proffering advice
to the North Star. (p. 383)
More contemporary observers, Weller (1965:29ff.) in West Virginia

and Hicks (1976:39) in western North Carolina, document the same trait

among the sections of the region they observed.

When one looks at White Pine, such comments have a remarkable

applicability. The people are basically non-social, sticking to small

groups, not often unified in acting except in opposition to a proposed

local change, on a special occasion such as Pine Festival Day, or with-

in the small sphere of the kin group. Regarding their inability to

work together, a twenty-nine-year-old male resident of White Pine says

of his fellow citizens

They, they want to know, but they don't want to come out of their
shell long enough to get enough of it to expand. They fear ex-
pansion. This town definitely doesn't, it, one day they'll be
shocked into the fact that Morristown will try to annex them,
and there will be a total bar-, that will be the common bond
that brings them together. They will be, that will be the sup-
portive element would be to stand against some cause.
This does not contradict the statement above (p. 9) about the cohesive-

ness of the community; it means the cohesiveness is passive and residual.









The self-reliance of the region's people has another side to it, as

Ford (p. 13) points out, in their attitude of acceptance toward federal

welfare and assistance programs. Odum remarked over forty years ago that

the people of Southern Appalachia, "reputed to be the most individualis-

tic of all the regions, they cooperate most fully with flew Deal tech-

niques' (1936:97). Notwithstanding scattered initial opposition (primar-

ilyby those whose lands would be claimed by the reservoirs), the people

in the White Pine area and throughout East Tennessee have accepted and

cooperated with TVA's many programs. There was little reluctance to

accept whatever federal assistance was proffered.

The traditionalism and fatalism characterizing Appalachian people
areattributed to their struggle of many generations with the harsh reali-

ties of life in a region with few resources and little available produc-

tive land, according to Ford (1962:16). Weller (1965:33-40) shows how

such themes flow throughout mountain songs, stories, and daily attitudes.

In White Pine it is manifest in the strong and widespread doubts about

the value of change, of progress, and of post-high school education (see

quotes above, pp. 12-3). The twenty-nine-year-old male quoted above

says, regarding the community's residents,


They're...blinded really to what's, what's going on, I think.
They don't see their youth. They don't see, they don't see life
as progressing. They just see it as, as a tolerable state to be
in, and you just tolerate it now and hope for the future, yet you
don't work for it. You don't do anything working hand in hand.

As indicated above (pp. 8-9), recent changes and pressures for more

changes have made life more complex and confusing for many in White Pine.









The traditional way and pace of life in the community have been con-

sidered as something of a birthright. Many other things, such as po-

litical and religious affiliations, are also considered immutable

birthrights. Many wish keenly to pass all this on to the next genera-

tion.

Little additional commentary here (see pp. 17-20 above) is needed

to show how applicable is the fourth characteristic trait Ford studied,

fundamentalist religion, to White Pine. The community has no non-fun-

damentalist churches and the great majority of its citizens are church

members. Religious fundamentalism is probably the most powerful and

pervasive force in the community.

With reference, then, to the four characteristic traits Ford claims

to be consistently typical of Southern Appalachian people, it is obvious

that White Pine should be considered a typical community in the region.



The Study of Appalachian Speech
As Wolfram notes, Mencken once pointed out that the speech of the

Southern mountains had been studied more than that of any other region

except New England (Wolfram 1977:92). However, the sum total of lit-

erature on Appalachian English does not actually tell us much about how

people in the region speak. At best, most of the literature is little

more than anecdotal, dealing with a scattering of exotic words and pro-

nunciations, many of which indeed are archaisms. Very few writers have

actually asserted that Shakespearean English is or was spoken in the

mountains. But many have taken extreme pains to indicate usages in Ap-










Appalachian speech corresponding to usages in Spenser, Chaucer, and

especially Shakespeare. This creates the impression that the language

of the mountains is little more than a carryover from a bygone era and

another country. This is especially true when a commentator devotes

little attention to anything but such analogues.

One might go so far as to speak of a "typical" treatment of Appala-

chain speech in the literature. This is represented by Berrey (1940),

Bray (1950), Carpenter (1933), Combs (1916, 1931), Watkins (1949),

West (1966), and many others. The typical treatment contains the

following:

1) A brief description of mountain culture and of the mountaineer

himself, especially mentioning the most salient traits of his

personality and how he has adopted himself to his mountain

environment.

2) Discussion of various peculiarities of pronunciation, syntax,

vocabulary, usage, onomastics, and similes and other figures

of speech and expressions, with an effort to compare such with

Old, Middle and Elizabethan English whenever possible.

3) (Optional) A few remarks on how mountain speech is/has been

passing away, despite its expressiveness and color.

Coleman (1936) is a perfect example of this typical treatment. Her
first chapter points out the ancestry of the mountain people, shows how

they developed self-reliance in their isolated habitat, emphasizes the

fierce independence of their character, and shows the low position of

women in mountain society. In her second chapter, she covers a wide










selection of grammatical, lexical, and pronunciation features in

mountain speech from North Georgia, with accompanying citations from

sixteenth-century British literature wherever possible. In chapter

three she bewails the inevitable passing of the mountaineer with the

coming of modern education and technology, which will "wipe away for-

ever our admirable mountaineer with his quaint and delightful manner

and speech" (p. 30).

Even well-known early scholars of the speech of the region are

given to such vague statements as "in general, most vowels may be used

interchangeably. Most vowels may replace 'e,' 'i,' and 'u'" (Combs

1931:1315).

But the most fundamental deficiency of almost all studies of

Appalachian speech to date is that they provide absolutely no context

for their data. It is impossible to understand a language fully unless

we have information regarding five different kinds of context for lin-

guistic data:

1) Which part of the speech community uses the formss? (Old or

yound, middle or lower class, rural or urban speakers, etc.)

2) How often is a form used? (Is it categorical or extremely

rare or the predominant usage?)

3) In what kind of social situation is the form used? (i.e., What

style of speech is involved?)

4) What communicative task is a speaker performing when a form

is used?







5) What is the linguistic context of the form? (Where in the sen-

tence and in the discourse does it appear?)

Only when we have answers to all these questions can we establish what

context a form appears in and whether it is typical verbal behavior.

The history of the study of Applachian English has been chiefly a

history of the study of extreme forms, and only rarely is any context

noted for data which are cited. Despite an occasional remark that a

certain form "is still heard" or is "generally" or "frequently" used,

the clear implication of most studies is that all the speakers in the

community under consideration use the forms cited all the time and in

every social situation.

The tendency to deal with the atypical (and hence unrepresenta-

tive) speakers and forms infects even works which are otherwise linguis-

tically sound. This temptation is very strong. Joseph S. Hall's tho-

rough and invaluable study of the phonetics of the speech of the few re-

maining residents of the Great Smoky Mountain National Park was under-

taken, he confesses, because he believes in this speech, which "has so

long been removed from the main currents of American culture, there would

remain vestiges of earlier stages of the growth of the English language"

and which also would not have "the deep impress of the schoolmaster's in-

fluence" (1942:4). Brewer and Brandes (1977) present only the extreme

features of Appalachian English for the benefit of schoolteachers who

will teach in Appalachia.

Two recent sociolinguistic studies, Hackenberg (1972) and Wolfram
and Christian (1975, 1976), take into account the first two contexts

(who uses the form and how often is it used?), but neither controls the





37

other three in their investigations. Wolfram and Christian's study,

based on interviews made in the mountains of southern West Virginia,

classified their speakers socioeconomically and carried out quantita-

tive analyses of many of the linguistic forms they discuss, thus pro-

viding the speaker and frequency contexts for their data. Yet they

do'not choose to base their work on a representative sample because

they were "primarily concerned with the language variety which might

be considered most divergent from some of the more mainstream varieties

of English" (1976:10). Their work is also written primarily for the *

educational community and one-fifth of their book deals with the educa-

tional implications of dialect diversity.

Hackenberg divides thirty nine residents of Nicholas County, West

Virginia, into four classes on the basis of education level attained and

occupation and quantifies the occurrence of '!non-standard" verb agree-

ment and subject relative pronoun deletion for his four classes. He

thus takes the first two contexts into consideration and even gives some

attention to the fifth context mentioned above, the linguistic context,

for his two linguistic features. The present writer demonstrates else-

where (Montgomery 1977b) that his treatment of the linguistic context is

too simplistic.

A description of speech is ethnographically justifiable only if all

five contexts are controlled and taken into account. Only then can it

be claimed that the behavior described is typical (see the discussion of

the Malinowski/Firth "context of situation" later in this chapter).

This present study has striven to control all five. Thus it makes a

unique contribution to the study of Appalachian English.









This study is based on a representative sample of forty infor-

mants in a Southern Appalachian community of approximately 3,000 per-

sons. It quantifies the linguistic features it examines both in terms

of which speaker used them and in terms of possible occurrence of the

feature (whenever possible). It controls the social situation in which

its data occur (see the discussion of the ethnographic parameters and

the role this fieldworker adopted for himself later in this chapter).

This study controls the task its informants were performing by elicit-

ing only expository (explanatory and descriptive) discourses, and it

controls the linguistic context of its data by considering the sentence

and especially the discourse contexts of data.


Fieldwork Procedures

The fieldwork upon which this dissertation is based was carried

out over a period of about two months, from December 1977 through Feb-

ruary 1978. The fieldworker/researcher spent most of his time in the

community of White Pine during this period. This section details the

process whereby he set up, undertook, and completed his fieldwork. It

also provides the rationale for the type of interview and the ethno-

graphic parameters adopted. Altogether six weeks were devoted to the

actual interviewing, one week (in the field) to preliminary fieldwork,

and a period of several weeks to preparation before he reached the

fieldwork site.

Two steps were taken before the researcher reached the White Pine

site. First, an ethnography of the White Pine and surrounding areas was









prepared. As thorough and as broad an ethnography as possible is in-

dispensable to successful and accurate anthropological and linguistic

fieldwork. It is necessary to explain this point in regard to linguis-

tic research in general and to this project in particular.

Speech is social behavior par excellence. As Hymes indicates, it

is the "nexus between language and social life" (1971:42). We cannot

accurately describe language Cor, as Malinowski would say, meaning)

without reference to its role in society and the setting, participants,

events, acts, norms, etc., which it involves. Speech must be viewed as

permeating and not just overlaying most social behavior; therefore, de-

scription of such behavior must make reference to language and its

forms.

No aspect of a culture can accurately be described without explica-

ting its interrelationships with and dependencies on other aspects of

the culture. This has been more or less axiomatic since Malinowski's

ethnographic work among the Trobriand Islanders more than sixty years

ago. Malinowski insisted that a researcher approach a way of life as

a whole, as full of interdependencies, as an integration of life. Only

the application of the most rigorous techniques and standards for the

collection of data of all kinds-material, linguistic, topographical,

personal, ceremonial, among others-could enable the adequate explica-

tion of the "context of situation," to use the term Malinowski (1923:306)

himself coined. All pertinent data must inform an understanding of a

context of situation before a generalization can be made, in other

words. For Malinowski, a context of situation is at the same time









broader and immensely more useful than the concept of a "linguistic

context": "Even as in the reality of spoken or written languages, a

word without 'linguistic context' is a mere figment and stands for

nothing by itself, so in the reality of a spoken living tongue, the

utterance has no meaning except in the 'context of situation'"

(1923:307).

In addition to the many kinds of data Malinowski considered useful,

healsoadvocated that a researcher keep two kinds of personal records

which should constitute parts of his field records (Kaberry 1957:78-9):

1) An ethnographic diary of daily activities and observations not-

ing what is the normal and the typical and what is not, the

kind of information that is not elicitable by questions and an-

swers of a community's residents. Malinowski called such infor-

mation the "imponderabilia of actual life." The present re-

searcher kept such a diary while in the field, a diary which

also recorded the accounts of his successes and failures and

mistakes in the field.

2) A "corpus inscriptionum," which "entailed the recording in the

native language of narratives, opinions, typical utterances,

myths, folk-lore, magical formulae, and native explanations and

interpretations of customs and beliefs" (Kaberry, p. 79).

Since this researcher collected only one type of discourse (ex-

pository, as defined in chapter two), his forty five hours of

taped explanations and descriptions constitute his "corpus in-

scriptionum."









Before returning to the explanation of how an ethnographic sum-

mary of the White Pine community was prepared, let us look more

closely at Malinowski's view of context and explore an instance from

White Pine where we must know the "context of situation" to fully un-

derstand certain linguistic behavior.

As a result of his fieldwork experience, Malinowski formulated the

two concepts "context of culture" and "context of situation," the lat-

ter mentioned above. Both have to do with how meaning is created by

language. A helpful redefinition of these concepts for linguistics has

been provided by J. R. Firth (1957b), one of Malinowski's students, to

express more satisfactorily how language means. According to Firth,


The context of culture is the environment for the total set of
[the options in behavior that are available to the individual
in his existence as social man], while the context of situation
is the environment of any selection that is made from within
them. (Halliday 1974:49)

For Firth, then, the two types of context formulated by Malinowski pre-

sent the distinction between what is potential and what is actual.

Again according to Halliday, "the context of culture defines the poten-

tial, the range of possibilities that are open. The actual choice among

these possibilities takes place within a given context of situation"

(1974:49) Firth's version consequently emphasizes the value of context

of situation for linguistics while at the same time implying that con-

text of culture has little value for the linguist.

The great usefulness of the concept of context of situation, in
Firth's definition, for linguists lies in its stress on the typical









rather than the accidental as an appropriate object of study. Firth's

view also makes the important implication that there is no necessary

one-to-one relationship between an utterance and its meaning. This

implication is most valuable when we examine typical behavior.

The typical behavior (an utterance, in this case) from the White

Pine corpus to be explored briefly here is the expression "I don't

know." At times in the interviews this expression became almost annoy-

ingly frequent. Consider an excerpt from an interview with a thirty-

nine-year-old housewife:

Fieldworker: Why should people be proud of White Pine? What
would you say?
Informant: Well, as a community, I think it's uh, it's a good
community to live in, to raise your children in. Well, I
don't know. The churches, the schools, everything. I think
it's just a, I don't know. I like it. I think it's, that's
why I think it's a good community is because I like it,
(uh-huh, uh-huh) and I don't know.

Of course, it may be that the informant does not know how to artic-

ulate her feelings here, or perhaps she in fact does not know why she

should be proud of White Pine, even though she had just stated, before

the fieldworker's questions, that the residents of the community should

be proud of White Pine. But this can hardly account for there being so

many occurrences in this excerpt of so deceptively short an expression

(three times in about twenty seconds). There are at least three differ-

ent "I don't know's" to be found in the corpus, two of which are exem-

plified here. Although this cannot be as clearly seen from the written

transcript, the researcher clearly identified at least three distinct

meanings for this expression:









1) A hesitation "I don't know" which frequently occurs, as in

our excerpt, with "well." This expression buys a little time

while the speaker decides exactly how to phrase an answer.

2) An ignorance "I don't know," exemplified by the other two in-

stances, which indicates the speaker's recognition of his/her

inability to answer or express him/herself.

3) A reluctance "I don't know" which is most often used as an ini-

tial reply to a question in order to show an informant's imme-

diate reluctance to reveal his/her thoughts on a subject. This

differs from the first "I don't know" above. This "I don't

know" is not a refusal, however; informants almost invariably

went on to give a direct answer after showing initial reluc-

tance.


The preparation of the ethnography of White Pine presented above

included the major sources on the Southern Appalachian culture and re-

gion (Kephart 1922, Campbell 1921, Weller 1965, Ford 1962, Vance 1935,

Hicks 1976, and others). It also included major sources on the East

Tennessee area (Folmsbee et al. 1969, Dykeman 1977, and others). Spe-

cial attention was given to the few sources on Jefferson County, Ten-

nessee, available in local East Tennessee libraries. From the East

Tennessee Development District office in Knoxville were obtained sum-

maries of the 1970 census reports and other vital statistics on the

White Pine community. In addition, much useful information was gath-

ered from the informants regarding local affairs.

The second step, taken after the preparation of as much of an eth-
nography as was possible from written sources, was the researcher's












making arrangements with the contact persons in White Pine to spend

an indefinite period of time with them and his explaining what kind

of investigation he would undertake and who he would like to interview.1

A contact person is an invaluable bridge between a fieldworker and a

community. In it a fieldworker is initially a total stranger. He can-

not reside long enough to become a familiar member of the community

(in White Pine or another typical small town in East Tennessee, this

would probably take several years in any case). This researcher's con-

tact persons introduced him to many of his informants and gave him many

insights into the life of the local community.

In White Pine itself, four problems had to be resolved before

actual fieldwork (i.e. interviewing) could begin. First, the field-

worker had to be thoroughly familiar with the layout of the town and

the most important basic facts of daily life in the community. In this

he was primarily assisted by his contact persons. This included learn-

ing the principal activities of the townspeople and the principal social

institutions in the community. He also had to determine the principal

concerns of life for the community. He was able to do this in the

course of the interviews.

Second, he had to establish the ethnographic parameters of his

work. This included his role in the community during his stay and in

the interview situation. Third, he had to choose a location for the

interviews, specifically, where they could be conducted comfortably.





45
Fourth, he had to determine the type of discourse to elicit and how

to devise a standard interview format which would best elicit it.

The ethnographic parameters were adopted according to Malinowski's

suggestions. He states that a fieldworker should make every attempt to

be an active participant rather than a passive observer-partial partic-

ipant: in the culture he is studying. He draws on his own experience

among the Trobrianders in showing why this is necessary:


There is all the difference between a sporadic plunging into the
company of natives, and being really in contact with them. What
does this latter mean? On the Ethnographer's side, it means
that his life in the village, which at first is a strange, some-
times intensely interesting adventure, soon adopts quite a nat-
ural course very much in harmony with his surroundings. Soon
after I had established myself in Omarakana, I began to take
part, in a way, in the village life, to look forward to the im-
portant or restive events, to take personal interest in the
gossip and the developments of the small village occurrences; to
wake up every morning to a day, presenting itself to me more or
less as it does to the native....As I went on my morning walk
through the village, I could see intimate details of family
life;...I could see the arrangements for the day's work, people
starting on their errands, or groups of men and women busy at
some manufacturing tasks. Quarrels, jokes, family scenes,
events usually trivial, sometimes dramatic, but always signifi-
cant, formed the atmosphere of my daily life, as well as of
theirs. (1922:7)

This researcher, of course, is neither a native nor a resident of

the White Pine area. He was a stranger to his informants initially.

At the same time, he is a native of East Tennessee and especially that

he is a relative of long-time White Pine residents enabled him to estab-

lish immediate rapport with his informants. It was also recognized

that the interview situation would necessarily be an artificial one.

Goldstein defines an "artificial situation" as one in which an informant

provides information "at the instigation of the collector" (1964:82).









The artificial interview situation, however, is, he adds, "the

only natural context for eliciting information from an informant"

(pp. 104-5). In other words, some type of directive elicitation is nec-
essary, however willing the informants are to participate, since they

do not always spontaneously volunteer discourse, especially exposi-

tions. Given these realities, the researcher decided to adopt the

role of a college student collecting general information for a.research

paper from a cross-section of the people in the town, stating that he

was interested in White Pine because he viewed it as a typical East

Tennessee small town. This is similar to the role of historian Gold-

stein advises for folklorists. In this role the researcher states that

he is not interested in official opinions or just in the opinions of

the oldest and/or most respected members of the community, but that

he is seeking instead the ideas and views of a cross-section of the com-

munity's average citizens. He emphasizes that he wants responses from

both native and non-native residents, from old and young, from members

of the whole community, whatever their social standing.

The researcher decided to have the interviews in the informant's

home rather than in the contact persons' home. He also discarded the

idea of interviewing in the informant's place of business or at some

other neutral, public site. Most informants readily invited the re-

searcher into their own homes, and for this reason and others it ap-

peared that informants would be most comfortable there. Moreover, a

fieldworker actually has more control over the interview in the infor-

mant's home. There he has greater discretion over the pacing of the









interview and over signaling the end of the interview and his impending

departure. It is easier to tell an informant that the interview is

over, to pack up the equipment, and to say good-bye than it is to tell

an informant that he/she can go home now. It is also less of an impo-

sition if the fieldworker takes care of his own transportation instead

of asking the informant to arrange his/her own transportation. Fi-

nally, informants appeared more likely to venture information in the

securest andmost familiar environment for them, their own homes.

The fourth thing that the researcher/fieldworker had to determine

before beginning full-scale interviewing was the type(s) of discourse

he would elicit and the interview format to elicit it (them). His ap-

proach was based on the typology of discourse genres (narrative, expo-

sitory, procedural, and hortatory) in Longacre (1976). Given the time

limitations and the ease of devising questions, the researcher decided

to concentrate on expository discourse. A standard set of questions

was asked of all informants, regardless of age, socioeconomic back-

ground, or level of educational attainment. No difficulties were ex-

perienced in elicitation according to this plan.

These limitations also had several advantages. First, using stan-

dard questions gives the study strict controls. Then, expository dis-

course seems to be the most plausible to elicit, given the ethnographic

parameters of this study. Most of the standard questions had to do

with everyday life in the town. Others dealt with situations that were

of widespread and frequent comment about town (such as the reputation

of Jefferson County's neighbor, Cocke County, for violence and










lawlessness). These are safe topics, unlikely to give offense, yet of

universal interest.

Before the interview itself and the sample are discussed, a few

more comments need to be made about how the fieldworker established

rapport with his informants. As Goldstein points out, the fieldworker

should take every opportunity to participate in the daily lives of the

people and the community he is studying. This researcher visited local

churches and various gathering places and meetings, chatting with peo-

ple frequently on the streets and at lunch counters. He continually

emphasized that he was interested in interviewing a cross-section of

townspeople.

The interview itself is a directive one, but it is loose and flex-

ible. It was directive insofar as it required specific vital informa-

tion from informants (age, length of residence in White Pine and else-

where, level of education attained, occupation held, and family back-

ground). The fixed questions asked of all forty informants require

expository responses:

1) How does White Pine react to newcomers in town?

2) What have been the most important changes you have seen in

White Pine while you have lived here?

3) What's a typical day like for you?

4) What do you think most people in White Pine want the town to

become?

5) What all goes on on Pine Day?

6) What do you do on Pine Day?









7) Why does Cocke County have a bad reputation?

8) How do local people feel about outsiders making fun of East

Tennesseans and the way they talk?

In this way, some uniformity in interviews was established.

However, the interviews were not rigid. For example, no infor-

mant was asked only these eight questions. Each informant was asked

additional questions depending on age or background. For instance, in-

formants over fifty who were native of the area were asked how local

people felt when TVA came into the county in the 1930's and bought up

prize farmland to be covered by reservoirs. Similarly, high school age

informants were asked how the new county-wide comprehensive high school

compared to the old high school in White Pine. Other questions were

asked, depending on the flow of the conversation.

One non-directive principle was adhered to throughout theinterview-

ing, however. This concerns the lack of a time limit. Once the field-

worker had asked a question on some subject, the informant was allowed

to give as long a response as desired. The fieldworker took pains nei-

ther to interrupt nor to influence the informant's pattern of thought.

Only when the informant signaled an end to a response or when an infor-

mant's response was clearly running out of steam did he intrude with a

comment or another question.

The interviews averaged slightly more than an hour in length. But

the fieldworker spent at least twice that amount of time with the aver-

age informant, exchanging small talk and establishing a rapport with in-

formants before the actual taping began. The time taken by the field-









worker to introduce himself and to tell about his family was indispen-

sable in establishing the uninhibited rapport that he desired. The in-

terviewer's family was of more concern to the informants than his mo-

tives for interviewing.

In his view, the fieldworker was able to establish quite good rap-

port with each of his informants and excellent rapport with nearly all

of them. This resulted in not only a good corpus but also a number of

good friendships with individuals in White Pine.



The Sample

According to the 1970 United States Census, the population of the

town of White Pine was 1,532. It had grown to an estimated 1,830 by

1977. But the community of White Pine as defined by its residents in-

cludes both the area outside the city limits and the adjacent neighbor-

hood of Leadvale, bringing the total population of the area in which the

interviews were conducted to approximately 3,000 persons in 1978. Of

this base population of 3,000 a sample of forty residents was inter-

viewed. This constitutes 1.3 percent of the whole. The interviews

average roughly one hour each, the shortest being thirty-five minutes

and the longest more than two hours.

The sample of forty informants is both a broad and a representative

cross-section of the White Pine community. The informants represented

all post-adolescent age groups and a wide variety of socioeconomic and

educational backgrounds. In age they range from sixteen to eighty-seven

as distributed among the following cells:










16-19 20-39 40-64 65 and over Total

Male 3 6 5 4 18

Female 3 5 8 6 22


As the sample was taken, care was exercised not to have too many in-

formants from a given age group or from either sex. The fieldworker's

overriding concern, as he spent his time in the White Pine community

and had contact with all sectors of the population, was to interview as

wide a variety of its residents as possible. It was not his primary

goal to investigate social or subgroup differences in discourse organi-

zation among his informants.

The breakdown of the forty informants according to their levels of

education is as follows:2


8th Grade Some High High School Some College College
or Less School Graduate Work Graduate

Female 1 7 5 5 4

Male 3 5 1 6 3

According to the 1970 census, only 6.3 percent of White Pine's popu-

lation depended on agriculture for its livelihood, while 74.1 percent

depended on manufacturing, trade, service, and finance industries.

Approximately 19 percent of the population was employed in construc-

tion and all other areas. Following is a breakdown of the White Pine

sample according to their employment by type of industry. Three who

were unemployed are ignored.









Agriculture Construc- Manufac- Trade, Service, Other
tion turning and Finance

# % #% % # % # %

1970 Census 89 6.3 129 9.1 720 51.0 327 23.1 142 10.0

Present
Study Infor- 3 8.1 3 8.1 6 16.2 21 56.8 4 10.8
mants


If females were not members of the labor force, they were categorized

according to the employment of the breadwinner of their families. The

present study's higher percentage of informants supported by trade, ser-

vice, and finance than the percentage of those depending on manufactur-

ing contrasts with the percentages for the overall White Pine area.

This is due to their availability. There were more people in trade,

service, and finance positions who were available for interviews than

in manufacturing.

Although no census figures are available on the lengths of resi-

dence of White Pine inhabitants, it is clear that the great majority

of them are lifelong natives of White Pine or the nearby area. Of the

forty informants, twenty-two were natives of White Pine and another

thirteen were natives of nearby counties in East Tennessee, nine of whom

had spent more than half their lives in White Pine. Four were natives

of elsewhere in Tennessee or of another state, and one was a native of

Great Britain. Informants were chosen on the bases of their ages and

educational backgrounds and not their nativities. Even so, the cohe-

sive devices analyzed in chapter three through five were examined for

differences in function between natives of White Pine and non-natives.

None but incidental differences in frequency were found.










In the White Pine community, only 1.7 percent of the population

was non-White in 1970; none of the forty informants for this study,

it happens, were non-White.


Transcription Procedures

The approximately forty-five hours of field tape recordings are

the primary documents for this study and are viewed as the ultimate

authority and repository for the data upon which this study is based.

All forty-five hours of interviews were transcribed by the fieldworker/

researcher himself. The transcription was done in slightly modified

orthography in order to result in as close to exact a form as possible

of what the informants said. As Labov (1969, 1972) has shown, careful

transcription is as important in considering grammatical features

as it is for phonological ones.

It must be realized, however, that any transcription can never

be more than an imperfect representation, and an interpretation as well,

of what is on tape. This is certainly true when a transcriber attempts

phonetic or a modified phonetic transcription, and it was true in the

present case when the transcriber had to determine the presence of cer-

tain deletions and especially when he tried to be consistent in punc-

tuating a transcript or in ascertaining sentence boundaries. A tran-

scriber quickly recognizes that deictics and especially conjunctions

mark the beginnings of sentences. But an informant's pausing and pacing






54
are also crucial to sentence, clause, and phrase boundaries. This

is why the tape recording itself remains the ultimate authority.

The transcription employed here makes no attempt at eye dialect.

Orthography was modified only to indicate several minor instances of

deletion such as

1) The voiced interdental in "them" and "that," if deleted, is re-

placed by an apostrophe (as in "with 'em" or "like 'at");

2) Clause initial "that's" and "it's," often reduced phonetically

to [s] or [s:], and therefore indistinguishable, are represented

frequently as "s";

3) Initial syllable deletion, as for "of course," represented as

"course," and "because," represented as "cause."


In no instances were such conventions used to indicate the deletion of

morphemes. Dealing with deletions in general was frequently difficult.

Basically any apparent morpheme that had a clear trace was represented

in full orthography, In fact, clear hesitation phenomena such as "uh"

and "hmm," the affirming "uh-huh" and the negation "huh-uh" and the like

were represented whenever audible. But the many standard English words

without such a trace (such as copulas, relative pronouns, and others)

were not represented.

Hesitation phenomena (as well as the parenthetical "you know," "I

mean," and other phrases adding nothing to the content of the interview)

are represented in this study's interview transcripts for the sake of

making the transcript as accurate as possible. It is also possible

that such phenomena might have some discourse significance, as markers

of sentences or of paragraphs (i.e. changes or returns to themes). In

any case, possibly significant data should not be discarded a prior.









The use of brackets and parentheses in the interview transcripts

follows the suggestions of Ives (1976). Parentheses serve to enclose

short interviewer remarks (rarely more than the occasional "uh-huh"

or "hmm" which insures an informant that the interviewer is intently

following the informant's account) which did not cut short or deflect

an informant's train of thought. Occasionally parentheses were also

employed to suggest an alternate transcription, when the speech on

the tape was not quite intelligible. Brackets enclose either inter-

viewer comments on the informant's behavior or short summaries of the

interviewer's speech. In the interview excerpt presented below, the

two different uses of parentheses are exemplified.

Fieldworker: In general, what have been the most significant
changes that you've seen in the community of White Pine it-
self during your lifetime?
Informant (74 year-old male): Oh, I guess this, what I remember
as farmland is, has been being and is being covered up with
houses and housing developments and such as that, (uh-huh)
and around the town. Most of those, I remember when, when
that was farmland, the ones out this side of town and well,
that's all around town. And that uh, and that and inflation.
(hmm) Course, we've gone from, as far as uh, as transporta-
tion, we've gone from horses and buggies and mules for to
farm with, and buggies to ride in and, and trains hauled the
produce, to now it's gone to trucks. The trains are, people
ride in cars and, and uh, and produce being hauled in trucks,
right off the farm from there on, from the time it's produced
til (to?) the, the time it's gets to the grocery store. Rail-
road used to use a good many people. That was pretty big bus-
iness. Of course, it hauled lots of people in those days.
That was quite a, quite an event when a train came to
town. .

As mentioned, all the interviews were entirely transcribed by the in-

terviewer himself. Inevitably there are sentences, phrases, and even









words (especially names) that only the interviewer can comprehend by

virtue of his having remembering the context of the interview.


Fieldwork Journal and Interview Catalog

In addition to the primary document, the tapes, and the secondary

document, the transcripts, two other document were assembled in this

project. A fieldwork journal of daily activities and observations was

maintained by the fieldworker. After a day's fieldwork was completed,

a record of the day was entered into the journal. As mentioned earlier

in this chapter, this emulated the ethnographic diary that Malinowski

kept while among the Trobriand Islanders. Not only does such a journal,

maintained daily, allow a fieldworker to profit from his own experi-

ences, his mistakes and his frustrations, but it also enables him to

make note of particularly successful and effective approaches and ques-

tions for his later use.

The other document assembled was the interview catalog, compiled

for the researcher's ease of reference. It contains four kinds of in-

formation:

1) The vital statistics on each informant (age, sex, level of edu-

cational attainment, etc.);

2) The physical details of the interview: when and where it was

conducted;

3) A summary of how the interview itself went, how good the rap-

port between fieldworker and informant was, and how helpful

the informant was;

4) A note on how the informant was contacted.









Explanation to Informants

The members of the White Pine community, when approached by the

fieldworker/researcher as potential informants, were almost invariably

most willing and even flattered to be asked to participate in an inter-

view. While a few of them were initially concerned that they were

being "singled out" to be interviewed for some special reason (the

fieldworker had to assure them that he wanted to interview typical cit-

izens), the fieldworker was actually turned down for an interview only

once (over the telephone, where, he is convinced, he was mistaken for a

salesman). Most informants were more interested in knowing why they

(and White Pine) were chosen than what the fieldworker actually wanted

to interview them for, and that the series of interviews was a part of

his schoolwork was sufficient explanation for most informants to be

quite willing to be interviewed. The usual explanation provided pro-

spective informants was as follows:


For a school assignment I would like to interview a cross-section
of people in White Pine and ask some general questions about
what life is like here. What I want to do is compare the an-
swers I get from a variety of people, both old and young, male
and female. I wondered if you would be willing, if you have
some time, to help me by answering a few questions.

The fieldworker willingly gave whatever additional information was de-

sired-that the interviews were part of his dissertation, that he was

working on his doctorate, or that his project was a linguistic one-

but this was rarely called for.

It should, of course, be pointed out again that the initial rap-
port that the fieldworker was able to establish with informants would









almost certainly not have been so successful without the assistance

of the contact persons in White Pine,

Informants for this study were assured that their names would

not be revealed and that they would remain anonymous to all but the

fieldworker.


Taping of the Interviews

The interviews for this study were recorded on a four-track, reel-

to-reel Voice of Music model 733 Tape o Matic tape recorder at 3 3/4

inches per second. The tapes used were Scotch No. 207, 1800 foot

tapes, which, at 3 3/4 inches per second, allowed up to ninety minutes

of recording per track. In all eleven reels of tape were required.

Because of logistical difficulties, two interviews had to be re-

corded on cassette tapes. These interviews are coded as tapes XII and

XIII in Appendix A.


The Value of Fieldwork

Since the ultimate value of doing linguistic fieldwork is to help

us better understand how people employ language, the fieldworker must

always remember that the interview is a very human event involving di-

rect contact between two or more people. As Goldstein declares, "the

problem of fieldwork is essentially human" (1964:22), and more than any-

thing else fieldwork is a venture in human relations. The fieldworker

cannot be detached from his informants, especially when subjects as

deep-felt as some of those covered in the present study's interviews









are involved. Fieldwork is extremely hard and tiring work, but its

rewards in terms of friendships gained justify the effort put in by

the fieldworker.

From this researcher's perspective, at least three skills are re-

quired of the linguistic fieldworker in such a study as this one:

1) The ability and willingness of a fieldworker to relate to his

informants as humans who are willing to share their personal, and some-

times intimate, experiences with him. The fieldworker must value his

informants as people and not as sources of data, which means that it

is probably impossible to do good fieldwork without genuinely enjoying

human contact.

2) A resourcefulness to adapt to new situations and people in ways

that the fieldworker could not have foreseen before going into the

field. The fieldworker must take care and even pains to approach peo-

ple as individuals without preconceptions and prejudices. He should ap-

proach his community as a microcosm having a great variety of unique in-

dividuals.

3) A disposition to persevere against frustrations and inconveni-

ences and ever-present delays. In short, he must be so committed to

doing his fieldwork that he develops a kind of salespitch for it.3 He

must be willing to solicit assistance from often-hesitant and sometimes-

skeptical people while at the same time endeavoring to establish a fa-

vorable rapport with them.

Though the most demanding work he had ever done, this writer found

it immensely rewarding. It gave him many fond memories and several

good friendships which have developed from his interviews.








Notes


1. The choices of the community to work in and of the contact persons
were entirely practical ones. This writer's brother-in-law is a
native of White Pine and his brother-in-law's parents are longtime
residents there. It is they who were chosen to be the primary con-
tact persons since they knew nearly everyone in White Pine and could
assist the fieldworker in making the necessary initial contacts.
More important than facilitating contacts, however, was that his
family relationship with the contact persons enabled people in White
Pine to identify him with people they knew well.
2. The category "Some College Work" includes all informants who had any
kind or length of post high school education, such as business or
other vocational school. According to the 1970 census, only 6.8 per-
cent of the residents of the White Pine division of Jefferson County
(figures are not available for the town of White Pine itself) over the
age of 25 had attended some college. However, three factors account
for a high percentage (45 percent) of this study's sample having col-
lege work: 1) the census area (with 2,085 people over 25) includes
rural area whose residents have a lower average educational level
than do the town's residents; 2) in 1970 a junior college opened five
miles from White Pine; it has attracted many part-time students, and
informants having attended as few as one or two courses are classi-
fied as having some college work; and 3) nearly all this study's in-
formants who enrolled in college for a short time dropped out (for
many in this generation, "trying a quarter of college" was the ex-
pected bridge from high school to the workaday world).

3. This same view, that the fieldworker must be willing to be a sales-
man, was expressed to the researcher by Vance Randolph, the dean of
American folklore collectors, in a personal conversation with the
writer on December 11, 1978, in Fayetteville, Arkansas.












CHAPTER II

BACKGROUND AND METHODOLOGY


Preface

The first chapter of this dissertation presents the non-linguistic

context of this study; this chapter discusses its linguistic context.

In the first part of the chapter several approaches to the analysis of

discourse structure are reviewed to put the analysis of oral expository

monologues into perspective. Then the methodology of this dissertation

is explained. The second section examines the importance of studying

monologue expositions. Then some remarks are made regarding the nature

of oral paragraphs, followed by a review of relevant studies of oral

and written paragraph structure. Finally Longacre's typology of dis-

course genres is discussed.

The control of the five different contexts mentioned in chapter

one (speaker type, frequency of occurrence, social situation, task, and

linguistic position) makes this study a novel one. Also, as far as

this writer can determine, this study is the only analysis of the dis-

course structure of oral expositions for any variety of American Eng-

lish. Although much has been written about discourse structure in Eng-

lish, oral expositions have been seriously studied only for other lan-

guages (see especially Longacre 1968). This dissertation represents

the first time that the task of eliciting, analyzing, and then gener-

alizing about a large corpus of expositions has been undertaken for

English. It therefore is an exploratory project.
61













Discourse Studies and the Approach of This Study

The decision to study the discourse level of oral expositions orig-
inally derived from this writer's growing conviction that the construc-

tion known as left dislocation, an often-used construction in speech,

could only be understood if its function were viewed as supersentential,

on the discourse level (Montgomery 1977a, 1978). If such a frequent but

little studied construction seemed to tell us something about how speak-

ers organize their linguistic behavior, it seemed likely that other fea-

tures and constructions of oral discourse should also be studied in

their discourse contexts. There is much to recommend the view that

oral expositions are preliminary to written ones. At least, understand-

ing how speakers negotiate and structure ongoing expositions might im-

prove our awareness of the communication system that people bring to the

writing process. Since linguists have for two generations made the case

that speech is primary and writing derivative, it seems crucial that a

study should begin with the structure of oral expositions rather than

written ones.

It is no longer necessary to justify the value of doing discourse

analysis, which may be broadly defined as the study of any segment of

verbal behavior larger than the sentence. Grimes (1975), the watershed

work in the field, has shown that discourse analysis is indeed possible

and that it rightly lies within the domain of linguistics (rather than

rhetoric or another field exclusively). Grimes' work is such a vast









cornucopia of observations on discourse that it should no longer be

asked "Is it worthwhile to do discourse analysis?" Rather we should

ask "What kind of discourse analysis will address significant ques-

tions about how people organize their linguistic behavior?" Since

Grimes provides neither a methodology nor a point of departure for a

principled approach to the analysis of expository discourse (but nei-

ther is his aim), we must look elsewhere for an approach to oral expo-

sitions.

According to Widdowson, discourse analysts have generally taken

one of two basic points of departure. Either they have started with

the text or discourse as a whole or they have begun with sentences or

subsentences in an effort to identify their discourse functionss:

One general approach to discourse analysis, then, begins with
instances of discourse, with actual data, and moves towards lin-
guistic units to the extent that this appears to be necessary
for the purpose of the description. The second approach moves
outwards, as it were, from the sentence, and deals with lin-
guistic expressions as realized in discourse but with the ab-
stract potential of linguistic forms. (1977:241)

The emphasis in the first approach is on the hierarchical nature of

discourse, in the second on the linear nature of discourse. There are

good reasons for choosing either point of departure and in many in-

stances discourse analysts have taken one or the other and elected not

to move toward the other. Propp (1958) and a number of other writers

on folkloristics do not move toward linguistic units at all. Several of

the modern-day text grammarians (Petbfi and Rieser 1976, Van Dijk 1972)

also rely on abstractions from a text for analysis and are primarily









concerned with the abstract and rhetorical structure of discourses.

Text grammarians for the most part view discourse or text as a unit of

competence and not of performance.

Other analysts approaching discourse (e.g. Williams 1964, Halliday

and Hasan 1976) choose to study exclusively the various cohesive re-

sources of English to fit one sentence after another into a context,

to produce acceptable sentence sequences. Williams examines deletion

and a wide variety of elements such as deictics and sentence adverbs.

Halliday and Hasan are concerned with such semantic resources of Eng-

lish as substitution, ellipsis, reference, and conjunction. This lin-

ear approach to discourse is devoted to explicating intersentential

linkage.

Several discourse studies attempt to bridge the gap between these

two points of departure and to balance the hierarchical and linear ap-

proaches to the study of discourse organization. They have varying re-

levance to the present study, however, since none of them (except, in

part, Longacre 1968) are based on a corpus of spontaneous oral exposi-

tions. Several studies, e.g. Grimes and Glock (1970), Klammer (1971),

Klammer and Compton (1970), and Van Dijk (1977), deal with only the

narrative structure of one or a small sample of texts. The basic point

of many narrative studies is that a discourse view of a text can indeed

make some interesting observations more so than it is to generalize

about a body of texts. In studies of narrative structure we encounter

again and again statements similar to that of Van Dijk: "The remarks

made about a well-known type of discourse, the story, could be extended









for other discourse types" (1977:155). Very few linguists have at-

tempted to do this, however. Longacre, in his Philippine study (1968),

has been the only one to do this. He posits formulae for the gramma-

tical structure of each type of discourse and paragraph. This comment

also applies to his extension of the concept of plot to types of dis-

course other than narratives.

Nearly all discourse analyses have investigated only written or

very highly formularized (i.e. the sermon or the folktale) oral texts.

The discourse features which they consider are not necessarily, even for

expositions, the same as those in spontaneous oral discourses, which we

will be concerned with here. As an example, we can compare Jones (1977)

with the present work. In her study of written expositions, Jones con-

tends that thematizing devices (repetition, marked word order, rhetori-

cal questions, clefting and pseudoclefting, among others) can fruitfully

be studied in written texts. But all the devices she finds are ex-

tremely rare in our White Pine corpus of oral expositions. The sole ex-

ception to this statement is conjunctions, which are of secondary impor-

tance in communicating themes. Moreover, the two most frequently used

thematic constructions from our corpus (left dislocations and existen-

tials) she does not mention. Thematic and other devices should be stud-

ied, but we should not assume that they are identical for written and

oral texts.

Other analysts have suggested other types of devices as crucial to

study in a discourse context: punctuating or transitional devices (Mer-

ritt 1972), various kinds of anaphora (Gutwinski 1976), equivalence









chains (Harris 1963), and those of Halliday and Hasan and of Williams

mentioned above. As valid as it may be to study each of these types,

the necessary heuristic framework for the comprehensive study of co-

hesive devices in discourse is provided only in Longacre and Levin-

sohn (1978). Our approach to the paragraph structure of oral exposi-

tions is largely based on the ideas in this article.

As suggested in their title, "Field Analysis of Discourse," Long-

acre and Levinsohn offer a framework for discourse analysis for lin-

guists working in the field with little-studied languages (such as

those which Summer Institute of Linguistics workers frequently encoun-

ter). What the present writer has chosen to do is to adopt it for

fieldwork with a much-studied language-English. Longacre and Levin-

sohn are determined to bridge the gap between the abstract structure

and the specific cohesive devices of a discourse but insist that the

analyst can do this only by first identifying and describing the "sur-

face structure cohesive devices" and discourse constituents: "The job

of the analyst is to...look through the flesh and the skin to the

skeletal structure beneath and to perceive the fundamental structure

of the whole" (1978:105). The aims of the analyst are to show how

these cohesive devices operate and then to display an outline of the

discourse in a schematic fashion.

Longacre and Levinsohn present eight surface structure cohesive

devices which permit a discourse to be outlined:

1) The role of tense and voice;

2) Particles and affixes;









3) Participant anaphora;

4) Deictics;

5) Lexical ties and paraphrase;

6) Summary and preview;

7) Conjunctions and introducers;

8) Backreference.

Given the time limitations of this study, four of these devices (the

role of tense and voice, particles and affixes, lexical ties and para-

phrase, and backreference) are not considered in this dissertation.

The remaining four (participant anaphora, deictics, summary and pre-

view, and conjunctions and introducers) are the concerns of chapters

three through five herein.

The second hypothesis of this study is that the grammatical struc-

ture of oral expository paragraphs is organized to a significant degree

by certain cohesive devices-the other four discussed by Longacre and

Levinsohn. It investigates this hypothesis by examining the functions)

and distribution of these devices.

The device called "participant anaphora" has to do with the identi-

fication of participants in a discourse: "Participants can be identified

by name, by a common noun, by pronoun, by an affix, or just by zero.

Such variations in anaphora are never unmotivated. Most commonly, the

domain of the participant anaphora chain is the paragraph" (Longacre and

Levinsohn 1978:108). We will here be concerned with only one link in

this chain-the first-(many linguists have studied anaphora: Gutwin-

ski (1976), Brinegar (1977), Hinds (1977) and others) and will call this









first link the "theme." The theme is the piece of information (not

necessarily a human participant) which usually initiates an anaphora

chain and which a block (or paragraph, as we will see) of information

is about. The primary surface structure devices which introduce theme

or "thematize" a piece of information are of two sorts in our corpus:

left dislocations and existentials. In the following exposition we un-

derline two left dislocations and two existentials, all four of which

initiate anaphora chains, to illustrate our approach.1

Fieldworker: What have been the major changes you've seen in
White Pine during your lifetime? Let's say, the most signifi-
cant changes?
Informant: Well, they restructured the bank. I wished they'd uh
left it the old way, but I guess they did that for, you know,
security measures and stuff, and then uh the mall, the mall I
was talking about, the shopping center. You know, they've
paved it all the way through, so that's, that was the major
thing right there that, that they've done. And of course,
the housing project, they put that down there, and I wish they
hadn't have. I like White Pine the way it was, you know.
Well, simpleton little town, you know, but it seems to be ex-
panding quite, quite a bit. (hmm) And there's a lot of
trailer courts going in. There's one right in the center of
town. I don't know how they ever got that admitted. Course,
they have restrictions now, but it was in there before the re-
strictions came, so (uh-huh) and, when I grew up, there, we
had one policeman, and that was Chief of Police Dick Reed, and
he had a deputy, and that was it, and they would drive around,
you know, most of the day, or if you needed him or anything to
around twelve o'clock, and then they'd go home, you know, so
the town was just, just nobody out, you know, or few people
out, but there was no crime, you know, nobody breaking into
anything. (uh-huh) That was just the way it was,... (V2-11)

In this exposition thematic devices introduce four themes (the mall,

the housing project, a lot of trailer courts, and no crime, although

the fourth one is not developed like the others) and at least four

paragraphs in this exposition. Theme is the one obligatory element









of a paragraph's grammatical structure and to a significant extent in

our White Pine expositions a paragraph's theme is marked by one of the

two constructions mentioned. In Longacre's most recent view of the

grammar of the paragraph, he indicates that "we find the paragraph

built around a theme that is not different in kind from a thematic par-

ticipant" (Forthcoming:118). Our chapter three confirms this for expo-

sitions.

Two other types of devices, deictics and summary, are considered

together in chapter four because deictics are frequently employed to

make summaries in our expositions. In general, deictics (especially

the word "that") help keep track of thematic information in a paragraph.

We will examine four uses of deictics in chapter four: 1) as the prono-

minal reference in left dislocation; 2) in extrapositions; 3) in iden-

tifying sentences; and 4) in what we will term "peak" sentences. An

example of a peak sentence is underlined below:


Fieldworker: What would you say have been the most significant
changes you've seen in the town in your few years here?
Informant: Most significant change I've seen is the, the intru-
sion or the coming in of, of new people. White Pine is no
longer the same town it was eight and a half years ago because
of its growth, uh because of people with different backgrounds
having moved into our town. (uh-huh) Uh therefore it has
changed the complex of our town tremendously. Uh I think this
is the greatest change that has happened in our town is to
have new people coming in with new blood, new thoughts, new
ideas, new approach to things. (X4-4)

We will see that such sentences represent another unit in the grammati-

cal structure of the paragraph-the peak-and we will confirm the view
of Longacre and Levinsohn that, in expository discourse, the peak is

'the culminating explanation" (1978:105).











The concern of chapter five is an eighth type of device-conjunc-

tions and introducers, insofar as they occur in a set of 200 of the

corpus' expositions. A rough taxonomy of the most frequent conjunc-

tions accompanying left dislocation is presented. Finally, some ob-

servations about subparagraphs, especially introduced by "like," are

made.

Longacre and Levinsohn thus elaborate eight types of surface

structure cohesive devices. Our object here is to show that, for one

corpus of English, four of them make complete sense only if viewed in

a discourse context. Although we examine the underlying regularities

of several of the specific devices (left dislocation and deictics which

participate in peak sentences, especially), our primary aim is to de-

termine their distribution and functions) across extended texts, which

also shows their periodic character. This periodic character reflects

the paragraph nature of spontaneous oral expositions. To a significant

degree, surface structure cohesive devices implement the grammatical

units of paragraphs, especially the theme and peak.

For convenience, parts of the analysis are performed on a subset of

the corpus' expositions, a random selection of 200 sample expositions,

five from each informant. In some cases, the analysis is based on a

smaller set of 80 expositions, two from each informant.

Based on the expositions of forty people, our findings have a

broad generality with respect to the community investigated, White

Pine, Tennessee. It might reasonably be asked why so little concern









and attention is given in this study to variation in cohesive devices

between subgroups of the informant sample, according to either the

age, the educational level, or the socioeconomic background of the in-

formants. Except for only a few scattered remarks (regarding the in-

troducer "like," for example) nothing is said about differences between

individual or groups of speakers. If White Pine had been presumed to

be a homogeneous speech community, the care to interview a large and rep-

resentative cross-section would not therefore have been taken. It was

found that the discourse devices under consideration in this study

showed inconsequential differences in function across the entire sample

of forty informants. None of the features were restricted to one subset

of the sample. The generalizations in this study resulted from their

obviously consistent functioning for all forty speakers.

This is not to say that no variation occurred in the frequency of

the realization of such devices as left dislocation. It was realized in

at least three instances for each speaker. But one speaker, a seventy-

year-old woman, had seventy instances. Hers was, however, by far the

longest interview-nearly three hours. Neither for the four informants

not native of Tennessee nor for the one native to Great Britain did left

dislocation function differently from White Pine natives. Peak sen-

tences, with fifty instances, were used by twenty three speakers of many

different educational levels and socioeconomic backgrounds and thus dis-

tribute broadly throughout the sample. Other devices may have varied in

frequency too. But none vary in function. More important, there is

little to suggest that any of the devices considered in this disserta-









tion are diagnostic of any of the speakers. A closer look might re-

veal some kind of variation, but that would be the object of another

study.


The Study of Monologue Expositions

Linguists interested in supersentential segments of speech have

only recently heeded Firth's injunction of nearly a half-century ago

that "neither linguists nor psychologists have begun the study of con-

versation; but it is here we shall find the key to a better understand-

ing of what language really is and how it works" (1957a:32). Only if it

is recognized that the function and manifestation of language is ba-

sically conversation can language best be understood. Recent studies

on turn-taking, as by Sacks, Schegloff, and Jefferson (1974), recognize

the importance of studying blocks of conversational dialogue and have

shown that they have cohesion, just as other discourse blocks do.

Klammer (1971) is a study of written dialogue paragraphs and is also an

important contribution in this respect.

The present study does not take into account turn-taking phenomena

between the fieldworker and the informants. But it is consistent with

the conversational frame of reference for studying language. It does

not artificially cut up conversation and thus de-contextualize the

monologue discourses it analyzes. It views expository monologues as

parts of conversational exchange, the context for which was controlled

(as indicated in chapter one) by the fieldworker. Pike and Lowe see

monologue as occurring when "one individual of the cast stays fixed









in the role of addressee" (1969:74). Klammer defines monologue more

loosely as "one uninterrupted speech within a dialogue context"

(1971:38).

The monologues we encounter in our corpus are usually isomorphic

with turns in conversation, but not always. In some instances it is

obvious that monologues continue across an exchange in turns. That is,

we cannot say that a speaker needs to reintroduce, at each turn in the

conversation, all the antecedents from an earlier turn.

Good reasons for studying the structure of monologues can be

found. Longacre views monologue as a "special lopsided development of

conversation" (Longacre 1976:165). But he stresses its "universal cul-

tural importance" (p. 197). His typology of discourse genres (Long-

acre 1968, 1976, Longacre and Levinsohn 1978), which he believes to

have universal applications, is actually a typology of monologues (nar-

rative, expository, procedural, and hortatory). What Longacre consi-

ders the deep structure or universal features of expository monologues

is discussed later in this chapter.

There is another good reason for studying the structure of mono-

logues. The expository monologue type of discourse on which this dis-

sertation is based has all the features of what Hymes (1971) calls a

"speech act": setting and scene, participants, ends, act sequence, key,

instrumentalities, norm, and genre. Viewing monologue as a unit, more-

over, allows us to examine thematizing devices, deictics, and other de-

vices in this dissertation. This is because these devices help orga-

nize monologue conversational turns into one or more self-contained









thematic units-paragraphs-which represent the natural thematic struc-

tural units in which speech is used.


Oral Paragraphs and a Review of Paragraph Studies

In analyzing the structure of discourses, linguists have often rec-

ognized distinct supersentential units smaller than the discourse it-

self. This is the case with Van Dijk (1977) for narrative, Klammer

(1971) for dialogue, and Longacre (1968) for dramatic, procedural, and

other types of discourse. Van Dijk uses the term "macro-structure" for

a sequence of sentences which belong together and which frequently has

"Morpho-Phonological markings" (1977:152-3). He mentions pauses, in-

tonation, and particles as examples of such markings. Klammer, Long-

acre, and others adopt the less original but more useful term "para-

graph," the traditional term referring to an orthographic unit of lan-

guage. Herein we will also use the term paragraph for these interme-

diate structural units in oral discourses that are unified in more ways

than simply forming a block of print on the page.

For Longacre, any string of two or more sentences which belong to-

gether constitute a paragraph: "Unstructured sequences, or strings of

sentences which do not constitute a paragraph structure, do not exist"

(1968:53). Most often the paragraph has been defined as any self-con-
tained thematic unit. Pike and Pike call it "the minimum unit in which

a theme is developed" (1977:488). We consider a paragraph to be any

thematic block of language, i.e. a unit of language organized around

a theme or piece of information.









Thematic unity is not something merely imputed to a paragraph in

speech. It is there and is reflected to one extent or another by var-

ious lexemic, grammatical, and phonological features. In a narrative

discourse, paragraph themes are usually characters; in exposition the

themes may be characters or many other kinds of information. Examples

will be presented in chapter three.

We can view these thematically unified units called paragraphs as

having three different modes of structure-phonological, lexemic, and

grammatical.2 Although of most interest to us is the grammatical mode,

all three modes to a great extent reinforce one another. It is doubt-

ful that the grammatical structure of a paragraph can be understood

without studying the other two.

Paragraphs have not often been understood as having characteristic

phonological features. But according to Klammer, "phonologically, in a

complete description of dialogue paragraphs, it would be necessary to

account for such phenomena as stress placement, juncture, and intona-

tion in relation to the paragraph level constraints affecting them"

(1971:33). Even earlier, Becker pointed out the existence of phonolo-

gical markers of paragraph structure:

Paragraph tagmemes seem to be marked by shifts in pitch regis-
ter, tempo, and volume when paragraphs are read aloud. While
these signals can be perceived by a trained phonetician, they
have not been adequately described in the laboratory, and
their written counter parts have not been identified. (1965:242)

Unfortunately research into such features and markers, as by

Bridgeman (1966) and Lehiste (1978, 1979), has been either too recent









or too sparse to have developed well-tested tools to examine them.

Bridgeman's pioneering work on Kaiwa claims that oral paragraphs in

that language have one of a set of six phonological onsets and one of

three phonological codas, with an optional peak preceding the coda of

a paragraph. Lehiste has attempted to find, with as yet little suc-

cess, the phonological correlates of paragraph boundaries by measuring

pausing, lengthening, and pharyngealization at the beginning and end of

sentences and paragraphs.

The isolation of the phonological features of paragraphs is the

subject for another dissertation. Only occasional observations regard-

ing phonological matters are advanced in this study, when these features

are concomitant with the surface structure cohesive devices that we

analyze. More than likely a characteristic paragraph-initial lengthen-

ing and/or pausing will eventually be shown to aid in giving paragraphs

coherence. One suspects, however, that no phonological signal will cat-

egorically mark some part of a paragraph's structure. What one should

expect to find is that a significant coincidence of phonological and

grammatical signals co-occur at thematic breaks in a discourse. In

other words, phonological paragraphs should be to a great extent iso-

morphic with grammatical and lexemic ones.

Paragraph structure can also be viewed as lexemic, involving how

the semantic resources of the language organize information into a the-

matic unit for a given kind of discourse. Although a principled means

of approaching the lexemic paragraph has yet to be satisfactorily de-

vised, quite a few lexemic functions which participate in paragraph









structure have been pointed out. Klammer (1971), largely adopting

his functions from Wise (1968), mentions addition of participant, elab-

oration,, comment, specification, contrast, reason, summation, conclu-

sion, setting, and result. But it has been difficult to show how these

functions are organized linearly, as Longacre has done for the grammat-

ical units of the paragraph. The set of such functions has seemed

open-ended and at the same time we have not yet found, according to

Pike and Pike, "a way to treat some of them in a linear fashion, in

order to be able to perceive them as something 'going on"' (1977:364).

This difficulty of formulating the lexemic structure of actual

paragraphs has in recent years caused lexemic structure to become

viewed as "deep" in contrast to the "surface" structure of paragraphs

being grammatical. The problem has become one of showing how lexemic

functions or units were realized by grammatical ones. Otherwise, lex-

emic paragraphs would be only hypothetical. The problem of the extent

of isomorphism was tackled by Wise (1968) for oral narratives in Noma-

tsiguenga and for written dialogues in two English texts (I Henry IV

and Great Expectations) by Klammer (1971). It is useful to quote at

length the remarks of Wise:


Lexemic structure, as the term is used here, is a variety of
deep structure. For example, "logical subject" and "logical
object," i.e., agent and goal, are constituents of lexemic con-
structions on the clause level. In levels beyond the sentence
the lexemic order of constituents is, in general, the chronolo-
gical order of events in the narrative. Paraphrase and syno-
nymy are important for the analysis of lexemic sentences and
phrases as well as for the analysis of lexemes which are posited
as the minimum lexemic units.









In contrast, grammatical structure is a variety of surface
structure. For example, "grammatical subject" and "grammatical
object," i.e., subject and complement, are constituents of gram-
matical constructions on the clause level. In levels beyond the
sentence, the grammatical order of constituents is the actual
surface order in which the narrative is told. Morphemes are po-
sited as the minimum grammatical units.
The nature of constituent units in lexemic constructions is ra-
dically different from that of constituent units in grammatical
constructions, but the boundaries of the two kinds of construc-
tions frequently coincide. However, the boundaries are not always
isomorphic so that the lexemic and grammatical constructions can-
not be mapped onto each other in a direct, one-to-one manner.
(1968:2-3)

This kind of mapping can be easily done in many cases. For example, the

lexemic function of the introduction of a participant into a paragraph is

usually equivalent to and realized in the theme and this is further rein-

forced by one of the thematizing cohesive devices-left dislocation in

many cases. The lexemic function of summation is sometimes indicated by

the kind of structure we are calling peak sentences, but this is not al-

ways the case. Such sentences always represent the peaks of paragraphs

grammatically but only sometimes are these peaks the realizations of such

sentences.

The third and most discussed way to view paragraph structure is gram-

matically. The grammar of a paragraph is the system of units (Longacre's

current term) or constituents (Wise's term) whose order is the actual

surface structure order of the paragraph. Longacre (1968, 1976, Forth-

coming) especially has long made the case that paragraphs are grammatical

entities and that each type of paragraph (narrative, explanatory, horta-

tory, dialogue, precatory, etc.) has its own grammar. He contends that

the units (in his 1968 work he calls them "tagmemes" in positing his









paragraph formulae, as we will see below) which constitute such enti-

ties are ordered linearly, and he posits a tagmemic formula for each

major and minor type of paragraph in Longacre (1968).

As Becker (1965) rightly points out, the description of paragraphs

as grammatical entities owes much to traditional rhetoric. Longacre's

and especially Becker's formulae for expository paragraphs resemble

closely long-employed formulae for paragraph development in rhetoric

texts. Becker's contention is that the majority of expository para-

graphs are organized grammatically in either a TRI (Topic-Restriction-

Illustration) or a PS (Problem-Solution) fashion.

Longacre's major contributions to the study of the grammar of para-

graphs lie in the aforementioned formulae and in his systematic delinea-

tion of paragraph types (Longacre 1968) and discourse types (Longacre

1968, 1976) on the original basis of studying paragraphs and discourses

in twenty-five Philippine languages and dialects. His general formula

for the tagmemes constituting explanatory paragraphs in the various

languages is3

PRELIM (+TEXT *EXPO *REASON *RESULT *WARNING) *TERMINAL
(1968:109)

With TEXT as the only obligatory tagmeme, this formula can clearly ac-

comodate an indefinitely large number of different patterns, since EXPO

(exposition) "can occur an indefinite number of times in paragraphs"

(p. 109). Longacre views TEXT as "much like the topic sentence of tra-

ditional rhetoric" (p. 109). We take the formularization of the gram-

mar of general paragraph types to be a desirable goal, but it is not









clear yet how tightly one can be devised for expositions in English.

Longacre's quest for a formula for the linear organization of para-

graphs serves as a practical model for us; our study focuses on two of

the units Longacre posits in his formula for expository/explanatory

paragraphs.

Longacre has revised his views on two of his units (or tagmemes,

as he called them in 1968), Text and Peak. Originally likening Text

to a paragraph's topic sentence, he indicates that the obligatory unit

in a paragraph is the theme, that piece of information which a para-

graph is "built around" and which, for an expository paragraph, "is not

different in kind from a thematic participant" in a narrative para-

graph (Forthcoming:118). The distinction between a topic sentence and

a thematic piece of information is an invaluable one; especially it is

the piece of information (a participant or whatever) which, through an

anaphora chain, gives thematic units to a paragraph. This most recent

view of Longacre's that a paragraph has an obligatory theme is consis-

tent with the one we maintain here.

In his earlier view, Longacre considered Peak to be only one as-

pect of Text. He states that Text "is the PEAK grammatically and lex-

ically of its paragraph. As PEAK of its paragraph the TEXT contrasts

in placement with the PEAK (BU or STEP ) of NARRATIVE and PROCEDURAL

PARAGRAPHS which come later in the paragraph" (1968:109).4 In a later

work, he indicates that Peak is the plot-like element optionally pres-

ent in non-narrative as well as narrative paragraphs (Longacre 1976:

228-9). It comes near the end of and represents a kind of climax or









denouement to the paragraph. Longacre's most recent view of Peak

is an even more refined one:

For discourses that are not narrative but still have a climax
of development, the peak may mark "target procedure" in a
procedural discourse, "climactic exhortation" in a behavioral
discourse of the hortatory variety, and a most satisfactory
or "culminating explanation" in expository discourse. .
(Longacre and Levinsohn 1978:105)

If we look back to the exposition above at the sen-

tence therein ("I think this is the greatest change that has happened

in our town is to have new people coming in with new blood, new

thoughts, new ideas, new approach to things"), we see clearly that this

sentence represents the culminating explanation and that it is the

peak of this one-paragraph exposition. Our chapter four below is

largely devoted to showing how such sentences indicate the peaks of

paragraphs.

In the remainder of this study we show the extent to which we can

find that certain surface structure cohesive devices are associated

with grammatical units of expository paragraphs.

At this point we should ask why we find surface structure cohe-

sive devices such as thematizing constructions and peak sentences at

all, since their use is not categorical. Left dislocation, as we will

see in chapter three, is employed in only slightly more than one-fifth

of its potential cases. If left dislocation is optional, as are other

thematizing constructions in English, why do speakers use it at all?

Why is it that grammatical units of the paragraph are sometimes

overtly marked by a specific structural reflex? We will suggest here









that such devices occur because they serve certain of a speaker's

pragmatic purposes.

The speaker's primary purpose in exposition is to convey new infor-

mation to the hearer. For this reason the speaker/producer of a dis-

course or text takes into account the presuppositions and expectations

of the hearer. The hearer expects the speaker to supply the necessary

antecedents and background for him/her to follow what is being said. In

other words, they have what is referred to as the

"Given-New Contract" of cooperation. The speaker must be aware of what

the hearer knows, which information has already been given to the hearer,

and which information will be new to the hearer. The hearer must make

the effort to follow the progress of the exposition.

It follows that any means which facilitates the fulfillment of this

contract and which makes an exposition easier for a hearer to follow may

well be used. While such thematic devices as left dislocation can appa-

rently be used also for other pragmatic reasons, basically they seem to

facilitate this given-new contract between the speaker and the hearer.

Thematic devices present to the hearer what new information will be com-

mented on in an ensuing message. This piece of information represents

what will be thematic for what follows and has the potential for being

the antecedent for the following block of discourse. It is only if the

hearer has such a piece of information, the theme, clearly in mind that

he/she can compute what the speaker says.

It is probable that the use of cohesive devices is especially fa-

vored by a speaker in certain pragmatic situations. Left dislocation,









for instance, seems more likely to be used when a speaker wishes to be

emphatic, to show a contrast, or to make a particularly abrupt point.

Although the primary function of left dislocations is to introduce new

themes, our corpus has thirty five instances of left dislocation invol-

ving given information, when a speaker wishes to reemphasize a piece of

information emphatically. In nearly ten percent of our left disloca-

tions (45 out of 606) we find the dislocated NP immediately preceded by

"but" or "however," showing that the speaker contrasts those NPs with

earlier ones. In seven instances, we find that a left dislocation is

immediately followed by an identifying sentence, as in

1) Threshing floors, that's what they had. (III2-27)

This shows that a speaker wishes to make an especailly abrupt point

about a theme.

In the remaining chapters of this dissertation a strong case is

made that we cannot understand how expository discourse is organized

without considering the supersentential or discourse context of cer-

tain cohesive devices. If this is true, it seems to imply something

about how a speaker communicates to a hearer. Cohesive devices, the

formal signals of the language in which a discourse is expressed,

seem to aid a speaker in communicating the hierarchical structure of

a discourse linearly. This is another way of saying that such devices

communicate the outline (implicitly, of course) of an ongoing discourse

to a hearer.









Longacre's Typology of Discourse Genres

It was indicated in chapter one that for practical reasons the

fieldworker chose to elicit expositions rather than other types of oral

discourse in the White Pine community. It was apparent that exposi-

tions were the type most readily elicitable from the cross-section of

people that he wished to interview. Also it was decided that the col-

lection of a large corpus of expositions rather than smaller corpora of

two or more types of discourse would be more conducive to generaliza-

tions about expositions.

That exposition is a distinguishable discourse type is not an ad

hoc presumption on our part. It is a hypothesis based on the typology

of discourses developed by Longacre (1968, 1976). In his study of

twenty-five languages and dialects in the Philippines, Longacre and his

colleagues undertook the basic spadework of sorting out which discourse

types could be described across the languages studied. The typology of

discourses which resulted has provided a practical framework for fur-

ther investigations with other languages, as this study is an investiga-

tion into the English of the Southern Appalachian region.

Limiting himself to prose monologues, Longacre indicates that four

genres of such monologues (narrative, expository, procedural, and hor-

tatory) have common characteristics cross-linguistically and that

within any given langauge there are a finite number of other genres.

We may characterize the four main genres as follows: narrative dis-

course relates a story of some kind; expository discourse explains

something or exposes a subject; procedural discourse (e.g. a recipe)









tells how to do something; and hortatory discourse endeavors to bring

about a change of some kind.

The major parameters along which these four genres are distin-

guished are shown in Figure 2.1 on the following page and adopted from

Longacre (1976). The rows in this figure distinguish those genres

which are +succession (having a sequence in time) and those which are

-succession (lacking a sequence in time). The columns distinguish the

+projection (occurring in the future) from the -projection (those hav-

ing accomplished time or in which the time frame is not focal) genres.

Longacre's scheme further indicates the person orientation of each of

the four genres.

According to Longacre's typology, narrative discourse characteris-

tically involves a sequence in time, occurs in an accomplished (past)

time frame, is agent-oriented, and is related from the first or third

person point of view. Procedural discourse also involves a sequence in

time, but occurs in a projected time frame, is patient-oriented, and is

not related from the point of view of a specific person. Hortatory

does not involve a sequence in time, is set in projected time, is ad-

dressee-oriented, and is related from second person point of view. Hor-

tatory discourse strictly speaking does not have a time frame but a

mode-since the speaker is wishing or urging that something take place.

Expository discourse need not involve a sequence in time nor does it

necessarily have either an accomplished or a projected time frame (time

is not focal). Expository discourse is subject matter-oriented and has

no necessary person reference (point of view). It is essentially im-
personal and may involve either inanimate or animate entities. If




















S NARRATIVE PROCEDURAL
U 1. 1/3 person. 1. Non-specific person.
C 2. Agent oriented. 2. Patient oriented.
C 3. Accomplished time. 3. Projected time.
+ E 4. Chronological linkage. 4. Chronological linkage.
S
S
I
O
N

S EXPOSITORY HORTATORY
U 1. No necessary person reference. I. 2 person.
C 2. (Subject matter oriented). 2. Addressee oriented.
C 3. Time not focal. 3. (Mode, not time).
-E 4. Logical linkage. 4. Logical linkage.
S
S
I
0
O
N

Diagmm I. Deep structure genre


Figure 2.1 Deep Structure Genre
(From Longacre 1976:200)


-PROJECrED


+PROJECTED











it brings in people," Longacre says, "they are simply subjects of

explanation and analysis" (1976:199).

More recently Longacre has furthered elaborated his scheme by

adding another parameter tensionon. What he says about tension char-

acterizing some discourses is worth repeating:


The final parameter *tension also applies to all the above dis-
course genres. Narrative discourse is purely episodic if it
contains essentially no struggle or plot (for example Willa
Catcher's novel Shadow on the Rock). Such discourse is consi-
dered to be -tension. Most stories, however, involve some
sort of struggle or plot and are clearly +tension. Similarly,
in procedural discourse, there are some discourses which are
more or less routine, while others involve struggle and alter-
natives and are therefore +tension. Both behavioral [horta-
tory] and expository discourse have varieties in which argu-
mentation is assumed. These are also +tension. Discourses
of these genre which do not have this characteristic are
-tension. (Longacre and Levinsohn 1978:104)

we mention this third parameter because it concerns the presence or

absence of a peak in the discourse: "If a discourse is plus tension,

there will most likely be some kind of climax of development, some

marked surface structure 'peak' (Longacre, 1978 ). This is a pri-

mary concern of chapter four.

Longacre's characterization of expositions appears to have consi-

derable validity for the expositions constituting our corpus, although

it must be pointed out that Longacre's types are "pure." In actuality

discourses often of different types are embedded within one another.

For example, within an exposition may be embedded a short narrative

for background or illustrative purposes. Longacre recognizes this in

viewing his genres as deep structure ones which are sometimes skewed in

the surface structure (1976:206-9).









Notes



1. The parenthetical notation at the end of these and all further
data in this dissertation refers to the page of typed transcript
on which they are found. For information on which speaker any
data are from, consult the list of informants in Appendix A.

2. Terminological obfuscation notwithstanding, we use lexemic to
refer to concepts such as those of Klammer and Wise, including
contrast, comment, addition of participant, specification, which
are manifested in linguistic structure in various ways. Other
tagmemicists (e.g. Pike and Pike 1977) use "referential" to
refer to such concepts. Stratificationalists use the term "seme-
mic."

3. For Longacre, terms "explanatory" and "expository" are equivalent
in his 1968 study.

4. Step and BU (Build Up) are the obligatory tagmemes in procedural
and narrative paragraphs.

5. In his 1976 work, Longacre does not distinguish description from
exposition and we will consider description a subtype of exposi-
tion here.












CHAPTER III

THEMATIC DEVICES


General Remarks
It is indicated in chapter two that a speaker may use certain

constructions known as cohesive devices to signal to a hearer what

the theme of a following block of the speaker's message will be.

Theme is the obligatory unit of such a block or paragraph and is de-

fined as the piece of information (not necessarily a human participant)

which initiates an anaphora chain and which the block is about. Al-

though it must be generalized to apply to all paragraphs, this narrow

definition of theme demonstrably applies to those paragraphs in our

corpus beginning with thematic devices.

In this chapter our concept of theme is compared with others

(Halliday 1967-8, Firbas 1966, Jones 1977) and we discuss briefly those

constructions that these treatments (especially Jones) consider to be

thematic devices. The main portion of the chapter explores the nature,

functionss, and distribution of one particular device-left disloca-

tion. This is one very common device through which paragraphs are

formed in this corpus of expository monologues. Finally we look at

existential sentences in the corpus and see how they also function as

thematic devices. They also complement left dislocation.

More often than not themes are reflected by anaphora chains ex-

tending for more than one sentence. An anaphora chain is a string of









substitutes (pronouns, for the most part) in a paragraph of language

which refer to a common piece of information or theme and is a special

kind of equivalence class. According to Becker (1965), an equivalence

class includes not only pronouns and demonstratives but also the repe-

tition of words and their synonyms. Although not every paragraph in

our corpus has an anaphora chain, those which begin with a "left dis-

located" NP nearly always do. In the two following examples the links

of the anaphora chains are underlined:

1) Fieldworker: Haveyou participated in any of the Pine Day cele-
brations here in town? Have you attended any of them?
Informant: Yes, I usually go, because the ladies of our church,
we usually have a booth. We bake things, and so we have a
sile. That's a big day for us, you know. That's a way to
make uh some money for our uh projects, and things like that,
and that's how we were able to help pay for our uh paint that
recently that we uh, we painted the inside t1e-church. The
ladies did a little painting, but, but we did have enough
cashto buy all the paint. (III4-15)

2) ...This property that he, he let us have then for the parsonage,
build a parsonage of it, it, at one time there was, was a school
there. I was going to reTte that to you. There, that's where
that I went to school, and my first eight years of schooling
was right there. And at this old school, that's where this par-
sonage now sets across from our church. (11I2-1)

In the first example, "the ladies of our church" is clearly the theme

of a block of information several sentences beyond its introduction. We

find ten pronouns referring back to it and one repetition of "the la-

dies." In the second example, the theme is "this property. .."; it is

around the property that the remainder of the exposition here is orga-

nized. The anaphora chain contains such a pro-word as "there," since

the piece of information which is left dislocated and which is thema-
tic is a locus. We see in these examples that theme is not just an











intuitive notion but that it demonstrably organizes a block of dis-

course.

Our view of theme as that information which unifies a block of dis-

ourse and which often is reflected by an anaphora chain is not incon-

sistent with other views found in the literature. For Halliday,


Theme is concerned with the information structure of the clause;
with the status of the elements not as participants in extralin-
guistic processes but as components of a message, with the rela-
tion of what is being said to what has gone before in the dis-
course, and its internal organization into an act of communica-
tion. (1967:199)


Halliday's position is that theme is involved in both clause and dis-

course structure. Theme is "what comes first in a clause" and around

which the remainder of a clause is structured (1967:212). But also

theme is that information which "is being talked about" (p. 212) and

which assigns


.to the discourse a structure which is independent of sen-
tence structure and through which the speaker both organizes
the act of communication into a chain or message blocks, the
'information units', and specifics within each message block
the value of the components in the progression of the dis-
course. (p. 196)


Halliday's express concern, however, is not to show how a theme parti-

cipates in discourse but to explore the many choices that a speaker

has for theme and how a clause is organized around such choices. To

support his clain that "the sneaker has within certain limits the op-

tion of selecting any element in the clause as thematic" (p. 212), he

shows the variety of different elements which may be fronted and thus









thematized in sentences with inverted word order (pp. 214-5). In sen-

tences 3 and 4 below, for example, a direct object and a temporal ad-

verb are thematized:

3) These houses my grandfather sold.

4) Tomorrow John's taking me to the theatre.

Sentences with such inversions are very infrequent in our data, but Hal-

iday's general claim about the kinds of elements which may be themes

matches closely those thematized by left dislocation shown later in

this chapter.

The present study takes a step beyond Halliday by examining how a

theme relates to a context beyond its immediate sentence, but our view

of theme is consistent with Halliday's. Nowhere does he state that a

theme's domain is bounded by its immediate sentence, and we may reason-

ably infer that he would agree with our finding that theme plays an

important role in organizing a paragraph.

In the Prague School view, theme has to do with the information

structure of a sentence. The elements) of a sentence which are

"given" (previously mentioned in a discourse) constitute that sen-

tence's theme. The theme need not appear in initial or in any other

sentence position and a sentence need not have any given information.

If it does not, it has no theme (Firbas 1966:268). Danes and Firbas,

contemporary members of the Prague School, continue this tradition.

Firbas' view of theme, however, is much less similar to the one we

take here because it is explicitly confined to the limits of a sen-

tence. His view is a refinement of the widely known theme-rheme




Full Text
100
"I mean," "I guess," and "I think." Such clauses add little or no in
formation to the theme of a passage. They differ insignificantly
from dislocations from matrix clauses. If we add these to all the dis
locations of matrix subjects, objects of verbs and of prepositions, we
find that 92 percent (506 of 550) of the dislocations are of the matrix
kind. Our contention in this chapter is that left dislocation intro
duces thematic elements into the stream of discourse. Since it is
quite reasonable that newly-introduced participants and thematic pieces
of information are most likely found in the central part of a sentence,
it is no surprise that matrix clause NPs are by far the preferred ones
for left dislocation. However, the non-matrix dislocations are of the
matic elements. But clearly it is usual that they first appear in a ma
trix clause.
Three other types of NPs which can be dislocated and replaced by
an appropriate pro-word are locatives (22 instances, 3.6 percent of the
total), as in 19-20 below, temporals (4 instances, 0.7 percent), as in
21 below, and possessives (23 instances, 3.8 percent), as in 22-3 below:
19) Well, at White Pine, you know, you're really more of an indivi
dual there. (VI11-1)
20) Now where Earl Atchley was raised, I mean over at Rankin side.
there was a big coal chute over there that furnished. .
(II12-6)
21) Back when that thing was a-running, well, I didn't know nothing
about it then. (X2-28)
22) This little James boy lived right down the road from here, just
the othere day my pastor had his funeral. (VI14-16)
23) Some of these very people who would criticize uh the rural areas
of East Tennessee, their ancestors came out of them. (X4-11)


19
Throughout all of the churches one finds a firm fundamentalism
(Puritanical in kind for most churches). One finds also a strong com
mitment to evangelistic activity and to the need for the individual to
undergo a conversion experience and for the congregation to support
missionary activities and frequent week-long revivals.
All the White Pine inhabitants refer to each one of the ten min
isters in the community as "Preacher" regardless of an inhabitant's own
church affiliation. Undoubtedly this reflects the fact that until a
generation ago, local churches usually did not have their own ministers.
Rather, they shared a minister with one or more churches outside the
community (the days of the "circuit riders"). It also seems to reflect
the present-day notion that preaching the Word of God is the most impor
tant duty a minister has. The epithet thus continues even though cler
gymen have adopted all the everyday duties of the profession (counsel
ing, visitation, etc.) in recent years. Preachers are also invested
with a great deal of moral authority in the community. They can wield
considerable political influence as well, if they choose, on many is
sues of local import such as the granting of beer permits or of permits
for businesses to open on Sunday. These are quite important issues in
White Pine.
One finds little evidence of sectarianism in White Pine. There
seems little difference between the Baptists, the Methodists, and the
Presbyterians in either their theology or in their way of conducting
business. Indeed, they coexist quite harmoniously. One's church af
filiation is viewed generally as dependent on the affiliation of one's
parents or one's spouse, not on one's personal convictions.


29
Throughout the literature on Southern Appalachia runs the view
that its people and culture are distinctive in at least a broad sense.
Most comnentators on the region point out that a combination of such
traits as "fierce" individualism and mistrustfulness of all strangers,
among others, characterize the Southern Appalachian native and make him/
her different from other Americans. White Pine is more of a valley than
a mountain community, and such characterizations have been more accu
rate of the isolated communities in the region than of the urban and
valley sections of Southern Appalachia. This has sometimes been recog
nized, as by John C. Campbell, who lived and taught in the region for a
quarter century. Campbell was concerned that some of his statements in
his The Southern Highlander and His Homeland would not be understood as
"applicable to the remote rural folk who were the particular object of
his study [and] were not true of their urban and valley kinsfolk"
(1921:xiv). But such characterizations, while sometimes extreme,
rarely claim that Southern Appalachia is culturally or socially homoge
neous. White Pine itself is certainly neither. Yet several broad cul
tural traits have for so long been pointed out as typical of Southern
Appalachian people that it is worthwhile to determine if they are ap
plicable to White Pine and can enable us better to understand it as a
Southern Appalachian community. We will ocnsider the four traits or
themes mentioned above that Ford chose to examine by conducting a
region-wide survey of Southern Appalachia. Ford found these traits to
be significantly apposite to both urban and metropolitan communities,
although less so than to rural communities.


6
Figure 1.1 Jefferson County, Tennessee
(Source: East Tennessee Development District)


55
The use of brackets and parentheses in the interview transcripts
follows the suggestions of Ives (1976). Parentheses serve to enclose
short interviewer remarks (rarely more than the occasional "uh-huh"
or "ham" which insures an informant that the interviewer is intently
following the informant's account) which did not cut short or deflect
an informant's train of thought. Occasionally parentheses were also
employed to suggest an alternate transcription, when the speech on
the tape was not quite intelligible. Brackets enclose either inter
viewer comments on the informant's behavior or short summaries of the
interviewer's speech. In the interview excerpt presented below, the
two different uses of parentheses are exemplified.
Fieldworker: In general, what have been the most significant
changes that you've seen in the community of White Pine it
self during your lifetime?
Informant (74 year-old male): Oh, I guess this, what I remember
as farmland is, has been being and is being covered up with
houses and housing developments and such as that, (uh-huh)
and around the town. Most of those, I remember when, when
that was farmland, the ones out this side of town and well,
that's all around town. And that uh, and that and inflation,
(hmm) Course, we've gone from, as far as uh, as transporta
tion, we've gone from horses and buggies and mules for to
farm with, and buggies to ride in and, and trains hauled the
produce, to now it's gone to trucks. The trains are, people
ride in cars and, and uh, and produce being hauled in trucks,
right off the farm from there on, from the time it's produced
til (to?) the, the time it's gets to the grocery store. Rail
road used to use a good many people. That was pretty big bus
iness. Of course, it hauled lots of people in those days.
That was quite a, quite an event when a train came to
town....
As mentioned, all the interviews were entirely transcribed by the in
terviewer himself. Inevitably there are sentences, phrases, and even


10
possible. As one sixty-year-old farmer who witnessed the relocation
of the thirties put it,
...the landowners were against it. They didn't want uh, they
didn't want to move out....Of course, we realize now that TVA
is our, our main uh source of energy and power and so, I don't
know what we'd do without it.
TVA, through its water management and its provision of electricity, set
the stage for the ensuing economic development of the area, for the
profitability of agriculture, and especially for the coming of indus
try. Since the war several dozen major factories have relocated to
East Tennessee.
Until after the Second World War and the locating of the facto
ries White Pine was very much an isolated farming community. Its
sphere of interest hardly extended beyond its boundaries. A half dozen
stores provided the necessities and there was little pressure to de
velop the town residentially, commercially, or industrially. The sen
timent was to keep the town as quiet as possible and to keep any con
scious development out. This is how the state of affairs in White Pine
remained, too, until only a few years ago. Eventually the preoccupa
tion with prohibiting development was buffeted by demands for more
housing, for recreational facilities, and for planning in general, so
the town has had to undergo considerable growing pains in meeting these
demands. These pains will persist for a long time because the pres
sures, primarily from the younger generation, will continue.
The most profound development in White Pine over the past thirty
years has been its evolution into a "bedroom" or almost strictly


176
Hinds, John V. 1977. Paragraph Structure and Pronominalization.
Papers in Linguistics. 10. 77-99.
Hinds, John V. 1978. Levels of Structure Within the Paragraph.
BLS 4. 598-609.
Hymes, Dell. 1971. The Contribution of Folklore to Sociolinguistic
Research. Journal of American Folklore. 84. 42-50.
Ives, Edward 0. 1976. A Manual for Field Workers. Volume 15 of
Northeast Folklore. Orono, Maine: The University Press.
Johnson, Linda and Richard Bayless. 1976. Cohesion in a Discourse-
Based Linguistic Theory. Manuscript.
Jones, Linda Kay. 1977. Theme in English Expository Discourse.
Lake Bluff, Illinois: Jupiter.
Kaberry, Phyllis. 1957. Malinowski's Contribution to Fieldwork Me
thods and the Writing of Ethnography. In Man and Culture: An
Evaluation of the Work of Brontislaw Malinowski, by Raymond
Firth. New York: Harper and Row. 71-91.
Kantor, Robert Neal. 1977. The Management and Comprehension of
Discourse Connection by Pronouns in English. Ohio State Uni
versity dissertation.
Keenan, Elinor Ochs and Bambi B. Schieffelin. 1976. Foregrounding
Referents: A Reconsideration of Left Dislocation in Discourse.
BLS 2. 240-57.
Kephart, Horace. 1976[1922]. Our Southern Highlanders. Knoxville:
University of Tennessee.
Ketner, Kenneth Laine. 1973. The Role of Hypotheses in Folkloristics.
Journal of American Folklore.
Klammer, Thomas Paul. 1971. The Structure of Dialogue Paragraphs
in Written English Dramatic and Narrative Discourse. Univer
sity of Michigan dissertation.
Klammer, Thomas P. and Carol J. Compton. 1970. Some Recent Contri
butions to Tagmemic Analysis of Discourse. Glossa. 4. 212-22.
Koen, Frank M., Alton L. Becker and Richard E. Young. 1968. The
Psychological Reality of the Paragraph. In Proceedings of the
Conference on Language and Language Behavior, edited by Eric M.
Zale. New York: Appleton-Century Crofts. 174-87.


140
In such sentences the deictic is anaphoric (except in a small
handful of cases we discuss later). It points backward in the
discourse and identifies something which has gone before with a
clause of information following it. The something ("beer joint,"
"shopping center," and "fight back") usually, but not always,
appears in the preceding sentence. Halliday (1967) indicates the
communicate effect of anaphoric identifying sentences:
33) ... the combination of deixis with identification,
particularly when the deixis is anaphoric, being highly
effective as a form of communication. The speaker
represents one part of the message as to be identified,
and then identifies it with something that is shown deic-
tically to be recoverable from the preceding discourse.
Furthermore since the demonstratives may have extended
text reference the identifier "that" in most instances is
not restricted to an element functioning in the clause;
any stretch of discourse may participate in the identifying
relation in this way. (p. 232)
If the identification is not with an NP as in examples 30-2 above
but with the entirety of a foregoing block of discourse, the effect
is of a summing up and tying together of the block. Showing this is
the following extended excerpt:
34) Fieldworker: What's a funeral like in a small town where
everybody knows everybody?
Informant: I think it's more important to the people. Yeah,
a funeral to me, it's not for the dead person. It is for the
people who are living, (uh-huh) that have lost somebody, (uh-
huh) In here, you know, if you know the person, you, every
body feels a sadness and, you know, you comfort 'em, and
we're we're closer. You know, the biggest thing to me is
like my Aunt, we viere talking a little bit about these funer
als, that you send invitations to a funeral, and that (hmm)
to me would just be, you know, how could you send an invita
tion to something like that? I mean you send invitations
to parties and things. Uh it* to. me, you know, a funeral
where you come if you want to. You know, you don't come just
because you got that invitation, come by invitation, then


13
of its college-educated citizens settle in town despite its lack of
opportunities. One twenty-five-year-old male college senior puts it
this way:
It's hard to say, you know, when the settling down comes, you
know, when, when I might settle down. I might settle down next
year, you know, or else it might be five more years from now,
maybe longer. But I like to think that I'd come back to this
town, you know. I, I think right now that it's still a fairly
good town, want to plant your roots.
Others choose to remain in the town because of a feeling of belonging
to the community, because they value the cohesiveness (although not as
tight as it once was) of White Pine and live without many of the ameni
ties of larger communities.
As indicated earlier, White Pine is an underdeveloped community.
Thw town itself, in contrast with the surrounding area, has virtually
no industry and relatively few commercial outlets. It has no motel, no
entertainment outlet, and no restaurant. It has none of these things
by design and promises to resist them in the near future, content to
maintain its "bedroom" status. Although it is not accurate to call the
town entirely unprogressive, one finds that even the young and college-
educated members of the community sometimes have reservations about the
town's growth and expansion. The college senior quoted above says,
regarding "progress,"
. . progress is on its way. It's on its way, you know, but for
some reason I can't but feel a kind of remorse, you know, . .
it's a reaction that I really do feel sad sometimes and regret and,


2
essential claim of this dissertation is that oral discourse structure
must be considered primary to written discourse, a principle which
has long been axiomatic for other levels of linguistic description
(Bloomfield 1933:21ff.). In studying oral discourse linguists have
heretofore dealt with highly formularized types (folktales and other
narratives) in English and not with spontaneous expository discourse.
This study is based on the premise that speech is hierarchical
as well as linear in structure, that it has, in other words, struc
tural units larger than the sentence and that it has structures whose
domains extend to groups of sentences. Further it leads to the
hypothesis that speech (expository in this case) has describable
paragraph structure.
Oral paragraphs may be viewed as having three different kinds
of structure: phonological, grammatical, and lexemic (Klammer 1971).
This study examines their grarmiatical structure and its second hypoth
esis is that the grammatical structure of oral expository paragraphs
is organized by certain surface structure cohesive devices (as des
cribed by Longacre and Levinsohn 1978). It investigates this hypoth
esis by examining the function(s) and distribution of the following
devices:
1) Introducers of themes into the stream of discourse. Such
introducers include left dislocation (as in "My mother, she's a
great cook"), existentials (as "There's a man here to see you"), it-
clefts (as "It was my brother who called") and WH-clefts (as "What I
did was to completely strike out").
2) Deictics (pointing words such as "that" and "this"), both


14
and feel that progress is not, is not as good maybe as a lot of
people think it is. It has its disadvantages as well as its ad
vantages, you know.
The effect of White Pine's resistance to growth and its sense of
community is that it is very slow to accept new residents, which causes
the outsider to view its people as clannish. With the accelerating im
migration into White Pine, however, of people with a diversity of back
grounds, the residents have had to learn to assimilate new neighbors
more quickly. The minister of the town's largest Baptist church moved
to White Pine from Kentucky in 1969. He describes the communitys re
ception of new people as follows:
I think the people in White Pine are very slow to accept you
within the community. They're very friendly, but a distant
friendliness. Uh they keep you at a distance, not only when
you move, but throughout your stay in White Pine to some de
gree. Uh I find the people in White Pine very, very friendly
outwardly, but uh not across the board as in some communities
They tend to stay within family groups or little cliques of
three or four families. There seems to be no broad base of,
of acceptance,...
This reluctance to accept newcomers manifests two salient traitstra
ditionalism and individualismlong considered characteristic of South
ern Appalachian culture, as pointed out by Ford (1962:11). Later in
this chapter we examine four basic traits discussed by Ford and see how
they aptly characterize White Pine as an Appalachian community. But
first we survey the economic and religious sides of White Pine's life
and then put this study into the perspective of others on Appalachian
English.


65
for other discourse types" (1977:155). Very few linguists have at
tempted to do this, however. Longacre, in his Philippine study (1968),
has been the only one to do this. He posits formulae for the gramma
tical structure of each type of discourse and paragraph. This comment
also applies to his extension of the concept of plot to types of dis
course other than narratives.
Nearly all discourse analyses have investigated only written or
very highly formularized (i.e. the sermon or the folktale) oral texts.
The discourse features which they consider are not necessarily, even for
expositions, the same as those in spontaneous oral discourses, which we
will be concerned with here. As an example, we can compare Jones (1977)
with the present work. In her study of written expositions, Jones con
tends that thematizing devices (repetition, marked word order, rhetori
cal questions, clefting and pseudoclefting, among others) can fruitfully
be studied in written texts. But all the devices she finds are ex
tremely rare in our White Pine corpus of oral expositions. The sole ex
ception to this statement is conjunctions, which are of secondary impor
tance in communicating themes. Moreover, the two most frequently used
thematic constructions from our corpus (left dislocations and existen-
tials) she does not mention. Thematic and other devices should be stud
ied, but we should not assume that they are identical for written and
oral texts.
Other analysts have suggested other types of devices as crucial to
study in a discourse context: punctuating or transitional devices (Mer
ritt 1972), various kinds of anaphora (Gutwinski 1976), equivalence


66
chains (Harris 1963), and those of Halliday and Hasan and of Williams
mentioned above. As valid as it may be to study each of these types,
the necessary heuristic framework for the comprehensive study of co
hesive devices in discourse is provided only in Longacre and Levin-
sohn (1978). Our approach to the paragraph structure of oral exposi
tions is largely based on the ideas in this article.
As suggested in their title, "Field Analysis of Discourse," Long-
acre and Levinsohn offer a framework for discourse analysis for lin
guists working in the field with little-studied languages (such as
those which Summer Institute of Linguistics workers frequently encoun
ter). What the present writer has chosen to do is to adopt it for
fieldwork with a much-studied languageEnglish. Longacre and Levin
sohn are determined to bridge the gap between the abstract structure
and the specific cohesive devices of a discourse but insist that the
analyst can do this only by first identifying and describing the "sur
face structure cohesive devices" and discourse constituents: "The job
of the analyst is to...look through the flesh and the skin to the
skeletal structure beneath and to perceive the fundamental structure
of the whole" (1978:105). The aims of the analyst are to show how
these cohesive devices operate and then to display an outline of the
discourse in a schematic fashion.
Longacre and Levinsohn present eight surface structure cohesive
devices which permit a discourse to be outlined:
1) The role of tense and voice;
2) Particles and affixes;


43
1) A hesitation "I don't know" which frequently occurs, as in
our excerpt, with "well." This expression buys a little time
while the speaker decides exactly how to phrase an answer.
2) An ignorance "I don't know," exemplified by the other two in
stances, which indicates the speaker's recognition of his/her
inability to answer or express him/herself.
3) A reluctance "I don't know" which is most often used as an ini
tial reply to a question in order to show an informant's imme
diate reluctance to reveal his/her thoughts on a subject. This
differs from the first "I don't know" above. This "I don't
know" is not a refusal, however; informants almost invariably
went on to give a direct answer after showing initial reluc
tance.
The preparation of the ethnography of White Pine presented above
included the major sources on the Southern Appalachian culture and re
gion (Kephart 1922, Campbell 1921, Weller 1965, Ford 1962, Vance 1935,
Hicks 1976, and others). It also included major sources on the East
Tennessee area (Folmsbee et al. 1969, Dykeman 1977, and others). Spe
cial attention was given to the few sources on Jefferson County, Ten
nessee, available in local East Tennessee libraries. From the East
Tennessee Development District office in Knoxville were obtained sum
maries of the 1970 census reports and other vital statistics on the
White Pine community. In addition, much useful information was gath
ered from the informants regarding local affairs.
The second step, taken after the preparation of as much of an eth
nography as was possible from written sources, was the researcher's


A DISCOURSE ANALYSIS OF
EXPOSITORY APPALACHIAN ENGLISH
BY
MICHAEL BRYANT MONTGOMERY
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1979

Copyright 1979
by
Michael Bryant Montgomery

ACKNOWLEDGMENTS
For their assistance to me throughout my graduate studies and
especially in undertaking the fieldwork for this dissertation I am
indebted to my parents. In many ways this dissertation would have
been impossible without them, and to them I dedicate it.
The fieldwork for this dissertation could not have been carried
out without the gracious assistance of my contact persons in White
Pine, the Atchleys — Earl, Emogene, Steve, and Philip—who enabled
me to make the necessary contacts with and to establish a rapport
with my informants. To them I will always be grateful.
I also wish to thank the people of White Pine for their hospital¬
ity. Their open-hearted welcome and open invitations to return will
long be appreciated.
To the chairman of my committee, Professor William J. Sullivan, I
am thankful for patient direction and many invaluable suggestions.
To the other members of my committee, Professors Jean Casagrande,
Robert Thomson, Roger Thompson, and Norman Markel, I am grateful for
patient tolerance of an erratic and hurried submission of my work.
My colleagues at the Department of English at the University of
Arkansas at Little Rock deserve thanks for their encouragements to
finish this work this past year.
Finally I wish to thank Professor Robert Longacre of the Summer
Institute of Linguistics for his invaluable personal advice to me.
iii

TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS. iii
ABSTRACT vi
CHAPTER I INTRODUCTION 1
General Remarks 1
The White Pine Setting 4
The Economic Life of White Pine 15
The Religious Life of White Pine 17
The Southern Appalachian Region 20
The Study of Appalachian Speech 33
Fieldwork Procedures 38
The Sample 50
Transcription Procedure 53
Fieldwork Journal and Interview Catalog 56
Explanation to Informants 57
Taping of the Interviews 58
The Value of Fieldwork 58
Notes 60
CHAPTER II BACKGROUND AND METHODOLOGY 61
Preface 61
Discourse Studies and the Approach of This Study 62
The Study of Monologue Expositions 72
Oral Paragraph and a Review of Paragraph Studies 74
Longacre's Typology of Discourse Genres 84
Notes 88
CHAPTER III THEMATIC DEVICES 89
Introduction 89
Left Dislocation 96
The Variety of Forms 97
Specifications 104
Functions 109
Relationship to Subsequent Discourse 117
Existential Sentences as Thematic Devices 124
CHAPTER IV DEICTICS AND SUMMARY DEVICES. . . 131
General Remarks 131
Left Dislocation with Deictics 136
Extrapositions with Deictics 137
iv

Page
Identifying Clauses with Deictics 139
Peak Sentences 143
CHAPTER V CONJUNCTIONS AND INTRODUCERS 154
CHAPTER VI SUMMARY AND CONCLUSIONS 161
APPENDIX A LIST OF INFORMANTS 163
APPENDIX B SAMPLE INTERVIEW 164
BIBLIOGRAPHY 172
BIOGRAPHICAL SKETCH 182
v

Abstract of Dissertation Presented to the Graduate Council of
the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
A DISCOURSE ANALYSIS OF
EXPOSITORY APPALACHIAN ENGLISH
By
Michael Bryant Montgomery
August 1979
Chairman: William J. Sullivan III
Major Department: Linguistics
This dissertation contributes to the fields of discourse anal¬
ysis and Southern Appalachian studies by describing the discourse
functions of several linguistic structures from a large sample of
spontaneous expository (explanatory and descriptive) discourses for
a cross-section of informants from a small East Tennessee hill commu¬
nity. The community, White Pine, is a typical Southern Appalachian
one.
Five different contexts are taken into consideration for the
examined linguistic structures: 1) Speaker Type (a representative sam¬
ple of forty informants was interviewed); 2) Frequency of Occurrence
(data are quantified when possible); 3) Social Situation (researcher
assumed one role with his informants); 4) Task (a standard interview
was controlled to elicit only expository discourse); and 5) Linguis¬
tic (the sentence and discourse position of all data was considered).
This study is based on the premise that speech is hierarchical
vi

as well as linear in structure, that it has, in other words, struc¬
tural units larger than the sentence. It hypothesizes in particular
that speech has paragraph structure. If this first hypothesis is
true, oral paragraphs should be describable to some degree of pre¬
cision and generality.
Paragraphs may be viewed as having phonological, lexemic, and
grammatical structure. This study examines their grammatical struc¬
ture and makes a second hypothesis: that the grammatical structure
of oral expository paragraphs is organized by certain surface struc¬
ture cohesive devices. It investigates this hypothesis by considering
the function(s) and distribution of the following devices:
1) Introducers of themes into the stream of discourse. Such
introducers include left dislocation (as in "My mother, she's a great
cook"), existentials (as "There's a man here to see you"), it-clefts
(as "It was my brother who called") and WH-clefts (as "What I did was
to completely strike out").
2) Deictics (pointing words such as "that" and "this"), both
cataphoric (forward-pointing) and anaphoric (backward-pointing).
Structures examined are that-extraposition (as "That is good that John
finally left"), identifying sentences (as "That's what I meant"), and
peak sentences (as "That's the best thing is to leave early").
3) Introducers and conjunctions ("and," "but," "now," "then,"
"like," and others) when they accompany left dislocation and assist
in introducing themes.
vii

All structures are examined with the view of identifying what is
typical about them, which is why they are quantified numerically when¬
ever possible. This is to insure that the data are not chance findings
of random observation.
It was found that spontaneous expository monologues contain super-
sentential units that are thematically unified and that various surface
structure cohesive devices play significant roles in creating such
unity for the benefit of a hearer. These units are shown to be oral
paragraphs.

CHAPTER 1
INTRODUCTION
General Comments
This dissertation describes the discourse functions of several
linguistic structures from a large sample of spontaneous expository
discourses for a cross-section of informants from a small East Ten¬
nessee hill community. The community, White Pine, is viewed as ty¬
pically Southern Appalachian. This dissertation thus contributes to
both discourse analysis and Southern Appalachian studies.
This study is based on a set of forty interviews conducted by the
researcher between December 1977 and February 1978 and averaging
slightly more than an hour each in length (forty-five hours total).
The interviews consisted of a small core of questions and were de¬
signed to elicit large amounts of expository (explanatory and descrip¬
tive) discourse from the informants in the form of spontaneous mono¬
logues. Subsequently the interviews were transcribed orthographically
and constitute a 1,070 page typed corpus from which data in this dis¬
sertation are taken.
Language, as a social instrument, consists of a set of systematiz-
able patterns of social behavior. These patterns, although presumed
to be repetitive and systematizable on all levels, have been largely
ignored for English at levels above that of the sentence. The
1

2
essential claim of this dissertation is that oral discourse structure
must be considered primary to written discourse, a principle which
has long been axiomatic for other levels of linguistic description
(Bloomfield 1933:21ff.). In studying oral discourse linguists have
heretofore dealt with highly formularized types (folktales and other
narratives) in English and not with spontaneous expository discourse.
This study is based on the premise that speech is hierarchical
as well as linear in structure, that it has, in other words, struc¬
tural units larger than the sentence and that it has structures whose
domains extend to groups of sentences. Further it leads to the
hypothesis that speech (expository in this case) has describable
paragraph structure.
Oral paragraphs may be viewed as having three different kinds
of structure: phonological, grammatical, and lexemic (Klammer 1971).
This study examines their granmatical structure and its second hypoth¬
esis is that the grammatical structure of oral expository paragraphs
is organized by certain surface structure cohesive devices (as des¬
cribed by Longacre and Levinsohn 1978). It investigates this hypoth¬
esis by examining the function(s) and distribution of the following
devices:
1) Introducers of themes into the stream of discourse. Such
introducers include left dislocation (as in "My mother, she's a
great cook"), existentials (as "There's a man here to see you"), it-
clefts (as "It was my brother who called") and WH-clefts (as "What I
did was to completely strike out").
2) Deictics (pointing words such as "that" and "this"), both

3
cataphoric (forward-pointing) and anaphoric (backward-pointing).
Structures examined are that-exptraposition (as “That is good that
John finally left"), identifying sentences (as “That's what I meant"),
and peak sentences (as "That's the best thing is to leave early").
3)Introducers and conjunctions ("and," "but," "now," "then,"
“like," and others) when they accompany left dislocation and assist
in introducing themes.
All devices are examined with the view of identifying their
typical function(s) and distribution within greater discourse block
structures. Hence they are quantified numerically whenever possible.
This is to insure that the data are not chance findings of random
observation. Also to insure the nature of the data gathered, five
different parameters are controlled in this study:
1) The speaker context (a representative sample of forty infor¬
mants was interviewed);
2) The frequency context (data are quantified whenever possible);
3) Social context (a matter of controlling the speech style
elicited; the fieldworker-researcher assumed one role with all infor¬
mants) ;
4) Task control (interview was controlled to elicit only explana¬
tions and descriptions from informants);
5) Linguistic context (the sentence and discourse position of
all data was taken into account).
For the most part (except the fourth and, in part, the fifth),
these have been controlled in the studies of Labov. Labov

4
(1966, 1972) has controlled the first two by adopting random sampling
techniques and quantitative analysis of his data. He has controlled
the third by deliberately eliciting several different speech styles
(four in his New York City investigations). And he has taken into ac¬
count the sentence context of data by devising the variable rule (Labov
1969). The present study, however, controls both the task context and
the discourse position context. Recent studies in the ethnography of
communication approach take the task and discourse position into ac¬
count but tend to ignore the others.
This study, in investigating a selection of surface structure co¬
hesive devices, found that they were to a significant extent used by
speakers to communicate thematically unified blocks of information to
hearers and that they reflected the grammatical structure of para¬
graphs. This confirms the hypothesis that oral expository speech
(Southern Appalachian in particular) has describable paragraph and
discourse structure.
The White Pine Setting
The small East Tennessee hill town of White Pine (White P'ne to
its inhabitants), in which the interviews for this dissertation were
conducted, is no longer the rural village it was only a little more
than a generation ago, and no longer is farming the principal occupa¬
tion of its residents. Today the town of White Pine is primarily res-
idéntial, a "bedroom" community (in 1970, eight times as many of its
people derived their livelihoods from manufacturing as from farming),

5
and a majority of its citizens are supported by employment in facto¬
ries which located in the surrounding area after the Second World War
because of the labor surplus and the low taxes. The factories re¬
tarded the emigration so disastrously prevalent throughout East Ten¬
nessee as well as other areas of Southern Appalachia in the past two
generations and attracted workers from the foothill and mountain areas
farther east, so that White Pine's population has continued to grow
steadily since the war (from 1,035 in 1950 to 1,530 in 1970 to 1,622
in 1975 to—unofficially—1 ,830 in 1977). It remains, however, a
cozy, closeknit (many would say "clannish") community despite the in¬
flux of new residents in recent years. Never has it needed more than
one stoplight, and its mayor still dutifully accepts his annual one
hundred fifty dollar check for his thirty hour a week job.
In this study, "White Pine" designates the White Pine community,
consisting of the town and adjacent neighborhoods and having a popula¬
tion of approximately 3,000 persons. It is the White Pine community
wherein the forty interviews for this study were conducted.
White Pine (refer to the map on page 6) lies at the extreme
northeastern corner of Jefferson County, forty miles east northeast
of Knoxville and within view of the Great Smoky Mountains. Seven
miles north on Highway 25E is Morristown (population 30,000+) and
five miles south is Cocke County (infamous for its rumrunning in the
20’s and sometimes called the "Moonshining Capital of the World" dur¬
ing that era). East of White Pine are its sister and erstwhile rival,
but incorporated, community Leadvale and the Nolichucky River; south
and west is Douglas Lake, a Tennessee Valley Authority reservoir.

6
Figure 1.1 Jefferson County, Tennessee
(Source: East Tennessee Development District)

7
Jefferson County (1960 population 21,493, 1970 population 24,940)
lies in the ridge and valley section of the Great Valley of East Ten¬
nessee between the Cumberland Plateau and the Blue Ridge. Today it is
sandwiched between the TVA reservoirs, Cherokee Lake (the dammed Hol-
ston River) and Douglas Lake (the dammed French Broad River) built in
the 1930's. The county is one of the earliest settled areas of what
was to become the state of Tennessee. The Holston, which formed the
county's northern border, brought settlers in from the northeast,
from upper East Tennessee and from Virginia (through the Shenandoah
Valley), while the French Broad, which largely determined Jefferson
County's southern extent and which flows west from Asheville after
flowing north, routed newcomers from the east into the area. Many of
these had received dispensations of land for Revolutionary services.
The county's first white residents arrived in 1783, thirteen years
before Tennessee became a state. The county seat, Dandridge, was
incorporated in 1793.
Settlement in the White Pine vicinity goes back nearly two hun¬
dred years, as shown by the fact that two of its present-day churches,
Westminister Presbyterian and Beth Car Methodist, were first organized
in 1787. Although the town of White Pine was not incorporated and
named until 1893, it achieved the status of a community and the char¬
acter of a town (and was called "Dandridge Crossing") when a railroad
stop and a post office were established there shortly after the Civil War.
A number of today's White Pine residents can trace back their ancestries

8
many generations in White Pine itself; some are descendants of the
original settlers in the area.
The rolling terrain of the White Pine area is better than average
farm and pasture land, with corn, tobacco, and wheat being the staple
crops. Although not in the mountains itself, White Pine lies in the
center of the region referred to as Southern Appalachia and is in many
respects a typical small town of the region. Although it is sometimes
supposed that all Appalachian communities are socially and culturally
homogeneous, only the extremely isolated areas actually are. There¬
fore White Pine is perhaps more typical of Southern Appalachia because
it is neither homogeneous nor isolated.
To the outsider, White Pine seems a slow-paced and unassuming
small town, all of whose residents know one another. It seems to be a
town with few pretentions. On a not atypical day one can get a good
flavor of the town's personality by visiting the town's center of ac¬
tivity and one of its chief public forums, the drugstore at Main and
Maple, where the mayor, the chief of police, one or two city alderman,
the town's physician, and a local minister hobnob over coffee.
Such an appearance would be rather deceptive, for in recent years
White Pine has entered a period of great transition socially and eco¬
nomically. It has also considerably changed in the way it views it¬
self (the near doubling of the town's population since 1960 through
annexation and immigration would suggest this anyway). Life has be¬
come more complex for its residents as the changes have come. For
such a long time White Pine resisted change of any kind that many of

9
its citizens feel that the current pace of change in housing, in the
economic pressures on the town, and in the arrival of new people de¬
manding more from the town than it can offer is accelerating out of
control. Not too many years ago White Pine residents did pretty much
know everyone else in town, but with new residents recently moving in
faster than ever, they feel uneasy and sense a lack of cohesiveness
they once took for granted. Many of the longtime residents have a
strong sense of the symbiosis of the community, of how everyone plays
a part in supporting the general welfare. A typical expression of
this was made by a thirty nine year old female:
It's a feeling of closeness and connentuity [sic]. If I could
draw you a picture, I would draw you a big ball of people all
wrapped up together, like worms. You know, like you're fishing
with worms, and all those worms are intertwined. That's the
way this town is.
For many, though, this sense of closeness and symbiosis is inseparable
from the "traditional" way of doing things, so that change is threat¬
ening.
Probably the most important event in the past fifty years for
White Pine and East Tennessee is the arrival of TVA. For White Pine,
the coming of TVA has in a sense been responsible for most of the
changes during the past forty-five years (TVA began in 1933, and the
construction of Douglas Dam near White Pine began shortly thereafter).
TVA's initial effect was to dislocate many farmers in Jefferson County
from good bottom land and to resettle them, which was the cause of some
resentment. Today few people in the county regret the changes TVA made

10
possible. As one sixty-year-old farmer who witnessed the relocation
of the thirties put it,
...the landowners were against it. They didn't want uh, they
didn't want to move out....Of course, we realize now that TVA
is our, our main uh source of energy and power and so, I don't
know what we'd do without it.
TVA, through its water management and its provision of electricity, set
the stage for the ensuing economic development of the area, for the
profitability of agriculture, and especially for the coming of indus¬
try. Since the war several dozen major factories have relocated to
East Tennessee.
Until after the Second World War and the locating of the facto¬
ries White Pine was very much an isolated farming community. Its
sphere of interest hardly extended beyond its boundaries. A half dozen
stores provided the necessities and there was little pressure to de¬
velop the town residentially, commercially, or industrially. The sen¬
timent was to keep the town as quiet as possible and to keep any con¬
scious development out. This is how the state of affairs in White Pine
remained, too, until only a few years ago. Eventually the preoccupa¬
tion with prohibiting development was buffeted by demands for more
housing, for recreational facilities, and for planning in general, so
the town has had to undergo considerable growing pains in meeting these
demands. These pains will persist for a long time because the pres¬
sures, primarily from the younger generation, will continue.
The most profound development in White Pine over the past thirty
years has been its evolution into a "bedroom" or almost strictly

11
residential community. Until a generation ago, most people in White
Pine grew up on the farm, but with the coming of industry came rising
economic expectations and consequently a revolution of rising demands
for private residential housing. These demands have partially been met
by the parceling up of farmland into lots and a mobile home park, but
as another generation has grown up, it has become acutely obvious that
the town especially lacks social activities.
In White Pine the principal social institutions are the churches
(ten in the community), and more than likely any social get-together or
meeting will be church-sponsored. Although there are several organized
civic groups (Ruritans, Lions, Beta Sigma Phi) in town, there is little
regular social activity. The occasional special event (church revival,
political candidate speech) may attract a group of citizens, and the
annual Pine Festival is an event that meets some of the social needs of
the town.
The Pine Festival, sponsored by the local Beta Sigma Phi and begun
in 1967, is a one-day collection of shows, contests, and displays held
on July 4 each year. The festival has become the focal event of the
community's life because it is the one occasion when a majority of its
people take part. The festival integrates the community and gives it
a greater sense of self-identity than it has ever had because it is hea¬
vily patronized by people from surrounding counties and communities.
Although there is tremendous socioeconomic disparity between the
very rich and the very poor in the community, one finds little overt
residential segregation by economic status. As the thirty-nine-year-

12
old female quoted earlier put it,
...there are very, very wealthy people in this town, and there
are very, very poor people in this town. But to save my life,
I can't tell you what the difference is. I mean I can't point
to one person and say "Okay, this is a poor person and this is
a wealthy person.” They, they blend.
This sentiment evidences the symbiosis referred to earlier.
One major division in the community is that between the middle
and older generations on the one hand and a large part of the younger
generation on the other. This division is a product of the rising ex¬
pectations of the town referred to above. In general the community is
more optimistic about its future than a generation ago because of im¬
proved local education. A comprehensive high school has been added and
a junior college has been opened five miles up the road. A more diver¬
sified job market has also developed. But most youths with professional
and career aspirations still see inadequate prospects in the nearby
area. Many of them, once they leave White Pine, are embarrassed to be
from the town. Often White Pine is for them a place to escape, a place
that can and does trap its ambitious youth. One twenty-nine-year-old
male characterized the town's subtle pressures not to aspire:
...most of the people prior to uh the sixties, when I got out,
they, they were settled. The idea was to, when you got out of
that high school was to get married, cause marriage was stabi¬
lization. You got married, you had children right then, as
quick as you could, because that made stay at home. You didn't
carouse. You didn't go on. You weren't apt to, to leave the
area and get involved in things that way.
This desire to escape is usually tempered by an attachment to the
familiar surroundings and personality of the town, though, so that some

13
of its college-educated citizens settle in town despite its lack of
opportunities. One twenty-five-year-old male college senior puts it
this way:
It's hard to say, you know, when the settling down comes, you
know, when, when I might settle down. I might settle down next
year, you know, or else it might be five more years from now,
maybe longer. But I like to think that I'd come back to this
town, you know. I, I think right now that it's still a fairly
good town, want to plant your roots.
Others choose to remain in the town because of a feeling of belonging
to the community, because they value the cohesiveness (although not as
tight as it once was) of White Pine and live without many of the ameni¬
ties of larger communities.
As indicated earlier, White Pine is an underdeveloped community.
Thw town itself, in contrast with the surrounding area, has virtually
no industry and relatively few commercial outlets. It has no motel, no
entertainment outlet, and no restaurant. It has none of these things
by design and promises to resist them in the near future, content to
maintain its "bedroom" status. Although it is not accurate to call the
town entirely unprogressive, one finds that even the young and college-
educated members of the community sometimes have reservations about the
town's growth and expansion. The college senior quoted above says,
regarding "progress,"
. . . progress is on its way. It's on its way, you know, but for
some reason I can't but feel a kind of remorse, you know, . . .
it's a reaction that I really do feel sad sometimes and regret and,

14
and feel that progress is not, is not as good maybe as a lot of
people think it is. It has its disadvantages as well as its ad¬
vantages, you know.
The effect of White Pine's rosistance to growth and its sense of
community is that it is very slow to accept new residents, which causes
the outsider to view its people as clannish. With the accelerating im¬
migration into White Pine, however, of people with a diversity of back¬
grounds, the residents have had to learn to assimilate new neighbors
more quickly. The minister of the town's largest Baptist church moved
to White Pine from Kentucky in 1969. He describes the community’s re¬
ception of new people as follows:
I think the people in White Pine are very slow to accept you
within the community. They're very friendly, but a distant
friendliness. Uh they keep you at a distance, not only when
you move, but throughout your stay in White Pine to some de¬
gree. Uh I find the people in White Pine very, very friendly
outwardly, but uh not across the board as in some communities
They tend to stay within family groups or little cliques of
three or four families. There seems to be no broad base of,
of acceptance,...
This reluctance to accept newcomers manifests two salient traits—tra¬
ditionalism and individualism—long considered characteristic of South¬
ern Appalachian culture, as pointed out by Ford (1962:11). Later in
this chapter we examine four basic traits discussed by Ford and see how
they aptly characterize White Pine as an Appalachian community. But
first we survey the economic and religious sides of White Pine's life
and then put this study into the perspective of others on Appalachian
English.

15
The Economic Life of White Pine
White Pine has made significant economic strides in the past two
decades and has undergone a thorough economic transformation in the
past generation, but it remains in many important ways undeveloped.
The percentage of its population below the poverty line (19.4) is well
above the national average (approximately 12). Host of its citizens
must look beyond the city limits for their livelihoods.
Fortunately the surrounding ten mile radius of White Pine has been
able to provide considerable employment opportunities for the last
three decades. This employment has not only stemmed but also largely
reversed the serious problem of emigration of the labor force which
plagued East Tennessee and White Pine in the 30's, 40's, and 50‘s, al¬
though it was not quite as severe as elsewhere in Appalachia, espe¬
cially in the coal mining regions of Kentucky and West Virginia. This
reversal has largely been accomplished through the benefits accruing
from the development of TVA. In the 1940's, Jefferson County's emigra¬
tion rate surpassed ten percent of its eligible adult work force, and
in the 1950's it was almost as high. But in the 1960's, immigrants
began to outnumber emigrants and the migration patterns underwent a
dramatic turnaround. In the 1970's, the county and especially White
Pine have seen a significant infusion of new residents and a decrease
in emigration. Employment prospects for incoming members of the job
market, although still limited, are better than twenty or thirty years
ago.
The dislocations brought about by the emigration over the years
have left their marks, though. White Pine has lost much of its Black

16
population (only 1.7 percent of its 1970 population was Black), and
many families have been split because the children have had to mi¬
grate north to find jobs. This need to leave home for employment
was aggravated for many natives of East Tennessee and Southern Appa¬
lachia by their inability to adjust to the urban North and their fond
attachment to home in the hills. Instances of their choosing to return
to certain unemployment or underemployment in the hills have been
frequently documented, as by Hicks (1976:29-30). This is perhaps not
so common in White Pine as in other areas in the region. A similar
attachment to native ground continues today; relatively few graduating
high school students and other new members of the labor force wish to
settle down elsewhere.
Almost entirely responsible for the improved economic profile for
Jefferson County and White Pine has been the tremendous growth of
local factories, especially those of the American Enka Corporation and
several major furniture firms. Attracting such industry to the area
have been the labor surplus, various tax incentives, and the availa¬
bility of cheap electricity. By 1970, more than half (51 percent) of
White Pine's employed population was in manufacturing, eight times as
many as in agriculture, forestry, and fisheries combined (6.3 percent).
In the same year, trade, service, and finance supported another 23.1
percent of White Pine. Construction employed 9.1 percent, mining 0.4
percent, and all other types of employment, 10.0 percent. The move¬
ment of workers into manufacturing has entailed the dissolution of the

17
tenant farming system and the development of more specialized agri¬
culture in the area.
This is not to say that farming is no longer a major concern to
most of the people of White Pine. It is. It must be remembered that
most of its citizens are a generation or less off the farm; many of
them still religiously tend kitchen gardens of an acre or two. But
few citizens must still depend on it as their primary source of income.
The new sources of income, the factory payroll primarily, have
brought in their wake profound changes in White Pine (above p. 10).
The better income is more dependable than one might derive from farm¬
ing. It has caused White Pine to become a bedroom community. At the
same time, the improved standard of living has raised the optimism of
White Pine's citizens for their future and especially for the future of
their children. For those wishing to break the cycle of having to go
to work immediately after leaving public school, there are more oppor¬
tunities than ever before. Within the past decade a junior college has
opened only five miles away, and many of White Pine's youth have de¬
cided to undertake some college work rather than enter the job market
right away. The area's economy, although still not providing suffi¬
cient skilled positions, continues healthy at the same time that White
Pine, a self-sufficient community a generation ago, finds itself more
and more dependent on the surrounding area.
The Religious Life of White Pine
The community of White Pine lies in the region often referred to
as the "Bible Belt," so called because much of its population still

18
adheres to fundamentalist beliefs in the Bible and because of the role
of its clergy as both "preachers of the Word" and as community spokes¬
men commanding political and social influence. The community of White
Pine is typical of the region. Not only are the churches by far the
most prominent institutions in the community, but also religious values
permeate so thoroughly every aspect of its life that it is simply quite
impossible to deal meaningfully with most aspects of the community's
existence without taking them into consideration.
Churches have long been numerous in Appalachia. A 1935 report of
the United States Department of Agriculture reported there were in 1926
as many as 3.25 churches for each 1,000 inhabitants. It might be
thought that such a high ratio would be accounted for by the fact that
few people could motor to church fifty years ago. But the present-day
ratio in White Pine is the same: ten churches for its approximately
3,000 inhabitants (3.3 per 1,000 people). Church membership, if not
necessarily participation in church activities, is almost required for
a White Pine citizen. Perhaps as much as eighty percent of the popula¬
tion in the community are church members.
The two largest churches in White Pine, one a Southern Baptist and
the other a Missionary Baptist, have approximately one thousand members
between them, and the other eight are by no means all small congrega¬
tions. Besides five Baptist churches (of several different affilia¬
tions), there are two Methodist, two Pentecostal, and one Presbyterian
church. There is not a single non-Protestant church in the entire
county.

19
Throughout all of the churches one finds a firm fundamentalism
(Puritanical in kind for most churches). One finds also a strong com¬
mitment to evangelistic activity and to the need for the individual to
undergo a conversion experience and for the congregation to support
missionary activities and frequent week-long revivals.
All the White Pine inhabitants refer to each one of the ten min¬
isters in the community as "Preacher" regardless of an inhabitant's own
church affiliation. Undoubtedly this reflects the fact that until a
generation ago, local churches usually did not have their own ministers.
Rather, they shared a minister with one or more churches outside the
community (the days of the "circuit riders"). It also seems to reflect
the present-day notion that preaching the Word of God is the most impor¬
tant duty a minister has. The epithet thus continues even though cler¬
gymen have adopted all the everyday duties of the profession (counsel¬
ing, visitation, etc.) in recent years. Preachers are also invested
with a great deal of moral authority in the community. They can wield
considerable political influence as well, if they choose, on many is¬
sues of local import such as the granting of beer permits or of permits
for businesses to open on Sunday. These are quite important issues in
White Pine.
One finds little evidence of sectarianism in White Pine. There
seems little difference between the Baptists, the Methodists, and the
Presbyterians in either their theology or in their way of conducting
business. Indeed, they coexist quite harmoniously. One's church af¬
filiation is viewed generally as dependent on the affiliation of one's
parents or one's spouse, not on one's personal convictions.

20
The foregoing does not strictly hold for the Pentecostals, who do
consider themselves to an extent to be a people apart. The extremes
of Pentecostal ism in Southern Appalachia have long been pointed out
(as in La Barre's study (1964) on the snake-handling cult). One can
find congregations of "snake handlers" not more than twenty miles east
of White Pine in Cocke County. But people in the area overwhelmingly
view such practices as quite objectionable and offensive.
The primary functions of the church itself are to provide relig¬
ious services and training for its members, yet in a community where
such a large percentage of the inhabitants are members, churches inevi¬
tably offer more than the fulfillment of spiritual needs. Churches in
White Pine provide most of the organized social activities. Scheduled
weekend activities are relatively infrequent in White Pine, but any
given activity is more than likely a church-related one. Since 1975,
when the county's high schools were consolidated, the only organized
youth activities in White Pine are those that are church-sponsored.
In short, then, the influence of the organized religion in White
Pine is pervasive. The churces are viewed as having an indispensable
and rightful place in the life of the corrcnunity and in the life of the
individual.
The Southern Appalachian Region
Alghough the Appalachian Mountains extend thirteen hundred miles
from Vermont to Alabama, the region usually called "Appalachia" and

21
viewed as sharing certain distinctive cultural characteristics and
social and economic conditions encompasses the southern half of the
mountain range and includes as its core a sliver of North Georgia, the
eastern thirds of Tennessee and Kentucky, the western thirds of
Virginia and North Carolina, and the entire state of West Virginia.
This is in addition to a variety of fringe sections, depending upon
which demarcation of the region one chooses. Perhaps the most exten¬
sive demarcation is the political one made by the Appalachian Regional
Commission. The Commission defines Appalachia by county, based on
economic conditions. Their "Appalachia" includes parts of Alabama,
Mississippi, South Carolina, Ohio, Maryland, Pennsylvania, and New
York, in addition to the core area mentioned above.
Gastil, whose regions are determined by sociocultural features,
considers the core area outlined above and a small corner of South
Carolina to be the "Mountain" area of the South (Gastil 1975:174ff).
Vance (1962) views the "Southern Appalachians" as a distinctive region
sharing agricultural and economic conditions, which covers parts of
West Virginia, Kentucky, Tennessee, North Carolina, South Carolina,
Georgia, and Alabama. The demarcation of the "Southern Highland" re¬
gion by John C. Campbell (see map on the following page), a lifelong
student of the region, comprises the core area of portions of six states
and also the fringes of three others, a 112,000 square mile territory.
For Campbell, this area exhibits similar economic, educational, and ref-
igious characteristics. What is important for us is that, whatever the
delimitation of Southern Appalachia, the community of White Pine has always

22
Figure 1.2 Campbell's Southern Highland Region
(Source: Campbell 1921:xxxviii)

23
been considered in its heart. Further reason for designating White
Pine as an "Appalachian" community is presented below.
The linguistic analogue of Southern Appalachia is the South Mid¬
land area (Kurath 1949, Kurath and McDavid 1961), which encompasses
the mountain districts and also most of Tennessee and considerable por¬
tions of Mississippi, Alabama, and Georgia. This area, determined by a
combination of lexical, phonological, and morphological isoglosses, has
been traditionally viewed as more closely related linguistically to the
North Midland area (Pennsylvania and parts of Maryland, Delaware, and
New Jersey) than to the South.
The question of the relative affinity of linguistic and other
traits in Southern Appalachia (or the South Midland) to regions north
and south is still very much unanswered, in fact, but it will not overly
concern us here. In a provocative reassessment of the linguistic dis¬
tinctiveness of the area, Bailey (1968) concludes there is no "Midland"
area at all which has a core of distinctive features and that North Mid¬
land should be retermed 'Lower Northern' while South Midland be redesig¬
nated 'Outer Southern.' Gastil views the cultural affinities as
lying closer to the South: "Culturally the mountains are an extreme
version of the Upland [South]" (1975:194).
All this is not to say, of course, that even the most restricted
definition of Southern Appalachia implies that it is socially homogene¬
ous or that its people have a common background because they now may
share certain cultural or linguistic features or even a common outlook
on life. Yet commentators on the area, both native and non-native,

24
infer that its inhabitants are alike in ancestry and cultural back¬
ground, especially on the basis of a number of linguistic relics
apparently distinctive of its people. The most extreme forms of such
inferences issue an assertions that Appalachian people are "Elizabe¬
than" or that they preserve the "purest Anglo-Saxon blood in the United
States" or that they are "almost pure Scotch-Irish" (note the contra¬
diction here). These inferences have formed a part of the chronic
misconceptions about the inhabitants of Southern Appalachia.
Even such a serious commentator as Campbell has abetted such
inferences by his statement that Southern Appalachian people form "a
definite . . . racial group" (1921:xiv), although he conceived of
their having developed into such a distinct group of people by virtue
of "their common interests, hardships, and struggles" (p. 71) and not
by virtue of their lineal descent from one stock of Old World people.
His lengthy study of the surnames of early mountain settlers revealed
almost equal proportions of Scotch-Irish and English names, and a
large number of Germans as well:
Without a doubt the Scotch-Irish and English elements are the
strongest in the mountain population, though the Highland
people are not different from the Lowland Southerners in this
respect. The Scotch-Irish strain is strongest in some mountain
sections, the English in others; and in some communities may be
surmised an influence of German ancestry, (p. 71)
Several other groups, most especially the French Huguenots, were numer¬
ous among the early settlers in Appalachia too. Thus the region's
present-day descendants are a hybrid, whencever came their language (a

25
handful of its features have indeed been shown to be carryovers from
Northern England and Scotland) or its musical or other traditions.
For a variety of unscientific and often impressionistic reasons,
however, the people of Southern Appalachia are reputed to be quite dif¬
ferent from all other Americans. If they do not descend "directly from
the Scotch-Irish,” there is a quaintness in their approach to life, it
is said, harking back to an earlier era, or their speech has a quality
and character indisputably unique. Not only does Appalachian speech
preserve various relics, that is, but it also has a distinctive quality
in a more general sense. The literature on the subject abounds in de¬
scriptions of how mountain folk talk, more often than not written by
college-educated children remembering fondly earlier days in the hills.
Here are three representative descriptions:
There are certain peculiarities of enunciation which it might be
well to speak of here. For the Smoky Mountaineer the nose is as
much an organ of speech as the larynx. This is particularly no¬
ticeable in old people, and may be because of the catarrh
brought on by constant exposure and diseased tonsils. The moun¬
taineer drawls to an alarming extent, even more so than his
neighbors of the plantations. His voice is utterly without ca¬
dence, almost a droning monotone. On the other hand, it has a
deep, resonant quality which catarrh and embarrassment never
wholly obliterate. (Walker 1939:3-4)
The mountaineer has clung to the Shakespearean word these many
years because he has not learned the modern word. Indeed, he
has had little contact with those who speak the modern word; and
since his surroundings and habits and thoughts have been largely
the same as they were in Shakespeare's day, he has had no need
for new or changed speech. (Coleman 1936:30)
As a race, the people of our southern mountains speak softly.
Even their most commonplace remarks somehow are made to sound
secret, and of greatest importance. The women especially speak
softly, their voices often plaintively sweet, pitched in a
minor key and lifting upward, so that the last word of every
sentence is almost sung. (Hannum 1969:29)

26
The contention in these excerpts is that the speech of Southern Appa¬
lachian people has a marked flavor and character all its own, beyond
the presence of archaic expressions and words therein. Commentators
have long been at pains to account for this on the basis of one extra-
linguistic factor or another (isolation, pace of life, etc.). Cratis
Williams, lifelong resident of the mountains, attributes in part the
distinctiveness of the way people in the region speak to "the highland¬
er's habit of thrusting his chin forward rigidly when he speaks."
(1961:10) Brewer and Brandes try to explain why Southern Appalachian
English impresses outsiders as so different by describing the kinesics
of its speakers, detailing the posture and the eye and body contact of
the typical speaker in the region.
Almost all the studies and comments on the speech of Southern Ap¬
palachia have had the unfortunate effect, however, of lumping the en¬
tire region together linguistically and assuming that it is homogeneous
within itself as well as distinct from other regions of the country.
To this date, very little attention has been devoted to regional lin¬
guistic variation within Appalachia. Only the two linguistic atlas
projects, the Linguistic Atlas of the Middle and South Atlantic States
and the Linguistic Atlas of the Gulf States, have given significant
consideration to such variation. This presumption of underlying uni¬
formity has been espoused frequently, as in Carpenter (1933). He con¬
tends that the only "variation in the vernacular of the southern and
central Appalachians has been due to the difference in the degree of
isolation of the various sections" (p. 22). Most writers ignore the

27
subject of geographical variation entirely. This creates the impres¬
sion that there is none. No attention was given until very recently
to variation by age, socioeconomic background, or any other variable.
This presumption of little, if any, variation is directly at odds
with the view of Southern Appalachian speakers themselves. They be¬
lieve that Southern Appalachia has much subregional variation. Speak¬
ers in the region as a rule (and White Pine exemplifies this well) con
tend that they speak differently from neighboring communities only a
few miles down the road or one or two hills over, although it is diffi
cult to tell what differences they have in mind. These subjective
views have occasionally been supported by statements in the linguistic
literature, as by Berrey: "the dialect may vary slightly with the lo¬
cality, and even from family to family" (1940:46), or by Walker: "It
must be realized that the speech of the mountaineer varies from one
district to another" (1939:6). Wolfram and Christian, in commenting
on vowel variation in southern West Virginia, state that "there are
apparent vowel differences from region to region within Appalachia.
Even within the restricted locale we have studied here, there is evi¬
dence that several different vowel systems must be recognized"
(1976:64).
To what extent the term "Appalachian English" refers to a deter¬
minable linguistic entity is then unsettled. On the one hand, some
commentators and Appalachian people recognize an almost hollow-to-hol-
low and community-to-community variation. On the other, such long¬
time students of American English as Mencken (1963) and Wise (1957)

28
consider the dialect of the Southern Appalachian area distinct and
identifiable. One reasonable approach to the issue of whether the
term "Appalachian English" should be used is that of Wolfram, who
adopts the term and applies it to a "unique combination of linguistic
characteristics rather than . . . structures as such" (Wolfram 1977:
97). For him, Appalachian English refers to a common denominator of
non-unique features which aré still subject to variation for any given
district within Appalachia or for subgroups within such a district.
Since this study is designated as one on "Appalachian" speech,
this must now be justified and explained. This study does not inves¬
tigate linguistic features dealt with systematically in earlier studies
on the speech of the Southern Appalachian region or of the speech of
any other variety of American English. It thus lacks the comparability
of evidence required to distinguish the speech of White Pine as "Appa¬
lachian" solely on the basis of linguistic criteria, although nearly
all the features—-phonological, lexical, and grammatical—usually
termed "Appalachian," as by Brewer and Brandes (1977), are found in
the speech of White Pine inhabitants. Nor does it adopt the design¬
ation solely on the basis of White Pine's geographical location. If
it did, "East Tennessee" would have been used. Rather, the speech of
the inhabitants of White Pine is called "Appalachian" because the
community is, ethnographically speaking, an Appalachian community, on
the basis of several of those salient characteristics which have long
been associated with Southern Appalachia—individualism and self-
reliance, traditionalism, fatalism, and fundamentalist religion
(Ford 1962:1 Iff.).

29
Throughout the literature on Southern Appalachia runs the view
that its people and culture are distinctive in at least a broad sense.
Most comnentators on the region point out that a combination of such
traits as "fierce" individualism and mistrustfulness of all strangers,
among others, characterize the Southern Appalachian native and make him/
her different from other Americans. White Pine is more of a valley than
a mountain community, and such characterizations have been more accu¬
rate of the isolated communities in the region than of the urban and
valley sections of Southern Appalachia. This has sometimes been recog¬
nized, as by John C. Campbell, who lived and taught in the region for a
quarter century. Campbell was concerned that some of his statements in
his The Southern Highlander and His Homeland would not be understood as
"applicable to the remote rural folk who were the particular object of
his study [and] were not true of their urban and valley kinsfolk"
(1921:xiv). But such characterizations, while sometimes extreme,
rarely claim that Southern Appalachia is culturally or socially homoge¬
neous. White Pine itself is certainly neither. Yet several broad cul¬
tural traits have for so long been pointed out as typical of Southern
Appalachian people that it is worthwhile to determine if they are ap¬
plicable to White Pine and can enable us better to understand it as a
Southern Appalachian community. We will ocnsider the four traits or
themes mentioned above that Ford chose to examine by conducting a
region-wide survey of Southern Appalachia. Ford found these traits to
be significantly apposite to both urban and metropolitan communities,
although less so than to rural communities.

30
Ford surveyed nearly 1,500 homes in a region closely coterminous
with Campbell's Southern Highland region (page 22 above) in three dif¬
ferent kinds of community—rural, urban, and metropolitan—with a
questionnaire devised to indicate the strength of four prominent
traits—individualism and self-reliance, traditionalism, fatalism, and
fundamentalist religion. It was hypothesized that, because of the eco¬
nomic and social changes which had taken place after the Second World
War, a disparity between rural and non-rural manifestation of these
traits would be found. Although he did find some disparity, his results
show that the four traits he chose form a common denominator for Southern
Appalachia in general. Ford outlines his choice of traits in this man¬
ner:
In examining the web of mountain life, one finds these themes
intertwined and generally, though not always, mutually support¬
ing. Most so-called "mountain traits" are to be found in one
form or another throughout the nation, particularly in rural
areas. At the same time, each of them has its antithesis in
contemporary industrial society. The self-reliant individual¬
ist, at least as an "ideal type," stands at the far end of the
scale from the much berated "organization man." Traditionalism,
not only in the sense of clinging to an earlier heritage but
also in the exaltation of resistance to social change, is viewed
as both anachronistic and vaguely immoral by a larger society
that values progress through rational, scientific endeavor.
Even more reprehensible to a culture that stresses achievement,
self-betterment, and mastery ovor nature is a passive resigna¬
tion to one's situation in life, particularly if it is a situa¬
tion viewed as both undesirable and remediable. Less subject to
censure by the larger society, perhaps, but contrasting as
sharply with its dominant values—and not immune from ridicule—
is the rigid, pervasive religious ethos of the [Southern Appala¬
chian] Region. (p. 11)
Commentators on Southern Appalachia have long characterized its
people as individualistic, self-reliant, and non-social. Campbell

31
devotes an entire chapter (1921:90-122) to analyzing their individualism
and views their self-reliance as a result of their extreme isolation.
Horace Kephart, who lived in western North Carolina the first two dec¬
ades of this century, characterizes the people as having "fiery indi¬
vidualism" and as being "non-social" (1922:382-3) and goes on to say
Except as kinsmen or partisans they cannot pull together. Speak
to them of community of interests, try to show them the advan¬
tages of cooperation, and you might as well be proffering advice
to the North Star. (p'. 383)
More contemporary observers, Weller (1965:29ff.) in West Virginia
and Hicks (1976:39) in western North Carolina, document the same trait
among the sections of the region they observed.
When one looks at White Pine, such comments have a remarkable
applicability. The people are basically non-social, sticking to small
groups, not often unified in acting except in opposition to a proposed
local change, on a special occasion such as Pine Festival Day, or with¬
in the small sphere of the kin group. Regarding their inability to
work together, a twenty-nine-year-old male resident of White Pine says
of his fellow citizens
They, they want to know, but they don't want to come out of their
shell long enough to get enough of it to expand. They fear ex¬
pansion. This town definitely doesn't, it, one day they'll be
shocked into the fact that Morristown will try to annex them,
and there will be a total bar-, that will be the common bond
that brings them together. They will be, that will be the sup¬
portive element would be to stand against some cause.
This does not contradict the statement above (p. 9) about the cohesive¬
ness of the community; it means the cohesiveness is passive and residual.

32
The self-reliance of the region's people has another side to it, as
Ford (p. 13) points out, in their attitude of acceptance toward federal
welfare and assistance programs. Odum remarked over forty years ago that
the people of Southern Appalachia, "reputed to be the most individualis¬
tic of all the regions, they cooperate most fully with Mew Deal tech¬
niques" (1936:97). Notwithstanding scattered initial opposition (primar¬
ily by those whose lands would be claimed by the reservoirs), the people
in the White Pine area and throughout East Tennessee have accepted and
cooperated with TVA's many programs. There was little reluctance to
accept whatever federal assistance was proffered.
The traditionalism and fatalism characterizing Appalachian people
are attributed to their struggle of many generations with the harsh reali¬
ties of life in a region with few resources and little available produc¬
tive land, according to Ford (1962:16). Weller (1965:33-40) shows how
such themes flow throughout mountain songs, stories, and daily attitudes.
In White Pine it is manifest in the strong and widespread doubts about
the value of change, of progress, and of post-high school education (see
quotes above, pp. 12-3). The twenty-nine-year-old male quoted above
says, regarding the community's residents,
They're...blinded really to what's, what's going on, I think.
They don't see their youth. They don't see, they don't see life
as progressing. They just see it as, as a tolerable state to be
in, and you just tolerate it now and hope for the future, yet you
don't work for it. You don't do anything working hand in hand.
As indicated above (pp. 8-9), recent changes and pressures for more
changes have made life more complex and confusing for many in White Pine.

33
The traditional way and pace of life in the community have been con¬
sidered as something of a birthright. Many other things, such as po¬
litical and religious affiliations, are also considered immutable
birthrights. Many wish keenly to pass all this on to the next genera¬
tion.
Little additional commentary here (see pp. 17-20 above) is needed
to show how applicable is the fourth characteristic trait Ford studied,
fundamentalist religion, to White Pine. The community has no non-fun¬
damentalist churches and the great majority of its citizens are church
members. Religious fundamentalism is probably the most powerful and
pervasive force in the community.
With reference, then, to the four characteristic traits Ford claims
to be consistently typical of Southern Appalachian people, it is obvious
that White Pine should be considered a typical community in the region.
The Study of Appalachian Speech
As Wolfram notes, Mencken once pointed out that the speech of the
Southern mountains had been studied more than that of any other region
except New England (Wolfram 1977:92). However, the sum total of lit¬
erature on Appalachian English does not actually tell us much about how
people in the region speak. At best, most of the literature is little
more than anecdotal, dealing with a scattering of exotic words and pro¬
nunciations, many of which indeed are archaisms. Very few writers have
actually asserted that Shakespearean English is or was spoken in the
mountains. But many have taken extreme pains to indicate usages in Ap-

34
Appalachian speech corresponding to usages in Spenser, Chaucer, and
especially Shakespeare. This creates the impression that the language
of the mountains is little more than a carryover from a bygone era and
another country. This is especially true when a commentator devotes
little attention to anything but such analogues.
One might go so far as to speak of a "typical" treatment of Appala-
chain speech in the literature. This is represented by Berrey (1940),
Bray (1950), Carpenter (1933), Combs (1916, 1931), Watkins (1949),
West (1966), and many others. The typical treatment contains the
following:
1) A brief description of mountain culture and of the mountaineer
himself, especially mentioning the most salient traits of his
personality and how he has adopted himself to his mountain
environment.
2) Discussion of various peculiarities of pronunciation, syntax,
vocabulary, usage, onomastics, and similes and other figures
of speech and expressions, with an effort to compare such with
Old, Middle and Elizabethan English whenever possible.
3) (Optional) A few remarks on how mountain speech is/has been
passing away, despite its expressiveness and color.
Coleman (1936) is a perfect example of this typical treatment. Her
first chapter points out the ancestry of the mountain people, shows how
they developed self-reliance in their isolated habitat, emphasizes the
fierce independence of their character, and shows the low position of
women in mountain society. In her second chapter, she covers a wide

35
selection of grammatical, lexical, and pronunciation features in
mountain speech from North Georgia, v/ith accompanying citations from
sixteenth-century British literature wherever possible. In chapter
three she bewails the inevitable passing of the mountaineer with the
coming of modern education and technology, which will "wipe away for¬
ever our admirable mountaineer with his quaint and delightful manner
and speech" (p. 30).
Even well-known early scholars of the speech of the region are
given to such vague statements as "in general, most vowels may be used
interchangeably. Most vowels may replace 'e,‘ 'i,1 and 'u'" (Combs
1931:1315).
But the most fundamental deficiency of almost all studies of
Appalachian speech to date is that they provide absolutely no context
for their data. It is impossible to understand a language fully unless
we have information regarding five different kinds of context for lin¬
guistic data:
1) Which part of the speech community uses the form(s)? (Old or
yound, middle or lower class, rural or urban speakers, etc.)
2) How often is a form used? (Is it categorical or extremely
rare or the predominant usage?)
3) In what kind of social situation is the form used? (i.e., What
style of speech is involved?)
4) What communicative task is a speaker performing when a form
is used?

36
5) What is the linguistic context of the form? (Where in the sen¬
tence and in the discourse does it appear?)
Only when we have answers to all these questions can we establish what
context a form appears in and whether it is typical verbal behavior.
The history of the study of Applachian English has been chiefly a
history of the study of extreme forms, and only rarely is any context
noted for data which are cited. Despite an occasional remark that a
certain form "is still heard" or is "generally" or "frequently" used,
the clear implication of most studies is that all the speakers in the
community under consideration use the forms cited all the time and in
every social situation.
The tendency to deal with the atypical (and hence unrepresenta¬
tive) speakers and forms infects even works which are otherwise linguis¬
tically sound. This temptation is very strong. Joseph S. Hall's tho¬
rough and invaluable study of the phonetics of the speech of the few re¬
maining residents of the Great Smoky Mountain National Park was under¬
taken, he confesses, because he believes in this speech, which "has so
long been removed from the main currents of American culture, there would
remain vestiges of earlier stages of the growth of the English language"
and which also would not have "the deep impress of the schoolmaster's in¬
fluence" (1942:4). Brewer and Brandes (1977) present only the extreme
features of Appalachian English for the benefit of schoolteachers who
will teach in Appalachia.
Two recent sociolinguistic studies, Hackenberg (1972) and Wolfram
and Christian (1975, 1976), take into account the first two contexts
(who uses the form and how often is it used?), but neither controls the

37
other three in their investigations. Wolfram and Christian's study,
based on interviews made in the mountains of southern West Virginia,
classified their speakers socioeconomically and carried out quantita¬
tive analyses of many of the linguistic forms they discuss, thus pro¬
viding the speaker and frequency contexts for their data. Yet they
do not choose to base their work on a representative sample because
they were "primarily concerned with the language variety which might
be considered most divergent from some of the more mainstream varieties
of English" (1976:10). Their work is also written primarily for the
educational community and one-fifth of their book deals with the educa¬
tional implications of dialect diversity.
Hackenberg divides thirty nine residents of Nicholas County, West
Virginia, into four classes on the basis of education level attained and
occupation and quantifies the occurrence of "non-standard" verb agree¬
ment and subject relative pronoun deletion for his four classes. He
thus takes the first two contexts into consideration and even gives some
attention to the fifth context mentioned above, the linguistic context,
for his two linguistic features. The present writer demonstrates else¬
where (Montgomery 1977b) that his treatment of the linguistic context is
too simplistic.
A description of speech is ethnographically justifiable only if all
five contexts are controlled and taken into account. Only then can it
be claimed that the behavior described is typical (see the discussion of
the Malinowski/Firth "context of situation" later in this chapter).
This present study has striven to control all five. Thus it makes a
unique contribution to the study of Appalachian English.

38
This study is based on a representative sample of forty infor¬
mants in a Southern Appalachian community of approximately 3,000 per¬
sons. It quantifies the linguistic features it examines both in terms
of which speaker used them and in terms of possible occurrence of the
feature (whenever possible). It controls the social situation in which
its data occur (see the discussion of the ethnographic parameters and
the role this fieldworker adopted for himself later in this chapter).
This study controls the task its informants were performing by elicit¬
ing only expository (explanatory and descriptive) discourses, and it
controls the linguistic context of its data by considering the sentence
and especially the discourse contexts of data.
Fieldwork Procedures
The fieldwork upon which this dissertation is based was carried
out over a period of about two months, from December 1977 through Feb¬
ruary 1978. The fieldworker/researcher spent most of his time in the
community of White Pine during this period. This section details the
process whereby he set up, undertook, and completed his fieldwork. It
also provides the rationale for the type of interview and the ethno¬
graphic parameters adopted. Altogether six weeks were devoted to the
actual interviewing, one week (in the field) to preliminary fieldwork,
and a period of several weeks to preparation before he reached the
fieldwork site.
Two steps were taken before the researcher reached the White Pine
site. First, an ethnography of the White Pine and surrounding areas was

39
prepared. As thorough and as broad an ethnography as possible is in¬
dispensable to successful and accurate anthropological and linguistic
fieldwork. It is necessary to explain this point in regard to linguis¬
tic research in general and to this project in particular.
Speech is social behavior par excellence. As Hymes indicates, it
is the "nexus between language and social life" (1971:42). We cannot
accurately describe language (or, as Malinowski would say, meaning)
without reference to its role in society and the setting, participants,
events, acts, norms, etc., which it involves. Speech must be viewed as
permeating and not just overlaying most social behavior; therefore, de¬
scription of such behavior must make reference to language and its
forms.
No aspect of a culture can accurately be described without explica¬
ting its interrelationships with and dependencies on other aspects of
the culture. This has been more or less axiomatic since Malinowski's
ethnographic work among the Trobriand Islanders more than sixty years
ago. Malinowski insisted that a researcher approach a way of life as
a whole, as full of interdependencies, as an integration of life. Only
the application of the most rigorous techniques and standards for the
collection of data of all kinds—material, linguistic, topographical,
personal, ceremonial, among others—could enable the adequate explica¬
tion of the "context of situation," to use the term Malinowski (1923:306)
himself coined. All pertinent data must inform an understanding of a
context of situation before a generalization can be made, in other
words. For Malinowski, a context of situation is at the same time

40
broader and immensely more useful than the concept of a "linguistic
context": "Even as in the reality of spoken or written languages, a
word without 'linguistic context' is a mere figment and stands for
nothing by itself, so in the reality of a spoken living tongue, the
utterance has no meaning except in the 'context of situation'"
(1923:307).
In addition to the many kinds of data Malinowski considered useful
he also advocated that a researcher keep two kinds of personal records
which should constitute parts of his field records (Kaberry 1957:78-9):
1) An ethnographic diary of daily activities and observations not¬
ing what is the normal and the typical and what is not, the
kind of information that is not elicitable by questions and an¬
swers of a community's residents. Malinowski called such infor
mation the "imponderabilia of actual life." The present re¬
searcher kept such a diary while in the field, a diary which
also recorded the accounts of his successes and failures and
mistakes in the field.
2) A "corpus inscriptionum," which "entailed the recording in the
native language of narratives, opinions, typical utterances,
myths, folk-lore, magical formulae, and native explanations and
interpretations of customs and beliefs" (Kaberry, p. 79).
Since this researcher collected only one type of discourse (ex¬
pository, as defined in chapter two), his forty five hours of
taped explanations and descriptions constitute his "corpus in¬
scriptionum."

41
Before returning to the explanation of how an ethnographic sum¬
mary of the White Pine community was prepared, let us look more
closely at Malinowski's view of context and explore an instance from
White Pine where we must know the "context of situation" to fully un¬
derstand certain linguistic behavior.
As a result of his fieldwork experience, Malinowski formulated the
two concepts "context of culture" and "context of situation," the lat¬
ter mentioned above. Both have to do with how meaning is created by
language. A helpful redefinition of these concepts for linguistics has
been provided by J. R. Firth (1957b), one of Malinowski's students, to
express more satisfactorily how language means. According to Firth,
The context of culture is the environment for the total set of
[the options in behavior that are available to the individual
in his existence as social man], while the context of situation
is the environment of any selection that is made from within
them. (Hal 1iday 1974:49)
For Firth, then, the two types of context formulated by Malinowski pre¬
sent the distinction between what is potential and what is actual.
Again according to Halliday, "the context of culture defines the poten¬
tial, the range of possibilities that are open. The actual choice among
these possibilities takes place within a given context of situation"
(1974:49) Firth’s version consequently emphasizes the value of context
of situation for linguistics while at the same time implying that con¬
text of culture has little value for the linguist.
The great usefulness of the concept of context of situation, in
Firth's definition, for linguists lies in its stress on the typical

42
rather than the accidental as an appropriate object of study. Firth's
view also makes the important implication that there is no necessary
one-to-one relationship between an utterance and its meaning. This
implication is most valuable when we examine typical behavior.
The typical behavior (an utterance, in this case) from the White
Pine corpus to be explored briefly here is the expression "I don't
know." At times in the interviews this expression became almost annoy¬
ingly frequent. Consider an excerpt from an interview with a thirty-
nine-year-old housewife:
Fieldworker: Why should people be proud of White Pine? What
would you say?
Informant: Well, as a community, I think it's uh, it's a good
community to live in, to raise your children in. Well, 1
don't know. The churches, the schools, everything. I think
it's just a, I don't know. I like it. I think it's, that's
why I think it's a good community is because I like it,
(uh-huh, uh-huh) and I don't know.
Of course, it may be that the informant does not know how to artic¬
ulate her feelings here, or perhaps she in fact does not know why she
should be proud of White Pine, even though she had just stated, before
the fieldworker's questions, that the residents of the community should
be proud of White Pine. But this can hardly account for there being so
many occurrences in this excerpt of so deceptively short an expression
(three times in about twenty seconds). There are at least three differ¬
ent “I don't know's" to be found in the corpus, two of which are exem¬
plified here. Although this cannot be as clearly seen from the written
transcript, the researcher clearly identified at least three distinct
meanings for this expression:

43
1) A hesitation "I don't know" which frequently occurs, as in
our excerpt, with "well." This expression buys a little time
while the speaker decides exactly how to phrase an answer.
2) An ignorance "I don't know," exemplified by the other two in¬
stances, which indicates the speaker's recognition of his/her
inability to answer or express him/herself.
3) A reluctance "I don't know" which is most often used as an ini¬
tial reply to a question in order to show an informant's imme¬
diate reluctance to reveal his/her thoughts on a subject. This
differs from the first "I don't know" above. This "I don't
know" is not a refusal, however; informants almost invariably
went on to give a direct answer after showing initial reluc¬
tance.
The preparation of the ethnography of White Pine presented above
included the major sources on the Southern Appalachian culture and re¬
gion (Kephart 1922, Campbell 1921, Weller 1965, Ford 1962, Vance 1935,
Hicks 1976, and others). It also included major sources on the East
Tennessee area (Folmsbee et al. 1969, Dykeman 1977, and others). Spe¬
cial attention was given to the few sources on Jefferson County, Ten¬
nessee, available in local East Tennessee libraries. From the East
Tennessee Development District office in Knoxville were obtained sum¬
maries of the 1970 census reports and other vital statistics on the
White Pine community. In addition, much useful information was gath¬
ered from the informants regarding local affairs.
The second step, taken after the preparation of as much of an eth¬
nography as was possible from written sources, was the researcher's

44
making arrangements with the contact persons in White Pine to spend
an indefinite period of time with them and his explaining what kind
of investigation he would undertake and who he would like to interview.^
A contact person is an invaluable bridge between a fieldworker and a
community. In it a fieldworker is initially a total stranger. He can¬
not reside long enough to become a familiar member of the community
(in White Pine or another typical small town in East Tennessee, this
would probably take several years in any case). This researcher's con¬
tact persons introduced him to many of his informants and gave him many
insights into the life of the local community.
In White Pine itself, four problems had to be resolved before
actual fieldwork (i.e. interviewing) could begin. First, the field-
worker had to be thoroughly familiar with the layout of the town and
the most important basic facts of daily life in the community. In this
he was primarily assisted by his contact persons. This included learn¬
ing the principal activities of the townspeople and the principal social
institutions in the community. He also had to determine the principal
concerns of life for the community. He was able to do this in the
course of the interviews.
Second, he had to establish the ethnographic parameters of his
work. This included his role in the community during his stay and in
the interview situation. Third, he had to choose a location for the
interviews, specifically, where they could be conducted comfortably.

45
Fourth, he had to determine the type of discourse to elicit and how
to devise a standard interview format which would best elicit it.
The ethnographic parameters were adopted according to Malinowski's
suggestions. He states that a fieldworker should make every attempt to
be an active participant rather than a passive observer-partial partic¬
ipant in the culture he is studying. He draws on his own experience
among the Trobrianders in showing why this is necessary:
There is all the difference between a sporadic plunging into the
company of natives, and being really in contact with them. What
does this latter mean? On the Ethnographer's side, it means
that his life in the village, which at first is a strange, some¬
times intensely interesting adventure, soon adopts quite a nat¬
ural course very much in harmony with his surroundings. Soon
after I had established myself in Omarakana, I began to take
part, in a way, in the village life, to look forward to the im¬
portant or restive events, to take personal interest in the
gossip and the developments of the small village occurrences; to
wake up every morning to a day, presenting itself to me more or
less as it does to the native....As I went on my morning walk
through the village, I could see intimate details of family
life;...I could see the arrangements for the day's work, people
starting on their errands, or groups of men and women busy at
some manufacturing tasks. Quarrels, jokes, family scenes,
events usually trivial, sometimes dramatic, but always signifi¬
cant, formed the atmosphere of my daily life, as well as of
theirs. (1922:7)
This researcher, of course, is neither a native nor a resident of
the White Pine area. He was a stranger to his informants initially.
At the same time, he is a native of East Tennessee and especially that
he is a relative of long-time White Pine residents enabled him to estab¬
lish immediate rapport with his informants. It was also recognized
that the interview situation would necessarily be an artifical one.
Goldstein defines an "artificial situation" as one in which an informant
provides information "at the instigation of the collector" (1964:82).

46
The artificial interview situation, however, is, he adds, "the
only natural context for eliciting information from an informant"
(pp. 104-5). In other words, some type of directive elicitation is nec¬
essary, however willing the informants are to participate, since they
do not always spontaneously volunteer discourse, especially exposi¬
tions. Given these realitites, the researcher decided to adopt the
role of a college student collecting general information for a research
paper from a cross-section of the people in the town, stating that he
was interested in White Pine because he viewed it as a typical East
Tennessee small town. This is similar to the role of historian Gold¬
stein advises for folklorists. In this role the researcher states that
he is not interested in official opinions or just in the opinions of
the oldest and/or most respected members of the community, but that
he is seeking instead the ideas and views of a cross-section of the com¬
munity's average citizens. He emphasizes that he wants responses from
both native and non-native residents, from old and young, from members
of the whole community, whatever their social standing.
The researcher decided to have the interviews in the informant's
home rather than in the contact persons' home. He also discarded the
idea of interviewing in the informant's place of business or at some
other neutral, public site. Most informants readily invited the re¬
searcher into their own homes, and for this reason and others it ap¬
peared that informants would be most comfortable there. Moreover, a
fieldworker actually has more control over the interview in the infor¬
mant's home. There he has greater discretion over the pacing of the

47
interview and over signaling the end of the interview and his impending
departure. It is easier to tell an informant that the interview is
over, to pack up the equipment, and to say good-bye than it is to tell
an informant that he/she can go home now. It is also less of an impo¬
sition if the fieldworker takes care of his own transportation instead
of asking the informant to arrange his/her own transportation. Fi¬
nally, informants appeared more likely to venture information in the
securest and most familiar environment for them, their own homes.
The fourth thing that the researcher/fieldworker had to determine
before beginning full-scale interviewing was the type(s) of discourse
he would elicit and the interview format to elicit it (them). His ap¬
proach was based on the typology of discourse genres (narrative, expo¬
sitory, procedural, and hortatory) in Longacre (1976). Given the time
limitations and the ease of devising questions, the researcher decided
to concentrate on expository discourse. A standard set of questions
was asked of all informants, regardless of age, socioeconomic back¬
ground, or level of educational attainment. No difficulties were ex¬
perienced in elicitation according to this plan.
These limitations also had several advantages. First, using stan¬
dard questions gives the study strict controls. Then, expository dis¬
course seems to be the most plausible to elicit, given the ethnographic
parameters of this study. Most of the standard questions had to do
with everyday life in the town. Others dealt with situations that were
of widespread and frequent comment about town (such as the reputation
of Jefferson County's neighbor, Cocke County, for violence and

48
lawlessness). These are safe topics, unlikely to give offense, yet of
universal interest.
Before the interview itself and the sample are discussed, a few
more comments need to be made about how the fieldworker established
rapport with his informants. As Goldstein points out, the fieldworker
should take every opportunity to participate in the daily lives of the
people and the community he is studying. This researcher visited local
churches and various gathering places and meetings, chatting with peo¬
ple frequently on the streets and at lunch counters. He continually
emphasized that he was interested in interviewing a cross-section of
townspeople.
The interview itself is a directive one, but it is loose and flex¬
ible. It was directive insofar as it required specific vital informa¬
tion from informants (age, length of residence in White Pine and else¬
where, level of education attained, occupation held, and family back¬
ground). The fixed questions asked of all forty informants require
expository responses:
1) How does White Pine react to newcomers in town?
2) What have been the most important changes you have seen in
White Pine while you have lived here?
3) What's a typical day like for you?
4) What do you think most people in White Pine want the town to
become?
5) What all goes on on Pine Day?
6) What do you do on Pine Day?

49
7) Why does Cocke County have a bad reputation?
8) How do local people feel about outsiders making fun of East
Tennesseans and the way they talk?
In this way, some uniformity in interviews was established.
However, the interviews were not rigid. For example, no infor¬
mant was asked only these eight questions. Each informant was asked
additional questions depending on age or background. For instance, in¬
formants over fifty who were native of the area were asked how local
people felt when TVA came into the county in the 1930's and bought up
prize farmland to be covered by reservoirs. Similarly, high school age
informants were asked how the new county-wide comprehensive high school
compared to the old high school in White Pine. Other questions were
asked, depending on the flow of the conversation.
One non-directive principle was adhered to throughout the interview¬
ing, however. This concerns the lack of a time limit. Once the field-
worker had asked a question on some subject, the informant was allowed
to give as long a response as desired. The fieldworker took pains nei¬
ther to interrupt nor to influence the informant's pattern of thought.
Only when the informant signaled an end to a response or when an infor¬
mant's response was clearly running out of steam did he intrude with a
comment or another question.
The interviews averaged slightly more than an hour in length. But
the fieldworker spent at least twice that amount of time with the aver¬
age informant, exchanging small talk and establishing a rapport with in¬
formants before the actual taping began. The time taken by the field-

50
worker to introduce himself and to tell about his family was indispen¬
sable in establishing the uninhibited rapport that he desired. The in¬
terviewer's family was of more concern to the informants than his mo¬
tives for interviewing.
In his view, the fieldworker was able to establish quite good rap¬
port with each of his informants and excellent rapport with nearly all
of them. This resulted in not only a good corpus but also a number of
good friendships with individuals in White Pine.
The Sample
According to the 1970 United States Census, the population of the
town of White Pine was 1,532. It had grown to an estimated 1,830 by
1977. But the community of White Pine as defined by its residents in¬
cludes both the area outside the city limits and the adjacent neighbor¬
hood of Leadvale, bringing the total population of the area in which the
interviews were conducted to approximately 3,000 persons in 1978. Of
this base population of 3,000 a sample of forty residents was inter¬
viewed. This constitutes 1.3 percent of the whole. The interviews
average roughly one hour each, the shortest being thirty-five minutes
and the longest more than two hours.
The sample of forty informants is both a broad and a representative
cross-section of the White Pine community. The informants represented
all post-adolescent age groups and a wide variety of socioeconomic and
educational backgrounds. In age they range from sixteen to eighty-seven
as distributed among the following cells:

51
16-19
20-39
40-64
65 and over
Total
Male
3
6
5
4
18
Female
3
5
8
6
22
As the sample was taken, care was exercised not to have too many in¬
formants from a given age group or from either sex. The fieldworker's
overriding concern, as he spent his time in the White Pine community
and had contact with all sectors of the population, was to interview as
wide a variety of its residents as possible. It was not his primary
goal to investigate social or subgroup differences in discourse organi¬
zation among his informants.
The breakdown of the forty informants according to their levels of
education
is as
2
follows:
8th
Grade
Some High
High School
Some College
College
or
Less
School
Graduate
Work
Graduate
Female
1
7
5
5
4
Male
3
5
1
6
3
According to the 1970 census, only 6.3 percent of White Pine's popu¬
lation depended on agriculture for its livelihood, while 74.1 percent
depended on manufacturing, trade, service, and finance industries.
Approximately 19 percent of the population was employed in construc¬
tion and all other areas. Following is a breakdown of the White Pine
sample according to their employment by type of industry. Three who
were unemployed are ignored.

52
Agriculture
Construc¬
tion
Manufac¬
turing
Trade,
and
, Service
Finance
, Other
#
%
#
%
# %
#
%
#
%
1970 Census 89
6.3
129
9.1
720 51.0
327
23.1
142
10.0
Present
Study Infor- 3
8.1
3
8.1
6 16.2
21
56.8
4
10.8
mants
If females were not members of the labor force, they were categorized
according to the employment of the breadwinner of their families. The
present study's higher percentage of informants supported by trade, ser¬
vice, and finance than the percentage of those depending on manufactur¬
ing contrasts with the percentages for the overall White Pine area.
This is due to their availability. There were more people in trade,
service, and finance positions who were available for interviews than
in manufacturing.
Although no census figures are available on the lengths of resi¬
dence of White Pine inhabitants, it is clear that the great majority
of them are lifelong natives of White Pine or the nearby area. Of the
forty informants, twenty-two were natives of White Pine and another
thirteen were natives of nearby counties in East Tennessee, nine of whom
had spent more than half their lives in White Pine. Four were natives
of elsewhere in Tennessee or of another state, and one was a native of
Great Britain. Informants were chosen on the bases of their ages and
educational backgrounds and not their nativities. Even so, the cohe¬
sive devices analyzed in chapter three through five were examined for
differences in function between natives of White Pine and non-natives.
None but incidental differences in frequency were found.

53
In the White Pine community, only 1.7 percent of the population
was non-White in 1970; none of the forty informants for this study,
it happens, were non-White.
Transcription Procedures
The approximately forty-five hours of field tape recordings are
the primary documents for this study and are viewed as the ultimate
authority and repository for the data upon which this study is based.
All forty-five hours of interviews were transcribed by the fieldworker/
researcher himself. The transcription was done in slightly modified
orthography in order to result in as close to exact a form as possible
of what the informants said. As Labov (1969, 1972) has shown, careful
transcription is as important in considering grammatical features
as it is for phonological ones.
It must be realized, however, that any transcription can never
be more than an imperfect representation, and an interpretation as well,
of what is on tape. This is certainly true when a transcriber attempts
phonetic or a modified phonetic transcription, and it was true in the
present case when the transcriber had to determine the presence of cer¬
tain deletions and especially when he tried to be consistent in punc¬
tuating a transcript or in ascertaining sentence boundaries. A tran¬
scriber quickly recognizes that deictics and especially conjunctions
mark the beginnings of sentences. But an informant's pausing and pacing

54
are also crucial to sentence, clause, and phrase boundaries. This
is why the tape recording itself remains the ultimate authority.
The transcription employed here makes no attempt at eye dialect.
Orthography was modified only to indicate several minor instances of
deletion such as
1) The voiced interdental in "them" and "that," if deleted, is re¬
placed by an apostrophe (as in "with 'em" or "like "at");
2) Clause initial "that's" and "it's," often reduced phonetically
to [s] or [s:], and therefore indistinguishable, are represented
frequently as '"s";
3) Initial syllable deletion, as for "of course," represented as
"course," and "because," represented as "cause."
In no instances were such conventions used to indicate the deletion of
morphemes. Dealing with deletions in general was frequently difficult.
Basically any apparent morpheme that had a clear trace was represented
in full orthography. In fact, clear hesitation phenomena such as "uh"
and "hmm," the affirming "uh-huh" and the negation "huh-uh" and the like
were represented whenever audible. But the many standard English words
without such a trace (such as copulas, relative pronouns, and others)
were not represented.
Hesitation phenomena (as well as the parenthetical "you know," "I
mean," and other phrases adding nothing to the content of the interview)
are represented in this study's interview transcripts for the sake of
making the transcript as accurate as possible. It is also possible
that such phenomena might have some discourse significance, as markers
of sentences or of paragraphs (i.e, changes or returns to themes). In
any case, possibly significant data should not be discarded a priori.

55
The use of brackets and parentheses in the interview transcripts
follows the suggestions of Ives (1976). Parentheses serve to enclose
short interviewer remarks (rarely more than the occasional "uh-huh"
or "ham" which insures an informant that the interviewer is intently
following the informant's account) which did not cut short or deflect
an informant's train of thought. Occasionally parentheses were also
employed to suggest an alternate transcription, when the speech on
the tape was not quite intelligible. Brackets enclose either inter¬
viewer comments on the informant's behavior or short summaries of the
interviewer's speech. In the interview excerpt presented below, the
two different uses of parentheses are exemplified.
Fieldworker: In general, what have been the most significant
changes that you've seen in the community of White Pine it¬
self during your lifetime?
Informant (74 year-old male): Oh, I guess this, what I remember
as farmland is, has been being and is being covered up with
houses and housing developments and such as that, (uh-huh)
and around the town. Most of those, I remember when, when
that was farmland, the ones out this side of town and well,
that's all around town. And that uh, and that and inflation,
(hmm) Course, we've gone from, as far as uh, as transporta¬
tion, we've gone from horses and buggies and mules for to
farm with, and buggies to ride in and, and trains hauled the
produce, to now it's gone to trucks. The trains are, people
ride in cars and, and uh, and produce being hauled in trucks,
right off the farm from there on, from the time it's produced
til (to?) the, the time it's gets to the grocery store. Rail¬
road used to use a good many people. That was pretty big bus¬
iness. Of course, it hauled lots of people in those days.
That was quite a, quite an event when a train came to
town....
As mentioned, all the interviews were entirely transcribed by the in¬
terviewer himself. Inevitably there are sentences, phrases, and even

56
words (especially names) that only the interviewer can comprehend by
virtue of his having remembering the context of the interview.
Fieldwork Journal and Interview Catalog
In addition to the primary document, the tapes, and the secondary
document, the transcripts, two other document were assembled in this
project. A fieldwork journal of daily activities and observations was
maintained by the fieldworker. After a day’s fieldwork was completed,
a record of the day was entered into the journal. As mentioned earlier
in this chapter, this emulated the ethnographic diary that Malinowski
kept while among the Trobriand Islanders. Not only does such a journal,
maintained daily, allow a fieldworker to profit from his own experi¬
ences, his mistakes and his frustrations, but it also enables him to
make note of particularly successful and effective approaches and ques¬
tions for his later use.
The other document assembled was the interview catalog, compiled
for the researcher's ease of reference. It contains four kinds of in¬
formation:
1) The vital statistics on each informant (age, sex, level of edu¬
cational attainment, etc.);
2) The physical details of the interview: when and where it was
conducted;
3) A summary of how the interview itself went, how good the rap¬
port between fieldworker and informant was, and how helpful
the informant was;
4) A note on how the informant was contacted.

57
Explanation to Informants
The members of the White Pine community, when approached by the
fieldworker/researcher as potential informants, were almost invariably
most willing and even flattered to be asked to participate in an inter¬
view. While a few of them were initially concerned that they were
being "singled out" to be interviewed for some special reason (the
fieldworker had to assure them that he wanted to interview typical cit¬
izens), the fieldworker was actually turned down for an interview only
once (over the telephone, where, he is convinced, he was mistaken for a
salesman). Most informants were more interested in knowing why they
(and White Pine) were chosen than what the fieldworker actually wanted
to interview them for, and that the series of interviews was a part of
his schoolwork was sufficient explanation for most informants to be
quite willing to be interviewed. The usual explanation provided pro¬
spective informants was as follows:
For a school assignment I would like to interview a cross-section
of people in White Pine and ask some general questions about
what life is like here. What I want to do is compare the an¬
swers I get from a variety of people, both old and young, male
and female. I wondered if you would be willing, if you have
some time, to help me by answering a few questions.
The fieldworker willingly gave whatever additional information was de¬
sired—that the interviews were part of his dissertation, that he was
working on his doctorate, or that his project was a linguistic one—
but this was rarely called for.
It should, of course, be pointed out again that the initial rap¬
port that the fieldworker was able to establish with informants would

58
almost certainly not have been so successful without the assistance
of the contact persons in White Pine.
Informants for this study were assured that their names would
not be revealed and that they would remain anonymous to all but the
fieldworker.
Taping of the Interviews
The interviews for this study were recorded on a four-track, reel-
to-reel Voice of Music model 733 Tape o Matic tape recorder at 3 3/4
inches per second. The tapes used were Scotch No. 207, 1800 foot
tapes, which, at 3 3/4 inches per second, allowed up to ninety minutes
of recording per track. In all eleven reels of tape were required.
Because of logistical difficulties, two interviews had to be re¬
corded on cassette tapes. These interviews are coded as tapes XII and
XIII in Appendix A.
The Value of Fieldwork
Since the ultimate value of doing linguistic fieldwork is to help
us better understand how people employ language, the fieldworker must
always remember that the interview is a very human event involving di¬
rect contact between two or more people. As Goldstein declares, "the
problem of fieldwork is essentially human" (1964:22), and more than any¬
thing else fieldwork is a venture in human relations. The fieldworker
cannot be detached from his informants, especially when subjects as
deep-felt as some of those covered in the present study's interviews

59
are involved. Fieldwork is extremely hard and tiring work, but its
rewards in terms of friendships gained justify the effort put in by
the fieldworker.
From this researcher's perspective, at least three skills are re¬
quired of the linguistic fieldworker in such a study as this one:
1) The ability and willingness of a fieldworker to relate to his
informants as humans who are willing to share their personal, and some¬
times intimate, experiences with him. The fieldworker must value his
informants as people and not as sources of data, which means that it
is probably impossible to do good fieldwork without genuinely enjoying
human contact.
2) A resourcefulness to adapt to new situations and people in ways
that the fieldworker could not have foreseen before going into the
field. The fieldworker must take care and even pains to approach peo¬
ple as individuals without preconceptions and prejudices. He should ap
proach his corrmunity as a microcosm having a great variety of unique in
dividuals.
3)A disposition to persevere against frustrations and inconveni¬
ences and ever-present delays. In short, he must be so committed to
3
doing his fieldwork that he develops a kind of salespitch for it. He
must be willing to solicit assistance from often-hesitant and sometimes
skeptical people while at the same time endeavoring to establish a fa¬
vorable rapport with them.
Though the most demanding work he had ever done, this writer found
it immensely rewarding. It gave him many fond memories and several
good friendships which have developed from his interviews.

60
Notes
1. The choices of the community to work in and of the contact persons
were entirely practical ones. This writer's brother-in-law is a
native of White Pine and his brother-in-law's parents are longtime
residents there. It is they who were chosen to be the primary con¬
tact persons since they knew nearly everyone in White Pine and could
assist the fieldworker in making the necessary initial contacts.
More important than facilitating contacts, however, was that his
family relationship with the contact persons enabled people in White
Pine to identify him with people they knew well.
2. The category "Some College Work" includes all informants who had any
kind or length of post high school education, such as business or
other vocational school. According to the 1970 census, only 6.8 per¬
cent of the residents of the White Pine division of Jefferson County
(figures are not available for the town of White Pine itself) over the
age of 25 had attended some colleqe. However, three factors account
for a high percentage (45 percent) of this study's sample having col¬
lege work: 1) the census area (with 2,085 people over 25) includes
rural area whose residents have a lower averaqe educational level
than do the town's residents; 2) in 1970 a junior college opened five
miles from White Pine; it has attracted many part-time students, and
informants having attended as few as one or two courses are classi¬
fied as having some college work; and 3) nearly all this study's in¬
formants who enrolled in college for a short time dropped out (for
many in this generation, "trying a quarter of college" was the ex¬
pected bridge from high school to the workaday world).
3. This same view, that the fieldworker must be willing to be a sales¬
man, was expressed to the researcher by Vance Randolph, the dean of
American folklore collectors, in a personal conversation with the
writer on December 11, 1978, in Fayetteville, Arkansas.

CHAPTER II
BACKGROUND AND METHODOLOGY
Preface
The first chapter of this dissertation presents the non-linguistic
context of this study; this chapter discusses its linguistic context.
In the first part of the chapter several approaches to the analysis of
discourse structure are reviewed to put the analysis of oral expository
monologues into perspective. Then the methodology of this dissertation
is explained. The second section examines the importance of studying
monologue expositions. Then some remarks are made regarding the nature
of oral paragraphs, followed by a review of relevant studies of oral
and written paragraph structure. Finally Longacre's typology of dis¬
course genres is discussed.
The control of the five different contexts mentioned in chapter
one (speaker type, frequency of occurrence, social situation, task, and
linguistic position) makes this study a novel one. Also, as far as
this writer can determine, this study is the only analysis of the dis¬
course structure of oral expositions for any variety of American Eng¬
lish. Although much has been written about discourse structure in Eng¬
lish, oral expositions have been seriously studied only for other lan¬
guages (see especially Longacre 1968). This dissertation represents
the first time that the task of eliciting, analyzing, and then gener¬
alizing about a large corpus of expositions has been undertaken for
English. It therefore is an exploratory project.
61

62
Discourse Studies and the Approach of This Study
The decision to study the discourse level of oral expositions orig¬
inally derived from this writer's growing conviction that the construc¬
tion known as left dislocation, an often-used construction in speech,
could only be understood if its function were viewed as supersentential,
on the discourse level (Montgomery 1977a, 1978). If such a frequent but
little studied construction seemed to tell us something about how speak¬
ers organize their linguistic behavior, it seemed likely that other fea¬
tures and constructions of oral discourse should also be studied in
their discourse contexts. There is much to recommend the view that
oral expositions are preliminary to written ones. At least, understand¬
ing how speakers negotiate and structure ongoing expositions might im¬
prove our awareness of the communication system that people bring to the
writing process. Since linguists have for two generations made the case
that speech is primary and writing derivative, it seems crucial that a
study should begin with the structure of oral expositions rather than
written ones.
It is no longer necessary to justify the value of doing discourse
analysis, which may be broadly defined as the study of any segment of
verbal behavior larger than the sentence. Grimes (1975), the watershed
work in the field, has shown that discourse analysis is indeed possible
and that it rightly lies within the domain of linguistics (rather than
rhetoric or another field exclusively). Grimes' work is such a vast

63
cornucopia of observations on discourse that it should no longer be
asked "Is it worthwhile to do discourse analysis?" Rather we should
ask "What kind of discourse analysis will address significant ques¬
tions about how people organize their linguistic behavior?" Since
Grimes provides neither a methodology nor a point of departure for a
principled approach to the analysis of expository discourse (but nei¬
ther is his aim), we must look elsewhere for an approach to oral expo¬
sitions.
According to Widdowson, discourse analysts have generally taken
one of two basic points of departure. Either they have started with
the text or discourse as a whole or they have begun with sentences or
subsentences in an effort to identify their discourse function(s):
One general approach to discourse analysis, then, begins with
instances of discourse, with actual data, and moves towards lin¬
guistic units to the extent that this appears to be necessary
for the purpose of the description. The second approach moves
outwards, as it were, from the sentence, and deals with lin¬
guistic expressions as realized in discourse but with the ab¬
stract potential of linguistic forms. (1977:241)
The emphasis in the first approach is on the hierarchical nature of
discourse, in the second on the linear nature of discourse. There are
good reasons for choosing either point of departure and in many in¬
stances discourse analysts have taken one or the other and elected not
to move toward the other. Propp (1958) and a number of other writers
on folkloristics do not move toward linguistic units at all. Several of
the modern-day text grammarians (Pettifi and Rieser 1976, Van Dijk 1972)
also rely on abstractions from a text for analysis and are primarily

64
concerned with the abstract and rhetorical structure of discourses.
Text grammarians for the most part view discourse or text as a unit of
competence and not of performance.
Other analysts approaching discourse (e.g. Williams 1964, Halliday
and Hasan 1976) choose to study exclusively the various cohesive re¬
sources of English to fit one sentence after another into a context,
to produce acceptable sentence sequences. Williams examines deletion
and a wide variety of elements such as deictics and sentence adverbs.
Halliday and Hasan are concerned with such semantic resources of Eng¬
lish as substitution, ellipsis, reference, and conjunction. This lin¬
ear approach to discourse is devoted to explicating intersentential
linkage.
Several discourse studies attempt to bridge the gap between these
two points of departure and to balance the hierarchical and linear ap¬
proaches to the study of discourse organization. They have varying re¬
levance to the present study, however, since none of them (except, in
part, Longacre 1968) are based on a corpus of spontaneous oral exposi¬
tions. Several studies, e.g. Grimes and Glock (1970), Klammer (1971),
Klammer and Compton (1970), and Van Dijk (1977), deal with only the
narrative structure of one or a small sample of texts. The basic point
of many narrative studies is that a discourse view of a text can indeed
make some interesting observations more so than it is to generalize
about a body of texts. In studies of narrative structure we encounter
again and again statements similar to that of Van Dijk: "The remarks
made about a well-known type of discourse, the story, could be extended

65
for other discourse types" (1977:155). Very few linguists have at¬
tempted to do this, however. Longacre, in his Philippine study (1968),
has been the only one to do this. He posits formulae for the gramma¬
tical structure of each type of discourse and paragraph. This comment
also applies to his extension of the concept of plot to types of dis¬
course other than narratives.
Nearly all discourse analyses have investigated only written or
very highly formularized (i.e. the sermon or the folktale) oral texts.
The discourse features which they consider are not necessarily, even for
expositions, the same as those in spontaneous oral discourses, which we
will be concerned with here. As an example, we can compare Jones (1977)
with the present work. In her study of written expositions, Jones con¬
tends that thematizing devices (repetition, marked word order, rhetori¬
cal questions, clefting and pseudoclefting, among others) can fruitfully
be studied in written texts. But all the devices she finds are ex¬
tremely rare in our White Pine corpus of oral expositions. The sole ex¬
ception to this statement is conjunctions, which are of secondary impor¬
tance in communicating themes. Moreover, the two most frequently used
thematic constructions from our corpus (left dislocations and existen-
tials) she does not mention. Thematic and other devices should be stud¬
ied, but we should not assume that they are identical for written and
oral texts.
Other analysts have suggested other types of devices as crucial to
study in a discourse context: punctuating or transitional devices (Mer¬
ritt 1972), various kinds of anaphora (Gutwinski 1976), equivalence

66
chains (Harris 1963), and those of Halliday and Hasan and of Williams
mentioned above. As valid as it may be to study each of these types,
the necessary heuristic framework for the comprehensive study of co¬
hesive devices in discourse is provided only in Longacre and Levin-
sohn (1978). Our approach to the paragraph structure of oral exposi¬
tions is largely based on the ideas in this article.
As suggested in their title, "Field Analysis of Discourse," Long-
acre and Levinsohn offer a framework for discourse analysis for lin¬
guists working in the field with little-studied languages (such as
those which Summer Institute of Linguistics workers frequently encoun¬
ter). What the present writer has chosen to do is to adopt it for
fieldwork with a much-studied language—English. Longacre and Levin¬
sohn are determined to bridge the gap between the abstract structure
and the specific cohesive devices of a discourse but insist that the
analyst can do this only by first identifying and describing the "sur¬
face structure cohesive devices" and discourse constituents: "The job
of the analyst is to...look through the flesh and the skin to the
skeletal structure beneath and to perceive the fundamental structure
of the whole" (1978:105). The aims of the analyst are to show how
these cohesive devices operate and then to display an outline of the
discourse in a schematic fashion.
Longacre and Levinsohn present eight surface structure cohesive
devices which permit a discourse to be outlined:
1) The role of tense and voice;
2) Particles and affixes;

67
3) Participant anaphora;
4) Deictics;
5) Lexical ties and paraphrase;
6) Summary and preview;
7) Conjunctions and introducers;
8) Backreference.
Given the time limitations of this study, four of these devices (the
role of tense and voice, particles and affixes, lexical ties and para¬
phrase, and backreference) are not considered in this dissertation.
The remaining four (participant anaphora, deictics, summary and pre¬
view, and conjunctions and introducers) are the concerns of chapters
three through five herein.
The second hypothesis of this study is that the grammatical struc¬
ture of oral expository paragraphs is organized to a significant degree
by certain cohesive devices—the other four discussed by Longacre and
Levinsohn. It investigates this hypothesis by examining the function(s)
and distribution of these devices.
The device called "participant anaphora" has to do with the identi¬
fication of participants in a discourse: "Participants can be identified
by name, by a common noun, by pronoun, by an affix, or just by zero.
Such variations in anaphora are never unmotivated. Most commonly, the
domain of the participant anaphora chain is the paragraph" (Longacre and
Levinsohn 1978:108). We will here be concerned with only one link in
this chain—the first—(many linguists have studied anaphora: Gutwin-
ski (1976), Brinegar (1977), Hinds (1977) and others) and will call this

68
first link the "theme." The theme is the piece of information (not
necessarily a human participant) which usually initiates an anaphora
chain and which a block (or paragraph, as we will see) of information
is about. The primary surface structure devices which introduce theme
or "thematize" a piece of information are of two sorts in our corpus:
left dislocations and existentials. In the following exposition we un¬
derline two left dislocations and two existentials, all four of which
initiate anaphora chains, to illustrate our approach.^
Fieldworker: What have been the major changes you've seen in
White Pine during your lifetime? Let's say, the most signifi¬
cant changes?
Informant: Well, they restructured the bank. I wished they'd uh
left it the old way, but I guess they did that for, you know,
security measures and stuff, and then uh the mall, the mall I
was talking about, the shopping center. You know, they've
paved it all the way throuqh, so that's, that was the major
thing right there that, that they've done. And of course,
the housing project, they put that down there, and I wish they
hadn't have. I like White Pine the way it was, you know.
Well, simpleton little town, you know, but it seems to be ex¬
panding quite, quite a bit. (hmm) And there's a lot of
trailer courts going in. There's one right in the center of
town. I don't know how they ever got that admitted. Course,
they have restrictions now, but it was in there before the re¬
strictions came, so (uh-huh) and, when I grew up, there, we
had one policeman, and that was Chief of Police Dick Reed, and
he had a deputy, and that was it, and they would drive around,
you know, most of the day, or if you needed him or anything to
around twelve o'clock, and then they'd go home, you know, so
the town was just, just nobody out, you know, or few people
out, but there was no crime, you know, nobody breaking into
anything, (uh-huh) That was just the way it was,... (V2-11)
In this exposition thematic devices introduce four themes (the mall,
the housing project, a lot of trailer courts, and no crime, although
the fourth one is not developed like the others) and at least four
paragraphs in this exposition. Theme is the one obligatory element

69
of a paragraph's grammatical structure and to a significant extent in
our White Pine expositions a paragraph's theme is marked by one of the
two constructions mentioned. In Longacre's most recent view of the
grammar of the paragraph, he indicates that "we find the paragraph
built around a theme that is not different in kind from a thematic par¬
ticipant" (Forthcoming:118). Our chapter three confirms this for expo¬
sitions.
Two other types of devices, deictics and summary, are considered
together in chapter four because deictics are frequently employed to
make summaries in our expositions. In general, deictics (especially
the word "that") help keep track of thematic information in a paragraph.
We will examine four uses of deictics in chapter four: 1) as the prono¬
minal reference in left dislocation; 2) in extrapositions; 3) in iden¬
tifying sentences; and 4) in what we will term "peak" sentences. An
example of a peak sentence is underlined below:
Fieldworker: What would you say have been the most significant
changes you've seen in the town in your few years here?
Informant: Most significant change I've seen is the, the intru¬
sion or the coming in of, of new people. White Pine is no
longer the same town it was eight and a half years ago because
of its growth, uh because of people with different backgrounds
having moved into our town, (uh-huh) Uh therefore it has
changed the complex of our town tremendously. Uh I think this
is the greatest change that has happened in our town is to
have new people coming in with new blood, new thoughts, new
ideas, new approach to things. (X4-4)
We will see that such sentences represent another unit in the grammati¬
cal structure of the paragraph—the peak—and we will confirm the view
of Longacre and Levinsohn that, in expository discourse, the peak is
'the culminating explanation" (1978:105).

70
The concern of chapter five is an eighth type of device—conjunc¬
tions and introducers, insofar as they occur in a set of 200 of the
corpus1 expositions. A rough taxonomy of the most frequent conjunc¬
tions accompanying left dislocation is presented. Finally, some ob¬
servations about subparagraphs, especially introduced by "like," are
made.
Longacre and Levinsohn thus elaborate eight types of surface
structure cohesive devices. Our object here is to show that, for one
corpus of English, four of them make complete sense only if viewed in
a discourse context. Although we examine the underlying regularities
of several of the specific devices (left dislocation and deictics which
participate in peak sentences, especially), our primary aim is to de¬
termine their distribution and function(s) across extended texts, which
also shows their periodic character. This periodic character reflects
the paragraph nature of spontaneous oral expositions. To a significant
degree, surface structure cohesive devices implement the grammatical
units of paragraphs, especially the theme and peak.
For convenience, parts of the analysis are performed on a subset of
the corpus1 expositions, a random selection of 200 sample expositions,
five from each informant. In some cases, the analysis is based on a
smaller set of 80 expositions, two from each informant.
Based on the expositions of forty people, our findings have a
broad generality with respect to the community investigated, White
Pine, Tennessee. It might reasonably be asked why so little concern

71
and attention is given in this study to variation in cohesive devices
between subgroups of the informant sample, according to either the
age, the educational level, or the socioeconomic background of the in¬
formants. Except for only a few scattered remarks (regarding the in¬
troducer "like," for example) nothing is said about differences between
individual or groups of speakers. If White Pine had been presumed to
be a homogeneous speech community, the care to interview a large and rep¬
resentative cross-section would not therefore have been taken. It was
found that the discourse devices under consideration in this study
showed inconsequential differences in function across the entire sample
of forty informants. None of the features were restricted to one subset
of the sample. The generalizations in this study resulted from their
obviously consistent functioning for all forty speakers.
This is not to say that no variation occurred in the frequency of
the realization of such devices as left dislocation. It was realized in
at least three instances for each speaker. But one speaker, a seventy-
year-old woman, had seventy instances. Hers was, however, by far the
longest interview—nearly three hours. Neither for the four informants
not native of Tennessee nor for the one native to Great Britain did left
dislocation function differently from White Pine natives. Peak sen¬
tences, with fifty instances, were used by twenty three speakers of many
different educational levels and socioeconomic backgrounds and thus dis¬
tribute broadly throughout the sample. Other devices may have varied in
frequency too. But none vary in function. More important, there is
little to suggest that any of the devices considered in this disserta-

72
tion are diagnostic of any of the speakers. A closer look might re¬
veal some kind of variation, but that would be the object of another
study.
The Study of Monologue Expositions
Linguists interested in supersentential segments of speech have
only recently heeded Firth's injunction of nearly a half-century ago
that "neither linguists nor psychologists have begun the study of con¬
versation; but it is here we shall find the key to a better understand¬
ing of what language really is and how it works" (1957a; 32). Only if it
is recognized that the function and manifestation of language is ba¬
sically conversation can language best be understood. Recent studies
on turn-taking, as by Sacks, Schegloff, and Jefferson (1974), recognize
the importance of studying blocks of conversational dialogue and have
shown that they have cohesion, just as other discourse blocks do.
Klammer (1971) is a study of written dialogue paragraphs and is also an
important contribution in this respect.
The present study does not take into account turn-taking phenomena
between the fieldworker and the informants. But it is consistent with
the conversational frame of reference for studying language. It does
not artificially cut up conversation and thus de-contextualize the
monologue discourses it analyzes. It views expository monologues as
parts of conversational exchange, the context for which was controlled
(as indicated in chapter one) by the fieldworker. Pike and Lowe see
monologue as occurring when "one individual of the cast stays fixed

73
in the role of addressee" (1969:74). Klammer defines monologue more
loosely as "one uninterrupted speech within a dialogue context"
(1971:38).
The monologues we encounter in our corpus are usually isomorphic
with turns in conversation, but not always. In some instances it is
obvious that monologues continue across an exchange in turns. That is,
we cannot say that a speaker needs to reintroduce, at each turn in the
conversation, all the antecedents from an earlier turn.
Good reasons for studying the structure of monologues can be
found. Longacre views monologue as a "special lopsided development of
conversation" (Longacre 1976:165). But he stresses its "universal cul¬
tural importance" (p. 197). His typology of discourse genres (Long¬
acre 1968, 1976, Longacre and Levinsohn 1978), which he believes to
have universal applications, is actually a typology of monologues (nar¬
rative, expository, procedural, and hortatory). What Longacre consi¬
ders the deep structure or universal features of expository monologues
is discussed later in this chapter.
There is another good reason for studying the structure of mono¬
logues. The expository monologue type of discourse on which this dis¬
sertation is based has all the features of what Hymes (1971) calls a
"speech act": setting and scene, participants, ends, act sequence, key,
instrumentalities, norm, and genre. Viewing monologue as a unit, more¬
over, allows us to examine thematizing devices, deictics, and other de¬
vices in this dissertation. This is because these devices help orga¬
nize monologue conversational turns into one or more self-contained

74
thematic units—paragraphs—which represent the natural thematic struc¬
tural units in which speech is used.
Oral Paragraphs and a Review of Paragraph Studies
In analyzing the structure of discourses, linguists have often rec¬
ognized distinct supersentential units smaller than the discourse it¬
self. This is the case with Van Dijk (1977) for narrative, Klammer
(1971) for dialogue, and Longacre (1968) for dramatic, procedural, and
other types of discourse. Van Dijk uses the term "macro-structure" for
a sequence of sentences which belong together and which frequently has
"Morpho-Phonological markings" (1977:152-3). He mentions pauses, in¬
tonation, and particles as examples of such markings. Klammer, Long-
acre, and others adopt the less original but more useful term "para¬
graph," the traditional term referring to an orthographic unit of lan¬
guage. Herein we will also use the term paragraph for these interme¬
diate structural units in oral discourses that are unified in more ways
than simply forming a block of print on the page.
For Longacre, any string of two or more sentences which belong to¬
gether constitute a paragraph: "Unstructured sequences, or strings of
sentences which do not constitute a paragraph structure, do not exist"
(1968:53). Most often the paragraph has been defined as any self-con¬
tained thematic unit. Pike and Pike call it "the minimum unit in which
a theme is developed" (1977:488). We consider a paragraph to be any
thematic block of language, i.e. a unit of language organized around
a theme or piece of information.

75
Thematic unity is not something merely imputed to a paragraph in
speech. It is there and is reflected to one extent or another by var¬
ious lexemic, grammatical, and phonological features. In a narrative
discourse, paragraph themes are usually characters; in exposition the
themes may be characters or many other kinds of information. Examples
will be presented in chapter three.
We can view these thematically unified units called paragraphs as
having three different modes of structure—phonological, lexemic, and
2
grammatical. Although of most interest to us is the grammatical mode,
all three modes to a great extent reinforce one another. It is doubt¬
ful that the grammatical structure of a paragraph can be understood
without studying the other two.
Paragraphs have not often been understood as having characteristic
phonological features. But according to Klammer, "phonologically,in a
complete description of dialogue paragraphs, it would be necessary to
account for such phenomena as stress placement, juncture, and intona¬
tion in relation to the paragraph level constraints affecting them"
(1971:33). Even earlier, Becker pointed out the existence of phonolo¬
gical markers of paragraph structure:
Paragraph tagmemes seem to be marked by shifts in pitch regis¬
ter, tempo, and volume when paragraphs are read aloud. While
these signals can be perceived by a trained phonetician, they
have not been adequately described in the laboratory, and
their written counter parts have not been identified. (1965:242)
Unfortunately research into such features and markers, as by
Bridgeman (1966) and Lehiste (1978, 1979), has been either too recent

76
or too sparse to have developed well-tested tools to examine them.
Bridgeman's pioneering work on Kaiwa claims that oral paragraphs in
that language have one of a set of six phonological onsets and one of
three phonological codas, with an optional peak preceding the coda of
a paragraph. Lehiste has attempted to find, with as yet little suc¬
cess, the phonological correlates of paragraph boundaries by measuring
pausing, lengthening, and pharyngealization at the beginning and end of
sentences and paragraphs.
The isolation of the phonological features of paragraphs is the
subject for another dissertation. Only occasional observations regard¬
ing phonological matters are advanced in this study, when these features
are concomitant with the surface structure cohesive devices that we
analyze. More than likely a characteristic paragraph-initial lengthen¬
ing and/or pausing will eventually be shown to aid in giving paragraphs
coherence. One suspects, however, that no phonological signal will cat¬
egorically mark some part of a paragraph's structure. What one should
expect to find is that a significant coincidence of phonological and
grammatical signals co-occur at thematic breaks in a discourse. In
other words, phonological paragraphs should be to a great extent iso¬
morphic with grammatical and lexemic ones.
Paragraph structure can also be viewed as lexemic, involving how
the semantic resources of the language organize information into a the¬
matic unit for a given kind of discourse. Although a principled means
of approaching the lexemic paragraph has yet to be satisfactorily de¬
vised, quite a few lexemic functions which participate in paragraph

77
structure have been pointed out. Klammer (1971), largely adopting
his functions from Wise (1968), mentions addition of participant, elab¬
oration, comment, specification, contrast, reason, summation, conclu¬
sion, setting, and result. But it has been difficult to show how these
functions are organized linearly, as Longacre has done for the grammat¬
ical units of the paragraph. The set of such functions has seemed
open-ended and at the same time we have not yet found, according to
Pike and Pike, "a way to treat some of them in a linear fashion, in
order to be able to perceive them as something 'going on"1 (1977:364).
This difficulty of formulating the lexemic structure of actual
paragraphs has in recent years caused lexemic structure to become
viewed as "deep" in contrast to the "surface" structure of paragraphs
being grammatical. The problem has become one of showing how lexemic
functions or units were realized by grammatical ones. Otherwise, lex¬
emic paragraphs would be only hypothetical. The problem of the extent
of isomorphism was tackled by Wise (1968) for oral narratives in Noma-
tsiguenga and for written dialogues in two English texts (I Henry IV
and Great Expectations) by Klammer (1971). It is useful to quote at
length the remarks of Wise:
Lexemic structure, as the term is used here, is a variety of
deep structure. For example, "logical subject" and "logical
object," i.e., agent and goal, are constituents of lexemic con¬
structions on the clause level. In levels beyond the sentence
the lexemic order of constituents is, in general, the chronolo¬
gical order of events in the narrative. Paraphrase and syno¬
nymy are important for the analysis of lexemic sentences and
phrases as well as for the analysis of lexemes which are posited
as the minimum lexemic units.

78
In contrast, grammatical structure is a variety of surface
structure. For example, "grammatical subject" and "grammatical
object," i.e., subject and complement, are constitutents of gram¬
matical constructions on the clause level. In levels beyond the
sentence, the grammatical order of constitutents is the actual
surface order in which the narrative is told. Morphemes are po¬
sited as the minimum grammatical units.
The nature of constituent units in lexemic constructions is ra¬
dically different from that of constituent units in grammatical
constructions, but the boundaries of the two kinds of construc¬
tions frequently coincide. However, the boundaries are not always
isomorphic so that the lexemic and grammatical constructions can¬
not be mapped onto each other in a direct, one-to-one manner.
(1968:2-3)
This kind of mapping can be easily done in many cases. For example, the
lexemic function of the introduction of a participant into a paragraph is
usually equivalent to and realized in the theme and this is further rein¬
forced by one of the thematizing cohesive devices—left dislocation in
many cases. The lexemic function of summation is sometimes indicated by
the kind of structure we are calling peak sentences, but this is not al¬
ways the case. Such sentences always represent the peaks of paragraphs
grammatically but only sometimes are these peaks the realizations of such
sentences.
The third and most discussed way to view paragraph structure is gram¬
matically. The grammar of a paragraph is the system of units (Longacre's
current term) or constituents (Wise's term) whose order is the actual
surface structure order of the paragraph. Longacre (1968, 1976, Forth¬
coming) especially has long made the case that paragraphs are grammatical
entities and that each type of paragraph (narrative, explanatory, horta¬
tory, dialogue, precatory, etc.) has its own grammar. He contends that
the units (in his 1968 work he calls them "tagmemes" in positing his

79
paragraph formulae, as we will see below) which constitute such enti¬
ties are ordered linearly, and he posits a tagmemic formula for each
major and minor type of paragraph in Longacre (1968).
As Becker (1965) rightly points out, the description of paragraphs
as grammatical entities owes much to traditional rhetoric. Longacre's
and especially Becker's formulae for expository paragraphs resemble
closely long-employed formulae for paragraph development in rhetoric
texts. Becker's contention is that the majority of expository para¬
graphs are organized grammatically in either a TRI (Topic-Restriction-
11 lustration) or a PS (Problem-Solution) fashion.
Longacre's major contributions to the study of the grammar of para
graphs lie in the aforementioned formulae and in his systematic delinea
tion of paragraph types (Longacre 1968) and discourse types (Longacre
1968, 1976) on the original basis of studying paragraphs and discourses
in twenty-five Philippine languages and dialects. His general formula
for the tagmemes constituting explanatory paragraphs in the various
languages is3
iPRELIM (+TEXT ±EXP0 TREASON ¿RESULT ¿WARNING) ¿TERMINAL
(1968:109)
With TEXT as the only obligatory tagmeme, this formula can clearly ac¬
comodate an indefinitely large number of different patterns, since EXPO
(exposition) "can occur an indefinite number of times in paragraphs"
(p. 109). Longacre views TEXT as "much like the topic sentence of tra¬
ditional rhetoric" (p. 109). We take the formularization of the gram¬
mar of general paragraph types to be a desirable goal, but it is not

80
clear yet how tightly one can be devised for expositions in English.
Longacre's quest for a formula for the linear organization of para¬
graphs serves as a practical model for us; our study focuses on two of
the units Longacre posits in his formula for expository/explanatory
paragraphs.
Longacre has revised his views on two of his units (or tagmemes,
as he called them in 1968), Text and Peak. Originally likening Text
to a paragraph's topic sentence, he indicates that the obligatory unit
in a paragraph is the theme, that piece of information which a para¬
graph is "built around" and which, for an expository paragraph, "is not
different in kind from a thematic participant" in a narrative para¬
graph (Forthcoming: 118). The distinction betv/een a topic sentence and
a thematic piece of information is an invaluable one; especially it is
the piece of information (a participant or whatever) which, through an
anaphora chain, gives thematic units to a paragraph. This most recent
view of Longacre's that a paragraph has an obligatory theme is consis¬
tent with the one we maintain here.
In his earlier view, Longacre considered Peak to be only one as¬
pect of Text. He states that Text "is the PEAK grammatically and lex¬
ically of its paragraph. As PEAK of its paragraph the TEXT contrasts
in placement with the PEAK (BUn or STEPp) of NARRATIVE and PROCEDURAL
PARAGRAPHS which come later in the paragraph" (1968:109).^ In a later
work, he indicates that Peak is the plot-like element optionally pres¬
ent in non-narrative as well as narrative paragraphs (Longacre 1976:
228-9). It comes near the end of and represents a kind of climax or

81
denouement to the paragraph. Longacre's most recent view of Peak
is an even more refined one:
For discourses that are not narrative but still have a climax
of development, the peak may mark "target procedure" in a
procedural discourse, "climactic exhortation" in a behavioral
discourse of the hortatory variety, and a most satisfactory
or "culminating explanation" in expository discourse. . .
(Longacre and Levinsohn 1978:105)
If we look back to the exposition above at the sen¬
tence therein ("I think this is the greatest change that has happened
in our town is to have new people coming in with new blood, new
thoughts, new ideas, new approach to things"), we see clearly that this
sentence represents the culminating explanation and that it is the
peak of this one-paragraph exposition. Our chapter four below is
largely devoted to showing how such sentences indicate the peaks of
paragraphs.
In the remainder of this study we show the extent to which we can
find that certain surface structure cohesive devices are associated
with grammatical units of expository paragraphs.
At this point we should ask why we find surface structure cohe¬
sive devices such as thematizing constructions and peak sentences at
all, since their use is not categorical. Left dislocation, as we will
see in chapter three, is employed in only slightly more than one-fifth
of its potential cases. If left dislocation is optional, as are other
thematizing constructions in English, why do speakers use it at all?
Why is it that grammatical units of the paragraph are sometimes
overtly marked by a specific structural reflex? We will suggest here

82
that such devices occur because they serve certain of a speaker's
pragmatic purposes.
The speaker's primary purpose in exposition is to convey new infor¬
mation to the hearer. For this reason the speaker/producer of a dis¬
course or text takes into account the presuppositions and expectations
of the hearer. The hearer expects the speaker to supply the necessary
antecedents and background for him/her to follow what is being said. In
other words, they have what is referred to as the
"Given-New Contract" of cooperation. The speaker must be aware of what
the hearer knows, which information has already been given to the hearer,
and which information will be new to the hearer. The hearer must make
the effort to follow the progress of the exposition.
It follows that any means which facilitates the fulfillment of this
contract and which makes an exposition easier for a hearer to follow may
well be used. While such thematic devices as left dislocation can appa¬
rently be used also for other pragmatic reasons, basically they seem to
facilitate this given-new contract between the speaker and the hearer.
Thematic devices present to the hearer what new information will be com¬
mented on in an ensuing message. This piece of information represents
what will be thematic for what follows and has the potential for being
the antecedent for the following block of discourse. It is only if the
hearer has such a piece of information, the theme, clearly in mind that
he/she can compute what the speaker says.
It is probable that the use of cohesive devices is especially fa¬
vored by a speaker in certain pragmatic situations. Left dislocation,

83
for instance, seems more likely to be used when a speaker wishes to be
emphatic, to show a contrast, or to make a particularly abrupt point.
Although the primary function of left dislocations is to introduce new
themes, our corpus has thirty five instances of left dislocation invol¬
ving given information, when a speaker wishes to reemphasize a piece of
information emphatically. In nearly ten percent of our left disloca¬
tions (45 out of 606) we find the dislocated NP immediately preceded by
"but" or "however," showing that the speaker contrasts those NPs with
earlier ones. In seven instances, we find that a left dislocation is
immediately followed by an identifying sentence, as in
1) Threshing floors, that's what they had. (Ill2-27)
This shows that a speaker wishes to make an especailly abrupt point
about a theme.
In the remaining chapters of this dissertation a strong case is
made that we cannot understand how expository discourse is organized
without considering the supersentential or discourse context of cer¬
tain cohesive devices. If this is true, it seems to imply something
about how a speaker communicates to a hearer. Cohesive devices, the
formal signals of the language in which a discourse is expressed,
seem to aid a speaker in communicating the hierarchical structure of
a discourse linearly. This is another way of saying that such devices
communicate the outline (implicitly, of course) of an ongoing discourse
to a hearer.

84
Lonqacre's Typology of Discourse Genres
It was indicated in chapter one that for practical reasons the
fieldworker chose to elicit expositions rather than other types of oral
discourse in the White Pine community. It was apparent that exposi¬
tions were the type most readily elicitable from the cross-section of
people that he wished to interview. Also it was decided that the col¬
lection of a large corpus of expositions rather than smaller corpora of
two or more types of discourse would be more conducive to generaliza¬
tions about expositions.
That exposition is a distinguishable discourse type is not an ad
hoc presumption on our part. It is a hypothesis based on the typology
of discourses developed by Longacre (1968, 1976). In his study of
twenty-five languages and dialects in the Philippines, Longacre and his
colleagues undertook the basic spadework of sorting out which discourse
types could be described across the languages studied. The typology of
discourses which resulted has provided a practical framework for fur¬
ther investigations with other languages, as this study is an investiga¬
tion into the English of the Southern Appalachian region.
Limiting himself to prose monologues, Longacre indicates that four
genres of such monologues (narrative, expository, procedural, and hor¬
tatory) have common characteristics cross-1inguistically and that
5
within any given langauge there are a finite number of other genres.
We may characterize the four main genres as follows: narrative dis¬
course relates a story of some kind; expository discourse explains
something or exposes a subject; procedural discourse (e.g. a recipe)

85
tells how to do something; and hortatory discourse endeavors to bring
about a change of some kind.
The major parameters along which these four genres are distin¬
guished are shown in Figure 2.1 on the following page and adopted from
Longacre (1976). The rows in this figure distinguish those genres
which are +succession (having a sequence in time) and those which are
-succession (lacking a sequence in time). The columns distinguish the
+projection (occurring in the future) from the -projection (those hav¬
ing accomplished time or in which the time frame is not focal) genres,
longacre's scheme further indicates the person orientation of each of
the four genres.
According to Longacre's typology, narrative discourse characteris¬
tically involves a sequence in time, occurs in an accomplished (past)
time frame, is agent-oriented, and is related from the first or third
person point of view. Procedural discourse also involves a sequence in
time, but occurs in a projected time frame, is patient-oriented, and is
not related from the point of view of a specific person. Hortatory
does not involve a sequence in time, is set in projected time, is ad¬
dressee-oriented, and is related from second person point of view. Hor¬
tatory discourse strictly speaking does not have a time frame but a
mode—since the speaker is wishing or urging that something take place.
Expository discourse need not involve a sequence in time nor does it
necessarily have either an accomplished or a projected time frame (time
is not focal). Expository discourse is subject matter-oriented and has
no necessary person reference (point of view). It is essentially im¬
personal and may involve either inanimate or animate entities. If

86
-PROJECTED
+PROJECTED
s
NARRATIVE
PROCEDURAL
u
. 1. 1/3 person.
I. Non-specific person.
c
2- Agent oriented.
2. Patient oriented.
c
3. Accomplished time.
3. Projected time.
+ E
4. Chronological linkages
4. Chronological linkage.
s
.
s
I
0
N
s
EXPOSITORY
HORTATORY
u
1. No necessary person reference.
1. 2 person.
c
2. (Subject matter oriented).
2. Addressee oriented.
c
3. Time not focal.
3. (Mode, not time).
-E
4. Logical linkage.
4. Logical linkage.
s
s
I
o
N
Diagram I. Deep structure genre
Figure 2.1 Deep Structure Genre
(From Longacre 1976:200)

87
it brings in people," Longacre says, "they are simply subjects of
explanation and analysis" (1976:199).
More recently Longacre has furthered elaborated his scheme by
adding another parameter (±tension). What he says about tension char¬
acterizing some discourses is worth repeating:
The final parameter ±tension also applies to all the above dis¬
course genres. Narrative discourse is purely episodic if it
contains essentially no struggle or plot (for example Willa
Cather's novel Shadow on the Rock). Such discourse is consi¬
dered to be -tension. Most stories, however, involve some
sort of struggle or plot and are clearly +tension. Similarly,
in procedural discourse, there are some discourses which are
more or less routine, while others involve struggle and alter¬
natives and are therefore +tension. Both behavioral [horta¬
tory] and expository discourse have varieties in which argu¬
mentation is assumed. These are also +tension. Discourses
of these genre which do not have this characteristic are
-tension. (Longacre and Levinsohn 1978:104)
we mention this third parameter because it concerns the presence or
absence of a peak in the discourse: “If a discourse is plus tension,
there will most likely be some kind of climax of development, some
marked surface structure ‘peak1 (Longacre, 1978 ). This is a pri¬
mary concern of chapter four.
Longacre's characterization of expositions appears to have consi¬
derable validity for the expositions constituting our corpus, although
it must be pointed out that Longacre's types are "pure." In actuality
discourses often of different types are embedded within one another.
For example, within an exposition may be embedded a short narrative
for background or illustrative purposes. Longacre recognizes this in
viewing his genres as deep structure ones which are sometimes skewed in
the surface structure (1976:206-9).

88
Notes
1. The parenthetical notation at the end of these and all further
data in this dissertation refers to the page of typed transcript
on which they are found. For information on which speaker any
data are from, consult the list of informants in Appendix A.
2. Terminological obfuscation notwithstanding, we use lexemic to
refer to concepts such as those of Klammer and Wise, including
contrast, comment, addition of participant, specification, which
are manifested in linguistic structure in various ways. Other
tagmemicists (e.g. Pike and Pike 1977) use "referential" to
refer to such concepts. Stratificationalists use the term "seme-
mi c. "
3. For Longacre, terms "explanatory" and "expository" are equivalent
in his 1968 study.
4. Step and BU (Build Up) are the obligatory tagmemes in procedural
and narrative paragraphs.
5. In his 1976 work, Longacre does not distinguish description from
exposition and we will consider description a subtype of exposi¬
tion here.

CHAPTER III
THEMATIC DEVICES
General Remarks
It is indicated in chapter two that a speaker may use certain
constructions known as cohesive devices to signal to a hearer what
the theme of a following block of the speaker's message will be.
Theme is the obligatory unit of such a block or paragraph and is de¬
fined as the piece of information (not necessarily a human participant)
which initiates an anaphora chain and which the block is about. Al¬
though it must be generalized to aDply to all paragraphs, this narrow
definition of theme demonstrably applies to those paragraphs in our
corpus beginning with thematic devices.
In this chapter our concept of theme is compared with others
(Halliday 1967-8, Firbas 1966, Jones 1977) and we discuss briefly those
constructions that these treatments (especially Jones) consider to be
thematic devices. The main portion of the chapter explores the nature,
function(s), and distribution of one particular device—left disloca¬
tion. This is one very common device through which paragraphs are
formed in this corpus of expository monologues. Finally we look at
existential sentences in the corpus and see how they also function as
thematic devices. They also complement left dislocation.
More often than not themes are reflected by anaphora chains ex¬
tending for more than one sentence. An anaphora chain is a string of
89

90
substitutes (pronouns, for the most part) in a paragraph of language
which refer to a common piece of information or theme and is a special
kind of equivalence class. According to Becker (1965), an equivalence
class includes not only pronouns and demonstratives but also the repe¬
tition of words and their synonyms. Although not every paragraph in
our corpus has an anaphora chain, those which begin with a "left dis¬
located" NP nearly always do. In the two following examples the links
of the anaphora chains are underlined:
1) Fieldworker: Have you participated in any of the Pine Day cele¬
brations here in town? Have you attended any of them?
Informant: Yes, I usually go, because the ladies of our church,
we usually have a booth. We bake things, and so we have a
sale. That's a big day for us^, you know. That's a way to
make uh some money for our uh projects, and thinqs like that,
and that's how we were able to help pay for our uh paint that
recently that we uh, we painted the inside the church. The
ladies did a little painting, but, but we did have enough
cash to buy all the paint. (II14-15)
2) ...This property that he, he let us have then for the parsonage,
build a parsonage of i_t, vt, at one time there was, was a school
there. I was going to relate that to you. There, that's where
that I went to school, and my first eight years of schooling
was right there. And at this old school, that's where this par¬
sonage now sets across from our church. (II12-1)
In the first example, "the ladies of our church" is clearly the theme
of a block of information several sentences beyond its introduction. We
find ten pronouns referring back to it and one repetition of "the la¬
dies." In the second example, the theme is "this property. . it is
around the property that the remainder of the exposition here is orga¬
nized. The anaphora chain contains such a pro-word as "there," since
the piece of information which is left dislocated and which is thema¬
tic is a locus. We see in these examples that theme is not just an

91
intuitive notion but that it demonstrably organizes a block of dis¬
course.
Our view of theme as that information which unifies a block of di
ourse and which often is reflected by an anaphora chain is not incon¬
sistent with other views found in the literature. For Hallidav,
Theme is concerned with the information structure of the clause
with the status of the elements not as participants in extralin
guistic processes but as components of a message, with the rela
tion of what is being said to what has gone before in the dis¬
course, and its internal organization into an act of communica¬
tion. . . (1967:199)
Halliday's position is that theme is involved in both clause and dis¬
course structure. Theme is "what comes first in a clause" and around
which the remainder of a clause is structured (1967:212). But also
theme is that information which "is being talked about" (p. 212) and
which assigns
. . .to the discourse a structure which is independent of sen¬
tence structure and through which the speaker both organizes
the act of communication into a chain or message blocks, the
'information units', and specifics within each message block
the value of the components in the progression of the dis¬
course. (p. 196)
Halliday's express concern, however, is not to show how a theme parti¬
cipates in discourse but to explore the many choices that a speaker
has for theme and how a clause is organized around such choices. To
support his clain that "the soeaker has within certain limits the op¬
tion of selecting any element in the clause as thematic" (p. 212), he
shows the variety of different elements which may be fronted and thus

92
thematized in sentences with inverted word order (pp. 214-5). In sen¬
tences 3 and 4 below, for example, a direct object and a temporal ad¬
verb are thematized:
3) These houses my grandfather sold.
4) Tomorrow John's taking me to the theatre.
Sentences with such inversions are very infrequent in our data, but Hal-
iday's general claim about the kinds of elements which may be themes
matches closely those thematized by left dislocation shown later in
this chapter.
The present study takes a step beyond Halliday by examinina how a
theme relates to a context beyond its immediate sentence, but our view
of theme is consistent with Halliday's. Nowhere does he state that a
theme's domain is bounded by its immediate sentence, and we may reason¬
ably infer that he would agree with our finding that theme plays an
important role in organizing a paragraph.
In the Prague School view, theme has to do with the information
structure of a sentence. The element(s) of a sentence which are
"given" (previously mentioned in a discourse) constitute that sen¬
tence's theme. The theme need not appear in initial or in any other
sentence position and a sentence need not have any given information.
If it does not, it has no theme (Firbas 1966:268). Danes and Firbas,
contemporary members of the Prague School, continue this tradition.
Firbas' view of theme, however, is much less similar to the one we
take here because it is explicitly confined to the limits of a sen¬
tence. His view is a refinement of the widely known theme-rheme

93
dichotomy of sentence organization originally proposed by one of the
founders of the Prague School of linguistics, Vilem Mathesius, a half
century ago.
Firbas' refinement is that theme is "constituted by the sentence
element (or elements) carrying the lowest degree(s) of Communicative
Dynamism within the sentence" (1966:272). He defines "degree of Com¬
municative Dynamism" as "the extent to which the sentence element con¬
tributes to the development of the communication, to which it 'pushes
the communication forward,1 as it were" (p. 270). Firbas thus views
sentences as organized not in a strict dichotomy but as having parts
with various degrees of dynamism. Since he does not provide a clear
way to measure the dynamism of sentence elements, however, his criter¬
ion for theme is not readily applicable.
Firbas1 conception of theme is different from this study's on sev¬
eral accounts. We view theme as that constituent around which a block
or paragraph of language is organized, whether it be one or more sen¬
tences long. This constituent may be new information (not previously
mentioned) or given information (previously mentioned) in the context
of the discourse as a whole. Firbas' conception restricts theme to a
sentential domain and to being only given information. This conception
is too limited and puzzling, since obviously an element's discourse con
text must be considered before determining its newness or givenness.
The work of Jones (1977) on theme in expository written English
discourse is similar to this study's approach in many respects. Insist
ing that theme should not be characterized in clausal or sentential

terms and that it should not be tied to old (given) information or to
a specific syntactic position, she defines theme as the "main idea" or
"central thread" or "nuclear constituent" at each level of language or¬
ganization, especially at the paragraph level (pp. vi-vii). Deriving
from Pike and Pike (1977), her view of theme is more intuitive than
ours. Although she believes each paragraph has a theme, she takes it
for granted that the written paragraphs she examines are thematically
unified and does not show how an anaphora or lexical chain identifies
a paragraph's theme. In our view, such a chain is the best evidence
for what a paragraph's nuclear constituent is.
Jones' primary concern is the examination of many "grammatical de¬
vices for highlighting theme" (pp. 169-220). Such devices which often
convey the theme of the written expository paragraphs she considers in¬
clude
1) Repetition of a word, phrase, or sentence;
2) Inversion of declarative word order;
3) Special constructions, including passive sentences and sentences
with an initial "as for. . phrase;
4) Rhetorical questions;
5) Pseudo-cleft sentences (as "What is unique about milk is its
richness in vitamins and minerals"; we refer to such sentences
as "WH-clefts" herein);
6) Cleft sentences (as "It is the very mingling of races, dedicated
to common ideals, which creates and recreates our vitality"; we

95
call these "it-clefts" herein); and
7) Conjunctions (six different types: summary, contrastive,
enumerative, continuation, comparison, example).
These devices are employed by authors, according to Jones, to give
special prominence to (and not necessarily to introduce) the theme of a
paragraph and they may be found anywhere in a given paragraph. The
thematic devices we focus on in this study introduce new themes into
discourse as well as give prominence to them. They are always found at
the beginning of paragraphs.
Jones devotes most attention to showing how rhetorical questions,
cleft sentences, and pseudo-cleft sentences highlight paragraph themes.
It appears that these and the other devices she discusses are devices
found primarily in written expositions and thus we should expect to
find a different set of devices in oral ones. In a selection of 200
expositions from the White Pine corpus (five from each informant), we
find only three it-clefts and only four WH-clefts. All seven instances
introduce new information but none establishes an anaphora chain. We
can hardly generalize, however, on the basis of so few instances.
Jones does not mention existential sentences as one device pre¬
senting themes, and she only makes a very brief mention of left dislo¬
cation as a device which "directs special attention" to an element
(p. 172). Even the three devices she gives special attention to are
found to be rather uncommon in written expositions: "Except for con¬
junctions. . . , the other devices all occur rather infrequently in

96
texts. Indeed, the devices would lose their strong impact if they
occurred with great frequency. It is their very occasional usage that
makes them espeically forceful in indicating thematic patterns'1 (p. 220).
Although the devices she discusses are similar in function to the the¬
matic devices we deal with here, the differences in kind and in fre¬
quency for such devices probably indicate that strategies employed in
oral expositions differ from those used in written expositions. The
great frequency of such a thematic device as left dislocation(606 in¬
stances for our corpus) indicates a more prominent role for thematic
devices in era! than in written expositions.
The present study differs from other treatments of theme in being
a data-based study of the nature, function(s), and distribution of the¬
matic devices in oral expository English. For the remainder of this
chapter we examine two such devices, left dislocation and existentials,
and show how they mark a paragraph's theme and indicate to a hearer
what a paragraph will be organized around.
In several cases the analysis in this chapter is based on a selec¬
tion of 200 expositions (five from each informant) and in other cases
on 80 expositions (two from each informant).
Left Dislocation
Although this chapter's primary concern is showing that left dis¬
location is a thematic device rather than displaying its various forms,
such a display yields strong evidence that left dislocation is properly
understood in discourse rather than in sentential terms.

97
Put simply, left dislocation is a construction in which an NP or
noun phrase (simple, complex, or compound) appears before the begin¬
ning of a sentence (i.e. leftmost) and has an appropriate pro-word
copy in the sentence itself, as in 5-7 below:^
1) And my boy, he froze so bad, he had what they call "buck
fever." (V3-27)
2) Well, these arbitraty rules that people set up for themself,
that's something else I don't like, you know. (113-32)
3) Like last year, me and a couple or few of my friends uh, we
watched the queen and everything out there. (XI2-8)
Left dislocation is not a repair construction which represents a
speaker's change of mind and a second attempt at phrasing a thought
after beginning a sentence. It is very rare to find any sort of au¬
dible hesitation or an overly long pause, for instance, between a dis¬
located NP and the sentence which follows, but a short, discernible
pause does characteristically occur. The intonation contour on left
dislocated NPs, no matter how long, is very consistently level with a
slight fall at the end for our data. In another study of left dislo¬
cations, Keenan and Schieffelin found that the NPs were "uttered with
a slight rising intonation. . , .This is then often followed by a pause
or by a hesitation marker (e.g. uhh)" (1976:254). Our findings differ
from theirs on both accounts.
The Variety of Forms
An NP in any syntactic position may be dislocated, whether it is
in the matrix or main clause or in an embedded clause. Most often

98
(356 of the total 606 left dislocations in the data, or 58.7 percent),
as in 5-7 above, it is the matrix subject which is dislocated. This
is not surprising. Halliday indicates that in "the declarative, the
subject if the unmarked theme" (1967:213). He defines theme as that,
in a sentence, which "is being talked about" (p. 212). Therefore we
should expect themes, which information blocks are about, to most repre¬
sent subjects of sentences. -
But the NP may represent the subject of other clauses, such as an
introductory adverbial one (21 instances, or 3.5 percent), as in 8 be¬
low, of an object clause (36 instances, 5.9 percent of the total), as
in 9-10 below, or of a relative clause (9 instances, 1.5 percent of the
total), as in 11:
8) Children., if they. took a cold, why they'd make 'em, some
boneset tea. (il12-34)
9) Just anything, even the commercials on the, on the television,
we thought they were great. (VI3-8)
10) And the newest hospital, I think they were running short of
funds at one time, (XIII-8)
11) And then Jake Butcher, you know the bank that he's the head of.
(11-18)
In thirteen instances, the subject of an object clause is dislocated to
the front of the clause and not of the sentence, as in 12:
12) And I reckon the man that made it, he thought he couldn't make
a living of it. (IV2-13)
Left dislocation thus can involve the subject of any clause in a sen¬
tence. It is just as wide-ranging for objects of verbs, although it

99
is not as frequent. For example, in sentence 13 the object is dis¬
located from the matrix clause and in 14-5 from embedded clauses:
13) And whatever's to do, she takes it. (VI14-19)
14) Even public records, there's no place to display 'em. (14-14)
15) And my mother, I mean I've heard her tell lots about Mr.
Lawson Franklin, (III2-2)
In the data are 53 instances (8.7 percent of the total) of matrix ob¬
ject dislocations and 14 instances (2.3 percent of the total) of embed¬
ded object dislocations.
Left dislocations can also involve the object of a preposition
from the matrix clause (36 instance, 5.9 percent of the total), as in
16 below, from an object clause (8 instances, 1.3 percent), as in 17,
or an introductory adverbial clause (1 instance, 0.2 percent), as in
18:
16) My first car I had, I done most all the work on ft. (113-5)
17) All that white lightning and everything, you know they got
into vt somehow. (VII3-23)
18) And our husbands, if it wasn't for them, there's no way. (V1 -10)
This various display of forms illustrates that left disclocation can
involve NPs in many positions (more are given below). But it is not
meant to imply an essential difference between matrix and embedded
clause NPs with regard to dislocation. Indeed, such a distinction is
sometimes misleading because examples such as 8 above can be viewed as
dislocations from either the matrix or the introductory adverbial
clause. Those like 10 above have matrix clauses such as "you know,"

100
"I mean," "I guess," and "I think." Such clauses add little or no in¬
formation to the theme of a passage. They differ insignificantly
from dislocations from matrix clauses. If we add these to all the dis¬
locations of matrix subjects, objects of verbs and of prepositions, we
find that 92 percent (506 of 550) of the dislocations are of the matrix
kind. Our contention in this chapter is that left dislocation intro¬
duces thematic elements into the stream of discourse. Since it is
quite reasonable that newly-introduced participants and thematic pieces
of information are most likely found in the central part of a sentence,
it is no surprise that matrix clause NPs are by far the preferred ones
for left dislocation. However, the non-matrix dislocations are of the¬
matic elements. But clearly it is usual that they first appear in a ma¬
trix clause.
Three other types of NPs which can be dislocated and replaced by
an appropriate pro-word are locatives (22 instances, 3.6 percent of the
total), as in 19-20 below, temporals (4 instances, 0.7 percent), as in
21 below, and possessives (23 instances, 3.8 percent), as in 22-3 below:
19) Well, at White Pine, you know, you're really more of an indivi¬
dual there. (VI11-1)
20) Now where Earl Atchley was raised, I mean over at Rankin side.
there was a big coal chute over there that furnished. , .
(II12-6)
21) Back when that thing was a-running, well, I didn't know nothing
about it then. (X2-28)
22) This little James boy lived right down the road from here, just
the othere day my pastor had his funeral. (VI14-16)
23) Some of these very people who would criticize uh the rural areas
of East Tennessee, their ancestors came out of them. (X4-11)

101
The dislocation of two NPs from one sentence, as in 23 above, is quite
rare (two instances in the data). Since we contend in this chapter
that left dislocation introduces new themes into a discourse and that
only one theme occurs per paragraph, it might be asked whether these
double left dislocations introduce two themes and are thus counterex¬
amples to our contention. Such sentences are clearly not counterexam¬
ples, however, once their contexts are examined. The context of sen¬
tence 23 shows that "rural areas of Tennessee" is not a theme newly
introduced by the dislocation in that sentence.
24) . . .Uh each community has its own brogue, its own way of com¬
municating uh to one another, and uh I know, I hear this cri¬
ticism all the time. Well, the people in, in the small towns
are, are illiterate. Uh they're not polished. They're unedu¬
cated, they're uncouth, uh which is really not so. Uh we have
some very brilliant people in, in the small rural areas of
Tennessee. Uh some of these very people who would criticize
uh the rural areas of East Tennessee, their ancestors came out
of them, and in many ways they're criticizing their own back¬
ground, their own family tree. (X4-11)
In this case we have two pieces of information very closely bound togeth¬
er, but this sentence represents a shift in theme. The new theme is
clearly "some of these very people. . ."
Left dislocation sometimes involves a clause intervening between
the NP and the sentence itself. In these instances the speaker appa¬
rently chooses to provide further information or background so that the
dislocated NP may be clearly understood by the hearer. This apposi-
tional kind of information may be a restrictive relative clause, as in
22 above or in 25 below, or it may be an entire sentence or two of in¬
tervening information, as in 26-7:

1C2
25) Well, this house I moved in, _U's an old house. (IV2-5)
26) You know, a man that uh, that had been raised up in a, an envi¬
ronment, you know, where he uh, where they had no real val¬
ues, you know, where he didn't really know his parents, his,
and that was offered no real guidance or whatever, you know,
uh he's more apt to, to fall back on the things that. . .
(112-17)
27) And Frances and I, the one that lives in Florida is ill, we
were swinging and singing songs. (IV1-8)
In addition to 26-7, five further examples of left dislocation
working supersententially are found in the data, further evidence that we
cannot understand left dislocation as a sentential process or operation.
Some linguists have offered a sentential analysis, however, and we
briefly examine their claims, Ross posits an optional rule for left
dislocation (1967:232ff.). His formulation is as follows:
X - NP - Y
1 2 3
2# [1 2 3]
+Pro
Optional
He indicates that, since left dislocation can apply to NPs indetermi¬
nately far from the beginning of a sentence, the "distance that the dis¬
located NP has traveled. . .suggests that the statement of the rule
must make crucial use of a variable" (p. 234). This variable is repre¬
sented by "X" in his rule, Ross adheres to the view that left disloca¬
tion is a transformation operating not only within the sentence but
also within the main clause: "Left Dislocation must be restricted to
main clauses" (p. 237).
Postal takes the sentential point of view as well (1971:135-7).
He claims that left dislocation is not bound by cross-over constraints

103
and that it is derived from reduced introductory phrases such as "As
for. . His analysis of the restrictions on left dislocation is in- â– 
correct in the restriction to only one NP per sentence (cf. 23 above)
and to only main clauses.
An accurate account of left dislocation thus confounds those wish¬
ing to give it a sentential analysis. Any data-based account (and Ross
and Postal's are not) discovers that we must consider left dislocation
in its discourse context, that we must take more than the immediate sen¬
tence into consideration. This is especially true since the amount of
appositional information a speaker chooses to provide the hearer is in¬
determinate. Example 28 below shows the apposition may consist of sev¬
eral sentences; In it "there" refers to "at Rankin side. . ."
28) Now over at the coal chutes, now where Earl Atchley was
raised, I mean over at Rankin side, there was a big coal
chute over there that furnished the coal for the steam en¬
gines, you know, and uh, well, we had the two stores here
before the lake come that furnished people, I mean, all of
the rural area. That's where we depended on our livelihood,
on such things as we didn't grow on the farm. Why, we went
to there, and uh. . . (Ill2-6)
If the dislocated NP represents the thematic piece of information
of a paragraph, there should be an upper limit or threshold to the
amount of information constituting the apposition (in order for a
speaker to assume that the theme is still in the hearer's conscious¬
ness). This is a matter, however, of studying the limits of short-term
memory and of the cognitive basis of thematization and paragraph struc¬
ture. The relevant issues involved are reviewed by Van Dijk
(1977:155ff.).

104
Specifications
In the above section it is shown that any NP in a sentence has the
potential for being dislocated. Earlier it is noted that the NP dislo¬
cated must be the unifying piece of information, or theme, of a block
of discourse (this is an important point and is demonstrated in the
section "Relationship to Subsequent Discourse" below). A theme, which
in a well-formed block of discourse or paragraph often establishes an
anaphora chain, represents something a speaker wishes a hearer to have
in mind so that a body of information can then be built around it. The
theme thus is information the speaker wishes to be in the forefront of
the hearer's consciousness. It is NEW information, which Chafe regards
as knowledge which : speaker assumes to be not in the "consciousness of
the addressee at the time of the utterance" (1976:30).
All thematic devices such as left dislocation, existentialization,
and clefting constructions usually introduce new information into a
stream of discourse. But all such devices do not introduce exactly the
same kind of information. Moreover, they are not interchangeable.
As a rule, existentials do not introduce definite NPs. The excep¬
tions to this are treated in Reed (1979). Reed claims the criterion for
a noun being existentialized is whether it has a "presupposition of
identifiable reference." In Chafean terms, this means "knowable" to the
hearer. Without exception, left dislocations do not introduce nonspecif¬
ic indefinite pieces of information into discourse. Knowing this, we
not only can relate these two thematic devices to various comments in
the literature, but we also can point out the important differences and
overlappings between them.

105
New information may be either definite or indefinite; if indefi¬
nite, it may be specific, nonspecific, or generic (see Stockwell et
al. 1973:67-95 for discussion of these). In the view of Li and Thomp¬
son (1976), a left dislocated NP represents the "topic" of a sentence,
the part of a sentence regarding which the rest of the sentence makes
a "comment." Although they avoid considering topics in their dis¬
course contexts, they maintain that across languages topics are always
either definite or specific NPs. They state that "one of the primary
characteristics of topics, then, is that they must be definite"
(1976:461). Givon agrees with this characterization of topics and
points out that they also cannot be referential-indefinite" (1976:154).
Here we view left dislocated NPs as discourse themes rather than sen¬
tence topics. But the left dislocations from the White Pine data agree
closely with their characterization of topics.
The overwhelming majority of the 606 instances of left disloca¬
tion in our corpus are of definite NPs. The themes which are left dis¬
located are so often [+definite] that it can be said without reserva¬
tion that left dislocated NPs are normally definite. In a relatively
small minority of cases they are generic, as in sentences 29 and 30
below:
29) Yeah, a funeral to me, it's not for the dead person. (XI1-18)
30) Children,, if they. took a cold, why they'd make 'em- some
boneset tea.' 0112-34) 1
Or they are specific indefinites, as in 31 and 32:

106
31) One of the biaqer Blacks, he's about twenty four years old.
(13-2)
32) People my age that went North and worked, some of 'em have
come back [to] the territory. (IV4—18)
But not in one instance is the left dislocated NP a nonspecific indef¬
inite.:. In her study, Prince (1978b) agrees that nonspecific indefi¬
nites cannot be dislocated. In general, her contention that specific
indefinite NPs. which are dislocated represent subjects is consistent
with the data from our corpus. Of the several dozen dislocated specif¬
ic NPs, only three are not subjects, two of which are 33 and 34 below:
33) Certain people who have money in this town, people look up to
'em. (.13-6)
34) Well, each, each section, I think they call it a pod. (VI-6)
Such examples are equivalent to passive sentences with specific indefi¬
nites dislocated from the subject position.
Another contention of Prince, however, is disaffirmed by the data
from White Pine,. She claims that left dislocated subjects are usually
new information while left dislocated non-subjects reintroduce old in¬
formation.. It is shown below that nearly all left dislocations repre¬
sent new information, regardless of which syntactic position in the sen¬
tence they are related to.
Further evidence for all left dislocated NPs not being nonspecific
indefinites, is available. There was not a single instance of an indefi¬
nite pronoun serving as the copy of the dislocated NP. Our data leave
no doubt that left dislocation is a thematic device constrained not to

107
operate on certain NPs, those indefinites which may be characterized
as [-specific], even though one major study (Gundel 1974) argues other¬
wise, in fact. Gundel argues for the acceptability of sentences with
nonspecific indefinites, as 35 and 36:
35) A misprint, the proofreader didn't see one.
36) A phone call, he didn't charge me for one.
but these judgments are entirely her own. Her analysis of the con¬
straints on left dislocation, based wholly on data not taken from actual
speech, contends erroneously that noun phrases which contain quanti-
fier-like determiners (with the exception of ONLY and EVEN) cannot be
dislocated (PP. 62-3). Many of the instances from our corpus, including
37 and 33 below, show this cannot be true:
37) Most of my friends, the people with whom I was raised, we all
went to church. (IVI-8)
38) Well, each, each section, I think they call vt a pod. (VI-6)
A non-data-based analysis thus can only guess at the constraints on left
dislocation. Gundel also goes for a sentence-domain treatment.
Obviously it is only a large corpus such as ours which can show
what kinds of NPs can be dislocated. There is no reason to believe that
nonspecific indefinites are thematized by left dislocation. It is shown
below that they are introduced by existentials when thematized.
Left dislocation, as we define it, involves a pro-word copy of the
NP which is thematized. There is evidence from our corpus that a
fronted [+definite] thematic NP need not have such a copy and that

108
therefore left dislocation is one type of a more general construction.
Our corpus has 61 instances of constructions without the copy, and
these are of two different sorts. One such construction involves the
repetition or embellished repetition of the fronted NP, as in 39 and
40 below:
39) Well, this town, I love this little old town. (112-31)
40) And then Tate Springs, I know you've heard of Tate Springs.
(IV1-10)
These sentences show that a speaker may choose to repeat a fronted NP
rather than use a copy. Example 41 and 42 show a second construction
a speaker may employ. Here a theme is announced, and its second ref¬
erence is a paraphrase of the first.
41) And our uh supermarkets, we have two or three nice grocery
stores now. (II14-17)
42) Uh housing, seems like everybody has a new house, you know.
(V4-15)
Such patterns do not relate the fronted NP to the rest of the sentence
syntactically in the same way as left dislocation does. But they re¬
semble left dislocation in introducing a new piece of information, a
potential theme, into discourse. Also the intonation of such patterns
is identical to that of left dislocations.
As mentioned earlier, left dislocations are obviously not repaired
sentences representing a speaker's second attempt at phrasing. Nor are
the paraphrasing patterns of embellishing or narrowing down an NP such
repairs. In all examples with fronted NPs we find a characteristic

109
juncture between the NP and what comes after, but rarely is there
any sort of hesitation phenomena ('tih" or a long pause, for instance)
which would indicate that a speaker might be changing course after
beginning with an NP. Other than appositional information, it is also
rare for anything to intervene between the NP and the sentence in which
it is thematic.
Functions
To understand the functioning of left dislocation in oral exposi¬
tions, we must determine 1) the relationship of left dislocated NPs to
preceding portions of discourse; 2) the relationship of left dislocated
NPs to subsequent portions of discourse; and 3) the distribution of
left dislocated NPs over entire expositions.
As is remarked above (p. 105), left dislocation usually involves
pieces of information which are NEW, i.e. knowable to or recognizable
by the hearer from outside the immediate verbal context. Of the 606
instances of left dislocation identified in the White Pine corpus, 571
involve such NEW information (94.2 percent of the total, that is).
Thirty-five instances (5.8 percent of the total) involve GIVEN informa¬
tion (which the speaker would probably presume to be in the conscious¬
ness of the hearer because it had been mentioned in a closely preceding
context). These seem to involve the speaker's giving a special empha¬
sis to the left dislocated piece of information.
Four types of dislocated NEW information suggest themselves.
These types are not mutually exclusive. For example, a left dislocated

no
NP may be both contrastive and reintroduced.
1) The majority of left dislocated NPs represent new partici¬
pants or other pieces of information which a speaker had not pre¬
viously mentioned. Of the total of 606 instances, 488 (80.5 percent)
are primarily of this type. In introducing new participants or char¬
acters or in indicating the principal object or location of discus¬
sion for the ensuing sentence (and further discourse usually), left
dislocation is thus a stage-setting device. We see three examples
of this in the following discussion of his immediate family by a sixty-
two-year-old man:
43) Mother, she's been dead fifty two year, (hmm) and Dad, he's
been dead about uh fourteen year, (hmm) So therefore uh,
I had a sister, she's been dead, oh, I'd say twenty five
year, and I had a half-brother died nine months old. He
was a twin, and the other one, ire died of meningitis. He
choked to death. That's, that's been uh thirty, thirty
years, I guess, or maybe a little longer than that, just
not knowing the time. (IV2-2)
Introducing new information for the first time is thus the primary func¬
tion of left dislocation. This is true whether the left dislocated NP
represents a subject or an object, whether it repersents an element in
the matrix or an embedded clause. Prince (1978) claims that left dis¬
located subjects tend to represent new information while left disloca¬
ted non-subjects in general represent old information. There is no
support for these contentions in this corpus' data. Nearly 95 percent
of left dislocations serve to introduce new information, regardless of
whether they are subjects or not.
2) Another type of left dislocation represents a special kind of

Ill
new information. The left dislocated element will name a general class
of items, and the following sentences will specify which and/or how
many members of the class are involved. In other words, some left dis¬
locations involve the focusing on or narrowing down to a portion of the
new information which is introduced. There are 34 instances (5.6 per¬
cent of the total) of this type in the corpus. Two short examples of
this are given below:
44) People my age that went North and worked, some of 'em have
come back [to] the territory. (IV4-18)
45) These students, I know several of 'em because I’ve helped the
teacher average their grades. (VIII1-18)
We note the similarity between such constructions and those like sen¬
tences 39 and 40 above.
3) On occasion, a speaker wishes to introduce a piece of informa¬
tion to a hearer abruptly in order to contrast it with a foregone piece.
In such instances the new information is often preceded by "but" or "how¬
ever" and frequently it is left dislocated. In our corpus are ten in¬
stances (1.7 percent of the total) which are primarily contrastive and
several dozen more whose constrastiveness is evident but apparently sec¬
ondary. Although usually preceded by one of the conjunctions mentioned
above, their primary characteristic is the heightened stress on the left
dislocated element which they have. In the following examples we have
contrastive stress on "younger" and on "old":
46) . . .and I think the older people think there’s no use in all
this change, you know. Why, we've gone like this for years.
Why not continue? (uh-huh) But I think the uh younger,
younger people in town uh, I think they want change. I, I
think they want it to grow and develop, you know, . . .
(VI -4)

112
47) . . .of course, as I say, the young ones I don't know. But
them old ones, if they liked you, they'd bend over back¬
wards for you. They'd do anything in the world that, for
you, if they liked you. (X2-16)
Such left dislocations serve not only to indicate a contrast with a pre¬
ceding piece of information. They also serve to introduce thematic ele¬
ments for the succeeding sentence(s) like other types of left disloca-
ti on.
4) Often a speaker in the course of an exposition wishes to rein¬
troduce a piece of information back into the stream of discourse after
it has fallen into the background of the hearer's consciousness. The
information may have been mentioned in an earlier turn in the conversa¬
tion or earlier in the same turn. But when a speaker wishes to rein¬
troduce it, it is often found in a left dislocation structure. Appa¬
rently the constraint on this type of left dislocation is that at least
one thematic piece of information has intervened since the last mention
of the reintroduced one. This intervening material may be from one
sentence to many dozens of sentences long. In our corpus, 29 of the
606 instances (4.8 percent of the total) are primarily of this type.
In the following excerpt, "sulfur" is reintroduced as a thematic piece
of information:
48) . . .my aunt used to buy calama. You used to could get calama
just uh loose, I mean, at the drugstore, and she would just
take what calama, she said that calama in the spring, that
everybody needed a dose of calama to, to clean their system,
and so she would take, and just what would lay on the end of
a, of a knife, table, that we would, why, she'd give us, all
of us children a dose of calama in the spring, and we'd uh,
and uh sulfur and molasses. That's another thing. She

113
would mix a half a teaspoon of sulfur and uh something like
a tablespoon of molasses, mix that sulfur in that. She said
there's nothing as good, and she was the, she was the doctor
of our family, I reckon, for she was always prescribing some¬
thing, you know. If we'd get anything, what to do. (uh-huh)
And uh, but now that sulfur, I mean, she said that purified
the blood, and so that, I didn't, I still don't understand
that. But I took a lot of it, I'll tell you. It didn't hurt
me. If it didn't help, it didn't hurt. (II12-36)
''Sulfur" subordinates and replaces "my aunt" as the thematic piece of
information towards the end of this passage.
As mentioned above, thirty-five of the 606 left dislocations repre¬
sent given pieces of information. When these thiry-five pieces of in¬
formation are dislocated, they are neither mentioned in the discourse
for the first time nor reintroduced after the intervention of a differ¬
ent thematic piece. They are not mentioned to bring something into the
forefront of a hearer's consciousness; they are mentioned to keep that
piece of information in the forefront. In other words, these disloca¬
tions represent especially emphasized pieces. In the following excerpt
"Mrs. Seabrook" is left dislocated the second time mentioned (obviously
for emphasis) rather than when she is introduced:
49) . . .they had their own Black children went to this school (uh-
huh) and their Black teacher. And uh I'll tell you a teacher
that used to teach over there was Mrs. Seabrook. She died
two years ago, up here in this school. Oh Mrs. Seabrook, we
all liked her very much. She was, that was, that was a sit¬
uation that I thought that worked just fine when it came to
integration and all like that. (11-36)
"Mrs. Seabrook" is definitely not new information when it is dislocated.
Given NPs, when dislocated, are often accompanied by interjections such
as "oh," "why," "boy," and others.

114
Before examining how left dislocated NPs are related to subse¬
quent portions of discourse, we examine here more closely which types
of NPs are dislocated, and their distribution is also examined here.
If left dislocation usually introduces new information and is a
stage-setting device for the ensuing discourse, we would expect numer-
our dislocated NPs to come at the very beginning of an informant's
response to a question, where the speaker is probably most likely to
make introductions. Of the 606 instances of left dislocation in our
corpus, 67 (11 percent) are of this kind. In the example below, "at
White Pine" is dislocated in order to set a stage or frame of refer¬
ence for what comes after:
50) Fieldworker: How do the two schools compare?
Informant: Well, at White Pine, you know, you're really more
of an individual there. When we moved out to our new high
school, you were just like a person in a crowd, but as you
get older and stuff like that. . . (VI11-1)
In general left dislocation does tend to involve NPs coming early in
expositions rather than late, but it is not a very strong function of
discourse-initial position.
If left dislocation is primarily a character-introducing device,
we would expect a large majority of dislocated NPs to be animate ones.
Because left dislocation can involve locative and temporal pieces of
information, it is clearly not restricted to only animate NPs. About
two-thirds (402 of the 606, or 66.3 percent of the total) of the dis¬
locations are animate, a strong if not overwhelming .. majority. One-
third then introduce inanimate NPs newly into discourse.
If left dislocation were a device for introducing active partici¬
pants rather than any thematic character, we would expect most

115
dislocated animate NPs to represent agents. On the basis of the data,
however, we find that left dislocated animate nouns are likely to be
agents only about half the time (215 of the 402 animate NPs, or 53.5
percent).
The large number of dislocated NPs which are inanimates or non¬
agents in exposition should not be surprising. Longacre claims that
in its deep structure expository discourse is subject matter-oriented
and not agent-, patient-, or addressee-oriented (1976:200). In a cor¬
pus of narratives, on the other hand, we would probably find a much
higher percentage of agents among dislocated NPs. Longacre also
claims that the subject matter orientation of exposition is generally
realized by equative and descriptive clauses (p. 202). If we examine
the 402 instances of dislocated animate nouns, we find that 94 of them
(23.4 of the animate ones, that is) are from such clauses.
Probably the large number of dislocated NPs which are animates
and agents in our expositions should not be surprising either. It is
much too simplistic to assume that any given genre of discourse con¬
sists primarily of just one or two types of clause. Also, we must
remember that discourses are frequently embedded into one another. In
our corpus' expositions, we frequently find embedded narrative and
other episodes which are person- rather than subject matter- oriented.
So far we have examined the kinds of NPs dislocated, their posi¬
tions, and their functions only with reference to the total number of
left dislocations identified in the corpus — 606. An examination with

116
reference to actual potential for dislocation would probably better
show us which NPs are most likely to be dislocated. What percentage
of the animate nouns introduced by speakers are left dislocated? What
percentage of the inanimate ones? Only after answering such questions
can we say that animate nouns are more likely to be left dislocated.
Counting potentially dislocatable NPs has been done here for three
kinds of NPs: names, thematic NPs which are not nonspecific indefinites,
and such NPs which are animate (to determine if animate NPs are more
likely to be dislocated than inanimate ones). In this way, two ques¬
tions about potential for dislocation can be answered: how likely are
names (persons and non-persons) to be dislocated in expository dis¬
course, and how likely are thematic pieces of information of all kinds
likely to be dislocated in expositions?
To answer the first question, the names in a subset of 200 exposi¬
tions (fiveofrom each informant) were counted. Ninety-three instances
of names were found, 27 of which were dislocated. Twenty-nine percent
of the potential names, therefore, were dislocated.
To answer the second question, the thematic pieces of information
in a subset of eighty expositions (two from each informant) were
counted. In many cases, thematic elements were difficult to identify
because the speaker changes direction or introduces further, unrelated
information before they are clearly established (and before they ini¬
tiate an anaphora chain. As a result, we can only approximate the num¬
ber of potential cases. In the eighty expositions approximately 280
such potential cases were found, and fifty-nine of them were left dis¬
located (21.1 percent). Roughly speaking, then, between one-fifth and

117
one-fourth of the potentially dislocatable NPs are dislocated. In
addition, the subset of animate themes were counted to see if a
higher percentage of them were dislocated than inanimate ones. If
they were, this would have indicated that left dislocation is a
character-introducing device. It was found, however, that a slightly
lower percentage (33 of 160 instances, or 20.63 percent) of the ani¬
mate NPs were dislocated than the inanimate ones (26 of 120, or 21.67
percent).
Although the constraint is slightly stronger with names, it is
apparent that the overriding factors accounting for left dislocation
are the functions we have discussed (contrastiveness, emphaticness,
newness especially), given that the NP is thematic. At least for our
corpus of expositions, it makes no significant difference whether the
NP is animate or not.
Relationship to Subsequent Discourse
It has been seen that left dislocation is a device which usually
introduces new information into the stream of discourse. Also it has
been shown that left dislocated NPs may initiate an anaphora chain and
represent the information around which subsequent discourse is orga¬
nized. But how typically do such NPs do this? Before it is concluded
that left dislocation is a thematic device which normally organizes
paragraphs, we must see whether the discourse beyond the immediate sen¬
tence where the left dislocation occurs is still oriented around that
NP. If left dislocation introduces themes, it should be atypical for

118
such NPs not to be the center of a subsequent discourse organization.
But it is also necessary to define the relationship between left dis¬
located NPs and the portions of discourse which follow them. To de¬
termine this, we look at the subset of 200 expositions from the cor¬
pus we also used for earlier analyses.
In the 200 expositions, there are 190 instances of left disloca¬
tion. Over two-thirds of the left dislocated NPs are thematic. They
constitute the central bit of information around which the message is
organized beyond one sentence. There are good reasons for the remain¬
ing themes of only one sentence, a point to which we return below.
More than one-half of the total (97 of the 190) are thematic for at
least three sentences. Figure 3.1 on the following page summarizes the
extent to which the 190 dislocated NPs are thematic and how many are
thematic to each extent.
Figure 3.1 shows that left dislocated NPs may be thematic for
quite extensive lengths; nearly one-tenth of the 190 instances are the¬
matic for at least ten sentences. The greatest extent to which one is
thematic is for thirty-eight sentences.
As a thematic device, left dislocation represents a speaker's
strategy of presenting to a hearer what an ensuing block of discourse
will be about. It seems unlikely that a speaker would signal the in¬
troduction of a theme by such a device which would turn out to be the¬
matic for only one sentence without there being overriding reasons for
doing so or an explicit signal that such was the case. Once a theme
has been introduced, it seems likely that it would be maintained for

119
No. of Sentences No. of NPs Thematic Percentage of Total
Beyond Introduction To This Extent
For Which NP Is
Thematic
1
61
32.1
2
32
16.8
3
22
11.6
4
17
8.9
5
14
7.4
6
6
3.1
7
5
2.6
8
8
4.2
9
7
3.7
10
5
2.6
10+
13
6.9
190
100.0
Figure 3.1 Extent To Which Left Dislocated NPs Are Thematic

120
several sentences or at least until a speaker signaled to a hearer
that it were no longer the theme.
In looking at the sixty-one instances of left dislocated NPs in
our corpus which are thematic for only one sentence, we find that in
the majority (46 of 61) of them a speaker employs one or two of the _
cohesive deivces (thematic devices and conjunctions especially) to in¬
dicate to a hearer that a theme will not be maintained. In the re¬
maining fifteen instances the speaker provides some other other expli¬
cit signal. Most often a speaker will use a thematic device, either
another left dislocation or an existential, or one or more conjunc¬
tions, another kind of surface structure cohesive device which will be
our concern in chapter five. These devices indicate to a hearer that
a change in subject or in direction of a speaker's exposition is tak¬
ing place. They are part of a speaker's repertoire of devices for
structuring an ongoing monologue.
In eighteen of the sixty one instances we find another theme ex¬
plicitly introduced in the following sentence. Ten of these involve
another left dislocation, two an existential sentence, and six the in¬
troduction of another character into the discourse. In twenty-eight
instances one or more conjunctions or sequencers or interjections are
used to show the hearer a change in direction. These conjunctions and
other structures are shown in Figure 3.2 on the following page. More
often than not, these conjunctions are preceded or followed by a pause
or a hesitational "uh" which reinforces the change of direction they
signal to a hearer. Also, they often accompany thematic devices and
further strengthen a speaker's change of theme signal.

Sentence-Initial Conjunction(s) Frequency of Use
and + then 8
but 5
and 6
of course 3
well 3
so + then 1
like 1
except 1
28
Figure 3.2 Conjunctions Signaling Change in Theme

122
In the following excerpt, the dislocated NP, "this one son that's
with me" is thematic for only one sentence. The following sentence
begins with both a conjunction and another left dislocation to indi¬
cate a change in theme, for contrast in this case.
51)This one son that’s lives here with me, he'll just sit and
die laughing at stuff that goes on and, and enjoy it, you
know. But the other two, they don't fence with that stuff,
so I don't know. (VIII3-20)
The remaining fifteen instances of a left dislocated NP thematic
for only one sentence are of four different kinds. In five instances,
the left dislocation occurs in the last sentence of the exposition, in¬
dicating the speaker did not wish to pursue the subject further. In
four instances the NP is copied by "that" in an identify sentence, as
in
52) Threshing floors, that's what they had. (II13-27)
We will discuss identifying sentences in chapter four. They present a
summary identification regarding something and it is hardly surprising
that a speaker would introduce a new theme after making such a point.
Another kind of summary effect is given when the sentence follow¬
ing the one from which a dislocation takes place has an intensive
"just" in it, as in 53 below:
53) . . .but the other two, they don't fence with that stuff, so I
don't know. I reckon it's just the way you think. (VIII3-20)
The summarizing effect here indicates this is well-constructed thema¬
tically. Four of the sixty-one instances have this signal.

123
The remaining two instances of the sixty-one which are thematic
for only one sentence involve a speaker's change in perspective. This
is signaled by a phrase such as "I think" revealing an explicit first
person statement. One of the two instance follows:
54) . . .and the others, you know, we might draw them into our
group, and again we might not. We'd make a really good
sociological study, I think. (XI2-14)
The kinds of devices used by speakers in our expository monologues
to indicate a change in theme after one sentence are exactly the same
as those used after longer discourse blocks unified thematically. The
only difference is that we find a greater variery of conjunctions if
we look at 190 instances.
For about a third of the cases (62 of 190) we find an explicit in¬
troduction of a new theme, thirty-three times by another left disloca¬
tion, sixteen times by an existential sentence, and thirteen times by
the introduction of another character into the exposition. In forty in-
tances of the 190 the theme discontinues because the end of the mono¬
logue is reached.
The most common conjunctions used to indicate a change in theme for
the 190 cases are the same as were found for the 61 above. Several con¬
junctions not found (e.g. "anyway") for the sample of 61 instances ap¬
peared, and there are several instances among the larger sample of com¬
binations of conjunctions, as "and then of course" for the example be¬
low, which intensify the speaker's signal:
55) Uncle John Harrison, he uh run that livery stable. And then
of course, they did away with that part of it. . . (VI3-13)

124
it is obvious, however, that conjunctions for our larger sample are
used in exactly the same way as those in the smaller one by speakers:
to indicate in many cases a change in theme for an ongoing discourse.
Conjunctions have functions other than indicating a paragraph's theme,
of course.
What then is the relationship between left dislocated NPs and sub¬
sequent discourse? An answer can now be given. Such NPs are normally
thematic for more than one sentence. After introducing an NP repre¬
senting new information into discourse through left dislocation, a
speaker normally maintains it as the theme until he explicitly indicates
btherwise. After a left dislocation, a hearer expects the ensuing dis¬
course to be thematically unified around what has been dislocated. This
expectation on the part of the hearer is frequently facilitated by a
speaker's use of anaphora referring to the theme. When a speaker wishes
to turn to a new theme, he/she uses one or another device for the hear¬
er's benefit to indicate this is happening. A writer uses an indenta¬
tion to signal a change in theme and hence the beginning of a new para¬
graph; a speaker uses an analogous cohesive device which is verbal to
indicate a change in theme and the beginning of a new (oral) paragraph.
We may conclude then that, at least for our corpus, the fundamen¬
tal function of left dislocation is the creation of paragraphs.
Existential Sentences as Thematic Devices
The other thematic device which is widely found in our corpus of
expositions is the existential sentence, a kind of sentence which nor¬
mally begins with "there" and indicates the existence or non-existence

125
of something. The information introduced by existentials is typically
new to the stream of discourse, as we see in the extended example be¬
low:
56) . . .if that canning factory hadn't burned, it'd give a lot of
people jobs in that area too. You know there's quite a few
people in around Bybee. (uh-huh) We had a couple of stores
and a post office and things like that, but it was just, com¬
paring it to here, to White Pine, just about like White Pine.
Cause at one time or other there was a, an old hotel here in
White Pine, and a post office, and uh I, even before that
they might just uh been a country store, not even a drug¬
store, but of course see, they added the drugstore and the
bank and then different and, and different things like that,
and then the railroad. Now there was no railroad that went
through Bybee. It was out in the country more. I mean, of
course, railroad track in the country too, but I mean it was
off the railroad line, and uh in White Pine uh here wasn't.
There used to be a depot up here. You could get off. You
could get on the train and get off the train right there.
It's been torn down not too long ago. (hmm) There's an
express office up there. I don't think it's been torn down
uh more than, uh more than maybe about six or eight years.
Little gray building. You know, you could still have things
come in express but no passenger, you know, there's not any
passenger trains even in Tennessee now. (11-13)
In this exposition we find seven existentials (each underlined) which
introduce new information. Two ofthem(". . .there was no railroad. . ."
and "There's not any passenger trains. . .")donot introduce new (not
previously mentioned) NPs like the other five. But they do add new in¬
formation (that something previously mentioned does not exist) to a mes¬
sage and thus push forward the progress of a discourse. We have more to
say about the functions of existential sentences after examining the
kinds of information they introduce.
Unlike left dislocation, which normally introduces definite NPs,
existential sentences normally introduce information which is

126
indefinite. It may be specific, generic, or nonspecific indefinite,
but in only one instance, given below, out of 320 (in the set of 200
expositions) is it definite:
57) Because there's the Seventh Day Advents [sic], they have
their services on . . . (IV2-16)
Left dislocation and existential sentences thus complement one another.
The former usually does not introduce indefinites; the latter very
rarely does otherwise. Both can introduce specific indefinites. In
the following example, for instance, we see that the informant begins
the sentence with a definite NP. But when she realizes that she will
introduce a specific indefinite one instead, she does so with an exis¬
tential and not through left dislocation:
58) And now then the Smith, there's a handicapped boy that
lives in that building now, Kenneth Smith. (II12-2)
This éxample shows a switch in strategy by the speaker. In the
data we find a number of sentences which begin with an existential
and then appear to be left dislocations, such as
59) And there was an old gentleman named Arwood, he lived down
the road just a little ways down there. . . (X2-8)
which also involve specific indefinites and look like changes in
speaker strategy. They are well-formed constructions, however, and
we will look at them in detail at the end of this section. They
represent a fusion of left dislocation and existential patterns.

127
As a thematic device, the existential sentence does not have the
force of left dislocation, for perhaps three reasons. For one, it is
not a marked construction (in terms of information structure) as left
dislocation is. Not only is it normal, generally speaking, for the
first mention of a noun (except proper nouns) in a discourse to be in¬
definite. Left dislocation overrides this. But left dislocation also
has the effect of yanking an NP from its expected syntactic position
and bracketing it at the front of a sentence. Both left dislocation
and the existential are thematic devices and introduce themes, but the
former is a marked construction, the latter not.
Another reason is that an existentialized NP does not initiate an
anaphora chain nearly so often as a left dislocated one. In the exposi¬
tion above (p. 126), for example, four existentialized NPs ("people,"
"railroad," "depot," and "express office") are followed by at least one
anaphoric element each, but the other three are not. If these seven
NPs are typical, then themes introduced by existentials are not on the
average thematic for as long as those introduced by left dislocation.
A third reason is that existentials have another function which
competes with the function of thematization. This competing function
is that they are used simply to communicate whether something does or
does not exist or how much of something exists. Typically, existen¬
tialized NPs have quantifiers, as do the following:
60) And then there's a lot of flu going around. . . (VIII2-16)
61) They's only one thing I'd like for this town to get. (VIII2-23)
Existential sentences with quantifiers in the data rarely initiate
anaphora chains.

128
Existent!ais have been widely recognized in the literature as
by Bayless and Johnson, as signaling "that new information is being
introduced without directly relating it to some old information"
(n.d.:37). But this function and the function of indicating that
something does or does not exist seem to be the only basic functions
of existentials. They do not possess the range of functions that left
dislocation has. Existentials are not used to indicate a contrast,
nor do they emphatically reintroduce an NP back into a discourse. Oc¬
casionally an existential will specify a subgroup of an NP which has
been previously mentioned (cf. left dislocation, pp 111-2 above), as
in 61 below:
62) The only opposition I would think uh that would come from
within this town to any one thing I guess would be maybe
opposition to alcohol. There'd be some opposition there.
TTTT2-26)
In two instances of the 320 an existential is repeated, as in 63 below,
to reinforce the effect of the first one:
63) . . .there's more difference in the atmosphere, the people
and all from here to there than there was from, when I went
to Philadelphia. Yes, there really was. . . (VI3-5)
It was indicated earlier that eleven percent of our corpus' left
dislocations occur at the beginning of an informant's response. Be¬
cause existentials also introduce pieces of information and partici¬
pants into discourse we should expect to find many of them occurring
exposition-initially also. Of the 200 selected expositions, 21 (10.5
percent of the total) begin with an existential. This is more than
begin with a left dislocation (16 of the 200, or 8 percent).

129
We have seen that left dislocation and existential sentences
have different ranges of functions but that both introduce new, the¬
matic information into discourse as a rule, albeit of different kinds.
A close look at our White Pine reveals a number of sentences which
combine left dislocation and existentialization. In such sentences
the differences between the two disappear, as in sentences 64-6 below:
64) There's not a teacher I don't think up there that I wouldn't
want my child to be in their room. (VI-4)
65) But they's some people I wouldn't want 'em to grow up in
White Pine like it is. (VIIII-23)
66) There are people in this town that I would love to step on
their face. (VIII4-24)
On balance, these sentences seem less like left dislocations than ex-
istentials. In these we do not find the characteristic juncture
which follows a left dislocated NP: a slight pause and level or only
very slightly falling intonation (in contrast to the falling tone char¬
acteristic of the end of a declarative sentence). But in seventeen
instances in our corpus as a whole there are sentences such as 67-9
67) I know they was some people uh, their cars got stalled between
here and there. (XI1-6)
68) And there was an old gentleman named Arwood, he lived down the
road just a little ways down there. . . (X2-5)
69) Well, there was a lady lived in this community, she was an
educated somebody, a school teacher. (1X1-29)
in which an NP is existentialized and then copied by a pronoun and
which have the characteristic juncture between the NP and the rest of
the sentence. On these NPs steady pitch is maintained and does not

130
fall off as at the end of a sentence. The three examples above and
the fourteen others cannot be regarded as anything but fused sen¬
tences.
It is here that our two types of thematic devices, left dislo¬
cation and existentialization, considered separately in this chapter,
fall together and show that they are indeed equivalent structures
when it comes to introducing thematic NPs which are either specific
or generic indefinite. Such NPs may be introduced by either or both
devices. Especially with generic NPs, the two devices most clearly
merge.

CHAPTER IV
DEICTICS AND SUMMARY DEVICES
General Remarks
In this chapter we consider two additional surface structure
cohesive devices, deictics and summary devices, which Longacre and
Levinsohn suggest as crucial to an investigation of the organization
of discourse. They indicate that deictics (words with a specific ref¬
erent which point elsewhere in discourse, especially the demonstra¬
tives "this" and "that") are frequently used in languages to keep
track of a discourse’s participants. They indicate that summary is an
important "type of paraphrase which makes use of very generic predi¬
cates and substitutes" (1978:108). The two devices are considered to¬
gether because, for our data, a speaker usually employs a deictic in
summarizing, as in identifying sentences such as "That's what I
meant."
As a structure which points, a deictic may be either endophoric
or exophoric. If endophoric, it points elsewhere within a text; if exo-
phoric, it points outside a text. An endophoric deictic may be either
anaphoric (pointing backward) or cataphoric (pointing forward). In this
study we will examine only endophoric deictics and specifically will ex¬
amine their use in four different structures: 1) left dislocation with a
deictic as the pronominal reference (as in "Sulfur and molasses, that's
another thing"); 2) extraposition of a subject (as in "That wasn't pos¬
sible to get there"); 3) identifying sentences (as in "That's what
131

132
really happened to us."); and 4) what we call peak sentences (as in
"I guess that's all we can expect out of life is a living."). The
latter two structures are frequently used by a speaker to summarize
a block of discourse.
We are concerned with only the two deictics "this" and "that" in
our data. Generally speaking, "this" may be either anaphoric or cata¬
phoric, although usually the latter. "That" very rarely is cataphoric
We see, in the following examples from the data, that the two deictics
usually point in opposite directions:
1) He really means to do good. This was one thing about the
Kennedies. They had definite goals that they were going to
try to, . . . (113-33)
2) So this is a Knoxville man that come to Morristown, and I, I
went out there to see him. . . . (V1113-18)
3) Well, industry I'd just as soon stay out anyway. As far as,
you know, all industry's doing's polluting the air and
ruining the environment and, you know, that's, like the big
city, that's another reason I wouldn't want to live in a big
city. TV2-13)
In the first example, "this" points forward to a point the speaker
wishes to make about the Kennedies. In the second example, the
speaker uses "this" to point forward to an individual she is giving
information about. Although she had earlier mentioned him, in this
sentence she indicates he is from Knoxville, a new piece of informa¬
tion which "this" points forward to. In the third example, "that"
makes a broad reference backward to the effects industry has on cities
according to the speaker.
These two deictics have much in common with pronouns. Both deic¬
tics and pronouns participate in anaphora chains, and both are referen
tial to other parts of a discourse. But the reference of deictics

133
is generally less clear and much broader than it is for pronouns. None
of the three deictics in the examples above, for instance, clearly
refer to a specific NP or other element. More important than the
referential function of deictics is the function of pointing, of giving
the hearer directions in an ongoing discourse. Deictics serve primarily
to keep a discourse going, by pointing forward or backward, and they
play a large part in the organization of the expositions in our corpus.
The four structures we will examine are the most interesting involving
deictics but represent only a few which are important in organizing
ongoing expositions.
The other functions of deictics derive from the basic function to
organize a discourse. One of the most important of these is to signal
to a hearer a conclusion or summary statement. A speaker will often
use a sentence with "that" to indicate the end of a paragraph or dis¬
course has arrived, as in the following example:
4) My mother was born in Corbin, Kentucky. She was born in Corbin,
Kentucky. Uh, at a very early age, they moved to Kingsport,
and uh I don't know why, never asked. I didn't ask the reason
why, but they moved to Kingsport, and of course she was raised
in Kingsport. And of course that's where my father met her
was in Kingsport. And of course, when they, since they've
been married now for, well, he's lived down here now thirty,
it'll be thirty years in March. Same time when they all moved
down here. We all moved together (un-huh) thirty years ago.
That's a long time, making me feel old. (113-3)
In the two extended examples in chapter two above (pp. 69-70), we
also find deictics having this function. Such deictics mark the ends
of thematically self-contained units.
The third and fourth structures we will examine, identifying sen¬
tences and peak sentences, frequently display this summarizing function.

134
But they may also perform another common function of deictics — to serve
as a bridge from one sentence or portion of discourse to another. A
deictic may be the bridge between information earlier in the discourse
and an elaboration, an identification, a paraphrase, or a repetition
of this information. In the examples below deictics serve as a
bridge to a repetition (in 5) and an elaboration (in 6):
5) Well, as a community, I think it's a, it's a good community
to live in, to raise your children in. Well, I don't know.
The churches, the schools, everything. I think it's just a,
I don’t know, I like it. I think it’s, that's why I think
it's a good community is because I like it, and I don't know.
(VI—7)
6) ... if you're not doing something, you haven't had any
business for a long time, you know, so there's, there's never
really, never really that much time, but uh we enjoy it and
uh it gives us a living, and uh I guess that's all we can
expect out of life is a living. (V14-211
A third basic of deictics in our corpus of expositions is to
maintain a level of tension in the stream of discourse. Most often
this is done by a speaker's use of the deictic "that" when, at least
in written English, "it" would be used. The tendency for speakers to
use "that," a device which is more dynamic than "it" by virtue of its
pointing function, is quite widespread. In many left dislocations in
our corpus (74 of the 606) we find a deictic (usually "that," some¬
times "those") rather than the third person pronoun as the copy of
the dislocated NP. Two examples follow:
7) That old concrete block there, that was my garage. (V12-24)
8) To go to Morristown to buy a pair of socks, that was kind of
aggravating. (X3-12)
In many extrapositions in our corpus "that" appears as the filler
subject rather than "it," as in

135
9)And you know, that makes you feel awful good for a man to be
that nice to you, you know. (X2-16)
10) We had to meet a bus, so that was sort of inconvenient to get
to school, you know, . . ^ OII4-2)
There are 34 instances of extraposition with "that" in the data, about
half as many as with "it." We will examine more closely this use of a
deictic later in the chapter. There are countless additional instances
in the data of the employment of "that,"as in 11-13 below:
11) Take it up with your immediate supervisor or take it up with
your department head. That's the same way with complaints
within the city. (113-20)
12) That's been ten, eleven, eleven to twelve years ago now that
that was done, . . . (I13-9)
13) It took about, that took about three days to get the power
back, . . . (1X1-14)
In 12, "that" is used rather than "it" in a cleft sentence. In 13
we see the speaker changes strategies to use the stronger deictic
and more emphatically makes the point.
The deictic is more forceful than a pronoun, but the use of "that"
imparts greater tension because of the greater stress which normally
accompanies it. In examples 7-13 above, "that" carries more stress
than "it" would in the same context. A different and perhaps more
accurate way to look at the variation between the two is to say that
the speaker decides whether to give tension to a context. If tension
is given, then "that" is used. If it is not, then “it" is used. It
seems probable that "it" under normal circumstances cannot carry stress.
In the next three sections we examine the form and nature of left
dislocations with "that," extrapositions with "that," and identifying

136
sentences. Then the fourth structure mentioned earlier, peak sentences,
is examined in light of Longacre's ideas on the peak of expository
discourses/paragraphs. The form and especially the distribution of
peak sentences in expositions is studied. At the end of the chapter is
a short discussion of minor summary devices.
Left Dislocations with Deictics
As noted above, we find seventy four instances in our data of
deictics copying left dislocated NPs rather than pronouns. This rep¬
resents 12.2 percent of the total left dislocations but, more impor¬
tantly, approximately one-third of all potential cases. That is,
about 200 of the dislocated NPs in the data are inanimate. A close
look at the data reveals left dislocations for which deictics are the
only, or at least the preferred, copy. These seem to indicate that a
speaker does not always have the option of using either the stressed
deictic or the unstressed pronoun.
One such kind is the left dislocation of an NP from an identi¬
fying sentence (7 instances of this), as in 14 and 15:
14) But now in that enclosure, that was where the railroad
station was. (VI3-12)
15) Threshing floors, that's what they had. (II13-27)
When a subordinate clause is dislocated, as in 16 and 17, the deictic
seems to be the preferred copy.
16) To go to Morristown to buy a pair of socks, that was kind of
aggravating. (X3-12)
17) But to turn high, be real warm and then turn cold, that
hurts. (1X1-19)

137
In a number of further instances, the use of "it" to copy the dis¬
located NP seems completely unacceptable, as in 18 and 19.
18) Classical music, we played a little of that when I . . .
(Xl-12)
19) ... a commercial blended whiskey or a Kentucky bourbon,
something like that would on, on the whole would be either
eighty to . . . (13-25)
In most instances of left dislocations with deictics, as in
20 and 21 below, the speaker has obviously chosen the deictic for
its more forceful, emphatic quality.
20) And of course, the housing project, they put that down
there. (V2-11)
21) The old Atchley house, now you saw that down at Leadvale.
(V2-22)
The left dislocations with deictics in our corpus represent the use
of two cohesive devices to introduce themes. The deictic enhances
the effect of the left dislocation.
Extrapositions with Deictics
In the usual transformational treatment of a sentence involving
extraposition, a clausal subject has been moved beyond the predicate
and replaced by "it," as in
22) It is possible that this sentence looks familiar,
which is the transformed version of
23) That this sentence looks familiar is possible.
In our data we find thirty four instances of extrapositions of
clauses in which "that" is used rather than "it" as the filler
subject of the transformed sentence. There are about twice as many
"it"-extrapositions in the data. Examples of extraposition with "that"
are

138
24) And you know, that makes you feel very good to see a young
man turn over a new leaf. (IVI-22)
25) The only way I'd vote for Ray Blanton is whether, if they's
voting for him to leave the country, and that's about the
same way with the other Democrats, . . . (VIII1-14)
26) That's been a question I think in a lot of people's minds
whether it would be feasible to build one here or not.
(1X2-14)
In these instances the deictic is cataphoric or exophoric. It
neither refers or points backward in the discourse, as "that" usually
does. This cataphoric nature of deictics in such sentences is more
clearly seen in five of the extrapositions, as in 27 and 28 below:
27) That would cost too much money, to clean the lines and
things like that. (II14-14)
28) That's kind of a traditional thing for us, just to go up
there and have a big dinner. . . (V2-19)
in which we have a pause (represented by the commas) before the
extraposed clause. Although such sentences are sometimes simply
termed "right dislocations" or "afterthought constructions," we can
understand them by considering their discourse contexts. The deic¬
tics in such sentences are not anaphoric. They point to a situation
outside the stream of discourse. When the speaker realizes he/she
must provide the information to which they broadly refer, he/she
attaches the information at the end of the sentence, after a slight
pause.
Right dislocation is a kind of apposition and represents a
speaker's addition of new information to a discourse context. It is
a widespread strategy found in our corpus, especially with structures
having deictics. In the next section we examine right dislocation with
identifying sentences.

139
An extraposition, then, is a structure with information shifted
to the end of a sentence. When a speaker uses "that" rather than "it"
in an extraposition, he/she explicitly points forward to this infor¬
mation. An especially strong cataphoric effect occurs in the one
extraposition in our data with "this":
29)This don't sound feasible that they would be from all areas
of the country. (114-10)
Identifying Clauses with Deictics
In chapter three the infrequency of it-clefts (three instances)
and WH-clefts (four instances) was pointed out. Such constructions
seem to be more widely employed in written than in spoken English.
In those same 200 selected expositions in the White Pine corpus which
were searched for cleft constructions, however, we find another kind
of construction superficially similar to clefts. There are 27
instances of this construction, which is termed, after Halliday
(1967:231), "anaphoric identifying clauses." Their frequency substan¬
tiates Halliday's statement (p. 232) that they "are probably at
least as frequent, in dialogue" as clefts are. In the following
three excerpts the identifying clauses are underlined:
30) They're putting in I call it a, I don't know whether this
is the proper term to use, it's the term I use, a beer
joint. And that's what everybody else around here says,
right down our road down here. (V4-19)
31) They set it up out in the shopping center there. They just
have that whole thing roped off, and, and that's where they
have their . . . (II13-15)
32) So I mean, what more can you do but fight back? When you've
already turned the other cheek, there's only one thing to do,
and that's fight back, and even the Bible'll tell you to do
that. So that's, that's what we're doing. (VII12-27)

140
In such sentences the deictic is anaphoric (except in a small
handful of cases we discuss later). It points backward in the
discourse and identifies something which has gone before with a
clause of information following it. The something ("beer joint,"
"shopping center," and "fight back") usually, but not always,
appears in the preceding sentence. Halliday (1967) indicates the
communicate effect of anaphoric identifying sentences:
33) ... the combination of deixis with identification,
particularly when the deixis is anaphoric, being highly
effective as a form of communication. The speaker
represents one part of the message as to be identified,
and then identifies it with something that is shown deic-
tically to be recoverable from the preceding discourse.
Furthermore since the demonstratives may have extended
text reference the identifier "that" in most instances is
not restricted to an element functioning in the clause;
any stretch of discourse may participate in the identifying
relation in this way. (p. 232)
If the identification is not with an NP as in examples 30-2 above
but with the entirety of a foregoing block of discourse, the effect
is of a summing up and tying together of the block. Showing this is
the following extended excerpt:
34) Fieldworker: What's a funeral like in a small town where
everybody knows everybody?
Informant: I think it's more important to the people. Yeah,
a funeral to me, it's not for the dead person. It is for the
people who are living, (uh-huh) that have lost somebody, (uh-
huh) In here, you know, if you know the person, you, every¬
body feels a sadness and, you know, you comfort 'em, and
we're we're closer. You know, the biggest thing to me is
like my Aunt, we viere talking a little bit about these funer¬
als, that you send invitations to a funeral, and that (hmm)
to me would just be, you know, how could you send an invita¬
tion to something like that? I mean you send invitations
to parties and things. Uh it* to. me, you know, a funeral
where you come if you want to. You know, you don't come just
because you got that invitation, come by invitation, then

141
it's not as, it's not really you that's coming. It's that
invitation. I think it's uh more important, and it helps
people around here more than it would a, you know, in
other places, because then we're, we're closer and a lot
of people feel the loss. You know, it's not just the fam¬
ily. I've never been to any funerals but (hmm) that's
just what it seems like to me. (XI-18,19)
In this exposition, the deictic "that" in the last sentence points
back to and sums up the entire preceding discourse. It also signals
to a hearer that an end of a whole or part of a discourse has been
reached. In the cases which have left dislocation from an identi¬
fying clause, discussed above on page 134, an especially strong
note of finality is conveyed. Since identifying clauses with
deictics are frequently used to conclude paragraphs, left disloca¬
tions from identifying clauses often represent exceptionally short
and abrupt paragraphs.
The identifying sentences in our corpus differ in two important
respects from what Halliday says about them, however. First, Halliday
claims that only in contrastive cases or when the deictic is exophoric
(referring to something in the speech situation, outside the text) is
the deictic stressed. In our data, the deictic, although usually
anaphoric, does receive focal stress. Focal stress is not as strong
as contrastive but it is quite obvious, especially with the instances
involving left dislocation.
The second exception to Halliday's characterization of identifying
sentences is that the deictic cannot be cataphoric. Although the
identifying sentences we will now examine briefly differ from the
twenty instances in which'"that" identifies something in the preceding
discourse with a WH-clause, they are identifying sentences in function
and according to Halliday. These sentences involve a right dislocation,

142
as those extra-positions discussed above on page 136 do. In these
cases the deictic points forward to the information which has not
been previously mentioned by a speaker, as in
35) The girls are just, you know, they're, they're so strong
and you can see their weak points, but you can see how
they overcome them, and they're outgoing. And 1 guess maybe
that's the wav I like to be, you know, to be able to get
out and to be, so I won't be so nervous and everything.
(XI-6)
36) It's a friendly town. It really is. It's small, and you
just kind of hate to leave it, you know, really, but I was
surprised. I didn't really think that that many people
stay. I like White Pine a lot, but I don't know if I want,
you know, to stay here for rest of my life or whatever, but
a lot of them's gotten married and they're still living
here in White Pine, and so I think that's the biggest reason
White Pine's really grown a lot, because all the, you know,
the kids are growing up and everything and they're staying
here. (VII3-13)
In the first example, "to be able . . . everything" is appositive to
the identifying sentence; in the second, "because . . . here" is.
We perhaps cannot say that the deictics in such contexts are over¬
whelmingly cataphoric, but although they do weakly point backward,
they do not clearly point to any specific portion of the preceding
discourse and, more importantly, the discourse would be incomplete
in these instances without the appositional information. The force of
these deictics is predominantly cataphoric.
On the other hand, we also find that identifying sentences with
WH-clauses may also have appositional information attached afterward,
as in
37) And then I came back to Morristown. My brother-in-law and
sister opened a, an office in Morristown and so I came down
to live with them and go to Morristown Business School, and

143
then that's where I wet m.y husband, in Morristown. (Vl-1)
38) Now that's something they've done so much here, housing
development. That's what's turned White Pine around, that
low-rent housing development area, and, and uh, and uh, the
apartment complexes. (11-34)
In such instances, the appositions are exact or near exact repetitions
of something previously mentioned. "In Morristown" is nothing more
than an afterthought here. "That low-rent . . . complexes" is also
unnecessary for the discourse to be well-formed.
We may tentatively hypothesize than that identification sentences
with WH-clauses are anaphoric and that those without are cataphoric
to right dislocated or appositional information at the end of the
sentence. Why this might be so is not clear, but further evidence
of this distinction will be found in the next section's discussion
of peak sentences, to which we now turn.
Peak Sentences
We come now to a kind of structure found in our corpus which has
not been described in the linguistic literature to date. This struc¬
ture will be called a "peak" sentence and five examples of it are
39-43 below:
39) That's another thing we have through here's the Appalachian
Trail. (V3-4)
40) That's all he's ever got done in Washington is politics.
(VI2-20)
41) But that was the reason I gave was that I would like to go
to vote and give an opinion. (VI12-10)
42) That's the best way I ever found to think is just to walk
back up through the woods or something like that . . .
(VII12-10)

144
43) That's the biggest change that's been made since I remember
living over here is that shopping center. (X2-21)
Such sentences appear to be fusions of two identifying clauses. In
39, for instance, "That's another thing we have through here" appears
to be an anaphoric identifying clause which is fused with “another
thing we have through here's the Appalachian Trail." By itself, the
first part seems to be anaphoric to something previously mentioned.
Combined with the second, it is not clearly anaphoric or cataphoric.
In our corpus there are fifty instances of such sentences in the
expositions of 23 different speakers. Forty-three begin with the
deictic "that," five begin with "this," and two begin with "it" and
will be considered with the forty-eight beginning with deictics. An
example with "this" and one with "it" follow:
44) This is what I do not like about the low-rent housing in
White Pine is that the people in White Pine really have no
say over it whatsoever. (X4-6)
45) It's the biggest thing that ever hit close to the town was
the truck stop, . . . (VIII2-19)
Although such sentences do serve to identify one part of a discourse
with another, like the anaphoric identifying sentences, we refer to
them as peak sentences because they are employed by a speaker to indi¬
cate a peak or culmination in a discourse. We return to this point
below.
In ten of the fifty instances, a WH-clause follows the intro¬
ducing deictic, as in 44 above, so that these peak sentences appear
to be fusions of anaphoric identifying clauses and WH-clefts. In all
the remaining forty instances peak sentences also represent

145
combinations of well-known patterns. Peak sentences are clearly well-
formed fusions and are not faulty constructions.
It is not altogether clear, however, how the internal structure
of such sentences should be analyzed. What is the subject of a
sentence such as 44? Most likely it and other peak sentences are
best viewed as having sentential subjects and being representable
in such a tree-diagram as
46)
UP VP
,s WUeW
But if we look at a peak sentence in isolation from its discourse
context, there is no reason not to propose a structure like 47:
47)
5
V up
. /
IS
\aJ1m4? í'iwe is ^

146
Neither analysis tells us much about peak sentences. Having an
unusual structure and being employed by speakers with relative
infrequency, they would seem to have special properties that we can
determine from considering their discourse context. We will consider
this context by asking two questions: Are the introducing deictics
of peak sentences anaphoric or cataphoric? How are peak sentences
distributed in discourse blocks?
As was indicated at the beginning of this chapter, deictics
frequently do not clearly refer to another portion of discourse. In
many cases they merely point. If we look closely at peak sentences,
it is usually quite difficult to determine in which direction the
initial deictics point. If we examine the entire exposition in
which 42 above appears, is the deictic "that" anaphoric or cata¬
phoric?
48) Fieldworker: In White Pine what do you do for a good time?
Informant: Without leaving White Pine? Watch TV. Maybe,
maybe walk back, you know, like one time, it's, you know,
peaceful back there in the barn, and like uh, like I said,
if you're standing back there by yourself in the barn, like
one time when I went back yonder, and you, I had, you know,
it's, it's not really a seat really. It's part of the manger
that comes out, and it's a hard position to get in cause you
got to sit with your legs way up in the air, but Kathy was
up here watching TV or doing something and uh, talking to
Momma, I guess, and, and I went back there that day. You
know, it's just a good thinking place or it's a good place
to relax. I mean you just set there and the cows, all the,
if you ever set close to a bunch of cows chewing, well, it's
just, you know, relaxes you. And then your dogs are all
laying in there, you know, a couple of cats, and the rabbits
are right outside. They're all peeking in at you, and that,
that's a good way, that's the best way I ever found to think
is just to walk back up through the woods or something like

147
that. But for a good time and not leave White Pine, it's
either ride around or watch TV. Usually watch TV, because
they a lot of good stuff been coming on TV here lately.
(VII12-9)
Or is it anaphoric or cataphoric in the context of the following
exposition?
49) Fieldworker: What would you say have been the most signif¬
icant changes you've seen in the town in your few years here?
Informant: Most signficiant change I've seen is the, the
intrusion or the coming in or, of new people. White Pine is
no longer the same town it was eight and a half years ago
because of its growth, uh because of people with different
backgrounds having moved into our town, (uh-huh) Uh there¬
fore it has changed the complex of our town tremendously.
Uh I think this is the greatest change that has happened in
our town is to have new people coming in with new blood, new
thoughts, new ideas, new approach to things" (X4-4)
On the one hand, the deictics point backward in the discourse; on the
other, they point forward, to information coming at the end of the
sentence which sums up the entire preceding discourse. In a sense,
we may call such deictics "ambiphoric," pointing in both directions.
Peak sentences resemble identifying sentences with right dis¬
location (as in 37 and 38 above). But the deictics in peak sentences
are more clearly bridges between foregoing and following portions of
discourse. Normally the initial deictic in peak sentences (although
a speaker retains the option for such stressing in cases of emphasis)
does not receive the focal stress we find with deictics in anaphoric
identifying sentences, indicating that peak sentences are on the whole
more forward looking. Thus they more strongly tell a hearer that
important information is coming up.
An examination of all fifty peak sentences shows that all of them
point both backwards and forwards. In some instances, as in 50 below,

148
the anaphoricity predominates; in others, as in 51, the cataphoricity
clearly is more prominent:
50) ... and I think that uh the main thing I think in the candy-
pulling that we, course we like the candy and that was a
good way to have it, but we liked the fellowship and together
of all the young people. That was, that was the main thing,
I mean, that we loved to get together. And I said like this,
so many, well in fact, that's where I met my husband was at
a candy-pulling. (II12-12}
51) I cannot pass, well, bratwurst is my favorite, those large
white sausage, you know, on hot, hard sour dough bread, uh
with hyacinth, very hot mustard. I just, I can't pass one
up. I went with my father-in-law uh on one trip while we
was there, and we stopped for lunch, why, that's all I
could think about was getting me some potato salad and
getting me a couple of bratwurst on bread. (ÍI3-18)
In the discussion of identification sentences with right dislocation
above (pp. 140-1), it was indicated that those with WH-clauses tended
to have initial deictics that are anaphoric and that those without
such clauses tended to have cataphoric ones. Notice that the ana¬
phoric deictic in 51 above is followed by a WH-clause and that the
cataphoric one in 52 is not. Of the ten instances of peak sentences
with WH-clauses in the data, only three are not clearly anaphoric.
The peak sentences without such clauses have deictics which are more
clearly ambiphoric. The seven strongly anaphoric sentences, like
50 above and 52 below, repeat information verbatim, as does 37 above.
52) Well, as a community, I think it's a, it's a good community
to live in, to raise your children in. Well, I don't know.
The churches, the schools, everything. I think it's just
a, I don't know, I like it. I think it's, that's why I think
it's a good communitv is because I like it. and I don't know.
(VI—7)
The forty remaining instances have deictics which are cataphoric to
varying degrees; none of them involve a repetition as in 52. The

149
initial deictic seems especially cataphoric when it is "this," as in:
53) ... We have a volunteer fire department, which I might
say is outstanding. It's terribly proficient. It's very,
very, very good, (hmm) And we've improved on that as times
come along, (uh-huh) Uh, we have two trucks. We have a
large engine. Have a small brush-fire, brush-fire fighter,
uh to handle any small jobs, and we now, as of this year,
have good communication. Uh, they have portable radio units,
uh volunteer fire department, (hmm) and all times. They're
on the job from somewhere. We, we have quite a few within
the town itself that, that have their own business, that
are capable of being able to go on, to go out at any time,
(hmm) Uh, this doesn't seem to matter as far as uh insur¬
ance is concerned. This is something that's bothered me
for a long time is that we're in uh grade ten in insurance,
as far as homeowners is concerned. (113-12)
Although peak sentences and identifying sentences have been
treated separately in this chapter, they are quite similar in many
respects. When we look at their distribution in discourse, we will
see they are very close. At the end of chapter three we examined
structures which were the fusions of two structures —left disloca¬
tion of an NP and existentialization of an NP — that had been viewed
as distinct until that point. In the same way as thematic devices are
not discreet from one another, we find a continuum of structures
involving deictics. Peak sentences are identical to anaphoric identi¬
fying sentences except that the latter have a pause where the former
have a copulative verb ("is") and that the latter have a focal stress
on the initial deictic whereas the former may or may not have. In
addition, anaphoric identifying clauses with right dislocation such
as in 35-6 above are very similar to that-extrapositions, as in 27-8,
the latter are never anaphoric and never identify one part of a
discourse with another.

150
Viewed in isolation, instances of the three kinds of deictics
(those in peak sentences, in anaphoric identifying clauses, and in
extrapositions) may not then be readily distinguishable. It is only
when we examine their anaphoricity or cataphoricity in their dis¬
course context that they can be distinguished.
The peak sentences (examples 48-53 above) we have looked at in
their discourse contexts come either at the end or towards the end
of a speaker's exposition of a point. Moreover, they clearly have
the effect of summing up and culminating a foregoing portion of
discourse sometimes of considerable length, as in 48, where a teenage
speaker explains the best way he knows to get away and clear his
thoughts, or in 49, where a Baptist minister explains the one most
important change he has observed in the town of White Pine. In only
one case of fifty does a peak sentence come at the beginning of an
informant's response:
54) Fieldworker: The more I listen to disco music, the more
it seems to me it all sounds the same.
Informant: It does. It's got the same beat, same loudness.
That's what a lot of people do too is go to places like
Flanagan's and Uncle Sam's and dance.
In this example, "that" is not anaphoric and the peak sentence does
not sum up any portion of foregoing conversation. It may be that it
sums up information implicit in the conversation or in the mind of the
informant, but it clearly sums up what the informant has to say on the
subject of entertainment. It does not really constitute a counterexam¬
ple to our generalization that peak sentences follow and sum up blocks
of discourse.

151
That we should find structures in discourse like peak sentences
which mark the peak of a speaker's relation of information to a
hearer is pointed out by Longacre and Levinsohn (1978:108). More
than a decade ago, in his Philippine study, Longacre indicated that
paragraphs may optionally have units called "peaks" around which
an entire paragraph is focused. His views on the nature of peak
have evolved somewhat since 1968, but his latest comments on peak
(Longacre and Levinsohn 1978:105) apply perfectly to the function
in discourse of what we have been calling peak sentences.
In his 1968 work Longacre did not distinguish peak from text
(the one obligatory unit in the grammar of any explanatory para¬
graph). Considering peak as only one aspect of text, he states
(p. 109) that text "is the PEAK grammatically and lexically of its
paragraph." In Longacre 1976, he comes to view peak as an optional
unit of exposition but believes that much good exposition, in
exhibiting a tension or progress toward a clear exposing of a
subject, is characterized by having a peak. He views (p. 228) peak
in exposition as analogous to climax and denouement in narrative and
goes on to explain
It seems evident to me that an expository discourse of the
better sort reflects a certain struggle, a struggle to achieve
clarity in the face of recalaitrant elements in the subject
itself and possibly in the lack of background on the part of
those who are to hear the discourse. Exposition of the better
sort is able to clarify the main outlines of its subject in
spite of these difficulties within the subject matter and
within the receptors. ... It would seem therefore that an
artful expository or hortatory discourse will have a meaningful
cumulative thrust. This should correlate in at least some
discourses with a marked surface structure Peak. I believe
that of the various devices open for marking of surface struc¬
ture Peak in expository and hortatory discourse, rhetorical

152
Underlining is probably the most frequently used. (1976:228-9)
(According to Longacre (1976:217-8), rhetorical underlining comprises
"parallelism, paraphrase, and tautologies of various sorts" used by
an author to ensure that a reader or hearer does not miss a point.)
Peak then comes near the end of and represents a kind of climax to an
expository discourse or paragraph. In Longacre and Levinsohn (1978:
105) we find Longacre's most recent views on peak. Again he indicates
that only discourses which have tension have peaks:
If a discourse is plus tension, there will most likely be some
kind of climax of development, some marked surface structure
"peak". . . . For discourses that are not narrative but still
have a climax of development, the peak may mark "target pro¬
cedure" in a procedural discourse, "climactic exhortation"
in a behavioral discourse of the hortatory variety, and a most
satisfactory or "culminating explanation" in expository
discourse.
An examination of any of our discourses in which a speaker uses a
peak sentence reveals that precisely what a peak sentence does is
present the most satisfactory or culminating explanation regarding
a topic.
For a speaker to accomplish the task of adequately explaining a
subject, it is critical that his/her climactic or culminating point
get across to a hearer. Information therefore which represents the
peak or consummation of an expository discourse, sometimes of several
hundred words in length, must stand out to a hearer. A speaker must
make sure that the significance of such information is conveyed, even
if it must be explicitly pointed out. It is the pointing out of it
which the particular kind of peak sentence studied in this chapter
does, through the use of deictics, a cohesive device. A speaker need

153
not always employ a cohesive device, which makes a speaker's message
"stick together," to signal a peak, but he/she sometimes does.

CHAPTER V
CONJUNCTIONS AND INTRODUCERS
In this chapter conjunctions and introducers are treated together.
In our corpus, conjunctions very often introduce sentences, paragraphs,
and even entire expositions. This is consistent with what Longacre and
Levinsohn point out: "As introducers of sentences, conjunctions serve
to fit the sentence into paragraph context. They also serve to fit
the paragraph into discourse context" (1978:109). Some conjunctions
(especially "and," "but," "then," and "now") so frequently
begin sentences in our oral expositions that clearly one of the
primary functions of such words is to mark sentences. Frequently
conjunctions thus accompany the pauses which divide sentences. In the
following example exposition the sentence-initial conjunctions are
underlined:
1) Fieldworker: What do you think most people who live here want
the town to become?
Informant: Well, I think everybody wants us to have the new
hospital, which we're about on the verge of getting, I think,
which we really need is the new hospital. And everybody
wants it. Just, now White Pine's built up a lot over the
years. Of course, like I say, it's still a small town,
(uh-huh) But just other than that, well, they's not really
no great changes been, to me they's not, I mean, uh to other
people it might seem different. But to me it just seems
like the same, the way it's always been. (X11-14)
In addition to five sentence-initiating conjunctions here, we have two
instances of "well," and one of "of course," other introducers,
154

155
beginning sentences. Conjunctions and other introducers are the surface
structure cohesive device which constitutes what a speaker uses to
stitch the parts of a message together into a discourse.
Although they do not by themselves introduce themes, conjunc¬
tions and introducers frequently begin paragraphs and usually
accompany left dislocated NPs which introduce paragraph themes. It
is shown in chapter three above that a speaker often uses a conjunc¬
tion or introducer to signal a hearer that one theme is being super¬
seded by another. In this chapter we examine the conjunctions and
introducers which accompany the 190 dislocated NPs in the 200
selected expositions used in chapter three. In particular we look
at the introducer "like" and suggest that it is a sub-paragraphing
device. For the most part our analysis is suggestive rather than
definitive. Because of this study's limitations, the functions
and distributions of conjunctions are not yet entirely sorted out.
It has been widely pointed out in the literature that para¬
graphs and discourses characteristically begin with certain intro¬
ducers in English and in other languages as well. Regarding oral
paragraphs Van Dijk says "A new paragraph thus indicates (sub-)topic
change. In spoken language we have pauses, intonation and specific
particles like 'now1, 'well', etc. to indicate such paragraphs.
Other languages have specific morphemes in order to mark beginnings
and endings of stretches of discourse. . ." (1977:152-3). Longacre
states that "many languages have particles that indicate either the
beginning or the end of a paragraph" (Forthcoming:!17). He goes on
to cite examples from Huichol, Shipibo, and Capanahua.

156
In our data the presence of an introducer of some kind at
the beginning of expositions and paragraphs is typical. What
typically introduces our expositions is "well." We cannot claim
that it more commonly begins expositions than other kinds of dis¬
course or that its primary function may not be a general one to be¬
gin any conversation turn. But its great frequency (114, or 57 per¬
cent of the expositions examined begin with it) indicates that it is
a speaker's usual way of getting a monologue exposition underway.
In our data only two other introducers are used with any frequency at
all to begin discourses: "of course" (10 instances) and "now" (7 in¬
stances).
In chapter three it is shown that left dislocated NPs typically
begin paragraphs. If we look at the 190 left dislocations in the
200 selected expositions, we find that conjunctions typically accom¬
pany such NPs. In other words, conjunctions are frequent cues to
new paragraphs. Only 39 (20.5 percent of the total) begin with no
conjunction or other introducer. In such instances only the pause
accompanies the NP. Of the remaining 151, 60 left dislocated NPs
are preceded by either "and" or a combination of "and" and other
conjunctions or particles. Twenty-five are preceded by "but" or a
combination with "but." Seventeen NPs have "well" by itself or in
combination. Ten are preceded by "of course" by itself or in a
combination.
Twenty-nine of the dislocated NPs are preceded by a hesitation
"uh," either alone or with a pronoun. It is indicated in chapter
three above that left dislocation rarely involves a hesitation

157
between the dislocated NP and the ensuing sentence. But an audible
hesitation often precedes the NP, however, and apparently aids a
speaker in introducing a theme into discourse. We include the
hesitation "uh" in the table on the following page, which lists the
variety of introducers for the 190 left dislocations.
One introducer in the list here, "like," deserves our special
attention. Below are three examples of it accompanying left dis¬
located NPs:
2) Like Brooklyn, you know their lingo. (V3-16)
3) And like Lisa Williams who lives down the street here, she
got me interested in it, . . . (V113-11)
4) Like my cousins, they had to go to South Carolina because
of the job situation. (XI-8)
Seventeen of the 190 NPs are preceded by "like" or "like" in combin¬
ation with another conjunction. Of the total of 606 left disloca¬
tions in the entire corpus, 29 are accompanied by "like." "Like"
is apparently restricted primarily to younger speakers, and it dif¬
fers in function from other introducers. Rather than accompanying
left dislocated NPs which initiate paragraphs, it creates units
which we will call "sub-paragraphs."
Although only eight of this study's informants were twenty-five
years old or younger, these eight used twenty-one of the twenty-nine
instances of "like." That is, twenty percent of the informants
account for 72.4 percent of the occurences. "Like" not only accom¬
panies left dislocation, however. In general, it is a conjunction
which introduces an example to elaborate a point a speaker has made.

158
Introduced s)
Frequency of Occurrence
0
39
and
31
and + ih
6
and + then
6
and + then + uh
5
and + of course
3
and + even
1
and + even + uh
1
and + why
1
and + wel1
1
and + uh + so
1
but
16
but + uh
4
but + now
2
but + as far as
2
well
13
well + uh
1
well + of course
1
well + now + then
1
like
12
and + like
3
but + like
1
now + like
1
and + then + of course
1
of course
4
why
3
now
3
as far as
3
because
2
uh + because
1
of course + uh
1
SO
1
so + uh
1
uh
7
oh
1
naturally
1
where
1
that + uh
1
you know
3
okay
1
yeah
1
most of the time, though
1
190
Figure 5.1 Introducers Accompanying Left Dislocated NPs

159
Although older informants employ it to make such elaborations, we
find it much more common among the eight younger speakers, in whose
expositions it appears on 164 occasions. Host often used by one
seventeen-year-old high school student (47 instances), it was still
frequently used by all the remaining seven younger speakers.
The prevalence of "like" with left dislocated NPs might be inci¬
dental if it were not that it seems to have a slightly different func¬
tion when accompanying the thematic NPs of left dislocation. Often,
the NPs it accompanies are not thematic for more than one sentence.
Thirteen of the 29 (44.8 percent) are like this. More important, the
NPs seem to establish only limited themes which are not fully devel¬
oped. For instance, in the following examples, "our parents" and
"Brooklyn" are introduced only to make a single comment about them
before the speaker moves on. Like many other NPs introduced with
"like," they are not established as themes.
5) . . .Everyone thinks that they have a right to live their own
lives the way they want to, and they do. And I think we
see that more, you know, (hmm) Like our parents, they
think that we started the drugs and all that stuff, but it
was here before we got here, you know. That's how I
feel about it. (VII3-23)
6) Now in the World War Two I was, while I was in service, of
course you were with people from all over the United States,
(uh-huh) and I could tell by the accent within two or three
states or a thousand miles usually where someone was from,
(uh-huh) Like Brooklyn, you know their lingo, (uh-huh) and
you could tell if it's Alabama or Georgia. You could tell
by their talk, . . . (V3-16)
Even in expositions in which the like-accompanying NP is thematic for
longer than one sentence, the theme has this limited quality. In

160
other words, "like" seems to signal a size of unit which may be larger
than a sentence but is smaller than a paragraph.
In lieu of a readier term, we adopt "sub-paragraph" to refer to
such a unit intermediate between a sentence and a paragraph. The sub-
paragraph is a supersentential unit in which a theme is not fully de¬
veloped. Though a sub-paragraph has a theme, it does not have the the¬
matic nature of a paragraph. -There is no reason to believe that "like"
is the only sub-paragraph introducer, of course. Although they are not
examined here, the phrases "of course" and "you know" seem to operate
in the same manner as "like."
In this chapter we have briefly indicated that conjunctions and in¬
troducers play important roles in discourse. They piece paragraphs and
discourses together. They initiate paragraphs and discourses, espe¬
cially when assisting left dislocation in signaling to a hearer that a
new theme is being introduced. Conjunctions then are not only inti¬
mately involved in the linear structure of the stream of discourse, but
also in indicating new prominences and changes of direction and in ac¬
companying the beginning of new paragraphs. They are clearly an impor¬
tant device in both the hierarchical and the linear structure of dis¬
course.

CHAPTER VI
SUMMARY AND CONCLUSIONS
This dissertation was undertaken because of the long neglect
of the discourse level of analysis pertaining to American English.
Appalachian English was the variety of American English chosen to
study on this level because of its accessibility to the researcher.
The researcher undertook his own fieldwork and collected sufficient
data for this study to come to several important conclusions.
This study is about one aspect of how people in one small town
talk: how they organize what they have to say when they explain them¬
selves to a non-resident of that town. It shows that oral expositions
tend to be organized into blocks of thematically-unified information
called paragraphs and that paragraphs characteristically have one
theme each.
The organization of information into expository paragraphs is
shown in this dissertation to be reflected to an important degree by
certain devices called surface structure cohesive devices. It is
shown that speakers often rely on these devices to indicate to hearers
how a spontaneous monologue exposition is organized.
Extended speech thus normally has structure similar in kind to
extended segments of writing, although the devices used to reflect
this structure are different at least for the oral expositions exam¬
ined in this study.
161

162
Although this study is one of Appalachian English, its findings
are not claimed to be exclusive to it. Further investigations into
American English with the approach taken here should be carried out.

APPENDIX A
LIST OF INFORMANTS
t!l(l
Informant No.
Sex and Age of Informant
Interview Code
1
Female, age 63
11, 12
2
Male, age 29
13, 14
3
Male, age 25
112
4
Male, age 30
mi
5
Female, age 70
1112, 1113
6
Female, age 72
IV3
7
Female, age 62
IV1
8
Male, age 30
113, 114
9
Female, age 54
1114
10
Male, age 60
IV2
11
Male, age 74
IV4
12
Female, age 59
VII
13
Female, age 82
VI3
14
Female, age 39
VI
15
Male, age 53
V3
16
Female, age 44
V 4
17
Male, age 43
VI2(l-3), VI4
18
Male, age 73
VI2(4-30)
19
Female, age 17
VIII
20
Female, age 16
VI13
21
Female, age 43
VI12
22
Male, age 74
VI14
23
Male, age 17
VIII1
24
Female, age 85
VI113
25
Male, age 17
VI112
26
Female, age 39
VI114
27
Male, age 26
V2
28
Female, age 17
XI
29
Female, age 50
X3
30
Male, age 48
X4
31
Male, age 76
X2
32
Female, age 87
1X1
33
Male, age 34
1X3
34
Male, age 60
1X2
35
Female, age 76
1X4
36
Male, age 18
XI2(1-12)
37
Female, age 30
XIII
38
Female, age 28
XI2(13-8)
39
Female, age 53
XII
40
Female, age 33
XII
Informants are listed in the order interviewed. Code numbers refer to
pages of typed transcripts of interviews.
163

APPENDIX B
SAMPLE INTERVIEW
In order to exemplify the interview techniques employed in this
project, a long portion of a sample interview is here presented. It
is the interview with the thirty-four-year-old police chief of White
Pine, a native of Greene County, Tennessee, and a resident in White
Pine for three years. The code letters "F" and "I" here refer to the
fieldworker and the informant turns in the interview.
F: Well, just in your three years here, have there been major changes
in the way you've had to go about law enforcement in this town?
I: Well, I can't really answer that other than this way. The people
that I talked to, and this includes the city fathers, so to speak,
and the businesspeople, said there's a one hundred difference. Uh
for instance uh they used to congregate in the shopping center and
they'd pass their marijuana around or their pills or whatever,
just from one car to another, and uh they'd rip up and down the
highways. Well, since I've been here, we've got a speed gun. We
cut that out. (uh-huh) We cut the drug problem out in the shop¬
ping center. We don't allow no loafing at all uh once the busi¬
nesses are closed. (_hmm) People park their cars and leave them.
Fine. But they don't set in 'em. Uh that was a change we had to
make, so uh, uh to answer your question, as far as the citizens
of White Pine is concerned, they's been a tremendous change. We
just had to set our foot down, get a little tighter. The police
officer that was here, and he was just a one-man department, he
164

165
was born raised here. He knew everybody, and I guess it was just
hard for him to, you know, get real, real serious and crack down,
so to speak, (uh-huh) so change has been tremendous, from what I
can hear. Like I said, I'm not from here and the only way I can
answer your question honestly is to have lived here before I came,
or before we took over, so to speak. The new administration had
a lot to do with it. They gave us a little more power and a lit¬
tle more authority, and uh excuse me [The phone rings here for I]
and uh that helps. You know, you [again the phone rings for I)
F: I wanted to ask you about Cocke County. Why does it have the re¬
putation it does?
I: My opinion is they don't want to have any other type of reputation
as far as law enforcement is concerned. Uh, it's uh, it's just
wide open. Well, it's not as bad as it was. They've got a new at
1 torney general has done some things that uh I approve of. He's uh
cut out a lot of prostitution, (uh-huh) Not all of it, mind you,
but a lot of it. But I think that uh, course, you know as well as
I do, you have what the people want. Just like White Pine, if
White Pine wanted twenty police officers, they would have them,
(uh-huh) Uh if Cocke County wanted most, there's no way they
could cut it all out, but if Cocke County wanted most of their
vice and uh so- crime cut out, they could. It's a border, state
border county, (uh-huh) state line county, and that, that has a
lot to do with it. Plus the fact that it's a very mountainous
area, and uh a lot of the out of the way places, so to speak, and
they just have a lot of crime and a lot of, for the size of the

166
county, I guess they have more crime than any other county in the
United States. Now that's saying a mouthful, (yeah) but you can
uh, you can figure one or two homicides a week. Now, you consi¬
der the size of Cocke County, that would be one or two an hour in
the city of New York, and, you know, that's I guess about what
they have, so uh it's a very high rate of crime, but I believe
it's what they people wants. The majority of the people, now. Of
course, all of them don't, (uh-huh) and it's, it’s got a lot to do
with the law enforcement. Uh you don't have the men anywhere, I
don't care whether it's Cocke County, Jefferson County, or what
county it is. You don't ever have the men you need, and uh you
can't watch everybody. You've got eight officers to uh eighteen to
twenty thousand people, then you see what you're up against. I've
got three officers, and uh I've got twenty, well, two thousand peo¬
ple in the city alone, (uh-huh) so you can see what they're doing.
They're watching eight hundred to a thousand people, each officer
is, (uh-huh) and uh there's no way that you can uh, you can do that.
There's just uh no way at all, so uh that's, I think the biggest
reason they have the crime rate that they do.
F: But as far as I can remember back in Knoxville I've been reading
about strange goings on in Cocke County. You know, senseless mur¬
ders or
I: Well, they, the people feud among themselves. It goes back a long
ways I suppose. Uh they uh they've just got what you, I'd say
roots that uh entice 'em to do things like this. They, they're
just uh, there's a lot of mean people over there, uh plus a lot

167
of, there's good people over there, (uh-huh) Uh it's uh quite a
high crime rate for the size of the county. I mean it's tremen¬
dous. You just wouldn't believe. I have known, I have known, uh
known 'em to have uh as many as two or three homicides a day over
there, (hmm) and that's a lot. Uh eight or ten a week, and uh you
take a county that small, why, it's tremendous, (hmm) It's just
outrageous, just to put it bluntly, so it's uh really bad. But I,
like I said a while ago, I don't think it's as bad as it was, cause
we got a new attorney general, and he, (hmm) you know, he pretty
well put a little, quietened it down a little bit. [#114-31 is an
interruption of a prospective policeman coming in to fill out an ap¬
plication]
F: What do you personally do on Pine Day every year?
I: I don't uh, I don't really can't answer that unless, well, we have
a parade and I coordinate it. (uh-huh) Uh it's about the only,
really the only uh parcipi-, uh part that I take in it (uh-huh) is
uh the coordinating the parade and leading and handling the parade,
(uh-huh) Course, like I said, I've only been here a couple, well,
I think I've had, been here two or three Pine Days and really don't
know the meaning of it, to be quite honest with you. But it's a,
it's a relatively quiet, nice little town, uh comparing it with
other towns its size in East Tennessee, (uh-huh) I like it.
F: In the three years you've been here, what would you say have been
the most significant changes in the town?
I: Well, the, I think the most significant change is the, the things
that the city fathers have done, such as uh recreation, uh streets,

168
the sewer, the water, and the recreation. It was pretty bad,
and they worked very hard in uh doing the things of this type.
We've got a couple of hundred thousand dollars for recreation,
which we've got a new swimming pool we hoping to open this spring
(uh-huh) or early summer, and uh mostly things like that. Recre¬
ation, streets, water, sewer, and so forth and so on. (uh-huh)
That's been the major changes since I've been here.
F: Has the truck stop going up out there had any effect on town? If
so, what kind of effect?
I: I think it's had very little effect. Now, Roadway uh Trucking Com¬
pany has put a uh terminal, not a terminal, but a uh I guess you
could say an intermediate or a mid way changing point out there on
the Interstate. Now, it's brought quite a few families into uh, I
would say in the last six months to a year, it's brought ten,
twenty new families into the city of White Pine alone, (uh-huh) and
uh I think Roadway Trucking has had a small impact on the city, as
far as its growth is concerned. I don't, I can't see that the
truck stop has had any. I'm sure it's had some, but I would think
a amount small enough that you really wouldn't recognize it.
F: Hmm. What about the new housing project that's gone up?
I: Well, now it's was built when I came here. Just, they had just com¬
pleted its (yeah) and uh it gives us a lot of problems. You're
talking about the low-rent?
F: Uh-huh.
I: It gives us a lot of our domestic problems. Uh of course, you can
understand why. You have the type people that live there that do

169
that, (uh-huh) We have to answer a lot of calls down there and a
lot of neglected children and so forth and so on. (uh-huh) That's
the biggest problem, you know, with it.
F: What kind of town do you think most people around here want White
Pine to become?
I: I think a majority of the people want it to stay like it is. Uh
[Visiting ex-policeman filling out application: Nice, quiet, well-
run town.] Yeah, they, they don't like to see, especially your old
people. They don't like to see changes that would be major changes,
like uh I don't think that they'd want to see a housing project of
three or four times the size of that one go in, and I'm convinced
that they don't want to see too much industry, (uh-huh) I think the
statement was made, I won't say who made it, but a statement was
made uh some time back, 'sa year or so ago. We had a company, a
drug company, Davis Manufacturing, I think, that wanted to build a
sizable plant here in White Pine. I think one of the city officials
said "well, White Pine's a nice, quiet little town. Let's just
leave it like that." and the company didn't come in. So I think
that's the way most people feel about it. (hmm) Course you, I
guess you'd have to look at the people that make the decisions on
the city, such as your aldermen and the mayor, (uh-huh) Uh they're
the people that says whether or not anything's going to be here or
not, whether the citizens like it or not, and I think most of them,
well, not, maybe not most of 'em, but the majority of 'em feel, uh
minority of 'em feel thataway, one or two of 'em anyway, (uh-huh)
They's just as soon it stay like it is or go back to like it was
years ago. I honestly feel that way.

170
F: I've talked to a couple of younger people, say, in their mid twen¬
ties who said that, . . .every time he comes here he sees a new
house going up and he feels a little pang of something.
I: Well, it, it is. We're growing. Uh ten years from now, I, I
wouldn't really want to uh say how big we'll be. We may be ten,
fifteen thousand. Just depends of a lot of things, (hmm) but uh
it, it's true. There's a lot of new, new uh houses being built.
In fact, you can't drive in the city any time that you don't see
either one uh just being started or being completed or something.
F: Is this the kind of place you think you would like to settle down
in for a long while?
I: Yes, I have no, I have no uh desire to go anywhere, other than vaca¬
tions. Uh course, I'm from a small community, (uh-huh) just east of
here a few miles and it's what I'm used to. I have no objections
to the way uh, the way the people are here, with the exception of a
few. Course, that's always the problem, (uh-huh) Uh but I'd like
to see it grow, not tremendously, but a little at a time, (uh-huh)
and uh have more things to do for the, especially for the young. I
think they, they should, they deserve. Course I, I know, you and
I were both young once and we'd uh like to had a lot of things that
we didn't have, and it's not changed. Kids today feel the same
way we did. (uh-huh) There's a lot of things they'd like, and I'd
like to see it. Course, it takes money. You talking taxes. Peo¬
ple just scream when you holler raising taxes. White Pine has the
lowest tax rate in any city, I would dare say, in the United States.

171
A dollar and a quarter. Now you tell me any city in the state of
Tennessee that’s, or anywhere for that fact, that has a tax rate
that low. That's ridiculous. . . .

BIBLIOGRAPHY
Bailey, Charles-James N. 1968. Is There a Midland Dialect of
American English? ERIC 021 240.
Bayless, Richard L. and Linda K. Johnson, n.d. Function of Surface
Rearrangement Rules in English Discourse: Old and Mew Informa¬
tion. Manuscript.
Becker, Alton L. 1965. A Tagmemic Approach to Paragraph Analysis.
College Composition and Communication. 16. 237-42.
Becker, Alton L. 1966. Symposium on the Paragraph. College Compo¬
sition and Communication. 17. 67-72.
Berrey, Lester V. 1940. Southern Mountain Dialect. American
Speech. 15. 45-54.
Bicentennial Committee of Jefferson County, Tennessee. 1976. Heri¬
tage Jefferson County.
Bloomfield, Leonard. 1933. Language. New York: Holt.
Bond, George Foot. 1939. A Study of an Appalachian Dialect. Uni¬
versity of Florida thesis.
Bray, Rose Altizer. 1950. Disappearing Dialect. Antioch Review.
10. 279-88.
Brewer, Alberta and Carson. 1975. Valley So Wild: A Folk History.
Knoxville: East Tennessee Historical Society.
Brewer, Jeutonne and Paul D. Brandes. 1977. Dialect Clash in Amer¬
ica. Metuchen, Mew Jersey: Scarecrow.
Bridgeman, LoraineHall. 1966. Oral Paragraphs in Kaiwa (Guarani).
Indiana University dissertation.
Brinegar, Bonnie Carter. 1977. Analysis of Six Sequence Signals of
English Discourse. University of Southern Mississippi disserta¬
tion.
Brown, Calvin. 1889. Dialectal Survivals in Tennessee. Modern Lan¬
guage Notes. 4. 205-9.
Campbell, John C. 1969[1921], The Southern Highlander and His Home¬
land. Lexington: University Press of Kentucky.
172

173
Carpenter, Charles. 1933. Variation in the Southern Mountain Dia¬
lect. American Speech. 8. 22-5.
Caudill, Harry M. 1963. Night Comes to the Cumberlands: A Biography
of a Depressed Area. Boston: Little Brown.
Chafe, Wallace L. 1974. Language and Consciousness. Language. 50.
111-35.
Chafe, Wallace L. 1976. Givenness, Contrastiveness, Definiteness,
Subjects, and Point of View. In Subject and Topic, edited by
Charles N. Li. New York: Academic. 25-56.
Chapman, Maristan. 1929. American Speech as Practiced in the South¬
ern Mountains. Century. 117. 617-23.
Clark, Herbert H. and Susan E. Haviland. 1977. Comprehension and
the Given-New Contract. In Discourse Production and Comprehen¬
sion, edited by Roy A. Freedle. Norwood, New Jersey: Ablex.
1-4C|.
Coleman, Wilma. 1936. Mountain Dialect in North Georgia. University
of Georgia thesis.
Combs, Josiah H. 1916. Old, Early, and Elizabethan English in the
Southern Mountains. Dialect Notes. 4. 283-97.
Combs, Josiah H. 1931. The Language of the Southern Mountains.
PMLA. 46. 1302-22.
Coulthard, Malcolm. 1976. Discourse Analysis in English—A Short
Review of the Literature. Language Teaching and Linguistics
Abstracts. 8. 71-89.
Daneá, Frantisek. 1966. A Three-Level Approach to Syntax. In Tra-
vaux Linguistiques de Prague: Volume I, edited by Joseph Vachek.
Tuscaloosa: University of Alabama. 225-40.
DaneS, Frantisek, ed. 1974. Papers on Functional Sentence Perspec¬
tive. The Hague: Mouton.
Dressier, Wolfgang U. 1970. Towards a Semantic Deep Structure of
Discourse Grammar. CLS 6. 202-8.
Dressier, Wolfgang U., ed. 1978. Current Trends in Text Linguistics.
New York: Walter de Gruyter.
Dykeman, Wilma. 1965. The French Broad. Knoxville: University of
Tennessee.
Dykeman, Wilma. 1977. Tennessee. New York: Norton.

174
East Tennessee Development District. 1972. County Census Summary:
Jefferson County, Part I. Report No. ET-SS-9g-71. Knoxville.
East Tennessee Development District. 1973. County Census Summary:
Jefferson County, Part II. Report No. ET-SS-13b-73. Knox¬
ville.
East Tennessee Development District. 1977. Food Stamp Survey,
1976: A Summary Report. Report No. 163. Knoxville.
Firbas, Jan. 1966. On Defining the Theme in Functional Sentence
Analysis. In Travaux Linguistiques de Prague: Volume I, edited
by Joseph Vachek. Tuscaloosa: University of Alabama. 267-80.
Firth, J. R. 1957a[1935]. The Technique of Semantics. Reprinted
in Papers in Linguistics 1934-51, by J. R. Firth. London: Ox¬
ford University Press. 7-33.
Firth, J. R. 1957b. Ethnographic Analysis and Language with Refer¬
ence to Malinowski's Views. In Man and Culture: An Evaluation
of the Work of Brontislaw Malinowski, by Raymond Firth. New
York: Harper and Row. 93-118.
Folmsbee, Stanley J., Robert E. Corlew and Enoch L. Mitchell. 1969.
Tennessee: A Short History. Knoxville: University of Tennessee.
Ford, Thomas R., ed. 1962. The Southern Appalachian Region: A Sur¬
vey. Lexington: University of Kentucky.
Gastil, Raymond D. 1975. Cultural Regions of the United States.
Seattle: University of Washington.
Givon, Talmy. 1976. Topic, Pronoun, and Grammatical Agreement. In
Subject and Topic, edited by Charles N. Li. New York: Academic.
149-88.
Glassie, Henry. 1968. Patterns in the Material Folk Culture of the
Eastern United States. Philadelphia: University of Pennsylvania.
Gleason, Henry, n.d. Some Discourse Features. Manuscript.
Goldstein, Kenneth S. 1964. A Guide for Field Workers in Folklore.
Hatboro, Pennsylvania: Folklore Associates.
Grimes, Joseph E. 1971. Kinds of Information in Discourse. Kivung.
4. 64-74.
Grimes, Joseph E. 1972. Outlines and Overlays. Language. 48.
513-24.
Grimes, Joseph E. 1975. The Thread of Discourse. The Hague: Mouton.

175
Grimes, Joseph E. and Naomi Glock. 1970. A Saramaccan Narrative
Pattern. Language. 46. 408-25.
Gumperz, John and Dell Hymes, eds. 1972. Directions for Sociolin¬
guistics: The Ethnography of Communication. New York: Holt.
Gundel, Jeanette K. 1974. The Role of Topic and Comment in Linguis¬
tic Theory. Ohio State University dissertation.
Gutwinski, Waldemar. 1976. Cohesion in Literary Texts: A Study of
Some Grammatical and Lexical Features of English Discourse.
The Hague: Mouton.
Hackenberg, Robert. 1972. A Sociolinguistic Description of Appala¬
chian English. Georgetown University dissertation.
Hall, Joseph S. 1942. The Phonetics of Great Smoky Mountain Speech.
Morningside Heights, New York: King's Crown Press.
Halliday, M. A. K. 1967-8. Notes on Transitivity and Theme. Jour¬
nal of Linguistics. 3. 37-81. 3. 199-244. 4. 179-215.
Halliday, M. A. K. 1974. Explorations in the Function of Language.
London: Edward Arnold.
Halliday, M. A. K. 1975. Learning How to Mean: Explorations in the
Development of Language. London: Edward Arnold.
Halliday, M. A. K. and Ruqaiya Hasan. 1976. Cohesion in English.
London: Longman.
Hannum, Alberta Pierson. 1969. Look Back with Love: A Recollec¬
tion of the Blue Ridge. New York: Vanguard.
Harris, Zellig S. 1952. Discourse Analysis. Language. 28. 1-30.
Harris, Zellig S. 1963. Discourse Analysis Reprints. The Hague:
Mouton.
Helms, Mrs. W. B. 1930. History of White Pine. Manuscript.
Hendricks, William 0. 1967. On the Notion 'Beyond the Sentence.'
Linguistics. 37. 12-51.
Hendricks, William 0. 1972. Current Trends in Discourse Analysis.
In Current Trends in Stylistics, edited by Braj Kachru and H.
W. Stahlke. Edmonton: Linguistic Research. 83-95.
Hicks, George L. 1976. Appalachian Valley. New York: Holt.

176
Hinds, John V. 1977. Paragraph Structure and Pronominalization.
Papers in Linguistics. 10. 77-99.
Hinds, John V. 1978. Levels of Structure Within the Paragraph.
BLS 4. 598-609.
Hymes, Dell. 1971. The Contribution of Folklore to Sociolinguistic
Research. Journal of American Folklore. 84. 42-50.
Ives, Edward 0. 1976. A Manual for Field Workers. Volume 15 of
Northeast Folklore. Orono, Maine: The University Press.
Johnson, Linda and Richard Bayless. 1976. Cohesion in a Discourse-
Based Linguistic Theory. Manuscript.
Jones, Linda Kay. 1977. Theme in English Expository Discourse.
Lake Bluff, Illinois: Jupiter.
Kaberry, Phyllis. 1957. Malinowski's Contribution to Fieldwork Me¬
thods and the Writing of Ethnography. In Man and Culture: An
Evaluation of the Work of Brontislaw Malinowski, by Raymond
Firth. New York: Harper and Row. 71-91.
Kantor, Robert Neal. 1977. The Management and Comprehension of
Discourse Connection by Pronouns in English. Ohio State Uni¬
versity dissertation.
Keenan, Elinor Ochs and Bambi B. Schieffelin. 1976. Foregrounding
Referents: A Reconsideration of Left Dislocation in Discourse.
BLS 2. 240-57.
Kephart, Horace. 1976[1922]. Our Southern Highlanders. Knoxville:
University of Tennessee.
Ketner, Kenneth Laine. 1973. The Role of Hypotheses in Folkloristies.
Journal of American Folklore.
Klammer, Thomas Paul. 1971. The Structure of Dialogue Paragraphs
in Written English Dramatic and Narrative Discourse. Univer¬
sity of Michigan dissertation.
Klammer, Thomas P. and Carol J. Compton. 1970. Some Recent Contri¬
butions to Tagmemic Analysis of Discourse. Glossa. 4. 212-22.
Koen, Frank M., Alton L. Becker and Richard E. Young. 1968. The
Psychological Reality of the Paragraph. In Proceedings of the
Conference on Language and Language Behavior, edited by Eric M.
Zale. New York: Appleton-Century Crofts. 174-87.

177
Kurath, Hans. 1949. A Word Geography of the Eastern United States.
Ann Arbor: University of Michigan.
Kurath, Hans and Raven I. McDavid, Jr. 1961. The Pronunciation of
English in the Atlantic States. Ann Arbor: University of Mich¬
igan.
La Barre, Weston. 1964. The Snake-Handling Cult of the American
Southeast. In Cultural Anthropology, edited by Ward Goodenough.
New York: McGraw Hill. 309-33.
Labov, William. 1966. The Social Stratification of English in New
York City. Washington: Center for Applied Linguistics.
Labov, William. 1969. Contraction, Deletion, and Inherent Varia¬
bility of the English Copula. Language. 45. 715-62.
Labov, William. 1972. Sociolinguistic Patterns. Philadelphia:
University of Pennsylvania.
Lehiste, Ilse. 1978. Temporal Organization and Prosody-Perceptual
Aspects. Manuscript.
Lehiste, Ilse. 1979. Perception of Sentence and Paragraph Bound¬
aries. To appear in Frontiers of Speech Communication Research,
edited by B. Lindblom and S. Ohman. New York: Academic.
Li, Charles N. and Sandra A. Thompson. 1976. Subject and Topic:
A New Typology of Language. In Subject and Topic, edited by
Charles N. Li. New York: Academic. 457-90.
Linde, Charlotte. 1977. Information Structure in Discourse. In
Studies in Language Variation: Semantics, Syntax, Pragmatics,
Social Situations, Ethnographic Approaches, edited by Ralph W.
Fasold and Roger W. Shuy. Washington: Georgetown University.
226-36.
Longacre, Robert E. 1968. Philippine Languages: Discourse, Para¬
graph, and Sentence Structure. Santa Ana, California: Summer
Institute of Linguistics.
Longacre, Robert E. 1976. An Anatomy of Speech Notions. Lisse:
Peter de Ridder.
Longacre, Robert E. 1977a. A Discourse Manifesto. Manuscript.
Longacre, Robert E. 1977b. An Apparatus for the Identification of
Paragraph Types. Manuscript.
Longacre, Robert E. Forthcoming. The Paragraph as a Grammatical
Unit. To appear in Discourse and Syntax: Syntax and Semantics
12. New York: Academic. 115-34.

178
Longacre, Robert E. and Stephen Levinsohn. 1978. Field Methods of
Discourse. In Current Trends in Text Linguistics, edited by
Wolfgang U. Dressier. New York: Walter de Gruyter. 103-22.
Malinowski, Brontislaw. 1961[1922]. Argonauts of the Western Paci¬
fic. New York: Dutton.
Malinowski, Brontislaw. 1923. The Problem of Meaning in Primitive
Languages. Supplement I to The Meaning of Meaning, by C. K.
Ogden and I. A. Richards. New York: Harcourt.
Martin, Walt. 1977. Cultural Regions of Mid-Nineteenth Century
Tennessee. Manuscript.
Mathesius, Vilém. 1964. On the Potentiality of the Phenomena of
Language. In A Prague School Reader in Linguistics, edited
by Joseph Vachek. Bloomington: Indiana University. 1-32.
McMillan, James B. 1971. Annotated Bibliography of Southern Ameri¬
can English. Coral Gables: University of Miami.
Mencken, H. L. 1963. The American Language. The Fourth Edition
and the Two Supplements. Abridged, with Annotations and New
Material, by Raven I. McDavid, Jr. New York: Knopf.
Merritt, Marilyn. 1972. Punctuating Devices in Discourse: Prelim¬
inary Notes. Paper read at the 1972 LSA Winter Meeting, Atlanta.
Montgomery, Michael B. 1975. Attitudes of the Southern Mountain
People Toward Language. Manuscript.
Montgomery, Michael B. 1977a. Topics Are Dislocated All Around Us.
Manuscript.
Montgomery, Michael B. 1977b. Subject Relative Pronoun Deletion: A
Discourse Function. Paper Read at the South Atlantic American
Dialect Meeting, November 5, 1977.
Montgomery, Michael B. 1978. Left Dislocation: Its Mature in Appa¬
lachian English. The SECOL Bulletin. II. 2. 55-61.
Odum, Howard W. 1936. Southern Regions of the United States.
Chapel Hill: University of North Carolina.
Pederson, Lee, Raven I. McDavid, Jr., Charles W. Foster, and Charles
E. Billiard, eds. 1974. A Manual for Dialect Research in the
Southern States. Tuscaloosa: University of Alabama.
Petdfi, J. S. and H. Rieser. 1976. Studies in Text Grammar.
Boston: D. Reidel.

179
Pike, Kenneth L. 1964. Discourse Analysis and Tagmeme Matrices.
Oceanic Linguistics. 3. 5-25.
Pike, Kenneth L. 1967. Language in Relation to a Unified Theory
of the Structure of Human Behavior. The Hague: Mouton.
^Pike, Kenneth L. and Ivan Lowe. 1969. Pronominal Reference in
English Conversation and Discourse — A Group Treatment.
Folia Lingüistica. 3. 68-106.
Pike, Kenneth L. and Evelyn G. Pike. 1977. Grammatical Analysis.
Dallas: Summer Institute of Linguistics.
Postal, Paul M. 1971. Cross-Over Phenomena. New York: Harper & Row.
Prince, Ellen F. 1978a. A Comparison of WH-clefts and it-clefts in
Discourse. Language. 54. 883-906.
Prince, Ellen F. 1978b. Left Dislocation, It Makes Life Easier.
Paper read at the 1978 LSA Winter Meeting, Boston.
Propp, Vladimir. 1958. Morphology of the Folktale. Bloomington:
Indiana University.
Raine, James Watt. 1924. The Land of Saddle Bags. New York: Coun¬
cil of Women for Home Missions and Missionary Education Movement
of the United States and Canada.
Reed, Ann M. 1979. 'There' and Definites. The SECOL Bulletin.
III. 1. 9-24.
Ross, John Robert. 1967. Constraints on Variables in Syntax. MIT
dissertation.
Sacks, Harvey, Emanuel A. Schegloff and Gail Jefferson. 1974. A Sim¬
plest Systematises for the Organization of Turn-Taking for
Convestigation. Language. 50. 696-735.
Sinclair, J. McH. and R. M. Coulthard. 1974. Towards an Anatomy of ^
Discourse: The English Used by Teachers and Students. London:
Oxford University.
Stewart, William A. 1967. Language and Communication Problems in
Southern Appalachia. Washington: Center for Applied Linguistics.
Stockwell, Robert P., Paul Schachter, and Barbara Hall Partee. 1973.
The Major Syntactic Structures of English. New York: Holt.
United States Department of Agriculture. Bureau of Agricultural
Economics. Bureau of Home Economics, and Forest Service. 1935.
Economic and Social Problems and Conditions of the Southern
Appalachians. Miscellaneous Publication 205. Washington, D. C.

180
Vance, Rupert B. 1935. Human Geography of the South. Chapel Hill:
University of North Carolina.
Vance, Rupert B. 1962. The Region: A New Survey. In The Southern
Appalachian Region: A Survey, edited by Thomas R. Ford. Lexing¬
ton: University of Kentucky. 1-8.
Van Dijk, Teun A. 1972. Some Aspects of Text Grammars. The Hague:
Mouton.
Van Dijk, Teun A. 1977. Text and Context: Explorations in the Seman¬
tics and Pragmatics of Discourse. London: Longman.
Van Dijk, Teun A. and Janos S.' Petdfi, eds. 1977. Grammars and
Descriptions: Studies in Text Theory and Text Analysis. New
York: Walter de Gruyter.
Walker, Ralph S. 1939. A Mountaineer Looks at His Own Speech.
Tennessee Folklore Society Bulletin. 5. 1-13.
Watkins, Floyd. 1949. The Southern Mountaineers; Archaic English.
Georgia Review. 3. 219-25.
Weatherford, W. D. and Earl J. C. Brewer. 1962. Life and Religion
in Southern Appalachia. New York: Friendship.
Weller, Jack. 1965. Yesterday's People. Lexington: University
of Kentucky.
West, John Foster. 1966. Dialect of the Southern Mountains.
North Carolina Folklore. 14. 31-4.
Widdowson, Henry G. 1977. Approaches to Discourse. In Grundbe-
griffe und Hauptstrdmungen der Linguistik. Hamburg: Hoffman
und Campe. 236-60.
Williams, Cratis D. 1961. Vowels and Diphthongs in Mountain
Speech. Mountain Life and Work. 37. 8-11.
Williams, Joseph Marek. 1966. Some Grammatical Characteristics
of Continuous Discourse. University of Wisconsin dissertation.
Wise, C. M. 1957. Applied Phonetics. Englewood Cliffs, New Jersey:
Prentice-Hall.
Wise, Mary Ruth. 1968. Identification of Participants in Discourse:
A Study of Aspects of Forms and Meaning in Nomatsiguenga.
Ann Arbor: University of Michigan dissertation.

181
Wise, Mary Ruth. 1971 Identification of Participants in Discourse:
A Study of Aspects of Form and Meaning in flomatsiguenga. Norman,
Oklahoma: Summer Institute of Linguistics.
Wolfram, Walt. 1977. On the Linguistic Study of Appalachian Speech.
Appalachian Journal. 5. 92-102.
Wolfram, Walt and Donna Christian. 1975. Sociolinguistic Variation
in Appalachian Dialects. Final Report, NIE Grant Number 74-0026.
Wolfram, Walt and Donna Christian. 1976. Appalachian Speech.
Arlington: Center for Applied Linguistics.

BIOGRAPHICAL SKETCH
Michael Bryant Montgomery was born on May 15, 1950, in Knox¬
ville, Tennessee, and attended public school in that city, gradua¬
ting from Holston High as salutatorian of his class in 1968. There¬
after he entered Maryville College, Maryville, Tennessee, and gradu¬
ated cum laude with a Bachelor of Arts in English from Maryville in
1973. While pursuing a master of Arts in English at the University
of Tennessee at Knoxville, he determined to study linguistics. After
receiving his master's in 1975, he enrolled in the Program in Linguis¬
tics at the University of Florida to study for the doctorate. During
the 1978-9 academic year he was Instructor in English at the Univer¬
sity of Arkansas at Little Rock and coordinator of the Department of
English's Writer's Hotline Service. Beginning in the fall of 1979 he
will be Assistant Professor of English at Memphis State University in
Memphis, Tennessee. His favorite hobby is linguistics.
182

I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation and
is fully adequate, in scope and quality, as a dissertation for the
degree of Doctor of Philosophy.
William Sulliv^h,'Chairman
Associate Professor of
Linguistics
I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation and
is fully adequate, in scope and quality, as a dissertation for the
degree of Doctor of Philosophy.
yean Casagratfde
Associate Professor of
Linguistics
I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation and
is fully adequate, in scope and quality, as a dissertation for the
degree of Doctor of Philosophy.
fU,- V /-vtrtupSy
Roger W. Thompson J
Associate Professor of
Linguistics
I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation and
is fully adequate, in scope and quality, as a dissertation for the
degree of Doctor of Philosophy.
Norman N. Markel
Professor of Speech

I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation and
is fully adequate, in scope and quality, as a dissertation for the
degree of Doctor of Philosphy.
Robert S"irThomson
Associate Professor of
Engljsi)/
This dissertation was submitted to the Graduate Faculty of the Program
in Linguistics in the College of Liberal Arts and Sciences and to the
Graduate Council, and was accepted as partial fulfillment of the re¬
quirements for the degree of Doctor of Philosophy.
August 1979
Dean, Graduate School
ii V

UNIVERSITY OF Fl-ORIDA
ML 7106



109
juncture between the NP and what comes after, but rarely is there
any sort of hesitation phenomena ('tih" or a long pause, for instance)
which would indicate that a speaker might be changing course after
beginning with an NP. Other than appositional information, it is also
rare for anything to intervene between the NP and the sentence in which
it is thematic.
Functions
To understand the functioning of left dislocation in oral exposi
tions, we must determine 1) the relationship of left dislocated NPs to
preceding portions of discourse; 2) the relationship of left dislocated
NPs to subsequent portions of discourse; and 3) the distribution of
left dislocated NPs over entire expositions.
As is remarked above (p. 105), left dislocation usually involves
pieces of information which are NEW, i.e. knowable to or recognizable
by the hearer from outside the immediate verbal context. Of the 606
instances of left dislocation identified in the White Pine corpus, 571
involve such NEW information (94.2 percent of the total, that is).
Thirty-five instances (5.8 percent of the total) involve GIVEN informa
tion (which the speaker would probably presume to be in the conscious
ness of the hearer because it had been mentioned in a closely preceding
context). These seem to involve the speaker's giving a special empha
sis to the left dislocated piece of information.
Four types of dislocated NEW information suggest themselves.
These types are not mutually exclusive. For example, a left dislocated


39
prepared. As thorough and as broad an ethnography as possible is in
dispensable to successful and accurate anthropological and linguistic
fieldwork. It is necessary to explain this point in regard to linguis
tic research in general and to this project in particular.
Speech is social behavior par excellence. As Hymes indicates, it
is the "nexus between language and social life" (1971:42). We cannot
accurately describe language (or, as Malinowski would say, meaning)
without reference to its role in society and the setting, participants,
events, acts, norms, etc., which it involves. Speech must be viewed as
permeating and not just overlaying most social behavior; therefore, de
scription of such behavior must make reference to language and its
forms.
No aspect of a culture can accurately be described without explica
ting its interrelationships with and dependencies on other aspects of
the culture. This has been more or less axiomatic since Malinowski's
ethnographic work among the Trobriand Islanders more than sixty years
ago. Malinowski insisted that a researcher approach a way of life as
a whole, as full of interdependencies, as an integration of life. Only
the application of the most rigorous techniques and standards for the
collection of data of all kindsmaterial, linguistic, topographical,
personal, ceremonial, among otherscould enable the adequate explica
tion of the "context of situation," to use the term Malinowski (1923:306)
himself coined. All pertinent data must inform an understanding of a
context of situation before a generalization can be made, in other
words. For Malinowski, a context of situation is at the same time


142
as those extra-positions discussed above on page 136 do. In these
cases the deictic points forward to the information which has not
been previously mentioned by a speaker, as in
35) The girls are just, you know, they're, they're so strong
and you can see their weak points, but you can see how
they overcome them, and they're outgoing. And 1 guess maybe
that's the wav I like to be, you know, to be able to get
out and to be, so I won't be so nervous and everything.
(XI-6)
36) It's a friendly town. It really is. It's small, and you
just kind of hate to leave it, you know, really, but I was
surprised. I didn't really think that that many people
stay. I like White Pine a lot, but I don't know if I want,
you know, to stay here for rest of my life or whatever, but
a lot of them's gotten married and they're still living
here in White Pine, and so I think that's the biggest reason
White Pine's really grown a lot, because all the, you know,
the kids are growing up and everything and they're staying
here. (VII3-13)
In the first example, "to be able . everything" is appositive to
the identifying sentence; in the second, "because . here" is.
We perhaps cannot say that the deictics in such contexts are over
whelmingly cataphoric, but although they do weakly point backward,
they do not clearly point to any specific portion of the preceding
discourse and, more importantly, the discourse would be incomplete
in these instances without the appositional information. The force of
these deictics is predominantly cataphoric.
On the other hand, we also find that identifying sentences with
WH-clauses may also have appositional information attached afterward,
as in
37) And then I came back to Morristown. My brother-in-law and
sister opened a, an office in Morristown and so I came down
to live with them and go to Morristown Business School, and


52
Agriculture
Construc
tion
Manufac
turing
Trade,
and
, Service
Finance
, Other
#
%
#
%
# %
#
%
#
%
1970 Census 89
6.3
129
9.1
720 51.0
327
23.1
142
10.0
Present
Study Infor- 3
8.1
3
8.1
6 16.2
21
56.8
4
10.8
mants
If females were not members of the labor force, they were categorized
according to the employment of the breadwinner of their families. The
present study's higher percentage of informants supported by trade, ser
vice, and finance than the percentage of those depending on manufactur
ing contrasts with the percentages for the overall White Pine area.
This is due to their availability. There were more people in trade,
service, and finance positions who were available for interviews than
in manufacturing.
Although no census figures are available on the lengths of resi
dence of White Pine inhabitants, it is clear that the great majority
of them are lifelong natives of White Pine or the nearby area. Of the
forty informants, twenty-two were natives of White Pine and another
thirteen were natives of nearby counties in East Tennessee, nine of whom
had spent more than half their lives in White Pine. Four were natives
of elsewhere in Tennessee or of another state, and one was a native of
Great Britain. Informants were chosen on the bases of their ages and
educational backgrounds and not their nativities. Even so, the cohe
sive devices analyzed in chapter three through five were examined for
differences in function between natives of White Pine and non-natives.
None but incidental differences in frequency were found.


CHAPTER III
THEMATIC DEVICES
General Remarks
It is indicated in chapter two that a speaker may use certain
constructions known as cohesive devices to signal to a hearer what
the theme of a following block of the speaker's message will be.
Theme is the obligatory unit of such a block or paragraph and is de
fined as the piece of information (not necessarily a human participant)
which initiates an anaphora chain and which the block is about. Al
though it must be generalized to aDply to all paragraphs, this narrow
definition of theme demonstrably applies to those paragraphs in our
corpus beginning with thematic devices.
In this chapter our concept of theme is compared with others
(Halliday 1967-8, Firbas 1966, Jones 1977) and we discuss briefly those
constructions that these treatments (especially Jones) consider to be
thematic devices. The main portion of the chapter explores the nature,
function(s), and distribution of one particular deviceleft disloca
tion. This is one very common device through which paragraphs are
formed in this corpus of expository monologues. Finally we look at
existential sentences in the corpus and see how they also function as
thematic devices. They also complement left dislocation.
More often than not themes are reflected by anaphora chains ex
tending for more than one sentence. An anaphora chain is a string of
89


I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation and
is fully adequate, in scope and quality, as a dissertation for the
degree of Doctor of Philosphy.
Robert S"irThomson
Associate Professor of
Engljsi)/
This dissertation was submitted to the Graduate Faculty of the Program
in Linguistics in the College of Liberal Arts and Sciences and to the
Graduate Council, and was accepted as partial fulfillment of the re
quirements for the degree of Doctor of Philosophy.
August 1979
Dean, Graduate School
i)08l;


171
A dollar and a quarter. Now you tell me any city in the state of
Tennessee thats, or anywhere for that fact, that has a tax rate
that low. That's ridiculous. . .


30
Ford surveyed nearly 1,500 homes in a region closely coterminous
with Campbell's Southern Highland region (page 22 above) in three dif
ferent kinds of communityrural, urban, and metropolitanwith a
questionnaire devised to indicate the strength of four prominent
traitsindividualism and self-reliance, traditionalism, fatalism, and
fundamentalist religion. It was hypothesized that, because of the eco
nomic and social changes which had taken place after the Second World
War, a disparity between rural and non-rural manifestation of these
traits would be found. Although he did find some disparity, his results
show that the four traits he chose form a common denominator for Southern
Appalachia in general. Ford outlines his choice of traits in this man
ner:
In examining the web of mountain life, one finds these themes
intertwined and generally, though not always, mutually support
ing. Most so-called "mountain traits" are to be found in one
form or another throughout the nation, particularly in rural
areas. At the same time, each of them has its antithesis in
contemporary industrial society. The self-reliant individual
ist, at least as an "ideal type," stands at the far end of the
scale from the much berated "organization man." Traditionalism,
not only in the sense of clinging to an earlier heritage but
also in the exaltation of resistance to social change, is viewed
as both anachronistic and vaguely immoral by a larger society
that values progress through rational, scientific endeavor.
Even more reprehensible to a culture that stresses achievement,
self-betterment, and mastery ovor nature is a passive resigna
tion to one's situation in life, particularly if it is a situa
tion viewed as both undesirable and remediable. Less subject to
censure by the larger society, perhaps, but contrasting as
sharply with its dominant valuesand not immune from ridicule
is the rigid, pervasive religious ethos of the [Southern Appala
chian] Region. (p. 11)
Commentators on Southern Appalachia have long characterized its
people as individualistic, self-reliant, and non-social. Campbell


40
broader and immensely more useful than the concept of a "linguistic
context": "Even as in the reality of spoken or written languages, a
word without 'linguistic context' is a mere figment and stands for
nothing by itself, so in the reality of a spoken living tongue, the
utterance has no meaning except in the 'context of situation'"
(1923:307).
In addition to the many kinds of data Malinowski considered useful
he also advocated that a researcher keep two kinds of personal records
which should constitute parts of his field records (Kaberry 1957:78-9):
1) An ethnographic diary of daily activities and observations not
ing what is the normal and the typical and what is not, the
kind of information that is not elicitable by questions and an
swers of a community's residents. Malinowski called such infor
mation the "imponderabilia of actual life." The present re
searcher kept such a diary while in the field, a diary which
also recorded the accounts of his successes and failures and
mistakes in the field.
2) A "corpus inscriptionum," which "entailed the recording in the
native language of narratives, opinions, typical utterances,
myths, folk-lore, magical formulae, and native explanations and
interpretations of customs and beliefs" (Kaberry, p. 79).
Since this researcher collected only one type of discourse (ex
pository, as defined in chapter two), his forty five hours of
taped explanations and descriptions constitute his "corpus in
scriptionum."


42
rather than the accidental as an appropriate object of study. Firth's
view also makes the important implication that there is no necessary
one-to-one relationship between an utterance and its meaning. This
implication is most valuable when we examine typical behavior.
The typical behavior (an utterance, in this case) from the White
Pine corpus to be explored briefly here is the expression "I don't
know." At times in the interviews this expression became almost annoy
ingly frequent. Consider an excerpt from an interview with a thirty-
nine-year-old housewife:
Fieldworker: Why should people be proud of White Pine? What
would you say?
Informant: Well, as a community, I think it's uh, it's a good
community to live in, to raise your children in. Well, 1
don't know. The churches, the schools, everything. I think
it's just a, I don't know. I like it. I think it's, that's
why I think it's a good community is because I like it,
(uh-huh, uh-huh) and I don't know.
Of course, it may be that the informant does not know how to artic
ulate her feelings here, or perhaps she in fact does not know why she
should be proud of White Pine, even though she had just stated, before
the fieldworker's questions, that the residents of the community should
be proud of White Pine. But this can hardly account for there being so
many occurrences in this excerpt of so deceptively short an expression
(three times in about twenty seconds). There are at least three differ
ent I don't know's" to be found in the corpus, two of which are exem
plified here. Although this cannot be as clearly seen from the written
transcript, the researcher clearly identified at least three distinct
meanings for this expression:


no
NP may be both contrastive and reintroduced.
1) The majority of left dislocated NPs represent new partici
pants or other pieces of information which a speaker had not pre
viously mentioned. Of the total of 606 instances, 488 (80.5 percent)
are primarily of this type. In introducing new participants or char
acters or in indicating the principal object or location of discus
sion for the ensuing sentence (and further discourse usually), left
dislocation is thus a stage-setting device. We see three examples
of this in the following discussion of his immediate family by a sixty-
two-year-old man:
43) Mother, she's been dead fifty two year, (hmm) and Dad, he's
been dead about uh fourteen year, (hmm) So therefore uh,
I had a sister, she's been dead, oh, I'd say twenty five
year, and I had a half-brother died nine months old. He
was a twin, and the other one, ire died of meningitis. He
choked to death. That's, that's been uh thirty, thirty
years, I guess, or maybe a little longer than that, just
not knowing the time. (IV2-2)
Introducing new information for the first time is thus the primary func
tion of left dislocation. This is true whether the left dislocated NP
represents a subject or an object, whether it repersents an element in
the matrix or an embedded clause. Prince (1978) claims that left dis
located subjects tend to represent new information while left disloca
ted non-subjects in general represent old information. There is no
support for these contentions in this corpus' data. Nearly 95 percent
of left dislocations serve to introduce new information, regardless of
whether they are subjects or not.
2) Another type of left dislocation represents a special kind of


150
Viewed in isolation, instances of the three kinds of deictics
(those in peak sentences, in anaphoric identifying clauses, and in
extrapositions) may not then be readily distinguishable. It is only
when we examine their anaphoricity or cataphoricity in their dis
course context that they can be distinguished.
The peak sentences (examples 48-53 above) we have looked at in
their discourse contexts come either at the end or towards the end
of a speaker's exposition of a point. Moreover, they clearly have
the effect of summing up and culminating a foregoing portion of
discourse sometimes of considerable length, as in 48, where a teenage
speaker explains the best way he knows to get away and clear his
thoughts, or in 49, where a Baptist minister explains the one most
important change he has observed in the town of White Pine. In only
one case of fifty does a peak sentence come at the beginning of an
informant's response:
54) Fieldworker: The more I listen to disco music, the more
it seems to me it all sounds the same.
Informant: It does. It's got the same beat, same loudness.
That's what a lot of people do too is go to places like
Flanagan's and Uncle Sam's and dance.
In this example, "that" is not anaphoric and the peak sentence does
not sum up any portion of foregoing conversation. It may be that it
sums up information implicit in the conversation or in the mind of the
informant, but it clearly sums up what the informant has to say on the
subject of entertainment. It does not really constitute a counterexam
ple to our generalization that peak sentences follow and sum up blocks
of discourse.


All structures are examined with the view of identifying what is
typical about them, which is why they are quantified numerically when
ever possible. This is to insure that the data are not chance findings
of random observation.
It was found that spontaneous expository monologues contain super-
sentential units that are thematically unified and that various surface
structure cohesive devices play significant roles in creating such
unity for the benefit of a hearer. These units are shown to be oral
paragraphs.


59
are involved. Fieldwork is extremely hard and tiring work, but its
rewards in terms of friendships gained justify the effort put in by
the fieldworker.
From this researcher's perspective, at least three skills are re
quired of the linguistic fieldworker in such a study as this one:
1) The ability and willingness of a fieldworker to relate to his
informants as humans who are willing to share their personal, and some
times intimate, experiences with him. The fieldworker must value his
informants as people and not as sources of data, which means that it
is probably impossible to do good fieldwork without genuinely enjoying
human contact.
2) A resourcefulness to adapt to new situations and people in ways
that the fieldworker could not have foreseen before going into the
field. The fieldworker must take care and even pains to approach peo
ple as individuals without preconceptions and prejudices. He should ap
proach his corrmunity as a microcosm having a great variety of unique in
dividuals.
3)A disposition to persevere against frustrations and inconveni
ences and ever-present delays. In short, he must be so committed to
3
doing his fieldwork that he develops a kind of salespitch for it. He
must be willing to solicit assistance from often-hesitant and sometimes
skeptical people while at the same time endeavoring to establish a fa
vorable rapport with them.
Though the most demanding work he had ever done, this writer found
it immensely rewarding. It gave him many fond memories and several
good friendships which have developed from his interviews.


159
Although older informants employ it to make such elaborations, we
find it much more common among the eight younger speakers, in whose
expositions it appears on 164 occasions. Host often used by one
seventeen-year-old high school student (47 instances), it was still
frequently used by all the remaining seven younger speakers.
The prevalence of "like" with left dislocated NPs might be inci
dental if it were not that it seems to have a slightly different func
tion when accompanying the thematic NPs of left dislocation. Often,
the NPs it accompanies are not thematic for more than one sentence.
Thirteen of the 29 (44.8 percent) are like this. More important, the
NPs seem to establish only limited themes which are not fully devel
oped. For instance, in the following examples, "our parents" and
"Brooklyn" are introduced only to make a single comment about them
before the speaker moves on. Like many other NPs introduced with
"like," they are not established as themes.
5) . .Everyone thinks that they have a right to live their own
lives the way they want to, and they do. And I think we
see that more, you know, (hmm) Like our parents, they
think that we started the drugs and all that stuff, but it
was here before we got here, you know. That's how I
feel about it. (VII3-23)
6) Now in the World War Two I was, while I was in service, of
course you were with people from all over the United States,
(uh-huh) and I could tell by the accent within two or three
states or a thousand miles usually where someone was from,
(uh-huh) Like Brooklyn, you know their lingo, (uh-huh) and
you could tell if it's Alabama or Georgia. You could tell
by their talk, . (V3-16)
Even in expositions in which the like-accompanying NP is thematic for
longer than one sentence, the theme has this limited quality. In


86
-PROJECTED
+PROJECTED
s
NARRATIVE
PROCEDURAL
u
. 1. 1/3 person.
I. Non-specific person.
c
2- Agent oriented.
2. Patient oriented.
c
3. Accomplished time.
3. Projected time.
+ E
4. Chronological linkages
4. Chronological linkage.
s
.
s
I
0
N
s
EXPOSITORY
HORTATORY
u
1. No necessary person reference.
I. 2 person.
c
2. (Subject matter oriented).
2. Addressee oriented.
c
3. Time not focal.
3. (Mode, not time).
-E
4. Logical linkage.
4. Logical linkage.
s
s
I
o
N
Diagram I. Deep structure genre
Figure 2.1 Deep Structure Genre
(From Longacre 1976:200)


106
31) One of the biaqer Blacks, he's about twenty four years old.
(13-2)
32) People my age that went North and worked, some of 'em have
come back [to] the territory. (IV418)
But not in one instance is the left dislocated NP a nonspecific indef
inite.:. In her study, Prince (1978b) agrees that nonspecific indefi
nites cannot be dislocated. In general, her contention that specific
indefinite NPs. which are dislocated represent subjects is consistent
with the data from our corpus. Of the several dozen dislocated specif
ic NPs, only three are not subjects, two of which are 33 and 34 below:
33) Certain people who have money in this town, people look up to
'em. (.13-6)
34) Well, each, each section, I think they call it a pod. (VI-6)
Such examples are equivalent to passive sentences with specific indefi
nites dislocated from the subject position.
Another contention of Prince, however, is disaffirmed by the data
from White Pine,. She claims that left dislocated subjects are usually
new information while left dislocated non-subjects reintroduce old in
formation.. It is shown below that nearly all left dislocations repre
sent new information, regardless of which syntactic position in the sen
tence they are related to.
Further evidence for all left dislocated NPs not being nonspecific
indefinites, is available. There was not a single instance of an indefi
nite pronoun serving as the copy of the dislocated NP. Our data leave
no doubt that left dislocation is a thematic device constrained not to


8
many generations in White Pine itself; some are descendants of the
original settlers in the area.
The rolling terrain of the White Pine area is better than average
farm and pasture land, with corn, tobacco, and wheat being the staple
crops. Although not in the mountains itself, White Pine lies in the
center of the region referred to as Southern Appalachia and is in many
respects a typical small town of the region. Although it is sometimes
supposed that all Appalachian communities are socially and culturally
homogeneous, only the extremely isolated areas actually are. There
fore White Pine is perhaps more typical of Southern Appalachia because
it is neither homogeneous nor isolated.
To the outsider, White Pine seems a slow-paced and unassuming
small town, all of whose residents know one another. It seems to be a
town with few pretentions. On a not atypical day one can get a good
flavor of the town's personality by visiting the town's center of ac
tivity and one of its chief public forums, the drugstore at Main and
Maple, where the mayor, the chief of police, one or two city alderman,
the town's physician, and a local minister hobnob over coffee.
Such an appearance would be rather deceptive, for in recent years
White Pine has entered a period of great transition socially and eco
nomically. It has also considerably changed in the way it views it
self (the near doubling of the town's population since 1960 through
annexation and immigration would suggest this anyway). Life has be
come more complex for its residents as the changes have come. For
such a long time White Pine resisted change of any kind that many of


148
the anaphoricity predominates; in others, as in 51, the cataphoricity
clearly is more prominent:
50) ... and I think that uh the main thing I think in the candy-
pulling that we, course we like the candy and that was a
good way to have it, but we liked the fellowship and together
of all the young people. That was, that was the main thing,
I mean, that we loved to get together. And I said like this,
so many, well in fact, that's where I met my husband was at
a candy-pulling. (II12-12}
51) I cannot pass, well, bratwurst is my favorite, those large
white sausage, you know, on hot, hard sour dough bread, uh
with hyacinth, very hot mustard. I just, I can't pass one
up. I went with my father-in-law uh on one trip while we
was there, and we stopped for lunch, why, that's all I
could think about was getting me some potato salad and
getting me a couple of bratwurst on bread. (I3-18)
In the discussion of identification sentences with right dislocation
above (pp. 140-1), it was indicated that those with WH-clauses tended
to have initial deictics that are anaphoric and that those without
such clauses tended to have cataphoric ones. Notice that the ana
phoric deictic in 51 above is followed by a WH-clause and that the
cataphoric one in 52 is not. Of the ten instances of peak sentences
with WH-clauses in the data, only three are not clearly anaphoric.
The peak sentences without such clauses have deictics which are more
clearly ambiphoric. The seven strongly anaphoric sentences, like
50 above and 52 below, repeat information verbatim, as does 37 above.
52) Well, as a community, I think it's a, it's a good community
to live in, to raise your children in. Well, I don't know.
The churches, the schools, everything. I think it's just
a, 1 don't know, I like it. I think it's, that's why I think
it's a good communitv is because I like it. and I don't know.
(VI7)
The forty remaining instances have deictics which are cataphoric to
varying degrees; none of them involve a repetition as in 52. The


93
dichotomy of sentence organization originally proposed by one of the
founders of the Prague School of linguistics, Vilem Mathesius, a half
century ago.
Firbas' refinement is that theme is "constituted by the sentence
element (or elements) carrying the lowest degree(s) of Communicative
Dynamism within the sentence" (1966:272). He defines "degree of Com
municative Dynamism" as "the extent to which the sentence element con
tributes to the development of the communication, to which it 'pushes
the communication forward,1 as it were" (p. 270). Firbas thus views
sentences as organized not in a strict dichotomy but as having parts
with various degrees of dynamism. Since he does not provide a clear
way to measure the dynamism of sentence elements, however, his criter
ion for theme is not readily applicable.
Firbas1 conception of theme is different from this study's on sev
eral accounts. We view theme as that constituent around which a block
or paragraph of language is organized, whether it be one or more sen
tences long. This constituent may be new information (not previously
mentioned) or given information (previously mentioned) in the context
of the discourse as a whole. Firbas' conception restricts theme to a
sentential domain and to being only given information. This conception
is too limited and puzzling, since obviously an element's discourse con
text must be considered before determining its newness or givenness.
The work of Jones (1977) on theme in expository written English
discourse is similar to this study's approach in many respects. Insist
ing that theme should not be characterized in clausal or sentential


177
Kurath, Hans. 1949. A Word Geography of the Eastern United States.
Ann Arbor: University of Michigan.
Kurath, Hans and Raven I. McDavid, Jr. 1961. The Pronunciation of
English in the Atlantic States. Ann Arbor: University of Mich
igan.
La Barre, Weston. 1964. The Snake-Handling Cult of the American
Southeast. In Cultural Anthropology, edited by Ward Goodenough.
New York: McGraw Hill. 309-33.
Labov, William. 1966. The Social Stratification of English in New
York City. Washington: Center for Applied Linguistics.
Labov, William. 1969. Contraction, Deletion, and Inherent Varia
bility of the English Copula. Language. 45. 715-62.
Labov, William. 1972. Sociolinguistic Patterns. Philadelphia:
University of Pennsylvania.
Lehiste, Ilse. 1978. Temporal Organization and Prosody-Perceptual
Aspects. Manuscript.
Lehiste, Ilse. 1979. Perception of Sentence and Paragraph Bound
aries. To appear in Frontiers of Speech Communication Research,
edited by B. Lindblom and S. Ohman. New York: Academic.
Li, Charles N. and Sandra A. Thompson. 1976. Subject and Topic:
A New Typology of Language. In Subject and Topic, edited by
Charles N. Li. New York: Academic. 457-90.
Linde, Charlotte. 1977. Information Structure in Discourse. In
Studies in Language Variation: Semantics, Syntax, Pragmatics,
Social Situations, Ethnographic Approaches, edited by Ralph W.
Fasold and Roger W. Shuy. Washington: Georgetown University.
226-36.
Longacre, Robert E. 1968. Philippine Languages: Discourse, Para
graph, and Sentence Structure. Santa Ana, California: Summer
Institute of Linguistics.
Longacre, Robert E. 1976. An Anatomy of Speech Notions. Lisse:
Peter de Ridder.
Longacre, Robert E. 1977a. A Discourse Manifesto. Manuscript.
Longacre, Robert E. 1977b. An Apparatus for the Identification of
Paragraph Types. Manuscript.
Longacre, Robert E. Forthcoming. The Paragraph as a Grammatical
Unit. To appear in Discourse and Syntax: Syntax and Semantics
12. New York: Academic. 115-34.


175
Grimes, Joseph E. and Naomi Glock. 1970. A Saramaccan Narrative
Pattern. Language. 46. 408-25.
Gumperz, John and Dell Hymes, eds. 1972. Directions for Sociolin
guistics: The Ethnography of Communication. New York: Holt.
Gundel, Jeanette K. 1974. The Role of Topic and Comment in Linguis
tic Theory. Ohio State University dissertation.
Gutwinski, Waldemar. 1976. Cohesion in Literary Texts: A Study of
Some Grammatical and Lexical Features of English Discourse.
The Hague: Mouton.
Hackenberg, Robert. 1972. A Sociolinguistic Description of Appala
chian English. Georgetown University dissertation.
Hall, Joseph S. 1942. The Phonetics of Great Smoky Mountain Speech.
Morningside Heights, New York: King's Crown Press.
Halliday, M. A. K. 1967-8. Notes on Transitivity and Theme. Jour
nal of Linguistics. 3. 37-81. 3. 199-244. 4. 179-215.
Halliday, M. A. K. 1974. Explorations in the Function of Language.
London: Edward Arnold.
Halliday, M. A. K. 1975. Learning How to Mean: Explorations in the
Development of Language. London: Edward Arnold.
Halliday, M. A. K. and Ruqaiya Hasan. 1976. Cohesion in English.
London: Longman.
Hannum, Alberta Pierson. 1969. Look Back with Love: A Recollec
tion of the Blue Ridge. New York: Vanguard.
Harris, Zellig S. 1952. Discourse Analysis. Language. 28. 1-30.
Harris, Zellig S. 1963. Discourse Analysis Reprints. The Hague:
Mouton.
Helms, Mrs. W. B. 1930. History of White Pine. Manuscript.
Hendricks, William 0. 1967. On the Notion 'Beyond the Sentence.'
Linguistics. 37. 12-51.
Hendricks, William 0. 1972. Current Trends in Discourse Analysis.
In Current Trends in Stylistics, edited by Braj Kachru and H.
W. Stahlke. Edmonton: Linguistic Research. 83-95.
Hicks, George L. 1976. Appalachian Valley. New York: Holt.


as well as linear in structure, that it has, in other words, struc
tural units larger than the sentence. It hypothesizes in particular
that speech has paragraph structure. If this first hypothesis is
true, oral paragraphs should be describable to some degree of pre
cision and generality.
Paragraphs may be viewed as having phonological, lexemic, and
grammatical structure. This study examines their grammatical struc
ture and makes a second hypothesis: that the grammatical structure
of oral expository paragraphs is organized by certain surface struc
ture cohesive devices. It investigates this hypothesis by considering
the function(s) and distribution of the following devices:
1) Introducers of themes into the stream of discourse. Such
introducers include left dislocation (as in "My mother, she's a great
cook"), existentials (as "There's a man here to see you"), it-clefts
(as "It was my brother who called") and WH-clefts (as "What I did was
to completely strike out").
2) Deictics (pointing words such as "that" and "this"), both
cataphoric (forward-pointing) and anaphoric (backward-pointing).
Structures examined are that-extraposition (as "That is good that John
finally left"), identifying sentences (as "That's what I meant"), and
peak sentences (as "That's the best thing is to leave early").
3) Introducers and conjunctions ("and," "but," "now," "then,"
"like," and others) when they accompany left dislocation and assist
in introducing themes.
vii


3
cataphoric (forward-pointing) and anaphoric (backward-pointing).
Structures examined are that-exptraposition (as That is good that
John finally left"), identifying sentences (as That's what I meant"),
and peak sentences (as "That's the best thing is to leave early").
3)Introducers and conjunctions ("and," "but," "now," "then,"
like," and others) when they accompany left dislocation and assist
in introducing themes.
All devices are examined with the view of identifying their
typical function(s) and distribution within greater discourse block
structures. Hence they are quantified numerically whenever possible.
This is to insure that the data are not chance findings of random
observation. Also to insure the nature of the data gathered, five
different parameters are controlled in this study:
1) The speaker context (a representative sample of forty infor
mants was interviewed);
2) The frequency context (data are quantified whenever possible);
3) Social context (a matter of controlling the speech style
elicited; the fieldworker-researcher assumed one role with all infor
mants) ;
4) Task control (interview was controlled to elicit only explana
tions and descriptions from informants);
5) Linguistic context (the sentence and discourse position of
all data was taken into account).
For the most part (except the fourth and, in part, the fifth),
these have been controlled in the studies of Labov. Labov


181
Wise, Mary Ruth. 1971 Identification of Participants in Discourse:
A Study of Aspects of Form and Meaning in flomatsiguenga. Norman,
Oklahoma: Summer Institute of Linguistics.
Wolfram, Walt. 1977. On the Linguistic Study of Appalachian Speech.
Appalachian Journal. 5. 92-102.
Wolfram, Walt and Donna Christian. 1975. Sociolinguistic Variation
in Appalachian Dialects. Final Report, NIE Grant Number 74-0026.
Wolfram, Walt and Donna Christian. 1976. Appalachian Speech.
Arlington: Center for Applied Linguistics.


73
in the role of addressee" (1969:74). Klammer defines monologue more
loosely as "one uninterrupted speech within a dialogue context"
(1971:38).
The monologues we encounter in our corpus are usually isomorphic
with turns in conversation, but not always. In some instances it is
obvious that monologues continue across an exchange in turns. That is,
we cannot say that a speaker needs to reintroduce, at each turn in the
conversation, all the antecedents from an earlier turn.
Good reasons for studying the structure of monologues can be
found. Longacre views monologue as a "special lopsided development of
conversation" (Longacre 1976:165). But he stresses its "universal cul
tural importance" (p. 197). His typology of discourse genres (Long
acre 1968, 1976, Longacre and Levinsohn 1978), which he believes to
have universal applications, is actually a typology of monologues (nar
rative, expository, procedural, and hortatory). What Longacre consi
ders the deep structure or universal features of expository monologues
is discussed later in this chapter.
There is another good reason for studying the structure of mono
logues. The expository monologue type of discourse on which this dis
sertation is based has all the features of what Hymes (1971) calls a
"speech act": setting and scene, participants, ends, act sequence, key,
instrumentalities, norm, and genre. Viewing monologue as a unit, more
over, allows us to examine thematizing devices, deictics, and other de
vices in this dissertation. This is because these devices help orga
nize monologue conversational turns into one or more self-contained


83
for instance, seems more likely to be used when a speaker wishes to be
emphatic, to show a contrast, or to make a particularly abrupt point.
Although the primary function of left dislocations is to introduce new
themes, our corpus has thirty five instances of left dislocation invol
ving given information, when a speaker wishes to reemphasize a piece of
information emphatically. In nearly ten percent of our left disloca
tions (45 out of 606) we find the dislocated NP immediately preceded by
"but" or "however," showing that the speaker contrasts those NPs with
earlier ones. In seven instances, we find that a left dislocation is
immediately followed by an identifying sentence, as in
1) Threshing floors, that's what they had. (Ill2-27)
This shows that a speaker wishes to make an especailly abrupt point
about a theme.
In the remaining chapters of this dissertation a strong case is
made that we cannot understand how expository discourse is organized
without considering the supersentential or discourse context of cer
tain cohesive devices. If this is true, it seems to imply something
about how a speaker communicates to a hearer. Cohesive devices, the
formal signals of the language in which a discourse is expressed,
seem to aid a speaker in communicating the hierarchical structure of
a discourse linearly. This is another way of saying that such devices
communicate the outline (implicitly, of course) of an ongoing discourse
to a hearer.


102
25) Well, this house I moved in, _U's an old house. (IV2-5)
26) You know, a man that uh, that had been raised up in a, an envi
ronment, you know, where he uh, where they had no real val
ues, you know, where he didn't really know his parents, his,
and that was offered no real guidance or whatever, you know,
uh he's more apt to, to fall back on the things that. .
(112-17)
27) And Frances and I, the one that lives in Florida is ill, we
were swinging and singing songs. (IV1-8)
In addition to 26-7, five further examples of left dislocation
working supersententially are found in the data, further evidence that we
cannot understand left dislocation as a sentential process or operation.
Some linguists have offered a sentential analysis, however, and we
briefly examine their claims. Ross posits an optional rule for left
dislocation (1967:232ff.). His formulation is as follows:
X NP Y
1 2 3
2# [1 2 3]
+Pro
Optional
He indicates that, since left dislocation can apply to NPs indetermi
nately far from the beginning of a sentence, the "distance that the dis
located NP has traveled. .suggests that the statement of the rule
must make crucial use of a variable" (p. 234). This variable is repre
sented by "X" in his rule, Ross adheres to the view that left disloca
tion is a transformation operating not only within the sentence but
also within the main clause: "Left Dislocation must be restricted to
main clauses" (p. 237).
Postal takes the sentential point of view as well (1971:135-7).
He claims that left dislocation is not bound by cross-over constraints


5
and a majority of its citizens are supported by employment in facto
ries which located in the surrounding area after the Second World War
because of the labor surplus and the low taxes. The factories re
tarded the emigration so disastrously prevalent throughout East Ten
nessee as well as other areas of Southern Appalachia in the past two
generations and attracted workers from the foothill and mountain areas
farther east, so that White Pine's population has continued to grow
steadily since the war (from 1,035 in 1950 to 1,530 in 1970 to 1,622
in 1975 tounofficially1 ,830 in 1977). It remains, however, a
cozy, closeknit (many would say "clannish") community despite the in
flux of new residents in recent years. Never has it needed more than
one stoplight, and its mayor still dutifully accepts his annual one
hundred fifty dollar check for his thirty hour a week job.
In this study, "White Pine" designates the White Pine community,
consisting of the town and adjacent neighborhoods and having a popula
tion of approximately 3,000 persons. It is the White Pine community
wherein the forty interviews for this study were conducted.
White Pine (refer to the map on page 6) lies at the extreme
northeastern corner of Jefferson County, forty miles east northeast
of Knoxville and within view of the Great Smoky Mountains. Seven
miles north on Highway 25E is Morristown (population 30,000+) and
five miles south is Cocke County (infamous for its rumrunning in the
20's and sometimes called the "Moonshining Capital of the World" dur
ing that era). East of White Pine are its sister and erstwhile rival,
but incorporated, community Leadvale and the Nolichucky River; south
and west is Douglas Lake, a Tennessee Valley Authority reservoir.


18
adheres to fundamentalist beliefs in the Bible and because of the role
of its clergy as both "preachers of the Word" and as community spokes
men commanding political and social influence. The community of White
Pine is typical of the region. Not only are the churches by far the
most prominent institutions in the community, but also religious values
permeate so thoroughly every aspect of its life that it is simply quite
impossible to deal meaningfully with most aspects of the community's
existence without taking them into consideration.
Churches have long been numerous in Appalachia. A 1935 report of
the United States Department of Agriculture reported there were in 1926
as many as 3.25 churches for each 1,000 inhabitants. It might be
thought that such a high ratio would be accounted for by the fact that
few people could motor to church fifty years ago. But the present-day
ratio in White Pine is the same: ten churches for its approximately
3,000 inhabitants (3.3 per 1,000 people). Church membership, if not
necessarily participation in church activities, is almost required for
a White Pine citizen. Perhaps as much as eighty percent of the popula
tion in the community are church members.
The two largest churches in White Pine, one a Southern Baptist and
the other a Missionary Baptist, have approximately one thousand members
between them, and the other eight are by no means all small congrega
tions. Besides five Baptist churches (of several different affilia
tions), there are two Methodist, two Pentecostal, and one Presbyterian
church. There is not a single non-Protestant church in the entire
county.


Sentence-Initial Conjunction(s) Frequency of Use
and + then 8
but 5
and 6
of course 3
well 3
so + then 1
like 1
except 1
28
Figure 3.2 Conjunctions Signaling Change in Theme


41
Before returning to the explanation of how an ethnographic sum
mary of the White Pine community was prepared, let us look more
closely at Malinowski's view of context and explore an instance from
White Pine where we must know the "context of situation" to fully un
derstand certain linguistic behavior.
As a result of his fieldwork experience, Malinowski formulated the
two concepts "context of culture" and "context of situation," the lat
ter mentioned above. Both have to do with how meaning is created by
language. A helpful redefinition of these concepts for linguistics has
been provided by J. R. Firth (1957b), one of Malinowski's students, to
express more satisfactorily how language means. According to Firth,
The context of culture is the environment for the total set of
[the options in behavior that are available to the individual
in his existence as social man], while the context of situation
is the environment of any selection that is made from within
them. (Hal 1iday 1974:49)
For Firth, then, the two types of context formulated by Malinowski pre
sent the distinction between what is potential and what is actual.
Again according to Halliday, "the context of culture defines the poten
tial, the range of possibilities that are open. The actual choice among
these possibilities takes place within a given context of situation"
(1974:49) Firths version consequently emphasizes the value of context
of situation for linguistics while at the same time implying that con
text of culture has little value for the linguist.
The great usefulness of the concept of context of situation, in
Firth's definition, for linguists lies in its stress on the typical


146
Neither analysis tells us much about peak sentences. Having an
unusual structure and being employed by speakers with relative
infrequency, they would seem to have special properties that we can
determine from considering their discourse context. We will consider
this context by asking two questions: Are the introducing deictics
of peak sentences anaphoric or cataphoric? How are peak sentences
distributed in discourse blocks?
As was indicated at the beginning of this chapter, deictics
frequently do not clearly refer to another portion of discourse. In
many cases they merely point. If we look closely at peak sentences,
it is usually quite difficult to determine in which direction the
initial deictics point. If we examine the entire exposition in
which 42 above appears, is the deictic "that" anaphoric or cata
phoric?
48) Fieldworker: In White Pine what do you do for a good time?
Informant: Without leaving White Pine? Watch TV. Maybe,
maybe walk back, you know, like one time, it's, you know,
peaceful back there in the barn, and like uh, like I said,
if you're standing back there by yourself in the barn, like
one time when I went back yonder, and you, I had, you know,
it's, it's not really a seat really. It's part of the manger
that comes out, and it's a hard position to get in cause you
got to sit with your legs way up in the air, but Kathy was
up here watching TV or doing something and uh, talking to
Momma, I guess, and, and I went back there that day. You
know, it's just a good thinking place or it's a good place
to relax. I mean you just set there and the cows, all the,
if you ever set close to a bunch of cows chewing, well, it's
just, you know, relaxes you. And then your dogs are all
laying in there, you know, a couple of cats, and the rabbits
are right outside. They're all peeking in at you, and that,
that's a good way, that's the best way I ever found to think
is just to walk back up through the woods or something like


62
Discourse Studies and the Approach of This Study
The decision to study the discourse level of oral expositions orig
inally derived from this writer's growing conviction that the construc
tion known as left dislocation, an often-used construction in speech,
could only be understood if its function were viewed as supersentential,
on the discourse level (Montgomery 1977a, 1978). If such a frequent but
little studied construction seemed to tell us something about how speak
ers organize their linguistic behavior, it seemed likely that other fea
tures and constructions of oral discourse should also be studied in
their discourse contexts. There is much to recommend the view that
oral expositions are preliminary to written ones. At least, understand
ing how speakers negotiate and structure ongoing expositions might im
prove our awareness of the communication system that people bring to the
writing process. Since linguists have for two generations made the case
that speech is primary and writing derivative, it seems crucial that a
study should begin with the structure of oral expositions rather than
written ones.
It is no longer necessary to justify the value of doing discourse
analysis, which may be broadly defined as the study of any segment of
verbal behavior larger than the sentence. Grimes (1975), the watershed
work in the field, has shown that discourse analysis is indeed possible
and that it rightly lies within the domain of linguistics (rather than
rhetoric or another field exclusively). Grimes' work is such a vast


126
indefinite. It may be specific, generic, or nonspecific indefinite,
but in only one instance, given below, out of 320 (in the set of 200
expositions) is it definite:
57) Because there's the Seventh Day Advents [sic], they have
their services on . (IV2-16)
Left dislocation and existential sentences thus complement one another.
The former usually does not introduce indefinites; the latter very
rarely does otherwise. Both can introduce specific indefinites. In
the following example, for instance, we see that the informant begins
the sentence with a definite NP. But when she realizes that she will
introduce a specific indefinite one instead, she does so with an exis
tential and not through left dislocation:
58) And now then the Smith, there's a handicapped boy that
lives in that building now, Kenneth Smith. (II12-2)
This xample shows a switch in strategy by the speaker. In the
data we find a number of sentences which begin with an existential
and then appear to be left dislocations, such as
59) And there was an old gentleman named Arwood, he lived down
the road just a little ways down there. . (X2-8)
which also involve specific indefinites and look like changes in
speaker strategy. They are well-formed constructions, however, and
we will look at them in detail at the end of this section. They
represent a fusion of left dislocation and existential patterns.


28
consider the dialect of the Southern Appalachian area distinct and
identifiable. One reasonable approach to the issue of whether the
term "Appalachian English" should be used is that of Wolfram, who
adopts the term and applies it to a "unique combination of linguistic
characteristics rather than . structures as such" (Wolfram 1977:
97). For him, Appalachian English refers to a common denominator of
non-unique features which ar still subject to variation for any given
district within Appalachia or for subgroups within such a district.
Since this study is designated as one on "Appalachian" speech,
this must now be justified and explained. This study does not inves
tigate linguistic features dealt with systematically in earlier studies
on the speech of the Southern Appalachian region or of the speech of
any other variety of American English. It thus lacks the comparability
of evidence required to distinguish the speech of White Pine as "Appa
lachian" solely on the basis of linguistic criteria, although nearly
all the features-phonological, lexical, and grammaticalusually
termed "Appalachian," as by Brewer and Brandes (1977), are found in
the speech of White Pine inhabitants. Nor does it adopt the design
ation solely on the basis of White Pine's geographical location. If
it did, "East Tennessee" would have been used. Rather, the speech of
the inhabitants of White Pine is called "Appalachian" because the
community is, ethnographically speaking, an Appalachian community, on
the basis of several of those salient characteristics which have long
been associated with Southern Appalachiaindividualism and self-
reliance, traditionalism, fatalism, and fundamentalist religion
(Ford 1962:1 Iff.).


113
would mix a half a teaspoon of sulfur and uh something like
a tablespoon of molasses, mix that sulfur in that. She said
there's nothing as good, and she was the, she was the doctor
of our family, I reckon, for she was always prescribing some
thing, you know. If we'd get anything, what to do. (uh-huh)
And uh, but now that sulfur, I mean, she said that purified
the blood, and so that, I didn't, 1 still don't understand
that. But I took a lot of it, I'll tell you. It didn't hurt
me. If it didn't help, it didn't hurt. (II12-36)
''Sulfur" subordinates and replaces "my aunt" as the thematic piece of
information towards the end of this passage.
As mentioned above, thirty-five of the 606 left dislocations repre
sent given pieces of information. When these thiry-five pieces of in
formation are dislocated, they are neither mentioned in the discourse
for the first time nor reintroduced after the intervention of a differ
ent thematic piece. They are not mentioned to bring something into the
forefront of a hearer's consciousness; they are mentioned to keep that
piece of information in the forefront. In other words, these disloca
tions represent especially emphasized pieces. In the following excerpt
"Mrs. Seabrook" is left dislocated the second time mentioned (obviously
for emphasis) rather than when she is introduced:
49) . .they had their own Black children went to this school (uh-
huh) and their Black teacher. And uh I'll tell you a teacher
that used to teach over there was Mrs. Seabrook. She died
two years ago, up here in this school. Oh Mrs. Seabrook, we
all liked her very much. She was, that was, that was a sit
uation that I thought that worked just fine when it came to
integration and all like that. (11-36)
"Mrs. Seabrook" is definitely not new information when it is dislocated.
Given NPs, when dislocated, are often accompanied by interjections such
as "oh," "why," "boy," and others.


135
9)And you know, that makes you feel awful good for a man to be
that nice to you, you know. (X2-16)
10) We had to meet a bus, so that was sort of inconvenient to get
to school, you know, . "! OII4-2)
There are 34 instances of extraposition with "that" in the data, about
half as many as with "it." We will examine more closely this use of a
deictic later in the chapter. There are countless additional instances
in the data of the employment of "that,"as in 11-13 below:
11) Take it up with your immediate supervisor or take it up with
your department head. That's the same way with complaints
within the city. (113-20)
12) That's been ten, eleven, eleven to twelve years ago now that
that was done, . (I13-9)
13) It took about, that took about three days to get the power
back, . (1X1-14)
In 12, "that" is used rather than "it" in a cleft sentence. In 13
we see the speaker changes strategies to use the stronger deictic
and more emphatically makes the point.
The deictic is more forceful than a pronoun, but the use of "that"
imparts greater tension because of the greater stress which normally
accompanies it. In examples 7-13 above, "that" carries more stress
than "it" would in the same context. A different and perhaps more
accurate way to look at the variation between the two is to say that
the speaker decides whether to give tension to a context. If tension
is given, then "that" is used. If it is not, then it" is used. It
seems probable that "it" under normal circumstances cannot carry stress.
In the next three sections we examine the form and nature of left
dislocations with "that," extrapositions with "that," and identifying


180
Vance, Rupert B. 1935. Human Geography of the South. Chapel Hill:
University of North Carolina.
Vance, Rupert B. 1962. The Region: A New Survey. In The Southern
Appalachian Region: A Survey, edited by Thomas R. Ford. Lexing
ton: University of Kentucky. 1-8.
Van Dijk, Teun A. 1972. Some Aspects of Text Grammars. The Hague:
Nouton.
Van Dijk, Teun A. 1977. Text and Context: Explorations in the Seman
tics and Pragmatics of Discourse. London: Longman.
Van Dijk, Teun A. and Janos S. Petdfi, eds. 1977. Grammars and
Descriptions: Studies in Text Theory and Text Analysis. New
York: Walter de Gruyter.
Walker, Ralph S. 1939. A Mountaineer Looks at His Own Speech.
Tennessee Folklore Society Bulletin. 5. 1-13.
Watkins, Floyd. 1949. The Southern Mountaineers; Archaic English.
Georgia Review. 3. 219-25.
Weatherford, W. D. and Earl J. C. Brewer. 1962. Life and Religion
in Southern Appalachia. New York: Friendship.
Weller, Jack. 1965. Yesterday's People. Lexington: University
of Kentucky.
West, John Foster. 1966. Dialect of the Southern Mountains.
North Carolina Folklore. 14. 31-4.
Widdowson, Henry G. 1977. Approaches to Discourse. In Grundbe-
griffe und Hauptstrdmungen der Linguistik. Hamburg: Hoffman
und Campe. 236-60.
Williams, Cratis D. 1961. Vowels and Diphthongs in Mountain
Speech. Mountain Life and Work. 37. 8-11.
Williams, Joseph Marek. 1966. Some Grammatical Characteristics
of Continuous Discourse. University of Wisconsin dissertation.
Wise, C. M. 1957. Applied Phonetics. Englewood Cliffs, New Jersey:
Prentice-Hall.
Wise, Mary Ruth. 1968. Identification of Participants in Discourse:
A Study of Aspects of Forms and Meaning in Nomatsiguenga.
Ann Arbor: University of Michigan dissertation.


BIOGRAPHICAL SKETCH
Michael Bryant Montgomery was born on May 15, 1950, in Knox
ville, Tennessee, and attended public school in that city, gradua
ting from Holston High as salutatorian of his class in 1968. There
after he entered Maryville College, Maryville, Tennessee, and gradu
ated cum laude with a Bachelor of Arts in English from Maryville in
1973. While pursuing a master of Arts in English at the University
of Tennessee at Knoxville, he determined to study linguistics. After
receiving his master's in 1975, he enrolled in the Program in Linguis
tics at the University of Florida to study for the doctorate. During
the 1978-9 academic year he was Instructor in English at the Univer
sity of Arkansas at Little Rock and coordinator of the Department of
English's Writer's Hotline Service. Beginning in the fall of 1979 he
will be Assistant Professor of English at Memphis State University in
Memphis, Tennessee. His favorite hobby is linguistics.
182


115
dislocated animate NPs to represent agents. On the basis of the data,
however, we find that left dislocated animate nouns are likely to be
agents only about half the time (215 of the 402 animate NPs, or 53.5
percent).
The large number of dislocated NPs which are inanimates or non
agents in exposition should not be surprising. Longacre claims that
in its deep structure expository discourse is subject matter-oriented
and not agent-, patient-, or addressee-oriented (1976:200). In a cor
pus of narratives, on the other hand, we would probably find a much
higher percentage of agents among dislocated NPs. Longacre also
claims that the subject matter orientation of exposition is generally
realized by equative and descriptive clauses (p. 202). If we examine
the 402 instances of dislocated animate nouns, we find that 94 of them
(23.4 of the animate ones, that is) are from such clauses.
Probably the large number of dislocated NPs which are animates
and agents in our expositions should not be surprising either. It is
much too simplistic to assume that any given genre of discourse con
sists primarily of just one or two types of clause. Also, we must
remember that discourses are frequently embedded into one another. In
our corpus' expositions, we frequently find embedded narrative and
other episodes which are person- rather than subject matter- oriented.
So far we have examined the kinds of NPs dislocated, their posi
tions, and their functions only with reference to the total number of
left dislocations identified in the corpus 606. An examination with


15
The Economic Life of White Pine
White Pine has made significant economic strides in the past two
decades and has undergone a thorough economic transformation in the
past generation, but it remains in many important ways undeveloped.
The percentage of its population below the poverty line (19.4) is well
above the national average (approximately 12). Host of its citizens
must look beyond the city limits for their livelihoods.
Fortunately the surrounding ten mile radius of White Pine has been
able to provide considerable employment opportunities for the last
three decades. This employment has not only stemmed but also largely
reversed the serious problem of emigration of the labor force which
plagued East Tennessee and White Pine in the 30's, 40's, and 50s, al
though it was not quite as severe as elsewhere in Appalachia, espe
cially in the coal mining regions of Kentucky and West Virginia. This
reversal has largely been accomplished through the benefits accruing
from the development of TVA. In the 1940's, Jefferson County's emigra
tion rate surpassed ten percent of its eligible adult work force, and
in the 1950's it was almost as high. But in the 1960's, immigrants
began to outnumber emigrants and the migration patterns underwent a
dramatic turnaround. In the 1970's, the county and especially White
Pine have seen a significant infusion of new residents and a decrease
in emigration. Employment prospects for incoming members of the job
market, although still limited, are better than twenty or thirty years
ago.
The dislocations brought about by the emigration over the years
have left their marks, though. White Pine has lost much of its Black


UNIVERSITY OF Fl-ORIDA
ML 7106


137
In a number of further instances, the use of "it" to copy the dis
located NP seems completely unacceptable, as in 18 and 19.
18) Classical music, we played a little of that when I . .
(XI-12)
19) ... a commercial blended whiskey or a Kentucky bourbon,
something like that would on, on the whole would be either
eighty to . (13-25)
In most instances of left dislocations with deictics, as in
20 and 21 below, the speaker has obviously chosen the deictic for
its more forceful, emphatic quality.
20) And of course, the housing project, they put that down
there. (V2-11)
21) The old Atchley house, now you saw that down at Leadvale.
(V2-22)
The left dislocations with deictics in our corpus represent the use
of two cohesive devices to introduce themes. The deictic enhances
the effect of the left dislocation.
Extrapositions with Deictics
In the usual transformational treatment of a sentence involving
extraposition, a clausal subject has been moved beyond the predicate
and replaced by "it," as in
22) It is possible that this sentence looks familiar,
which is the transformed version of
23) That this sentence looks familiar is possible.
In our data we find thirty four instances of extrapositions of
clauses in which "that" is used rather than "it" as the filler
subject of the transformed sentence. There are about twice as many
"it"-extrapositions in the data. Examples of extraposition with "that"
are


38
This study is based on a representative sample of forty infor
mants in a Southern Appalachian community of approximately 3,000 per
sons. It quantifies the linguistic features it examines both in terms
of which speaker used them and in terms of possible occurrence of the
feature (whenever possible). It controls the social situation in which
its data occur (see the discussion of the ethnographic parameters and
the role this fieldworker adopted for himself later in this chapter).
This study controls the task its informants were performing by elicit
ing only expository (explanatory and descriptive) discourses, and it
controls the linguistic context of its data by considering the sentence
and especially the discourse contexts of data.
Fieldwork Procedures
The fieldwork upon which this dissertation is based was carried
out over a period of about two months, from December 1977 through Feb
ruary 1978. The fieldworker/researcher spent most of his time in the
community of White Pine during this period. This section details the
process whereby he set up, undertook, and completed his fieldwork. It
also provides the rationale for the type of interview and the ethno
graphic parameters adopted. Altogether six weeks were devoted to the
actual interviewing, one week (in the field) to preliminary fieldwork,
and a period of several weeks to preparation before he reached the
fieldwork site.
Two steps were taken before the researcher reached the White Pine
site. First, an ethnography of the White Pine and surrounding areas was


92
thematized in sentences with inverted word order (pp. 214-5). In sen
tences 3 and 4 below, for example, a direct object and a temporal ad
verb are thematized:
3) These houses my grandfather sold.
4) Tomorrow John's taking me to the theatre.
Sentences with such inversions are very infrequent in our data, but Hal-
iday's general claim about the kinds of elements which may be themes
matches closely those thematized by left dislocation shown later in
this chapter.
The present study takes a step beyond Halliday by examinina how a
theme relates to a context beyond its immediate sentence, but our view
of theme is consistent with Halliday's. Nowhere does he state that a
theme's domain is bounded by its immediate sentence, and we may reason
ably infer that he would agree with our finding that theme plays an
important role in organizing a paragraph.
In the Prague School view, theme has to do with the information
structure of a sentence. The element(s) of a sentence which are
"given" (previously mentioned in a discourse) constitute that sen
tence's theme. The theme need not appear in initial or in any other
sentence position and a sentence need not have any given information.
If it does not, it has no theme (Firbas 1966:268). Danes and Firbas,
contemporary members of the Prague School, continue this tradition.
Firbas' view of theme, however, is much less similar to the one we
take here because it is explicitly confined to the limits of a sen
tence. His view is a refinement of the widely known theme-rheme


157
between the dislocated NP and the ensuing sentence. But an audible
hesitation often precedes the NP, however, and apparently aids a
speaker in introducing a theme into discourse. We include the
hesitation "uh" in the table on the following page, which lists the
variety of introducers for the 190 left dislocations.
One introducer in the list here, "like," deserves our special
attention. Below are three examples of it accompanying left dis
located NPs:
2) Like Brooklyn, you know their lingo. (V3-16)
3) And like Lisa Williams who lives down the street here, she
got me interested in it, . (V113-11)
4) Like my cousins, they had to go to South Carolina because
of the job situation. (XI-8)
Seventeen of the 190 NPs are preceded by "like" or "like" in combin
ation with another conjunction. Of the total of 606 left disloca
tions in the entire corpus, 29 are accompanied by "like." "Like"
is apparently restricted primarily to younger speakers, and it dif
fers in function from other introducers. Rather than accompanying
left dislocated NPs which initiate paragraphs, it creates units
which we will call "sub-paragraphs."
Although only eight of this study's informants were twenty-five
years old or younger, these eight used twenty-one of the twenty-nine
instances of "like." That is, twenty percent of the informants
account for 72.4 percent of the occurences. "Like" not only accom
panies left dislocation, however. In general, it is a conjunction
which introduces an example to elaborate a point a speaker has made.


155
beginning sentences. Conjunctions and other introducers are the surface
structure cohesive device which constitutes what a speaker uses to
stitch the parts of a message together into a discourse.
Although they do not by themselves introduce themes, conjunc
tions and introducers frequently begin paragraphs and usually
accompany left dislocated NPs which introduce paragraph themes. It
is shown in chapter three above that a speaker often uses a conjunc
tion or introducer to signal a hearer that one theme is being super
seded by another. In this chapter we examine the conjunctions and
introducers which accompany the 190 dislocated NPs in the 200
selected expositions used in chapter three. In particular we look
at the introducer "like" and suggest that it is a sub-paragraphing
device. For the most part our analysis is suggestive rather than
definitive. Because of this study's limitations, the functions
and distributions of conjunctions are not yet entirely sorted out.
It has been widely pointed out in the literature that para
graphs and discourses characteristically begin with certain intro
ducers in English and in other languages as well. Regarding oral
paragraphs Van Dijk says "A new paragraph thus indicates (sub-)topic
change. In spoken language we have pauses, intonation and specific
particles like 'now1, 'well', etc. to indicate such paragraphs.
Other languages have specific morphemes in order to mark beginnings
and endings of stretches of discourse. ." (1977:152-3). Longacre
states that "many languages have particles that indicate either the
beginning or the end of a paragraph" (Forthcoming:!17). He goes on
to cite examples from Huichol, Shipibo, and Capanahua.


107
operate on certain NPs, those indefinites which may be characterized
as [-specific], even though one major study (Gundel 1974) argues other
wise, in fact. Gundel argues for the acceptability of sentences with
nonspecific indefinites, as 35 and 36:
35) A misprint, the proofreader didn't see one.
36) A phone call, he didn't charge me for one.
but these judgments are entirely her own. Her analysis of the con
straints on left dislocation, based wholly on data not taken from actual
speech, contends erroneously that noun phrases which contain quanti-
fier-like determiners (with the exception of ONLY and EVEN) cannot be
dislocated (PP. 62-3). Many of the instances from our corpus, including
37 and 33 below, show this cannot be true:
37) Most of my friends, the people with whom I was raised, we all
went to church. (IVI-8)
38) Well, each, each section, I think they call vt a pod. (VI-6)
A non-data-based analysis thus can only guess at the constraints on left
dislocation. Gundel also goes for a sentence-domain treatment.
Obviously it is only a large corpus such as ours which can show
what kinds of NPs can be dislocated. There is no reason to believe that
nonspecific indefinites are thematized by left dislocation. It is shown
below that they are introduced by existentials when thematized.
Left dislocation, as we define it, involves a pro-word copy of the
NP which is thematized. There is evidence from our corpus that a
fronted [+definite] thematic NP need not have such a copy and that


158
Introduced s)
Frequency of Occurrence
0
39
and
31
and + ih
6
and + then
6
and + then + uh
5
and + of course
3
and + even
1
and + even + uh
1
and + why
1
and + wel1
1
and + uh + so
1
but
16
but + uh
4
but + now
2
but + as far as
2
well
13
well + uh
1
well + of course
1
well + now + then
1
like
12
and + like
3
but + like
1
now + like
1
and + then + of course
1
of course
4
why
3
now
3
as far as
3
because
2
uh + because
1
of course + uh
1
SO
1
so + uh
1
uh
7
oh
1
naturally
1
where
1
that + uh
1
you know
3
okay
1
yeah
1
most of the time, though
1
190
Figure 5.1 Introducers Accompanying Left Dislocated NPs


11
residential community. Until a generation ago, most people in White
Pine grew up on the farm, but with the coming of industry came rising
economic expectations and consequently a revolution of rising demands
for private residential housing. These demands have partially been met
by the parceling up of farmland into lots and a mobile home park, but
as another generation has grown up, it has become acutely obvious that
the town especially lacks social activities.
In White Pine the principal social institutions are the churches
(ten in the community), and more than likely any social get-together or
meeting will be church-sponsored. Although there are several organized
civic groups (Ruritans, Lions, Beta Sigma Phi) in town, there is little
regular social activity. The occasional special event (church revival,
political candidate speech) may attract a group of citizens, and the
annual Pine Festival is an event that meets some of the social needs of
the town.
The Pine Festival, sponsored by the local Beta Sigma Phi and begun
in 1967, is a one-day collection of shows, contests, and displays held
on July 4 each year. The festival has become the focal event of the
community's life because it is the one occasion when a majority of its
people take part. The festival integrates the community and gives it
a greater sense of self-identity than it has ever had because it is hea
vily patronized by people from surrounding counties and communities.
Although there is tremendous socioeconomic disparity between the
very rich and the very poor in the community, one finds little overt
residential segregation by economic status. As the thirty-nine-year-


51
16-19
20-39
40-64
65 and over
Total
Male
3
6
5
4
18
Female
3
5
8
6
22
As the sample was taken, care was exercised not to have too many in
formants from a given age group or from either sex. The fieldworker's
overriding concern, as he spent his time in the White Pine community
and had contact with all sectors of the population, was to interview as
wide a variety of its residents as possible. It was not his primary
goal to investigate social or subgroup differences in discourse organi
zation among his informants.
The breakdown of the forty informants according to their levels of
education
is as
2
follows:
8th
Grade
Some High
High School
Some College
College
or
Less
School
Graduate
Work
Graduate
Female
1
7
5
5
4
Male
3
5
1
6
3
According to the 1970 census, only 6.3 percent of White Pine's popu
lation depended on agriculture for its livelihood, while 74.1 percent
depended on manufacturing, trade, service, and finance industries.
Approximately 19 percent of the population was employed in construc
tion and all other areas. Following is a breakdown of the White Pine
sample according to their employment by type of industry. Three who
were unemployed are ignored.


103
and that it is derived from reduced introductory phrases such as "As
for. His analysis of the restrictions on left dislocation is in-
correct in the restriction to only one NP per sentence (cf. 23 above)
and to only main clauses.
An accurate account of left dislocation thus confounds those wish
ing to give it a sentential analysis. Any data-based account (and Ross
and Postal's are not) discovers that we must consider left dislocation
in its discourse context, that we must take more than the immediate sen
tence into consideration. This is especially true since the amount of
appositional information a speaker chooses to provide the hearer is in
determinate. Example 28 below shows the apposition may consist of sev
eral sentences; In it "there" refers to "at Rankin side. ."
28) Now over at the coal chutes, now where Earl Atchley was
raised, I mean over at Rankin side, there was a big coal
chute over there that furnished the coal for the steam en
gines, you know, and uh, well, we had the two stores here
before the lake come that furnished people, I mean, all of
the rural area. That's where we depended on our livelihood,
on such things as we didn't grow on the farm. Why, we went
to there, and uh. . (Ill2-6)
If the dislocated NP represents the thematic piece of information
of a paragraph, there should be an upper limit or threshold to the
amount of information constituting the apposition (in order for a
speaker to assume that the theme is still in the hearer's conscious
ness). This is a matter, however, of studying the limits of short-term
memory and of the cognitive basis of thematization and paragraph struc
ture. The relevant issues involved are reviewed by Van Dijk
(1977:155ff.).


31
devotes an entire chapter (1921:90-122) to analyzing their individualism
and views their self-reliance as a result of their extreme isolation.
Horace Kephart, who lived in western North Carolina the first two dec
ades of this century, characterizes the people as having "fiery indi
vidualism" and as being "non-social" (1922:382-3) and goes on to say
Except as kinsmen or partisans they cannot pull together. Speak
to them of community of interests, try to show them the advan
tages of cooperation, and you might as well be proffering advice
to the North Star. (p'. 383)
More contemporary observers, Weller (1965:29ff.) in West Virginia
and Hicks (1976:39) in western North Carolina, document the same trait
among the sections of the region they observed.
When one looks at White Pine, such comments have a remarkable
applicability. The people are basically non-social, sticking to small
groups, not often unified in acting except in opposition to a proposed
local change, on a special occasion such as Pine Festival Day, or with
in the small sphere of the kin group. Regarding their inability to
work together, a twenty-nine-year-old male resident of White Pine says
of his fellow citizens
They, they want to know, but they don't want to come out of their
shell long enough to get enough of it to expand. They fear ex
pansion. This town definitely doesn't, it, one day they'll be
shocked into the fact that Morristown will try to annex them,
and there will be a total bar-, that will be the common bond
that brings them together. They will be, that will be the sup
portive element would be to stand against some cause.
This does not contradict the statement above (p. 9) about the cohesive
ness of the community; it means the cohesiveness is passive and residual.


7
Jefferson County (1960 population 21,493, 1970 population 24,940)
lies in the ridge and valley section of the Great Valley of East Ten
nessee between the Cumberland Plateau and the Blue Ridge. Today it is
sandwiched between the TVA reservoirs, Cherokee Lake (the dammed Hol-
ston River) and Douglas Lake (the dammed French Broad River) built in
the 1930's. The county is one of the earliest settled areas of what
was to become the state of Tennessee. The Holston, which formed the
county's northern border, brought settlers in from the northeast,
from upper East Tennessee and from Virginia (through the Shenandoah
Valley), while the French Broad, which largely determined Jefferson
County's southern extent and which flows west from Asheville after
flowing north, routed newcomers from the east into the area. Many of
these had received dispensations of land for Revolutionary services.
The county's first white residents arrived in 1783, thirteen years
before Tennessee became a state. The county seat, Dandridge, was
incorporated in 1793.
Settlement in the White Pine vicinity goes back nearly two hun
dred years, as shown by the fact that two of its present-day churches,
Westminister Presbyterian and Beth Car Methodist, were first organized
in 1787. Although the town of White Pine was not incorporated and
named until 1893, it achieved the status of a community and the char
acter of a town (and was called "Dandridge Crossing") when a railroad
stop and a post office were established there shortly after the Civil War.
A number of today's White Pine residents can trace back their ancestries


34
Appalachian speech corresponding to usages in Spenser, Chaucer, and
especially Shakespeare. This creates the impression that the language
of the mountains is little more than a carryover from a bygone era and
another country. This is especially true when a commentator devotes
little attention to anything but such analogues.
One might go so far as to speak of a "typical" treatment of Appala-
chain speech in the literature. This is represented by Berrey (1940),
Bray (1950), Carpenter (1933), Combs (1916, 1931), Watkins (1949),
West (1966), and many others. The typical treatment contains the
following:
1) A brief description of mountain culture and of the mountaineer
himself, especially mentioning the most salient traits of his
personality and how he has adopted himself to his mountain
environment.
2) Discussion of various peculiarities of pronunciation, syntax,
vocabulary, usage, onomastics, and similes and other figures
of speech and expressions, with an effort to compare such with
Old, Middle and Elizabethan English whenever possible.
3) (Optional) A few remarks on how mountain speech is/has been
passing away, despite its expressiveness and color.
Coleman (1936) is a perfect example of this typical treatment. Her
first chapter points out the ancestry of the mountain people, shows how
they developed self-reliance in their isolated habitat, emphasizes the
fierce independence of their character, and shows the low position of
women in mountain society. In her second chapter, she covers a wide


169
that, (uh-huh) We have to answer a lot of calls down there and a
lot of neglected children and so forth and so on. (uh-huh) That's
the biggest problem, you know, with it.
F: What kind of town do you think most people around here want White
Pine to become?
I: I think a majority of the people want it to stay like it is. Uh
[Visiting ex-policeman filling out application: Nice, quiet, well-
run town.] Yeah, they, they don't like to see, especially your old
people. They don't like to see changes that would be major changes,
like uh I don't think that they'd want to see a housing project of
three or four times the size of that one go in, and I'm convinced
that they don't want to see too much industry, (uh-huh) I think the
statement was made, I won't say who made it, but a statement was
made uh some time back, 'sa year or so ago. We had a company, a
drug company, Davis Manufacturing, I think, that wanted to build a
sizable plant here in White Pine. I think one of the city officials
said "well, White Pine's a nice, quiet little town. Let's just
leave it like that." and the company didn't come in. So I think
that's the way most people feel about it. (hum) Course you, I
guess you'd have to look at the people that make the decisions on
the city, such as your aldermen and the mayor, (uh-huh) Uh they're
the people that says whether or not anything's going to be here or
not, whether the citizens like it or not, and I think most of them,
well, not, maybe not most of 'em, but the majority of 'em feel, uh
minority of 'em feel thataway, one or two of 'em anyway, (uh-huh)
They's just as soon it stay like it is or go back to like it was
years ago. I honestly feel that way.


136
sentences. Then the fourth structure mentioned earlier, peak sentences,
is examined in light of Longacre's ideas on the peak of expository
discourses/paragraphs. The form and especially the distribution of
peak sentences in expositions is studied. At the end of the chapter is
a short discussion of minor summary devices.
Left Dislocations with Deictics
As noted above, we find seventy four instances in our data of
deictics copying left dislocated NPs rather than pronouns. This rep
resents 12.2 percent of the total left dislocations but, more impor
tantly, approximately one-third of all potential cases. That is,
about 200 of the dislocated NPs in the data are inanimate. A close
look at the data reveals left dislocations for which deictics are the
only, or at least the preferred, copy. These seem to indicate that a
speaker does not always have the option of using either the stressed
deictic or the unstressed pronoun.
One such kind is the left dislocation of an NP from an identi
fying sentence (7 instances of this), as in 14 and 15:
14) But now in that enclosure, that was where the railroad
station was. (VI3-12)
15) Threshing floors, that's what they had. (II13-27)
When a subordinate clause is dislocated, as in 16 and 17, the deictic
seems to be the preferred copy.
16) To go to Morristown to buy a pair of socks, that was kind of
aggravating. (X3-12)
17) But to turn high, be real warm and then turn cold, that
hurts. (1X1-19)


APPENDIX A
LIST OF INFORMANTS
mfi
Informant No.
Sex and Age of Informant
Interview Code
1
Female, age 63
11, 12
2
Male, age 29
13, 14
3
Male, age 25
112
4
Male, age 30
mi
5
Female, age 70
1112, 1113
6
Female, age 72
IV3
7
Female, age 62
IV1
8
Male, age 30
113, 114
9
Female, age 54
1114
10
Male, age 60
IV2
11
Male, age 74
IV4
12
Female, age 59
VII
13
Female, age 82
VI3
14
Female, age 39
VI
15
Male, age 53
V3
16
Female, age 44
V 4
17
Male, age 43
VI2(l-3), VI4
18
Male, age 73
VI2(4-30)
19
Female, age 17
VIII
20
Female, age 16
VI13
21
Female, age 43
VI12
22
Male, age 74
VI14
23
Male, age 17
VIII1
24
Female, age 85
VI113
25
Male, age 17
VI112
26
Female, age 39
VI114
27
Male, age 26
V2
28
Female, age 17
XI
29
Female, age 50
X3
30
Male, age 48
X4
31
Male, age 76
X2
32
Female, age 87
1X1
33
Male, age 34
1X3
34
Male, age 60
1X2
35
Female, age 76
1X4
36
Male, age 18
XI2(1-12)
37
Female, age 30
XIII
38
Female, age 28
XI2(13-8)
39
Female, age 53
XII
40
Female, age 33
XII
Informants are listed in the order interviewed. Code numbers refer to
pages of typed transcripts of interviews.
163


4
(1966, 1972) has controlled the first two by adopting random sampling
techniques and quantitative analysis of his data. He has controlled
the third by deliberately eliciting several different speech styles
(four in his New York City investigations). And he has taken into ac
count the sentence context of data by devising the variable rule (Labov
1969). The present study, however, controls both the task context and
the discourse position context. Recent studies in the ethnography of
communication approach take the task and discourse position into ac
count but tend to ignore the others.
This study, in investigating a selection of surface structure co
hesive devices, found that they were to a significant extent used by
speakers to communicate thematically unified blocks of information to
hearers and that they reflected the grammatical structure of para
graphs. This confirms the hypothesis that oral expository speech
(Southern Appalachian in particular) has describable paragraph and
discourse structure.
The White Pine Setting
The small East Tennessee hill town of White Pine (White P'ne to
its inhabitants), in which the interviews for this dissertation were
conducted, is no longer the rural village it was only a little more
than a generation ago, and no longer is farming the principal occupa
tion of its residents. Today the town of White Pine is primarily res-
idntial, a "bedroom" community (in 1970, eight times as many of its
people derived their livelihoods from manufacturing as from farming),


33
The traditional way and pace of life in the community have been con
sidered as something of a birthright. Many other things, such as po
litical and religious affiliations, are also considered immutable
birthrights. Many wish keenly to pass all this on to the next genera
tion.
Little additional commentary here (see pp. 17-20 above) is needed
to show how applicable is the fourth characteristic trait Ford studied,
fundamentalist religion, to White Pine. The community has no non-fun
damentalist churches and the great majority of its citizens are church
members. Religious fundamentalism is probably the most powerful and
pervasive force in the community.
With reference, then, to the four characteristic traits Ford claims
to be consistently typical of Southern Appalachian people, it is obvious
that White Pine should be considered a typical community in the region.
The Study of Appalachian Speech
As Wolfram notes, Mencken once pointed out that the speech of the
Southern mountains had been studied more than that of any other region
except New England (Wolfram 1977:92). However, the sum total of lit
erature on Appalachian English does not actually tell us much about how
people in the region speak. At best, most of the literature is little
more than anecdotal, dealing with a scattering of exotic words and pro
nunciations, many of which indeed are archaisms. Very few writers have
actually asserted that Shakespearean English is or was spoken in the
mountains. But many have taken extreme pains to indicate usages in Ap-


117
one-fourth of the potentially dislocatable NPs are dislocated. In
addition, the subset of animate themes were counted to see if a
higher percentage of them were dislocated than inanimate ones. If
they were, this would have indicated that left dislocation is a
character-introducing device. It was found, however, that a slightly
lower percentage (33 of 160 instances, or 20.63 percent) of the ani
mate NPs were dislocated than the inanimate ones (26 of 120, or 21.67
percent).
Although the constraint is slightly stronger with names, it is
apparent that the overriding factors accounting for left dislocation
are the functions we have discussed (contrastiveness, emphaticness,
newness especially), given that the NP is thematic. At least for our
corpus of expositions, it makes no significant difference whether the
NP is animate or not.
Relationship to Subsequent Discourse
It has been seen that left dislocation is a device which usually
introduces new information into the stream of discourse. Also it has
been shown that left dislocated NPs may initiate an anaphora chain and
represent the information around which subsequent discourse is orga
nized. But how typically do such NPs do this? Before it is concluded
that left dislocation is a thematic device which normally organizes
paragraphs, we must see whether the discourse beyond the immediate sen
tence where the left dislocation occurs is still oriented around that
NP. If left dislocation introduces themes, it should be atypical for


129
We have seen that left dislocation and existential sentences
have different ranges of functions but that both introduce new, the
matic information into discourse as a rule, albeit of different kinds.
A close look at our White Pine reveals a number of sentences which
combine left dislocation and existentialization. In such sentences
the differences between the two disappear, as in sentences 64-6 below:
64) There's not a teacher I don't think up there that I wouldn't
want my child to be in their room. (VI-4)
65) But they's some people I wouldn't want 'em to grow up in
White Pine like it is. (VIIII-23)
66) There are people in this town that I would love to step on
their face. (VIII4-24)
On balance, these sentences seem less like left dislocations than ex-
istentials. In these we do not find the characteristic juncture
which follows a left dislocated NP: a slight pause and level or only
very slightly falling intonation (in contrast to the falling tone char
acteristic of the end of a declarative sentence). But in seventeen
instances in our corpus as a whole there are sentences such as 67-9
67) I know they was some people uh, their cars got stalled between
here and there. (XI1-6)
68) And there was an old gentleman named Arwood, he lived down the
road just a little ways down there. . (X2-5)
69) Well, there was a lady lived in this community, she was an
educated somebody, a school teacher. (1X1-29)
in which an NP is existentialized and then copied by a pronoun and
which have the characteristic juncture between the NP and the rest of
the sentence. On these NPs steady pitch is maintained and does not


79
paragraph formulae, as we will see below) which constitute such enti
ties are ordered linearly, and he posits a tagmemic formula for each
major and minor type of paragraph in Longacre (1968).
As Becker (1965) rightly points out, the description of paragraphs
as grammatical entities owes much to traditional rhetoric. Longacre's
and especially Becker's formulae for expository paragraphs resemble
closely long-employed formulae for paragraph development in rhetoric
texts. Becker's contention is that the majority of expository para
graphs are organized grammatically in either a TRI (Topic-Restriction-
11 lustration) or a PS (Problem-Solution) fashion.
Longacre's major contributions to the study of the grammar of para
graphs lie in the aforementioned formulae and in his systematic delinea
tion of paragraph types (Longacre 1968) and discourse types (Longacre
1968, 1976) on the original basis of studying paragraphs and discourses
in twenty-five Philippine languages and dialects. His general formula
for the tagmemes constituting explanatory paragraphs in the various
languages is3
iPRELIM (+TEXT EXP0 TREASON RESULT ^WARNING) ^TERMINAL
(1968:109)
With TEXT as the only obligatory tagmeme, this formula can clearly ac
comodate an indefinitely large number of different patterns, since EXPO
(exposition) "can occur an indefinite number of times in paragraphs"
(p. 109). Longacre views TEXT as "much like the topic sentence of tra
ditional rhetoric" (p. 109). We take the formularization of the gram
mar of general paragraph types to be a desirable goal, but it is not


70
The concern of chapter five is an eighth type of deviceconjunc
tions and introducers, insofar as they occur in a set of 200 of the
corpus1 expositions. A rough taxonomy of the most frequent conjunc
tions accompanying left dislocation is presented. Finally, some ob
servations about subparagraphs, especially introduced by "like," are
made.
Longacre and Levinsohn thus elaborate eight types of surface
structure cohesive devices. Our object here is to show that, for one
corpus of English, four of them make complete sense only if viewed in
a discourse context. Although we examine the underlying regularities
of several of the specific devices (left dislocation and deictics which
participate in peak sentences, especially), our primary aim is to de
termine their distribution and function(s) across extended texts, which
also shows their periodic character. This periodic character reflects
the paragraph nature of spontaneous oral expositions. To a significant
degree, surface structure cohesive devices implement the grammatical
units of paragraphs, especially the theme and peak.
For convenience, parts of the analysis are performed on a subset of
the corpus' expositions, a random selection of 200 sample expositions,
five from each informant. In some cases, the analysis is based on a
smaller set of 80 expositions, two from each informant.
Based on the expositions of forty people, our findings have a
broad generality with respect to the community investigated, White
Pine, Tennessee. It might reasonably be asked why so little concern


80
clear yet how tightly one can be devised for expositions in English.
Longacre's quest for a formula for the linear organization of para
graphs serves as a practical model for us; our study focuses on two of
the units Longacre posits in his formula for expository/explanatory
paragraphs.
Longacre has revised his views on two of his units (or tagmemes,
as he called them in 1968), Text and Peak. Originally likening Text
to a paragraph's topic sentence, he indicates that the obligatory unit
in a paragraph is the theme, that piece of information which a para
graph is "built around" and which, for an expository paragraph, "is not
different in kind from a thematic participant" in a narrative para
graph (Forthcoming: 118). The distinction betv/een a topic sentence and
a thematic piece of information is an invaluable one; especially it is
the piece of information (a participant or whatever) which, through an
anaphora chain, gives thematic units to a paragraph. This most recent
view of Longacre's that a paragraph has an obligatory theme is consis
tent with the one we maintain here.
In his earlier view, Longacre considered Peak to be only one as
pect of Text. He states that Text "is the PEAK grammatically and lex
ically of its paragraph. As PEAK of its paragraph the TEXT contrasts
in placement with the PEAK (BUn or STEPp) of NARRATIVE and PROCEDURAL
PARAGRAPHS which come later in the paragraph" (1968:109).^ In a later
work, he indicates that Peak is the plot-like element optionally pres
ent in non-narrative as well as narrative paragraphs (Longacre 1976:
228-9). It comes near the end of and represents a kind of climax or


Copyright 1979
by
Michael Bryant Montgomery


105
New information may be either definite or indefinite; if indefi
nite, it may be specific, nonspecific, or generic (see Stockwell et
al. 1973:67-95 for discussion of these). In the view of Li and Thomp
son (1976), a left dislocated NP represents the "topic" of a sentence,
the part of a sentence regarding which the rest of the sentence makes
a "comment." Although they avoid considering topics in their dis
course contexts, they maintain that across languages topics are always
either definite or specific NPs. They state that "one of the primary
characteristics of topics, then, is that they must be definite"
(1976:461). Givon agrees with this characterization of topics and
points out that they also cannot be referential-indefinite" (1976:154).
Here we view left dislocated NPs as discourse themes rather than sen
tence topics. But the left dislocations from the White Pine data agree
closely with their characterization of topics.
The overwhelming majority of the 606 instances of left disloca
tion in our corpus are of definite NPs. The themes which are left dis
located are so often [+definite] that it can be said without reserva
tion that left dislocated NPs are normally definite. In a relatively
small minority of cases they are generic, as in sentences 29 and 30
below:
29) Yeah, a funeral to me, it's not for the dead person. (XII-18)
30) Children,, if they. took a cold, why they'd make 'em- some
boneset tea.' fl 112-34) 1
Or they are specific indefinites, as in 31 and 32:


108
therefore left dislocation is one type of a more general construction.
Our corpus has 61 instances of constructions without the copy, and
these are of two different sorts. One such construction involves the
repetition or embellished repetition of the fronted NP, as in 39 and
40 below:
39) Well, this town, I love this little old town. (112-31)
40) And then Tate Springs, I know you've heard of Tate Springs.
(IV1-10)
These sentences show that a speaker may choose to repeat a fronted NP
rather than use a copy. Example 41 and 42 show a second construction
a speaker may employ. Here a theme is announced, and its second ref
erence is a paraphrase of the first.
41) And our uh supermarkets, we have two or three nice grocery
stores now. (II14-17)
42) Uh housing, seems like everybody has a new house, you know.
(V4-15)
Such patterns do not relate the fronted NP to the rest of the sentence
syntactically in the same way as left dislocation does. But they re
semble left dislocation in introducing a new piece of information, a
potential theme, into discourse. Also the intonation of such patterns
is identical to that of left dislocations.
As mentioned earlier, left dislocations are obviously not repaired
sentences representing a speaker's second attempt at phrasing. Nor are
the paraphrasing patterns of embellishing or narrowing down an NP such
repairs. In all examples with fronted NPs we find a characteristic


101
The dislocation of two NPs from one sentence, as in 23 above, is quite
rare (two instances in the data). Since we contend in this chapter
that left dislocation introduces new themes into a discourse and that
only one theme occurs per paragraph, it might be asked whether these
double left dislocations introduce two themes and are thus counterex
amples to our contention. Such sentences are clearly not counterexam
ples, however, once their contexts are examined. The context of sen
tence 23 shows that "rural areas of Tennessee" is not a theme newly
introduced by the dislocation in that sentence.
24) . .Uh each community has its own brogue, its own way of com
municating uh to one another, and uh I know, I hear this cri
ticism all the time. Well, the people in, in the small towns
are, are illiterate. Uh they're not polished. They're unedu
cated, they're uncouth, uh which is really not so. Uh we have
some very brilliant people in, in the small rural areas of
Tennessee. Uh some of these very people who would criticize
uh the rural areas of East Tennessee, their ancestors came out
of them, and in many ways they're criticizing their own back
ground, their own family tree. (X4-11)
In this case we have two pieces of information very closely bound togeth
er, but this sentence represents a shift in theme. The new theme is
clearly "some of these very people. ."
Left dislocation sometimes involves a clause intervening between
the NP and the sentence itself. In these instances the speaker appa
rently chooses to provide further information or background so that the
dislocated NP may be clearly understood by the hearer. This apposi-
tional kind of information may be a restrictive relative clause, as in
22 above or in 25 below, or it may be an entire sentence or two of in
tervening information, as in 26-7:


168
the sewer, the water, and the recreation. It was pretty bad,
and they worked very hard in uh doing the things of this type.
We've got a couple of hundred thousand dollars for recreation,
which we've got a new swimming pool we hoping to open this spring
(uh-huh) or early summer, and uh mostly things like that. Recre
ation, streets, water, sewer, and so forth and so on. (uh-huh)
That's been the major changes since I've been here.
F: Has the truck stop going up out there had any effect on town? If
so, what kind of effect?
I: I think it's had very little effect. Now, Roadway uh Trucking Com
pany has put a uh terminal, not a terminal, but a uh I guess you
could say an intermediate or a mid way changing point out there on
the Interstate. Now, it's brought quite a few families into uh, I
would say in the last six months to a year, it's brought ten,
twenty new families into the city of White Pine alone, (uh-huh) and
uh I think Roadway Trucking has had a small impact on the city, as
far as its growth is concerned. I don't, I can't see that the
truck stop has had any. I'm sure it's had some, but I would think
a amount small enough that you really wouldn't recognize it.
F: Hmm. What about the new housing project that's gone up?
I: Well, now it's was built when I came here. Just, they had just com
pleted its (yeah) and uh it gives us a lot of problems. You're
talking about the low-rent?
F: Uh-huh.
I: It gives us a lot of our domestic problems. Uh of course, you can
understand why. You have the type people that live there that do


118
such NPs not to be the center of a subsequent discourse organization.
But it is also necessary to define the relationship between left dis
located NPs and the portions of discourse which follow them. To de
termine this, we look at the subset of 200 expositions from the cor
pus we also used for earlier analyses.
In the 200 expositions, there are 190 instances of left disloca
tion. Over two-thirds of the left dislocated NPs are thematic. They
constitute the central bit of information around which the message is
organized beyond one sentence. There are good reasons for the remain
ing themes of only one sentence, a point to which we return below.
More than one-half of the total (97 of the 190) are thematic for at
least three sentences. Figure 3.1 on the following page summarizes the
extent to which the 190 dislocated NPs are thematic and how many are
thematic to each extent.
Figure 3.1 shows that left dislocated NPs may be thematic for
quite extensive lengths; nearly one-tenth of the 190 instances are the
matic for at least ten sentences. The greatest extent to which one is
thematic is for thirty-eight sentences.
As a thematic device, left dislocation represents a speaker's
strategy of presenting to a hearer what an ensuing block of discourse
will be about. It seems unlikely that a speaker would signal the in
troduction of a theme by such a device which would turn out to be the
matic for only one sentence without there being overriding reasons for
doing so or an explicit signal that such was the case. Once a theme
has been introduced, it seems likely that it would be maintained for


APPENDIX B
SAMPLE INTERVIEW
In order to exemplify the interview techniques employed in this
project, a long portion of a sample interview is here presented. It
is the interview with the thirty-four-year-old police chief of White
Pine, a native of Greene County, Tennessee, and a resident in White
Pine for three years. The code letters "F" and "I" here refer to the
fieldworker and the informant turns in the interview.
F: Well, just in your three years here, have there been major changes
in the way you've had to go about law enforcement in this town?
I: Well, I can't really answer that other than this way. The people
that I talked to, and this includes the city fathers, so to speak,
and the businesspeople, said there's a one hundred difference. Uh
for instance uh they used to congregate in the shopping center and
they'd pass their marijuana around or their pills or whatever,
just from one car to another, and uh they'd rip up and down the
highways. Well, since I've been here, we've got a speed gun. We
cut that out. (uh-huh) We cut the drug problem out in the shop
ping center. We don't allow no loafing at all uh once the busi
nesses are closed. (_hmm) People park their cars and leave them.
Fine. But they don't set in 'em. Uh that was a change we had to
make, so uh, uh to answer your question, as far as the citizens
of White Pine is concerned, they's been a tremendous change. We
just had to set our foot down, get a little tighter. The police
officer that was here, and he was just a one-man department, he
164


53
In the White Pine community, only 1.7 percent of the population
was non-White in 1970; none of the forty informants for this study,
it happens, were non-White.
Transcription Procedures
The approximately forty-five hours of field tape recordings are
the primary documents for this study and are viewed as the ultimate
authority and repository for the data upon which this study is based.
All forty-five hours of interviews were transcribed by the fieldworker/
researcher himself. The transcription was done in slightly modified
orthography in order to result in as close to exact a form as possible
of what the informants said. As Labov (1969, 1972) has shown, careful
transcription is as important in considering grammatical features
as it is for phonological ones.
It must be realized, however, that any transcription can never
be more than an imperfect representation, and an interpretation as well,
of what is on tape. This is certainly true when a transcriber attempts
phonetic or a modified phonetic transcription, and it was true in the
present case when the transcriber had to determine the presence of cer
tain deletions and especially when he tried to be consistent in punc
tuating a transcript or in ascertaining sentence boundaries. A tran
scriber quickly recognizes that deictics and especially conjunctions
mark the beginnings of sentences. But an informant's pausing and pacing


160
other words, "like" seems to signal a size of unit which may be larger
than a sentence but is smaller than a paragraph.
In lieu of a readier term, we adopt "sub-paragraph" to refer to
such a unit intermediate between a sentence and a paragraph. The sub-
paragraph is a supersentential unit in which a theme is not fully de
veloped. Though a sub-paragraph has a theme, it does not have the the
matic nature of a paragraph. -There is no reason to believe that "like"
is the only sub-paragraph introducer, of course. Although they are not
examined here, the phrases "of course" and "you know" seem to operate
in the same manner as "like."
In this chapter we have briefly indicated that conjunctions and in
troducers play important roles in discourse. They piece paragraphs and
discourses together. They initiate paragraphs and discourses, espe
cially when assisting left dislocation in signaling to a hearer that a
new theme is being introduced. Conjunctions then are not only inti
mately involved in the linear structure of the stream of discourse, but
also in indicating new prominences and changes of direction and in ac
companying the beginning of new paragraphs. They are clearly an impor
tant device in both the hierarchical and the linear structure of dis
course.


174
East Tennessee Development District. 1972. County Census Summary:
Jefferson County, Part I. Report No. ET-SS-9g-71. Knoxville.
East Tennessee Development District. 1973. County Census Summary:
Jefferson County, Part II. Report No. ET-SS-13b-73. Knox
ville.
East Tennessee Development District. 1977. Food Stamp Survey,
1976: A Summary Report. Report No. 163. Knoxville.
Firbas, Jan. 1966. On Defining the Theme in Functional Sentence
Analysis. In Travaux Linguistiques de Prague: Volume I, edited
by Joseph Vachek. Tuscaloosa: University of Alabama. 267-80.
Firth, J. R. 1957a[1935]. The Technique of Semantics. Reprinted
in Papers in Linguistics 1934-51, by J. R. Firth. London: Ox
ford University Press. 7-33.
Firth, J. R. 1957b. Ethnographic Analysis and Language with Refer
ence to Malinowski's Views. In Man and Culture: An Evaluation
of the Work of Brontislaw Malinowski, by Raymond Firth. New
York: Harper and Row. 93-118.
Folmsbee, Stanley J., Robert E. Corlew and Enoch L. Mitchell. 1969.
Tennessee: A Short History. Knoxville: University of Tennessee.
Ford, Thomas R., ed. 1962. The Southern Appalachian Region: A Sur
vey. Lexington: University of Kentucky.
Gastil, Raymond D. 1975. Cultural Regions of the United States.
Seattle: University of Washington.
Givon, Talmy. 1976. Topic, Pronoun, and Grammatical Agreement. In
Subject and Topic, edited by Charles N. Li. New York: Academic.
149-88.
Glassie, Henry. 1968. Patterns in the Material Folk Culture of the
Eastern United States. Philadelphia: University of Pennsylvania.
Gleason, Henry, n.d. Some Discourse Features. Manuscript.
Goldstein, Kenneth S. 1964. A Guide for Field Workers in Folklore.
Hatboro, Pennsylvania: Folklore Associates.
Grimes, Joseph E. 1971. Kinds of Information in Discourse. Kivung.
4. 64-74.
Grimes, Joseph E. 1972. Outlines and Overlays. Language. 48.
513-24.
Grimes, Joseph E. 1975. The Thread of Discourse. The Hague: Mouton.


85
tells how to do something; and hortatory discourse endeavors to bring
about a change of some kind.
The major parameters along which these four genres are distin
guished are shown in Figure 2.1 on the following page and adopted from
Longacre (1976). The rows in this figure distinguish those genres
which are +succession (having a sequence in time) and those which are
-succession (lacking a sequence in time). The columns distinguish the
+projection (occurring in the future) from the -projection (those hav
ing accomplished time or in which the time frame is not focal) genres,
longacre's scheme further indicates the person orientation of each of
the four genres.
According to Longacre's typology, narrative discourse characteris
tically involves a sequence in time, occurs in an accomplished (past)
time frame, is agent-oriented, and is related from the first or third
person point of view. Procedural discourse also involves a sequence in
time, but occurs in a projected time frame, is patient-oriented, and is
not related from the point of view of a specific person. Hortatory
does not involve a sequence in time, is set in projected time, is ad
dressee-oriented, and is related from second person point of view. Hor
tatory discourse strictly speaking does not have a time frame but a
modesince the speaker is wishing or urging that something take place.
Expository discourse need not involve a sequence in time nor does it
necessarily have either an accomplished or a projected time frame (time
is not focal). Expository discourse is subject matter-oriented and has
no necessary person reference (point of view). It is essentially im
personal and may involve either inanimate or animate entities. If


132
really happened to us."); and 4) what we call peak sentences (as in
"I guess that's all we can expect out of life is a living."). The
latter two structures are frequently used by a speaker to summarize
a block of discourse.
We are concerned with only the two deictics "this" and "that" in
our data. Generally speaking, "this" may be either anaphoric or cata
phoric, although usually the latter. "That" very rarely is cataphoric
We see, in the following examples from the data, that the two deictics
usually point in opposite directions:
1) He really means to do good. This was one thing about the
Kennedies. They had definite goals that they were going to
try to, . (113-33)
2) So this is a Knoxville man that come to Morristown, and I, I
went out there to see him. . (V1113-18)
3) Well, industry I'd just as soon stay out anyway. As far as,
you know, all industry's doing's polluting the air and
ruining the environment and, you know, that's, like the big
city, that's another reason I wouldn't want to live in a big
city. ~Tv2-13)
In the first example, "this" points forward to a point the speaker
wishes to make about the Kennedies. In the second example, the
speaker uses "this" to point forward to an individual she is giving
information about. Although she had earlier mentioned him, in this
sentence she indicates he is from Knoxville, a new piece of informa
tion which "this" points forward to. In the third example, "that"
makes a broad reference backward to the effects industry has on cities
according to the speaker.
These two deictics have much in common with pronouns. Both deic
tics and pronouns participate in anaphora chains, and both are referen
tial to other parts of a discourse. But the reference of deictics


123
The remaining two instances of the sixty-one which are thematic
for only one sentence involve a speaker's change in perspective. This
is signaled by a phrase such as "I think" revealing an explicit first
person statement. One of the two instance follows:
54) . .and the others, you know, we might draw them into our
group, and again we might not. We'd make a really good
sociological study, I think. (XI2-14)
The kinds of devices used by speakers in our expository monologues
to indicate a change in theme after one sentence are exactly the same
as those used after longer discourse blocks unified thematically. The
only difference is that we find a greater variery of conjunctions if
we look at 190 instances.
For about a third of the cases (62 of 190) we find an explicit in
troduction of a new theme, thirty-three times by another left disloca
tion, sixteen times by an existential sentence, and thirteen times by
the introduction of another character into the exposition. In forty in-
tances of the 190 the theme discontinues because the end of the mono
logue is reached.
The most common conjunctions used to indicate a change in theme for
the 190 cases are the same as were found for the 61 above. Several con
junctions not found (e.g. "anyway") for the sample of 61 instances ap
peared, and there are several instances among the larger sample of com
binations of conjunctions, as "and then of course" for the example be
low, which intensify the speaker's signal:
55) Uncle John Harrison, he uh run that livery stable. And then
of course, they did away with that part of it. . (VI3-13)


60
Notes
1. The choices of the community to work in and of the contact persons
were entirely practical ones. This writer's brother-in-law is a
native of White Pine and his brother-in-law's parents are longtime
residents there. It is they who were chosen to be the primary con
tact persons since they knew nearly everyone in White Pine and could
assist the fieldworker in making the necessary initial contacts.
More important than facilitating contacts, however, was that his
family relationship with the contact persons enabled people in White
Pine to identify him with people they knew well.
2. The category "Some College Work" includes all informants who had any
kind or length of post high school education, such as business or
other vocational school. According to the 1970 census, only 6.8 per
cent of the residents of the White Pine division of Jefferson County
(figures are not available for the town of White Pine itself) over the
age of 25 had attended some colleqe. However, three factors account
for a high percentage (45 percent) of this study's sample having col
lege work: 1) the census area (with 2,085 people over 25) includes
rural area whose residents have a lower averaqe educational level
than do the town's residents; 2) in 1970 a junior college opened five
miles from White Pine; it has attracted many part-time students, and
informants having attended as few as one or two courses are classi
fied as having some college work; and 3) nearly all this study's in
formants who enrolled in college for a short time dropped out (for
many in this generation, "trying a quarter of college" was the ex
pected bridge from high school to the workaday world).
3. This same view, that the fieldworker must be willing to be a sales
man, was expressed to the researcher by Vance Randolph, the dean of
American folklore collectors, in a personal conversation with the
writer on December 11, 1978, in Fayetteville, Arkansas.


CHAPTER IV
DEICTICS AND SUMMARY DEVICES
General Remarks
In this chapter we consider two additional surface structure
cohesive devices, deictics and summary devices, which Longacre and
Levinsohn suggest as crucial to an investigation of the organization
of discourse. They indicate that deictics (words with a specific ref
erent which point elsewhere in discourse, especially the demonstra
tives "this" and "that") are frequently used in languages to keep
track of a discourse's participants. They indicate that summary is an
important "type of paraphrase which makes use of very generic predi
cates and substitutes" (1978:108). The two devices are considered to
gether because, for our data, a speaker usually employs a deictic in
summarizing, as in identifying sentences such as "That's what I
meant."
As a structure which points, a deictic may be either endophoric
or exophoric. If endophoric, it points elsewhere within a text; if exo-
phoric, it points outside a text. An endophoric deictic may be either
anaphoric (pointing backward) or cataphoric (pointing forward). In this
study we will examine only endophoric deictics and specifically will ex
amine their use in four different structures: 1) left dislocation with a
deictic as the pronominal reference (as in "Sulfur and molasses, that's
another thing"); 2) extraposition of a subject (as in "That wasn't pos
sible to get there"); 3) identifying sentences (as in "That's what
131


151
That we should find structures in discourse like peak sentences
which mark the peak of a speaker's relation of information to a
hearer is pointed out by Longacre and Levinsohn (1978:108). More
than a decade ago, in his Philippine study, Longacre indicated that
paragraphs may optionally have units called "peaks" around which
an entire paragraph is focused. His views on the nature of peak
have evolved somewhat since 1968, but his latest comments on peak
(Longacre and Levinsohn 1978:105) apply perfectly to the function
in discourse of what we have been calling peak sentences.
In his 1968 work Longacre did not distinguish peak from text
(the one obligatory unit in the grammar of any explanatory para
graph). Considering peak as only one aspect of text, he states
(p. 109) that text "is the PEAK grammatically and lexically of its
paragraph." In Longacre 1976, he comes to view peak as an optional
unit of exposition but believes that much good exposition, in
exhibiting a tension or progress toward a clear exposing of a
subject, is characterized by having a peak. He views (p. 228) peak
in exposition as analogous to climax and denouement in narrative and
goes on to explain
It seems evident to me that an expository discourse of the
better sort reflects a certain struggle, a struggle to achieve
clarity in the face of recalaitrant elements in the subject
itself and possibly in the lack of background on the part of
those who are to hear the discourse. Exposition of the better
sort is able to clarify the main outlines of its subject in
spite of these difficulties within the subject matter and
within the receptors. ... It would seem therefore that an
artful expository or hortatory discourse will have a meaningful
cumulative thrust. This should correlate in at least some
discourses with a marked surface structure Peak. I believe
that of the various devices open for marking of surface struc
ture Peak in expository and hortatory discourse, rhetorical


Page
Identifying Clauses with Deictics 139
Peak Sentences 143
CHAPTER V CONJUNCTIONS AND INTRODUCERS 154
CHAPTER VI SUMMARY AND CONCLUSIONS 161
APPENDIX A LIST OF INFORMANTS 163
APPENDIX B SAMPLE INTERVIEW 164
BIBLIOGRAPHY 172
BIOGRAPHICAL SKETCH 182
v


84
Lonqacre's Typology of Discourse Genres
It was indicated in chapter one that for practical reasons the
fieldworker chose to elicit expositions rather than other types of oral
discourse in the White Pine community. It was apparent that exposi
tions were the type most readily elicitable from the cross-section of
people that he wished to interview. Also it was decided that the col
lection of a large corpus of expositions rather than smaller corpora of
two or more types of discourse would be more conducive to generaliza
tions about expositions.
That exposition is a distinguishable discourse type is not an ad
hoc presumption on our part. It is a hypothesis based on the typology
of discourses developed by Longacre (1968, 1976). In his study of
twenty-five languages and dialects in the Philippines, Longacre and his
colleagues undertook the basic spadework of sorting out which discourse
types could be described across the languages studied. The typology of
discourses which resulted has provided a practical framework for fur
ther investigations with other languages, as this study is an investiga
tion into the English of the Southern Appalachian region.
Limiting himself to prose monologues, Longacre indicates that four
genres of such monologues (narrative, expository, procedural, and hor
tatory) have common characteristics cross-1inguistically and that
5
within any given langauge there are a finite number of other genres.
We may characterize the four main genres as follows: narrative dis
course relates a story of some kind; expository discourse explains
something or exposes a subject; procedural discourse (e.g. a recipe)


134
But they may also perform another common function of deictics to serve
as a bridge from one sentence or portion of discourse to another. A
deictic may be the bridge between information earlier in the discourse
and an elaboration, an identification, a paraphrase, or a repetition
of this information. In the examples below deictics serve as a
bridge to a repetition (in 5) and an elaboration (in 6):
5) Well, as a community, I think it's a, it's a good community
to live in, to raise your children in. Well, I don't know.
The churches, the schools, everything. I think it's just a,
I don't know, I like it. I think it's, that's why I think
it's a good community is because I like it, and I don't know.
(VI7)
6) ... if you're not doing something, you haven't had any
business for a long time, you know, so there's, there's never
really, never really that much time, but uh we enjoy it and
uh it gives us a living, and uh I guess that's all we can
expect out of life is a living. (VI4-21")
A third basic of deictics in our corpus of expositions is to
maintain a level of tension in the stream of discourse. Most often
this is done by a speaker's use of the deictic "that" when, at least
in written English, "it" would be used. The tendency for speakers to
use "that," a device which is more dynamic than "it" by virtue of its
pointing function, is quite widespread. In many left dislocations in
our corpus (74 of the 606) we find a deictic (usually "that," some
times "those") rather than the third person pronoun as the copy of
the dislocated NP. Two examples follow:
7) That old concrete block there, that was my garage. (V12-24)
8) To go to Morristown to buy a pair of socks, that was kind of
aggravating. (X3-12)
In many extrapositions in our corpus "that" appears as the filler
subject rather than "it," as in


lie
reference to actual potential for dislocation would probably better
show us which NPs are most likely to be dislocated. What percentage
of the animate nouns introduced by speakers are left dislocated? What
percentage of the inanimate ones? Only after answering such questions
can we say that animate nouns are more likely to be left dislocated.
Counting potentially dislocatable NPs has been done here for three
kinds of NPs: names, thematic NPs which are not nonspecific indefinites,
and such NPs which are animate (to determine if animate NPs are more
likely to be dislocated than inanimate ones). In this way, two ques
tions about potential for dislocation can be answered: how likely are
names (persons and non-persons) to be dislocated in expository dis
course, and how likely are thematic pieces of information of all kinds
likely to be dislocated in expositions?
To answer the first question, the names in a subset of 200 exposi
tions (fiveofrom each informant) were counted. Ninety-three instances
of names were found, 27 of which were dislocated. Twenty-nine percent
of the potential names, therefore, were dislocated.
To answer the second question, the thematic pieces of information
in a subset of eighty expositions (two from each informant) were
counted. In many cases, thematic elements were difficult to identify
because the speaker changes direction or introduces further, unrelated
information before they are clearly established (and before they ini
tiate an anaphora chain. As a result, we can only approximate the num
ber of potential cases. In the eighty expositions approximately 280
such potential cases were found, and fifty-nine of them were left dis
located (21.1 percent). Roughly speaking, then, between one-fifth and


75
Thematic unity is not something merely imputed to a paragraph in
speech. It is there and is reflected to one extent or another by var
ious lexemic, grammatical, and phonological features. In a narrative
discourse, paragraph themes are usually characters; in exposition the
themes may be characters or many other kinds of information. Examples
will be presented in chapter three.
We can view these thematically unified units called paragraphs as
having three different modes of structurephonological, lexemic, and
2
grammatical. Although of most interest to us is the grammatical mode,
all three modes to a great extent reinforce one another. It is doubt
ful that the grammatical structure of a paragraph can be understood
without studying the other two.
Paragraphs have not often been understood as having characteristic
phonological features. But according to Klammer, "phonologically,in a
complete description of dialogue paragraphs, it would be necessary to
account for such phenomena as stress placement, juncture, and intona
tion in relation to the paragraph level constraints affecting them"
(1971:33). Even earlier, Becker pointed out the existence of phonolo
gical markers of paragraph structure:
Paragraph tagmemes seem to be marked by shifts in pitch regis
ter, tempo, and volume when paragraphs are read aloud. While
these signals can be perceived by a trained phonetician, they
have not been adequately described in the laboratory, and
their written counter parts have not been identified. (1965:242)
Unfortunately research into such features and markers, as by
Bridgeman (1966) and Lehiste (1978, 1979), has been either too recent


47
interview and over signaling the end of the interview and his impending
departure. It is easier to tell an informant that the interview is
over, to pack up the equipment, and to say good-bye than it is to tell
an informant that he/she can go home now. It is also less of an impo
sition if the fieldworker takes care of his own transportation instead
of asking the informant to arrange his/her own transportation. Fi
nally, informants appeared more likely to venture information in the
securest and most familiar environment for them, their own homes.
The fourth thing that the researcher/fieldworker had to determine
before beginning full-scale interviewing was the type(s) of discourse
he would elicit and the interview format to elicit it (them). His ap
proach was based on the typology of discourse genres (narrative, expo
sitory, procedural, and hortatory) in Longacre (1976). Given the time
limitations and the ease of devising questions, the researcher decided
to concentrate on expository discourse. A standard set of questions
was asked of all informants, regardless of age, socioeconomic back
ground, or level of educational attainment. No difficulties were ex
perienced in elicitation according to this plan.
These limitations also had several advantages. First, using stan
dard questions gives the study strict controls. Then, expository dis
course seems to be the most plausible to elicit, given the ethnographic
parameters of this study. Most of the standard questions had to do
with everyday life in the town. Others dealt with situations that were
of widespread and frequent comment about town (such as the reputation
of Jefferson County's neighbor, Cocke County, for violence and


87
it brings in people," Longacre says, "they are simply subjects of
explanation and analysis" (1976:199).
More recently Longacre has furthered elaborated his scheme by
adding another parameter (ftension). What he says about tension char
acterizing some discourses is worth repeating:
The final parameter tension also applies to all the above dis
course genres. Narrative discourse is purely episodic if it
contains essentially no struggle or plot (for example Willa
Cather's novel Shadow on the Rock). Such discourse is consi
dered to be -tension. Most stories, however, involve some
sort of struggle or plot and are clearly +tension. Similarly,
in procedural discourse, there are some discourses which are
more or less routine, while others involve struggle and alter
natives and are therefore +tension. Both behavioral [horta
tory] and expository discourse have varieties in which argu
mentation is assumed. These are also +tension. Discourses
of these genre which do not have this characteristic are
-tension. (Longacre and Levinsohn 1978:104)
we mention this third parameter because it concerns the presence or
absence of a peak in the discourse: If a discourse is plus tension,
there will most likely be some kind of climax of development, some
marked surface structure peak1 (Longacre, 1978 ). This is a pri
mary concern of chapter four.
Longacre's characterization of expositions appears to have consi
derable validity for the expositions constituting our corpus, although
it must be pointed out that Longacre's types are "pure." In actuality
discourses often of different types are embedded within one another.
For example, within an exposition may be embedded a short narrative
for background or illustrative purposes. Longacre recognizes this in
viewing his genres as deep structure ones which are sometimes skewed in
the surface structure (1976:206-9).


88
Notes
1. The parenthetical notation at the end of these and all further
data in this dissertation refers to the page of typed transcript
on which they are found. For information on which speaker any
data are from, consult the list of informants in Appendix A.
2. Terminological obfuscation notwithstanding, we use lexemic to
refer to concepts such as those of Klammer and Wise, including
contrast, comment, addition of participant, specification, which
are manifested in linguistic structure in various ways. Other
tagmemicists (e.g. Pike and Pike 1977) use "referential" to
refer to such concepts. Stratificationalists use the term "seme-
mi c. "
3. For Longacre, terms "explanatory" and "expository" are equivalent
in his 1968 study.
4. Step and BU (Build Up) are the obligatory tagmemes in procedural
and narrative paragraphs.
5. In his 1976 work, Longacre does not distinguish description from
exposition and we will consider description a subtype of exposi
tion here.


16
population (only 1.7 percent of its 1970 population was Black), and
many families have been split because the children have had to mi
grate north to find jobs. This need to leave home for employment
was aggravated for many natives of East Tennessee and Southern Appa
lachia by their inability to adjust to the urban North and their fond
attachment to home in the hills. Instances of their choosing to return
to certain unemployment or underemployment in the hills have been
frequently documented, as by Hicks (1976:29-30). This is perhaps not
so common in White Pine as in other areas in the region. A similar
attachment to native ground continues today; relatively few graduating
high school students and other new members of the labor force wish to
settle down elsewhere.
Almost entirely responsible for the improved economic profile for
Jefferson County and White Pine has been the tremendous growth of
local factories, especially those of the American Enka Corporation and
several major furniture firms. Attracting such industry to the area
have been the labor surplus, various tax incentives, and the availa
bility of cheap electricity. By 1970, more than half (51 percent) of
White Pine's employed population was in manufacturing, eight times as
many as in agriculture, forestry, and fisheries combined (6.3 percent).
In the same year, trade, service, and finance supported another 23.1
percent of White Pine. Construction employed 9.1 percent, mining 0.4
percent, and all other types of employment, 10.0 percent. The move
ment of workers into manufacturing has entailed the dissolution of the


122
In the following excerpt, the dislocated NP, "this one son that's
with me" is thematic for only one sentence. The following sentence
begins with both a conjunction and another left dislocation to indi
cate a change in theme, for contrast in this case.
51)This one son thats lives here with me, he'll just sit and
die laughing at stuff that goes on and, and enjoy it, you
know. But the other two, they don't fence with that stuff,
so I don't know. (VIII3-20)
The remaining fifteen instances of a left dislocated NP thematic
for only one sentence are of four different kinds. In five instances,
the left dislocation occurs in the last sentence of the exposition, in
dicating the speaker did not wish to pursue the subject further. In
four instances the NP is copied by "that" in an identify sentence, as
in
52) Threshing floors, that's what they had. (II13-27)
We will discuss identifying sentences in chapter four. They present a
summary identification regarding something and it is hardly surprising
that a speaker would introduce a new theme after making such a point.
Another kind of summary effect is given when the sentence follow
ing the one from which a dislocation takes place has an intensive
"just" in it, as in 53 below:
53) . .but the other two, they don't fence with that stuff, so I
don't know. I reckon it's just the way you think. (VI113-20)
The summarizing effect here indicates this is well-constructed thema
tically. Four of the sixty-one instances have this signal.


125
of something. The information introduced by existentials is typically
new to the stream of discourse, as we see in the extended example be
low:
56) . .if that canning factory hadn't burned, it'd give a lot of
people jobs in that area too. You know there's quite a few
people in around Bybee. (uh-huh) We had a couple of stores
and a post office and things like that, but it was just, com
paring it to here, to White Pine, just about like White Pine.
Cause at one time or other there was a, an old hotel here in
White Pine, and a post office, and uh I, even before that
they might just uh been a country store, not even a drug
store, but of course see, they added the drugstore and the
bank and then different and, and different things like that,
and then the railroad. Now there was no railroad that went
through Bybee. It was out in the country more. I mean, of
course, railroad track in the country too, but I mean it was
off the railroad line, and uh in White Pine uh here wasn't.
There used to be a depot up here. You could get off. You
could get on the train and get off the train right there.
It's been torn down not too long ago. (hmm) There's an
express office up there. I don't think it's been torn down
uh more than, uh more than maybe about six or eight years.
Little gray building. You know, you could still have things
come in express but no passenger, you know, there's not any
passenger trains even in Tennessee now. (11-13)
In this exposition we find seven existentials (each underlined) which
introduce new information. Two ofthem(". .there was no railroad. ."
and "There's not any passenger trains. .")donot introduce new (not
previously mentioned) NPs like the other five. But they do add new in
formation (that something previously mentioned does not exist) to a mes
sage and thus push forward the progress of a discourse. We have more to
say about the functions of existential sentences after examining the
kinds of information they introduce.
Unlike left dislocation, which normally introduces definite NPs,
existential sentences normally introduce information which is


104
Specifications
In the above section it is shown that any NP in a sentence has the
potential for being dislocated. Earlier it is noted that the NP dislo
cated must be the unifying piece of information, or theme, of a block
of discourse (this is an important point and is demonstrated in the
section "Relationship to Subsequent Discourse" below). A theme, which
in a well-formed block of discourse or paragraph often establishes an
anaphora chain, represents something a speaker wishes a hearer to have
in mind so that a body of information can then be built around it. The
theme thus is information the speaker wishes to be in the forefront of
the hearer's consciousness. It is NEW information, which Chafe regards
as knowledge which : speaker assumes to be not in the "consciousness of
the addressee at the time of the utterance" (1976:30).
All thematic devices such as left dislocation, existentialization,
and clefting constructions usually introduce new information into a
stream of discourse. But all such devices do not introduce exactly the
same kind of information. Moreover, they are not interchangeable.
As a rule, existentials do not introduce definite NPs. The excep
tions to this are treated in Reed (1979). Reed claims the criterion for
a noun being existentialized is whether it has a "presupposition of
identifiable reference." In Chafean terms, this means "knowable" to the
hearer. Without exception, left dislocations do not introduce nonspecif
ic indefinite pieces of information into discourse. Knowing this, we
not only can relate these two thematic devices to various comments in
the literature, but we also can point out the important differences and
overlappings between them.


17
tenant farming system and the development of more specialized agri
culture in the area.
This is not to say that farming is no longer a major concern to
most of the people of White Pine. It is. It must be remembered that
most of its citizens are a generation or less off the farm; many of
them still religiously tend kitchen gardens of an acre or two. But
few citizens must still depend on it as their primary source of income.
The new sources of income, the factory payroll primarily, have
brought in their wake profound changes in White Pine (above p. 10).
The better income is more dependable than one might derive from farm
ing. It has caused White Pine to become a bedroom community. At the
same time, the improved standard of living has raised the optimism of
White Pine's citizens for their future and especially for the future of
their children. For those wishing to break the cycle of having to go
to work immediately after leaving public school, there are more oppor
tunities than ever before. Within the past decade a junior college has
opened only five miles away, and many of White Pine's youth have de
cided to undertake some college work rather than enter the job market
right away. The area's economy, although still not providing suffi
cient skilled positions, continues healthy at the same time that White
Pine, a self-sufficient community a generation ago, finds itself more
and more dependent on the surrounding area.
The Religious Life of White Pine
The community of White Pine lies in the region often referred to
as the "Bible Belt," so called because much of its population still


27
subject of geographical variation entirely. This creates the impres
sion that there is none. No attention was given until very recently
to variation by age, socioeconomic background, or any other variable.
This presumption of little, if any, variation is directly at odds
with the view of Southern Appalachian speakers themselves. They be
lieve that Southern Appalachia has much subregional variation. Speak
ers in the region as a rule (and White Pine exemplifies this well) con
tend that they speak differently from neighboring communities only a
few miles down the road or one or two hills over, although it is diffi
cult to tell what differences they have in mind. These subjective
views have occasionally been supported by statements in the linguistic
literature, as by Berrey: "the dialect may vary slightly with the lo
cality, and even from family to family" (1940:46), or by Walker: "It
must be realized that the speech of the mountaineer varies from one
district to another" (1939:6). Wolfram and Christian, in commenting
on vowel variation in southern West Virginia, state that "there are
apparent vowel differences from region to region within Appalachia.
Even within the restricted locale we have studied here, there is evi
dence that several different vowel systems must be recognized"
(1976:64).
To what extent the term "Appalachian English" refers to a deter
minable linguistic entity is then unsettled. On the one hand, some
commentators and Appalachian people recognize an almost hollow-to-hol-
low and community-to-community variation. On the other, such long
time students of American English as Mencken (1963) and Wise (1957)


68
first link the "theme." The theme is the piece of information (not
necessarily a human participant) which usually initiates an anaphora
chain and which a block (or paragraph, as we will see) of information
is about. The primary surface structure devices which introduce theme
or "thematize" a piece of information are of two sorts in our corpus:
left dislocations and existentials. In the following exposition we un
derline two left dislocations and two existentials, all four of which
initiate anaphora chains, to illustrate our approach.^
Fieldworker: What have been the major changes you've seen in
White Pine during your lifetime? Let's say, the most signifi
cant changes?
Informant: Well, they restructured the bank. I wished they'd uh
left it the old way, but I guess they did that for, you know,
security measures and stuff, and then uh the mall, the mall I
was talking about, the shopping center. You know, they've
paved it all the way throuqh, so that's, that was the major
thing right there that, that they've done. And of course,
the housing project, they put that down there, and I wish they
hadn't have. I like White Pine the way it was, you know.
Well, simpleton little town, you know, but it seems to be ex
panding quite, quite a bit. (hmm) And there's a lot of
trailer courts going in. There's one right in the center of
town. I don't know how they ever got that admitted. Course,
they have restrictions now, but it was in there before the re
strictions came, so (uh-huh) and, when I grew up, there, we
had one policeman, and that was Chief of Police Dick Reed, and
he had a deputy, and that was it, and they would drive around,
you know, most of the day, or if you needed him or anything to
around twelve o'clock, and then they'd go home, you know, so
the town was just, just nobody out, you know, or few people
out, but there was no crime, you know, nobody breaking into
anything, (uh-huh) That was just the way it was,... (V2-11)
In this exposition thematic devices introduce four themes (the mall,
the housing project, a lot of trailer courts, and no crime, although
the fourth one is not developed like the others) and at least four
paragraphs in this exposition. Theme is the one obligatory element


112
47) . .of course, as I say, the young ones I don't know. But
them old ones, if they liked you, they'd bend over back
wards for you. They'd do anything in the world that, for
you, if they liked you. (X2-16)
Such left dislocations serve not only to indicate a contrast with a pre
ceding piece of information. They also serve to introduce thematic ele
ments for the succeeding sentence(s) like other types of left disloca-
ti on.
4) Often a speaker in the course of an exposition wishes to rein
troduce a piece of information back into the stream of discourse after
it has fallen into the background of the hearer's consciousness. The
information may have been mentioned in an earlier turn in the conversa
tion or earlier in the same turn. But when a speaker wishes to rein
troduce it, it is often found in a left dislocation structure. Appa
rently the constraint on this type of left dislocation is that at least
one thematic piece of information has intervened since the last mention
of the reintroduced one. This intervening material may be from one
sentence to many dozens of sentences long. In our corpus, 29 of the
606 instances (4.8 percent of the total) are primarily of this type.
In the following excerpt, "sulfur" is reintroduced as a thematic piece
of information:
48) . .my aunt used to buy calama. You used to could get calama
just uh loose, I mean, at the drugstore, and she would just
take what calama, she said that calama in the spring, that
everybody needed a dose of calama to, to clean their system,
and so she would take, and just what would lay on the end of
a, of a knife, table, that we would, why, she'd give us, all
of us children a dose of calama in the spring, and we'd uh,
and uh sulfur and molasses. That's another thing. She


CHAPTER 1
INTRODUCTION
General Comments
This dissertation describes the discourse functions of several
linguistic structures from a large sample of spontaneous expository
discourses for a cross-section of informants from a small East Ten
nessee hill community. The community, White Pine, is viewed as ty
pically Southern Appalachian. This dissertation thus contributes to
both discourse analysis and Southern Appalachian studies.
This study is based on a set of forty interviews conducted by the
researcher between December 1977 and February 1978 and averaging
slightly more than an hour each in length (forty-five hours total).
The interviews consisted of a small core of questions and were de
signed to elicit large amounts of expository (explanatory and descrip
tive) discourse from the informants in the form of spontaneous mono
logues. Subsequently the interviews were transcribed orthographically
and constitute a 1,070 page typed corpus from which data in this dis
sertation are taken.
Language, as a social instrument, consists of a set of systematiz-
able patterns of social behavior. These patterns, although presumed
to be repetitive and systematizable on all levels, have been largely
ignored for English at levels above that of the sentence. The
1


147
that. But for a good time and not leave White Pine, it's
either ride around or watch TV. Usually watch TV, because
they a lot of good stuff been coming on TV here lately.
(VII12-9)
Or is it anaphoric or cataphoric in the context of the following
exposition?
49) Fieldworker: What would you say have been the most signif
icant changes you've seen in the town in your few years here?
Informant: Most signficiant change I've seen is the, the
intrusion or the coming in or, of new people. White Pine is
no longer the same town it was eight and a half years ago
because of its growth, uh because of people with different
backgrounds having moved into our town, (uh-huh) Uh there
fore it has changed the complex of our town tremendously.
Uh I think this is the greatest change that has happened in
our town is to have new people coming in with new blood, new
thoughts, new ideas, new approach to things" (X4-4)
On the one hand, the deictics point backward in the discourse; on the
other, they point forward, to information coming at the end of the
sentence which sums up the entire preceding discourse. In a sense,
we may call such deictics "ambiphoric," pointing in both directions.
Peak sentences resemble identifying sentences with right dis
location (as in 37 and 38 above). But the deictics in peak sentences
are more clearly bridges between foregoing and following portions of
discourse. Normally the initial deictic in peak sentences (although
a speaker retains the option for such stressing in cases of emphasis)
does not receive the focal stress we find with deictics in anaphoric
identifying sentences, indicating that peak sentences are on the whole
more forward looking. Thus they more strongly tell a hearer that
important information is coming up.
An examination of all fifty peak sentences shows that all of them
point both backwards and forwards. In some instances, as in 50 below,


149
initial deictic seems especially cataphoric when it is "this," as in:
53) ... We have a volunteer fire department, which I might
say is outstanding. It's terribly proficient. It's very,
very, very good, (hmm) And we've improved on that as times
come along, (uh-huh) Uh, we have two trucks. We have a
large engine. Have a small brush-fire, brush-fire fighter,
uh to handle any small jobs, and we now, as of this year,
have good communication. Uh, they have portable radio units,
uh volunteer fire department, (hmm) and all times. They're
on the job from somewhere. We, we have quite a few within
the town itself that, that have their own business, that
are capable of being able to go on, to go out at any time,
(hmm) Uh, this doesn't seem to matter as far as uh insur
ance is concerned. This is something that's bothered me
for a long time is that we're in uh grade ten in insurance,
as far as homeowners is concerned. (113-12)
Although peak sentences and identifying sentences have been
treated separately in this chapter, they are quite similar in many
respects. When we look at their distribution in discourse, we will
see they are very close. At the end of chapter three we examined
structures which were the fusions of two structures left disloca
tion of an NP and existentialization of an NP that had been viewed
as distinct until that point. In the same way as thematic devices are
not discreet from one another, we find a continuum of structures
involving deictics. Peak sentences are identical to anaphoric identi
fying sentences except that the latter have a pause where the former
have a copulative verb ("is") and that the latter have a focal stress
on the initial deictic whereas the former may or may not have. In
addition, anaphoric identifying clauses with right dislocation such
as in 35-6 above are very similar to that-extrapositions, as in 27-8,
the latter are never anaphoric and never identify one part of a
discourse with another.


91
intuitive notion but that it demonstrably organizes a block of dis
course.
Our view of theme as that information which unifies a block of di
ourse and which often is reflected by an anaphora chain is not incon
sistent with other views found in the literature. For Hallida.v,
Theme is concerned with the information structure of the clause
with the status of the elements not as participants in extralin
guistic processes but as components of a message, with the rela
tion of what is being said to what has gone before in the dis
course, and its internal organization into an act of communica
tion. . (1967:199)
Halliday's position is that theme is involved in both clause and dis
course structure. Theme is "what comes first in a clause" and around
which the remainder of a clause is structured (1967:212). But also
theme is that information which "is being talked about" (p. 212) and
which assigns
. .to the discourse a structure which is independent of sen
tence structure and through which the speaker both organizes
the act of communication into a chain or message blocks, the
'information units', and specifics within each message block
the value of the components in the progression of the dis
course. (p. 196)
Halliday's express concern, however, is not to show how a theme parti
cipates in discourse but to explore the many choices that a speaker
has for theme and how a clause is organized around such choices. To
support his clain that "the soeaker has within certain limits the op
tion of selecting any element in the clause as thematic" (p. 212), he
shows the variety of different elements which may be fronted and thus


21
viewed as sharing certain distinctive cultural characteristics and
social and economic conditions encompasses the southern half of the
mountain range and includes as its core a sliver of North Georgia, the
eastern thirds of Tennessee and Kentucky, the western thirds of
Virginia and North Carolina, and the entire state of West Virginia.
This is in addition to a variety of fringe sections, depending upon
which demarcation of the region one chooses. Perhaps the most exten
sive demarcation is the political one made by the Appalachian Regional
Commission. The Commission defines Appalachia by county, based on
economic conditions. Their "Appalachia" includes parts of Alabama,
Mississippi, South Carolina, Ohio, Maryland, Pennsylvania, and New
York, in addition to the core area mentioned above.
Gastil, whose regions are determined by sociocultural features,
considers the core area outlined above and a small corner of South
Carolina to be the "Mountain" area of the South (Gastil 1975:174ff).
Vance (1962) views the "Southern Appalachians" as a distinctive region
sharing agricultural and economic conditions, which covers parts of
West Virginia, Kentucky, Tennessee, North Carolina, South Carolina,
Georgia, and Alabama. The demarcation of the "Southern Highland" re
gion by John C. Campbell (see map on the following page), a lifelong
student of the region, comprises the core area of portions of six states
and also the fringes of three others, a 112,000 square mile territory.
For Campbell, this area exhibits similar economic, educational, and ref-
igious characteristics. What is important for us is that, whatever the
delimitation of Southern Appalachia, the community of White Pine has always


48
lawlessness). These are safe topics, unlikely to give offense, yet of
universal interest.
Before the interview itself and the sample are discussed, a few
more comments need to be made about how the fieldworker established
rapport with his informants. As Goldstein points out, the fieldworker
should take every opportunity to participate in the daily lives of the
people and the community he is studying. This researcher visited local
churches and various gathering places and meetings, chatting with peo
ple frequently on the streets and at lunch counters. He continually
emphasized that he was interested in interviewing a cross-section of
townspeople.
The interview itself is a directive one, but it is loose and flex
ible. It was directive insofar as it required specific vital informa
tion from informants (age, length of residence in White Pine and else
where, level of education attained, occupation held, and family back
ground). The fixed questions asked of all forty informants require
expository responses:
1) How does White Pine react to newcomers in town?
2) What have been the most important changes you have seen in
White Pine while you have lived here?
3) What's a typical day like for you?
4) What do you think most people in White Pine want the town to
become?
5) What all goes on on Pine Day?
6) What do you do on Pine Day?


64
concerned with the abstract and rhetorical structure of discourses.
Text grammarians for the most part view discourse or text as a unit of
competence and not of performance.
Other analysts approaching discourse (e.g. Williams 1964, Halliday
and Hasan 1976) choose to study exclusively the various cohesive re
sources of English to fit one sentence after another into a context,
to produce acceptable sentence sequences. Williams examines deletion
and a wide variety of elements such as deictics and sentence adverbs.
Halliday and Hasan are concerned with such semantic resources of Eng
lish as substitution, ellipsis, reference, and conjunction. This lin
ear approach to discourse is devoted to explicating intersentential
linkage.
Several discourse studies attempt to bridge the gap between these
two points of departure and to balance the hierarchical and linear ap
proaches to the study of discourse organization. They have varying re
levance to the present study, however, since none of them (except, in
part, Longacre 1968) are based on a corpus of spontaneous oral exposi
tions. Several studies, e.g. Grimes and Glock (1970), Klammer (1971),
Klammer and Compton (1970), and Van Dijk (1977), deal with only the
narrative structure of one or a small sample of texts. The basic point
of many narrative studies is that a discourse view of a text can indeed
make some interesting observations more so than it is to generalize
about a body of texts. In studies of narrative structure we encounter
again and again statements similar to that of Van Dijk: "The remarks
made about a well-known type of discourse, the story, could be extended


139
An extraposition, then, is a structure with information shifted
to the end of a sentence. When a speaker uses "that" rather than "it"
in an extraposition, he/she explicitly points forward to this infor
mation. An especially strong cataphoric effect occurs in the one
extraposition in our data with "this":
29)This don't sound feasible that they would be from all areas
of the country. (114-10)
Identifying Clauses with Deictics
In chapter three the infrequency of it-clefts (three instances)
and WH-clefts (four instances) was pointed out. Such constructions
seem to be more widely employed in written than in spoken English.
In those same 200 selected expositions in the White Pine corpus which
were searched for cleft constructions, however, we find another kind
of construction superficially similar to clefts. There are 27
instances of this construction, which is termed, after Halliday
(1967:231), "anaphoric identifying clauses." Their frequency substan
tiates Halliday's statement (p. 232) that they "are probably at
least as frequent, in dialogue" as clefts are. In the following
three excerpts the identifying clauses are underlined:
30) They're putting in I call it a, I don't know whether this
is the proper term to use, it's the term I use, a beer
joint. And that's what everybody else around here says,
right down our road down here. (V4-19)
31) They set it up out in the shopping center there. They just
have that whole thing roped off, and, and that's where they
have their . (II13-15)
32) So I mean, what more can you do but fight back? When you've
already turned the other cheek, there's only one thing to do,
and that's fight back, and even the Bible'll tell you to do
that. So that's, that's what we're doing. (VII12-27)


49
7) Why does Cocke County have a bad reputation?
8) How do local people feel about outsiders making fun of East
Tennesseans and the way they talk?
In this way, some uniformity in interviews was established.
However, the interviews were not rigid. For example, no infor
mant was asked only these eight questions. Each informant was asked
additional questions depending on age or background. For instance, in
formants over fifty who were native of the area were asked how local
people felt when TVA came into the county in the 1930's and bought up
prize farmland to be covered by reservoirs. Similarly, high school age
informants were asked how the new county-wide comprehensive high school
compared to the old high school in White Pine. Other questions were
asked, depending on the flow of the conversation.
One non-directive principle was adhered to throughout the interview
ing, however. This concerns the lack of a time limit. Once the field-
worker had asked a question on some subject, the informant was allowed
to give as long a response as desired. The fieldworker took pains nei
ther to interrupt nor to influence the informant's pattern of thought.
Only when the informant signaled an end to a response or when an infor
mant's response was clearly running out of steam did he intrude with a
comment or another question.
The interviews averaged slightly more than an hour in length. But
the fieldworker spent at least twice that amount of time with the aver
age informant, exchanging small talk and establishing a rapport with in
formants before the actual taping began. The time taken by the field-


179
Pike, Kenneth L. 1964. Discourse Analysis and Tagmeme Matrices.
Oceanic Linguistics. 3. 5-25.
Pike, Kenneth L. 1967. Language in Relation to a Unified Theory
of the Structure of Human Behavior. The Hague: Mouton.
^Pike, Kenneth L. and Ivan Lowe. 1969. Pronominal Reference in
English Conversation and Discourse A Group Treatment.
Folia Lingistica. 3. 68-106.
Pike, Kenneth L. and Evelyn G. Pike. 1977. Grammatical Analysis.
Dallas: Summer Institute of Linguistics.
Postal, Paul M. 1971. Cross-Over Phenomena. New York: Harper & Row.
Prince, Ellen F. 1978a. A Comparison of WH-clefts and it-clefts in
Discourse. Language. 54. 883-906.
Prince, Ellen F. 1978b. Left Dislocation, It Makes Life Easier.
Paper read at the 1978 LSA Winter Meeting, Boston.
Propp, Vladimir. 1958. Morphology of the Folktale. Bloomington:
Indiana University.
Raine, James Watt. 1924. The Land of Saddle Bags. New York: Coun
cil of Women for Home Missions and Missionary Education Movement
of the United States and Canada.
Reed, Ann M. 1979. 'There' and Definites. The SECOL Bulletin.
III. 1. 9-24.
Ross, John Robert. 1967. Constraints on Variables in Syntax. MIT
dissertation.
Sacks, Harvey, Emanuel A. Schegloff and Gail Jefferson. 1974. A Sim
plest Systematics for the Organization of Turn-Taking for
Convestigation. Language. 50. 696-735.
Sinclair, J. McH. and R. M. Coulthard. 1974. Towards an Anatomy of ^
Discourse: The English Used by Teachers and Students. London:
Oxford University.
Stewart, William A. 1967. Language and Communication Problems in
Southern Appalachia. Washington: Center for Applied Linguistics.
Stockwell, Robert P., Paul Schachter, and Barbara Hall Partee. 1973.
The Major Syntactic Structures of English. New York: Holt.
United States Department of Agriculture. Bureau of Agricultural
Economics. Bureau of Home Economics, and Forest Service. 1935.
Economic and Social Problems and Conditions of the Southern
Appalachians. Miscellaneous Publication 205. Washington, D. C.


57
Explanation to Informants
The members of the White Pine community, when approached by the
fieldworker/researcher as potential informants, were almost invariably
most willing and even flattered to be asked to participate in an inter
view. While a few of them were initially concerned that they were
being "singled out" to be interviewed for some special reason (the
fieldworker had to assure them that he wanted to interview typical cit
izens), the fieldworker was actually turned down for an interview only
once (over the telephone, where, he is convinced, he was mistaken for a
salesman). Most informants were more interested in knowing why they
(and White Pine) were chosen than what the fieldworker actually wanted
to interview them for, and that the series of interviews was a part of
his schoolwork was sufficient explanation for most informants to be
quite willing to be interviewed. The usual explanation provided pro
spective informants was as follows:
For a school assignment I would like to interview a cross-section
of people in White Pine and ask some general questions about
what life is like here. What I want to do is compare the an
swers I get from a variety of people, both old and young, male
and female. I wondered if you would be willing, if you have
some time, to help me by answering a few questions.
The fieldworker willingly gave whatever additional information was de
siredthat the interviews were part of his dissertation, that he was
working on his doctorate, or that his project was a linguistic one
but this was rarely called for.
It should, of course, be pointed out again that the initial rap
port that the fieldworker was able to establish with informants would


25
handful of its features have indeed been shown to be carryovers from
Northern England and Scotland) or its musical or other traditions.
For a variety of unscientific and often impressionistic reasons,
however, the people of Southern Appalachia are reputed to be quite dif
ferent from all other Americans. If they do not descend "directly from
the Scotch-Irish, there is a quaintness in their approach to life, it
is said, harking back to an earlier era, or their speech has a quality
and character indisputably unique. Not only does Appalachian speech
preserve various relics, that is, but it also has a distinctive quality
in a more general sense. The literature on the subject abounds in de
scriptions of how mountain folk talk, more often than not written by
college-educated children remembering fondly earlier days in the hills.
Here are three representative descriptions:
There are certain peculiarities of enunciation which it might be
well to speak of here. For the Smoky Mountaineer the nose is as
much an organ of speech as the larynx. This is particularly no
ticeable in old people, and may be because of the catarrh
brought on by constant exposure and diseased tonsils. The moun
taineer drawls to an alarming extent, even more so than his
neighbors of the plantations. His voice is utterly without ca
dence, almost a droning monotone. On the other hand, it has a
deep, resonant quality which catarrh and embarrassment never
wholly obliterate. (Walker 1939:3-4)
The mountaineer has clung to the Shakespearean word these many
years because he has not learned the modern word. Indeed, he
has had little contact with those who speak the modern word; and
since his surroundings and habits and thoughts have been largely
the same as they were in Shakespeare's day, he has had no need
for new or changed speech. (Coleman 1936:30)
As a race, the people of our southern mountains speak softly.
Even their most commonplace remarks somehow are made to sound
secret, and of greatest importance. The women especially speak
softly, their voices often plaintively sweet, pitched in a
minor key and lifting upward, so that the last word of every
sentence is almost sung. (Hannum 1969:29)


130
fall off as at the end of a sentence. The three examples above and
the fourteen others cannot be regarded as anything but fused sen
tences.
It is here that our two types of thematic devices, left dislo
cation and existentialization, considered separately in this chapter,
fall together and show that they are indeed equivalent structures
when it comes to introducing thematic NPs which are either specific
or generic indefinite. Such NPs may be introduced by either or both
devices. Especially with generic NPs, the two devices most clearly
merge.


Ill
new information. The left dislocated element will name a general class
of items, and the following sentences will specify which and/or how
many members of the class are involved. In other words, some left dis
locations involve the focusing on or narrowing down to a portion of the
new information which is introduced. There are 34 instances (5.6 per
cent of the total) of this type in the corpus. Two short examples of
this are given below:
44) People my age that went North and worked, some of 'em have
come back [to] the territory. (IV4-18)
45) These students, I know several of 'em because Ive helped the
teacher average their grades. (VIII1-18)
We note the similarity between such constructions and those like sen
tences 39 and 40 above.
3) On occasion, a speaker wishes to introduce a piece of informa
tion to a hearer abruptly in order to contrast it with a foregone piece.
In such instances the new information is often preceded by "but" or "how
ever" and frequently it is left dislocated. In our corpus are ten in
stances (1.7 percent of the total) which are primarily contrastive and
several dozen more whose constrastiveness is evident but apparently sec
ondary. Although usually preceded by one of the conjunctions mentioned
above, their primary characteristic is the heightened stress on the left
dislocated element which they have. In the following examples we have
contrastive stress on "younger" and on "old":
46) . .and I think the older people think theres no use in all
this change, you know. Why, we've gone like this for years.
Why not continue? (uh-huh) But I think the uh younger,
younger people in town uh, I think they want change. I, I
think they want it to grow and develop, you know, . .
(VI -4)


173
Carpenter, Charles. 1933. Variation in the Southern Mountain Dia
lect. American Speech. 8. 22-5.
Caudill, Harry M. 1963. Night Comes to the Cumberlands: A Biography
of a Depressed Area. Boston: Little Brown.
Chafe, Wallace L. 1974. Language and Consciousness. Language. 50.
111-35.
Chafe, Wallace L. 1976. Givenness, Contrastiveness, Definiteness,
Subjects, and Point of View. In Subject and Topic, edited by
Charles N. Li. New York: Academic. 25-56.
Chapman, Maristan. 1929. American Speech as Practiced in the South
ern Mountains. Century. 117. 617-23.
Clark, Herbert H. and Susan E. Haviland. 1977. Comprehension and
the Given-New Contract. In Discourse Production and Comprehen
sion, edited by Roy A. Freedle. Norwood, New Jersey: Ablex.
1-4C|.
Coleman, Wilma. 1936. Mountain Dialect in North Georgia. University
of Georgia thesis.
Combs, Josiah H. 1916. Old, Early, and Elizabethan English in the
Southern Mountains. Dialect Notes. 4. 283-97.
Combs, Josiah H. 1931. The Language of the Southern Mountains.
PMLA. 46. 1302-22.
Coulthard, Malcolm. 1976. Discourse Analysis in EnglishA Short
Review of the Literature. Language Teaching and Linguistics
Abstracts. 8. 71-89.
Dane, Frantisek. 1966. A Three-Level Approach to Syntax. In Tra-
vaux Linguistiques de Prague: Volume I, edited by Joseph Vachek.
Tuscaloosa: University of Alabama. 225-40.
DaneS, Frantisek, ed. 1974. Papers on Functional Sentence Perspec
tive. The Hague: Mouton.
Dressier, Wolfgang U. 1970. Towards a Semantic Deep Structure of
Discourse Grammar. CLS 6. 202-8.
Dressier, Wolfgang U., ed. 1978. Current Trends in Text Linguistics.
New York: Walter de Gruyter.
Dykeman, Wilma. 1965. The French Broad. Knoxville: University of
Tennessee.
Dykeman, Wilma. 1977. Tennessee. New York: Norton.


46
The artificial interview situation, however, is, he adds, "the
only natural context for eliciting information from an informant"
(pp. 104-5). In other words, some type of directive elicitation is nec
essary, however willing the informants are to participate, since they
do not always spontaneously volunteer discourse, especially exposi
tions. Given these realitites, the researcher decided to adopt the
role of a college student collecting general information for a research
paper from a cross-section of the people in the town, stating that he
was interested in White Pine because he viewed it as a typical East
Tennessee small town. This is similar to the role of historian Gold
stein advises for folklorists. In this role the researcher states that
he is not interested in official opinions or just in the opinions of
the oldest and/or most respected members of the community, but that
he is seeking instead the ideas and views of a cross-section of the com
munity's average citizens. He emphasizes that he wants responses from
both native and non-native residents, from old and young, from members
of the whole community, whatever their social standing.
The researcher decided to have the interviews in the informant's
home rather than in the contact persons' home. He also discarded the
idea of interviewing in the informant's place of business or at some
other neutral, public site. Most informants readily invited the re
searcher into their own homes, and for this reason and others it ap
peared that informants would be most comfortable there. Moreover, a
fieldworker actually has more control over the interview in the infor
mant's home. There he has greater discretion over the pacing of the


138
24) And you know, that makes you feel very good to see a young
man turn over a new leaf. (IVI-22)
25) The only way I'd vote for Ray Blanton is whether, if they's
voting for him to leave the country, and that's about the
same way with the other Democrats, . (VIII1-14)
26) That's been a question I think in a lot of people's minds
whether it would be feasible to build one here or not.
(1X2-14)
In these instances the deictic is cataphoric or exophoric. It
neither refers or points backward in the discourse, as "that" usually
does. This cataphoric nature of deictics in such sentences is more
clearly seen in five of the extrapositions, as in 27 and 28 below:
27) That would cost too much money, to clean the lines and
things like that. (II14-14)
28) That's kind of a traditional thing for us, just to go up
there and have a big dinner. . (V2-19)
in which we have a pause (represented by the commas) before the
extraposed clause. Although such sentences are sometimes simply
termed "right dislocations" or "afterthought constructions," we can
understand them by considering their discourse contexts. The deic
tics in such sentences are not anaphoric. They point to a situation
outside the stream of discourse. When the speaker realizes he/she
must provide the information to which they broadly refer, he/she
attaches the information at the end of the sentence, after a slight
pause.
Right dislocation is a kind of apposition and represents a
speaker's addition of new information to a discourse context. It is
a widespread strategy found in our corpus, especially with structures
having deictics. In the next section we examine right dislocation with
identifying sentences.


54
are also crucial to sentence, clause, and phrase boundaries. This
is why the tape recording itself remains the ultimate authority.
The transcription employed here makes no attempt at eye dialect.
Orthography was modified only to indicate several minor instances of
deletion such as
1) The voiced interdental in "them" and "that," if deleted, is re
placed by an apostrophe (as in "with 'em" or "like "at");
2) Clause initial "that's" and "it's," often reduced phonetically
to [s] or [s:], and therefore indistinguishable, are represented
frequently as '"s";
3) Initial syllable deletion, as for "of course," represented as
"course," and "because," represented as "cause."
In no instances were such conventions used to indicate the deletion of
morphemes. Dealing with deletions in general was frequently difficult.
Basically any apparent morpheme that had a clear trace was represented
in full orthography. In fact, clear hesitation phenomena such as "uh"
and "hmm," the affirming "uh-huh" and the negation "huh-uh" and the like
were represented whenever audible. But the many standard English words
without such a trace (such as copulas, relative pronouns, and others)
were not represented.
Hesitation phenomena (as well as the parenthetical "you know," "I
mean," and other phrases adding nothing to the content of the interview)
are represented in this study's interview transcripts for the sake of
making the transcript as accurate as possible. It is also possible
that such phenomena might have some discourse significance, as markers
of sentences or of paragraphs (i.e, changes or returns to themes). In
any case, possibly significant data should not be discarded a priori.


167
of, there's good people over there, (uh-huh) Uh it's uh quite a
high crime rate for the size of the county. I mean it's tremen
dous. You just wouldn't believe. I have known, I have known, uh
known 'em to have uh as many as two or three homicides a day over
there, (hmm) and that's a lot. Uh eight or ten a week, and uh you
take a county that small, why, it's tremendous, (hmm) It's just
outrageous, just to put it bluntly, so it's uh really bad. But I,
like I said a while ago, I don't think it's as bad as it was, cause
we got a new attorney general, and he, (hmm) you know, he pretty
well put a little, quietened it down a little bit. [#114-31 is an
interruption of a prospective policeman coming in to fill out an ap
plication]
F: What do you personally do on Pine Day every year?
I: I don't uh, I don't really can't answer that unless, well, we have
a parade and I coordinate it. (uh-huh) Uh it's about the only,
really the only uh parcipi-, uh part that I take in it (uh-huh) is
uh the coordinating the parade and leading and handling the parade,
(uh-huh) Course, like I said, I've only been here a couple, well,
I think I've had, been here two or three Pine Days and really don't
know the meaning of it, to be quite honest with you. But it's a,
it's a relatively quiet, nice little town, uh comparing it with
other towns its size in East Tennessee, (uh-huh) I like it.
F: In the three years you've been here, what would you say have been
the most significant changes in the town?
I: Well, the, I think the most significant change is the, the things
that the city fathers have done, such as uh recreation, uh streets,


81
denouement to the paragraph. Longacre's most recent view of Peak
is an even more refined one:
For discourses that are not narrative but still have a climax
of development, the peak may mark "target procedure" in a
procedural discourse, "climactic exhortation" in a behavioral
discourse of the hortatory variety, and a most satisfactory
or "culminating explanation" in expository discourse. .
(Longacre and Levinsohn 1978:105)
If we look back to the exposition above at the sen
tence therein ("I think this is the greatest change that has happened
in our town is to have new people coming in with new blood, new
thoughts, new ideas, new approach to things"), we see clearly that this
sentence represents the culminating explanation and that it is the
peak of this one-paragraph exposition. Our chapter four below is
largely devoted to showing how such sentences indicate the peaks of
paragraphs.
In the remainder of this study we show the extent to which we can
find that certain surface structure cohesive devices are associated
with grammatical units of expository paragraphs.
At this point we should ask why we find surface structure cohe
sive devices such as thematizing constructions and peak sentences at
all, since their use is not categorical. Left dislocation, as we will
see in chapter three, is employed in only slightly more than one-fifth
of its potential cases. If left dislocation is optional, as are other
thematizing constructions in English, why do speakers use it at all?
Why is it that grammatical units of the paragraph are sometimes
overtly marked by a specific structural reflex? We will suggest here


166
county, I guess they have more crime than any other county in the
United States. Now that's saying a mouthful, (yeah) but you can
uh, you can figure one or two homicides a week. Now, you consi
der the size of Cocke County, that would be one or two an hour in
the city of New York, and, you know, that's I guess about what
they have, so uh it's a very high rate of crime, but I believe
it's what they people wants. The majority of the people, now. Of
course, all of them don't, (uh-huh) and it's, its got a lot to do
with the law enforcement. Uh you don't have the men anywhere, I
don't care whether it's Cocke County, Jefferson County, or what
county it is. You don't ever have the men you need, and uh you
can't watch everybody. You've got eight officers to uh eighteen to
twenty thousand people, then you see what you're up against. I've
got three officers, and uh I've got twenty, well, two thousand peo
ple in the city alone, (uh-huh) so you can see what they're doing.
They're watching eight hundred to a thousand people, each officer
is, (uh-huh) and uh there's no way that you can uh, you can do that.
There's just uh no way at all, so uh that's, I think the biggest
reason they have the crime rate that they do.
F: But as far as I can remember back in Knoxville I've been reading
about strange goings on in Cocke County. You know, senseless mur
ders or
I: Well, they, the people feud among themselves. It goes back a long
ways I suppose. Uh they uh they've just got what you, I'd say
roots that uh entice 'em to do things like this. They, they're
just uh, there's a lot of mean people over there, uh plus a lot


178
Longacre, Robert E. and Stephen Levinsohn. 1978. Field Methods of
Discourse. In Current Trends in Text Linguistics, edited by
Wolfgang U. Dressier. New York: Walter de Gruyter. 103-22.
Malinowski, Brontislaw. 1961[1922]. Argonauts of the Western Paci
fic. New York: Dutton.
Malinowski, Brontislaw. 1923. The Problem of Meaning in Primitive
Languages. Supplement I to The Meaning of Meaning, by C. K.
Ogden and I. A. Richards. New York: Harcourt.
Martin, Walt. 1977. Cultural Regions of Mid-Nineteenth Century
Tennessee. Manuscript.
Mathesius, Vilm. 1964. On the Potentiality of the Phenomena of
Language. In A Prague School Reader in Linguistics, edited
by Joseph Vachek. Bloomington: Indiana University. 1-32.
McMillan, James B. 1971. Annotated Bibliography of Southern Ameri
can English. Coral Gables: University of Miami.
Mencken, H. L. 1963. The American Language. The Fourth Edition
and the Two Supplements. Abridged, with Annotations and New
Material, by Raven I. McDavid, Jr. New York: Knopf.
Merritt, Marilyn. 1972. Punctuating Devices in Discourse: Prelim
inary Notes. Paper read at the 1972 LSA Winter Meeting, Atlanta.
Montgomery, Michael B. 1975. Attitudes of the Southern Mountain
People Toward Language. Manuscript.
Montgomery, Michael B. 1977a. Topics Are Dislocated All Around Us.
Manuscript.
Montgomery, Michael B. 1977b. Subject Relative Pronoun Deletion: A
Discourse Function. Paper Read at the South Atlantic American
Dialect Meeting, November 5, 1977.
Montgomery, Michael B. 1978. Left Dislocation: Its Mature in Appa
lachian English. The SECOL Bulletin. II. 2. 55-61.
Odum, Howard W. 1936. Southern Regions of the United States.
Chapel Hill: University of North Carolina.
Pederson, Lee, Raven I. McDavid, Jr., Charles W. Foster, and Charles
E. Billiard, eds. 1974. A Manual for Dialect Research in the
Southern States. Tuscaloosa: University of Alabama.
Petdfi, J. S. and H. Rieser. 1976. Studies in Text Grammar.
Boston: D. Reidel.


22
Figure 1.2 Campbell's Southern Highland Region
(Source: Campbell 1921:xxxviii)


165
was born raised here. He knew everybody, and I guess it was just
hard for him to, you know, get real, real serious and crack down,
so to speak, (uh-huh) so change has been tremendous, from what I
can hear. Like I said, I'm not from here and the only way I can
answer your question honestly is to have lived here before I came,
or before we took over, so to speak. The new administration had
a lot to do with it. They gave us a little more power and a lit
tle more authority, and uh excuse me [The phone rings here for I]
and uh that helps. You know, you [again the phone rings for I)
F: I wanted to ask you about Cocke County. Why does it have the re
putation it does?
I: My opinion is they don't want to have any other type of reputation
as far as law enforcement is concerned. Uh, it's uh, it's just
wide open. Well, it's not as bad as it was. They've got a new at
1 torney general has done some things that uh I approve of. He's uh
cut out a lot of prostitution, (uh-huh) Not all of it, mind you,
but a lot of it. But I think that uh, course, you know as well as
I do, you have what the people want. Just like White Pine, if
White Pine wanted twenty police officers, they would have them,
(uh-huh) Uh if Cocke County wanted most, there's no way they
could cut it all out, but if Cocke County wanted most of their
vice and uh so- crime cut out, they could. It's a border, state
border county, (uh-huh) state line county, and that, that has a
lot to do with it. Plus the fact that it's a very mountainous
area, and uh a lot of the out of the way places, so to speak, and
they just have a lot of crime and a lot of, for the size of the


78
In contrast, grammatical structure is a variety of surface
structure. For example, "grammatical subject" and "grammatical
object," i.e., subject and complement, are constitutents of gram
matical constructions on the clause level. In levels beyond the
sentence, the grammatical order of constitutents is the actual
surface order in which the narrative is told. Morphemes are po
sited as the minimum grammatical units.
The nature of constituent units in lexemic constructions is ra
dically different from that of constituent units in grammatical
constructions, but the boundaries of the two kinds of construc
tions frequently coincide. However, the boundaries are not always
isomorphic so that the lexemic and grammatical constructions can
not be mapped onto each other in a direct, one-to-one manner.
(1968:2-3)
This kind of mapping can be easily done in many cases. For example, the
lexemic function of the introduction of a participant into a paragraph is
usually equivalent to and realized in the theme and this is further rein
forced by one of the thematizing cohesive devicesleft dislocation in
many cases. The lexemic function of summation is sometimes indicated by
the kind of structure we are calling peak sentences, but this is not al
ways the case. Such sentences always represent the peaks of paragraphs
grammatically but only sometimes are these peaks the realizations of such
sentences.
The third and most discussed way to view paragraph structure is gram
matically. The grammar of a paragraph is the system of units (Longacre's
current term) or constituents (Wise's term) whose order is the actual
surface structure order of the paragraph. Longacre (1968, 1976, Forth
coming) especially has long made the case that paragraphs are grammatical
entities and that each type of paragraph (narrative, explanatory, horta
tory, dialogue, precatory, etc.) has its own grammar. He contends that
the units (in his 1968 work he calls them "tagmemes" in positing his


CHAPTER V
CONJUNCTIONS AND INTRODUCERS
In this chapter conjunctions and introducers are treated together.
In our corpus, conjunctions very often introduce sentences, paragraphs,
and even entire expositions. This is consistent with what Longacre and
Levinsohn point out: "As introducers of sentences, conjunctions serve
to fit the sentence into paragraph context. They also serve to fit
the paragraph into discourse context" (1978:109). Some conjunctions
(especially "and," "but," "then," and "now") so frequently
begin sentences in our oral expositions that clearly one of the
primary functions of such words is to mark sentences. Frequently
conjunctions thus accompany the pauses which divide sentences. In the
following example exposition the sentence-initial conjunctions are
underlined:
1) Fieldworker: What do you think most people who live here want
the town to become?
Informant: Well, I think everybody wants us to have the new
hospital, which we're about on the verge of getting, I think,
which we really need is the new hospital. And everybody
wants it. Just, now White Pine's built up a lot over the
years. Of course, like I say, it's still a small town,
(uh-huh) But just other than that, well, they's not really
no great changes been, to me they's not, I mean, uh to other
people it might seem different. But to me it just seems
like the same, the way it's always been. (X11-14)
In addition to five sentence-initiating conjunctions here, we have two
instances of "well," and one of "of course," other introducers,
154


74
thematic unitsparagraphswhich represent the natural thematic struc
tural units in which speech is used.
Oral Paragraphs and a Review of Paragraph Studies
In analyzing the structure of discourses, linguists have often rec
ognized distinct supersentential units smaller than the discourse it
self. This is the case with Van Dijk (1977) for narrative, Klammer
(1971) for dialogue, and Longacre (1968) for dramatic, procedural, and
other types of discourse. Van Dijk uses the term "macro-structure" for
a sequence of sentences which belong together and which frequently has
"Morpho-Phonological markings" (1977:152-3). He mentions pauses, in
tonation, and particles as examples of such markings. Klammer, Long-
acre, and others adopt the less original but more useful term "para
graph," the traditional term referring to an orthographic unit of lan
guage. Herein we will also use the term paragraph for these interme
diate structural units in oral discourses that are unified in more ways
than simply forming a block of print on the page.
For Longacre, any string of two or more sentences which belong to
gether constitute a paragraph: "Unstructured sequences, or strings of
sentences which do not constitute a paragraph structure, do not exist"
(1968:53). Most often the paragraph has been defined as any self-con
tained thematic unit. Pike and Pike call it "the minimum unit in which
a theme is developed" (1977:488). We consider a paragraph to be any
thematic block of language, i.e. a unit of language organized around
a theme or piece of information.


ACKNOWLEDGMENTS
For their assistance to me throughout my graduate studies and
especially in undertaking the fieldwork for this dissertation I am
indebted to my parents. In many ways this dissertation would have
been impossible without them, and to them I dedicate it.
The fieldwork for this dissertation could not have been carried
out without the gracious assistance of my contact persons in White
Pine, the Atchleys Earl, Emogene, Steve, and Philipwho enabled
me to make the necessary contacts with and to establish a rapport
with my informants. To them I will always be grateful.
I also wish to thank the people of White Pine for their hospital
ity. Their open-hearted welcome and open invitations to return will
long be appreciated.
To the chairman of my committee, Professor William J. Sullivan, I
am thankful for patient direction and many invaluable suggestions.
To the other members of my committee, Professors Jean Casagrande,
Robert Thomson, Roger Thompson, and Norman Markel, I am grateful for
patient tolerance of an erratic and hurried submission of my work.
My colleagues at the Department of English at the University of
Arkansas at Little Rock deserve thanks for their encouragements to
finish this work this past year.
Finally I wish to thank Professor Robert Longacre of the Summer
Institute of Linguistics for his invaluable personal advice to me.
iii


36
5) What is the linguistic context of the form? (Where in the sen
tence and in the discourse does it appear?)
Only when we have answers to all these questions can we establish what
context a form appears in and whether it is typical verbal behavior.
The history of the study of Applachian English has been chiefly a
history of the study of extreme forms, and only rarely is any context
noted for data which are cited. Despite an occasional remark that a
certain form "is still heard" or is "generally" or "frequently" used,
the clear implication of most studies is that all the speakers in the
community under consideration use the forms cited all the time and in
every social situation.
The tendency to deal with the atypical (and hence unrepresenta
tive) speakers and forms infects even works which are otherwise linguis
tically sound. This temptation is very strong. Joseph S. Hall's tho
rough and invaluable study of the phonetics of the speech of the few re
maining residents of the Great Smoky Mountain National Park was under
taken, he confesses, because he believes in this speech, which "has so
long been removed from the main currents of American culture, there would
remain vestiges of earlier stages of the growth of the English language"
and which also would not have "the deep impress of the schoolmaster's in
fluence" (1942:4). Brewer and Brandes (1977) present only the extreme
features of Appalachian English for the benefit of schoolteachers who
will teach in Appalachia.
Two recent sociolinguistic studies, Hackenberg (1972) and Wolfram
and Christian (1975, 1976), take into account the first two contexts
(who uses the form and how often is it used?), but neither controls the


26
The contention in these excerpts is that the speech of Southern Appa
lachian people has a marked flavor and character all its own, beyond
the presence of archaic expressions and words therein. Commentators
have long been at pains to account for this on the basis of one extra-
linguistic factor or another (isolation, pace of life, etc.). Cratis
Williams, lifelong resident of the mountains, attributes in part the
distinctiveness of the way people in the region speak to "the highland
er's habit of thrusting his chin forward rigidly when he speaks."
(1961:10) Brewer and Brandes try to explain why Southern Appalachian
English impresses outsiders as so different by describing the kinesics
of its speakers, detailing the posture and the eye and body contact of
the typical speaker in the region.
Almost all the studies and comments on the speech of Southern Ap
palachia have had the unfortunate effect, however, of lumping the en
tire region together linguistically and assuming that it is homogeneous
within itself as well as distinct from other regions of the country.
To this date, very little attention has been devoted to regional lin
guistic variation within Appalachia. Only the two linguistic atlas
projects, the Linguistic Atlas of the Middle and South Atlantic States
and the Linguistic Atlas of the Gulf States, have given significant
consideration to such variation. This presumption of underlying uni
formity has been espoused frequently, as in Carpenter (1933). He con
tends that the only "variation in the vernacular of the southern and
central Appalachians has been due to the difference in the degree of
isolation of the various sections" (p. 22). Most writers ignore the


141
it's not as, it's not really you that's coming. It's that
invitation. I think it's uh more important, and it helps
people around here more than it would a, you know, in
other places, because then we're, we're closer and a lot
of people feel the loss. You know, it's not just the fam
ily. I've never been to any funerals but (hmm) that's
just what it seems like to me. (XI-18,19)
In this exposition, the deictic "that" in the last sentence points
back to and sums up the entire preceding discourse. It also signals
to a hearer that an end of a whole or part of a discourse has been
reached. In the cases which have left dislocation from an identi
fying clause, discussed above on page 134, an especially strong
note of finality is conveyed. Since identifying clauses with
deictics are frequently used to conclude paragraphs, left disloca
tions from identifying clauses often represent exceptionally short
and abrupt paragraphs.
The identifying sentences in our corpus differ in two important
respects from what Halliday says about them, however. First, Halliday
claims that only in contrastive cases or when the deictic is exophoric
(referring to something in the speech situation, outside the text) is
the deictic stressed. In our data, the deictic, although usually
anaphoric, does receive focal stress. Focal stress is not as strong
as contrastive but it is quite obvious, especially with the instances
involving left dislocation.
The second exception to Halliday's characterization of identifying
sentences is that the deictic cannot be cataphoric. Although the
identifying sentences we will now examine briefly differ from the
twenty instances in which'"that" identifies something in the preceding
discourse with a WH-clause, they are identifying sentences in function
and according to Halliday. These sentences involve a right dislocation,


95
call these "it-clefts" herein); and
7) Conjunctions (six different types: summary, contrastive,
enumerative, continuation, comparison, example).
These devices are employed by authors, according to Jones, to give
special prominence to (and not necessarily to introduce) the theme of a
paragraph and they may be found anywhere in a given paragraph. The
thematic devices we focus on in this study introduce new themes into
discourse as well as give prominence to them. They are always found at
the beginning of paragraphs.
Jones devotes most attention to showing how rhetorical questions,
cleft sentences, and pseudo-cleft sentences highlight paragraph themes.
It appears that these and the other devices she discusses are devices
found primarily in written expositions and thus we should expect to
find a different set of devices in oral ones. In a selection of 200
expositions from the White Pine corpus (five from each informant), we
find only three it-clefts and only four WH-clefts. All seven instances
introduce new information but none establishes an anaphora chain. We
can hardly generalize, however, on the basis of so few instances.
Jones does not mention existential sentences as one device pre
senting themes, and she only makes a very brief mention of left dislo
cation as a device which "directs special attention" to an element
(p. 172). Even the three devices she gives special attention to are
found to be rather uncommon in written expositions: "Except for con
junctions. . the other devices all occur rather infrequently in


97
Put simply, left dislocation is a construction in which an NP or
noun phrase (simple, complex, or compound) appears before the begin
ning of a sentence (i.e. leftmost) and has an appropriate pro-word
copy in the sentence itself, as in 5-7 below:^
1) And my boy, he froze so bad, he had what they call "buck
fever." (V3-27)
2) Well, these arbitraty rules that people set up for themself,
that's something else I don't like, you know. (113-32)
3) Like last year, me and a couple or few of my friends uh, we
watched the queen and everything out there. (XI2-8)
Left dislocation is not a repair construction which represents a
speaker's change of mind and a second attempt at phrasing a thought
after beginning a sentence. It is very rare to find any sort of au
dible hesitation or an overly long pause, for instance, between a dis
located NP and the sentence which follows, but a short, discernible
pause does characteristically occur. The intonation contour on left
dislocated NPs, no matter how long, is very consistently level with a
slight fall at the end for our data. In another study of left dislo
cations, Keenan and Schieffelin found that the NPs were "uttered with
a slight rising intonation. .This is then often followed by a pause
or by a hesitation marker (e.g. uhh)" (1976:254). Our findings differ
from theirs on both accounts.
The Variety of Forms
An NP in any syntactic position may be dislocated, whether it is
in the matrix or main clause or in an embedded clause. Most often


153
not always employ a cohesive device, which makes a speaker's message
"stick together," to signal a peak, but he/she sometimes does.


98
(356 of the total 606 left dislocations in the data, or 58.7 percent),
as in 5-7 above, it is the matrix subject which is dislocated. This
is not surprising. Halliday indicates that in "the declarative, the
subject if the unmarked theme" (1967:213). He defines theme as that,
in a sentence, which "is being talked about" (p. 212). Therefore we
should expect themes, which information blocks are about, to most repre
sent subjects of sentences. -
But the NP may represent the subject of other clauses, such as an
introductory adverbial one (21 instances, or 3.5 percent), as in 8 be
low, of an object clause (36 instances, 5.9 percent of the total), as
in 9-10 below, or of a relative clause (9 instances, 1.5 percent of the
total), as in 11:
8) Children., if they. took a cold, why they'd make 'em, some
boneset tea. (II2-34)
9) Just anything, even the commercials on the, on the television,
we thought they were great. (VI3-8)
10) And the newest hospital, I think they were running short of
funds at one time, (XIII-8)
11) And then Jake Butcher, you know the bank that he's the head of.
(11-18)
In thirteen instances, the subject of an object clause is dislocated to
the front of the clause and not of the sentence, as in 12:
12) And I reckon the man that made it, he thought he couldn't make
a living of it. (IV2-13)
Left dislocation thus can involve the subject of any clause in a sen
tence. It is just as wide-ranging for objects of verbs, although it


24
infer that its inhabitants are alike in ancestry and cultural back
ground, especially on the basis of a number of linguistic relics
apparently distinctive of its people. The most extreme forms of such
inferences issue an assertions that Appalachian people are "Elizabe
than" or that they preserve the "purest Anglo-Saxon blood in the United
States" or that they are "almost pure Scotch-Irish" (note the contra
diction here). These inferences have formed a part of the chronic
misconceptions about the inhabitants of Southern Appalachia.
Even such a serious commentator as Campbell has abetted such
inferences by his statement that Southern Appalachian people form "a
definite . racial group" (1921:xiv), although he conceived of
their having developed into such a distinct group of people by virtue
of "their common interests, hardships, and struggles" (p. 71) and not
by virtue of their lineal descent from one stock of Old World people.
His lengthy study of the surnames of early mountain settlers revealed
almost equal proportions of Scotch-Irish and English names, and a
large number of Germans as well:
Without a doubt the Scotch-Irish and English elements are the
strongest in the mountain population, though the Highland
people are not different from the Lowland Southerners in this
respect. The Scotch-Irish strain is strongest in some mountain
sections, the English in others; and in some communities may be
surmised an influence of German ancestry, (p. 71)
Several other groups, most especially the French Huguenots, were numer
ous among the early settlers in Appalachia too. Thus the region's
present-day descendants are a hybrid, whencever came their language (a


A DISCOURSE ANALYSIS OF
EXPOSITORY APPALACHIAN ENGLISH
BY
MICHAEL BRYANT MONTGOMERY
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1979


128
Existent!ais have been widely recognized in the literature as
by Bayless and Johnson, as signaling "that new information is being
introduced without directly relating it to some old information"
(n.d.:37). But this function and the function of indicating that
something does or does not exist seem to be the only basic functions
of existentials. They do not possess the range of functions that left
dislocation has. Existentials are not used to indicate a contrast,
nor do they emphatically reintroduce an NP back into a discourse. Oc
casionally an existential will specify a subgroup of an NP which has
been previously mentioned (cf. left dislocation, pp 111-2 above), as
in 61 below:
62) The only opposition I would think uh that would come from
within this town to any one thing I guess would be maybe
opposition to alcohol. There'd be some opposition there.
TTTT2-26)
In two instances of the 320 an existential is repeated, as in 63 below,
to reinforce the effect of the first one:
63) . .there's more difference in the atmosphere, the people
and all from here to there than there was from, when I went
to Philadelphia. Yes, there really was. . (VI3-5)
It was indicated earlier that eleven percent of our corpus' left
dislocations occur at the beginning of an informant's response. Be
cause existentials also introduce pieces of information and partici
pants into discourse we should expect to find many of them occurring
exposition-initially also. Of the 200 selected expositions, 21 (10.5
percent of the total) begin with an existential. This is more than
begin with a left dislocation (16 of the 200, or 8 percent).


I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation and
is fully adequate, in scope and quality, as a dissertation for the
degree of Doctor of Philosophy.
Mho*
William Sulliv^h,'Chairman
Associate Professor of
Linguistics
I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation and
is fully adequate, in scope and quality, as a dissertation for the
degree of Doctor of Philosophy.
yean Casagrate
Associate Professor of
Linguistics
I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation and
is fully adequate, in scope and quality, as a dissertation for the
degree of Doctor of Philosophy.
I'U- V /-MVU
Roger ff. Thompson
Associate Professor of
Linguistics
I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation and
is fully adequate, in scope and quality, as a dissertation for the
degree of Doctor of Philosophy.
Norman N. Markel
Professor of Speech


156
In our data the presence of an introducer of some kind at
the beginning of expositions and paragraphs is typical. What
typically introduces our expositions is "well." We cannot claim
that it more commonly begins expositions than other kinds of dis
course or that its primary function may not be a general one to be
gin any conversation turn. But its great frequency (114, or 57 per
cent of the expositions examined begin with it) indicates that it is
a speaker's usual way of getting a monologue exposition underway.
In our data only two other introducers are used with any frequency at
all to begin discourses: "of course" (10 instances) and "now" (7 in
stances).
In chapter three it is shown that left dislocated NPs typically
begin paragraphs. If we look at the 190 left dislocations in the
200 selected expositions, we find that conjunctions typically accom
pany such NPs. In other words, conjunctions are frequent cues to
new paragraphs. Only 39 (20.5 percent of the total) begin with no
conjunction or other introducer. In such instances only the pause
accompanies the NP. Of the remaining 151, 60 left dislocated NPs
are preceded by either "and" or a combination of "and" and other
conjunctions or particles. Twenty-five are preceded by "but" or a
combination with "but." Seventeen NPs have "well" by itself or in
combination. Ten are preceded by "of course" by itself or in a
combination.
Twenty-nine of the dislocated NPs are preceded by a hesitation
"uh," either alone or with a pronoun. It is indicated in chapter
three above that left dislocation rarely involves a hesitation


CHAPTER II
BACKGROUND AND METHODOLOGY
Preface
The first chapter of this dissertation presents the non-linguistic
context of this study; this chapter discusses its linguistic context.
In the first part of the chapter several approaches to the analysis of
discourse structure are reviewed to put the analysis of oral expository
monologues into perspective. Then the methodology of this dissertation
is explained. The second section examines the importance of studying
monologue expositions. Then some remarks are made regarding the nature
of oral paragraphs, followed by a review of relevant studies of oral
and written paragraph structure. Finally Longacre's typology of dis
course genres is discussed.
The control of the five different contexts mentioned in chapter
one (speaker type, frequency of occurrence, social situation, task, and
linguistic position) makes this study a novel one. Also, as far as
this writer can determine, this study is the only analysis of the dis
course structure of oral expositions for any variety of American Eng
lish. Although much has been written about discourse structure in Eng
lish, oral expositions have been seriously studied only for other lan
guages (see especially Longacre 1968). This dissertation represents
the first time that the task of eliciting, analyzing, and then gener
alizing about a large corpus of expositions has been undertaken for
English. It therefore is an exploratory project.
61


119
No. of Sentences No. of NPs Thematic Percentage of Total
Beyond Introduction To This Extent
For Which NP Is
Thematic
1
61
32.1
2
32
16.8
3
22
11.6
4
17
8.9
5
14
7.4
6
6
3.1
7
5
2.6
8
8
4.2
9
7
3.7
10
5
2.6
10+
13
6.9
190
100.0
Figure 3.1 Extent To Which Left Dislocated NPs Are Thematic


58
almost certainly not have been so successful without the assistance
of the contact persons in White Pine.
Informants for this study were assured that their names would
not be revealed and that they would remain anonymous to all but the
fieldworker.
Taping of the Interviews
The interviews for this study were recorded on a four-track, reel-
to-reel Voice of Music model 733 Tape o Matic tape recorder at 3 3/4
inches per second. The tapes used were Scotch No. 207, 1800 foot
tapes, which, at 3 3/4 inches per second, allowed up to ninety minutes
of recording per track. In all eleven reels of tape were required.
Because of logistical difficulties, two interviews had to be re
corded on cassette tapes. These interviews are coded as tapes XII and
XIII in Appendix A.
The Value of Fieldwork
Since the ultimate value of doing linguistic fieldwork is to help
us better understand how people employ language, the fieldworker must
always remember that the interview is a very human event involving di
rect contact between two or more people. As Goldstein declares, "the
problem of fieldwork is essentially human" (1964:22), and more than any
thing else fieldwork is a venture in human relations. The fieldworker
cannot be detached from his informants, especially when subjects as
deep-felt as some of those covered in the present study's interviews


terms and that it should not be tied to old (given) information or to
a specific syntactic position, she defines theme as the "main idea" or
"central thread" or "nuclear constituent" at each level of language or
ganization, especially at the paragraph level (pp. vi-vii). Deriving
from Pike and Pike (1977), her view of theme is more intuitive than
ours. Although she believes each paragraph has a theme, she takes it
for granted that the written paragraphs she examines are thematically
unified and does not show how an anaphora or lexical chain identifies
a paragraph's theme. In our view, such a chain is the best evidence
for what a paragraph's nuclear constituent is.
Jones' primary concern is the examination of many "grammatical de
vices for highlighting theme" (pp. 169-220). Such devices which often
convey the theme of the written expository paragraphs she considers in
clude
1) Repetition of a word, phrase, or sentence;
2) Inversion of declarative word order;
3) Special constructions, including passive sentences and sentences
with an initial "as for. phrase;
4) Rhetorical questions;
5) Pseudo-cleft sentences (as "What is unique about milk is its
richness in vitamins and minerals"; we refer to such sentences
as "WH-clefts" herein);
6) Cleft sentences (as "It is the very mingling of races, dedicated
to common ideals, which creates and recreates our vitality"; we


152
Underlining is probably the most frequently used. (1976:228-9)
(According to Longacre (1976:217-8), rhetorical underlining comprises
"parallelism, paraphrase, and tautologies of various sorts" used by
an author to ensure that a reader or hearer does not miss a point.)
Peak then comes near the end of and represents a kind of climax to an
expository discourse or paragraph. In Longacre and Levinsohn (1978:
105) we find Longacre's most recent views on peak. Again he indicates
that only discourses which have tension have peaks:
If a discourse is plus tension, there will most likely be some
kind of climax of development, some marked surface structure
"peak". . For discourses that are not narrative but still
have a climax of development, the peak may mark "target pro
cedure" in a procedural discourse, "climactic exhortation"
in a behavioral discourse of the hortatory variety, and a most
satisfactory or "culminating explanation" in expository
discourse.
An examination of any of our discourses in which a speaker uses a
peak sentence reveals that precisely what a peak sentence does is
present the most satisfactory or culminating explanation regarding
a topic.
For a speaker to accomplish the task of adequately explaining a
subject, it is critical that his/her climactic or culminating point
get across to a hearer. Information therefore which represents the
peak or consummation of an expository discourse, sometimes of several
hundred words in length, must stand out to a hearer. A speaker must
make sure that the significance of such information is conveyed, even
if it must be explicitly pointed out. It is the pointing out of it
which the particular kind of peak sentence studied in this chapter
does, through the use of deictics, a cohesive device. A speaker need


99
is not as frequent. For example, in sentence 13 the object is dis
located from the matrix clause and in 14-5 from embedded clauses:
13) And whatever's to do, she takes it. (VI14-19)
14) Even public records, there's no place to display 'em. (14-14)
15) And my mother, I mean I've heard her tell lots about Mr.
Lawson Franklin, (III2-2)
In the data are 53 instances (8.7 percent of the total) of matrix ob
ject dislocations and 14 instances (2.3 percent of the total) of embed
ded object dislocations.
Left dislocations can also involve the object of a preposition
from the matrix clause (36 instance, 5.9 percent of the total), as in
16 below, from an object clause (8 instances, 1.3 percent), as in 17,
or an introductory adverbial clause (1 instance, 0.2 percent), as in
18:
16) My first car I had, I done most all the work on ft. (113-5)
17) All that white lightning and everything, you know they got
into vt somehow. (VII3-23)
18) And our husbands, if it wasn't for them, there's no way. (V1 -10)
This various display of forms illustrates that left disclocation can
involve NPs in many positions (more are given below). But it is not
meant to imply an essential difference between matrix and embedded
clause NPs with regard to dislocation. Indeed, such a distinction is
sometimes misleading because examples such as 8 above can be viewed as
dislocations from either the matrix or the introductory adverbial
clause. Those like 10 above have matrix clauses such as "you know,"


71
and attention is given in this study to variation in cohesive devices
between subgroups of the informant sample, according to either the
age, the educational level, or the socioeconomic background of the in
formants. Except for only a few scattered remarks (regarding the in
troducer "like," for example) nothing is said about differences between
individual or groups of speakers. If White Pine had been presumed to
be a homogeneous speech community, the care to interview a large and rep
resentative cross-section would not therefore have been taken. It was
found that the discourse devices under consideration in this study
showed inconsequential differences in function across the entire sample
of forty informants. None of the features were restricted to one subset
of the sample. The generalizations in this study resulted from their
obviously consistent functioning for all forty speakers.
This is not to say that no variation occurred in the frequency of
the realization of such devices as left dislocation. It was realized in
at least three instances for each speaker. But one speaker, a seventy-
year-old woman, had seventy instances. Hers was, however, by far the
longest interviewnearly three hours. Neither for the four informants
not native of Tennessee nor for the one native to Great Britain did left
dislocation function differently from White Pine natives. Peak sen
tences, with fifty instances, were used by twenty three speakers of many
different educational levels and socioeconomic backgrounds and thus dis
tribute broadly throughout the sample. Other devices may have varied in
frequency too. But none vary in function. More important, there is
little to suggest that any of the devices considered in this disserta-


143
then that's where I wet m.y husband, in Morristown. (Vl-1)
38) Now that's something they've done so much here, housing
development. That's what's turned White Pine around, that
low-rent housing development area, and, and uh, and uh, the
apartment complexes. (11-34)
In such instances, the appositions are exact or near exact repetitions
of something previously mentioned. "In Morristown" is nothing more
than an afterthought here. "That low-rent . complexes" is also
unnecessary for the discourse to be well-formed.
We may tentatively hypothesize than that identification sentences
with WH-clauses are anaphoric and that those without are cataphoric
to right dislocated or appositional information at the end of the
sentence. Why this might be so is not clear, but further evidence
of this distinction will be found in the next section's discussion
of peak sentences, to which we now turn.
Peak Sentences
We come now to a kind of structure found in our corpus which has
not been described in the linguistic literature to date. This struc
ture will be called a "peak" sentence and five examples of it are
39-43 below:
39) That's another thing we have through here's the Appalachian
Trail. (V3-4)
40) That's all he's ever got done in Washington is politics.
(VI2-20)
41) But that was the reason I gave was that I would like to go
to vote and give an opinion. (VI12-10)
42) That's the best way I ever found to think is just to walk
back up through the woods or something like that . .
(VII12-10)


20
The foregoing does not strictly hold for the Pentecostals, who do
consider themselves to an extent to be a people apart. The extremes
of Pentecostal ism in Southern Appalachia have long been pointed out
(as in La Barre's study (1964) on the snake-handling cult). One can
find congregations of "snake handlers" not more than twenty miles east
of White Pine in Cocke County. But people in the area overwhelmingly
view such practices as quite objectionable and offensive.
The primary functions of the church itself are to provide relig
ious services and training for its members, yet in a community where
such a large percentage of the inhabitants are members, churches inevi
tably offer more than the fulfillment of spiritual needs. Churches in
White Pine provide most of the organized social activities. Scheduled
weekend activities are relatively infrequent in White Pine, but any
given activity is more than likely a church-related one. Since 1975,
when the county's high schools were consolidated, the only organized
youth activities in White Pine are those that are church-sponsored.
In short, then, the influence of the organized religion in White
Pine is pervasive. The churces are viewed as having an indispensable
and rightful place in the life of the corrcnunity and in the life of the
individual.
The Southern Appalachian Region
Alghough the Appalachian Mountains extend thirteen hundred miles
from Vermont to Alabama, the region usually called "Appalachia" and


82
that such devices occur because they serve certain of a speaker's
pragmatic purposes.
The speaker's primary purpose in exposition is to convey new infor
mation to the hearer. For this reason the speaker/producer of a dis
course or text takes into account the presuppositions and expectations
of the hearer. The hearer expects the speaker to supply the necessary
antecedents and background for him/her to follow what is being said. In
other words, they have what is referred to as the
"Given-New Contract" of cooperation. The speaker must be aware of what
the hearer knows, which information has already been given to the hearer,
and which information will be new to the hearer. The hearer must make
the effort to follow the progress of the exposition.
It follows that any means which facilitates the fulfillment of this
contract and which makes an exposition easier for a hearer to follow may
well be used. While such thematic devices as left dislocation can appa
rently be used also for other pragmatic reasons, basically they seem to
facilitate this given-new contract between the speaker and the hearer.
Thematic devices present to the hearer what new information will be com
mented on in an ensuing message. This piece of information represents
what will be thematic for what follows and has the potential for being
the antecedent for the following block of discourse. It is only if the
hearer has such a piece of information, the theme, clearly in mind that
he/she can compute what the speaker says.
It is probable that the use of cohesive devices is especially fa
vored by a speaker in certain pragmatic situations. Left dislocation,


BIBLIOGRAPHY
Bailey, Charles-James N. 1968. Is There a Midland Dialect of
American English? ERIC 021 240.
Bayless, Richard L. and Linda K. Johnson, n.d. Function of Surface
Rearrangement Rules in English Discourse: Old and Mew Informa
tion. Manuscript.
Becker, Alton L. 1965. A Tagmemic Approach to Paragraph Analysis.
College Composition and Communication. 16. 237-42.
Becker, Alton L. 1966. Symposium on the Paragraph. College Compo
sition and Communication. 17. 67-72.
Berrey, Lester V. 1940. Southern Mountain Dialect. American
Speech. 15. 45-54.
Bicentennial Committee of Jefferson County, Tennessee. 1976. Heri
tage Jefferson County.
Bloomfield, Leonard. 1933. Language. New York: Holt.
Bond, George Foot. 1939. A Study of an Appalachian Dialect. Uni
versity of Florida thesis.
Bray, Rose Altizer. 1950. Disappearing Dialect. Antioch Review.
10. 279-88.
Brewer, Alberta and Carson. 1975. Valley So Wild: A Folk History.
Knoxville: East Tennessee Historical Society.
Brewer, Jeutonne and Paul D. Brandes. 1977. Dialect Clash in Amer
ica. Metuchen, Mew Jersey: Scarecrow.
Bridgeman, LoraineHall. 1966. Oral Paragraphs in Kaiwa (Guarani).
Indiana University dissertation.
Brinegar, Bonnie Carter. 1977. Analysis of Six Sequence Signals of
English Discourse. University of Southern Mississippi disserta
tion.
Brown, Calvin. 1889. Dialectal Survivals in Tennessee. Modern Lan
guage Notes. 4. 205-9.
Campbell, John C. 1969[1921], The Southern Highlander and His Home
land. Lexington: University Press of Kentucky.
172


45
Fourth, he had to determine the type of discourse to elicit and how
to devise a standard interview format which would best elicit it.
The ethnographic parameters were adopted according to Malinowski's
suggestions. He states that a fieldworker should make every attempt to
be an active participant rather than a passive observer-partial partic
ipant in the culture he is studying. He draws on his own experience
among the Trobrianders in showing why this is necessary:
There is all the difference between a sporadic plunging into the
company of natives, and being really in contact with them. What
does this latter mean? On the Ethnographer's side, it means
that his life in the village, which at first is a strange, some
times intensely interesting adventure, soon adopts quite a nat
ural course very much in harmony with his surroundings. Soon
after I had established myself in Omarakana, I began to take
part, in a way, in the village life, to look forward to the im
portant or restive events, to take personal interest in the
gossip and the developments of the small village occurrences; to
wake up every morning to a day, presenting itself to me more or
less as it does to the native....As I went on my morning walk
through the village, I could see intimate details of family
life;...I could see the arrangements for the day's work, people
starting on their errands, or groups of men and women busy at
some manufacturing tasks. Quarrels, jokes, family scenes,
events usually trivial, sometimes dramatic, but always signifi
cant, formed the atmosphere of my daily life, as well as of
theirs. (1922:7)
This researcher, of course, is neither a native nor a resident of
the White Pine area. He was a stranger to his informants initially.
At the same time, he is a native of East Tennessee and especially that
he is a relative of long-time White Pine residents enabled him to estab
lish immediate rapport with his informants. It was also recognized
that the interview situation would necessarily be an artifical one.
Goldstein defines an "artificial situation" as one in which an informant
provides information "at the instigation of the collector" (1964:82).


96
texts. Indeed, the devices would lose their strong impact if they
occurred with great frequency. It is their very occasional usage that
makes them espeically forceful in indicating thematic patterns'1 (p. 220).
Although the devices she discusses are similar in function to the the
matic devices we deal with here, the differences in kind and in fre
quency for such devices probably indicate that strategies employed in
oral expositions differ from those used in written expositions. The
great frequency of such a thematic device as left dislocation(606 in
stances for our corpus) indicates a more prominent role for thematic
devices in era! than in written expositions.
The present study differs from other treatments of theme in being
a data-based study of the nature, function(s), and distribution of the
matic devices in oral expository English. For the remainder of this
chapter we examine two such devices, left dislocation and existentials,
and show how they mark a paragraph's theme and indicate to a hearer
what a paragraph will be organized around.
In several cases the analysis in this chapter is based on a selec
tion of 200 expositions (five from each informant) and in other cases
on 80 expositions (two from each informant).
Left Dislocation
Although this chapter's primary concern is showing that left dis
location is a thematic device rather than displaying its various forms,
such a display yields strong evidence that left dislocation is properly
understood in discourse rather than in sentential terms.


37
other three in their investigations. Wolfram and Christian's study,
based on interviews made in the mountains of southern West Virginia,
classified their speakers socioeconomically and carried out quantita
tive analyses of many of the linguistic forms they discuss, thus pro
viding the speaker and frequency contexts for their data. Yet they
do not choose to base their work on a representative sample because
they were "primarily concerned with the language variety which might
be considered most divergent from some of the more mainstream varieties
of English" (1976:10). Their work is also written primarily for the
educational community and one-fifth of their book deals with the educa
tional implications of dialect diversity.
Hackenberg divides thirty nine residents of Nicholas County, West
Virginia, into four classes on the basis of education level attained and
occupation and quantifies the occurrence of "non-standard" verb agree
ment and subject relative pronoun deletion for his four classes. He
thus takes the first two contexts into consideration and even gives some
attention to the fifth context mentioned above, the linguistic context,
for his two linguistic features. The present writer demonstrates else
where (Montgomery 1977b) that his treatment of the linguistic context is
too simplistic.
A description of speech is ethnographically justifiable only if all
five contexts are controlled and taken into account. Only then can it
be claimed that the behavior described is typical (see the discussion of
the Malinowski/Firth "context of situation" later in this chapter).
This present study has striven to control all five. Thus it makes a
unique contribution to the study of Appalachian English.


72
tion are diagnostic of any of the speakers. A closer look might re
veal some kind of variation, but that would be the object of another
study.
The Study of Monologue Expositions
Linguists interested in supersentential segments of speech have
only recently heeded Firth's injunction of nearly a half-century ago
that "neither linguists nor psychologists have begun the study of con
versation; but it is here we shall find the key to a better understand
ing of what language really is and how it works" (1957a; 32). Only if it
is recognized that the function and manifestation of language is ba
sically conversation can language best be understood. Recent studies
on turn-taking, as by Sacks, Schegloff, and Jefferson (1974), recognize
the importance of studying blocks of conversational dialogue and have
shown that they have cohesion, just as other discourse blocks do.
Klammer (1971) is a study of written dialogue paragraphs and is also an
important contribution in this respect.
The present study does not take into account turn-taking phenomena
between the fieldworker and the informants. But it is consistent with
the conversational frame of reference for studying language. It does
not artificially cut up conversation and thus de-contextualize the
monologue discourses it analyzes. It views expository monologues as
parts of conversational exchange, the context for which was controlled
(as indicated in chapter one) by the fieldworker. Pike and Lowe see
monologue as occurring when "one individual of the cast stays fixed


162
Although this study is one of Appalachian English, its findings
are not claimed to be exclusive to it. Further investigations into
American English with the approach taken here should be carried out.


23
been considered in its heart. Further reason for designating White
Pine as an "Appalachian" community is presented below.
The linguistic analogue of Southern Appalachia is the South Mid
land area (Kurath 1949, Kurath and McDavid 1961), which encompasses
the mountain districts and also most of Tennessee and considerable por
tions of Mississippi, Alabama, and Georgia. This area, determined by a
combination of lexical, phonological, and morphological isoglosses, has
been traditionally viewed as more closely related linguistically to the
North Midland area (Pennsylvania and parts of Maryland, Delaware, and
New Jersey) than to the South.
The question of the relative affinity of linguistic and other
traits in Southern Appalachia (or the South Midland) to regions north
and south is still very much unanswered, in fact, but it will not overly
concern us here. In a provocative reassessment of the linguistic dis
tinctiveness of the area, Bailey (1968) concludes there is no "Midland"
area at all which has a core of distinctive features and that North Mid
land should be retermed 'Lower Northern' while South Midland be redesig
nated 'Outer Southern.' Gastil views the cultural affinities as
lying closer to the South: "Culturally the mountains are an extreme
version of the Upland [South]" (1975:194).
All this is not to say, of course, that even the most restricted
definition of Southern Appalachia implies that it is socially homogene
ous or that its people have a common background because they now may
share certain cultural or linguistic features or even a common outlook
on life. Yet commentators on the area, both native and non-native,


133
is generally less clear and much broader than it is for pronouns. None
of the three deictics in the examples above, for instance, clearly
refer to a specific NP or other element. More important than the
referential function of deictics is the function of pointing, of giving
the hearer directions in an ongoing discourse. Deictics serve primarily
to keep a discourse going, by pointing forward or backward, and they
play a large part in the organization of the expositions in our corpus.
The four structures we will examine are the most interesting involving
deictics but represent only a few which are important in organizing
ongoing expositions.
The other functions of deictics derive from the basic function to
organize a discourse. One of the most important of these is to signal
to a hearer a conclusion or summary statement. A speaker will often
use a sentence with "that" to indicate the end of a paragraph or dis
course has arrived, as in the following example:
4) My mother was born in Corbin, Kentucky. She was born in Corbin,
Kentucky. Uh, at a very early age, they moved to Kingsport,
and uh I don't know why, never asked. I didn't ask the reason
why, but they moved to Kingsport, and of course she was raised
in Kingsport. And of course that's where my father met her
was in Kingsport. And of course, when they, since they've
been married now for, well, he's lived down here now thirty,
it'll be thirty years in March. Same time when they all moved
down here. We all moved together (un-huh) thirty years ago.
That's a long time, making me feel old. (113-3)
In the two extended examples in chapter two above (pp. 69-70), we
also find deictics having this function. Such deictics mark the ends
of thematically self-contained units.
The third and fourth structures we will examine, identifying sen
tences and peak sentences, frequently display this summarizing function.


76
or too sparse to have developed well-tested tools to examine them.
Bridgeman's pioneering work on Kaiwa claims that oral paragraphs in
that language have one of a set of six phonological onsets and one of
three phonological codas, with an optional peak preceding the coda of
a paragraph. Lehiste has attempted to find, with as yet little suc
cess, the phonological correlates of paragraph boundaries by measuring
pausing, lengthening, and pharyngealization at the beginning and end of
sentences and paragraphs.
The isolation of the phonological features of paragraphs is the
subject for another dissertation. Only occasional observations regard
ing phonological matters are advanced in this study, when these features
are concomitant with the surface structure cohesive devices that we
analyze. More than likely a characteristic paragraph-initial lengthen
ing and/or pausing will eventually be shown to aid in giving paragraphs
coherence. One suspects, however, that no phonological signal will cat
egorically mark some part of a paragraph's structure. What one should
expect to find is that a significant coincidence of phonological and
grammatical signals co-occur at thematic breaks in a discourse. In
other words, phonological paragraphs should be to a great extent iso
morphic with grammatical and lexemic ones.
Paragraph structure can also be viewed as lexemic, involving how
the semantic resources of the language organize information into a the
matic unit for a given kind of discourse. Although a principled means
of approaching the lexemic paragraph has yet to be satisfactorily de
vised, quite a few lexemic functions which participate in paragraph


144
43) That's the biggest change that's been made since I remember
living over here is that shopping center. (X2-21)
Such sentences appear to be fusions of two identifying clauses. In
39, for instance, "That's another thing we have through here" appears
to be an anaphoric identifying clause which is fused with another
thing we have through here's the Appalachian Trail." By itself, the
first part seems to be anaphoric to something previously mentioned.
Combined with the second, it is not clearly anaphoric or cataphoric.
In our corpus there are fifty instances of such sentences in the
expositions of 23 different speakers. Forty-three begin with the
deictic "that," five begin with "this," and two begin with "it" and
will be considered with the forty-eight beginning with deictics. An
example with "this" and one with "it" follow:
44) This is what I do not like about the low-rent housing in
White Pine is that the people in White Pine really have no
say over it whatsoever. (X4-6)
45) It's the biggest thing that ever hit close to the town was
the truck stop, . (VIII2-19)
Although such sentences do serve to identify one part of a discourse
with another, like the anaphoric identifying sentences, we refer to
them as peak sentences because they are employed by a speaker to indi
cate a peak or culmination in a discourse. We return to this point
below.
In ten of the fifty instances, a WH-clause follows the intro
ducing deictic, as in 44 above, so that these peak sentences appear
to be fusions of anaphoric identifying clauses and WH-clefts. In all
the remaining forty instances peak sentences also represent


TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS. iii
ABSTRACT vi
CHAPTER I INTRODUCTION 1
General Remarks 1
The White Pine Setting 4
The Economic Life of White Pine 15
The Religious Life of White Pine 17
The Southern Appalachian Region 20
The Study of Appalachian Speech 33
Fieldwork Procedures 38
The Sample 50
Transcription Procedure 53
Fieldwork Journal and Interview Catalog 56
Explanation to Informants 57
Taping of the Interviews 58
The Value of Fieldwork 58
Notes 60
CHAPTER II BACKGROUND AND METHODOLOGY 61
Preface 61
Discourse Studies and the Approach of This Study 62
The Study of Monologue Expositions 72
Oral Paragraph and a Review of Paragraph Studies 74
Longacre's Typology of Discourse Genres 84
Notes 88
CHAPTER III THEMATIC DEVICES 89
Introduction 89
Left Dislocation 96
The Variety of Forms 97
Specifications 104
Functions 109
Relationship to Subsequent Discourse 117
Existential Sentences as Thematic Devices 124
CHAPTER IV DEICTICS AND SUMMARY DEVICES. . 131
General Remarks 131
Left Dislocation with Deictics 136
Extrapositions with Deictics 137
iv


127
As a thematic device, the existential sentence does not have the
force of left dislocation, for perhaps three reasons. For one, it is
not a marked construction (in terms of information structure) as left
dislocation is. Not only is it normal, generally speaking, for the
first mention of a noun (except proper nouns) in a discourse to be in
definite. Left dislocation overrides this. But left dislocation also
has the effect of yanking an NP from its expected syntactic position
and bracketing it at the front of a sentence. Both left dislocation
and the existential are thematic devices and introduce themes, but the
former is a marked construction, the latter not.
Another reason is that an existentialized NP does not initiate an
anaphora chain nearly so often as a left dislocated one. In the exposi
tion above (p. 126), for example, four existentialized NPs ("people,"
"railroad," "depot," and "express office") are followed by at least one
anaphoric element each, but the other three are not. If these seven
NPs are typical, then themes introduced by existentials are not on the
average thematic for as long as those introduced by left dislocation.
A third reason is that existentials have another function which
competes with the function of thematization. This competing function
is that they are used simply to communicate whether something does or
does not exist or how much of something exists. Typically, existen
tialized NPs have quantifiers, as do the following:
60) And then there's a lot of flu going around. . (VIII2-16)
61) They's only one thing I'd like for this town to get. (VIII2-23)
Existential sentences with quantifiers in the data rarely initiate
anaphora chains.


12
old female quoted earlier put it,
...there are very, very wealthy people in this town, and there
are very, very poor people in this town. But to save my life,
I cant tell you what the difference is. I mean I can't point
to one person and say "Okay, this is a poor person and this is
a wealthy person. They, they blend.
This sentiment evidences the symbiosis referred to earlier.
One major division in the community is that between the middle
and older generations on the one hand and a large part of the younger
generation on the other. This division is a product of the rising ex
pectations of the town referred to above. In general the community is
more optimistic about its future than a generation ago because of im
proved local education. A comprehensive high school has been added and
a junior college has been opened five miles up the road. A more diver
sified job market has also developed. But most youths with professional
and career aspirations still see inadequate prospects in the nearby
area. Many of them, once they leave White Pine, are embarrassed to be
from the town. Often White Pine is for them a place to escape, a place
that can and does trap its ambitious youth. One twenty-nine-year-old
male characterized the town's subtle pressures not to aspire:
...most of the people prior to uh the sixties, when I got out,
they, they were settled. The idea was to, when you got out of
that high school was to get married, cause marriage was stabi
lization. You got married, you had children right then, as
quick as you could, because that made stay at home. You didn't
carouse. You didn't go on. You weren't apt to, to leave the
area and get involved in things that way.
This desire to escape is usually tempered by an attachment to the
familiar surroundings and personality of the town, though, so that some


Abstract of Dissertation Presented to the Graduate Council of
the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
A DISCOURSE ANALYSIS OF
EXPOSITORY APPALACHIAN ENGLISH
By
Michael Bryant Montgomery
August 1979
Chairman: William J. Sullivan III
Major Department: Linguistics
This dissertation contributes to the fields of discourse anal
ysis and Southern Appalachian studies by describing the discourse
functions of several linguistic structures from a large sample of
spontaneous expository (explanatory and descriptive) discourses for
a cross-section of informants from a small East Tennessee hill commu
nity. The community, White Pine, is a typical Southern Appalachian
one.
Five different contexts are taken into consideration for the
examined linguistic structures: 1) Speaker Type (a representative sam
ple of forty informants was interviewed); 2) Frequency of Occurrence
(data are quantified when possible); 3) Social Situation (researcher
assumed one role with his informants); 4) Task (a standard interview
was controlled to elicit only expository discourse); and 5) Linguis
tic (the sentence and discourse position of all data was considered).
This study is based on the premise that speech is hierarchical
vi


56
words (especially names) that only the interviewer can comprehend by
virtue of his having remembering the context of the interview.
Fieldwork Journal and Interview Catalog
In addition to the primary document, the tapes, and the secondary
document, the transcripts, two other document were assembled in this
project. A fieldwork journal of daily activities and observations was
maintained by the fieldworker. After a days fieldwork was completed,
a record of the day was entered into the journal. As mentioned earlier
in this chapter, this emulated the ethnographic diary that Malinowski
kept while among the Trobriand Islanders. Not only does such a journal,
maintained daily, allow a fieldworker to profit from his own experi
ences, his mistakes and his frustrations, but it also enables him to
make note of particularly successful and effective approaches and ques
tions for his later use.
The other document assembled was the interview catalog, compiled
for the researcher's ease of reference. It contains four kinds of in
formation:
1) The vital statistics on each informant (age, sex, level of edu
cational attainment, etc.);
2) The physical details of the interview: when and where it was
conducted;
3) A summary of how the interview itself went, how good the rap
port between fieldworker and informant was, and how helpful
the informant was;
4) A note on how the informant was contacted.


44
making arrangements with the contact persons in White Pine to spend
an indefinite period of time with them and his explaining what kind
of investigation he would undertake and who he would like to interview.^
A contact person is an invaluable bridge between a fieldworker and a
community. In it a fieldworker is initially a total stranger. He can
not reside long enough to become a familiar member of the community
(in White Pine or another typical small town in East Tennessee, this
would probably take several years in any case). This researcher's con
tact persons introduced him to many of his informants and gave him many
insights into the life of the local community.
In White Pine itself, four problems had to be resolved before
actual fieldwork (i.e. interviewing) could begin. First, the field-
worker had to be thoroughly familiar with the layout of the town and
the most important basic facts of daily life in the community. In this
he was primarily assisted by his contact persons. This included learn
ing the principal activities of the townspeople and the principal social
institutions in the community. He also had to determine the principal
concerns of life for the community. He was able to do this in the
course of the interviews.
Second, he had to establish the ethnographic parameters of his
work. This included his role in the community during his stay and in
the interview situation. Third, he had to choose a location for the
interviews, specifically, where they could be conducted comfortably.


63
cornucopia of observations on discourse that it should no longer be
asked "Is it worthwhile to do discourse analysis?" Rather we should
ask "What kind of discourse analysis will address significant ques
tions about how people organize their linguistic behavior?" Since
Grimes provides neither a methodology nor a point of departure for a
principled approach to the analysis of expository discourse (but nei
ther is his aim), we must look elsewhere for an approach to oral expo
sitions.
According to Widdowson, discourse analysts have generally taken
one of two basic points of departure. Either they have started with
the text or discourse as a whole or they have begun with sentences or
subsentences in an effort to identify their discourse function(s):
One general approach to discourse analysis, then, begins with
instances of discourse, with actual data, and moves towards lin
guistic units to the extent that this appears to be necessary
for the purpose of the description. The second approach moves
outwards, as it were, from the sentence, and deals with lin
guistic expressions as realized in discourse but with the ab
stract potential of linguistic forms. (1977:241)
The emphasis in the first approach is on the hierarchical nature of
discourse, in the second on the linear nature of discourse. There are
good reasons for choosing either point of departure and in many in
stances discourse analysts have taken one or the other and elected not
to move toward the other. Propp (1958) and a number of other writers
on folkloristics do not move toward linguistic units at all. Several of
the modern-day text grammarians (Pettifi and Rieser 1976, Van Dijk 1972)
also rely on abstractions from a text for analysis and are primarily


124
it is obvious, however, that conjunctions for our larger sample are
used in exactly the same way as those in the smaller one by speakers:
to indicate in many cases a change in theme for an ongoing discourse.
Conjunctions have functions other than indicating a paragraph's theme,
of course.
What then is the relationship between left dislocated NPs and sub
sequent discourse? An answer can now be given. Such NPs are normally
thematic for more than one sentence. After introducing an NP repre
senting new information into discourse through left dislocation, a
speaker normally maintains it as the theme until he explicitly indicates
btherwise. After a left dislocation, a hearer expects the ensuing dis
course to be thematically unified around what has been dislocated. This
expectation on the part of the hearer is frequently facilitated by a
speaker's use of anaphora referring to the theme. When a speaker wishes
to turn to a new theme, he/she uses one or another device for the hear
er's benefit to indicate this is happening. A writer uses an indenta
tion to signal a change in theme and hence the beginning of a new para
graph; a speaker uses an analogous cohesive device which is verbal to
indicate a change in theme and the beginning of a new (oral) paragraph.
We may conclude then that, at least for our corpus, the fundamen
tal function of left dislocation is the creation of paragraphs.
Existential Sentences as Thematic Devices
The other thematic device which is widely found in our corpus of
expositions is the existential sentence, a kind of sentence which nor
mally begins with "there" and indicates the existence or non-existence


CHAPTER VI
SUMMARY AND CONCLUSIONS
This dissertation was undertaken because of the long neglect
of the discourse level of analysis pertaining to American English.
Appalachian English was the variety of American English chosen to
study on this level because of its accessibility to the researcher.
The researcher undertook his own fieldwork and collected sufficient
data for this study to come to several important conclusions.
This study is about one aspect of how people in one small town
talk: how they organize what they have to say when they explain them
selves to a non-resident of that town. It shows that oral expositions
tend to be organized into blocks of thematically-unified information
called paragraphs and that paragraphs characteristically have one
theme each.
The organization of information into expository paragraphs is
shown in this dissertation to be reflected to an important degree by
certain devices called surface structure cohesive devices. It is
shown that speakers often rely on these devices to indicate to hearers
how a spontaneous monologue exposition is organized.
Extended speech thus normally has structure similar in kind to
extended segments of writing, although the devices used to reflect
this structure are different at least for the oral expositions exam
ined in this study.
161


114
Before examining how left dislocated NPs are related to subse
quent portions of discourse, we examine here more closely which types
of NPs are dislocated, and their distribution is also examined here.
If left dislocation usually introduces new information and is a
stage-setting device for the ensuing discourse, we would expect numer-
our dislocated NPs to come at the very beginning of an informant's
response to a question, where the speaker is probably most likely to
make introductions. Of the 606 instances of left dislocation in our
corpus, 67 (11 percent) are of this kind. In the example below, "at
White Pine" is dislocated in order to set a stage or frame of refer
ence for what comes after:
50) Fieldworker: How do the two schools compare?
Informant: Well, at White Pine, you know, you're really more
of an individual there. When we moved out to our new high
school, you were just like a person in a crowd, but as you
get older and stuff like that. . (VI11-1)
In general left dislocation does tend to involve NPs coming early in
expositions rather than late, but it is not a very strong function of
discourse-initial position.
If left dislocation is primarily a character-introducing device,
we would expect a large majority of dislocated NPs to be animate ones.
Because left dislocation can involve locative and temporal pieces of
information, it is clearly not restricted to only animate NPs. About
two-thirds (402 of the 606, or 66.3 percent of the total) of the dis
locations are animate, a strong if not overwhelming .. majority. One-
third then introduce inanimate NPs newly into discourse.
If left dislocation were a device for introducing active partici
pants rather than any thematic character, we would expect most


32
The self-reliance of the region's people has another side to it, as
Ford (p. 13) points out, in their attitude of acceptance toward federal
welfare and assistance programs. Odum remarked over forty years ago that
the people of Southern Appalachia, "reputed to be the most individualis
tic of all the regions, they cooperate most fully with Mew Deal tech
niques" (1936:97). Notwithstanding scattered initial opposition (primar
ily by those whose lands would be claimed by the reservoirs), the people
in the White Pine area and throughout East Tennessee have accepted and
cooperated with TVA's many programs. There was little reluctance to
accept whatever federal assistance was proffered.
The traditionalism and fatalism characterizing Appalachian people
are attributed to their struggle of many generations with the harsh reali
ties of life in a region with few resources and little available produc
tive land, according to Ford (1962:16). Weller (1965:33-40) shows how
such themes flow throughout mountain songs, stories, and daily attitudes.
In White Pine it is manifest in the strong and widespread doubts about
the value of change, of progress, and of post-high school education (see
quotes above, pp. 12-3). The twenty-nine-year-old male quoted above
says, regarding the community's residents,
They're...blinded really to what's, what's going on, I think.
They don't see their youth. They don't see, they don't see life
as progressing. They just see it as, as a tolerable state to be
in, and you just tolerate it now and hope for the future, yet you
don't work for it. You don't do anything working hand in hand.
As indicated above (pp. 8-9), recent changes and pressures for more
changes have made life more complex and confusing for many in White Pine.


77
structure have been pointed out. Klammer (1971), largely adopting
his functions from Wise (1968), mentions addition of participant, elab
oration, comment, specification, contrast, reason, summation, conclu
sion, setting, and result. But it has been difficult to show how these
functions are organized linearly, as Longacre has done for the grammat
ical units of the paragraph. The set of such functions has seemed
open-ended and at the same time we have not yet found, according to
Pike and Pike, "a way to treat some of them in a linear fashion, in
order to be able to perceive them as something 'going on"1 (1977:364).
This difficulty of formulating the lexemic structure of actual
paragraphs has in recent years caused lexemic structure to become
viewed as "deep" in contrast to the "surface" structure of paragraphs
being grammatical. The problem has become one of showing how lexemic
functions or units were realized by grammatical ones. Otherwise, lex
emic paragraphs would be only hypothetical. The problem of the extent
of isomorphism was tackled by Wise (1968) for oral narratives in Noma-
tsiguenga and for written dialogues in two English texts (I Henry IV
and Great Expectations) by Klammer (1971). It is useful to quote at
length the remarks of Wise:
Lexemic structure, as the term is used here, is a variety of
deep structure. For example, "logical subject" and "logical
object," i.e., agent and goal, are constituents of lexemic con
structions on the clause level. In levels beyond the sentence
the lexemic order of constituents is, in general, the chronolo
gical order of events in the narrative. Paraphrase and syno
nymy are important for the analysis of lexemic sentences and
phrases as well as for the analysis of lexemes which are posited
as the minimum lexemic units.


120
several sentences or at least until a speaker signaled to a hearer
that it were no longer the theme.
In looking at the sixty-one instances of left dislocated NPs in
our corpus which are thematic for only one sentence, we find that in
the majority (46 of 61) of them a speaker employs one or two of the _
cohesive deivces (thematic devices and conjunctions especially) to in
dicate to a hearer that a theme will not be maintained. In the re
maining fifteen instances the speaker provides some other other expli
cit signal. Most often a speaker will use a thematic device, either
another left dislocation or an existential, or one or more conjunc
tions, another kind of surface structure cohesive device which will be
our concern in chapter five. These devices indicate to a hearer that
a change in subject or in direction of a speaker's exposition is tak
ing place. They are part of a speaker's repertoire of devices for
structuring an ongoing monologue.
In eighteen of the sixty one instances we find another theme ex
plicitly introduced in the following sentence. Ten of these involve
another left dislocation, two an existential sentence, and six the in
troduction of another character into the discourse. In twenty-eight
instances one or more conjunctions or sequencers or interjections are
used to show the hearer a change in direction. These conjunctions and
other structures are shown in Figure 3.2 on the following page. More
often than not, these conjunctions are preceded or followed by a pause
or a hesitational "uh" which reinforces the change of direction they
signal to a hearer. Also, they often accompany thematic devices and
further strengthen a speaker's change of theme signal.


69
of a paragraph's grammatical structure and to a significant extent in
our White Pine expositions a paragraph's theme is marked by one of the
two constructions mentioned. In Longacre's most recent view of the
grammar of the paragraph, he indicates that "we find the paragraph
built around a theme that is not different in kind from a thematic par
ticipant" (Forthcoming:118). Our chapter three confirms this for expo
sitions.
Two other types of devices, deictics and summary, are considered
together in chapter four because deictics are frequently employed to
make summaries in our expositions. In general, deictics (especially
the word "that") help keep track of thematic information in a paragraph.
We will examine four uses of deictics in chapter four: 1) as the prono
minal reference in left dislocation; 2) in extrapositions; 3) in iden
tifying sentences; and 4) in what we will term "peak" sentences. An
example of a peak sentence is underlined below:
Fieldworker: What would you say have been the most significant
changes you've seen in the town in your few years here?
Informant: Most significant change I've seen is the, the intru
sion or the coming in of, of new people. White Pine is no
longer the same town it was eight and a half years ago because
of its growth, uh because of people with different backgrounds
having moved into our town, (uh-huh) Uh therefore it has
changed the complex of our town tremendously. Uh I think this
is the greatest change that has happened in our town is to
have new people coming in with new blood, new thoughts, new
ideas, new approach to things. (X4-4)
We will see that such sentences represent another unit in the grammati
cal structure of the paragraphthe peakand we will confirm the view
of Longacre and Levinsohn that, in expository discourse, the peak is
'the culminating explanation" (1978:105).


67
3) Participant anaphora;
4) Deictics;
5) Lexical ties and paraphrase;
6) Summary and preview;
7) Conjunctions and introducers;
8) Backreference.
Given the time limitations of this study, four of these devices (the
role of tense and voice, particles and affixes, lexical ties and para
phrase, and backreference) are not considered in this dissertation.
The remaining four (participant anaphora, deictics, summary and pre
view, and conjunctions and introducers) are the concerns of chapters
three through five herein.
The second hypothesis of this study is that the grammatical struc
ture of oral expository paragraphs is organized to a significant degree
by certain cohesive devicesthe other four discussed by Longacre and
Levinsohn. It investigates this hypothesis by examining the function(s)
and distribution of these devices.
The device called "participant anaphora" has to do with the identi
fication of participants in a discourse: "Participants can be identified
by name, by a common noun, by pronoun, by an affix, or just by zero.
Such variations in anaphora are never unmotivated. Most commonly, the
domain of the participant anaphora chain is the paragraph" (Longacre and
Levinsohn 1978:108). We will here be concerned with only one link in
this chainthe first(many linguists have studied anaphora: Gutwin-
ski (1976), Brinegar (1977), Hinds (1977) and others) and will call this


50
worker to introduce himself and to tell about his family was indispen
sable in establishing the uninhibited rapport that he desired. The in
terviewer's family was of more concern to the informants than his mo
tives for interviewing.
In his view, the fieldworker was able to establish quite good rap
port with each of his informants and excellent rapport with nearly all
of them. This resulted in not only a good corpus but also a number of
good friendships with individuals in White Pine.
The Sample
According to the 1970 United States Census, the population of the
town of White Pine was 1,532. It had grown to an estimated 1,830 by
1977. But the community of White Pine as defined by its residents in
cludes both the area outside the city limits and the adjacent neighbor
hood of Leadvale, bringing the total population of the area in which the
interviews were conducted to approximately 3,000 persons in 1978. Of
this base population of 3,000 a sample of forty residents was inter
viewed. This constitutes 1.3 percent of the whole. The interviews
average roughly one hour each, the shortest being thirty-five minutes
and the longest more than two hours.
The sample of forty informants is both a broad and a representative
cross-section of the White Pine community. The informants represented
all post-adolescent age groups and a wide variety of socioeconomic and
educational backgrounds. In age they range from sixteen to eighty-seven
as distributed among the following cells:


9
its citizens feel that the current pace of change in housing, in the
economic pressures on the town, and in the arrival of new people de
manding more from the town than it can offer is accelerating out of
control. Not too many years ago White Pine residents did pretty much
know everyone else in town, but with new residents recently moving in
faster than ever, they feel uneasy and sense a lack of cohesiveness
they once took for granted. Many of the longtime residents have a
strong sense of the symbiosis of the community, of how everyone plays
a part in supporting the general welfare. A typical expression of
this was made by a thirty nine year old female:
It's a feeling of closeness and connentuity [sic]. If I could
draw you a picture, I would draw you a big ball of people all
wrapped up together, like worms. You know, like you're fishing
with worms, and all those worms are intertwined. That's the
way this town is.
For many, though, this sense of closeness and symbiosis is inseparable
from the "traditional" way of doing things, so that change is threat
ening.
Probably the most important event in the past fifty years for
White Pine and East Tennessee is the arrival of TVA. For White Pine,
the coming of TVA has in a sense been responsible for most of the
changes during the past forty-five years (TVA began in 1933, and the
construction of Douglas Dam near White Pine began shortly thereafter).
TVA's initial effect was to dislocate many farmers in Jefferson County
from good bottom land and to resettle them, which was the cause of some
resentment. Today few people in the county regret the changes TVA made


145
combinations of well-known patterns. Peak sentences are clearly well-
formed fusions and are not faulty constructions.
It is not altogether clear, however, how the internal structure
of such sentences should be analyzed. What is the subject of a
sentence such as 44? Most likely it and other peak sentences are
best viewed as having sentential subjects and being representable
in such a tree-diagram as
46)
UP VP
But if we look at a peak sentence in isolation from its discourse
context, there is no reason not to propose a structure like 47:
47)
5
V up
. /
IS
\aJ1m4? is u1


90
substitutes (pronouns, for the most part) in a paragraph of language
which refer to a common piece of information or theme and is a special
kind of equivalence class. According to Becker (1965), an equivalence
class includes not only pronouns and demonstratives but also the repe
tition of words and their synonyms. Although not every paragraph in
our corpus has an anaphora chain, those which begin with a "left dis
located" NP nearly always do. In the two following examples the links
of the anaphora chains are underlined:
1) Fieldworker: Have you participated in any of the Pine Day cele
brations here in town? Have you attended any of them?
Informant: Yes, I usually go, because the ladies of our church,
we usually have a booth. We bake things, and so we have a
sale. That's a big day for ijs, you know. That's a way to
make uh some money for our uh projects, and thinqs like that,
and that's how we were able to help pay for our uh paint that
recently that we uh, we painted the inside the church. The
ladies did a little painting, but, but we did have enough
cash to buy all the paint. (II14-15)
2) ...This property that he, he let us have then for the parsonage,
build a parsonage of ft, ft, at one time there was, was a school
there. I was going to relate that to you. There, that's where
that I went to school, and my first eight years of schooling
was right there. And at this old school, that's where this par
sonage now sets across from our church. (II12-1)
In the first example, "the ladies of our church" is clearly the theme
of a block of information several sentences beyond its introduction. We
find ten pronouns referring back to it and one repetition of "the la
dies." In the second example, the theme is "this property. it is
around the property that the remainder of the exposition here is orga
nized. The anaphora chain contains such a pro-word as "there," since
the piece of information which is left dislocated and which is thema
tic is a locus. We see in these examples that theme is not just an


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EVBK1YOW4_BJMZCN INGEST_TIME 2012-12-07T20:48:44Z PACKAGE AA00012898_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


170
F: I've talked to a couple of younger people, say, in their mid twen
ties who said that, . .every time he comes here he sees a new
house going up and he feels a little pang of something.
I: Well, it, it is. We're growing. Uh ten years from now, I, I
wouldn't really want to uh say how big we'll be. We may be ten,
fifteen thousand. Just depends of a lot of things, (hmm) but uh
it, it's true. There's a lot of new, new uh houses being built.
In fact, you can't drive in the city any time that you don't see
either one uh just being started or being completed or something.
F: Is this the kind of place you think you would like to settle down
in for a long while?
I: Yes, I have no, I have no uh desire to go anywhere, other than vaca
tions. Uh course, I'm from a small community, (uh-huh) just east of
here a few miles and it's what I'm used to. I have no objections
to the way uh, the way the people are here, with the exception of a
few. Course, that's always the problem, (uh-huh) Uh but I'd like
to see it grow, not tremendously, but a little at a time, (uh-huh)
and uh have more things to do for the, especially for the young. I
think they, they should, they deserve. Course I, I know, you and
I were both young once and we'd uh like to had a lot of things that
we didn't have, and it's not changed. Kids today feel the same
way we did. (uh-huh) There's a lot of things they'd like, and I'd
like to see it. Course, it takes money. You talking taxes. Peo
ple just scream when you holler raising taxes. White Pine has the
lowest tax rate in any city, I would dare say, in the United States.


35
selection of grammatical, lexical, and pronunciation features in
mountain speech from North Georgia, v/ith accompanying citations from
sixteenth-century British literature wherever possible. In chapter
three she bewails the inevitable passing of the mountaineer with the
coming of modern education and technology, which will "wipe away for
ever our admirable mountaineer with his quaint and delightful manner
and speech" (p. 30).
Even well-known early scholars of the speech of the region are
given to such vague statements as "in general, most vowels may be used
interchangeably. Most vowels may replace 'e, 1 i,1 and V" (Combs
1931:1315).
But the most fundamental deficiency of almost all studies of
Appalachian speech to date is that they provide absolutely no context
for their data. It is impossible to understand a language fully unless
we have information regarding five different kinds of context for lin
guistic data:
1) Which part of the speech community uses the form(s)? (Old or
yound, middle or lower class, rural or urban speakers, etc.)
2) How often is a form used? (Is it categorical or extremely
rare or the predominant usage?)
3) In what kind of social situation is the form used? (i.e., What
style of speech is involved?)
4) What communicative task is a speaker performing when a form
is used?