Studies In Expressive Movement

CHAPTER I

THE PROBLEM OF CONSISTENCY IN EXPRESSIVE MOVEMENT

Every thing in man is progressive; every thing congenial; form, stature, complexion, hair, skin, veins, nerves, bones, voice, walk, manner, style, passion, love, hatred. One and the same spirit is manifest in all.

— LAVATER

Literary and Speculative Approaches. The most famous of physiognomists, Johann Kaspar Lavater, considered the basic principle of expression to lie in the fact that no feature of the body contradicts any other feature. It is not, he assures us, the eyes that laugh nor the shoulders that laugh; it is the man that laughs. Lavater was not an experimentalist, and according to modern standards did not offer proof of this fundamental theorem of physiognomy. Even though common sense may dispose us favorably toward Lavater’s position, it still remains in the absence of accurate and controlled study, sheer dogma. And if like many dogmas this one contains both truth and falsehood, nothing excepting analysis and measurement can tell us the proportion of each.

Essayists, more frequently than experimentalists, have reflected upon the consistency of the expressive acts of an individual, and their significance in representing his personality. C. H. Cooley, for example, is impressed by the interdependence of an author’s modes of expression. He writes

Can anyone . . . read Macaulay and think of a soft and delicately inflected voice? I imagine not: these periods must be connected with a sonorous and somewhat mechanical utterance; the sort of person that speaks softly and with delicate inflections would have written otherwise. On the other hand, in reading Robert Louis Stevenson, it is impossible, I should say, not to get the impression of a sensitive and flexible speech. Such impressions are mostly vague and may be incorrect, but for sympathetic readers they exist and constitute a real, though subtle, physiognomy (30 pp. 112 f.).

In spite of Cooley’s warning that such impressions are mostly vague and may be incorrect, the belief that the style of expression is the signature of personality, always consistent with itself, is very prevalent. In an essay on Eugene O’Neill, Barrett Clark describes certain extraordinary physical features and expressive gestures of the playwright in order to make more intelligible the extraordinary quality of his work. In his preface to the letters of Franz Schubert,¹ Ernest Newman writes, Schubert’s letters are a true counterpart of his music: the style is simple—melodic and diatonic, we may almost call it—without any involutions or complexities . . . Newman tells us that the uncouth insufficiency of some of Beethoven’s first sketches for a musical work has its counterpart in the general shapelessness of his literary style, and that the flamboyant exaggerations of the earlier Berlioz are reproduced in his prose. He points also to the obvious parallel between Wagner’s musical compositions and his highly involved prose, with its maddeningly long sentences, its syntactical involutions, its enclosure of one parenthetic clause in another . . .

In our program notes at the symphony, we read concerning Handel that his personality and his music alike were characterized by the same ineffable majesty. Biographers and critics seem to consider it only natural that the diverse expressions of a single personality should show striking similarity.

Psychologists themselves have occasionally been interested in this problem, though seldom has the interest led to direct experimentation. Implicitly, Woodworth’s adverbial definition of personality (202 pp. 552 ff.) contains a statement of the issue. If, as he says, personality is revealed in the way in which an individual talks, remembers, thinks, or loves, the question naturally arises as to the degree of uniformity in the individual’s manner of doing all these things. More direct is Roback’s (142) discussion of the personal idiom, wherein he, like Newman, shows how the same quality often pervades the different creative works of an artist. Roback (141) and Krout (93), like Freud (55), are interested in the symbolic value of gestures. Idiosyncrasies of movement, they hold, proceed from complex inner dispositions and can never be treated as isolated motor phenomena.

In a study of the voices of nine speakers over the radio Pear (130) approaches our problem very closely. Many of his listeners ventured to give physical descriptions of the speakers who were known to them only from their reading over the radio of a passage from Dickens. Although these descriptions were difficult to score, they seemed on the whole to be more right than wrong. In one case, where the speaker was a detective-sergeant he was almost invariably described as being a robust man of heavy build, stout and burly. In another connection Pear observes that, since speech is a form of action, it is natural that some persons’ speech-style should reflect their general behavior. Again, in our estimation no speaker’s appearance was at variance with his or her voice. These last two observations, though unsupported by experimental evidence, are at least interesting statements of our problem. Suggestive illustrations of consistency are found also in the writings of the criminologist, Hans Gross (62), and an eminently practical test of the matter is promised in the modus operandi system of classifying crimes (188). According to this system it is assumed that each criminal employs methods peculiar to himself which if analyzed and recorded will aid in identifying him as the malefactor in a new case.

None of these antecedent studies gives our problem an altogether satisfactory formulation. Most of them are subject to criticism for their confusion of two quite distinct issues. In the first place, there is the problem of the meaning or diagnostic significance of expression (i.e., whether an idiosyncrasy in manner is a valid indicator of some personal complex, prejudice, or interest). Secondly, there is the problem of the inter-agreement of an individual’s various expressive habits (e.g., whether the heaviness of Macaulay’s prose actually did have some recognizable counterpart in his vocal expression). It is clear that the first problem cannot be satisfactorily attacked until some solution of the second problem is available. Unless we can discover what statisticians call the reliability of our indicators, we should not venture by the aid of these indicators to diagnose inner qualities of personality. For example, before attempting to verify Friedrich Gerstäcker’s character-readings from a man’s manner of wearing a hat,¹ it is necessary to know whether the position of the hat is a characteristic one for the wearer. It is necessary likewise to know whether the man who wears his hat, say, on the back of his head exhibits other movements supposed to be indicative of laxness and extravagance.

Experimental Approaches. Many American investigators have crusaded against the popular type of physiognomy which claims to establish parallels between single physical features and traits of personality. However serviceable to the public as a warning their iconoclasm may be, it has led to little constructive research, and has contributed nothing to a psychological understanding of the tangled issues involved. On the contrary, it has actually obscured legitimate problems within the field of expression. It is not uncommon, for example, for American psychologists to classify graphology with astrology, and to condemn both with one breath. In so doing these psychologists simply confuse a bona fide problem of motor psychology with occultism. Again, the passion for correlating isolated physical characteristics with some ill-conceived trait blinds many investigators to more promising avenues of research.

One of the most widely cited investigations is that of Cleeton and Knight (28). Although useful as a caution against popular physiognomists, it was not designed to advance the serious study of expression. For example, the fact that seventy judges made reliable judgments based on a few minutes’ observation of the behavior and appearance of the subjects is too lightly passed over. The writers were interested in showing that these judges could not have based their decisions on the legendary signs employed by certain popular physiognomists; but the challenge of the fact that some other aspects of expression are interpreted consistently by judges is not recognized.¹

Contrasting with the majority of American studies, research on the Continent takes for granted that no case can be made for the older physiognomy, and proceeds hopefully to new and rather daring formulations. From the Institute of Child Research at Leningrad, for example, Oseretzky (127) has issued a comprehensive plan for the study of Psychomotorik. In his scheme this large field is subdivided into three sections. Motoskopie aims to analyze and classify all the significant types of expressive movement (posture, pose, facial expression, gesticulation, handshake, gait, speech, handwriting, automatic movement, pathological movement). Motometrie is concerned with the measurements of movements, and Motographie with their recording. Actually the chief aim of this Russian work is the development of an age scale for the determination of the motor maturity of children. Unfortunately for our purpose, the inter-agreement of the test items in this scale is not given, so that no light is shed upon our problem of motor consistency. We are, however, indebted to Oseretzky for the broad perspective he gives to the whole subject of expression; several suggestions from his Motoskopie are included in our survey of the forms of expressive movement on pp. 24–35.

The work of Enke (48) on psychomotor types is directly concerned with the problem of consistency. His purpose is to determine whether there are features of movement which differentiate Kretschmer’s pyknic and non-pyknic types. His findings are positive: (1) whether the times are measured by spontaneous tapping or by the congenial pace of work, pyknics are essentially slower in their movements than leptosomes and athletics; (2) in tests for freedom of movement pyknics are more irregular and variable, the athletics and leptosomes are more mechanical, automatized, and stereotyped; (3) leptosomes and athletics are given to perseveration, i.e., they find it difficult to change their personal rhythm, and are slowed by distraction; (4) pyknics are adaptable to external rhythms and their action is speeded by distraction; (5) finely coördinated activity is best carried out by leptosomes, less well by pyknics, and least well by athletics; (6) pyknics are fluid, free, soft, rounded, uninhibited in their actions, leptosomes are stiff and angular; (7) pyknics fatigue gradually, leptosomes suddenly; (8) pyknics have a lower average handwriting pressure but greater variability than do non-pyknics. Taking the results together, Enke concludes that the movements of pyknics tend to be slow, free, adaptable, uninhibited, easygoing but variable. The schizothymic temperament which is associated with non-pyknic physiques seems, on the contrary, to express itself in movement that is hesitant, cautious, critical, tense, stereotyped. For our purposes, of course, the important point is that these results were arrived at on the basis of the inter-agreement of several tests. According to Enke it does not matter, for example, whether the typical leptosome is writing with a pen, carrying a glass of water, or reacting to music, he is found to be uniformly tense and cautious.

The present investigation, like Enke’s, administers a large number of motor tests and examines their agreement with one another. In other respects, however, the methods of these two studies differ. Enke employed 500 subjects; the present study, 25. But Enke’s advantage in numbers is somewhat offset by his failure to determine the reliability of his measures. From his results it is possible to conclude only that, by and large, pyknics and leptosomes differ in various aspects of expressive movement; the amount of the differences is stated merely in percentages (without reference to probable errors). Enke entirely omits the problem of personal (intra-individual) consistency. From his data it is impossible to tell whether a given leptosome possesses all, some, or even none of the expressive habits discovered to be characteristic of his type. The present study, on the other hand, makes no use whatever of typological assumptions. It is interested only in the question whether each of our 25 single subjects is consistent with himself.

Another German publication possessing sweep and originality is Gang und Charakter (18), a collection of essays submitted in a prize competition. Only one of the contributions (Wolff’s) is genuinely experimental, and must be considered separately. The rest are for the most part programmatic and critical. In Wilsmann’s essay, for example, we find gait analyzed into one subjective or qualitative characteristic (rhythm) and seven objective or quantitative characteristics (regularity, speed, weight, length, elasticity, definiteness of direction, changeableness). Each of these objective features, he claims, needs to be interpreted in the light of the unanalyzable factor of rhythm, a procedure very similar to Klages’ in the field of handwriting.¹ Most of the contributors to Gang und Charakter seem to hold with Wilsmann that there are unanalyzable features in gait, and that experiment is foredoomed to failure.

Wolff’s study (18), on the other hand, is close in spirit and in conception to our own. His statement of the methodological dilemma facing investigators is worth quoting:

This is the standpoint of characterology to-day: Either we have followers of the statistical-mechanical procedure which yields a crude succession of single elements which never give a living picture; or else we have the intuitive penetration which, it is true, can delineate brilliantly the living processes of human beings, but cannot be communicated, verified, or utilized for a knowledge of genetic relationships (18 p. 109).

To resolve this conflict between a scientific but inadequate characterology and an intuitive but incommunicable characterology, Wolff proposes the experimental-dynamic methodology which involves

(1) a surrender of the total personality, but a retention of interest in complex and structured forms of expression (to be studied either by themselves or in relation to one another);

(2) a standardization of outer situations, so that the resulting differences of behavior may be referred only to variables in personality;

(3) an exclusion of subjective factors such as those resulting when the experimenters have unequal acquaintance with the subjects;

(4) the use of several judges, when judges are necessary, to control reliability;

(5) the study of movement that is spontaneous and unaffected by self-consciousness of the subject.

In an early experiment (200) Wolff obtained five sets of data from each of his subjects: (a) the subject’s vocal expression recorded without his knowledge, (b) a profile picture of the subject’s face, (c) a specimen of his handwriting, (d) a picture of his hands, (e) his manner of retelling a folk-story, stenographically recorded. Wolff endeavored to determine whether judges could match correctly the various records taken from any single subject. His principal results may be summarized as follows: (a) profiles and hands were not matched more frequently than chance would allow with handwriting, voice, or Nacherzählung; (b) voice records and handwriting were matched correctly about one and a half times as frequently as would be expected by chance; (c) the same frequency obtained in the matching of specimens of handwriting with styles of retelling stories; (d) when a judge was not required to match records but to characterize them, his descriptions of the different records from a given subject tended to be identical, (for example, if the retelling of a story was described as colorless, the handwriting of the subject was also likely to be described as colorless by the same judge). Wolff is disinclined to give a more strictly statistical statement of his results, for statistics, he holds, cannot express adequately all the consistency discovered by his method.

In his experiments on gait he employed the moving picture camera. His subjects were mostly students, of both sexes, dressed in a standard, loose fitting garment. He assigned them the simple task of picking up rings and tossing them over a stake. A week after the films were made they were shown to eight judges who knew the subjects and to five who did not. The heads in the films were blocked out, and the judges who knew the subjects were first asked to identify them if possible. Of these indentifications 30% were correct and 70% false. The subjects, who also acted as judges, were able in 100% of the cases to make correct self-identifications. This last result contrasts with Wolff’s previous discovery that, from records of voice, hands, profile, and retelling, identification of others was easier than identification of self. From this curious result it appears that one’s expression is recognizable to oneself only when it involves the large muscles of the trunk and limbs.

The judges were also asked to describe each personality from gait. No validation for these descriptions is given, but Wolff notes that many of them correspond exactly to characterizations given for the same subjects in his earlier experiments. That is to say, the impression made by the gait of a given subject was likely to be reported in the same terms as the impressions made by his style of narration, handwriting, or voice. This fact is perhaps more remarkable since the experiments were separated by an interval of two years. No quantitative statement of this identity of descriptive terms was attempted. Wolff, however, has made a definite experimental attack upon the problem of intra-individual consistency in expressive movement. With the aid not of impression but of correlation we hope in the present study to improve upon Wolff’s method, if not upon his clear-cut perception of the problem.

A study employing a matching method very similar to Wolff’s is that of Arnheim (7). Starting with a simple discovery by Wertheimer that the handwritings of Michelangelo, da Vinci, and Raphael could be rather accurately identified by observers who had no acquaintance with the artists’ chirography, Arnheim proceeded to invent numerous problems of a similar order. Employing over 75 sets of material, and 65 judges, he assembled voluminous data on the consistency of expression. Quotations were matched with photographs, handwritings with sketches of personalities, silhouettes with single descriptive terms, and so on. The matchings were made in all variety of combinations (e.g., three photographs with one, two, three, four, or six sketches). It is difficult to present Arnheim’s very miscellaneous results in any standard fashion. In practically every case, however, the number of correct matchings exceeds chance, often by a very large margin. In only two experiments was the number right less than probability, and in only one case was the number wrong greater than chance. Dividing his half right answers equally between the rights and wrongs, we may summarize rather roughly some of his findings as follows: In all experiments in which there were two choices, the average number of correct matchings was 70.9% instead of the 50% expected by chance; where three alternatives existed the results were more favorable (83.6% instead of 33.3%); where six alternatives existed for a given standard, correct matchings were 51% (instead of 16.7%). Arnheim claims that for all the experiments together the average correct matching is about two and a half times that expected by chance.

It should be noted that although both Arnheim’s and Wolff’s experiments are explicitly aimed at intra-individual consistency, their matching methods depend largely upon the factor of subjective impression. Having discovered consistency in impression they conclude by inference that there is a corresponding consistency in expression. Excepting for the graphological studies (Chapters X and XI), the present investigation employs almost entirely direct objective measurement of consistency.

It seems that the only American investigation which closely resembles our own is that of Downey (39). In 1919 she published an account of a study designed to test the proposition that graphic individuality is but a specific example of a pattern that is impressed upon all the expressive movements of a given person (39 p. 97). She herself classified the handwriting of 12 subjects according to her judgment of the scripts as rapid or slow, light or heavy, loose or compact, expansive or restrained, adroit or maladroit, fluent or jerky, angular or supple, conventional or individual, impulsive or deliberate, concentric or eccentric. Eleven judges were asked to classify the same 12 subjects under the same rubrics, basing their judgments not upon handwriting but upon their impression of the subjects’ gait, carriage, and gestures. The method of scoring was simple. If the majority of judges chose the same rubric for a given subject as Downey did, an agreement was recorded. Chance would allow 50% agreement; the actual correspondence is 60.5%. Her final conclusion is cautious: It may be stated that the outcome of the experiment is slightly in favor of an agreement between graphic and expressive movement. . . . Certainly no sweeping assertions of general similarity can be ventured, although for a few traits there is strong evidence of such harmony (39 p. 104). If the three pairs of rubrics which gave the greatest subjective difficulty, viz., light—heavy, individual—conventional, adroit—maladroit, are eliminated, her results indicate more than 70% agreement.

Downey’s method is more controlled than Wolff’s and Arnheim’s, since her conditions excluded the free intuitive perception of similarities in different fields of expression. In her work there was no direct matching of two sets of records by a single judge, but rather two entirely independent sets of judgments which were compared objectively. Likewise she allowed no freedom for spontaneous characterization of the record; all her terms were prescribed in advance, and insofar as they were ambiguous or not understood they affected her results adversely. Finally, she did not use complete records of expression but only ten traits of movement, so that we cannot say from her results whether or not handwriting as a whole shows similarities with carriage and gesture considered as a whole. It is quite possible that all these conditions, though making for control, resulted in a method that was after all unsuited to the problem she was attacking.

In three respects Downey’s experiment anticipates our own. (1) She formulates precisely the same problem, though her method gives relatively more prominence to graphic movement. (2) She has evolved an objective method for what at first sight seems to be a hopelessly subjective problem. Her technique resembles somewhat that which is employed in Chapter XI of this monograph. (3) Her examination of individual records shows that even when her judgments and those of her collaborators seem hopelessly at variance in a quantitative sense, there are frequently meaningful connections between them which elude statistical statement. This last point anticipates our own discovery of congruence which is discussed in detail in Chapter VI.

The Genetic Approach. Popular lore has helped us to identify our problem, and a few antecedent experiments have provided us with suggestions of method. But the issues with which we are dealing appear more