Title: Personality patterns and psycholinguistic differences in response to music
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00097839/00001
 Material Information
Title: Personality patterns and psycholinguistic differences in response to music
Physical Description: vi, 59 leaves. : illus. ; 28 cm.
Language: English
Creator: Resnick, Robert William, 1937-
Publication Date: 1967
Copyright Date: 1967
 Subjects
Subject: Psycholinguistics   ( lcsh )
Music -- Psychological aspects   ( 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 56-58.
Additional Physical Form: Also available on World Wide Web
General Note: Manuscript copy.
General Note: Vita.
 Record Information
Bibliographic ID: UF00097839
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 - 000559024
oclc - 13432897
notis - ACY4470

Downloads

This item has the following downloads:

personalitypatte00resnrich ( PDF )


Full Text












PERSONALITY PATTERNS AND

PSYCHOLINGUISTIC DIFFERENCES

IN RESPONSE TO MUSIC









By
ROBERT WILLIAM RESNICK













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










UNIVERSITY OF FLORIDA
December, 1967
































"TO LIZ AND CHRISTI"












ACKNOWLEDGMENTS


The author would like to express appreciation to the members

of the supervisory committee, Drs. Richard Anderson, Ben Barger, Hugh

Davis, Ira Gordon, and C. M. Levy, for their encouragement and suggestions

rendered during this study.

To Dr. Audrey Schumacher, Chairman, the author would like to

express his special thanks for her time, effort, and wisdom which was

both needed and appreciated.

The author would also like to thank Drs. H. T. Martin, Harry

Grater, and Pincus Gross for their help and counsel throughout his

graduate career.

Great appreciation is extended to Linda Harris, both for her

typing of this manuscript and her sincere interest and help in pre-

paring this study.

The author would also like to thank his mother, Sylvia Resnick,

the pianist who recorded the stimulus materials for this study.

Finally, the author would like to thank his wife, Liz, who

has given her help, her support, and her criticisms. Without Liz,

this study should probably still be in process.
















TABLE OF CONTENTS


ACKNOWLEDGMENTS . . . . .

LIST OF TABLES . . . . .

CHAPTER

I INTRODUCTION . . .

IT STATEMENT OF THE PROBLEM

III HYPOTHESES . . . .

IV METHOD . . . . .

V RESULTS . . . . .

VI DISCUSSION . . . .

VII SUMMARY . . . .

APPEN IX A . . . . . .

APPFNDIX B . . . . . .

APPE DI C . . . . . .

BIBLIOGRAPHY . . . . . .

BIOGRArIIC;AL SKETCH . . ...


Page

. iii

* V


. . . . . . . .

. . . . . . . .













LIST OF TABLES


Table

1 Directional Predictions of Linguistic Scores on the
Eight Linguistic Continua for the "Anxiety Neurotics"
(High A, High R) and "Psychopaths" (Low A, Lo- R)
Groups . . . . . . . . . . . . .

2 Means and Standard Deviations of the Raw Scores Made
by Groups High, Middle, and Low to the Twelve
Linguistic Variabls . . . . . . ... ..


Page


3 One-way Analysis of Variance
Words . . . . . .

4 One-way Analysis of Variance
per Sentence . . . .

5 One-way Analysis of Variance
Pronouns . . . . .

6 One-way Analysis of Variance
Referens/Total . . .

7 One-way Analysis of Variance
Referents/Personal Pronouns

8 One-way Analysis of Variance
Terms . . . . . .

9 One-way Analysis of Variance
Terms . . . . . .

10 One-way Analysis of Variance
Words . . . . . .

11 One-way Analysis of Variance

12 One-way Analysis of Variance


Total Number of
. . . . . . 28

Number of lords
. . . . . . 28

Per Cent Fersonal
. . . . . . 29

Per Cent Personal
. . . . . . 29

Per Cent Personal
. . . . . . 29

Per Cent Qualifying
. . . . . . 30

Per Cent All-ress
. . . . . . 30

Per Cent Negative
. . . . . . . 30

Per Cent Adjicatives . 31

Per Cent Verbs. ... . 31










Table Page

13 One-way Analysis of Variance for Verb/Adjective
Quotient. . . . . . . . .. . .31

14 One-way Analysis of Variance for Type Token Ratio . 32

15 Mixed Factorial Analysis of Variance for Total
Number of Words .. ... . . . . . . .34

16 Mixed Factorial Analysis of Variance for Nu-iber
of Words Per Sentence . . . . . . . .. 34

17 Mixed Factorial Analysis of Variance for Per Cent
Personal Pronouns . . . . . . . . ... 35

18 Mixed Factorial Analysis of Variance for Per Cent
Personal Referents/Total . . . . ... . .35

19 Mixed Factorial Analysis of Variance for Per Cent
Personal Referents/Personal Pronouns . . . . .. 36

20 Mixed Factorial Analysis of Variance for Per Cent
Qualifying Terms . . . . . . . . . 36

21 Mixed Factorial Analysis of Variance for Per Cent
Allness Terms . . . . . . . ... . . 37

22 Mixed Factorial Analysis of Variance for Per Cent
Negative Words . . . . . . . .... .. 37

23 Mixed Factorial Analysis of Variance for Per Cent
Adjectives . . . . . . . . ... ... . .38

24 Mixed Factorial Analysis of Variance for Pec Cent
Verbs . . . . . . . . ... .. . .38

25 Mixed Factorial Analysis of Variance for Ver'b/
Adjective Quotient . . . . .... ...... 39

26 Mixed Factorial Analysis of Variance for Tyrp
Token Ratio . . . . . . . . ... . . 39














CHAPTER I


INTRODUCTION


Relating linguistic behavior to personality has been termed

psycholinguistics by Osgood and Scboek (1954) who defined it as being

concerned with . relations between messages and the character-

istics of human individuals who select and interpret them." The

present study attempts to investigate whether personality differences

among subjects from an operationally defined normal population (i.e.,

male college students) can be identified by distinctive linguistic

responses to musical stimuli. The focus of this study is on the

relationship between personality and linguistic behavior although it

is recognized that culture too is importantly related to personality

and language.


Culture and language

Language has been recognized both as a means by which attitudes

and modes of life are molded, as well as being molded by the individual's

interactions with his environment. Edward Sapir, who is considered by

many to be the father of modern linguistics, asserted that . lan-

guage does not exist apart from culture, that is, from the socially

inherited ac~setblage of practices and beliefs that determine the texture

of oui lives" (Sapir, 1921; P. 221). A few years later Sapir wrote a












forceful synthesis of ideas describing the relationship between lan-

guage and culture.

Language is a guide to "social reality." . Human
beings do not live in the objective world alone, nor
alone in the world of social activity ordinarily
understood, but are very much at the mercy of the
particular language which has become the medium of
expression for their society. It is quite an illusion
to imagine that one adjusts to reality essentially
without the use of language and that language is merely
an incidental means of solving specific problems of
communication or reflection. The fact of the matter is
that the "real world" is to a large extent unconsciously
built up on the language habits of the group . .
Even comparatively simple acts of perception are very
much more at the mercy of the social patterns called
words than we might suppose. . We see and hear and
otherwise experience very largely as we do because the
language habits of our community predispose certain
choices of interpretation (Sapir, 1929; P. 209-210).

That the interaction works both ways seems also to be tenable

in that although a hypothetical ideal structure of a language exists at

any point in time, this structure is constantly in the process of

changing due to the influence of the culture and probably the influence

of individuals within that culture.

Benjamin L. Whorf, a former student of Sapir, has emphasized

the hypothesis that cross-cultural linguistic differences systematically

effect cross-cultural perceptual and cognitive ideational patterns.

This concept has a long history dating to the latter part of the eigh-

teenth century with Herder and Von Humboldt, and later with the philosopher

Ernst Cassirer and the linguists Johann Leo Weisgerber and Jost Trier

(Greenberg, 1961). Whorf rejected the concept of the natural logic of

language which states that language is merely a formalized index system

where parts and pieces are selected at will to build cognitions and












perceptions. Rather, it is the linguistic vessels available to the

individual which determine the limits of his cognitive and perceptual

fields (Whorf, 1956). Brown and Lenneberg summarize the Whorfian

hypothesis by saying that . language is not a cloak following

the contours of thought. Languages are molds into which infant minds

are poured" (Brown & Lenneberg, 1961; P. 481).

In testing the Whorfian hypothesis that linguistic habits deter-

mine how the world is viewed, Brown and Lenneberg (1961) had Harvard

and Radcliffe undergraduates name colors which were projected on a

screen. The stimuli (colors) were subdivided according to their wave-

lengths into those which were pure (primary) and those whose wavelengths

were in between the standard colors, the latter being referred to as

"colors-between-names." the length of the name given to each stimulus

by the subject, his response latency, and the inter-subject reliability

were measured. The "colors-between-names" had generally longer latencies,

less inter-subject reliability, and longer names. The latter was sig-

nificantly correlated with the lntency and reliability measures both

for inter- and intra-subject responses. Factor analysis revealed a

single factor, codability, which the authors stated is highly related

to cognitive and recognitional processes. Simply put: It is easier

to package a cognition for which one has a linguistic container. Other

studies (Carroll & Casagrande, 1958; Lorenz, 1953) also support the

Whorfian hypothesis.

Gestalt field theory maintains that language incorporates unique

ways of organizing the world. Their basic assumption is . .a con-

ceptual intermediary world in each language which is not simply a










reflection of the external world but an ideational organization of

this world, and which involves the speaker intimately and irrevocably

in the culture of his group. This Zwischenwelt is the product of long

years of cultural development" (Lafral, 1965; P. 112).

Language and personality

Nunnally suggests that . .individual differences in word

usage relate importantly to individual differences in learning, percep-

tion, and personality" (Nunnally, 1965; P. 203). Nunnally goes on to

assert that children learn words partly to solve problems. "The child

learns to say 'water' because this will provide the required form of

reinforcement more rapidly than will crying, random motions, or the use

of any other word" (Nunnally, 1965; P. 205-206). The more appropriate

responses (words) would therefore gain habit strength with inappropriate

responses losing habit strength. It is important to note that the

appropriateness of a response is based on the consequences contingent

upon that response. Whether a child learns linguistic behavior in a

classical or instrumental model, the behaviors become parts of his

instrumental repertoire which differentially elicit responses from his

environment.

With respect to the English language in particular, Alfred

Korzybski (1951) asserted that our Aristotelian, subject-predicate

thinking gives us a warped and rigid world outlook. He posited non-

verbal thinking within a non-Aristotelian system as the answer which

would loosen our linguistic shackles. (This position, in part, appears

to be a specific application of the psychoanalytic "regression in the











service of the ego," where an individual returns to primary, non-verbal

processes.) Korzybski's theory of General Semantics implicitly states

that our linguistic system not only orders our perceptions but stultifies

and stagnates them. It is the synthesis of some of the thinking of

Sapir (1921, 1929), Whorf (1956), and more tangentially, Korzybski

(1951) which has been the impetus for this present investigation.

Since linguistic patterns seem to be inexorably linked to percep-

tion and cognition, it would seem tenable to hypothesize that individual

personality differences would lead to linguistic differences within a

culture. Berg and Adams state that "the relationship between speech

and personality has long been recognized" (1962; P. 66). They point

out that most experimental attention has been addressed to the area

of what (content) is said, rather than how (structure) it is said.

Sanford (1949, as cited in Berg & Adams, 1962; P. 66) states that

" . language traditionally, has been regarded as the 'vehicle of

thought' with the thought attracting more attention than the vehicle."

It would seem that structural characteristics of language are much less

in the subject's awareness and hence are less subject to his volitional

manipulations. Eisenson posited that "speech is most importantly used

by man to make social adjustments, to elicit those responses from his

environment which further his ends" (1938; P. 172). He stated that

if it accepted that personality development consists essentially of

the adjustments made by persons to their environment, and that if it

is true that linguistic development is used most importantly to make

social adjustments, then language and personality are inseparable

(Eiscnson, 1938).











One of the earliest experimental investigations of the struc-

tural characteristics of language as related to personality was carried

out by Buscmarnin 1927 (as cited in Berg & Adams, 1962). He found that

an increase in verbs as compared to the number of adjectives (verb/

adjective quotient) was positively correlated with teachers' ratings

of "emotional instability," while the percentage of adjectives correlated

negatively with the same external criteria. Active (verb) responsive-

ness was therefore associated with instability while qualifying and

individuating (adjective) responsiveness was associated with the reciprocal.

Balken and Masserman (1940) used the verb/adjective ratio to in-

vestigate the language of psychiatric patients in three diagnostic cate-

gories. The stimulus was a standard presentation of the T.A.T. cards,

then called the Morgan-Murray cards. They then selected five obsessive-

compulsives, five conversion hysterics, and five anxiety states,

matching all for I.Q. The verb/adjective quotients obtained were 2.17,

1.35, and 3.11 for the compulsives, hysterics, and anxious patients,

respectively. These three quotients arc significantly different from

each other, and in that respect at least, discriminate among diagnostic

categories. Congruent with generally accepted psychological theory,

the anxious subjects were dramatic and active, while the hysterics

were less active and more descriptive.

Lorenz and Cobb (1953) compared the spontaneous speech accom-

panying responses to the T.A.T. for ten hysterical patients and ten

control subjects. The hysterics used more verbs and pronouns, fewer

adjectives (and therefore a higher verb/adjective quotient), conjunc-

tions, prepositions, and articles. In another study by the same











investigators (Lorenz & Cobb, 1953), they reported the language habits

of ten manic and ten control subjects. The manics, like the hysterics,

used more pronouns, verbs,- and fewer adjectives. Berg and Adams sum-

marize the latter Lorenz and Cobb study as showing that . the manic

speech appeared to be relatively repetitive and homogeneous, unlike nor-

mal speech which qualifies and individualizes" (1953; P. 67).

Mann (1944) found a higher adjective/verb quotient for 30 under-

graduate freshmen (.51) than for 30 schizophrenic hospitalized patients

(.43). This ratio is the reciprocal of the one mentioned in the pre-

ceding studies and therefore is in agreement. That is, the verb/adjective

quotient was again higher for the group with pathology.

The Type Token Ratio (TTR) as devised by Wendell Johnson (1944)

compares the number of different words (the types) with the total number

of words (the tokens). Mann (1944) has shown that normal subjects have

a higher TTR than schizophrenic patients for spoken language. Fairbanks

(1944) has shown the same thing for written language samples. She also

found that the patients used personal pronouns for 10.4 per cent of

their total words, compared to only 3.7 per cent for the freshmen. In

addition, the patients used significantly more verbs, but fewer nouns,

adjectives, conjunctions, prepositions, and articles. The patients

used 1,087 negative words compared to only 484 for the freshmen, as

well as doubling the freshmen frequency for the first person singular

pronoun "I" (Fairbanks, 1944).

Osbood and Walker (1959) found significantly rore "Allness" and

"Qualifying" terms in suicide notes than in their controls (written let-

ters). "Aliness" terms are extreme and polarized statements or words












such as always and never. "Qualifying" terms are words which tend to

modify and reduce the author's intensity of commitment, such as aDproxi-

mately and perhaDs. Balken and Masserman (1940) found their hysterical

group was highest on Allness terms and lowest on Qualifying terms, while

the obsessive-compulsive group used the fewest Allness terms and the

greatest number of Qualifying terms. Anxiety state patients were in

the middle for both dimensions. The authors interpreted these results

as supportive to the theory that hysterical patients characteristically

are not vague or ambivalent in their adjustment, whatever else they

may be. They saw the obsessive-compulsives' performance as indicative

of the ambivalences and uncertainties manifest in their fantasies.


The stimulus situation and language

Perhaps inconsistencies of linguistic performance within and

between studies would seem to be at least in part a function of the

divergent stimuli employed. Brodsky (1964) empirically supported this

hypothesis by showing significantly different linguistic scores for a

self description and an impersonal description of a picture of a room.

Other studies (Boder, 1940) have demonstrated that differences in

linguistic performance are a function of the type of writing being done

(e.g., legal documents versus novels).

In this current investigation, the particular nature of the

auditory stimulus was such that the subject hopefully chose to be either

"ego involved" or "objective." Music was chosen as the stimulus in this

investigation for its emotional "pull" and meaningfulness which are,

'however, for the most part non-specific. Musical stimuli have the











unique attribute of being over-learned yet mostly non-specific. Musical

stimuli are not so easily labelled as are visual stimuli. We have more

verbal labels for describing visual phenomena, and since music is less

related to linguistic labels, there is consequently less specific pull

from such stimuli. It would follow that since musical experiences are

not able to be named so easily as visual experiences, differences in

description would be more a function of the subject's responding.

In terms of human ontcgeny, response to sound is one of the first

response modalities noted in the neonate. Indeed, there seems to be

some evidence for even prenatal responses to auditory and tactile stimuli.

Music, historically, has played a great role in the emotional lives of

people. History points out that music is one of the forms that man has

used to express his joy, his sorrow, his hope, and his love. "The power-

ful and immediate connection of musical experience, and the many indica-

tions that unconscious needs gain satisfaction through this medium,

have long pointed to measures of musical preference as effective avenues

to deeper aspects of personality" (Cattell & McMichael, 1960). This

basis for selection of the specific music employed in this study will

be discussed in the methodological section containing a description of

the stimulus materials.


Personality characteristics and linguistic characteristics

With the empirical evidence regarding the relationship between

language and personality reported above, the theoretical rationale for

the personality groups employed in this current study will be explored.

Fenichel (1945) distinguishes between psychoneurotics and

character disorders largely on the respective presence or absence of













anxiety and repressive mechanisms. Psychoneurosis is grounded in con-

flict, the impetus for anxiety, while character disorders are characterized

by their lack of internal conflict and hence lack of anxiety and repres-

sion. Welsh (1956) empirically supported this hypothesis when he found

via factor analysis that his 'A' scale (Anxiety) had its highest loading

on scale 7 (psychasthenia) of the Minnesota Multiphasic Personality

Inventory, while scale 4 (psychopathy) had no significant loading. In

addition, while character disorders are seemingly immune to social pres-

sures, psychoneurotics are continually brought into conflict by them.

The psychoneurotics seem to be more "other directed," while the character

disorders appear to be almost totally "inner directed." Theoretically,

it would seem that personalities with high anxiety and need for social

approval components would be qualitatively and quantitatively different

from non-anxious, socially disdainful personalities, in terms of lin-

guistic habits.

It is evident that people sometimes differ as to their linguistic

habits both cross-culturally and according to psychological diagnoses.

Clearly this is to a considerable extent environmental, cross-culturally.

All members of a culture are exposed to approximately the same language

system. Within a culture, however, individual members may differ in

their "selection" of linguistic habits. To say that a culture's language

habits are homogeneous is not to say that there are no individual dif-

ferences. Just as different languages have some commonalities, yet

retain very real differences, individuals or groups within a single lin-

guistic system may also retain some differences.












More specifically, if psychoneurotics (e.g., High A, High R on

the 1DPI) have found by selective learning that certain linguistic

habits are more efficient (reinforcing) in terms of satisfying their

systems of needs, defenses, anxieties, and satisfactions, then their

linguistic characteristics may so demonstrate. While language habits

may be commonly available to all members of a linguistic group, it does

not follow that all members of that group will embrace all of the habits.

Explicitly, members of a linguistic culture are exposed to approximately

the same linguistic habits. However, individuals respond to the same

constellation of linguistic stimuli in terms of their idiosyncratic

system of needs, defenses, and anxieties in order to maximize adaptation.

Therefore, it might be expected that people with similar needs, defenses,

and anxieties (personalities) are more like other individuals with

similar need systems than they are to individuals with divergent sys-

tems. In experimental terms, the variance between personality goups

would be expected to significantly exceed the variance within personality

groups. People who share personality patterns may also share linguistic

habits due not necessarily to sharing needs at that time, but due to

the relative homogeneity of their past and present environments.


Anxiety and repressor scales

Welsh (1956) reports that dozens of factor analytic studies

essentially agreed as to the loadings of the first two factors on the

Minnesota Multiphasic Personality Inventory although the interpretations

and the names assigned to these factors vary. The first factor shows

very high loadings on scale 7 (Pt) and scale 8 (Sc) with negative











loadings on K whenever it is used. The second factor shows fairly high

loadings on the neurotic triad (scales 1, 2, and 3), especially scale 3

(Hy) and scale 2 (D), with negative loadings frequently appearing on

scale 9 (Ma). It is important to note that the .. similarity of

loadings on these two factors has appeared despite the varying popula-

tions employed" (Welsh, 1956; P. 264). Welsh calls these first two

factor scales on the MMPI Anxiety (A) and Repressor (R). The A and R

dimensions were employed to select subjects for the current study due

to their high reliability and their more total representation of variance

on the MMPI than almost any combination of clinical scales. The dis-

advantage of using'factor analytically derived scales lies in the

uncertainty of what dimensions this neat cluster of loadings represents.

Pragmatically, the factor becomes really useful after it has been iso-

lated and then studied both retrospectively and lontittudinally.

Dahlstrom and Welsh (1960) and Hathaway and Meehl (1951, as

cited in Dahlstrom & Welsh, 1960) describe High A, Kigh R's on the MMPI

as anxious, depressed, tense, nervous, and obsessive-compulsive. Guthrie

(as cited in Dahlstrom & Welsh, 1960), in studying medical patients with

the above symptoms, found 40 items on the MPI which were characteristic

of this group. Almost all of the 40 items are from clinical scales 2

and 7, particularly the items on those scales which also appear on

Welsh's A scale. Welsh (1956) describes people with elevated A and R

scores as usually diagnosed as anxiety states, manic-depressive, de-

pressed, or reactive-depressive.

Welsh (1956) describes people with low A and R scores as usually

diagnosed as behavior and/or character disorders, alcoholics, or








13



manic-depressive, manic. Hathaway and Meehl (1951, as cited in Dahlstrom

& Welsh, 1960) state that most psychiatric patients with this pattern

are diagnosed as psychotic with manic disorders predominating. Although

a significant subgroup were categorized as conduct disorders, neurotic

disorders were almost non-existent. It is important to note that

Hathaway and Meehl are describing a psychiatric population and not a

college population.















CHAPTER II


STATEMENT OF THE PROBLEM


Linguistic differences have been shown by cross-cultural studies

to be related to cultural differences in cognitions and perceptions and,

inferentially, to attitudes (Brown & Lenneberg, 1961; Whorf, 1956;

Carroll & Casagrande, 1958). While a common language tends to make

for common attitudes, differences in attitudes might be expected to

make for language differences within a common language.

Certain language differences have been related to psychological

diagnoses (Busemann, 1927; Balkan & Masserman, 1940; Chotlos, 1944;

Fairbanks, 1944; Mann, 1944; Lorenz & Cobb, 1953, 1953, 1954). The

current research used subjects selected from the normal population ac-

cording to personality characteristics as measured by the A and R scales

on the MMPI. Age, sex, and educational history of the Ss were held

constant across conditions. Musical themes were selected as the stimuli,

since as, previously noted, it was expected that music would lend more

freedom from specific verbal labels due to music's being mostly over-

learned yet non-specific. Structural linguistic scores were selected

for the dependent variables due to their previous fruitful use, high

reliability, and the fact that subjects are largely unaware of the

structure of their language and therefore less able to volitionally

manipulate it.









Linguistic dimensions

The linguistic dimensions employed in this investigation were:

1. Total Number of Words

2. Number of Words Per Sentence

3. Per Cent Personal Pronouns

4. Per Cent Qualifying Terms

5. Per Cent Allness Terms

6. Per Cent Negative Words

7. VAQ (verb/adjective quotient)

8. TTR (type token ratio)

The criteria and procedure for judging, coding, and tabulating

these dimensions are described in the Methods section of this paper.

For the purposes of this investigation, the term "anxiety neurotic" will

be synonomous with the High A, High R factors on the MMPI, while the

term "psychopath" will be synonymous with the Low A, Low R pattern on

the IyPI. Anxiety neurotics and psychopaths are therefore operationally

defined by their A and R scores on the E=I. Persons who differed on

these two personality dimensions were compared on a number of linguistic

dimensions with the prediction being made that they differ significantly

along these dimensions. Each of the linguistic dimensions will be dis-

cussed in turn, in terms of the theory underlying the particular direc-

tional prediction. The predictions are as follows:

Total Number of Words.--Since psychopathic personalities are

described as those finding it difficult to expend large amounts of

effort in tasks imposed upon then by others, it is predicted that they













will score lower on this dimension than the anxiety neurotics who will

probably feel obliged to do as much as they have time for, once they

get started. 0

Number of Words Per Sentence.--Although it is predicted that

psychopathic personalities will use a lower total number of words, it

is felt that they will use more words per sentence. In addition to the

anxiety neurotic's need for closure and need to restrict external stimu-

lation, the psychopath may not write many sentences for the experimenter

but will probably write longer individual sentences since this IS a task

he has decided to accept. Once he has decided to accept a single

sentence as a personal task, it is felt that he will perform that task

as if it were self-imposed.

Personal Pronouns.--Since the psychopathic personality is seen

as being more "inner directed" and would seem to mold his perceptions

egocentrically, and since the anxiety neurotic appears to be more "other

directed" and less egocentric at least when dealing with the outside

world, it is predicted that psychopathic personalities will use more

personal pronouns than anxiety neurotics.

Qualifying Terms.--Since this dimension may be viewed as a

function of the VAQ, it is predicted that anxiety neurotics will score

higher on this dimension than will psychopathic personalities. Quali-

fying terms restrict commitment, and would therefore seem to be appro-

priate tools for the anxiety neurotic.

Allness Terms.--The prediction for this dimension is that anxiety

neurotics will score lower than psychopaths. Since allness terms are











described as an index of commitment, it would seem that the cautious

predilection of the anxiety neurotic would negate a high percentage of

illness terms. The psychopathic personality, on the other hand, views

his world in more black and white terms, and therefore is expected to

score higher on this dimension.

Negative Words.--For this linguistic dimension, it is predicted

that anxiety neurotics will score higher than psychopaths. The former

sees his world as having faults and therefore to be negated. Psychopaths,

on the other hand, seem to be more active upon their world than reactive

to it.

Verb/Adiective Quotient.--The prediction for this linguistic dimen-

sion is that anxiety neurotics will have a lower VAQ than will psycho-

pathic personalities. As with qualifying terms, anxiety neurotics

are concerned with modifying their actions so as to offend as few as

possible.

Type Token Ratio.--Psychopathic personalities are predicted to

be higher on this dimension than anxiety neurotics since they anxiety

component of the latter would tend to induce perserveration and restric-

tiveness. Since the TTR is the number of different words per unit of

words, then the relatively "uninhibited" psychopath would tend to score

higher.














CHAPTER III
*

HYPOTHESES


Primary hypothesis

The general hypothesis proposed is: Persons in a normal

population with anxiety neurotic and psychopathic personality patterns

as measured by the A and R scales on the MMPI can be differentiated by

psycholinguistic analyses of responses to musical stimuli.


Secondary hvyotheses

More specifically, the following direction of differences on

the eight linguistic continue previously discussed are summarized in

Table 1.


Table 1

Directional Predictions of Linguistic Scores on the Eight
Linguistic Continua for the "Anxiety Neurotics" (High A,
High R) and "Psychopaths" (Low A, Low R) Groups




Groups

Linguistic Variables "Anxiety Neurotics" "Psychopaths"
High A, High R Low A, Low R


1. Total Number of Words High Low
2. Number of Words/Sentence Low High
3. Personal Pronouns Low High
4. Qualifying Terms High Low
5. Allness Terms Low High
6. Negative Words High Low
7. Verb/Adjective Quotient Low High
8. Type Token Ratio Low High











CHAPTER IV


METHOD


Subjects

The subjects [or this investigation were 50 male students in

the University College of the University of Florida. When they entered

the University, all incoming freshmen were routinely given the Minnesota

Multiphasic Personality Inventory. The three groups were selected

according to their Anxiety (A) and Repressor (R) scores on the MMPI. A

and R are the first two factors of the MPI and together account for

most of the variance on that instrument (Welsh, 1956). The High, Middle,

and Low scoring groups consisted of 20, 12, and 18 Ss, respectively.

The High A and R ("anxiety neurotics") group consisted of Ss

whose A score on the Y>PI was at least one standard deviation above the

population mean of A and whose R score was above the population mean of

R. The mean A and R scores for this experimental group were 64.9 and

53.6, respectively, with standard deviations of 4.2 and 4.3. The popu-

lation means for A and R are both 50.

The Middle A and R normalsls") group consisted of Ss whose A and

R scores were both at (or near) the A and R means. The mean A and R

scores for this group were 49.9 and 49.9, respectively, with standard

deviations of 1.2 and 2.1. This group was included in order to provide

base line data.










The Low A and R ("psychopathic characteristics") group consisted

of Ss whose R score on the IMPI was at least one standard deviation

below the mean of R and whose A core was below the mean of A. The mean

A and R scores for this experimental group were 42.5 and 36.6, respec-

tively, with standard deviations of 5.4 and 3.4.

Freshmen at the University of Florida are relatively homogeneous

in age, educational background, and socio-economic status, and the three

groups were considered to be matched on these variables. Occasional

differences were assumed to occur randomly among the three groups.


Stimulus materials

The stimulus constellation consisted of two presentations of

seven discrete musical themes played on a piano and recorded on audio

tape. The seven themes were recorded with an inter-stimulus interval of

ten seconds. The themes were adaptations of those which accompanied

silent movies. This music was chosen for the following reasons based

on pilot work done by this writer in 1962:

1. Each there in the set seemed to express an exaggerated
emotion.

2. The themes were arranged so as to follow a sequence which
lends itself to either a continuous story or several dis-
crete stories.

3. The same music was used for many different plots, since it
attempted to capture a feeling tone rather than an event.

4. Subjects found it relatively easy to respond to these stimuli.

5. As a free response stimulus, movie music appeared "safe" to
Ss since many of them did not feel they were putting them-
selves into the stories--they were just re-telling the
"traditional old time movie story" (if they recognized that
the stimuli were movie music).











The

employed in

as follows:


seven musical themes generally follow the basic sequence

most silent movies. They are, as specifically as possible,


1. Opening them. Non-specific, ethereal. "Sets the stage"
for any action to follow.

2. Statement of female (usually heroine). Adaptation of
"Sweetheart of Old Sigma Chi."

3. Statement of male (usually hero). Adaptation of "My Hero."

4. Statement of conflict (usually villain or situational).

5. Statement of action (usually chase and/or flight) terminating
in resolution of conflict.

6. Reunion of hero and heroine, no inference as to future.
Adaptation of "Kiss iMe Again."

7. Statement of future (positive or negative). Adaptation of
"Hearts and Flowers."


Procedure

The Ss were invited by letter to participate in an experiment in

the "Psychology of Music." A copy of the letter sent can be found in

Appendix A. They were asked to report to the Language Laboratory at the

University of Florida. This laboratory contains 154 individual cubicles

equipped with earphones to monitor any of three channels transmitted

from a small room in the rear, housing tape playback equipment. Each S

was given a booklet containing a mimeographed identification cover sheet

and ten pieces of lined loose-leaf type paper. After checking earphones,

tape equipment, etc., all S were read the following instructions by the

writer, who stood at the front of the room:












You are going to hear a group of musical pieces
played twice. The first time you hear then I want you
just to listen. Do not write anything until you are
told. The second time yo'u will hear the musical pieces
in exactly the same order and you will have time between
each piece to write a story that the music suggests to
you.* I will tell you when the music is about to be
played for the second time. You may write one story or
several short stories. Remember, the first time you
just listen to the music, and the second time write a
story based on what the music suggests to you. Any
story will be fine. Please do not remove your earphones
or talk to anybody else. If you have any questions,
please raise your hand now.

At the end of the session, all Ss were asked to indicate, without con-

ferring with anyone, what kind of music they had heard. The purpose

of this was to determine whether recognition of the stimulus materials

as old time movie music varied among the three groups. In addition,

Ss were asked to indicate whether they were "tone deaf" or had any

hearing problems of which they were aware. After the booklets were

collected, interested Ss were asked to remain to hear an explanation

of the experiment.


Tabulation of dependent variables

A total of about 25,000 responses (words) were elicited from

the Ss of this investigation. Each of these responses were judged,

coded, and tabulated twelve times, once for each of the linguistic

dimensions which were the dependent variables in the analyses. The

procedure for judging, coding, and tabulating is outlined below.

Total Number of Words.--Each group of letters separated by

spaces on both sides from adjacent groups of letters was counted as



*The time allowed for writing after each musical theme was four minutes.











a word, even though it might be part of a place name, as in Des Moines

(two words), an initial, as in James A. Brown (three words), or a neolo-

gism coined by the subject. Any number was counted as one word; for

example, 125 was tabulated as one word. A hyphenated word was counted

as one word. Each time a word was used it was counted as an additional

word. Contractions were treated as one word; for example, can't was

tabulated as a single word. Abbreviations were also treated as one word

such as M.D.

Number of Words per Sentence.--Sentences were defined by their

punctuation, whether or not they were grammatically correct. The Total

Number of Words was divided by the Total Number of Sentences to yield

the Number of Words per Sentence.

Per Cent of Personal Pronouns.--Personal Pronouns were operationally

defined to be: I, me, my,myself, we our, us, you, she, he, her, his,

hers, him, your, they, their, them, herself, himself, themselves, yourself,

and ourselves. The Personal Pronouns were color coded, summed, then

divided by that S's Total Number of Words, and multiplied by 100 to

yield the Per Cent of Personal Pronouns.

Per Cent of Personal Referents.--Personal Referents were

operationally defined to be: I, my, me, myself, we, us, and our. The

Personal Referents were color coded, summed, then divided by that S's

Total Number of Words or Total Number of Personal Pronouns, and multiple

by 100 to yield Per Cent of Personal Referents/Total and Per Cent of

Personal Referents/Personal Pronouns.

Oualifying Terms.--Qualifying Terms were operationally defined

to be: almost, approximately, about, apparently, more-or-less, maybe,









especially, perhaps, somewhat, evidently, sometimes, could, probably,

possibly, except, around, may, at times, generally, slight, and appears.

The Qualifying Terms were color aded, summed, divided by that S's

Total Number of Words, and multiplied by 100 to yield the Per Cent of

Qualifying Terms.

Allness Terms.--Allness Terms were operationally defined to be:

all, never, completely, always, anything, nothing, entirely, nobody,

everybody, forever, utter, exactly, undoubtedly, anybody, wholly, every,

absolutely, everyone, no-one, no doubt, eternity, and ultimately. The

Allness Terms were color coded, summed, then divided by that S's Total

Number of Words and multiplied by 100 to yield the Per Cent of Allness

Terms.

Negative Terms.--Negative Terms were operationally defined to

be: no, not, never, none, neither, nothing, nobody, nowhere, and unable.

The Negative Terms were color coded, summed, then divided by that S's

Total Number of Words, and multiplied by 100 to yield the Per Cent of

Negative Terms.

Adjectives.--Adjectives were defined by regular classification,

and any verb form (i.e., participle) which the dictionary recognizes as

an adjective (Webster's Collegiate). The Adjectives were color coded,

summed, then divided by that S's Total Number of Words, and multiplied

by 100 to yield the Per Cent of Adjectives.

Verbs.--Verbs were defined as all simple verbs, participles plus

auxiliaries, gerunds and participles unless the dictionary recognizes

them as nouns and adjectives, as the case may be (Webster's Collegiate).

The verbs were color coded, summed, and then divided by that S's Total

Number of Words, and multiplied by 100 to yield the Per Cent of Verbs.








25


Verb/Adjective Quotient.--The VAQ was derived by dividing that

S's total number of Verbs by his total number of Adjectives.

Tyne Token Ratio.--The TTR was defined as the number of different

words per 100 words. These 100 word samples were then averaged to yield

a mean segmental TTR.














CHAPTER V


RESULTS


Four additional linguistic variables were culled from the data

since one of the measures (Per Cent of Personal Pronouns) was judged

to be too global in nature, hiding, rather than revealing relevant data.

Specifically, two additional variables dealing with personal pronouns

were added, namely, Per Cent of Personal Referents which was then com-

puted both as a percentage of the total number of words as well as a

percentage of all personal pronouns. The VAQ dimension was also

studied further by adding Per Cent of Verbs and Per Cent of Adjectives

to the variables. Thus, the total number of linguistic variables in

this study was twelve.

Table 2 shows the means and standard deviations of the raw

scores of the High, Middle, and Low groups on the A and R scales for

all twelve linguistic variables. As can be seen from Table 2, the

standard deviations for these linguistic continue are quite large, in

some cases exceeding their means (e.g., Per Cent Personal Referents/

Personal Pronouns). These large standard deviations indicate consider-

able overlap among the three distributions on any particular linguistic

variable.

























CC D -. 0 -(
'2) 4- -4 2) 4-


CD o

'-O


4-. 2)
44r 4 CC

~-4 -l
2) )


4 ) o .0 0 -4 0s







CA 00 rl co
ti CD)


C- 4 4-2 0


)a





02




o 0
o '-




-m2


u
2) ri

a "

CJ
0








a) -





0o



--4







n--4



r-4

C)C"


co
0 0


vn r


N,


r-


E-4


I-
0 O *
0 0 -
p 2 2 '



2) o)2 H 2)



2) C 2 0 0 ,-i 2) 4-' 2) 2)



H d C 0 B 2 5* C1 5 >
r8 Ol a ~ i T M N N N N N N


-n o

r-


C-) C-) 42
-f 0 C-

4- 4)


C .-4
0


2

a
2);

2)


C-




2)
-4
-2)




2)
-4
'-4
2)l
2)
(4

--4

2)
2)

2)
-4


'2) -4
Dc CC C

4-- 4.2











Tables 3 through 14 represent the one-way analyses of variance

for each of the twelve linguistic variables among the High, Middle, and

Low A and R groups. These analyses compared each group's mean on a

variable to the grand mean for that variable across all three groups.




Table 3

One-way Analysis of Variance for
Total Number of Words


Source d.f. Mean Square F*


Between 2 3973.53 .143
Within 47 27704.34
Total 49


*F .05 (2,47) = 3.53



Table 4

One-way Analysis of Variance for
Number of Words per Sentence




Source d.f. Mean Square F


Between 2 5.23 .205
Within 47 25.54
Total 49










Table 5

One-way Analysis of Variance for
Per Cent Personal Pronouns



Source d.f. Mean Square F


Between 2 9.02 .081


Within
Total


11.26


Table 6

One-way Analysis of Variance for
Per Cent Personal Referents/Total


Source


Mean Square


F


2.24


Betucen 2 11,23
Within 47 5.02
Toial 49


Table 7

One-way Analysis of Variance for
Per Cent Personal Referents/Personal Pronouns



Source d.f. Mean Square F


BEtween 2 887.14 1.07


824.25


WiLhin
Total










Table 8

One-way Analysis of Variance for
Per Cent Qualifying Terms


Source d.f. Mean Square F


Between 2 .151 A59
Within 47 .329
Total 49



Table 9

One-way Analysis of Variance for
Per Cent Allness Terms




Source d.f. Mean Square F


Between 2 .069 .082
Within 47 .848
Total 49



Table 10

One-way Analysis of Variance for
Per Cent Negative Words




Source d.f. Mean Square F


Between 2 .032 .030
Within 47 1.047
Total 49










Table 11


One-way Analysis of Variance for
Per Cent Adjectives


Source d.f. Mean Square F


Between 2 14.99 .553
Within 47 27.11
Total 49




Table 12

One-way Analysis of Variance for
Per Cent Verbs




Source d.f. Mean Square F


Between 2 2.93 .245
Within 47 11.95
Total 49




Table 13

One-way Analysis of Variance for
Verb/Adjective Quotient




Source d.f. Mean Square F


Between 2 .034 .191
Within 47 .179
Total 49










Table 14

One-way Analysis of Variance for
Type Token Ratio




Source d.f. Mean Square F


Between 2 5.48 .538
Within 47 10.20
Total 49


None of the Fs in Tables 3 through 14 indicates statistical

significance, suggesting that the variance among the three groups is no

greater than one would expect from chance fluctuations. Thus, any group's

mean on any linguistic variable does not differ significantly from the

grand mean for that linguistic variable. Finally, since none of the Fs

were significant, we assume that all observed differences among the three

groups are attributable to chance alone (Lindquist, 1953). It is note-

worthy, however, that the two variables related to Per Cent of Personal

Referents came closest to discriminating among the groups. This may be

related to previous research which indicated that psychiatric patients

used double the number of the pronoun 'I' than did college students

(Fairbanks, 1944). However, since the differences in the findings did

not depart from change, these comments are only raised very speculatively.

It is significant to note that none of the Ss in this

investigation indicated any hearing difficulties of which they were aware.

Furthermore, recognition of the stimulus materials as old time movie

music occurred in 30, 25, and 22 per cent of the High, Middle, and Low

A and R groups. These differences revealed a non-significant chi-square.











Thus, there appeared to be no differential recognition of the themes by

the three groups.


Additional analyses

The statistical analyses employed thus far were based on each

S's total scoreson each of the twelve linguistic dimensions. That is,

,although the Ss responded to seven discrete musical themes, their score

for any single linguistic variable were summed for all seven stimuli.

The question was raised as to whether the High, Middle, and Low groups

produced significantly different scores on a single or several musical

themes on any of the twelve variables. The approximately 25,000 responses

were retabulated and recomputed for each subject on each variable

separately within each of the seven themes, yielding 84 scores per S.

These data were then analyzed for each of the twelve variables separately,

using a mixed factorial design with two factors, one. of which has

repeated measures (Lindquist, 1953). This design tests the interaction

between musical themes and groups for each of the linguistic continue.

Tables 15 through 26 presents the analyses of variance for each of the

twelve linguistic variables.












Table 15

Mixed Factorial Analysis of Variance for
Total Number of Words


Source d.f. Mean Square F



Groups (G) 2 611.18 0.154
Error Between 47 3957.33 10.654
Themes (T) 6 714.79 1.924
G by T 12 521.22 1.403
Error Within 282 371.45
Total 349

* .05 (2,47) = 3.23
**F .05 (6, 282) = 2.10
***F .05 (12, 282) = 1.80



Table 16

Mixed Factorial Analysis of Variance for
Number of Words Per Sentence



Source d.f. Mean Square F


Groups 2 7623.69 0.365
Error Between 47 20873.26 4.883
Themes 6 8702.42 2.036
G by T 12 1913.51 0.448
Error Within 282 4275.02
Total 349


*F .05 (2, 47) = 3.23
**F .05 (6, 282) = 2.10
***F .05 (12, 282) = 1.80













Table 17

Mixed Factorial Analysis of Variance for
Per Cent Personal Pronouns


Source d.f. Mean Square F


Groups 2 106.01 1.357
Error Between 47 78.13 4.324
Themes 6 64.36 3.562**
G by T 12 12.87 0.712
Error Within 282 18.07
Total 349


'*F .05 (2, 47) = 3.23
**F .05 (6, 282) = 2.10
'**F .05 (12, 282) = 1.80



Table 18

Mixed Factorial Analysis of Variance for
Per Cent Personal Referents/Total



Source d.f. Mean Square F


Groups 2 80.84 2.425
Error Between 47 33.34 6.746
Themes 6 7.41 1.499
G by T 12 2.06 .418
Error Within 282 4.94
Total 349


.05 (2, 47) = 3.23
**, .05 (6, 282) = 2.10
***P .05 (12, 282) = 1.80













Table 19

Mixed Factorial Analysis of Variance for
Per Cent Personal Refesents/Personal Pronouns


Source d.f. Mean Square


Groups 2 6275.62 1.409
Error Between 47 4453.59 8.456
Themes 6 517.29 0.982
G by T 12 444.81 0.845
Error Within 282 526.69
Total 349


*F .05 (2, 47) = 3.23
**F .05 (6, 282) = 2.10
***F .05 (12, 282) = 1.80



Table 20

Mixed Factorial Analysis of Variance for
Per Cent Qualifying Terms



Source d.f. Mean Square F


Groups 2 0.98 0.802
Error Between 47 1.22 1.315
Themes 6 0.57 0.612
G by T 12 0.68 0.734
Error Within 282 0.92
Total 349


'* .05 (2, 47) = 3.23
**F .05 (6, 282) = 2.10
***F .05 (12, 282) = 1.80













Table 21

Mixed Factorial Analysis of Variance for
Per Cent Allness Terms


Source d.f. Mean Square F


Groups 2 1.44 0.339
Error Between 47 4.25 1.553
Themes 6 3.74 1.367
G by T 12 3.31 1.211
Error Within 282
Total 349


*F .05 (2, 47) = 3.23
**F .05 (6, 282) = 2.10
**F_ .05 (12, 282) = 1.80



Table 22

Mixed Factorial Analysis of Variance for
Per Cent Negative Words



Source d.f. Mean Square F


Groups 2 2.75 0.337
Error Between 47 8.15 2.062
Themes 6 1.70 0.431
G by T 12 1.63 0.411
Error Within 282 3.95
Total 349

*1 .05 (2, 47) = 3.23
**F .05 (6, 282)= 2.10
***1 .05 (12, 282) = 1.80













Table 23

Mixed Factorial Analysis of Variance for
Per Cent Adjectives


Source d.f. Mean Square F


Groups 2 31.86 0.179
Error Between 47 177.64 4.656
Themes 6 92.46 2.423**
G by T 12 42.69 1.119
Error Within 282 38.15
Total 349

*F .05 (2, 47) = 3.23
**F .05 (6, 282) = 2.10
***F .05 (12, 282)= 1.80



Table 24

Mixed Factorial Analysis of Variance for
Per Cent Verbs



Source d.f. Mean Square F

Groups 2 6.89 0.071
Error Between 47 96.65 3.893
Themes 6 57.84 2.330**
G by T 12 25.89 1.043
Error Within 282 24.83
Total 349


*F .05 (2, 47) = 3.23
**F .05 (6, 282) = 2.10
***F .05 (12, 282) = 1.80













Table 25

Mixed Factorial Analysis of Variance for
Verb/Adjective Quotient


Source d.f. Mean Square F


Groups 2 16554.50 .711
Error Between 47 23288.93 3.069
Themes 6 15930.50 2.099
G by T 12 4191.08 0.552
Error Within 282 7589.27
Total 349

*F .05 (2, 47) = 3.23
**F .05 (6, 282) = 2.10
***F .05 (12, 282) = 1.80



Table 26

Mixed Factorial Analysis of Variance for
Type Token Ratio



Source d.f. Mean Square F

Groups 2 357.68 1.162
Error Between 47 307.68 4.320
Themes 6 64.44 0.905
C by T 12 72.94 1.024
Error Within 282 71.22
Total 349

*F .05 2, 47) = 3.23
**~ .05 (6, 282) = 2.10
***F .05 (12, 282) = 1.80










Tables 15 through 26 show that none of the interactions between

Groups and Themes are significantly different from chance, indicating

that the High, Middle, and Low Awand R groups did not differ significantly

on any of the twelve variables within any of the seven stimuli.

Three variables, Per Cent Personal Pronouns, Per Cent Adjectives,

and Per Cent Verbs, did yield significant Fs at the .05 level for Thenes

(See Appendix C). This suggests that the three groups combined differed

in their linguistic behavior across themes on those variables. However,

the three groups did not differ from each other in their linguistic

behavior across themes as indicated by their interactions which did not

depart significantly from chance on any of the twelve linguistic variables.

It should be noted that the three significant Fs found for Themes con-

stitute only eight per cent of the 36 Fs computed and therefore their

significance may be due to chance factors. That possibility is reduced,

however, since all three significant Fs appeared on the same factor

(Themes).












CHAPTER VI


DISCUSSION

The results of this investigation as presented in the previous

chapter fail to reject the primary null hypothesis that linguistic

differences among the High, Middle, and Low A and R groups do not differ

significantly from chance. It therefore follows that the various

secondary null hypotheses are also not rejected, that is, none of the

directional differences predicted depart significantly from chance.

Since the latter hypotheses are secondary both in the ordinal and con-

notative sense, it would be superflous to comment about specific

directional predictions when the primary hypothesis predicting differ-

ences in any direction was not supported.

The results of this study suggest that male college groups with

the magnitude of discrepancy on the A and R scales of the M.PI such as

these do not show significantly different responses to old time movie

music on selected, formal, structural linguistic characteristics. The

negative findings of this investigation suggest the possibility of an

inherent deficit in the stimulus materials, population sampled, linguis-

tic variables or some combination of these three entities. Since these

language variables have been previously demonstrated to discriminate

significantly among groups (Balkan & Masserman, 1940; Fairbanks, 1944;

Mann, 1944; Lorenz & Cobb, 1953; Osgood & Walker, 1959), it seems reason-

able to assume that their lack of discriminative power in this study may

be a function of some other variable or variables.












The stimulus materials employed in this study, old time movie

music, may in part be related to the non-significant results of this

investigation. Brodsky (1964), has demonstrated that linguistic scores

are significantly effected by the particular stimuli employed. Boder

(1940), has also shown that linguistic scores differ depending on the

explicit purposes of the particular linguistic performance (e.g., novels

versus scientific writings versus legal documents). To the best of

this writer's knowledge there are no empirical studies published dealing

with linguistic scores as a function of music, although auditory pro-

jective tests have been developed by Skinner (1936), and Stone (1953).

Perhaps old time movie music, being a very specific example of the class

of auditory stimuli, may elicit structural linguistic scores which are

very different from visual stimuli such as T.A.T. cards. Specifically,

the frequency and range of exposure to old time movie music is some-

what restricted and possibly stereotypic. The differential exposure of

old tine movie music may result in shared and learned modal responses

across many sub-populations of the culture. It may be that language

describing old time movie music is as homogeneous as, for example,

biblical language or the linguistic idiom of jazz. Even though young

adults may never have seen a silent movie with piano accompaniment,

possibly many of them have heard this type of music satirically accom-

panying television skits and/or movie vignettes. Although frequency of

exposure to old time movie music probably varies, the sources of such

music are quite restricted, particularly since they have in some ways

become part of our present-day cultural mythology. Thus, old time movie











music may be too "labelled" and stereotypic to elicit differential

linguistic responses. The fact that on three variables, Per Cent

Personal Pronouns, Per Cent Adjectives, and Per Cent Verbs, the High,

Middle, and Low A and R groups did differ in their linguistic behavior

across themes although not among groups further supports the hypothesis

that old time movie music may be more stereotypic and have more labels

than was originally expected. This may not be, however, representative

of all music. Other types of music may prove extremely useful in the

future as stimuli in personality research, just as various types of

visual stimuli are differentially useful.

The second variable which may have heavily influenced the results

of this investigation is the location of the populations sampled on the

normal/deviant continuum. The samples employed in this study were

roughly equivalent to a trichotomous division of the normal range in

many other studies. For example, Mann (1944), found significant linguistic

differences between college freshmen and hospitalized schizophrenic

patients. This present study attempted to discriminate among personality

types with a population similar to the freshman half of Mann's Ss.

Clearly the total range of pathology for all of Mann's Ss is far larger

than the range for Ss in this current study. In other words, the range

of Ss sampled for this study spanned only one-half of that involved in

former studies. Perhaps samples drawn from the operationally defined

normal population (e.g., students) may not demonstrate behavioral dif-

ferences which have been demonstrated when these samples were combined

and then compared to deviant populations (e.g., hospitalized schizo-

phrenics) as Mann did in her study.












A third variable which may have influenced the results of the

present investigation is the relative cultural homogeneity of the popu-

lations sampled. Specifically, twenty-year-old college students may

share many more mores, habits, attitudes, and interests with each other

than they probably do with other twenty-year-old males who did not go

to college. Speculatively, college males and other males may have been

differentially reinforced for their linguistic behavior and therefore

at any point in time would probably respond differentially to stimuli

eliciting a linguistic response. The linguistic homogeneity of various

young peoples' sub-cultures (peer groups) within the entire population

may be seen as a rough indication of their values, mores, and attitudes.

For example, the language of an urban street gang member is generally

expected to differ from a college student.

Behavioral responses, in this case linguistic performance, are

determined by situational, peer reference, and cultural factors as well

as personality factors. The Ss in this investigation were selected to

be alike in sex, age, culture, and educational level in order to minimize

the possible variance from these sources, thereby attributing any vari-

ance which did occur in excess of chance to their personality differences.

It may be, however, that controlling such factors as sex, age, culture,

and educational level resulted in a group of subjects who may be a good

deal alike on many personality characteristics aside from A and R. Thus,

the homogeneity of responses (linguistic behavior) of the college students

in this investigation may be but a single expression of their more total

sub-culture homogeneity.











Thus far, the three hypothesized problem areas in this

investigation have included the properties of the stimulus materials,

the limited range sampled on the normal/deviant continuum, and the

relative cultural homogeneity of the populations sampled. While these

three possibilities surely do not exhaust the possible explanations of

negative results, they do seem to this writer to represent the most

basic, first order problems.

The relationships among culture, personality, and linguistics are

clearly complex. The results of this current study suggest that within

the normal population, culture may be a more potent determinant of strucural

linguistic behavior than personality. This seems tenable since the

three groups in this study did not differ from each other acorss themes

on any of the twelve linguistic variables, yet collectively differed

across themes on three of the variables. By contrast, some studies using

hospitalized psychiatric patients have previously demonstrated substan-

tially different linguistic behavior than normals. However, it is again

difficult to know if the source of their linguistic differences is due

to personality and/or cultural factors. In a very real sense, hospital-

ized psychiatric patients are as culturally deviant as they are "emotion-

ally" deviant. Operationally, people are hospitalized when they cannot

be efficiently maintained in their culture. Speculatively, it is sug-

gested that the more culturally deviant a person is, the more his language

may be determined by his personal needs rather than the cultural and/or

situational factors present.











Since the linguistic variables have previously demonstrated

their ability to discriminate between clearly divergent populations

with other stimulus materials, then their failure to so discriminate

in this study may be due to the difference in populations and/or

stimulus materials. Therefore, to replicate this study using the same

stimulus materials (i.e., old time movie music) with groups who have

already demonstrated linguistic differences in response to other types

of stimulus materials (i.e., T.A.T. cards) would probably shed some

light on whether the non-significant results in the current investi-

gation were a function of the stimulus materials. Another, and perhaps

easier investigation of this problem would be to administer T.A.T. cards

(which have already demonstrated their ability to elicit differential

linguistic scores from more divergent populations) to the same student

population employed in this current investigation. The results of

such an investigation would suggest if the findings of this present

study were a function of the stimulus materials or the population

samples.













CHAPTER VII


SULTARY


This study investigated whether selected, formal aspects of

linguistic behavior in response to musical stimuli could reflect person-

ality differences in anxiety and repression in a young adult, non-

patient population.

The primary hypothesis proposed was that subjects in a normal

population with anxiety neurotic and psychopathic personality character-

istics as measured by Welsh's Anxiety and Repressor scales can be

differentiated by psycholinguistic analysis of responses to musical

stimuli. The secondary hypotheses concerned directional predictions

of the differences.

Three groups of male Ss were selected from a college population

based on their scores on Aelsh's Anxiety and Repressor scales on the

Minnesota Multiphasic Personality Inventory. These 50 Ss were charac-

terized by their High, Middle, and Low A and R scores. The stimulus

materials consisted of seven musical themes which have been used in

the past to accompany silent movies. Old time movie music was selected

because the various themes seemed to elicit effective responses while

at the same time being relatively easy for Ss to respond to. In addition,

old time movie music had been used to depict feeling tones rather than

specific events. All Ss were asked to write stories that the music

suggested to them and linguistic scores were derived front these responses.










The linguistic variables employed were Total Number of Words, Number of

Words Per Sentence, Per Cent Personal Pronouns, Per Cent Personal

Referents/Total, Per Cent Personal Referents/Personal Pronouns, Per Cent

Qualifying Terms, Per Cent Allness Terms, Per Cent Negative Words, Per

Cent Adjectives, Per Cent Verbs, Verb/Adjective Quotient, and Type Token

Ratio.

Twelve one-way analyses of variance indicated that none of the

observed linguistic differences among the three groups departed signifi-

cantly from chance. Twelve mixed factorial analyses of variance indi-

cated that the three groups did not differ significantly from each other

in their linguistic behavior across the seven musical themes. On three

of the linguistic variables, however, the three groups collectively

demonstrated significantly different linguistic behavior across the

seven musical themes at the .05 level. As possible explanations of

these findings it was suggested that:

1. The limited range of Ss on the normal/deviant continuum was
too restricted to demonstrate differences as have already
been demonstrated with clearly divergent groups.

2. Certain properties of the stimulus materials themselves led
to stereotypic responses.

3. The relative homogeneity of the populations sampled in terms
of sex, age, culture, and educational history may have re-
sulted in Ss who were very much alike on many personality
factors and behavioral predilections aside from A and R.

It was speculatively suggested that the more culturally deviant

a person is, the more his language may be determined by his personal

needs rather than the cultural and/or situational factors present








49


Proposals for further research were made which would suggest

which of these explanations, if any, were influential to the results

of this investigation.



































APPENDIX A


LETTER OF INVITATION TO SUBJECTS














May 15, 1965



Dear Mr.

As you well know, the University of Florida encourages a great many
research projects which depend upon the participation of students re-
quested to serve as subjects. Such an investigation is presently being
conducted within the Depart-ent of Psychology and selected subjects
are urged to participate.

From the tests you took last September, we have determined that you are
one of the students whose brief cooperation is necessary for our current
research. The particular experiment in which you are urged to be a
subject is concerned with the "Psychology of Music". It involves only
about 40 minutes of your time and will prove, we are sure, to be of
the utmost interest to you personally. The experiment will be held in
the Language Laboratory in the basement of Anderson Hall (Near the
Library).

We have enclosed a postal card for you to return promptly on which we
have indicated y our appointment. Also enclosed is a card which you
may keep in your wallet as a reminder of time, place, and date. It is
important that you keep your appointment, but if it is not possible to
do so, contact the project director, Mr. Robert Resnick, for another
appointment. He nay be contacted in Building E (near Walker Hall),
Room 125, or called at Extention 2425.

Again, may we remind you that the prompt return of your postal card
within 24 hours is requested. We appreciate your cooperation and will
make appropriate note of it.

Yours truly,


Dr. Audrey S. Schumacher
Professor of Psychology


































APPENDIX B

A AND R SCORES AND LINGUISTIC RAW SCORES FOR
EACH SUBJECT IN THE H, M, AND L GROUPS




















II I I .. I i, l. H I


1,2








2I 7 7
,{ 1


N 1

.. i ]; l *,
, 2 2
0'.+t ? 2

ain` lo.'*




,i? .

'ir, +} n

,' ,.: } +


S. '. 1 7 2
*>2 ++i ;'. b I

4.2 30 .2 1. 7 b'




S1. 24 1/ 1 10

4,4 2.



49 2, .4,' .>.2 2

,6 -42 15.1 9
S 3 552 15.3 .
49 30 578 1,1.7 14.


2b i3 26.0 124

46 26 94 10.3 10


S 17 1








2 ri i)

,7 1


2 20



0 2 0)

2 0 i


I :, 2' 22






















1 22. .2





















2 1 1.7 1























.1 0 1 .1 31., i. l
.i 1, . ..'












































2 2 1.4 2 1 1 .22
: 1 1 7.
2. 2l..'5 ii.






2 6i 1 vi




















1 1 .8 20.5 1 .




S1 22., 2,.i .1

































APPENDIX C

HIGH, MID AND LOW GROUP, AND COMBINED GROUP MEANS
FOR THE THREE SIGNIFICANT LINGUISTIC VARIABLES ACROSS
MUSICAL THEMES









Mean Scores for Three Linguistic Variables
Across Musical Themes for the High A and R Group

Musical Themes


% Pers. Pronouns
% Adjectives
7 Verbs


1 2 3 4 5 6 7
4.55 8.85 8.85 7.70 7.10 6.95 7.95
25.65 24.CO 20.20 23.80 22.25 2".35 21.60
17.85 20.30 21.75 20.00 19.20 20.40 22.70


Mean Scores for Three Linguistic Variables
Across Musical Themes for the Mid A and R Groups

Musical Themes


% Pers. Pronouns
% Adjectives
% Verbs


1 2 3 4 5 6 7
5.16 8.00 6.75 6.58 7.16 8.33 8.08
23.41 23.41 24.50 25.00 22.82 23.50 20.66
17.66 20.64 20.83 21.83 23.25 20.66 19.50


Mean Scores for Three Linguistic Variables
Across Musical Themes for the Low A and R Group

Musical Themes


% Pers. Pronouns
% Adjectives
Z Verbs


1 2 3 4 5 6 7
7.66 9.50 9.11 7.77 7.55 10.22 10.05
25.66 22.50 22.33 20.27 20.77 21.22 22.00
18.83 22.22 21.16 22.05 21.16 21.22 19.05


Mean Scores for the Three Linguistic Variables
Across Musical Themes for the High, Mid
and Low A and R Grouos Combined

Musical Themes


% Pers. Pronouns
% Adjectives
% Verbs


1 2 3 4 5 6 7
5.80 8.88 8.54 7.46 7.22 8.64 8.74
25.48 23.32 22.00 22.82 21.86 23.02 21.52
18.17 21.06 21.31 21.18 20.88 20.76 20.62










BIBLIOGRAPHY


Balkan, Eva and Masserman, J. H. The language of phantasy: III
The language phantasies of patients with conversion hysteria,
anxiety states, and obsessive compulsive neurosis.
J. Psychol., 1940, 10.

Berg, I. A. and Adams, H. E. The experimental basis of personality
assessment. In A. Bachrach (Ed.), Experimental foundations
of clinical psychology. New York: Basic Books, 1962.

Boder, D. F. The adjective-verb quotient. Psychol. Rec., 1940,
3, 310-343.

Brodsky, S. L. Language patterns of repressors and sensitizers in
personal and impersonal descriptions. Unpublished doctoral
dissertation, U. of Florida, 1964.

Brown, R. W. Linguistic determinism and the part of speech. J. abnorm.
soc. Psychol., 1954, 55.

Words and things. Glencoe, Ill.: The Free Press, 1958.

Brown, R. W. and Lenneberg, E. H. A study in language and cognition.
In S. Saporta (Ed.), Psycholinguistics. New York: Holt,
Rhinehart and Winston, 1961.

Busemann, A. Uber typisch und phasiche unterscheide der categoricalen
aprachform. Z. padag. Psychol., 27, as cited in Berg, I. A.
and Adams, H. E. The experimental basis of personality
assessment. In A. Bachrach (Ed.), Experimental foundations
of clinical psychology. New York: Basic Books, 1962.

Carroll, J. B. and Casagrande, J. B. The function of language classi-
fications in behavior. In Maccoby, Newcomb, and Hartley
(Eds.), Readings in social psychology. (3rd ed.) New York:
Henry Holt, 1958.

Cattell, R. B. and McMichael, R. E. Clinical diagnosis by the IPAT
music preference test. J. consul. Psychol., 1960, 24.

Chomsky, N. Review of "verbal behavior" by B. F. Skinner. Language,
1959, 35.

Chotlos, J. W. A statistical and comparative analysis of individual
written language samples. Psychol. Monogr., 1944, 56, No. 2
(whole No. 255).








57


Dahlstrom, W. G. and Welsh, G. S. An M1%PI handbook, a guide to use in
clinical practice and research. Minneapolis: U. of Minn.
Press, 1960.

Eisenson, J. The psychology of speech. New York: F. S. Crofts and
Co., 1938.

Fairbanks, Helen. The quantitative differentiation of samples of
spoken language. Psychol. Monogr., 1944, 56, No. 2 (whole
No. 255).

Fenichel, O. The psychoanalytic theory of neurosis. New York: Norton,
1945.

Greenberg, J. H. Concerning inferences from linguistic to non-linguistic
data. In S. Saporta (Ed.), Psycholinguistics. New York: Holt,
Rhinehart and Winston, 1961.

Johnson, W. Studies in language behavior. Psychol. Monogr., 1944, 56,
No. 2 (whole No. 255).

Korzybski, A. The role of language in the perceptual process. In
R. R. Blake and G. V. Ramsey (Eds.), Perception: An approach
to personality. New York: Ronald Press, 1951.

Laffal, J. Pathological and normal language. New York: Atherton
Press, 1965.

Lindquist, E. F. Design and analysis of experiments in psychology and
education. Boston: Houghton Mifflin, 1953.

Lorenz, Maria. Language behavior in psychoneurotic patients. Arch.
Neurol. Psychiat., 1953, 69.

Lorenz, Maria and Cobb, S. Language behavior in manic patients. Arch.
Neurol. Psychiat., 1953, 69.

Language behavior in psychotic and neurotic subjects.
Arch. Neurol. Psychiat., 1954, 72.

Mann, Mary B. The quantitative differentiation of samples of written
language. Psychol. Monogr., 1944, 56, No. 2 (whole No. 255).

Nunnally, J. Individual differences in word usage. In S. Rosenberg
(Ed.), Directions in psycholinguistics. New York: The
Macmillan Co., 1965.

Osgood, C. E. and Seboek, T. A. (Eds.) Psycholinguistics: A survey of
theory and research problems. International Journal of American
Linguists, 1954, 20, No. 4. Also supplement to J. abnorm. soc.
Psychol., 1954, 49.











Osgood, C. E., Suci, G., and Tennebaum, P. The logic of semantic
differentiation. In:Measurement of Meaning. Urbana: U. of
Ill. Press, 1957, as reprinted in S. Saporta (Ed.),
Psycholinguistics, 1961.

Osgood, C. E. and Walker, Evelyn. Motivation and language behavior: A
content analysis of suicide notes. J. abnorm. soc. Psychol.,
1959, 59.

Rubenstein, H. and Aborn, M. Psycholinguistics. In Annual review
of psychology, 11. Palo Alto, Calif.: Annual Reviews, Inc.,
1960.

Sapir, E. Language. New York: Harcourt, Brace and Co., 1921.

The status of linguistics as a science. Language, 1929, 5.

Skinner, B. F. The verbal summator and a method for the study of latent
speech. J. of Psychol., 1936, 2, 71-107.

SA functional analysis of verbal behavior. In S. Saporta
(Ed.), Psycholinguistics. New York: Holt, Rhinehart and
Winston, 1961.

Stone, D. R. The Auditory Apperception Test. Western Psychological
Services, 1953.

Webster's New Collegiate Dictionary. Springfield, Mass.: G. & C.
Merriam Co., Publishers, 1961.

Welsh, G. S. and Dahlstrom, W. B. Basic readings on the MMPI in
psychology and medicine. Minneapolis: U. of Minn. Press, 1956.

Whorf, B. L. Language, thought, and reality. Boston: Institute of
Technology, 1956.

___ Science and linguistics. In S. Saporta (Ed.), Psycholinguistics.
New York: Holt, Rhinehart and Winston, 1961.

Zipf, G. K. Psycho-biology of language. Boston: Houghton Mifflin Co.,
1935.












BIOGRAPHICAL SKETCH


Robert William Resnick was born on July 10, 1937, at New York

City, New York. He attended Far Rockaway High School and in 1955 was

graduated from Croton-Harnon High School in Croton-on-Hudson, New York.

From 1955 to 1961 he attended the City College of New York and in 1961

received a Bachelor of Arts degree in Psychology. From 1961 to 1963

he attended Columbia University, where in 1962 he received a Master of

Arts degree in Psychology. From 1963 to the present he has been a

graduate student at the University of Florida. He was employed as a

graduate research assistant in the University Counseling Center from

January, 1964, to June, 1964. During the Summer of 1964 he was employed

as an Instructor in Project CAUSE. During the academic year 1964-1965

he held a USPHS Fellowship. In 1965-1966 he interned in clinical psy-

chology at the U.C.L.A. Neuropsychiatric Institite. From September of

1966 to the present he has been an Assistant Counseling Psychologist in the

University Counseling Center at the University of Florida.

Robert William Resnick is married to the former Elizabeth Alba

Ann Estrup. They have a three-year-old son, Christopher, and currently

reside in Gainesville, Florida.









This dissertation was prepared under the direction of the

chairman of the candidate's supervisory committee and has been approved

by all members of that committee. It was submitted to the Dean of the

College of Arts and Sciences and to the Graduate Council, and was

approved as partial fulfillment of the requirements for the degree of

Doctor of Philosophy.


December 1967






Dean, Coll pf-Ar s and Sciences




Dean, Graduate School


Supervisory Committee:



Chairman







^ IQ. 2-^'-
g^/w i ^-/~




























































































































A. _




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