Title: Temporal factors of talk in unconstrained conversation : personal and situational relationships
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00098327/00001
 Material Information
Title: Temporal factors of talk in unconstrained conversation : personal and situational relationships
Physical Description: xi, 128 leaves. : illus. ; 28 cm.
Language: English
Creator: Bein, Monte Francis, 1946-
Copyright Date: 1974
 Subjects
Subject: Sociolinguistics   ( lcsh )
Verbal behavior   ( lcsh )
Psychology thesis Ph. D   ( lcsh )
Dissertations, Academic -- Psychology -- UF   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Thesis: Thesis -- University of Florida.
Bibliography: Bibliography: leaves 118-127.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00098327
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000582452
oclc - 14107900
notis - ADB0827

Downloads

This item has the following downloads:

temporalfactorso00bein ( PDF )


Full Text














TEMPORAL FACTORS OF TALK IN UNCONSTRAINED CONVERSATION:
PERSONAL AND SITUATIONAL RELATIONSHIPS








By





MONTE FRANCIS BEIN


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
1974












This dissertation is dedicated to my
mother and father and their support of learning.













ACKNOWLEDGMENTS


Several people have been involved, at various points,

in the course of this research project. Ms. Madelyn Levine

helped me keep the tape recordings and papers in order, and

the subjects happy, throughout the process of data collec-

tion. Ms. Cathy Cook assisted with the scoring of the

personality and situational measures. Dr. Layne Prebor,

and the staff of the Equipment Shop, Communication Sciences
Laboratory, shared generously of their time and knowledge,

-with regard to the various problems of instrumentation.

Mr. Sanford Garner was an able and diligent chronographic

:assistant, and was paid with funds distributed by the Commu-

nication Sciences Laboratory. Mr. Wally Campbell, through

-several months of data reduction, patiently taught me the,

"language" of those vast machines at the University Com-

puting Center. Ms. Sunny Hughes transformed my "illegible"
handwriting into a communicable form.
Special thanks are due my committee members. My deepest

-appreciation goes to my committee chairman, friend and men-

tor, Dr. Norman Markel, without whose patient support and

gentle direction this research would not have been possible.

Dr. Franz Epting provided me with the initial opportunity

and feedback in my approach to the "word disease." Over the

years, Dr. Vernon Van De Riet, has remained the most important








influence on the development of my thought about clinical

psychology. Dr. Donald Teas and Dr. Edward Hutchinson have

contributed greatly to the pleasant, stimulating atmosphere

surrounding my work at the Communication Sciences Laboratory.


I













TABLE OF CONTENTS


Page

ACKNOWLEDGMENTS............................. .......... ii

LIST OF TABLES........................................vii

LIST OF FIGURES.......................................viii

ABSTRACT.............................................. ix

CHAPTER

I INTRODUCTION.................................... 1
Amount of Speech Channel Activity:
A Review of the Experimental Literature......3
Duration Measures of Speech Channel
Activity.................................. 3
Word Count Measures of Speech Channel
Activity............................... 11
Dyadic Units of Speech Channel Activity..14
Impressions from the Literature Review...17
The Person: Fundamental Interpersonal
Relations Orientation........................22
Compatibility........................... 24
The Situation: Evaluative and Dynamic
Dimensions of Conversation and Partner......26
Conversation Type........................29
Summary and Propositions....................29



II METHOD...................................... ..31
Subjects ...................... ...... ............31
Procedure ................................. 31
Raw Conversational Measures.................33
Inter-Chronographer Reliability.............35
Derived Conversational Measures.............35
Monadic Conversational Measures..........36
Dyadic Conversational Measures...........36
Measures of Person and Situation............39
Monadic Measures.........................39
Dyadic Measures..........................39


__ II








TABLE OF CONTENTS (CONTINUED)

CHAPTER Page

III RESULTS........................ ...........41
Temporal Factors of Talk: Inter-Correlations
and Normative Findings.... ................41
Monadic Conversational Measures............43
Talk Density and the Person..............43
Talk Density and the Situation...........43
Dyadic Conversational Measures..............53
Dyadic Compatibility.....................53
Conversation Type........................53


IV DISCUSSION..................................73
Individual Talk Density.....................73
Talk Density and Interpersonal
Orientation................ .............73
Talk Density and the Situation...........76
Dyadic Units of Speech Channel Activity.....79
Interactional Synchrony and
Compatibility............................79
Taxonomy of Conversation..................82
Minor Findings.......................... ......84
Summary and Conclusions....................87

APPENDICES.......................... ..........91

REFERENCES........................ ............. 118

BIOGRAPHICAL SKETCH...........................128













LIST OF'TABLES


Table Page
1 Conversational Measures ..........................37

2 Norms for Conversational Measures......... ........42

3 Analysis of Variance Summary for TDI in
Relation to Conversation-for-Speaker.............45

4 Analysis of Variance Summary for TD2 in
Relation to Conversation-for-Speaker.............46

5 Analysis of Variance Summary for TD2 in
Relation to Conversation-for-Partner.............50

6 Analysis of Variance Summary for TDI in
Relation to Speaker-for-Partner..................52

7 Analysis of Variance Summary for TCT in
Relation to Partner Evaluation...................56

8 Analysis of Variance Summary for TCT in
Relation to Partner Dynamism.....................59

9 Analysis of Variance:Summary for TCT in
Relation to Conversation Dynamism ................63

10, Analysis of Variance Summary for DTD in
Relation to Conversation Dynamism................65

11 Analysis of Variance Summary for STDI in
Relation to Conversation Dynamism................67

12 Analysis of Variance Summary for STRI in
-Relation to Conversation Dynamism................69

13 Analysis of Variance Summary for STDU in
-Relation to Conversation Dynamism.................71













LIST OF FIGURES

Figure Page

1 A Graphic Representation of TD1 and TD2
Across Conversation-for-Speaker...................48

2 A Graphic Representation of TD2
Across Conversation-for-Partner..................51

3 A Graphic Representation of TD1
Across Speaker-for-Partner......................54

4 A Graphic Representation of TCT
Across Partner Evaluation........................57

5 A Graphic Representation of TCT
Across Partner Dynamism..........................60

6 A Graphic Representation of TCT
Across Conversation Dynamism................. ...64

7 A Graphic Representation of DTD
Across Conversation Dynamism.....................66

8 A Graphic Representation of STDI
Across Conversation Dynamism........ .............68

9 A Graphic Representation of STRI
Across Conversation Dynamism....................70

10 A Graphic Representation of STDU
Across Conversation Dynamism....................72


viii


JI













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

TEMPORAL FACTORS OF TALK IN UNCONSTRAINED CONVERSATION:
PERSONAL AND SITUATIONAL RELATIONSHIPS

By
Monte Francis Bein

December, 1974
Chairman: Norman N. Markel, Ph.D.
Major Department: Psychology

This investigation was concerned with .the degree of

activation of the speech channel in unconstrained conversa-

tion. Amount of speech channel activity was examined both
in terms of the individual interlocutors, as well as in

dyadic terms, i.e., units describing the combined action

of the conversants. An extensive body of experimental

literature was reviewed and the following four hypotheses

.were offered: (1) Amount of talk was viewed as an expression

of internal states of interpersonal anxiety associated with

the needs for inclusion and control. (2) Amount of talk was

suggested as an index of certain situational factors of the

interaction, i.e., the evaluative and dynamic attributions

regarding the conversation and the conversational partner.

(3) Interactional synchrony, a condition of a high density
-of talk and a low density of simultaneous talk, was hypoth-
esized to be associated with the personal compatibility of







the two interlocutors. (4) The situational concepts

(partner and conversation) and dimensions (evaluation and

dynamism) were suggested as the basis of a taxonomy of

conversation, reflecting the temporal structure of the
interaction.

Tape recordings were made of 60 unconstrained conver-

sations, involving 120 unacquainted peers (college students).

Immediately subsequent to the interaction, the subjects

completed scales describing their characteristic inter-

personal orientation and their judgments regarding their

conversation and their conversational partner. Six person-

ality measures and four situational measures were derived

for each of the 120 subjects, and by mathematical combina-
tions of these measures, 16 compatibility measures and four
composite situational measures were derived for each of the

60 dyads. The tape recordings were subjected to a chrono-
graphic procedure, resulting in five monadic conversational

measures and seven dyadic conversational measures.

The results of the study indicated that an individual's
density of talk bore no relationship to his characteristic

interpersonal orientation. The results are discussed in

terms of the characteristics of unconstrained conversation.
As to situational relationships, participants described

as dynamic by their partners, participants in conversations

described as dynamic by their partners, and participants
who described their conversations as dynamic all exhibited

greater densities of talk. Thus, the dimension of dynamism,








indicating perceived power and energy and representing a

primary categorization of experience, is related to the

talk densities of individual interlocutors in

unconstrained conversation.

This study provided no supporting evidence for a.

conceptual model of conversation as an inter-dependence of

meshing processes, describable on two levels (conversational

synchrony and personal compatibility). The notion of

interactional synchrony is reconsidered, as well as the

amount of time necessary for the development of such

synchrony.
Regarding the question of a taxonomy of conversation,

Positive conversations, i.e., those characterized by a high

degree of interpersonal attraction between the conversants,

were associated with significantly increased total conver-

sation times. Dynamic conversations, i.e., interactions

mutually characterized by the interlocutors as powerful

and energetic, were associated with greater densities of

talk and simultaneous talk, an increased rate and longer

mean duration of simultaneous talk, as well as longer total

conversation times. It is concluded that interactions,

experientially differentiated by the interactors, are

associated with distinct temporal structures of talk.


II













C H A PT ER I
INTRODUCTION


The field of sociolinguistics (Ervin-Tripp, 1969) is

concerned with the relationship between characteristics of

the speech channel and characteristics of the communicators

and the communication situation. The speech channel, from

the vocal tract of an encoder to the ear of a decoder, is

one of six major channels of human "face-to-face" communi-

cation and includes both language and non-language sounds

(Markel, 1969). A basic characteristic of this channel is

that it is either activated or not activated, and the first

major question posed in this investigation is directed

toward an examination of individual psychological and situa-

tional correlates of the amount of activation of this-

channel.

As to psychological correlates, Argyle (1969), citing

several studies, suggests that dominance and affiliation are

the two major personality dimensions of consequence for

social interaction style. In the present study, Schutz's

(1966) FIRO (Fundamental Interpersonal Relations Orientation)
theory is employed to tap these dimensions of personality.

As to situational correlates, this investigation employs the

constructs of "semantic space" (Osgood, Suci, and Tannen-

baum, 1957) as a descriptive methodology.


I







This concern with correlates of speech channel activa-

tion directs one's attention to the phenomena of conversa-

tional interaction. Goldman-Eisler (1968) has pointed out
that time in its various manifestations of succession,

continuity, and duration is a basic datum of experience, and

that in spontaneous conversation, speech is spread over time,
thus revealing man'.s way of handling time. Time becomes a

quantity to be shared. Markel (1973) and other students of

human behavior have noted that the basic, most salient fea-
-ture of conversation is the alternation of action and

inaction. For a conversational interaction to occur, there
-must be co-ordination, synchrony, meshing, else the situa-

tion is that of two persons acting independently. Jaffee and
Feldstein (1970) defined a conversation as a succession of
sounds and silences generated by two interacting speakers.

If there is a failure of co-ordination of action (vocali-

zation, i.e., the production of sound) between the two
interlocutors there will be increased simultaneous vocali-

-zation. To the degree that there is a failure of co-ordin-

-ation of inaction, there will be decreased total dyadic
vocalization. In Sear's (1951) terms, these measurements

care dyadic units, i.e., units that describe the combined
action of two persons.
Chapple and Lindemann (1942), and more recently Argyle

(1969) have discussed these types of joint activity and
inactivity in terms of the notions of conversational "adjust-
-ment" and "equilibrium." In addition to meshing on the level


II








of dyadic units of speech channel activation, meshing may

be conceived of in terms of the personality characteristics

of the two interlocutors .(Argyle, 1969). The meshing of
personality characteristics is the notion of compatibility,

a central construct of Schutz's (1966) FIRO theory. A

second major focus of this investigation involves the deri-

vation of dyadic units of the amount of speech channel
activity, and examines their relationship to the personal

compatibility of the two interlocutors. In addition., the

semantic situational data of the two interlocutors is com-

bined and questions are raised regarding the relationship of

these dyadic units of speech channel activity to the notion
of conversation type.
Amount of Speech Channel Activity: A
Review ot the Experimental Literature

The amount of speech .channel activity has been examined

in two ways; as reflected by the duration of time spent in

speech production and as reflected in the total number of

words produced, i.e., verbal productivity (Mahl and Schulze,

1964). This operational distinction has been employed to
classify the extant experimental literature. There then
follows a section reviewing studies involving dyadic units
of speech channel activity.

Duration Measures of Speech Channel Activity
Eliot Chapple (1939) was among the first investigators

to isolate the time dimension for analysis as a method of
studying interactional behavior. Actually, Chapple's work
had been preceded by Norwine and Murphy (1938), in their


__







preliminary, voice-actuated machine analyses of the temporal

structure of telephonic conversation. Chapple was responding

to a felt need for objective measures in the description of
individuals, and his early work (Chapple, 1940; Chapple and

Arensberg, 1940) described personality differences as the

observed invariant properties of individuals in interaction.

His technique of interaction chronography assessed the prop-

erties of frequency, duration, and patterning in time of

periods of action and silence. Action was taken to.include

any form of message sending, thus subsuming vocalization.

For the assessment of these variables and combinations

thereof, Chapple (1949) developed the Interaction Chronograph,

a computer which is activated by a concealed observer during
the interaction. Chapple (1953) went on to devise a stan-

dardized interview for the employ of his Interaction

Chronograph and extended his inquiries into the interview

behavior of the mentally ill (Chapple, Chapple, Wood,
Miklowitz, Kline, and Saunders, 1960) and the relationships

of interview behavior to certain personality dimensions

(Chapple, Chapple, and Repp, 1954).
It was not until a decade later that the import of

Chapple's (1939) beginning became apparent to other investi-
gators. Verzeano and Finesinger (1949) reported the
development of their Automatic Speech Analyzer, thus

removing the observer from the system. Verzeano (1950, 1951)
went on to report time patterns of speech in normal subjects

and Lorenz and Cobb (1952) successfully applied the Automatic


I







Speech Analyzer in the differentiation of normal and

psychiatric patient populations. Goldman-Eisler (1951),
following Chapple's (1949) procedure, examined normal
conversational behavior. Other speech analyzers were
developed and reported by Kasl and Mahl (1956), Whittier

(1959) and Hargreaves and Starkweather (1959). Starkweather

(1959) reported data derived from role-playing situations
and Hargreaves (1960) examined data gathered from real-life
situations.
An excellent review of Chapple's work is presented by
Matarazzo, Saslow and Matarazzo (1956), and this article
represents the beginning of the most sustained and system-
atic inquiry into the temporal aspects of speech in inter-
action. Saslow and Matarazzo (1959) reported details of
five separate reliability studies which indicated that an
interviewee's durations of speech and silence were highly
reliable, despite large individual differences. Matarazzo
and his colleagues then embarked upon an extensive series
of validity studies involving planned changes in the inter-
viewer's behavior, which produced pronounced changes in the
speech behavior of.the interviewee. These planned changes

of interviewer behavior included increases and decreases in
his own average speech durations (Matarazzo, Weitman, Saslow
and Wiens, 1963); increases and decreases in his own speech
latencies (Matarazzo and Wiens, 1967); increases and
decreases in his speech interruptions of the interviewee
(Wiens, Saslow and Matarazzo, 1966); saying Mm-hmmm, Mm-hmmm







(Matarazzo, Wiens, Saslow, Alien and Weitman, 1964) and

nodding his head (Matarazzo, Saslow, Wiens, Weitman and

Allen, 1964). Over the years, a number of studies have

attempted to describe the relationship between content and

non-content speech variables (Kanfer, Phillips, Matarazzo

and Saslow, 1960; Phillips, Matarazzo, Matarazzo, Saslow and

Kanfer, 1961; Matarazzo, Weitman and Saslow, 1963; Matarazzo,

Wiens, Jackson and Manaugh, 1970a; Manaugh, Wiens and

Matarazzo, 1970; Matarazzo, Wiens, Jackson and Manaugh,

1970b). In addition, concurrect validity studies involving
normal and deviant groups (Matarazzo and Saslow, 1961) and

personality correlates (Matarazzo, Matarazzo, Saslow and

Phillips, 1958) have been reported. Matarazzo, Wiens and

Saslow (1965) present a fairly complete review of the first
eleven years of this research program.

Another research program, described and summarized by

Jaffee and Feldstein (1970), grows out of the Chapple tradi-

tion and, while lacking in seniority and extensity, the

level of methodological sophistication is comparable to the

work of the Matarazzo group. Their research program has
been implemented by the employ of the Automatic Vocal

Transaction Analyzer (Cassotta, Feldstein and Jaffee, 1964).
In contrast to the Interaction Recorder of Wiens, Matarazzo

and Saslow (1965), which is operated by a human observer,

the Automatic Vocal Transaction Analyzer has been electron-
ically designed to reproduce perceived temporal patterns
without the monitoring of the interaction by a human


I








observer. The investigations reported by Jaffee and

Feldstein (1970) have essentially replicated the findings
of Matarazzo et al. (1965), i.e., conversational behavior,

in terms of the on-off sequence of vocalization, is a stable

characteristic of the individual and is capable of modifying
and of being modified by the temporal patterning of other
individuals. The work of Jaffee and Feldstein (1970)

represents one major departure, i.e., an attempt to examine
conversational behavior in the absence.of role differences
between the interlocutors. Only recently, Matarazzo, Wiens,

Matarazzo and Saslow (1968) turned their attention to the
more naturalistic context of clinical psychotherapy, as
opposed to the standardized interview context of their
previous research efforts.
This long and grand research tradition, beginning with

Chapple (1939), has emphasized the more microscopic struc-
tural aspects of the patterning of vocalization and silence,
but some investigators have concerned themselves with the
significance of the relative amounts of vocalization and
silence. Recently,.Matarazzo, Wiens, Matarazzo and Saslow

(1968), have taken recourse to the more macroscopic measures
of amount of vocalization in an attempt to interpret their
findings regarding their datum of mean .utterance duration.

Their data, from seven series of psychotherapy sessions,
indicated that the mean percentage of session time used in
talking was approximately 67% for patients and 24% for
therapists.







Goldman-Eisler (1952, 1954), following Chapple's (.1949)
methodology with speech samples from psychiatric interviews,
derived the datum of action time (total action time as a
percentage of total interaction time). Goldman-Eisler

(1952) reported that, for interviewers, the amount of action
time depends on the type of patient interviewed, but that,
while interviewers adjust their action time, their relative
talkativeness remains. Patients, however, did not show
adjustment to different interviewers. Goldman-Eis.ler (1954),
re-analyzing this data, reported a significant positive cor-
relation between patient action time and the content variable

of self-reference. Re-analyzing data from an earlier study
of more natural conversational behavior (Goldman-Eisler,

1951), Goldman-Eisler (1954) included action time in a
measure of speaker dominance.
Cervin (1955a, 1955b, 1956) employed a similar measure
which he termed the participation quotient (total time spent
talking as a percentage of total interaction time), and
examined its relationship to solidarity (agreement/disagree-
ment by an experimental confederate) in a discussion

situation. The results indicated a higher participation

.quotient of Ss under conditions of solidarity. These find-
ings are generally consistent with the results.of interview
investigations by Kanfer and McBrearty (1962), Simkins (1963)
and Kanfer (1964) which examined the effects of various
forms of reinforcement on time spent in talking. Prebor

(1972) demonstrated that the participation quotient for







unconstrained conversation (.36) fell between the partici-

pation quotients.for solidarity (.40) and dissolidarity

(.25) reported by Cervin (1956).
Markel and Long (1974) examined the situational vari-

able of sex-of-listener and reported that both males and

females spoke more to female partners. Carment (1961)

examined another situational variable, strength of.opinion
regarding topic discussion areas, and found a positive rela-

tionship between this factor and the participation quotient.

One of the most interesting and novel examinations of the

effects of situational factors on amount of speech channel

activation is that of Soskin and John (1963). The investi-

gators collected their data from two married couples at a
weekend resort via radio transmitter, and derived a measure,

S%, i.e., the ratio of S total talking time to the total

talking time. The authors report data from one subject whose

S% varied from a low of 36% at breakfast with three other

people to a high of 69% in a situation of planning with his

wife.
Several investigators have examined the relationships

between organismic/personality factors and total duration

of time spent in talking. Cervin (1957), in the discussion
situation, found that Ss who scored high on a scale of
emotional responsiveness exhibited a greater participation

quotient. Carment, Miles and Cervin (1965), employing

psychometric measures of intelligence and extroversion,
found that Ss who scored high on both dimensions spoke the


I







greatest proportion of the interaction time. Levin, Bald-

win, Gallwey and Paivio (1960) indicated that children who
scored high on a measure of self-consciousness exhibited a

more drastic reduction in duration of speaking when the
communication was addressed to an audience, as compared to
a single target person. Axtell (1969), in an interview
situation, found that repressors, as measured by the R-S
scale, exhibited a lower duration of speaking. Kanfer, Bass

and Guyett (1963), examining the effect of social orienta-

tion on talking time in a laboratory condition of inter-
action without visual access, found that self-oriented Ss
took the most time in talking while interaction-oriented
Ss exhibited the least talking time. Markel, Bein, Campbell

and Shaw (in press), employing monologue speech samples,
observed that Ss with high.speaking times had significantly

higher scores on the.personality dimension of expressed
inclusion than did Ss with low speaking times.

Studies by Mahl (1956, 1961), Eisenman (1966), and
Pope, Blass, Siegman and Raher (1970) have been concerned
with the effects of anxiety. Mahl (1956) found that the
silence quotient (number of seconds of silence/number of
seconds available to patients for talk) increased in anxious

phases of psychotherapy sessions. This finding was not
replicated in a second study (Mahl, 1961). Pope et al. (1970)
derived a variation of the silence quotient from monologue
speech samples. They noted that speech samples, recorded
on days when patients were rated highly anxious, were







characterized by a lower silence quotient. By contrast, the

authors report that "high-depression speech" was character-
ized by a high silence quotient. Eisenman (1966) recorded

speaking times in group psychotherapy and noted that more
speech occurred among high anxious Ss (from the Taylor

Manifest Anxiety Scale) and also among first born subjects.

Carnes and Robinson (1948), employing a talk ratio
(client talk/total talk), made some interesting observations
in the context of the counseling interview. They noted
no difference in the talk ratio for experienced/inexperienced

counselors and positive correlations of the talk ratio to
scale ratings of growth in insight, working relationship,
and client responsibility for the interview. Anderson

(1960) reported that the interviewer talks more in employ-
ment interviews with applicants he accepts than in inter-
views with those he rejects. Simmons (1971) observed the
interaction of mother-child pairs and noted that mothers of
stutterers talked more than mothers of non-stutterers.

Word Count Measures of Speech Channel Activity

Research involving word count measures, i.e., verbal
productivity (Mahl and Schulze, 1964), finds its most

intensive focus in relationship to various constructs of
anxiety. It has been reported that high anxious subjects
(Taylor Manifest Anxiety Scale) exhibit a greater number of
words in the projective test situation (Benton, Hartman and
Sarason, 1955; Sauer and Marcuse, 1957) and in a structured
interview (Siegman and Pope, 1965b). For high anxious







subjects (Taylor Manifest Anxiety Scale), Westrope (1953)
reports a greater number of Rorschach responses and Davids
and Eriksen (1955) report a greater number of word associ-

ations. A series of investigations on the experimental

induction of anxiety in the structured interview (Siegman

and Pope, 1965a; Pope and Siegman, 1965; Pope, Siegman and.
Blass, 1970) has indicated that anxiety is associated with
increased verbal productivity. Incidentally, Siegman and
Pope.(1968) have also reported that the imposition of a
screen between the interlocutors results in decreased

productivity for the interviewee. Krause (1961), examining
segments from psychotherapy sessions, found a positive
significant relationship between the number of words pro-
duced and the verb-adjective ratio. Elsewhere, .Miller

(1951) reports that the verb-adjective ratio bears a posi-
tive correlation to emotional instability. Preston and
Gardner (1967), employing monologue speech samples,

observed that women tended to exhibit a greater verbal

productivity and that this seemed to be associated with an
increased need for social approval and an increased audience
anxiety.
On the other hand, there is some evidence for the
association.of anxiety with decreased productivity. Eichler

(1951) and Wharton (1953) present data which indicates
decreased Rorschach productivity when the Ss were made
anxious by threat of physical punishment. However,
Eichler's (1951) and Wharton's (1953) measures of







productivity are not directly comparable to word count

measures. Still, Lerea (1956) found that Ss exhibited a

lower verbal output when they reported considerable stage

fright, as compared to when they reported minimal stage

fright. West (1953), in an attempt to manipulate anxiety,

reported no differences in verbal output under conditions

of overt and covert tape recording.
Relationships between diagnostic status and verbal

productivity have been examined in studies by Feldstein

(1962) and Aronson and Weintraub (1967). Feldstein (1962),
in a structured interview situation, observed no differences

in verbal productivity between groups of schizophrenics and

medical students, but noted increased productivity for both

groups when the topic was affective and interpersonal.

Aronson and Weintraub (1967), analyzing monologue speech

samples, noted that divergencies from the "normal" number
of words decreased with improvement and increased with

deterioration, as measured by the depression scale of the

MMPI. Hawthorne (1934) .and Moore, Soderberg and Powell

(1952) examined the relationship between verbal productivity
and stuttering. Both investigations reported a positive
relationship between the variables.
Olson and Koetzle (1936) examined nursery school

children under conditions of free play and Winitz (1959)
observed kindergarten children in the projective test

situation. Both investigators reported that the number of
words employed by the girls was significantly greater, as


I




14


compared to the boys, and that older children spoke more

than younger children. Studies by Toman (1953) and Krauss,
Ruiz, Mozdzierz and Button (1967) asked Ss to tape record
statements of life history. Toman (1953) reported a mean
of.675 words and Krauss et al. (1967).reported a median of

436.words.
Mehrabian (1965) found that written communications

about liked objects were longer than communications about
disliked objects. Wiens, Jackson, Manaugh and Matarazzo

(1969) replicated Mehrabian's (1965) findings, and concluded
that the written channel may be as sensitive as the spoken
channel. Hirshman (1953), also exploring the written

channel with groups of psychotics and non-psychotics,
observed a significant rise in productivity for both groups
in reaction to interpersonal involvement.

Dyadic Units of Speech Channel Activity
The studies reviewed thus far have been concerned with
the relationship between various personal and situational

variables and the amount of speech channel activity of an
individual speaker. A second major focus of the present
investigation involves the explication of certain dyadic

units of the amount of speech channel activity and, in as
much as the former concern is reflected in a long and grand
research tradition, the latter concern represents somewhat of a
terra incognita. Dyadic units describe the combined action
of two persons (Sears, 1951). In the context of dyadic
interaction, dyadic units of amount of speech channel


I




15


activity would appropriately represent two distinct states,
i.e., a state of either interlocutor vocalizing dyadicc
vocalization time) and a state of both interlocutors

simultaneously vocalizing (simultaneous vocalization time).
Investigators have given these dyadic units of speech

channel activity scant attention. Anderson (1960) reported
that there is a greater amount of speech in employment
interviews with applicants who are accepted than in inter-
views with those who are rejected. Data from the psycho-

therapy series of Matarazzo et al. (1968) indicated that

dyadic vocalization represents about 87% of the total
session time. Markel and Long (1974) reported that dyadic
vocalization represents about 74% of the total interaction

time in unconstrained conversation. Soskin and John (1963),
in their "stream of talk" study reported total vocalization
proportions of interaction time ranging from .93 during

breakfast with four persons to .25 in a situation of husband
and wife making plans. Prebor (1972) observed that the
amount of talk per interval is lower in the beginning of a
conversation and that male-male dyads exhibited less speech
per interval.
As to simultaneous vocalization, Chapple and Lindemann
(1942), in comparison with data from psychiatric patients,
characterized a normal conversation as the complete absence
of double action and double silence, i.e., a smooth give
and take between the interlocutors. Chapple (1949) included
the frequency and duration of interruptions in the


I







interaction chronograph variables of "dominance" and

.'adjustment." However, Chapple purposefully confounded
his measure of action to include both vocal and gestural

activity. On the individual level, Wiens, Saslow and
Matarazzo (1966) reported that interruptions of the inter-
viewer by the interviewee occurred in about one of every

five interviewee utterances, that this was a highly stable
measure and was capable of being modified by changes in the
interviewer's speech behavior. Shaw and Sadler (1965), with

data from a discussion situation, observed that women
interrupted their partner more frequently than did men.

The Shaw and Sadler (1965) study also employed a dyadic
level of analysis, observing that more intimate couples were
less likely to interrupt each other. Argyle, Lalljee and
Cook. (1968) reported more interruption behavior in dyads
when both members wore dark glasses, indicating the impor-
tance of patterns of eye contact for interactional synchrony.

Markel and Long (1974) observed that simultaneous vocaliz-

ation represented about 1.9% of the total interaction time
in unconstrained conversation. Varying the distance between
conversants, they noted that the dyadic rate of simultaneous
speech was greatly increased for male-male dyads conversing
at a far distance (12 feet). Prebor (1972) indicated the
existence of more simultaneous vocalization in female-female
dyads. The only other known reference to dyadic units of
simultaneous vocalization appears in the psychotherapy
series of Matarazzo et al. (1968). Correlation coefficients




17


were computed between therapist and patient for the per-

centage of each interlocutor's utterances which were

interruptions of his partner. As such, these coefficients

are also dyadic units, and indicated the existence of a

striking degree of correspondence or tracking between the
participants.

There is one other obvious dyadic unit, total conver-
sation time, and, while not representing a state of speech

channel activity, it defines the unit of time into which
the above two states of speech channel activity must be

distributed, and is defined by the combined action of the

two participants. The literature review disclosed that

Anderson (1960) had examined this variable, reporting no

differences in total conversation time between employment
interviews in which the applicant was accepted and inter-

views in which the applicant was rejected. Prebor (1972)

reported that opposite sex dyads exhibited a greater total
conversation time than did same sex dyads.

Impressions from the Literature Review

The amount of speech channel activity has been operation-
alized as the number of words produced and as the duration

of time employed. Words are units of language, but speech

channel activity consists also of non-language sounds
(Markel, 1969). Eisenmann -(1966) reported a correlation of

.76 between the number of words spoken and the amount of
time spent in speech. Presumably, the variance unaccounted

for represents this distinction. The time dimension







represents a more accurate measurement of speech channel

activity. Furthermore, the context of dyadic interaction
aligns this investigation with the Chapple (1939)
methodology of interaction chronography.
Chapple's (1939) basic unit consisted of the duration

of individual actions, under which he included both vocal
and gestural activity. The extensive research program of
Matarazzo et al. (1965) took its impetus from Chapple's
beginning. These investigators confined their unit of

action, the utterance, to vocal activity, but there is a
confounding of speech and silence in the unit, as pauses
are included in the duration of a single utterance when the
context indicates that the speaker is not finished (Wiens,
Molde, Holman and Matarazzo,, 1966). As such, the utterance
does not represent a pure measure of speech channel activity.
Jaffee and Feldstein's (1970) unit, the vocalization,

appears a more appropriate index. They define a vocaliz-
ation as a segment of continuous sound, by one speaker,
that is bounded on each end by either silence or a

vocalization of the other speaker. The investigators
designed a voice-actuated machine (Cassotta, Feldstein and

Jaffee, 1964), the output of which is perceived temporal
patterns of dialogue. Such a system is a highly complex
and expensive undertaking. The system employed by Wiens,
Matarazzo and Saslow (1965), less electronically sophis-
ticated in that it requires monitoring of the interaction
by an observer, costs $15,000 and requires approximately







$4,000 per year for data processing.

For the more macroscopic purposes of this investi-
gation, exploring the significance of amount of speech

channel activity, an adaptation of a simple human-operated

device, described by Kas1 and Mahl (1956), meets the needs.

The vocalization (Jaffee and Feldstein, 1970), a segment of
perceived continuous sound by one speaker, is taken as the

basic datum. Boomer (1965) has indicated that the minimal

discontinuity in sound production perceivable by human
listeners is .20 seconds. This pause, or the simultaneous

cessation of vocalization by one speaker and onset of

vocalization by another speaker, demarcates the basic unit.

The sum of such units for an individual speaker is a

satisfactory operationalization of the amount of speech

channel activity, i.e., talking time. The literature review
suggested consideration of three different measures of talk

density, i.e., talking time.as a percentage of total

interaction time, e.g., Cervin (1956); individual talking

time as a percentage of total talking time, e.g., Soskin

and John (1963); talking time as a percentage of time
available for talk, e.g., Mahl (1956).

As to substantive knowledge regarding relationships
between amounts of speech channel activity and personality

factors, various constructs of anxiety have been the most
heavily researched, e.g., Pope, Siegman and Blass (1970).

Studies involving predispositional anxiety, experimentally

induced anxiety and naturalistic variations in anxiety




20


under clinical observation have all concurred in indicating

an activating effect of anxiety on the amount of speech
activity. The present study proposes to amplify the meaning
of this relationship through a consideration of Schutz's
(1966) FIRO theory.
Situational aspects which have been explored in rela-

tion to the amount of speech channel activity include
threat, reinforcement, sex-of-listener, various naturalistic
settings and topic type. The present investigation intends

a phenomenological approach, employing participant descrip-
tions of the situation derived from the constructs of
semantic space (Osgood et at, 1957).
The literature review indicated a paucity of studies

which had examined dyadic units, i.e., units describing the
combined action of two persons. Amount of speech channel
activity, in temporal dyadic units, indicates the amount of
time spent in the state of either interlocutor vocalizing

or the state of both interlocutors.simultaneously vocalizing.
A third dyadic unit, total conversation time, demarcates the
segment which is partitioned by the above two states.
As to dyadic vocalization time, some data has been

reported for employment .interviews (Anderson, 1960), psycho-
therapy (Matarazzo et al., 1968), conversation (Markel and

Long, 1974), and various naturalistic settings (Soskin and
John, 1963). Prebor (1972), and also Markel and Long (1974)
explored the relationships to dyadic sexual composition.

As to simultaneous vocalization time, relationships with







intimacy (Shaw and Sadler, 1965), interpersonal distance
(Markel and Long, 1974), disruption.of visual access
(Argyle, Lalljee and Cook, 1968). and dyadic sexual compo-
sition (Prebor, 1972) have been examined. Anderson (1960)
and Prebor (1972) have examined the variable of total
conversation time. The present study explores these dyadic
units as indices of interactional synchrony (Argyle, 1969;
Chapple and Lindemann, 1942), reflecting synchrony on
another level, that of the personal compatibility of the
interlocutors, i.e., Schutz's (1966) FIRO theory.
The review of the literature also indicated a near
absence of research involving unconstrained conversation.
Many of the studies have employed monologue speech samples,
most of the studies have employed data from interview
situations with its atypical roles and goals, and a few
studies have involved task orientations between.peers.
Studies involving interaction orientation between peers have
been reported only by Olson and Koetzle (1936), Goldman-
Eisler (1951), Hargreaves (1959), Soskin and John (1963),
Prebor (1972) and Markel and Long (1974). Jaffee and
Feldstein (1970) have indicated that it is not necessary
to provide topics for a lively conversation to ensue. This
investigation purports a more human ethological approach,
i.e., the descriptive study of human interaction in more or
less natural habitats and/or circumstances (Argyle, 1969).
Taxonomy is a prime concern in any unexplored area. The
present investigation examines the above described dyadic







units of speech channel activity as reflecting a taxonomy

of conversation based upon participant description of the
situation. The following sections briefly review and apply

FIRO theory (1966) and the theory of "semantic space"
(Osgood et at, 1957) to these concerns.

The Person: Fundamental Interpersonal
Relations Orientation
Schutz's (1966) FIRO theory is based on the postulate
that each individual has three interpersonal needs, i.e.,
conditions of the individual that are satisfied through the
attainment of satisfactory relations with others, and the
satisfaction of which is necessary to avoid undesirable

consequences (anxiety). These needs are as follows
(Schutz, 1966):

(1) Interpersonal need for Inclusion (I): the need to
establish and maintain a satisfactory relation with
people with respect to interaction and association.

(2) Interpersonal.need for Control (C): the need to
establish and maintain a satisfactory relation with
people with respect to control and power.

(3) Interpersonal need for Affection (A): the need
to establish and maintain a satisfactory relation with
people with respect to love and affection.
Schutz (1966) developed the FIRO-B questionnaire, con-
sisting of six.nine-item Guttman scales, to measure orienta-
tions to each of these three need areas. The procedure
relies on the self-report of individuals regarding the
behavior they express (e) to others and the behavior they
want (w) from others.
Schutz (1966) describes, for each need area, indivi-
duals whose behavior is consistently directed toward the







.satisfaction of that particular need. The oversocial is

characterized by excessive inclusion behavior, i.e.,

behavior designed to focus attention on the self, to be

noticed, to be listened to. The interpersonal behavior of

the autocrat tends toward the dominating, is manifested as
the desire for excessive power, authority, control over

others. The behavior of the overpersonal is excessively

directed toward being liked, toward becoming extremely

close to others.

The above conditions of excessive striving are taken

as representing discrepancies from an ideal state of need

satisfaction. In FIRO theory, this defines the presence of

interpersonal anxiety. The foregoing literature review

had indicated that "anxiety" was observed to have an

activating effect on the amount of speech channel activity.

The present investigation proposes to amplify this relation-

ship by.considering the amount of talking as an index of

this internal state of discrepancy between conditions of
actual and ideal interpersonal need satisfaction. Talking

is viewed as means of relieving certain internal states.

For the inclusion and control areas, the implied relation-

ships seem quite apparent. If a person has strong unsat-

isfied needs for inclusion, he may deal with this anxiety

by gaining attention through talking. If a person has

strong, unsatisfied needs for control, one obvious way of

dominating the other is to talk a great deal of the time.




24


Compatibility

Central to Schutz's (1966) theory is the notion of
compatibility, a property of a relation between two or more

persons that leads to mutual need satisfaction and harmo-

nious coexistence. Schutz (1966) states that the term is
best explicated sociometrically by the relation, "works
well with." Schutz (1966) describes three types of com-

patibility based upon the individuals' expressed (e) and
wanted (w) scores in each of the three need areas of
Inclusion (I), Control (C) and Affection (A).

Reciprocal compatibility (rk). A comparison of a's
description of how he likes to be acted toward with b's
description of how he likes to act toward people, and vice
versa, yields a measure of mutual need satisfaction or
reciprocal compatibility. Thus, rkab= ea-wbl +eb-wal.
Absolute measures are used as the concern is with the size
of the difference; the lower the score, the higher the

compatibility.

Originator compatibility (ok). Originator compatibility
is based upon each individual's preference for initiating,
as opposed to receiving, inclusion, control or affection
behavior. A measure of this preference (originator score)
is the difference between the expressed and wanted aspects
of a given need area, i.e., (ea-Wa). The measure of orig-
inator compatibility is obtained by adding the two origi-
nator scores, i.e., okab= (ea-wa) + (eb-wb). If there is
a preference for originating, the sum of the scores will be







positive, indicating competitive incompatibility. If

there is a preference for receiving, the sum of the scores

will be negative, indicating apathetic incompatibility.
Interchange compatibility (xk). Agreement as to the

general context or atmosphere of a relationship is described
by xk. The amount of interchange a person desires in a
given need area is described by combining his scores on the
expressed and wanted scales of that need area, i.e.,

(ea + wa = interchange score). The more similar two persons'
scores are, the more compatible, e.g., agreement as to the
degree of intimacy desired. To measure compatibility, one
score is simply subtracted from the other, i.e.,

xkab =I(ea + wa)l-I(eb + wb)jl The direction of the
difference is not important, so the absolute value of the
difference is sufficient; the smaller the value of xk, the
greater the interchange compatibility.

The present investigation examines the relationship of
these various formulations of dyadic compatibility to dyadic
units of speech channel activity, i.e., dyadic vocalization
time (DVT) and simultaneous vocalization time (SVT). In a
general sense, compatibility is taken as meaning "works well
with." In the context of an interaction orientation, "works
well with" is assumed to indicate a high DVT, i.e., nearly
all the time is filled with talk, and a low SVT, i.e., there
is little simultaneous vocalization (Chapple and Lindemann,
1942; Argyle, 1969).
As to evidence for FIRO theory, Schutz (1966) reports







satisfactory coefficients of stability for .FIRO-B measures

and concurrent validity studies based on political attitudes,

occupational choice and conformity behavior. In the preface
to the second edition, Schutz (1966) notes that the com-

patibility construct has been successfully applied in
marriage counseling, and as a method of group composition
for teaching teams., therapy, training and task groups.

Furthermore, Schutz (1966) reported the results of seven

large scale experiments which indicated that compatibility
was significantly related to sociometric choices, dyadic
productivity and group cohesiveness. Schutz (1967)

presents a complete bibliography which indicates the

present evidence for reliability and validity. Shaw and

Constanzo (1970), reviewing FIRO theory, concluded that the

general experimentalevidence for the theory is supportive,

that the theory is internally consistent and the predictions

amenable to testing.

The Situation: Evaluative and Dynamic Dimensions
ot Conversation and Partner
From the point of view of an individual, the situation
of conversational interaction may be viewed as consisting of

two concepts; a partner and their subsequent interaction,

the conversation. Following Osgood, Suci and Tannenbaum

(1957), the description, i.e., the connotative meaning, of
these concepts can be conceived of as their location on
experiential continue which are defined by a pair of polar

terms. Osgood et al. (1957) provide satisfactory evidence
from numerous factor analytic investigations which indicates







that the essential characteristics of "semantic space"

are described by a dimension of evaluation and a dimension

of dynamism. As such, these dimensions are taken as

representing the natural and spontaneous experiential
continue that people employ in describing their experience.

The studies reported indicate the possibility of as many

as seven dimensions, but the authors note that when the

measurement technique of semantic differentiation is.

applied to people, the potency and activity factors

coalesce into a factor most appropriately labeled dynamism,

and the additional factors (stability, tautness, novelty,

receptivity) are much less clearly defined.

Osgood et.al. (1957) present evidence that the evalua-
tive dimension is a most satisfactory measure of generalized

attitude. In Jungian terminology (Jung, 1968), the.evalua-

tive factor is isomorphic to the third psychological
function, feeling, i.e., the process of giving value. The

evaluative dimension when employed by the individual to

describe the concepts of conversation and partner may be

construed to represent measures of interaction satisfaction

and interpersonal attraction, respectively. A suggestion

that these dimensions are relevant to talk density is con-.
tained in the findings of Mehrabian (1965); that communi-

cations about liked.objects are longer than communications
about disliked objects.

The dynamism factor, including potency and activity,
represents a measure of strength and power, as well as a




28


measure of energy and excitement.. Carment's (1961) find-

ings, that opinion strength regarding discussion topic was
related to talk density, suggests the significance of the

dynamism factor. While evaluative ratings of conversation

and partner suggest a relationship to an individual's talk
density, the dynamism dimension, at least as applied to the

concept, partner, suggests a relationship between a partner's
talk density and the description of him by the individual.

Thus, in the present study, the relationship of an

individual's talk density is examined with respect both to

the situation as described.by the individual,.as well as to
the situation as described by that individual's partner.

There is a third apparent aspect of the situation worth

exploring. As to the concept, partner, his interpersonal

orientation (from the FIRO-B) might also be expected to

bear a relationship to an individual's talk density. The

question posed is whether or not the particular personality

structure of a conversational partner is related to the talk
density of an individual. Markel and Long (1974) have dem-

onstrated that an individual's density of talk bears a

relation to the sex of the partner.
There is an interesting isomorphism .between the proposed

critical dimensions of the person, i.e., dominance and

affiliation (Argyle, 1969), and the proposed critical

dimensions of the situation, i.e., dynamism and evaluation.
As the FIRO-B scales require subject responses in terms of

frequency, it seems wise to maintain this conceptual







consistency by employing the positive polar term, of

representative evaluative and dynamic adjective pairs, and

anchoring it to a similar seven-point scale of frequency.

Conversation Type

The semantic situational data of the two individual

interlocutors can be combined and questions raised regarding

the conversation and its relationship to the three dyadic

units of speech channel activity. The literature review.

had indicated that the phenomena.of unconstrained con-

versation remains virtually unexplored. The question of

taxonomy is an obvious initial one. Conversations can be

described as positive or negative, based either on the

factor of interpersonal attraction (partners' evaluations

of each other) or interaction satisfaction (partners'

evaluations of the conversation). The dynamism ratings,

both for partner and conversation, can be employed to

characterize conversations as dynamic or static.
Summary and Propositions

This investigation is intended as a contribution to

the field of sociolinguistics. The amount of speech channel

activity is examined in the context of unconstrained con-

versation, i.e., an interaction orientation between peers.

The amount of speech channel activity is described, in

individual units, by three temporal indices.of talk density,

and in dyadic units, by the proportions of total conversation

time spent in the states of either interlocutor vocalizing

dyadicc talk density) and both interlocutors simultaneously


29~-







vocalizing (simultaneous talk density). The following
relationships are suggested: (1) Individual talk density
is an index of a speaker's characteristic interpersonal
orientation. Talking is viewed as a means of relieving

states of anxiety associated with the interpersonal needs
for inclusion and control (Schutz, 1966). (2) Individual
talk density is an index of the situational meaning. The
situation is examined from the perspective of each inter-
locutor, in terms of the evaluative and dynamic attributions
regarding the concepts, partner and conversation. The

relationship of talk density to the interpersonal orienta-
tion of the partner is also examined. (3) Dyadic talk
density and simultaneous talk density are construed as.
indices of interactional synchrony, in the sense of a
synchronous conversation being one in which nearly all the
time is filled with talk and in which there is very little
simultaneous talk, i.e., there is a smooth give and take
between the interlocutors (Chapple and Lindemann, 1942;

Argyle, 1969). Synchrony on this level of interaction is
proposed to be accompanied by synchrony on the level of the
personal compatibility (Schutz, 1966) of the two inter-

locutors. (4) The dyadic units of total conversation
time,. dyadic talk density and simultaneous talk density
are suggested as reflecting a taxonomy of conversation
based on participant descriptions of the situation.












C H A P T ER II

METHOD

Subjects
Sixty male and 60 .female subjects were drawn from

introductory psychology courses at the University of

Florida. Their participation in the experiment fulfilled

one of the requirements of the psychology courses. Their

ages ranged from 17 to 39 years, with a mean of 20.1 years.

Prospective subjects were advised that the experiment

was concerned with certain aspects of how people talk to
each other. The sign-up procedure for the experiment was

arranged so that 60 dyads were formed; 20 male-male, 20

female-female, and 20 male-female dyads. The mixed sex

dyads were collected last so as to prevent the operation of

a selection bias. Subjects were requested to form dyads on

the sign-up sheet with persons with whom their acquaintance

was no more than that of a casual classroom basis. Dyad

members were requested to report to separate rooms at the

specified time to preclude the initiation of conversation

before the actual experiment had begun.

Procedure
When the subjects arrived they were received by the

experimenter or his assistant and, again, each person was

questioned about the level of his acquaintance with his








prospective partner. A short form of biographical infor-

mation was then filled out for each subject, and subject

numbers were assigned. When both dyad members had arrived

and the forms were completed, they were conducted from their

separate meeting rooms to the experimental room.
The experimental room was a 10' X 16' seminar room with

a 3' X 6' rectangular table and two chairs. Subject A of
each dyad was seated at one end of the table, and his

partner, Subject B, was seated to his/her right, on the
side of the table, at a distance of approximately three

feet. Sommer (1965) has indicated that this is the most

natural spatial arrangement for casual conversation. A

microphone was placed on the table between the subjects,

and so oriented that the recorded quality of Subject B's
voice was different from Subject A's. This was done to

facilitate subject identification from the tape recording.
The microphone was coupled to an Ampex AG-600 tape recorder

which was remote from the subjects. The biographical forms,

previously filled out, were placed on the table in front of
each subject.
The following instructions were then read to the subjects:

This experiment is concerned with how people talk to
each other. Your task is simply to engage in conversa-
tion with your partner for as long as you like, and on

whatever topics you like. Your conversation will be

tape recorded. To facilitate voice identification, I

would like each of you to read sentences one and two








on the top of the form.

These sentences.indicated each subject's number and the

current date. When each subject had been recorded, the

recorder was turned off, the forms collected, and the

following instructions read to the subjects:

Now we're ready to begin. Please remain seated until

you've completed your conversation and please do not

touch the microphone. Talk as long as you like, and

about whatever you like. When you've completed your

conversation, please return to the rooms you were in

before to fill out some more forms. Questions?

If there were no questions, the tape recorder was turned on

and this final sentence read to the subjects:

Remember you can talk about anything you like for as

long as you like.

The experimenter then left the room. Upon conclusion

of their conversation, the subjects returned to their

respective rooms and completed semantic differential scales,

describing their conversation and their partner, and the

FIRO-B (Schutz, 1967) scales. Copies of the semantic

differential scales and the FIRO-B scales appear in

Appendix A.
Raw Conversational Measures

The tape recordings of the conversations were then sub-

jected to a chronographic procedure by the experimenter and

an assistant. For this purpose, an apparatus was con-

structed which consisted of an electric stop watch (Precision







Scientific Instruments), an electric counter (Veeder-Root),

and a hand-held micro-switch. The counter-timer was
designed such that a depression of the micro-switch recorded
one unit on the counter and accumulated time until the

switch was released. This represents a simplified version

of an instrument described by-Kasl and Mahl (1956).
The tape recordings were monitored via headphones in a
quiet, distraction-free room. Following the advice of Kasl
and Mahl (1956), the chronographers fixed their eyes on a
neutral object in the room while monitoring the tapes, and

attempted to maintain a mental set of "responding to speech

as an irregular, intermittent auditory stimulus, rather than
to such linguistic units as phrases and sentences (p.390)."
Initially, a conversation was monitored for the period

of time necessary for the chronographer to become familiar
with the voices of the two interlocutors. Once this was

established, one of the subjects was targeted for first

monitoring. Each time this subject vocalized, the chron-

ographer depressed the micro-switch and released it when the
vocalization was terminated. This operation defined the
basic conversational unit of the study, the vocalization.
When the conversation was concluded, the chronographer
recorded the number of vocalizations (AN for Subject A) and
the total time of vocalization (AT for Subject A), and then

on a second monitoring obtained these measures for the other
participant (BN and BT for Subject B). On the third and
final monitoring, the chronographer.targeted those instances




35


when both participants were talking simultaneously and

obtained the number of simultaneous talks (SN) and the

total duration of the simultaneous talk (ST) in the con-

versation. Also, on the third monitoring, the total time

of the conversation (TCT), i.e., from the beginning of the

first vocalization to the end of the last vocalization,

was recorded by means of a second stop watch.
Inter-chronographer Reliability

After the two chronographers had become familiar with

the apparatus and procedure, nine conversations were

randomly selected from the sample of 60 conversations.

Each chronographer performed the chronographic operations

on the nine conversations, and then their respective

results were compared. Correlation coefficients of .87

for vocalizations and .99 for vocalization time were

obtained. Coefficients of r =.85 and r =.84 were observed,

respectively, for the number of simultaneous talks and the

total duration of simultaneous talk. When all conversations

had been chronographed, another reliability check was per-

formed on the final eight conversations. Reliability was

observed to have risen to r =.93 and r = .99, respectively,

for vocalizations and vocalization time. Reliabilities for

frequency and duration of simultaneous talk had increased

to .97 and .93, respectively.
Derived Conversational Measures

Upon conclusion of the chronographing, the raw conver-

sational measures were transferred from the data sheets to


~ I








computer punch cards. A computer program was devised

which calculated and printed five derived monadic measures

and seven derived dyadic measures.
Monadic Conversational Measures
The monadic measures consisted of three indices of

talk density;f ie., vocalization time as a percentage of

total conversation time (TD1); as a percentage of time

available for talk (TD2); and as a percentage of the total

vocalization time of the conversation (TD3). The remaining

two monadic measures were vocalization rate per minute

(VR), and mean vocalization duration (VD), reported in
seconds.

Dyadic Conversational Measures
The dyadic measures consisted of total conversation

time (TCT); the percentage of total conversation time

used in vocalizing by both interlocutors, dyadic talk
density (DTD); mean duration, in seconds, of simultaneous

talk (STDU); two measures of simultaneous talk density,

i.e., total simultaneous talk time as a percentage of

conversation time (STDI); and as a percentage of total

vocalization time (STD2); and two measures of simulta-
neous talk rate, i.e., the number of simultaneous talks
per minute of total conversation time (STRI) and the num-

ber per minute of total vocalization time (STR2). The

12 conversational measures and their computational formulas
appear in Table 1.








Table 1
Conversational Measures


Description Abbreviation Formulaa


A. Monadic Measures

1. Vocalization time as a percentage of
total conversation time (talk density)

2. Vocalization time as a percentage of
time available for talk (talk density)

3. Vocalization time as a percentage of
total vocalization time (talk density)

4. Vocalization rate per minute

5. Mean duration of vocalization in
seconds

B. Dyadic Measures

1. Total conversation time in seconds

2. Total vocalization time as a percentage
of total conversation time dyadicc talk
density)

3. Mean duration of simultaneous talk in
seconds


AT(BT)/TCT


AT(BT)/TCT-BT(AT)

AT(BT)/AT+BT-ST

AN(BN)/TCT x 60

AT(BT)/AN(BN)


AT+BT-ST/TCT


ST/SN


STDU







Table 1 continued


Description Abbreviation Formulaa

4. Simultaneous talk time as a percentage STD1 ST/TCT
of total conversation time

5. Simultaneous talk time as a percentage STD2 ST/AT+BT-ST
of total vocalization time

6. Simultaneous talk rate per minute of STR1 SN/TCT x 60
total conversation time

7. Simultaneous talk rate per minute of STR2 SN/AT+BT-ST x 60
total vocalization time


a. T= vocalization time in seconds

N= number of vocalizations

A= interlocutor A

.B= interlocutor B

S= simultaneous speech







Measures of Person and Situation
Monadic Measures

The FIRO-B protocols were hand-scored in accordance
with directions of Schutz (1967), resulting in six measures;

expressed and wanted inclusion (Ie and Iw); expressed and
wanted control (Ce and Cw); and expressed and wanted

affection (Ae and Aw).

The semantic differential scales, used to describe
the conversation and the conversational partner, consisted

of the same ten adjectives, and.each adjective was anchored
to a seven-point frequency scale (see Appendix A). Following
Osgood, Suci, and Tannenbaum (1957), five of the adjectives
were selected for their high and pure loading on the evalu-
ative dimension and the remaining five represented high
and pure loadings on the dynamism dimension (two potency,

three activity). By summing the ratings for each group of
five adjectives, four measures of the situation emerge;

conversational evaluation (CE), conversational dynamism (CD),
partner evaluation (PE), and.partner dynamism (PD).

Thus, five conversational measures (TD1, TD2, TD3, VR,
VD), six personality measures (Ie, Iw, Ce, Cw, Ae, Aw), and
four situational measures (CE, CD, PE, PD) were obtained for
each of the 120 subjects.

Dyadic Measures

By performing the mathematical operations described by
Schutz (1966), the FIRO-B scores of dyad members may be com-
bined to produce 16 distinct measures of dyadic compatibility.






A computer program was devised which calculated and printed
these 16 measures for each dyad. The measures included indi-

ces of reciprocal (rk), originator (ok), and interchange (xk)
compatibility in each of the three areas of inclusion (rkI,
okI, xkl), control (rkC, okC, xkC), and affection (rkA,
okA, xkA). The rationale and computational procedures for
each of these nine measures was described in the preceding
chapter. General measures of compatibility in each of the
three need areas (KI, KC, KA) were obtained by summing the
rk, ok, and xk scores for that specific need area, e.g.,
KI = rkI + okI + xkI. Measures of total reciprocal (rkT),
originator (okT), and interchange (xkT) compatibility repre-
sent sums for each type of compatibility over all need areas,
e.g., rkT = rkI + rkC + rkA. Total compatibility (K) is
obtained by combining either area compatibilities or com-
patibility types, e.g., K = KI + KC + KA = rkT + okT + xkT.

In addition, the four semantic differential scores of
each dyad member (A) were combined with the similar measures
of his partner (B) to produce dyadic descriptions (PED,
PDD, CED, CDD) of the evaluative and dynamic dimensions of
the two interlocutors (PED = PEA + PEB; PDD = PDA + PDB)
and their conversation (CED = CEA + CEB; CDD = CDA + CDB).
Thus, seven conversational measures (TCT, DTD, STDI,
STD2, STRI, STR2, STDU), sixteen indices of compatibility
(rkI, okI, xkI, rkC, okC, xkC, rkA, okA, xkA, KI, KC, KA,
rkT, okT, xkT, and K), and four composite situational measures
(PED,.P'DQ, CEO, CDD) were obtained for each of the 60 dyads.













C HA P T E R III

RESULTS


Temporal Factors of Talk: Inter-Correlations
and Normative Findings

The inter-correlations of the five monadic conversa-

tional measures and the inter-correlations of the seven

dyadic conversational measures are presented in Appendix B.

Monadic talk density has been computed as both a percentage

of interaction time (TDI) and as a percentage of total

vocalization time (TD3). The high positive correlation

(r=.91) of.the two measures suggested that this distinction

was superfluous and accordingly, TD3 was dropped from the

analysis. Simultaneous talk density and simultaneous talk

rate had also been computed with respect to both interaction

time and total vocalization time, and for similar reasons,

the indices based on vocalization time were dropped from

further analysis. The measures employing interaction time

as the common denominator were preferred in consideration

of comparability with previous research findings in the

area.

The over-all means and standard deviations of the

remaining four monadic measures and five dyadic measures

are presented in Table 2.







Table 2

Norms for Conversational Measuresa


Measures Mean Standard Deviation
A. Monadic Measures

1. TDI .41 .12
2. TD2 .68 .15

3. VR 11.16 3.05
4. VD 2.34 1.01

B. Dyadic Measures
1. TCT 1807.08 822.20
2. DTD .80 .09

3. STDU .40 .09
4. STDI .015 .010

5. STRI 2.14 1.11


a. The measures are described in Table 1.







Monadic Conversational Measures
Talk Density and the Person

Pearson correlation coefficients were computed between

the six indices of interpersonal orientation and the.four

conversational measures of-each subject over the sample of
120 subjects. Of the 24 computed coefficients, only one

was significant at the .05 level, a result to be expected

by chance alone. These coefficients of correlation appear

in Appendix C.

Talk Density and the Situation

Interpersonal orientation of partner. Pearson
correlation coefficients were computed between the four

conversational measures of each subject and the six indices

of interpersonal orientation of his partner. Of the 24
computed coefficients, again only one was significant at
the .05 level, a result not exceeding that expected by

chance. These correlation coefficients appear in Appendix D.

Situation-for-speaker. Each speaker's ratings of con-

versational dynamism, conversational evaluation, partner

dynamism and partner evaluation were correlated with his/her

four conversational measures. Employing the .05 level as a

cut-off, four of the 16 computed coefficients of correlation
were found to be significant (correlation coefficients

ranged from .20 to .39). These coefficients.of correlation

appear in Appendix E.

These results suggested that further information might.
be garnered by an analysis of variance, exploring the




44


effects of both the ratings of the conversation and the

ratings of the conversational partner. Accordingly, subjects
were first assigned to one of four groups on the basis of.
their dynamic and evaluative ratings of the conversation.
These ratings were employed to categorize subjects for
whom the conversation was static (below the dynamism
median), dynamic (above the dynamism median), positive
(above the evaluation median) and negative (below the
evaluation median). The normative findings regarding these
four situational measures are presented in Appendix F.
Four possible combinations (groups) emerge: Positive-
Dynamic, Negative-Dynamic, Positive-Static, Negative-Static.
After randomly removing subjects to equalize cell frequency,
the four groups of N=21 each were subjected to 2-way
analyses of variance (Winer, 1962) for each of the. four
monadic conversational measures.
For both TD1 and TD2, the dynamism factor was signif-
icant beyond the .01 level (F=9.34, df=1/80 and F=20.92,
df=1/80, respectively). Summaries of these analyses appear

in Tables 3 and 4.
Further statistical analysis, by way of the t-test,
indicated that TDI of the Negative-Dynamic group was signif-
icantly greater than that of the Negative-Static group
(t=2.07, p(.05), and that TD1 of the Positive-Dynamic group
was significantly greater than the Positive-Static group
(t=2.17, p<.05) and the Negative-Static group (t=2.42, P .02).
With regard to TD2, the Negative-Dynamic group was







Table 3
Analysis of Variance Summary for TDI
in Relation to Conversation-for-Speaker


Source SS df MS F

Evaluation .002 1 .002 .18

Dynamism .126 1 .126 9.34*

Interaction .000 1 .000 .00

Error 1.081 80 .014







Table 4

Analysis of Variance Summary for TD2
in Relation to Conversation-for-Speaker


Source SS df MS F
Evaluation .014 1 .014 .84

,Dynamism .340 1 .340 20.92*

Interaction .015 1 .015 .91

Error 1.299 80 .016







significantly greater than the Negative-Static group
(t=2.58, p9 .02) and the Positive-Static group (t=2.61, e<.02).

The Positive-Dynamic group was observed to have a TD2 sig-
nificantly greater than the Positive-Static group (t=3.94,

j14.01) and the Negative-Static group (t=3.91, p4.01). These
results are portrayed in Figure 1.

The remaining aspect of the situation-for-speaker
explored in this investigation involved the dynamic and
evaluative attributions the subject recorded with regard to
their conversational partner. Similar to the above analysis,
subjects were grouped according to whether they described
their partner as Positive-Dynamic, Positive-Static, Negative-

Dynamic or Negative-Static. After equalizing cell frequency,

the four groups of N=16 each were subjected to 2-way
analyses of variance (Winer, 1962) for each of the four
monadic conversational measures. No significant effects
were observed in any of the analyses of variance.

Situation-for-partner. Pearson correlation coefficients
were computed between the four conversational measures of

each subject and the dynamic and evaluative attributions
made by the subject's partner with respect to the subject
and their conversation. Employing the .05 level as a cut-
off, three of the 16 coefficients of correlation were found
to be significant (r ranged from .31 to .36). A table of
these coefficients appears in Appendix G.
These results suggested the existence of a relationship
between a.speaker's conversational measures and the






















% %
% %
U
Ur~


te


PD PS ND NS

Conversation-for-Speaker

Figure 1. A Graphic Representation of TDI and TD2
Across Conversation-for-Speaker


PD= Positive-Dynamic
PS= Positive-Static
ND= Negative-Dynamic
NS= Negative-Static


TDI
TD2


.80




.70


.60




.50



.40




.30




49


situation-for-partner, and further information was pursued

by means of analysis of variance. Subjects were first
assigned to one of four groups (as in the above, employing

the respective sample medians as cut-offs); subjects whose
partners described their conversation as (1) Positive-
Dynamic, (2) Positive-Static, (3) Negative-Dynamic or

(4) Negative-Static. After randomly removing subjects to
equalize cell frequency, the four groups of N=21 each were
subjected to 2-way analyses of variance (Winer, 1962) for

each of the four conversational measures.

The dynamism factor was observed to be significant
beyond the .05 level (F=4.57, df=1/80) for conversational
measure, TD2. A summary of this analysis appears in:

Table 5.
Further statistical analysis, by way of the t-test,
indicated that the Positive-Dynamic group had a significantly
greater TD2 than did subjects of the Positive-Static group

(t=2.73, a.O01). These results are presented in Figure 2.

Similar 2-way analyses of variance of the conversational
measures were performed with regard to the dynamic and
evaluative attributions accorded to the speaker by his part-

ner. Four similarly derived groups of N=16 each were
employed in separate analyses for each of the four monadic
conversational measures.

The dynamism factor was observed to be significant
beyond the .01 level in relation to TD1.(F=7.23, df=1/60).
A summary of this analysis appears in Table 6.




50


Table 5
Analysis of Variance Summary for TD2
in Relation to Conversation-for-Partner


Source SS df MS F
Evaluation .001 1 .001 .06

Dynamism .085 1 .085 4.57*

Interaction .058 1 .058 3.13

Error 1.481 80 .019


*P<.05






















.80




.70 ,


NS


Conversation-for-Partner

Figure 2. A Graphic Representation of TD2
Across Conversation-for-Partner


PD= Positive-Dynamic
PS= Positive-Static
ND= Negative-Dynamic
NS= Negative-Static


.60 .




.50 .


- II -- ---


W







Table 6
Analysis of Variance Summary for TD1
in Relation to Speaker-for Partner


Source SS df MS F
Evaluation .029 1 .029 2.08

Dynamism .100 1 .100 7.23*

Interaction .033 1 .033 2.35

Error .831 60 .014


*p.0 1







Further statistical analysis, by way of the t-test,

indicated that Negative-Dynamic subjects had a TD1 which
was greater than Negative-Static subjects (t=2.97, P4.01),

greater than the Positive-Static subjects (t=2.91, p4.01)

and greater than the Positive-Dynamic subjects (t=2.09,

pL.05). These results are shown in Figure 3.

Dyadic Conversational Measures
Dyadic Compatibil.ity
Pearson correlation coefficients were computed between

the 16 compatibility measures and the five conversational

measures of each dyad, over the sample of 60 dyads. Of the
80 computed coefficients of correlation, only two

coefficients attained a .05 level of significance, a result
to be expected by chance alone. These correlation coefficients
appear in Appendix H.
Conversation Type
Pearson correlation coefficients were computed between

the four composite situational measures (CED, CDD, PED, PDD)

and the five dyadic conversational measures over the sample
of 60 dyads. Employing the .05 level as a cut-off, 11 of the
20 computed coefficients were found to be significant (the
coefficients ranged from .27 to .46). These coefficients
of correlation appear in Appendix I.
The observed relationships between the composite

situational variables and the dyadic conversational measures
suggested that a more focused analysis, in terms of con-
versation type, be performed. The evaluative and dynamic




54
















.60



.50




.40



o .30

I-*
I-I

PD PS ND NS

Speaker-for-Partner

Figure 3. A Graphic Representation of TD1
Across Speaker-for-Partner

PD= Positive-Dynamic
PS= Positive-Static
ND= Negative-Dynamic
NS= Negative-Static








dimensions were each examined separately, both in relation

to the concepts, partner and conversation.
Evaluative dimension. With respect to the concept,

partner, conversations were classified as (1) Positive if

both members rated each other above the sample median for

partner evaluation, (2) Negative if both members rated
each other below the median, and (3) Ambivalent Evaluation

if one participant was rated above the median and the other

participant was rated below the median. After randomly

removing dyads to equalize group sizes, these three groups

of N=19 each were subjected to one-way analyses of variance

(Winer, 1962) for each of the five dyadic conversational

measures.
Partner evaluation was observed to be of significance

only with respect to total conversation time (TCT). For

this variable, partner evaluation was significant at the.

.05 level (F=3.29, df=2/54). A summary of this analysis
appears in Table 7. Further statistical analysis,.by way

.of the t-test, indicated that Positive dyads exhibited a
significantly greater total conversation time than did

Negative dyads (t=2.55, _<.02). These results are presented
in Figure 4.
Considering the concept, conversation, the sample con-
versation evaluation median was employed to characterize

conversations as (1) Positive, i.e., both members rating the

conversation above the sample median, (2) Negative, i.e.,
both members rating the conversation below the sample







Table 7
Analysis of Variance Summary for TCT
in Relation to Partner Evaluation


Source SS df MS F
Between Groups 4099230.0 2 2049615.0 3.29*

Within Groups 33594032.0 54 622111.7

Total 37693248.0 56




















2200

2100

2000
o
0
, 1900

S1800

1700
0
S 1600

L 1500

0

P AE N
0 Partner Evaluation

Figure 4. A Graphic Representation of TCT
Across Partner Evaluation

P= Positive
N= Negative
AE= Ambivalent Evaluation







median and, (3) Ambivalent Evaluation, i.e., one member

rating the conversation above the median and one member
rating the conversation below the median. Dyads were
randomly removed to equalize group sizes (N=19 each) and
then separate one-way analyses of variance were performed

for each of the five conversational measures. Conversation

evaluative type was not observed to be a significant
factor with respect to any of the conversational measures.

Dynamic dimension. Regarding the concept, partner,
conversations were classified as (1) Dynamic if both members
rated each other above the sample partner dynamism median,
(2) Static if both members rated each other below the sample

partner dynamism median, and (3) Ambivalent Dynamism if one
participant was rated above the sample dynamism median and
one was rated below the median. The groups were equalized

for sample size and then the three groups of N=14 each were
subjected to one-way analyses of variance for each of the
five dyadic, conversational measures.

Partner dynamism type was a significant factor only
in relation to total conversation time (TCT). For this
variable, partner dynamism type was significant at the .05
level (F=3.44, df=2/39). A summary of this analysis

appears in Table 8. Further statistical analysis, by way
of the t-test, indicated that Dynamic dyads had a signifi-
cantly greater total conversation time than did Static
dyads (t=2.62, pe.02). These results are presented in
Figure 5.







Table 8
Analysis of Variance Summary for TCT
in Relation to Partner Dynamism


Source SS df MS F
Between Groups 4230377.0 2 2115188.0 3.44*

Within Groups 23995360.0 39 615265.6

Total 28225728.0 41
















2300

2200

2100
0
8 2000

1900

1 1800

. 1700

S1600

o 1500,
o
- 1400
4.1
0
I-
IIi I I
D AD S

Partner Dynamism

Figure 5. A Graphic Representation of TCT
Across Partner Dynamism

D= Dynamic
S= Static
AD= Ambivalent Dynamism







As to the concept, conversation, similar to the above,
the sample conversation dynamism median was employed to

characterize conversations as (1) Dynamic, i.e., both mem-

bers rating the conversation above the sample median,
(2) Static, i.e., both members rating the conversation below
the sample median and (3).Ambivalent Dynamism, i.e., one
member rating the conversation above the median and one

member rating the conversation below the median. Dyads were
randomly removed to equalize group size (N=18 each), and

then the effect of conversational dynamism type was explored

by means of separate one-way analyses of variance for each
of the five dyadic conversational measures.

Conversational dynamism type was observed to be of
significant effect with respect to each of the five conver-

sational measures. Regarding total conversation time (TCT),
dynamism was significant at the .05 level (F=4.39, df=2/51),

and t-tests indicated that both the Dynamic group and the
Ambivalent Dynamism group had greater total conversation
times than did the Static group (t=2.32, p4.05 and t=2.76,

a1.01, respectively). As for dyadic talk density (DTD),
dynamism type was significant at the .01 level (F=6.42,
df=2/51), and t-tests indicated that the talk densities of

the Dynamic and Ambivalent Dynamism groups were both
greater than that of the Static conversation group (t=3.57,
p4.01 and t=2.04, p.05, respectively). Concerning simul-
taneous talk density (STDI), dynamism type was significant
at the .05 level (F=4.66, df=2/51), and the Dynamic







conversations had a significantly greater proportion of
simultaneous talk than did the Ambivalent Dynamism con-
versations (t=2.01, p<.05) and the Static conversations
(t=2.99, p4.01). Conversation dynamism type was signifi-
cant at the .05 level for simultaneous talk rate (STRI)
(F=4.71, df=2/51), and t-tests indicated Dynamic conver-
sations to have a rate of simultaneous talk significantly
greater than both the Ambivalent Dynamism group and the
Static group (t=2.33, p4.05 and t=2.89, 4C.01, respec-
tively). As to simultaneous talk duration (STDU), conver-
sation dynamism type was significant at the .05 level
(F=4.48, df=2/51), and t-tests indicated that both the
Dynamic conversations and the Ambivalent Dynamism con-
versations exhibited a mean duration of simultaneous talk
which was significantly greater than that of the Static
conversations (t=2.56, Ep.02 and t=2.62, pe.02, respec-
tively). Summaries of these analyses of variance appear
in Tables 9-13 and the results are portrayed graphically
in Figures 6-10.








Table 9
Analysis of Variance Summary for TCT
in Relation to Conversation Dynamism


Source SS df MS F
Between Groups 5192704.0 2 2596352.0 4.39*

Within Groups 30149328.0 51 591163.25

Total 35342032.0 53


*~4.05


















2100

o 2000

1-900

E 1800

o 1700
4.J
V 1600
L

S1500
o
u0
S1400

o 1300



D AD S

Conversation Dynamism

Figure 6. A Graphic Representation of TCT
Across Conversation Dynamism

D= Dynamic
S= Static
AD= Ambivalent Dynamism







Table 10

Analysis of Variance Summary for DTD
in Relation to Conversation Dynamism


Source SS df MS F
Between Groups .0910 2 .0455 6.42*

Within Groups .3613 51 .0071

Total .4523 53




66













.90



.85



.80



I .75
U
-U

7 .70



D. AD S

Conversation Dynamism

Figure 7. A Graphic Representation of DTD
Across Conversation Dynamism

D= Dynamic
S= Static
AD= Ambivalent Dynamism







Table 11

Analysis of Variance Summary for STDI
in Relation to Conversation Dynamism


Source SS df MS F

Between Groups .0009 2 .0004 4.66*

Within Groups .0048 51 .0001

Total .0057 53


*P<.05




68








.021

.020

.019

.018

.017

o .016

- .015









.010


D AD S

Conversation Dynamism

Figure 8. A Graphic Representation of STD1
Across Conversation Dynamism

D= Dynamic
S= Static
AD= Ambivalent Dynamism
AD= Ambivalent Dynamism








Table 12

Analysis of Variance Summary for STR1
in Relation to Conversation Dynamism


Source SS df MS F

Between Groups 10.7105 2 5.3552 4.71*


Within Groups 57.9888 51 1.1370


Total 68.6992 53



*4.o05


-----i~















2.8

2.7

02.6

E 2.5
L"
5 2.4
CL

e 2.3

2.2,

1- 2.1

0 2.0

* 1.9

.E 1.8

1.7


D AD S

Conversation Dynamism

Figure 9. A Graphic Representation of STRI
Across Conversation Dynamism

D= Dynamic
S= Static
AD= Ambivalent Dynamism







Table 13
Analysis of Variance Summary for STDU
in Relation to Conversation Dynamism


Source SS df MS F
Between Groups .0590 2 .0295 4.48*

Within Groups .3359 51 .0066

Total .3949 53
















.43 t


.35


:- ---I -- -- -j- -
D AD S

Conversation Dynamism


Figure 10.


A Graphic Representation of STDU
Across Conversation Dynamism


D= Dynamic
S= Static
AD= Ambivalent Dynamism












CHAPTER IV
DISCUSSION

Two major concerns are posed in this investigation.
The first involves an explication of the psychological
and situational significance of individual talk density.
The second focus explores the significance of dyadic units

of amount of speech channel activity.

Individual Talk Density
Talk Density and Interpersonal Orientation

No relationship was observed between the indices of
talk density and the interpersonal orientation of the
speaker. This analysis had been proposed as a means of
differentiating and amplifying the reported association
of anxiety with increased verbal/vocal output, through
examination of the concept of interpersonal anxiety.
Following Schutz (1966), anxiety is construed as the antic-

ipation of the non-satisfaction of a need. Two particular
states were hypothesized as being manifested in increased
talk density; the anticipation of being ignored (inclusion
anxiety) and the anticipation of not being influential
(control anxiety). Individuals may respond to the state of
inclusion anxiety with behavior designed to focus attention

on them, i.e., the oversocial (Schutz, 1966), and indi-
viduals may respond to the state of control anxiety with







dominating behavior, i.e., the autocrat (Schutz, 1966), but

this investigation disclosed that a heightened density of

talk is not a reliable behavioral indication of these

states.

Assuming the validity of the FIRO theory and measure-
ment (Shaw and Costanzo, 1970), and the demonstrated acti-

vating effect of anxiety, e.g., Pope, Siegman and Blass

(1970), one is led to speculation regarding the activating
effects of interpersonal anxiety upon other channels of

face-to-face .communication. A tentative hypothesis might

be that interpersonal anxiety is associated with an

increased density of kinesic (body movement) channel

behaviors, with the social significance of facial movements

being a likely area in which to focus analysis. For example,

looking behavior has apparently strong components of both

inclusion and control motivation. One may look, hoping to

"catch the eye" of the other, thus satisfying inclusion needs

of feeling attended to. Argyle (1967) suggests that people
who are more.concerned with being attended to, look more.

This observation speaks directly tothe concept of inclusion

anxiety. As to control components of looking, Argyle (1967)

further suggests that A may feel he has B under his power

when he is watching him, perhaps a transformation of passive

to active, regarding the child's experience of being watched
over by adults. Other communication channels (Markel, 1969),

i.e., proxemics (body placement, as.in the forward lean),
observation and odor (as in the wearing of startling,







unconventional clothes or potent perfumes) and touch might

represent avenues of lesser resistance for the expression

of interpersonal anxiety, and thus, be pre-potent over
speech channel activation.

That other communication channels might be pre-potent

for the expression of anxiety is suggested by a considera-
tion of an apparent difference between the speech channel and
the other channels of communication. Jaffee and Feldstein

(1970), noting the low incidence of simultaneous speech in
interaction, advance the idea of the operation of "terri-

toriality" with regard to speech channel behavior, i.e.,

speech by one interlocutor suppresses speech by the other

interlocutor. Simple reflection suggests that territori-

ality may not be such a prime characteristic of the other

channels of communication, where temporally concurrent
behaviors may be the rule, rather than the exception. In

this sense, other channels of communication might represent
avenues of decreased resistance for the expression of

anxiety. One may speculate that the extremes of inclusion

pathology (manic depressive psychosis, manic type) or con-

trol pathology (psychopathic personality) are necessary to
overcome the territorial imperative of the speech channel,
while lower levels of interpersonal anxiety are expressed
by an increased activation of other channels. Such an

explanation does not rule out the possibility that inter-
personal anxiety is expressed on the speech channel in ways

more subtle than sheer volume,.i.e., startling remarks as








expressions of inclusion concerns or patterns of voice

quality (loudness, tempo) for control concerns.

The experimental data, establishing the activating

effect of anxiety on speech, has been collected exclusively

in monologue or interview situations. In such situations,

the task of the subject is simply to fill a segment of time
with information, whereas, the "task of conversation," as

explored in this investigation, equally involves a gathering

of information. This implicit-instructional set, as well

as the territoriality of speech channel behavior,.would

seem to be operating to inhibit the effects of interpersonal

anxiety on the individual density of talk. One may conclude

that the relationship between talk density and anxiety is

dependent upon laboratory conditions of reduced reciprocity

and speech channel territoriality, and that the activating

effect of anxiety on speech is not of sufficient intensity

to overcome these stipulations in unconstrained two-person

interaction.

Talk Density and the Situation

While the first hypothesis .directed attention to an

examination of the relationship between interpersonal

anxiety level and talk density, a second proposition
suggested a relationship between the talk density of a

speaker and the specific situational significance of the

interaction. The situation was construed as two-faceted,

i.e., a partner and a conversation, and two dimensional,
i.e., evaluation and dynamism, and was explored in three







separate ways, i.e., the situation-for-speaker, situation-

for-partner, and the interpersonal orientation of the
partner.

Regarding the situation-for-speaker, talk density is
related only to the dynamic attributions made by the

speaker about the conversation itself, i.e., subjects
who described their conversations as dynamic, in compari-

son with participants describing their conversation as

static, spoke a significantly greater proportion of the

interaction time (TD1) and used a significantly greater

proportion of the time available for them to speak (TD2).

This finding indicates that the amount of power and energy

perceived in a conversation is related to the perceiver's
degree of participation regarding speech channel behavior.

What is surprising is the apparent irrelevance of the

evaluation dimension. It had been hypothesized that evalu-

ative ratings of the partner (interpersonal attraction) and
evaluative ratings of the conversation (interaction satis-

faction) would reflect a speaker's density of talk in that

conversation. This relationship had been suggested by
Mehrabian's (1965) findings that the factor of like-dislike

was related to the length of written communications. Again,
this may be understood as the result of contextual restraints.

In the Mehrabian (1965) study, persons were presented with

an unlimited space in which to behave (write), whereas in
conversation, the size of the behavioral space is partially

delimited by the conversational partner. Other channels of







communication might be expected to permit freer.expression

of such attitudinal concerns. The results of this study
indicate that individual talk density is not an index of

communicator attitude in unconstrained conversation.

As to the situation-for-partner, it had been hypoth-
esized that individual talk density is an indication of the

amount of dynamism attributed to a speaker by his partner.

This hypothesis was confirmed, i.e., speakers who were
described as dynamic by their partners spoke a signifi-

cantly greater proportion of the interaction time (TD1),

in comparison with speakers described as static. It may
be concluded that a heightened density of talk is one of

the behavioral correlates of those.persons perceived as

dynamic by their conversational partners. This is to say
that the variable, talk density, is associated with the

degree of power and energy that one person represents to

another. This analysis also indicated that speakers,
whose partners described their conversation as dynamic,

spoke a greater proportion of the time available for them

to speak (TD2), in comparison with speakers whose partners
judged their conversation as static. Thus, the amount of

power and energy attributed to a conversational interaction
is positively related both to the speaker's density of talk
and. to the partner's density of talk.

In answer to the second hypothesis, it may be asserted
that individual talk density.does represent an index of the

dynamic dimensions of the situation, but is unrelated to







the evaluative aspects of the situation. As such, the

dynamism dimension, representing one of two basic responses

to the environment, finds behavioral expression in the talk
densities of interlocutors in unconstrained conversation.

A third investigated aspect of the situation, the inter-
personal orientation of the conversational partner, provided

no further information. The talk density of a speaker bore

no relationship to the personality of his partner. This

still remains an intriguing question, perhaps indicating,

for example, if a particular kind of person would be more
effective in promoting talk with reticent, withdrawn

individuals. A suggestion comes from an observation by

Markel and Long (1974); that speakers of both sex speak more

to a female listener. Perhaps, listener personality dimen-

sions more directly associated with sex roles provide the

link to the talk density of a speaker.

Dyadic Units of Speech Channel Activity
Interactional Synchrony and Compatibility

The third hypothesis of the present investigation
represents a direct test of a conceptual model of two-

person interaction proposed by Argyle (1969). Two-person

interaction is conceptualized as an interdependence of
equilibrium processes describable at two distinct levels.

On the level of the interaction, equilibrium is described
by a condition of nearly all of the interaction time being
filled with speech dyadicc talk density) and only a small

proportion of the interaction time being occupied with








simultaneous speech (simultaneous talk density). The task

of the interactors is to achieve this state of equilibrium.
As such, this represents a condition of the two inter-

locutors working well together, i.e., establishing a

synchronous interaction. On another level, equilibrium
is describable in terms of the personal compatibility of

the two interlocutors, with the condition of compatibility

representing a steady state of harmonious co-existence and

mutual need satisfaction. It was hypothesized that there

was a relationship between these two levels of equilibrium,

but the results of the present investigation do not support
this contention.
The validity of Schutz's (1966) compatibility construct

has been satisfactorily demonstrated. This state of equil-

ibrium is simply and precisely defined by FIRO theory. Two

possible explanations for the negative results are apparent.
Either the notion of conversational equilibrium deserves

reconsideration or such a condition of equilibrium is not

achievable in initial interactions between strangers.
One of the specifications of conversational equilibrium

is the near absence of simultaneous speech. Further consid-

eration suggests a differentiation of simultaneous speech
into supportive and interruptive types. Supportive
simultaneous speech might consist of brief remarks of

agreement or interest which have the effect of maintaining
the partner's flow of talk. As such, this would not.seem
to indicate disequilibrium or poor meshing. However,







interruptive simultaneous speech, i.e., temporally con-

current remarks by one participant which result in the

partner's cessation of talk, seem an appropriate index of

dysynchrony. The present investigation confounded this

distinction. Regarding the stipulation that nearly all the

interaction time be filled with speech, there is an indica-
tion from the present investigation that an increased
density of talk is actually associated with one particular

type of incompatibility, rather than compatibility (total

control incompatibility bore a significant, positive

correlation with talk density). This suggests that Argyle's

(1969) notion of conversational equilibrium is, perhaps,
restricted to certain types of interlocutor compatibility.

It may be argued that the development of equilibrium
requires more intensive exposure to.the conversational

styles of the participants. Evidence that this may be so

comes from an investigation by Shaw and Sadler (1965).

Of their experimental groups (married-couples, dating

couples, unaffiliated couples), the married couples exhibited

a significantly smaller number of interruptions. Future

research might focus on repeated conversational encounters
between members of compatible and incompatible dyads. Per-

haps, compatible dyads would develop this equilibrium with
less exposure. In conclusion, there is no support for the

contention that the compatibility of interlocutors is

associated with a synchronous conversational interaction in
the initial encounters of unacquainted persons.







Taxonomy of Conversation

A fourth and final major hypothesis suggested that the
dimensions of dynamism and evaluation might further eluci-

date the psychological significance of the dyadic units of

concern. Speech channel relationships to evaluative ratings

were observed only insofar as they were applied to the

ratings of conversational partners. Evaluative ratings of

the conversation, implying interaction satisfaction, bore

no relationship to any of the five explored dyadic units.

The evaluative dimension, as applied to conversational

partners, may be interpreted as a measure of interpersonal

attraction, and the results of this investigation indicate
that the length of initial conversational encounters between

strangers is related to the degree of this interpersonal

attraction. Positive conversations, those high in inter-

personal attraction, were characterized by significantly

greater total conversation times than were Negative con-

versations. Dynamism ratings, as applied to partners, also

reflected the total conversation time. Dynamic conversa-

tions, in comparison to Static conversations, exhibited

greater total conversation times. It may be concluded that

the significance of one's conversational partner, i.e.,

both the joint attributions of power and energy as well as

the degree of interpersonal attraction, is related to the
total time of the conversation. In the writer's knowledge,

this represents the only known explication of the psycho-
logical significance of the length of initial encounters
between strangers.







Whereas, evaluative ratings of conversations did not

indicate any differentiations in terms of the dyadic con-
versational measures, dynamic ratings of conversations

indicated a clear cluster of relationships. Dynamic con-

versations, in comparison to Static conversations, exhibited

significantly greater densities of talk and simultaneous

talk, a significantly increased rate of simultaneous talk,

a significantly longer mean duration of simultaneous talk,
as well as a greater total conversation time. This
indicates that conversations, experientially differentiated

by the interlocutors, are associated with different temporal
structures. The relationship is between perceived conver-

sational power and energy and dyadic units of amount of

speech channel activity.

The results of the present study provide normative

criteria (Appendix F) for classifying conversations on a

dimension which specifies at least one behavioral difference.

As such, it represents an initial attempt at focusing and

differentiating the largely unexplored phenomena of mundane

conversation. Dynamic conversations may be construed
simply as lively, energetic conversations. Alternatively,

Dynamic conversations, in the sense of increased simultane-
ous talk, might be conceptualized as dysynchronous conversa-
tions. In the interests of specifying a holistic notion of
communication processes, obvious future questions might

focus on differences between Dynamic and Static conversa-
tions with regard to conversational content or the







characteristics of the other channels of face-to-face
communication.
Minor Findings

The methodology and normative findings associated with

the conversational measures deserves some comment. Regarding

the frequency and duration of vocalization, the reliabil-
ities reported in this investigation are appreciably higher
than those reported by other investigators (Kasl and Mahl,

1956; Jaffee and Feldstein, 1970). This is clearly the
result of the present chronographic technique, whereby each
interlocutor was individually monitored. Thus, increased

reliability of measurement is obtained at the price of
increased time for data reduction. As to frequency and

duration of simultaneous talk, Mahl (1959) had indicated
that the frequency of interruption could be reliably
obtained by listening to tape recordings. The present
investigation substantiates this claim, as well as demon-
strating that the duration of simultaneous talk is equally
amenable to techniques of perceptual .measurement. The

reliabilities observed for frequency and duration of
simultaneous talk in the present investigation are much
higher than those reported by Jaffee and Feldstein (1970).
Two measures of talk density were retained for the

final analyses in this investigation. TD1 is the obvious,
most popularly.employed measure of talk density, and the
observed mean value of .41 corresponds very closely with
that obtained in the Markel and Long (1974) investigation.







TD2 was suggested by Mahl's (1956) measure, the silence

quotient. The findings indicate that persons in uncon-
strained conversation speak an average of 68% of the time

available for them to speak, i.e., (TCT-partner vocalization
time). The present investigation does not suggest any
clear differential associations for these two measures,

but their observed correlation (.66) indicates that their
redundancy is not of sufficient strength to warrant the
dismissal of either measure. For the present, both are

considered viable operationalizations of talk.density and
future research might indicate their different significance.

The mean frequency and duration of vocalization were
included in the final analyses, though they were of concern

only for the derivation of the density measures. Still,
they are of interest in terms of comparison with previous
research findings. Verzeano (1950) examined monologue
speech samples via voice-actuated machine analysis, in
which a minimum pause of .5 seconds defined the speech
unit.- He noted an average speech unit duration of 2.82
seconds. The present study disclosed a mean vocalization

duration of 2.34 seconds. Since the speech units become
shorter as the size of the minimum defining pause is reduced,
it may be concluded that the chronographers of the present
study were responding to pauses of somewhat less than .5
seconds. The observed mean duration of vocalization is
consistent with Hargreaves's (1960) finding (2.6 seconds)
for friends in free discussion, but quite disparate from








Prebor's (1972) observed act duration of 1.15 seconds in

unconstrained conversation. The chronographers in the

Prebor (1972) study were apparently operating at a greatly
reduced reaction time. This variability suggests that the

highly reliable speech unit, described by Wiens et al.

(1966), might be more appropriate for research focused on
the analysis of individual speech acts, but the confounding

of speech and silence is not appropriate for concerns with

the total amount of speech channel activity.
Regarding the dyadic conversational measures, the

present study indicated that the state of either interloc-

utor vocalizing represents an average of 80% of the total

time of a conversation. This is somewhat higher than that

observed by Markel and Long (1974) for unconstrained con-

versation, but appreciably lower than the dyadic density

of talk reported for conversation constrained by role

differences (Matarazzo et al., 1968; Soskin and John, 1963).

The observed density of simultaneous talk (1.5%) is con-
sistent with that reported by Markel and Long (1974).

Jaffee and Feldstein (1970), employing an elaborately

sophisticated voice-actuated machine analysis, reported an

average duration of simultaneous speech of .4 seconds. The
present investigation, employing a simple human-operated
counter-timer, also disclosed an average simultaneous
speech duration of .40 seconds.
One final comment involves Prebor's (1972) reported

findings of an increased density of simultaneous talk in


~I~




87



female-female dyads and longer total conversation times for

male-female dyads. The effect of dyadic sexual composition

was examined in the present investigation by means of

analysis of variance, but.Prebor's (1972) findings were not

replicated, nor was dyadic sexual composition observed to
be related to any of the other dyadic conversational

measures.
Summary and Conclusions

Four major hypotheses, dealing with monadic and dyadic

units of amount of speech channel activity, were examined

in the present investigation. Regarding the first hypothe-

sis, a speaker's density of talk in unconstrained conver-

sation bears no relationship to his characteristic level

of interpersonal anxiety, as indicated by self-report of
interpersonal orientation. Assuming the validity of the

measurement of interpersonal anxiety (Schutz, 1966), and

the demonstrated activating effect of anxiety on speech

channel activity, an explanation is advanced based on the
characteristics of unconstrained conversation. The acti-

vating effect of anxiety has been demonstrated only with
.monologue speech samples and speech samples derived from

interview situations (e.g., Pope, Siegman and Blass, 1970).

In contrast, the task of unconstrained conversation is
characterized by an implicit demand for increased mutual

speech channel activity, i.e., reciprocity. Thus,
restrictions are imposed on the amount of talk in that
the speech channel is characterized by territoriality, i.e.,







speech by one interlocutor suppresses speech by the other

interlocutor. Given this situation, other channels of

face-to-face communication might be expected to be pre-

potent for the transmission of anxiety, in that temporally

concurrent behaviors of the interlocutors are the rule

rather than the exception. It is concluded that the acti-

vating effect of anxiety on speech production is not of

sufficient intensity to overcome these conditions of

reciprocity and speech channel territoriality in uncon-

strained conversation.

A second hypothesis represented an examination of the

relationship of individual talk density to certain situa-

tional factors of the interaction. It was observed that

participants who described their conversations as dynamic,

and who were described as dynamic by their conversational

partner, exhibited significantly greater densities of talk.

Furthermore, participants, in conversations described as

dynamic by their conversational partners, exhibited signif-

icantly greater talk densities. It is concluded that

individual talk density is an indication of the dynamic

dimensions of the situation, but that it is unrelated to

the evaluative aspects of the situation. The dynamism

dimension (Osgood et al., 1957), reflecting perceived power

and energy and representing one of two basic responses to

the environment, finds behavioral expression in the talk

densities of interlocutors in unconstrained conversation.

The present investigation provided no supporting


_s~l




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs