Interpersonal Judgments Based on Talkativeness: I Fact or Artifact? Donald P. Hayes; Leo Meltzer Sociometry, Vol. 35, No. 4. (Dec., 1972), pp. 538-561. Stable URL: http://links.jstor.org/sici?sici=0038-0431%28197212%2935%3A4%3C538%3AIJBOTI%3E2.0.CO%3B2-T Sociometry is currently published by American Sociological Association.

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/asa.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers, and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take advantage of advances in technology. For more information regarding JSTOR, please contact [email protected].

http://www.jstor.org Mon Jul 16 17:14:07 2007

Interpersonal Judgments Based on

Talkativeness: I Fact or Artifact?'

DONALD P. HAYES

AND

LEO MELTZER

Cornell University

Previous research has demonstrated that participation rates i n discussions relate t o m a n y asjects of interpersonal perception, but such studies have not determined whether the critical variable is sheer amount of talking or other variables (verbal and/or non-verbal) associated w i t h amount of talk. I n this paper, several techniques are developed for separating the purely temporal cues i n conversations from the rich set of cues w i t h which t h e y are ordinarily confounded. A series oJ experiments demonstrates that sheer amount of talk is strongly implicated i n the formation of interpersonal impressions. Consider this paradox: An impressive set of findings indicates that vocal productivity (how much people talk) is profoundly important in social interaction, yet most social psychologists have paid this variable only scant attention. ' ~ e f o r eproceeding to a possible resolution of this paradox, let us document its assertions. Talkative persons differ from quiet persons in significant respects. The quiet man, on taking a seat, is described as seeking out the less prominent position around a table (Dawe, 1934; Hare and Bales, 1963) ; he expects to be less successful in social encounters than more loquacious persons (Efran and Korn, 1 9 6 7 ) ; he has a tendency to have longer than usual latencies before entering conversations, with the presumed effect that he has a hard time getting the "floor" (Pope and Siegman, 1966; Matarazzo and Wiens, 1 9 6 6 ) ; he is relatively high in the positive reactive categories of the Bales l T h i s study was supported by Research Grants GS-2075 and GS-3066 from the National Science Foundation and Cornell's Center for International Studies. Grateful acknowledgment is given to: Sally Sievers for her expert help as Research Associate; Florence Hayes for many special computer programs and processing assistance; Karen Dempsey and Pamela Smith for editorial assistance; and Professor Ulric Neisser for suggesting a yoked control between a conversation and a mechanical analog of the conversation. 53 8

INTERPERSONAL JUDGMENTS AND TALKATIVENESS

539

IPA Code and relatively low in the task categories (Strodtbeck and Mann, 1956) ; he lacks skills in summarizing and developing task-oriented solutions to problems (Peterman, 1949; Knutson, 1960; Riecken, 1958); he is described by his associates as a "milquetoast" type, timid, and uninvolved (Knutson, 1960); he generally does not provoke reactions of strong positive or negative affect (Bales, 1956; Couch, 1960; North, 1957)) but he is occasionally the target of hostility, the scapegoat for groups in conflict (North, 1957). The reported attributes of the talkative person are often, but not necessarily, the inverse of those of the quiet man. Talkative persons sit in propitious locations for holding the floor (Hare and Bales, 1963); report successful social encounters (Efran and Korn, 1967); have short speech latencies which may help them gain control of the conversation (Pope and Siegman, 1966); tend to focus more than their share of contributions in task areas (Strodtbeck and Mann, 1956); contribute significantly in moving the group toward its objectives by stating ideas and by elaborating others' ideas; and manage to have their own ideas adopted more often than others (Riecken, 1958). There is some suggestion in the literature of affective ambivalence toward talkative persons. Active group members are most commonly described as having contributed to group goals and are among the most "liked," but the most talkative person is also frequently the target of the group's strongest negative affect (Bales, 1956). Part of this ambivalence may be traced to the most talkative person's general competence and energetic' manner (Couch, 1960)) on the one hand, and to his aggressiveness and intractability (Cervin, 1957; Carment, 1961)) on the other hand. Bavelas et al. (1965) moved a long step past merely correlational investigation when they demonstrated that sociometric status could be manipulated by experimentally raising and lowering the amounts of talk of participants in groups. Similarly Aiken (1965) and Zdep (1969) showed that interpersonal descriptions and leadership behavior also varied systematically with experimental changes in talk rates. These studies, together with those previously mentioned, suggest that the level of a person's participation during interaction is strongly implicated as a major cue for interpersonal judgments and evaluations, and hence is directly related to group properties such as status and sociometric structure. An even stronger claim can be made for the importance of the quantitative aspects of talk: Talkativeness appears to index directly, indeed. virtually to define, one of the primary factor analytic dimensions of interpersonill behavior or evaluation. Foa (1961) concluded, in his review of factor analytic studies of interpersonal behavior, that the various studies

540

SOCIOMETRY

all suggested "orthogonal axes of dominance-submission and love-hostility." In turn, there is strong evidence that the dominance-submission dimension is well indexed by the amount of vocal participation a person contributes during interaction. Couch (1960) showed that the number of "initiations" an individual produced (Bales code) loaded .96 on this dimension. "Dominance" is defined by Foa and Couch as synonymous with prominence and assertiveness (not with hostile dominance, which they view as a combination of high activity and negative affect). This dominance-submissiveness dimension is the most prominent one in the data derived from the Bales code, as well as from several other batteries of tests used in Couch's elaborate study (1960). The close link of vocal participation and perceived dominance held whether the data came from observers, from participants describing other participants or, to a lesser degree, from participants' descriptions of themselves. Similar findings have been reported by Mann (1961) in groups engaged in task-oriented or expressive-oriented conversations. Two further perspectives bear on the thesis that amount of vocal participation is a major dimension of interpersonal judgments. Osgood et al. (1957) labeled one of the persistent major factor-analytic dimensions of connotative meaning as "activity." I t does not seem too much of an inductive leap to suggest, therefore, that the activity of a person-prominently indexed by talkativeness-will also be strongly implicated in others' reaction to him. Finally, Chapple has for many years (see, for example, 1939 and 1970) maintained that interpersonal compatibility depends upon the "fit" of the interactants' activity patterns. While he is emphatic in including non-verbal, kinesic and postural behaviors in his concept of "activity," talkativeness is clearly a major component of activity in Chapple's sense. To our mind, the foregoing results and theoretical positions could easily explain and justify an intensive interest on the part of social psychologists in the quantitative aspects (including frequency, amount, timing and patterning) of that most social of social behaviors-talk. But the strange fact is that most social psychologists have been curiously uninterested in these non-verbal but vocal features of interaction. In many of the studies we have cited above, talk was an incidental rather than major focus. Vocal productivity per se is rarely mentioned by social psychological theorists, and is infrequently referenced directly in either the major abstracts or basic textbook indexes. Despite its pervasiveness as an event and its relation to so many social psychological concepts, it has not been central to research on interactional behavior. Why has this variable received less attention than it merits? Perhaps quantity-of-talk has been neglected because it is so obvious. Recall the fish

INTERPERSONAL JUDGMENTS AND TALKATIVENESS

541

who couldn't discover water because he was immersed in it and the man who was astounded to discover that he was speaking prose all his life. But we suspect that the real reason is that no one is convinced by the evidence. Those who are familiar with the cited results must feel that surely something more than sheer talk is involved. They must feel that associated with the quantity-of-talk variable are many other variables, e.g., the content of what is said, the quality and manner of its presentation, and all of the associated non-verbal behaviors which accompany speech-e.g., the proxemic, kinesic, and paralinguistic aspects of communication. Vocal output would then be viewed as simply the carrier for the other variables which themselves are the cues for judgments. This paper's position, however, is that vocal output per se is important -it is not merely a carrier for other variables like speech content and it deserves intensive study in its own right. Our hypothesis is that sheer total amount of vocal output constitutes a suficient cue for many interpersonal impressions. T o be sure, judgments of others are influenced by a large set of variables, but in this paper we demonstrate: ( a ) that a very important variable in this set is sheer talkativeness, and ( b ) that the effects of sheer amount of talk cannot be ascribed entirely to variables associated with talking (i.e., the content of what a person is saying as well as kinesic, vocal, gestural, proxemic and gaze behaviors). STUDY 1. TEMPORAL CUES VERSUS OTHER CUES 2

I n this section, a technique is described which isolates two classes of stimuli which ordinarily are thoroughly confounded: ( a ) information representing only the temporal pattern of participation in a conversation, and (b) the full set of vocal, verbal, temporal, kinesic, proxemic and other cues available in the same conversation. Also described is an investigation comparing judges' impressions of discussants observed under the above two types of conditions.

Method A conversation between three men took place before an unmanned television camera. I t was entirely spontaneous; the topic (the celibacy of priests) was unknown to the speakers until moments before the recording began. The discussion lasted twenty minutes. Each man wore a bone oscillator fixed to his forehead in order to enable each voice to be recorded on a We wish to thank Gary Bouma and John Rounds who helped to plan and to conduct Study 1.

542

SOCIOMETRY

separate track of a specially designed tape recorder. This bone oscillator recording system, developed by Hayes and Meltzer (1967)) is more effective than throat microphones in separating the speakers' voices, is comfortable to wear, and produces an intelligible signal. Two tapes were recorded simultaneously. One was a video tape with its own sound track and the other was a three-track audio tape which recorded each man's voice independently from the bone oscillators. These tapes were used to create two experimental conditions: ( a ) Video condition. Judges observed the video tape and heard the discussion. ( b ) Lights condition. Judges observed a representation of the conversation through a panel of lights blinking on and off. There was no sound. The panel contained three red lights which were connected to three voiceactuated relays. The relays were tripped by the signals on the special audiotape of the conversation. Whenever a man's voice rose to an audible level, the relay turned on that man's light. If more than one man was speaking simultaneously, more than one light was on. Even brief pauses amidst utterances were represented in the pattern of blinking lights. I n this condition, the lights constituted the sole stimulus. Only two levels of voice intensity were represented by a light-on and off. No additional information in the form of sound or text was available. Judges were told that the lights represented a conversation of three persons. Judges watched the lights in silence for a seven and one-half minute passage of the conversation. Following the presentation of the stimuli, judges in both conditions filled out the same questionnaire. Two sets of judges were utilized: the original experiment with 53 college students and the second with 38 Boy Scouts aged 12-14. The questionnaire contained ten semantic differential scales (reproduced in Table 1 ) : three evaluative items, four potency items and three activity items.

Results Despite the minimal information represented by the blinking lights, judges who watched produced descriptions surprisingly redundant with those produced by judges who saw and heard the videotape (see Table 1 ) . Other samples a t Berkeley, Laval University in Quebec, and Cornell (faculty and students) have also given similar evaluative judgments after having watched the same pattern of lights3 Results for the Berkeley, Laval, and additional Cornell samples are not shown in order to conserve space. Similarly, to conserve space, several other results described below in the text are not accompanied by detailed tabular presentation. Supplementary tables may be obtained upon request to either author.

544

SOCIOMETRY

These data suggest that this study's hypothesis is verified. There appears to be a marked tendency to give the same ratings to a person seen on videotape and to the lights blinking in synchrony to his vocalizing. Of the sixty pairs of lights versus video score comparisons in Table 1, only seven are statistically significant at .05 or better-not much more than the three or four which would be expected if the data had been drawn from a table of random numbers. These results suggest that the temporal information (which is all the lights indicate) is suficient for our subjects to draw inferences about what sort of person each speaker was during the conversation. However, several criticisms might be raised. First, it is hardly surprising that the temporal aspects of talk (as indexed by the lights) contain sufficient information for judgments of the three activity scales (passive-active, slow-fast, and inconspicuous-prominent). Second, Osgood et al. (1957, pages 73-74) have pointed out that in the domain of person perception their dimensions of activity and potency sometimes merge into a single dimension (which they label "dynamism"). Hence, the similarity of judgments in lights and video of the four potency scales (dull-sharp, weak-strong, softhard, quiet-loud) may not be considered much more surprising than those for the three activity scales. Finally, in the data for the three evaluative scales (bad-good, worthlessvaluable, unpleasant-pleasant) it is especially noteworthy that the lights conditions yielded judgments of approximately the same intensity and direction as was produced from the full range of information available on the videotape. However, the behavior of the participants was neither so endearing nor so hateful as to produce more than moderate levels of positive and negative evaluation in the judges-in both the video and lights conditions. I n short, the data of Study 1 are consistent with the initial hypothesis, but more evidence is required to be convincing. STUDY 2. NEW ADJECTIVES

While, in several diverse replications, substantial redundancy was found in semantic differential judgments between the lights and the video modes of stimulus presentation, the ten scales provide a limited range of interpersonal concepts. I t is possible that the results were somehow specific to the set of adjective scales used. I n particular, one would like to see the experiment replicated giving the respondents a set of adjectives less obviously loaded on activity and containing a wider range of terms connoting evaluation or affect.

INTERPERSONAL JUDGMENTS AND TALKATIVENESS

545

Method Twenty-three adjectives were selected from among those mentioned in Peabody's (1967) critical analysis of Osgood et al's (1957) work with the semantic differential. The 23 adjectives are listed in Table 2. These adjectives include a few terms with strong activity content (like belligerent) but most have little apparent activity content (like idealistic and flexible). Again judges (undergraduates) watched either the videotape or the three lights. The same videotape and lights display used in Study 1 were used again in Study 2 . The judges then indicated whether they found each adjective "characteristic," "neither characteristic nor uncharacteristic," or "uncharacteristicv of the man being described. For each term in the list, a mean was computed weighting "characteristic" as 1, "neither" as 2, and TABLE 2

Mean Ratings of Peabody Adjectives im Response to the Lights and Videotape Versions of the Celibacy Discussion Peabody Adjective

Idealistic Secretive Self-righteous Modest Firm Individualistic Cautious Skeptical Stupid Artificial Passive Vacillating Nosy Meditative Flexible Timid Discriminating Severe Belligerent Moral Committed Sarcastic Short-sighted

Person . 4 Light

1.97 2.29 1.97 2.35 1.48 1.90 2.10 1.87 2.58 2.45 2.16 2.10 2.58 1.97 2.10 2.13 1.97 2.32 1.94 2.06 1.26 2.23 2.35

Video

1.74 2.30 1.51 2.26 1.32 1.79 2.15 1.66 2.51 2.02 2.06 2.15 2.49 2.06 2.00 2.04 2.02 2.13 2.06 1.96 1.51 1.45 1.83

Person B Light

2.39 2.06 2.42 1.32 1.97 2.10 1.16 1.97 2.55 2.32 1.45 1.90 2.39 1.52 2.10 1.26 2.19 2.26 2.16 1.90 2.32 2.29 2.48

Video

2.21 1.94 2.40 1.30 2.15 2.26 1.28 2.32 2.62 2.57 1.15 2.21 2.40 1.53 2.40 1.15 2.34 2.13 2.09 2.40 2.38 2.26 2.66

Person C Light

1.61 2.19 1.48 2.19 1.39 1.84 2.29 1.97 2.39 2.23 2.13 2.23 2.32 2.29 2.23 2.06 2.32 2.06 1.81 2.13 1.45 2.10 2.42

Video

1.79 2.32 2.11 2.36 1.34 1.51 2.11 2.23 1.98 2.32 2.02 2.13 2.53 2.06 1.81 2.09 1.83 2.47 2.17 1.64 1.55 2.23 2.40

Notes: Peabody adjectives were rated as follows: l=characteristic; 2=neither characteristic nor uncharacteristic; 3=uncharacteristic. N=31 raters for lights and 47 raters for video.

SOCIOMETRY

"uncharacteristic" as 3. For each of the three stimulus persons in the videotape, a product moment correlation was computed between his set of scores derived from video and the corresponding scores derived from lights.

Results The results are shown in Table 2. Again there is considerable similarity between the interpersonal judgments based upon ordinary stimulus conditions (video) and the judgments based upon the analog to the purely temporal aspect of talking (lights). Correlations between lights judgments and video judgments, computed from the data of Table 2, are substantial; Mr. A, .72; Mr. B, .92; Mr. C, .55. The redundancy found here for a new set of descriptive terms, presented to a new set of subjects in a different manner, supports the results of Study 1. The implications of these results-should they be shown to have generality beyond the specific stimulus tapes used in both studies 1 and 2are: ( a ) Vocal activity provides a sufficient cue for a considerable range of interpersonal impressions, including (to a lesser degree) evaluative and affectual judgments. ( b ) Those variables which previous research has shown to be related to amount of participation (e.g., sociometric status as in Bavelas et al, 1965) may be represented in the temporal pattern of vocalization represented by the lights, including total amount of talk, since under the blinking lights condition, no content, kinesic or other nonverbal information is present to influence affective judgments. STUDY 3. NEW TAPES, NEW TOPICS

Study 3 investigated whether the prior results (indicating the importance of vocalization per se in influencing first impressions) could be confirmed, utilizing new stimulus conversations. An additional merit of the new stimulus tapes is that they dealt with an issue in which judges could be expected to be highly involved. Study 3 also utilized a new set of person perception adjectives, in order to extend further the generality of the set of attributes previously used.

Method Three pairs of paid undergraduates (strangers) were recruited to discuss the trial and conviction of Lt. Calley for his role in the My Lai massacre. The date of the discussions was approximately three weeks after the conviction. I t seems safe to say that most undergraduates were, a t the time of this study, interested and involved in this issue. The three pairs were asked

INTERPERSONAL JUDGMENTS AND TALKATIVENESS

547

to discuss this trial for seven minutes while being video- and audio-taped in the special manner described earlier. Three days later, a group of ten paid student judges were shown the following stimulus tapes in the order indicated: (1) videotape of the first pair of discussants (A and B ) ; (2) lights representation of the second pair (discussants C and D ) ; (3) videotape of the third pair (E and F ) ; and (4) lights representation of discussants A and B. Note that only one pair of discussants (A and B) was shown to judges in both lights and video mode. The judges were told that the four presentations were based on four different conversations. Following each presentation, the judges provided the following data for each stimulus person or light observed: ( I ) Semantic differential scales. These were the same as in Study 1. ( 2 ) Gough adjectives. Fifty items were drawn from Gough's (1965) list of adjectives. Half of our sample of terms was drawn a t random from Gough's sublist which contained words reported by one of his samples of students to be in the most "favorable" quartile or most "unfavorable" quartile; the other half of our words were drawn randomly from his more neutral terms. The judges checked each Gough adjective as either "characteristic" or "uncharacteristic" of each man (or light) rated, or checked that they had formed "no impression whatever." ( 3 ) Perceived activity item. After all other questions were asked, the judges were asked to write their estimate of the percentage of the total time that each person was talking.

Analysis and Results Two analyses were conducted, which we shall call Studies 3A and 3B. Study 3 A . For each stimulus person (or light) in the four presentations, the mean of the ten subjects' ratings was computed for each of the ten semantic differential scales. This was also done for the fifty Gough adjectives, weighting characteristic as 1, no impression as 2, and uncharacteristic as 3.4 The means for the two sets of adjectives are presented in Table 3. The data for persons A and B, who appeared to judges both in the lights and video modes, were correlated across modes, once for the semantic differential set and once for the Gough set. For semantic differential scales, the correlation for discussant A in video vs. lights is .90, and the same correla4The adjective checklist was scored in five other ways: a . number of judges who checked characteristic; b , number who checked uncharacteristic; c. number who checked no impression; d. ratio of a / ( a + b ) ; e. the ratio in d computed only for those words where fewer than six judges checked no impression. I n all cases, the results were similar; hence only one measure is reported in this paper.

SOCIOMETRY

548

TABLE 3

Mean Ratings ( N = 1 0 ) of Gough Adjectives and Osgood Semantic Diflerential Scales iot Response t o the Lights and Videotape Versions of Calley Trial Discussion Person B

Person A Gough

Adjective

Aggressive Aloof Anxious Appreciative Arrogant Conceited Cooperative Daring Demanding Dependent Dominant Emotional Forceful Frank Gentle Goodnatured Hardheaded Hostile Humorous Immature Independent Inhibited Irritable Mature Meek Methodical Opinionated Peaceable Persistent Reasonable Rude ' Self -confident Self-controlled Sensitive Sexy Shy Sincere Snobbish Sociable Spontaneous Strong Submissive

Light

Video

Light

Video

2.8 1.9 1.9 2.0 2.3 2.1 1.7 2.4 2.5 1.5 2.9 2.2 2.5 1.9 1.9 1.6 2.5 2.5 1.9 2 .O 2.5 1.7 2.2 1.9 1.5 1.5 2.2 1.7 1.7 1.8 2.3 2.1 1.7 1.8 2.0 1.8 1.8 2.2 2.2 2.3 2.6 1.4

2.6 2.7 2.1 1.9 2.7 2.4 1.5 2.7 2.5 1.9 2.8 1.8 2.6 1.4 1.7 1.5 2.6 2.9 2.2 2.2 2.3 2.0 2.5 1.8 2 .O 2.1 1.9 1.4 2.4 1.3 2.8 2.0 1.6 1.5 2.7 1.9 1.2 2.7 1.5 2.1 2.7 2.1

1.0 2 .O 2.2 2.1 1.7 1.9 2.1 1.9 1.3 2.8 1.0 2 .O 1.2 1.8 1.9 2.0 1.7 2.1 2.1 2.1 1.3 2.7 2.0 2.0 2.6 1.6 1.3 2 .O 1.0 1.9 1.9 1.0 1.9 1.9 2.1 2.8 1.8 2.1 1.9 1.5 1.2 3 .O

1.8 2.4 2.2 1.9 2.3 2.3 1.4 2.1 1.6 2.4 1.9 2 .O 2.1 1.2 1.8 1.2 2.4 2.9 1.4 2.6 1.2 2.7 2.7 1.4 2.6 1.8 1.8 1.3 1.9 1.2 2.7 1.3 1.3 1.5 2.5 2.4 1.2 2.7 1.4 1.4 1.5 2.7

INTERPERSONAL JUDGMENTS AND TALKATIVENESS

549

TABLE 3-Continued

Mean Ratings (N=10) of Gough Adjectives and Osgood Semantic Differential Scales i n Response t o the Lights and Videotape Versions of Galley Trial Discussion Person B

Person A Gough Adjective

Light

Video

Light

Video

3.2 3.8

3.1 3.6

5.5 4.7

5.2 4.3

3.3 4.2 4.0 4.1 3.9 3.4 3.5 3.3

3.2 4.4 4.5 4.5 3.2 3.5 3.1 3.3

5.8 4.6 4.7 4.3 4.8 5.2 4.9 5.1

5.4 4.7 4.9 4.4 5.3 4.5 4.3 4.8

Tactful Talkative Timid Understanding Uninhibited Unselfish Warm Withdrawn Semantic Differential Scales

Passive-Active Fast-Slow InconspicuousProminent Bad-Good Worthless-Valuable Unpleasant-Pleasant Dull-Sharp Weak-Strong Soft-Hard Quiet-Loud

Note: Gough adjectives were rated as follows: lzcharacteristic; 2=no impression whatever; 3=uncharacteristic.

tion for discussant B is .94. For Gough adjectives, the correlation for A is .55 and the correlation for B is .49. Clearly these data again provide support to the hypothesis that the temporal aspects of talk provide sufficient cues for assessing many attributes of the person speaking. Study 3B. For the semantic differential words only, correlation coefficients were computed to determine how similar the mean ratings were for each of the eight stimulus persons (or lights) in the four presentations. All possible correlations were computed yielding an eight by eight matrix. I t should be noted that this matrix includes the following types of correlations: ( a ) A person correlated with his fellow discussant, both shown in video. (b) A person correlated with his fellow discussant, both shown in their lights analogs. (c) A person shown in video correlated with the lights analog of the same person. ( d ) A person shown in video correlated with the lights analog of his fellow discussant. (e) A person or light representing

SOCIOMETRY

550

that person, in one discussion, correlated with another person or light, in another discussion. If vocal activity were the only factor controlling the judgments, then the correlations between any two stimulus persons (or lights) should be highest where vocal activity estimates are most similar. Inspection of the matrix shown in Table 4 indicates that this is generally, but by no means comTABLE 4

Intercorrelations Among Participants (or Their Lights Analogs) i n Three Dyadic

Discussions About the Calley Trial, for Mean Ratings on Tern Senzantic

Differential Scales

Mean Perceived Activity

Person and Treatment Mode

DL

Av

AL

Ev

Fv

Bv

BL

CL

D, lights ( D L )

A, video (Av)

A,lights (AL)

E,video (Ev)

F,video (Fv)

B,video (Bv)

B,lights (BL)

C ,lights (CL)

Notes: Decimal points omitted. The dyads were: A and B ; C and D ;E and F

pletely, the case. While vocal activity is strongly implicated in person perception, it is clearly not the only determinant. STUDY 4. IMAGINARY CONVERSATIONS

The combined results of the first three studies provide strong evidence that the information in total talk is not due to confounding with kinesic, proxemic, intonational, gestural or even content cues. There remains, however, a rich set of purely temporal cues, other than the total amount of talk, to which persons may be responding in the lights experiments. Such cues include: which person interrupted whom, how often, how long, who won or lost, who was slow or fast to respond when there was an opening in the conversation, and who was given to monologues as opposed to dialogues. Any or all of these might be considered as rival hypotheses to the hypothesis that persons evaluate the total amount of talk. Accordingly Study 4 was designed to test these rival hypotheses. I t s first objective was to inquire: Insofar as judgments of a speaker seem traceable to the temporal aspects of his vocal contribution, are the judgments due to the total amount of his contribution or to the remaining

INTERPERSONAL JUDGMENTS AND TALKATIVENESS

551

temporal aspects? A second objective was to provide some assurance that the results found for the discussions of the celibacy of priests and the Calley trial would generalize to conversations on other topics. Since only a "random sample of all types of conversations" would truly satisfy this issue of generalizability, a more indirect approach was devised. This approach, which we term the "numbers-only technique", was utilized in pursuing both objectives.

Method Two sets of respondents were told that they would be given an absolutely minimal amount of information about three men who carried on an actual conversation on the wisdom of permitting Catholic priests to marry. The only information they would know about each man was a number representing the share he contributed to all the talking which took place during the conversation. Their task would be to describe each man as best they could. One group of respondents was then given three numbers corresponding to the mean estimates of activity previously given by those who watched the conversation of Study 1 in its "lights" version (A = 38% ; B = 19% ; C = 4 8 % ) . T h e second group was given three numbers corresponding to the mean estimates given b y those who had watched the video version (A = 53%; B = 8 % ; C = 39% ) .5 A third set of respondents, acting as controls, watched the lights stimulus for seven and one-half minutes. Gough adjectives were rated by all three sets of respondents, and scored as in Study 3. The research question asked was: Would the judges rate a person in the imaginary conversation, about whom they knew nothing but a number describing his total share of talk, as did the judges who had watched the seven and one-half minute passage presented through the lights?

Results The data are presented in Table 5. The main result is that the fifty Gough adjectives were arrayed in similar ways by those who watched the lights and b y those who received just the three numbers. The correlations This discrepancy in shares of the conversation between those who watched the lights and videotape is replicable and as yet not fully understood. The differences however provide an opportunity to test a proposition from Hayes and Sievers (1972) and Hayes and Meltzer (1973). In those studies a clear curvilinearity is reported between vocal productivity and affective judgments. The proposition states that up to the maximum positive region of activity, increments in activity lead to more positive judgments; beyond that point further increments in activity yield negative evaluations. There were eighteen opportunities to test this hypothesis and all eighteen are consistent with the prediction (Hayes and Meltzer, 1973).

SOCIOMETRY

552

between the mean ratings based on lights and the mean ratings based on the

numbers-only stimuli, i.e., the data shown in adjacent columns of Table 5,

TABLE 5

Mean Ratings of Gough Adjectives i n Response to the Lights aazd Numbers-Only Versions of the Celibacy Discussion Person A Cough Adjective

Aggressive Aloof Anxious Appreciative Arrogant Conceited Cooperative Daring Demanding Dependent Dominant Emotional Forceful Frank Gentle Goodnatured Hardheaded Hostile Humorous Immature Independent Inhibited Irritable Mature Meek Methodical Opinionated Peaceable Persistent Reasonable Rude Self-confident Self -controlled Sensitive Sexy Shy Sincere Snobbish

Light

Number

Person B Light

Number

Person C Light

Number

INTERPERSONAL JUDGMENTS AND TALKATIVENESS

553

TABLE 5-Continued

Mean Ratings of Gough Adjectives i n Response t o the Lights and Numbers-Only Versions of the Celibacy Discussion Person A Gough Adjective

Sociable Spontaneous Strong Submissive Tactful Talkative Timid Understanding Uninhibited Unselfish Warm Withdrawn

Person B

Person C

Light

Number

Light

Number

Light

Number

1.42 1.42 1.21 2.83 1.83 1.25 2.92 1.83 1.46 2.00 1.75 3.00

1.23 1.20 1.50 2.80 2.30 1.33 2.90 2.10 1.30 2.13 2.00 2.87

2.25 2.58 2.63 1.33 1.71 3.00 1.29 1.58 2.83 1.71 1.88 1.38

2.53 2.67 2.33 1.73 1.63 2.90 1.43 1.53 2.67 1.80 1.97 1.43

1.75 1.08 1.13 2.96 2.50 1.08 2.96 2.29 1.21 2.54 2.17 3.00

1.10 1.20 1.43 2.67 1.97 1.20 2.87 1.77 1.37 1.83 1.77 2.83

Notes: Gough adjectives were rated as follows: l=characteristic; whatsoever ; 3=uncharacteristic. M=24 raters for lights, 30 raters for numbers-only.

2=no

impression

are: .78 for Mr. A; .96 for Mr. B; and .74 for Mr. C. The information available for person perception in temporal pattern (lights stimulus) appears relatively redundant with the information conveyed by knowledge of total participation alone (numbers-only stimulus). The evidence in this study shows that information about temporal pattern of participation is substantially redundant with the information derived from the total amount of talk. This is not to say that the temporal patterning does not or cannot contain information which might lead to conclusions other than those based on sheer verbal productivity, but clearly temporal patterning sometimes adds little new information to that available merely from total vocal productivity. This experiment, in conjunction with Studies 1-3, demonstrates that sheer verbal productivity, i.e., total amount of talking, is a suficient cue for many interpersonal judgments, in numerous situations. We feel emboldened to say "in numerous situations," in spite of using tapes drawn from only two types of conversations, because the hypothetical numbers-only "conversations" are quite independent of any stimulus conversations which might be recorded by us or any other researcher. Despite this, the results derived from so general a stimulus closely parallel those from specific conversations, suggesting that the results from these earlier studies may have wide generality.

SOCIOMETRY DISCUSSION

T h e reader would probably have been willing to grant, before he read this paper, that judgments of activity, dominance, prominence, or assertiveness might well be made from the total amount of a person's talking. If he has read T h e Measurement of Meaning (Osgood et al., 1957) he would probably not be surprised that sheer talk could also provide a good cue for estimating words highly loaded on Osgood's potency d i m e n ~ i o n .But ~ he may be somewhat puzzled a t the extent to which sheer talk provides a sufficient basis for judgments of an evaluative or affective nature. Besides being contrary to common sense, this fact seems contradictory to the reports of various factor theorists that activity and evaluation are orthogonal (Osgood et al., 1957) or that dominance (closely linked to activity) and interpersonal affect are orthogonal (Foa, 1961). How then is it possible for sheer talk to be related to evaluation and affect? Work presented elsewhere (Bales, 1956; Hayes and Meltzer, 1973; Hayes and Sievers, 1972) indicates that there is a U-shaped relationship between judgments of activity and judgments of evaluation, in the sphere of person who talk a great deal or very little are rated unfavorp e r ~ e p t i o n .Persons ~ ably and they are described as having predominantly unpleasant attributes. T h e most favorable evaluations are given to persons who contribute somewhat more than their share to the conversation. P u t this way, perhaps the results are not inconsistent with common sense after all. Whether or not the result is viewed as surprising, the fact is that total talk itself, not something associated artifactually with quantity of talk, is sometimes a sufficient basis for affective and evaluative judgments. This raises several questions. How is it that activity has acquired these evaluative and affective connotations? Why does the organism ever utilize this cue, rather than or in addition to other factors? One line of explanation is evolutionary. Perhaps man continues to use the same cues to communicate primitive affect as do some other animals. Vocal activity is one of the most obvious attributes differentiating one animal from another. I t would be surprising if animals did not utilize such an obvious cue for primitive information reception. I n the course of evolution there may have developed a built-in sensiT o quote Osgood et al. (1957, pages 73-74) : "When the sample of things being judged is restricted in some fashion, the nature and order of magnitude of the factors may change. For example, when judgments are limited to sociopolitical concepts (people and policies), there seems to be a coalescence of the second and third factors into what might be called a 'dynamism factor'. . . . I t is as though things in this frame of reference that are 'strong' are also necessarily 'active'. . . ." 7 Hayes and Sievers (1972) offer several explanations for the discrepancy between their results and those reported by factor analysts.

INTERPERSONAL JUDGMENTS AND TALKATIVENESS

555

tivity to this prominent feature of every animal's environment. A second line of explanation is ontogenetic. Over time, via operant conditioning, those socially inept people who have difficulty in speaking learn to be quiet because they have been punished for talking. Via secondary conditioning, quietness then becomes a cue to indicate ineptitude and various other unfavorable social and personal characteristics. Similarly, we suspect that most people have learned through their own experience that very high talk is associated with excessive concentration in the task area (Bales, 1956), insensitivity to others' needs, and attempts to dominate (Chapple, 1970; Kendon, 1963) .8 While we have shown that total talk is sufficient for judgments, the extent to which this cue is actually utilized in naturalistic situations is an open question. That we consistently find in our studies high redundancy of lights and video judgments suggests that sheer talk is often used. To ascertain how this cue is used when other cues are available will require a program of studies, systematically varying total talk, on the one hand, and other cues (content, proxemics, kinesics, etc.), on the other hand, as well as also taking into account the effects of situational and dispositional variables. Several researchers (e.g., Mehrabian and Ferris, 1967; Mehrabian and Wiener, 1967; Argyle et al., 1970) have conducted experiments contrasting the degree to which judgments are made from non-verbal (content-free) and from verbal (content) cues. I n their studies, non-verbal factors clearly predominate over verbal cues. Similar studies should be conducted comparing the cue of total vocal activity with the remainder of the cues in the non-verbal realm. The cue of total talk may be particularly important in initial impression formation situations such as those studied in our experiments, namely brief encounters. When first meeting or observing another person, we may size him up largely in terms of his vocal activity patterns. But as we have prolonged exposure to the person, we may come to evaluate him in terms of the degree to which his values are congruent with ours (Newcomb, 1961)) the instrumentality of his behavior for the enhancement of our values (Rosenberg, 1956)) the extent to which he and we are cognitively consistent (Abelson et at., 1968)) the number of reinforcements he provides to us per unit of time (Byrne and Nelson, 1965)) and so forth. Why do we suspect that sheer talk is a particularly potent cue when forming initial impressions? I n part, because it takes time to acquire some of the other kinds of information just mentioned, while activity level can be judged rapidly (though perhaps inaccurately). Also, in the initial stage 8 Brian Little has suggested (personal communication, 1971) that some overtalkative persons are defective in skills needed for listening whereas some quiet persons are defective in skills needed for talking.

556

SOCIOMETRY

of a relationship, one monitors oneself and the other person rather carefully (Thibaut and Kelley, 1959). The content of such discussions is prescribed by the social norms of politeness, and hence may be discounted. Without prolonged acquaintance, one may not pay too much attention to the verbal content of a conversation (whether observing or participating), for people ordinarily conceal personal information until some time has elapsed (Jourard, 1964). Since the content of the other person's talk may not be considered as very revealing, each interactant probably pays special attention to the other's non-verbal behavior. These speculations suggesting that vocal activity may be a particularly potent cue for early impression formation should not be considered as an adequate substitute for the research which obviously needs to be done. Our experiments showed different levels of redundancy (lights with video judgments) for different adjectives. How can these differences be explained? We assume that what is perceived in the lights is "pure activity." Therefore, the ease or difficulty of coordinating a word to the lights depends upon the word's activity "loading" in the factor analysis sense. If the activity loading is very low, it is impossible to judge the light using that word. But since this particular restriction is absent in applying the word in the video condition, there would be low redundancy of judgments across the lights and video modes for words with low activity loadings. Note that this argument holds for words with an affectual or evaluative coloring. There has to be some activity loading of the attribute judged in order to obtain similar evaluative or affectual judgments across the lights and video modes. Our own evidence of high consensus adjectives used to evaluate persons shows that such words also have a pronounced activity component (Hayes and Sievers, 1972; Hayes and Meltzer, 1973). This fact probably accounts for how respondents to the lights stimulus were able to deduce evaluative information from evidence of activity alone; variation in the extent of activity is curvilinearly related to evaluation. The experiments also showed different amounts of redundancy for the different stimulus persons judged. Why? In the video situation, the stimulus person may be perceived as active or inactive, and hence evaluated because of the U-shaped relation between activity and evaluation; but it is also possible that the stimulus person may be evaluated on other grounds. (We do not claim that activity is the ground for evaluation, only that it is a ground.) If evaluated on "other grounds" the judgment made in video may differ from the judgment made in lights, since these "other grounds" are excluded in the lights condition. But why do people pay particular attention to the activity of some persons they observe (as our judges in the celibacy discussion obviously did in re-

INTERPERSONAL JUDGMENTS AND TALKATIVENESS

557

spect to Mr. B, whose levels of redundancy were very high)? A facile answer is that there may be something which draws attention to the observed person's activity, e.g., his temporal pattern, or the intensity of his vocal performance, his speech rate, etc.; or there may be something which draws attention away from his activity, e.g., the profundity of his contribution or its insulting and obnoxious nature. But this is clearly a matter for future research. The primary implication of the present investigation is that the demonstrations reported should, when considered in conjunction with the literature cited a t the outset, lessen any tendencies to denigrate sheer talk as "mere talk." The effects of amount of talk cannot be explained away as artifactual. I t is clear that the medium (vocalization) is itself a message. I t would now seem appropriate for social psychologists to study the quantitative aspects of their main subject matter: social interaction, and in particular, talking. Our discipline has for too long considered speech, talking, communication as carriers of interesting variables rather than as phenomena of intrinsic interest. Talking is such a prominent behavior in social interaction that it is hardly surprising that human beings respond not only to the content of what is said but also to its manner (voice quality) and amount. These reactions to the quantity of talk include interpersonal feelings of affect and evaluation and group phenomena such as status hierarchies, leadership, and sociometric structures. A possible implication is in the area of therapy, particularly social skills training for neurotic and normal persons (Argyle, 1969). The amount of talk would appear to be a critical variable to manipulate if the intent is to have a person improve his self-perception. For self-conception is, in part, a form of perception of the way others react to oneself. Therapy would consist in increasing or decreasing a person's level of participation (through teaching, conditioning, role playing, practice, exhortation, whatnot) ; this new behavior would then be reacted to by significant others in the client's world; and finally the client would adjust his self concept as a function of the new way people are reacting to him. The Bavelas et a2 (1965) study suggests this type of training would have precisely such an effect. While we now know that amount of talk is related to evaluative judgments, it is not clear whether this relationship refers to actual (machine measured) or psychological (perceived) vocal activity. This would appear to be a fruitful area for various investigations. Two lines of research will be briefly outlined: ( 1 ) I t follows from the research reported here and in Hayes and Meltzer (1973) that a change in perceived vocal activity will be followed by a change in evaluation and a change in evaluation may well be followed by a change in perceived activity. If a machine records that a man talks 65%

SOCIOMETRY

of the time in a three-person conversation, we should expect that someone who likes this person very much (including the person himself) would underestimate the amount of time he talks, whereas someone who dislikes him very much would exaggerate the amount of time he talks. Given a discrepancy between the perceived activity and the evaluation (e.g., "He's very quiet, but I respect and admire him."), one would predict an incongruity, a dissonance, such that the perceiver would either tend to change his perception of one of these aspects or attempt to change the perceived object. Thus he might tend to draw out the liked but quiet person, or, conversely, to egg-on the moderately talkative but disliked person until such a person talks too much. ( 2 ) If perceived vocal activity is an important variable, then we should like to know: "What are the variables that make us perceive a person as more or less talkative?" The same objective amount of talk can be perceived as more or less by different persons; and two different speakers, talking an identical amount, may be perceived as differentially active by a single observer. What factors influence these activity estimates, and thereby the associated evaluative and affectual judgments? I n part, there are unknown experiential factors making for differential perception, but also, no doubt, there are measureable responses of the judged person which serve as stimuli to the response of the judge. If the purely physical cues we produce during the course of interaction have marked effects on the perceiver (see also Meltzer, Morris, and Hayes, 1971) then a program of social psychophysics, the study of one's social behavior and experience as a function of the cues produced by others in the situation is clearly in order. One needs to know: what stimuli do we produce in interaction? which of these are reacted to? and according to what functional relationships? If "perceived activity" turns out to be an important variable intervening between physically measureable amount of talk and evaluative responses, the same questions can be asked concerning the bases of perceived activity. Of course we recognize the importance of such cognitive principles as selectivity and interpretation of stimuli, but these principles have limits due to the nature of the stimuli. I t seems time for social psychology to pay much stronger attention to the direct measurement of the actual stimuli in interaction. The burgeoning literature on "nonverbal communication" suggests that social psychologists are beginning to see this need. CONCLUSIONS

The hypothesis posed a t the outset of this paper was that the sheer total amount of vocal output constitutes a sufficient cue for many interpersonal impressions. The steps in testing the hypothesis were: ( a ) find an analog to sheer talk which excludes the other cues with which vocal activity is

INTERPERSONAL JUDGMENTS AND TALKATIVENESS

559

usually confounded; ( b ) compare judgments made in the natural context with judgments made from the temporal analog; and (c) claim verification for the hypothesis if the judgments of the two types of stimuli are similar. The hypothesis was clearly verified, not only for terms heavily loaded with an activity connotation (e.g., dominant, modest, lively) but also for terms which have evaluative or affective connotations (e.g., valuable, friendly, skeptical, flexible). T o be sure, the latter type of person perception attribute is not as readily judged from the cue of verbal productivity as is the former type, but there is some basis in sheer talk even for the judgment of affective and evaluative terms. The authors hope that the demonstrations reported here, considered in conjunction with the literature cited a t the beginning of this paper, will lessen any tendencies to "explain away" the effects of amount of talk as artifactual, and will instead draw attention of the discipline to a variable which is both extremely easy to measure and of wide importance in human relations. REFERENCES Abelson, Robert P . , et al. 1968 Theories of Cognitive Consistency: A Sourcebook. Chicago: Rand McNally. Aiken, Edwin G. 1965 "Changes in interpersonal descriptions accompanying the operant conditioning of verbal frequency in groups." Journal of Verbal Learning and Verbal Behavior 4:243-247. Argyle, Michael 1969 "Training for social competence." Chapter 10 in Social Interaction. New York: Atherton Press. Argyle, Michael, Veronica Salter, Hilary Nicholson, Marylin Williams and Philip Burgess "The communication of inferior and superior attitudes by verbal and non1970 verbal signals." British Journal of Social and Clinical Psychology 9:222-231. Bales, Robert F. 1956 "Task status and likeability as a function of talking and listening in decisionmaking groups." Pp. 148-166 in L. D. White (Ed.), The State of Social Sciences. Chicago: University of Chicago Press. Bavelas, Alex, Albert H . Hastorf, Alan E . Gross and Richard W. Kite 1965 "Experiments on the alteration of group structure." Journal of Experimental Social Psychology 1 (January) : 55-70. Byrne, Donn and Don Nelson 1965 "Attitude as a linear function of proportion of positive reinforcements." Journal of Personality and Social Psychology 1 (June):659-663. Carment, David W. 1961 "Ascendant-submissive behaviour in pairs of human subjects as a function of their emotional responsiveness and opinion-strength." Canadian Journal of Psychology 15 (March) :45-51. Cervin, Vladimir B. 1967 "The relationship of ascendent-submissive behavior in dyadic groups of human subjects to their emotional responsiveness." Journal of Abnormal and Social Psychology 54 (March) :241-249.

560

SOCIOMETRY

Chapple, Eliot D . "Quantitative analysis of the interaction of individuals." Proceedings of the 1939 National Academy of Science 25 (February):%-67. Culture and Biological Man: Explorations in Behavioral Anthropology. New 1970 York: Holt, Rinehart and Winston. Couch, .4rthur A. Psychological Determinants of Interpersonal Behavior. Unpublished Ph.D. 1960 Thesis. Harvard University. Dawe, H . D . "The influence of the size of kindergarten group upon performance." Child 1934 Development 5:295-303. Efran, Jay S. and P . R. Korn 1967 "Some characteristics of reluctant ~articipants." Unpublished manuscript. Foa, Uriel G . "Convergences in the analysis of the structure of interpersonal behavior." 1961 Psychological Review 68 (September) :341-353. Gough, Harrison G . and A . B. Heilbrun, Jr. The Adjective Check List Manual. Palo Alto: Consulting Psychologists Press. 1965 Hare, A. Paul and Robert F. Bales "Seating position and small group interaction." Sociometry 26 (December): 1963 480-486. Hayes, Donald P . and Leo Meltzer 1967 "Bone-conducting microphones." American Journal of Psychology 80 (December) :619-624. 1973 "The parabolic relationship between evaluative judgments and activity." Submitted for publication. Hayes, Donald P . and Sally Sievers 1972 "A sociolinguistic investigation of the 'dimensions' of interpersonal behavior." Journal of Personality and Social Psychology, in press. Heise, David R . ''Semantic differential profiles for 1000 most frequent English words." Psy1965 chological Monographs 79 (8, Whole No. 601). Jourard, S. M. 1964 The Transparent Self. Princeton, N.J.: Van Nostrand. Kendon, Adam 1963 Temporal Aspects of the Social Performance in Two Person Encounters. Oxford D. Phil, thesis. Knutson, Andie L. 1960 "Quiet and vocal groups." Sociometry 23 (March):36-49. Mann, Richard D . 1961 '
INTERPERSONAL JUDGMENTS AND TALKATIVENESS

561

Mehrabian, Albert and Morton Wiener 1967 "Decoding of inconsistent communication." Journal of Personality and Social Psychology 6 (May): 109-114. Meltzer, Leo, William N . Morris and Donald P . Hayes 1971 "Interruption outcomes and vocal amplitude." Journal of Personality and Social Psychology 18 (No. 4) :392-402. Newcomb, Theodore M . 1961 The Acquaintance Process. New York: Holt, Rinehart and Winston. North, Alvin J . 1957 "Language and communication in group functioning." Group Psychotherapy 10 (December) : 300-3 18. Osgood, Charles E., George J. Suci and Percy H . Tannenbaum 1957 The Measurement of Meaning. Urbana: University of Illinois Press. Peabody, Dean "Trait inferences: Evaluative and descriptive aspects." Journal of Personality 1967 and Social Psychology 7 (4, Pt. 2) : 1-18. Peterman, Jack N. 1949 '[Verbal participation, its relation to decision satisfaction and the leader function in decision-making groups." Paper presented at the meeting of the Midwestern Psychological Association, Chicago, April. Pope, Benjamin and Aron W . Siegrnan 1966 "Interviewer-interviewee relationship and verbal behavior of interviewer in the initial interview." Psychotherapy: Theory, Research and Practice 3 (November) : 149-152. Riecken, Henry "The effects of talkativeness on ability to influence group solutions of prob1958 lems." Sociometry 21 (December):309-321. Rosenberg, Milton J. 1956 "Cognitive structure and attitudinal affect." Journal of Abnormal and Social Psychology 53 (November) :367-372. Strodtbeck, Fred and Richard D. Mann 1956 "Sex role differentiation in jury deliberation." Sociometry 19:3-11. Thibaut, John and Harold Kelley 1959 The Social Psychology of Groups. New York: Wiley. Zdep, Stanley M . 1969 "Intra group reinforcement and its effects on leadership behavior." Organizational Behavior and Human Performance 4 (3) :284-298.