Associative Tendencies in Psychoneurotics
- Author:
Tendler, Ph.D.
Psychological Clinic, Columbia University, New YorTc Post Graduate Medical School
The present study is an attempt to apply the free association method to a group of psychoneurotic adults in order to determine the association tendencies characteristic of psychoneurotic individuals. More specifically, the problem is to determine the various features shown, relations between these features, the differentiation of the psychoneurotic association pattern from that of the juvenile, the normal adult, the psychotic, and a consideration of the bearing of these results upon psychoneurotic behavior.
Subjects for the experiment were 50 adult psychoneurotics, 12 male and 38 female, ranging in age from 20 to 35 years. Subgroups of the general classification were disregarded, the purpose being a study of the association tendencies of the general group showing psychoneurotic behavior. This is necessitated by several considerations: the typical psychoneurotic is a mixture of a whole group of tendencies, pure types do not exist; symptoms tend to overlap, mental features certainly do overlap; examiners vary in the emphasis they give to one feature or another, some basing diagnoses upon behavior manifestations, others upon theoretical considerations. In view of these factors it was deemed best to study the group as a whole and to determine differentiating features in associative tendencies apart from the more subjective sub-classificatory diagnoses. An additional dimension of value for objective diagnosis may thus be gained. The test used was the KentRosanoff list of words, inasmuch as it offered the possibility of comparable data for normal adults, children, and psychotics. The significance of association tendencies for the general question of adjustment is stated by F. L. Wells as follows:2 “that a person who, under great possibilities of variation, reacts in the 1 Paper read at the annual meeting of the Association of Consulting Psychologists, Philadelphia, May 6, 1933. 2 Wells, F. L. Mental testa m clinical practice, New York, World Book Co., 1927, pp. 223-4. 108 association experiment as others do, will find himself in better tune with social situations generally than the person whose reactions tend to be peculiar to himself.” Kent and Rosanoff3 stated their general conclusion from studies on the associations of normals and psychotics as follows: ‘’ A comparison of our pathological with our normal material en masse reveals in the former evidence of a weakening of the normal tendency to respond by common reactions. To what extent this weakening occurs in psychoneurotics and what basic mental factor or factors underlie this weakening is the general problem of the present study. The specific problems studied and the results obtained follow. I. The Kent-Rosanoff study treats results objectively in terms of common responses (responses given by the normal group), their frequency of occurrence, individual responses (responses not given by the normal group), and doubtful responses (grammatical variants of common responses). Salient features are the average frequency of the common responses and the number of individual responses a person gives. The relationship between the two is of psychological interest. Are we dealing with two distinct tendencies, or are the two a single basic trend towards low or high community agreement? On the basis of findings in the psychoneurotic group here studied the latter appears to be the case. Average frequency scores for the psychoneurotic group run from 20 to 160.5, with a mean at 103.4 and an S.D. of 33. The number of individual responses for the group run from 1 to 37, with a mean of 10.3 and an S.D. of 7.3. The product moment r between the two is .74, P.E. 0.04. It would appear that the two measures are really a single^ basic trend of community agreement, high average frequency going with low number of individual responses, low average frequency scores going with high number of individual responses. Low average frequency scores have relatively the same significance for adjustment that individual responses have.
II. How do psychoneurotics, as a group, compare with normal adults and psychotics in common and individual responses? Table I gives the data upon which to draw for conclusions. Psychoneurotics tend to show a slight reduction in common associations and an increase in individual responses. Using as a pathological index the average of the three psychotic groups in common and individ3 Kent, G. H. and Kosanoff, A. J. A study of association in insanity, Amer. J. Insanity, 1910, 67, 317-390.
ual responses, 75 and 22.7 respectively, there is a drop of 16.7 in common responses from the normal adult, and an increase of 15.9 in individual responses. The psychoneurotic group gives a drop of 3.7 in common responses or 22 per cent of the psychotic drop, and an increase of 3.5 individual responses or 22 per cent of the psychotic increase in individual response. Table I Investigator Subjects Av. Common Av. Doubtful Av. Individual Kent-Rosanoff. Strong 4. Tendler. 1000 adults 247 psychotics 32 man. depress. 16 ” 50 psychoneurotic Average of 3 psychotic groups. 91.7 70.7 75.8 78.6 75. 1.5 2.5 3. 1. 1.7 6.8 26.8 21.5 19.7 10.3 22.7
4 Strong, E. K. A comparison between experimental data and clinical results in manic-depressive insanity, Amer. J. Psychol., 1913, 24, 66-98. III. An approach to the problem of what underlies the “weakening of the normal tendency to respond by common reactions” may be had through a study of the correlations between average frequency score and types of association, and between number of individual responses and types of association. The procedure was to classify the stimulus-response words as a unit into various association classes: superordination (table-furniture), subordination (fruit-apple), coordination (table-chair), contrast (dark-light), similarity (short-little), adjective-noun (deep-hole), noun-adjective (man-big), contiguity (table-dishes), noun-verb (manwalk). These classes cover the greater number of associations. Other classes were found but they give too few instances to be, valuable for correlation study. The number of each class of associations was found for each case, average frequency scores, and number of individual responses. Product moment correlations were determined for average frequency scores and score in each class of associations, and for individual response scores and score in each class of associations. Results are given in Table II. Several points are to be noted from these correlations: 1. Where average frequency gives plus correlations, individual response gives minus correlations, except for the noun-verb class of association which gives a negligible r. 2. Average frequency correlates positively with the following:
Table II Correlation bet. Av. Freq. and Contrast Assn. +.67 Coordinate Assn. +.55 Similarities Assn. +.50 Superordinate Assn. +.25 Noun-adjective ?.60 Adjective-noun ?.66 Sum adj.-noun and noun-adj. ?.79 Contiguity ? .42 Subordinates ?.21 Noun-verb ?.07 P.E. .055 ,072 .071 .089 .06 ,055 .034 .08 .09 ,09 Correlation bet. Nu. Ind. Response and Contrast Assn. ?.30 Coordinate Assn. ?.26 Similarity Assn. ?.53 Superordinates ?.40 Noun-adjective +.60 Adjective-noun +.39 Sum adj.-nounandnoun-adj. +.56 Contiguity +.14 Subordinates +.34 Noun-verb ?.03 P.E. .087 .089 .071 .08 .06 ,08 076 09 08 09
contrast, coordinates, similarities, superordinates. Individual response correlates negatively with the same classes of association. 3. Average frequency correlates negatively with the following. noun-adjective, adjective-noun, sum of noun-adjective and adjective-noun, contiguity, and subordinates. Individual response correlates positively with these. 4. Average frequency tends to give higher correlation coefficients than individual response, thus showing itself to be the more discriminative measure. Excluding the negligible r for the nounverh class, the average of the frequency correlations is .516, A.D. 0.142; the average of the individual response correlations is .391, A.D. 0.116.
The significance of positive and negative correlations is definitely indicated from another source. Woodrow and Lowell5 using 90 of the Kent-Rosanoff list of wor&s with 1000 children to study the difference between the juvenile and adult pattern of association found the following: 1. Types of association occurring more frequently in adults are: contrast, superordinates, coordinates. 2. Types of association occurring more frequently in children are: noun-adjective, adjective-noun, contiguity. 3. Indications were less definite for similarities and subordinates. We note that the positive correlations of average frequency found in the psychoneurotic group signify adult pattern of association, the negative correlations signify juvenile pattern of association. An opposite relationship holds for the individual correlations. In other words, low average frequency scores and high s Woodrow, H. and Lowell, F. Children’s association frequency tables, Psychol. Monog., 1916, 22, No. 5.
number of individual responses tend to give association types characteristic of the juvenile association pattern; high average frequency scores and low number of individual responses tend to give association types characteristic of the adult association pattern. The precise degree of correlation is indicated by the correlation coefficient of +.80, P.E. 0.034 between average frequency and sum of adult association types, a correlation coefficient of ?.77, P.E. 0.042 between average frequency and sum of juvenile association types. Between number of individual responses and adult association types the correlation coefficient is ?.53, P.E. 0.07; with the sum of juvenile association types a coefficient of +.53, P.E. 0.07. What was spoken of as group agreement turns out to be adult association pattern in contrast with juvenile association pattern. Average frequency suggests itself as an index of adult association pattern, individual response as an index of juvenile association pattern. IV. A study of the differential significance of certain selected features was undertaken. Using 8 individual responses as a critical score divides the group into two equal halves: type A, 25 cases with individual responses running from 1 to 8, type B, 25 cases with individual responses running from 9 to 37. The two were compared with respect to percentages of each falling at or above the mean of the entire group in number of individual responses, average frequency, sum of adult types of association, sum of Table III Feature Studied Mean Type A Per cent at or above Mean Type B Per cent at or above Mean Per cent Diff. S.D. D. S.D. Ind. responses… Aver, frequency. Sum adult ass .. . Sum juvenile ass. Contrasts Noun-ad j Adj.-noun Similarities Subordinates 10.3 103.4 41.7 36.0 12.2 7.6 8.3 9.5 3.36 0 80 72 28 60 16 16 72 0 72 16 32 60 40 64 52 32 32 72 64 40 32 20 48 36 40 32 0.09 .108 .124 .13 .13 .12 .12 .13 .09 8 6 3.2 2.46 1.54 4. 3.0 3. 3.4
juvenile types of association, contrasts, noun-adjective, adjectivenoun, similarities, subordinates. Results are given in Table III. All of the features presented show a differentiation between types A and B. Contrasts does not satisfy the statistical criteria of significance, sum of juvenile association classes is fairly satisfactory, all the rest are statistically satisfactory. On the basis of association tendencies, it would appear that there are two fairly distinct types or varieties of psychoneurotics one, variety A characterized by an adult association pattern, the other, variety B, characterized by a juvenile association pattern. This is of considerable significance, suggesting for the psychoneurotic^ behavior of variety B a constitutional basis and foi variety A a situational basis.
V. Is this finding of juvenile and adult association type pattern characteristic of psychoneurotics only, or does it hold for other groups? An analysis of the findings of investigators of other groups may throw light upon the problem. Percentages of the various association categories in different groups are presented in Table IV. Following the conclusions derived by Woodrow and Table IV Association type Adult type Contrast Superordinate. Coordinate. .. Total…. Intermediate type Subordinate. . Similarity… . Total Juvenile type Contiguity. Adj.-noun. Noun-adj.. Verbs Total. Grand total. Groups 26 14 8 48 1 23 24 7 3 1 5 16 10.6 7.6 10.9 29.1 1.6 8.9 10.5 6.0 6.9 4.3 6.4 23.6 63.2 1.3 3.7 6.0 11.0 2.1 8.6 10.7 15.3 11.2 7.8 10.2 44.5 66.2 8.4 5.9 3.0 17.3 1.1 5.0 6.1 2.7 3.7 2.0 0.6 9.0 32.4 8.6 4.9 3.6 17.1 1.3 3.9 5.2 3.7 5.0 2.2 1.1 12.0 34.3 16.5 8.1 18.4 43.0 1.6 11.6 13.2 6.8 5.6 3.8 7.1 23.3 79.5 8.2 3.8 15.3 27.3 3.2 7.6 10.8 7.9 12.2 10.0 8.4 38.5 76.6 12.3 5.9 16.8 35.0 2.4 9.6 12.0 7.3 8.9 6.9 7.7 30.8 77.8
Lowell we shall consider the association types of contrast, superordination, and coordination as belonging to the adult association pattern, the association types of subordinates and similarities as belonging to an intermediate pattern, and as characteristic of the juvenile pattern the association types of contiguity, adjectivenoun, noun-adjective, and verbs. Peicentages of these are given for the following groups:
a. An hypothetical standard of normality determined by classifying the 100 highest frequency responses to the 100 stimulus words, taken from the Kent-Rosanoff normal frequency tables. b. The percentages given by Woodrow and Lowell for the Kent-Rosanoff normal data. c. Woodrow and Lowell findings for the juvenile group. d. Murphy’s findings for Dementia Prascox.6 e. Murphy’s findings for Manic Depressives. /. The psychoneurotic group here studied, type A. g. The psychoneurotic group here studied, type B. h. The total psychoneurotic group here studied. To avoid errors of interpretation inherent in dealing with absolute percentages derived from varying sources a ratio device was adopted whereby the total for adult responses is taken as unity for each investigation and the total of intermediate and total of juvenile responses are treated as fractions or multiples of each unity. Comparisons are justifiable because of the fairly clear cut character of the categories of association. As far as the present investigation is concerned a deliberate effort was made for the classification of associations to be in accord with those of Woodrow and Lowell, who give ample illustration to their classifications. Murphy’s material appears to be comparable in the association classes here reported. The following tendencies may be indicated: 1. For the hypothetical standard of normality associations the ratio is 1: .5: .3 for adult, intermediate, and juvenile types respectively. There are about half as many intermediate, one-third as many juvenile responses as adult.
2. For the normal Kent-Rosanoff group the ratio is 1: .36: .8. There are about one-third as many intermediate, and about fourfifths as many juvenile responses as adult. 3. For the Woodrow-Lowell juvenile group the ratio is 1:1:4. There are about as many intermediate, and four times as many juvenile responses as adult.
4. For Murphy’s psychotics, groups d and e in Table IV, dementia prsecox and manic depressives, the indications are that there is a marked reduction in meaningful responses. Some 54 per cent were unclassifiable according to Murphy. Either marked deterioration or functional disability is at work. In the classifiable responses here reported the ratio of adult, intermediate, and juvenile e Murpliy, G. Types of word association in dementia prsecox, manicdepressives, and normal persons, Avier. J. Psycliiat., 1923, 79, 539-571. responses is 1: .35: .52 for the dementia prascox group. For the manic depressive group the ratio is 1: .3: .7. For the classifiable responses the pattern is seemingly that of the normal adult. As between the two psychotic groups there seems to be an increase in the juvenile responses for the manic depressives. 5. For the psychoneurotic, Type A, the ratios are 1: .3: .5 for adult, intermediate, and juvenile responses. The pattern appears to be that of the normal adult group. Ratios are derived from column (f’ in the table.
6. For the psychoneurotic, Type B, the ratios are 1: .4:1.4 for adult, intermediate, and juvenile responses. The significant feature here is the increase in and overbalancing of the juvenile type of association. Ratios are derived from column ?g’ in the table. 7 For the entire psychoneurotic group the ratios are 1: .34: .86, approximately a normal adult picture, which however masks the difference between Types A and B. Batios are derived from column ‘h’ in the table.
VI. What bearing do these facts have upon associative pathology ‘ For the psychotics, associative pathology lies in the unclassifiable responses and in the inability to respond. The most general statement that may be predicated is that a collapse has taken place in the function of meaningful association. Words lose their character of specific application and relation to other meaningful words. They become syllabic elements which may be attached, detached, and unattached in an associative flow devoid of meaning. Kent and Rosanoff analyzing individual responses m their pathological material direct attention to: the tendency to give neologisms, incoherent responses, stereotyped responses, sound reactions, pronounced perseveration. These illustrate definitely the general tendency here predicated.
For the psychoneurotic, Type B, the associative pathology lies in another direction, that of developmental status, in the predominance of juvenile association type. Of special interest in this connection is the occurrence of the noun-adjective, adjective-noun, and verb responses?30.6 per cent. Jung’s interpretation of these as a predicate type7 signifying emotional and subjective value does not seem to take into consideration its juvenile associative character. Neither does the concept of regression receive support from the present study. The mean average frequency for Type B on the Kent-Rosanoff adult frequency tables is 79.7. On the 7 Cited by Wells, F. L. Mental tests in clinical practice, pp. 199-200. Woodrow-Lowell children’s frequency tables it is 68. If associative regression had taken place, we should expect a higher frequency value in terms of the children’s frequency values than in terms of the adult frequency values. Such is not the case. What we find is a verbal equipment of the adult type with a mental association pattern of the juvenile type.
For Type A, no associative pathology seems to be indicated. Adult association types almost double the number of juvenile association types. The mean average frequency on the Kent-Rosanoff adult frequency tables is 126, while on the Woodrow-Lowell children’s frequency tables it is 89.9. Absence of associative pathology suggests the need for a closer analysis of this group in relation to situational difficulties, to behavior shown, to psychological processes, and responsiveness to treatment. A possible hypothesis is that under the general classification of psychoneurosis there are included two differing groups. One is characterized by a constitutional inadequacy of associative behavior. This group may be thought of as the true psychoneurotic whose symptoms develop and become fixed on the basis of connections of an inferior sort with resulting tensions. The other group showing no associative pathology may be regarded as showing an emotional responsiveness to situational difficulties. We must differentiate between the emotional disturbances of normal individuals and psychoneurotic behavior. The latter is generally stupid behavior,8 the former often involves genuine situational difficulties of considerable complexity. At any rate, the present study offers a suggestive approach for differentiating the psychotic from the psychoneurotic, for differentiation within the psychoneurotic group, and the possibility for tracing the relation of associative tendencies to symptomatology, responsiveness to treatment, and variations in treatment techniques.
8 Tendler, A. D. The mental status of psychoneurotics, Arch. Psychol., 60, Chap. 6, Intelligence and psychoneurosis, 55-71.
Disclaimer
The historical material in this project falls into one of three categories for clearances and permissions:
Material currently under copyright, made available with a Creative Commons license chosen by the publisher.
Material that is in the public domain
Material identified by the Welcome Trust as an Orphan Work, made available with a Creative Commons Attribution-NonCommercial 4.0 International License.
While we are in the process of adding metadata to the articles, please check the article at its original source for specific copyrights.