Achieving "Shared Reality" in the Communication Game:
A Social Action That Creates Meaning
E. Tory Higgins
Columbia University
Mailing Address: Department of Psychology
Columbia University
Schermerhorn Hall
New York, New York 10027
PUBLISHED AS:
Higgins, E. T. (1992). Achieving "shared reality" in the communication game: A social action that creates meaning. Journal of Language and Social Psychology, 11, 107-131.
Abstract
This paper reviews research on the "communication game" (Higgins, 1981a) that supports two sets of conclusions. The first set of conclusions concerns communication as social action: (a) Communicators tailor their summary of target information to suit their audience's knowledge of or attitudes toward the target; that is, they achieve "shared reality" with their audience and thereby perform a social action. (b) Communicators' different motivations to achieve "shared reality" with their audience influences the extent to which they tailor their message to suit the audience, as evident in "super-tuning", "anti-tuning", and "non-tuning". (c) When there is a delay between successive messages about a target, communicators' use their first message to construct their second message even though the two audiences have different characteristics. The second set of conclusions concerns how communication as social action creates meaning: (a) Communicators use their message summaries about a target as a direct source of information about the target even when the message distorted the original target information to suit the audience, and these message summaries in turn influence the communicators' own memory and impressions of the target. (b) As the delay since communicating about the target increases, communicators' use of their message as a source of target information persists or even increases. (c)Communicators' messages about a target to suit their audience can have either beneficial or detrimental effects on the accuracy of their memory and impressions of the target.
Key Words for Indexing Purposes: interpersonal communication; audience effects; social cognition
Achieving "Shared Reality" in the Communication Game:
A Social Action That Creates Meaning
A decade has passed since I presented an approach to communication, the "communication game", whose purpose was to explore the relation between communication and social cognition (Higgins, 1981a; see also McCann & Higgins, in press). Since that time my colleagues and I have conducted several studies testing various implications of this approach. It is time to consider both what we have learned from this research and what might be studied next.
The "communication game" approach to communication was developed as an alternative to the "information transmission" approach that dominated social psychology during the 1960's and 70's, particularly in the United States. This approach emphasized the transmission of information from a source or sender to a destination or recipient and was inspired by mathematical models of communication (e.g., Shannon & Weaver, 1949). Whether the concern was communication accuracy or communication and persuasion, the information transmission approach dominated (see, for example, Glucksberg, Krauss, & Higgins, 1975; Mehrabian & Reed, 1968; McGuire, 1969). In many ways this approach was a child of its era, the "cognitive revolution" that began in the 60's. This approach remained heavily cognitive even when it was applied to clearly social phenomena, such as considering interpersonal persuasion in terms of attention, comprehension, and retention.
The purpose behind proposing the "communciation game" approach was to provide an alternative "social science" perspective on interpersonal communication-- a perspective that took seriously the insights and findings on language by sociologists, anthropologists, and other social scientists and philosophers (e.g., Austin, 1962; Garfinkle, 1967; Goffman, 1959; Grice, 1971; Gumperz & Hymes, 1972; Mead, 1934; Morris, 1938; Watzlawick, Beavin, & Jackson, 1967). I believed that considering communication from a "social science" perspective might shed new light on its relation to social cognition. I had always been fascinated with the Whorf-Sapir notion of "linguistic determinism" concerning the possible interrelation among language, thought, and society, and believed that linguistic determinism at the level of pragmatics (rather than syntax) might have important implications for social cognition. I thought that such a "social science" perspective might reveal how the social action of communication produces meaning in our everyday lives. How does language use in a social context influence thought?
It was not only the study of interpersonal communication that was dominated by the "cognitive science" perspective. Social cognition in the 70's also emphasized cognitive structures and processes, such as schemata and inferential heuristics (see, for example, Fiske & Taylor, 1984). Indeed, the cognitive science perspective continues to dominate social cognition, with the major concern being how cognitive factors influence social phenomena (the cognition of social psychology). I believed that a "social science" perspective could also make a contribution to social cognition by considering how social factors influence cognition (the social psychology of cognition). But what exactly is a "social science" perspective? What makes something "social"?
One answer to these questions that I especially like was provided almost a century ago by the eminent social scientist, Max Weber (1967, pp. 156-157):
In "action" is included all human behavior when and in so far as the acting individual attaches a subjective meaning to it...Action is social in so far as, by virtue of the subjective meaning attached to by the acting individual (or individuals), it takes account of the behavior of others and is thereby oriented in its course.
Thus, action is "social" when its orientation takes account of other people (see also Thomas & Znaniecki, 1918). Within social psychology, social cognition is concerned with the "social" nature of "cognition". As Weber (1967) also pointed out, people assign meaning to the events in their lives and then respond to those meanings. The social psychology of cognition, then, is concerned with how the meaning that people assign to the events in their lives (e.g., their representation, interpretation, and evaluation of life events) is transformed because their actions take others into account.
To the extent that message production and reception take others into account, interpersonal communication is one kind of social action that can influence cognition. In contrast to the "information transmission" approach, a major concern of the "communication game" was how communication partners take each other into account and its social-cognitive consequences. Thus, by taking a social science perspective, the study of communication naturally addresses the fundamental question of how social action creates meaning.
Role Enactment and Role-Taking
Interpersonal communication is social action in at least two major ways. First, it involves role enactment. Second, it involves role-taking (or perspective-taking). Although there are other ways in which interpersonal communication is social action, our research on the communication game has been concerned mainly with these two, especially the second. Thus, the present paper will focus on the implications of these two variables.
Role Enactment
Role enactment is a quintessential example of "social" action. In the traditional view, role enactment relates to conduct that adheres to certain positions in the social structure rather than to individuals per se. Role expectations is the conceptual link between social structure and role enactment. As defined by Sarbin and Allen (1968), role expectations "are comprised of the rights and privileges, the duties and obligations, of any occupant of a social position in relation to persons occupying other positions in the social structure (pp. 497)". Role enactment in terms of role expectations, then, is social action. Sarbin and Allen (1968) also point out that role expectations operate as imperatives concerning a person's cognitions as well as his or her conduct during role enactment. Thus, role enactment is social action that can influence meaning.
Despite its promise as a window on the social psychology of cognition, relatively little attention has been paid by social psychologists to the effects of role enactment on cognition. In an early study, Jones and deCharms (1958) used instructions to establish different perceiver roles (e.g., member of a judicial board of inquiry versus member of a medical-psychological board of review) for the same target person (an ex-prisoner of war who had signed communistic propaganda statements). The study found that the subjects' attributions of personality characteristics to the target on the basis of the same target person information varied markedly depending on the role that they were assigned. A study by Ross, Amabile, and Steinmetz (1977) demonstrated the influence of role enactment on both performance and the interpretation of it. For each pair of subjects, one member was assigned to play the role of "questioner" and the other to play the role of "contestant" in a "question and answer" game. The "questioner" role permitted the actor to ask challenging questions that displayed esoteric knowledge whereas the "contestant" role caused most actors to display a lack of such knowledge. Ross et al. found that the subjects in the different roles evaluated their own knowledge and their partner's knowledge differently despite being exposed to the same target behavior, with the "contestant" subjects evaluating themselves most negatively.
In a classic communication study, Zajonc (1960) found that subjects assigned the role of "transmitter" of information represent the information in a more unified and organized way than subjects assigned the role of "recipient" even prior to the communication taking place. Several subsequent studies basically replicated this finding (for a review, see Higgins, 1981). The results of these studies suggest that even preparation for role enactment can influence how information is processed. Many of these earlier studies, however, confounded role assignment with expectations of receiving additional information about the target; specifically, only subjects assigned the role of "recipient" expected to receive later more information about the target. To control for this problem, role assignment and target information expectations were manipulated separately in a study by Higgins, McCann, and Fondacaro (1982). Consistent with previous conclusions about the "transmitter" or "speaker" role (but not the "recipient" role), the "transmitter" role was found to influence cognition independent of target information expectations. Together, the results of this literature suggest that speakers tend to polarize and distort stimulus information in preparation for message production in order to meet their role-related obligation to produce clear, concise messages. The "cognitive tuning" literature, then, provides an excellent example of how taking into account others, in this case others' expectations about one's role obligations, influences the representation of social information.
Role-Taking 1: Tailoring Messages to Suit the Audience's Knowledge
Successful role enactment requires taking into account the expectations and standards of others. Role-taking or perspective-taking involves inferences about how others would respond to a particular stimulus or situation. Successful role enactment, therefore, requires role-taking. [Role-taking, on the other hand, need not lead to role enactment). There are two basic kinds of role-taking: (a) "putting oneself in someone else's shoes", or situational role-taking, which involves infering how you would respond if you were in the situation of another person; and (b) "seeing the world through another person's eyes", or individual role-taking, which involves infering how another person would respond if he or she were in your situation (see Higgins, 1981b). Role-taking is a fundamental process underlying all social interaction, including interpersonal communication. A basic rule of interpersonal communication is that each participant should take their partner's characterisitics into account (Grice, 1971; Mead, 1934; Piaget, 1926).
In considering the nature and consequences of people following this rule, the communication literature has focused mostly on communicators taking into account their audience's knowledge of a topic (see, for example, Clark & Haviland, 1977; Fussell & Krauss, 1989; Glucksberg et al., 1975; Grice,1975; Higgins, 1977). We also considered this case in our research on the communication game). Our particular concern, however, was to examine how taking into account another's knowledge influenced the representation of social information; that is, how this type of social action creates meaning (see Higgins et al., 1982).
Half of the communicators believed that their audience had received basically the same information about the target person, and the other half of the communicators believed that their audience had received different information about the target person. In fact, all subjects received exactly the same target person information. All communicators were asked to describe and interpret the target person information for their audience. To follow the rule of saying something worthwhile (Grice, 1975), the communicators with the same target information as the audience should emphasize interpretation over description (i.e., give their personal opinion) whereas the communicators with different target information than the audience should emphasize description over interpretation. Indeed, the communicators who believed they had different target information than the audience were more likely than communicators with the same target information to "stick to the facts" (i.e., produce messages in which the target information was neither distorted nor deleted nor evaluatively polarized); that is, they were more likely to emphasize description over interpretation.
What was the effect of this social action on meaning? All subjects recalled the original target person information both before and after producing their message. As shown in Table 1, the recall of the communcators who emphasized interpretation over description (because their audience supposedly had the same information as them) was less accurate after they produced the message than before (i.e., more distortions, deletions, or evaluative polarizations). Some decrease in accuracy over time would be expected simply because of the increased delay between input and recall. Thus, it is all the more remarkable that the recall of the communicators who emphasized description over interpretation (because their audience supposedly had different target information than them) was, if anything, more accurate after they produced the message than before (see Table 1). Social action in this case supported accurate recall.
------------------------------
Insert Table 1 about here
------------------------------
These results indicate that not only do communicators take their audience's knowledge into account in producing their message, but that this social action impacts on their own later knowledge of the subject of the message (i.e., changes in memory). It should be noted that these effects on the communicators' own knowledge were not simply a direct effect of the audience's supposed knowledge because the information about the audience's knowledge was available at the time of both their post-message and pre-message recall. Thus, differences in message production that took into account differences in audience knowledge, i.e., social action, were critical in creating differences in meaning.
Although clearly important, the knowledge of one's audience about the subject of the message is not the only audience characteristic that communicators need to take into account. Indeed, from a "social" perspective other audience characteristics might be even more important. One such characteristic is the audience's attitude toward the subject of the message. Everyone has had the experience of talking about someone to another person who one believes likes or dislikes the person being talked about. Most people feel some pressure to describe the target person in a way that does not conflict with the audience's attitude. But beyond this social pressure, taking into account the audience's attitude in producing one's message simply follows the rule that communicators should take their audience's characteristics into account. And there is considerable evidence that communicators do, in fact, tailor their message to suit their audience's attitude or opinion about the subject of the message (e.g., Higgins & Rholes, 1978; Manis, Cornell, & Moore, 1974; Newtson & Czerlinsky, 1974). How does this common social action influence social cognition? Much of the research program on the communication game has explored the intricacies of this seemingly simple question. In so doing, some major issues about the relation between communication and social cognition have been considered. Let us turn now to these issues.
Role-Taking 2: Tailoring Messages to Suit the Audience's Attitude
Our first study on audience attitude effects involved undergraduates at Princeton who believed they were assistants in an experiment on communication as a source of information transmission by describing a target person to another student (the audience) who would use their message to try to identify the target person. Both the target person and the audience were supposedly in the same "eating club" at Princeton. Prior to giving the communicator some written information about the target person, the experimenter casually mentioned that other information indicated that the audience either liked or disliked the target person.
This audience-attitude manipulation had a major effect on how the communicators summarized the target person information. The target person information contained evaluatively ambiguous behaviors of the target (e.g., behaviors that could be equally labelled "confident" or "conceited"; behaviors that could be equally labelled "persistent" or "stubborn"), unambiguously positive behaviors of the target (e.g., behaviors that could only be labelled "athletic"), and unambiguously negative behaviors of the target (e.g., behaviors that could only be labelled "short-tempered"). As shown in Table 2, communicators labelled both the ambiguous and the unambiguous target information more positively in their messages when they believed that the audience liked the target person than when they believed that the audience disliked the target person.
------------------------------
Insert Table 2 about here
------------------------------
Did this social action of taking another's attitude into account in one's message production influence the communicators' own memory of the original target information? As predicted, it did. Table 3 shows that the communicators' reproductions of the target person information were significantly distorted in the evaluative direction represented by their audience's attitude. Additional correlational analyses comparing subjects' messages and recall also indicated that communicators tended to use their message more as a source of information two weeks later than immediately. If communicators attempt to recall the original target information almost immediately after reading it and producing their message (i.e., 20 minutes later), they can and should rely on their mental representation of the target details to the extent that this representation is relatively complete and accessible. Two weeks later this mental representation of the target details is unlikely to be complete or accessible. Now the message, as a summary of the target information, is useful and even necessary as an additional source of information for reconstructive memory. It is notable that this directional-distortion effect increased over time for the unambiguous information but was evident for the ambiguous information almost immediately. This makes sense because even after 20 minutes subjects may need to use their messages to reconstruct the ambiguous information.
------------------------------
Insert Table 3 about here
------------------------------
It was clear in this study as well that the effect of audience tuning on memory was not due simply to a direct influence of the audience's attitude on memory (e.g., some kind of informational conformity effect). First, any direct effect of audience attitude on recall should be greater after 20 minutes than after two weeks because both salience of the attitude and the need to be consistent with it would be greater after a brief delay. But the recall distortion for unambiguous information was found only after the two week delay. Second, because of a supposed "experimenter error", half of the subjects in the study were given the information about their audience's attitude and expected to communicate but did not actually communicate. These subjects did not distort their recall in the direction of the audience's attitude.
The strength of communicators' motivation to tailor their message about a target person to suit the attitude of their audience toward the target person is well illustrated in a study by Sedikides (1990; Study 2). This study used the same basic paradigm as the Higgins and Rholes (1978) study. But in this study the subjects read the target person information before they even received the communication instructions or learned about the audience's attitude. In addition, subjects' representation of the target person information had already been experimentally influenced by priming or activating particular trait constructs. Despite all this, the subjects still tailored their summaries of the target person information to suit the attitude of the audience. That is, the motivation to achieve "common ground" was so strong that it overrode prior effects of priming on the initial representation of the target information. Moreover, consistent with the results of our studies, the communicators' own personal impressions of the target person were more consistent with their audience-tailored message than with the primed constructs.
These memory and impression effects from tailoring messages to suit the audience's attitude are another example of how social action can create meaning. What is especially interesting about such cases is that the communcators take their audience into account in producing their message, but do not take this audience-tuning sufficiently into account when they later use their messages in reconstructive memory. This is precisely why social action can create meaning. If the communicators during recall took into account sufficiently their previous "taking into account" during communication, then the messages produced would facilitate and not distort recall. But it is very difficult, if not impossible, for communicators to calibrate the extent to which the source of their message is just the target information itself (i.e.,the data) versus their tailoring of the message to suit the audience (i.e., the communicative context). Communicators are likely to underestimate the extent to which their audience tuning is a source of the message, and, therefore, this social action is likely to create meaning (e.g, produce memory distortions).
Role-taking 3: Tailoring Messages to Suit Successive Audiences
In the Higgins and Rholes (1978) study, the communicators produced a single message to a single audience. Most interpersonal communication studies have used this basic paradigm. But interpersonal communication in everyday life often involves communicating about some subject to more than one person over time. Imagine that you had some knowledge about a person and summarized this knowledge for two different audiences, first for an audience you believed liked the person and then for an audience you believed disliked the person. We know from previous studies that you would probably summarize your knowledge of the target person in a somewhat positive manner for the first audience. But what would you do for the second audience? And would your impressions of the target person be influenced more by your message to the first audience or by your message to the second audience?
For a variety of reasons one might expect that the first audience would determine both subsequent messages and meaning construction. Traditional models of communication audience effects have stated that it is the first audience that is critical (e.g., Zimmerman & Bauer, 1956). And "consistency" models generally suggest that people would be motivated to respond later to a target person in a manner that was consistent with their initial response to the target (see Abelson, Aronson, McGuire, Newcomb, Rosenberg, and Tannenbaum, 1968). Recent social-cognitive models, however, raise the possibility that the time interval between the first and second messages might be critical in determining the impact of the second audience on message production and subsequent personal impressions.
If the stored details of the original target information become increasingly inaccessible over time, then communicators would increasingly rely on their first message as a source of target information-- both for their subsequent message and for their own impressions of the target (see Higgins & Rholes, 1978). There may also be an increasing consolidation or integration of the representation of the target details and the message representation of the target (see Wyer & Srull, 1986). If so, the impact of communicators' first message on their subsequent message and on their own impressions of the target would also increase over time. According to both of these information-processing models, then, when there is a long delay between the first and second messages, the impact of the first audience on communicators' first message should produce a primacy effect on both the second message and subsequent personal impressions. But when there is only a brief delay he first and second messages, the second message should be tuned to the second audience and have the greater impact on personal impressions by virtue of being both most accessible and the "final word" on the target.
This study modified the basic Higgins and Rholes (1978) paradigm to increase its ecological validity. In this study the communicators interacted face-to-face with each audience under circumstances in which they believed that their audience would personally benefit by their summaries of the target person's characteristics. [The two audiences were both confederates of the experimenter.] The communicators believed that their summaries would help the audience make a decision about whether to select the target person as a roommate [all participants were male]. Before the communicators gave their summary to the audience, the audience mentioned that he had met the target person briefly and that his initial impression of the target person was either positive or negative ("kinda liked" or "kinda disliked" the target). The communicators received an essay describing various characteristics of the target. They summarized this information to two different audiences, either only 15 minutes apart (the Brief Intermessage Delay) or one week apart (the Long Intermessage Delay). In an orthogonal design, the first audience either liked or disliked the target person and the second audience either liked or disliked the target person. The communicators' own impressions of the target person were obtained one week after their second message.
Consistent with the results of previous studies, the communicators tailored their message to suit the expressed attitude of their first audience. The communicators' labelling of the target person information was more positive when they communicated to an audience who supposedly liked the target person than when they communicated to an audiece who supposedly dislike the target person. It is notable that this audience-tuning effect occurred despite the audience stating that their expressed attitude toward the target person represented only an initial impression based on a brief encounter with the target person and despite the fact that the audiences stated that an accurate description of the target person would be most useful in helping them decide whether to select him as a roommate. These results thus demonstrate the powerful effect of audience attitude on communicators' social action.
As predicted, the labelling in communicators' message to the second audience was tailored to suit the second audience's attitude when there was only a brief delay between the first and second messages. But when there was a long delay (one week) between the first and second messages, the evaluative tone of the second message labelling was determined by the evaluative tone of the first message labelling (see Figure 1). The results indicate that after the long delay communicators used the first message to represent the target person information and then used this first message representation as the basis for their second message. Thus for information-processing reasons, the communicators' second social action, after a long delay, was determined by their first social action.
------------------------------
Insert Figure 1 about here
------------------------------
It is notable that there was no message tailoring in this study to suit the attitude of the second audience when there was a long delay between the first and second messages.These results indicate that there is a variable that constrains audience tuning in message production-- the target data. The target person information in these studies is evaluatively ambiguous, both by containing some behaviors that are inherently ambiguous and by containing equal numbers of positive and negative behaviors. Such ambiguous data minimally constrains the evaluative tone of the message. Thus, audience tuning can ordinarily be realized without misrepresenting the data. When there is a long delay between the first and second messages, however, the evaluative implications of the first message can polarize or disambiguate the represented target person information. This constructed or consolidated "data" would now be either unambiguously positive or unambiguoulsy negative. It would no longer be possible to summarize the target information in whichever direction suited the second audience without misrepresenting it. Thus, the target data sets constraints on the amount of audience tuning that will occur.
When audience tuning did occur, how did this social action influence communicators' own impressions of the target person one week later? Once again, the impact of social action on construction of meaning was most evident when there was there was the longest post-message delay-- for the first audience attitude/long intermessage delay condition. Communicators' delayed impressions in this condition were more positively distorted when they tailored their first message to an audience who supposedly liked the target person (M= .84) than disliked the target person (M= .32), and they were more negatively distorted when they tailored their first message to an audience who supposedly disliked the target person (M= 1.99) than liked the target person (M= .95).
This effect of first audience attitude was completely wiped out when there was only a brief delay between the first and second messages because in this condition the second audience's attitude determined the evaluative tone of the second, "final word" messages. Given that successive messages on some subject typically occur relatively far apart, the results of this study suggest that it may, indeed, be the first message on a subject, i.e. the first social action, that has the greatest impact on creating meaning over time (see Zimmerman & Bauer, 1956).
This initial series of studies investigating the communication game clearly demonstrated that communicators take into account the characteristics of their audience when producing their message, both their audience's knowledge and their audience's attitudes. This research also demonstrates that this type of social action influences the communicators' own knowledge of the subject of the message-- the "saying is believing" effect. The next series of studies on the communicaton game focused on the motivational underpinnings of communicators' social action. Does the extent to which communicators tailor their message to suit the audience, i.e., the extent of their audience-tuning, depend on their motivational orientation toward others? And how does this influence the construction of meaning? To answer these questions, it is necessary to consider the motivational underpinnings of interpersonal communication.
Achieving"Shared Reality" as a Goal of Interpersonal Communication
As just mentioned, our initial studies demonstrated that communicators will tailor their message to suit their audience's knowledge of or attitude toward the subject of the message-- a case of "basic tuning". Why does such basic tuning occur? The phenomenon of communicators tailoring their message to suit the knowledge of their audience can be understood, and has been understood (e.g., Higgins et al., 1982), in terms of Grice's (1975) maxim of "relevance"--while sticking to the topic at hand, say something worthwhile or "new" for the audience (see also Clark & Haviland, 1977). But the "relevance" maxim or "given-new" contract does not seem applicable to the case of communicators tailoring their message to suit the attitude of their audience. Indeed, if communicators wanted to provide the audience with "new" information they would express an attitude contrary to the audience's attitude. But, in fact, they tailor their messages to match the audience's atttitude. In addition, the communicators are not trying to be as "truthful" as possible. Their messages are tailored to suit their audience's attitude rather than just being summaries of the target person information.
Applying Grice's (1975) "cooperative" principle also fails to solve the problem because communicators could cooperate by providing the audience with as accurate and objective summaries as possible rather than tuning the message to suit the audience. It could be argued that subjects in the Higgins and Rholes (1978) study were cooperative because tailoring the message to suit their audience's attitude increased the likelihood that the audience would correctly identify the target person-- a from of "grounding" in the service of definite reference (see Clark & Brennan, 1991; Clark & Marshall, 1981). But in the McCann, Higgins, and Fondacaro (1991) study both the communicator and the audience knew that the communication was about the previously identified potential roommate. Moreover, the audience asked the communicator for his own personal impression. Nevertheless, the communicators in this study also tailored their messsage to suit the attitude of their audience. Why, then, did the communicators not cooperate by providing either objective summaries or their own impressions of the target person?
The "communication game" approach suggests that there are many ways to be "cooperative" depending on the goal or purpose of the interpersonal communication (see Higgins, 1981). "Cooperation" and "grounding" are basic procedural principles or "means" of communication. But what are the interpersonal goals or "ends" that they serve? The now classic work by Brown and Levinson (1978) on politeness phenomena demonstrates that communicators will sacrifice the maxim of quality, as well as the maxims of "quantity" and "manner", for the sake of politeness. That is, some forms of cooperation are sacrificed for other forms of cooperation. Interpersonal communication is rampant with such "trade-offs" between forms of cooperation, and it is the specific goals of a communication that determine which forms (i.e., procedural rules) are given priority (see Higgins, Fondacara, & McCann, 1981).
The "communication game" approach proposed that different "means" or strategies would be selected or emphasized depending on the communicators' social-interaction goals (see Higgins, 1981). These goals include "task" goals, "face" goals (i.e., impression management), "social relationship" goals, and "entertainment" goals. Regarding social relationship goals, for example, "speech accomodation" is a strategy in which communicators shift their speech style to be similar to (convergence)or different from (divergence)the speech style of their audience in order to associate themselves with or disassociate themselves from their audience, respectively. Convergence is likely to occur when communicators are attracted to or seek approval from the audience (see, for example, Giles & Smith, 1979).
As we have seen, communicators also tailor their message to suit their audience's knowledge on a subject. Communicators need not be attracted to or seek approval from their audience for such social tuning to occur. In addition, tailoring a message to suit the audience's attitude can be motivated by goals other than social relationship goals, such as "face" goals (i.e., make a positive impression on the audience). And the social tuning itself can involve modifying the denotative and connotative content of a message as well as shifting the speech style. Thus, there are a variety of different forms and goals of social tuning.
In our studies described above it is unlikely that the communicators were trying to fulfill social relationship or face goals because the relationship with the audience was minimal. Rather, it is possible that the communicators were attempting to achieve some "shared reality" with the audience. Communication is not just transmitting information. The "communication game" approach emphasizes the fact that participants in a communicative interaction typically attempt to construct a common or shared reality (see Rommetveit, 1974; Ruesch & Bateson, 1968). For interpersonal communication to work at all, regardless of its goal, the communicator and the audience must share some mutual knowledge, such as who or what is being referred to in the message, who or what is the message subject or topic (see Clark & Marshall, 1981). To begin with, then, the communicator and audience must establish some common ground as a basic "means" of communication. This does not imply, however, that the goal of the communicator is to achieve common ground with the audience in the sense of a "shared reality" about the characteristics of the message subject. The communicator and audience need to agree that the communication is about "X", such as a particular target person "X", but they need not be motivated to have similar knowlege or opinions of the characteristics of "X".
Communicators in the studies described earlier did attempt to achieve "shared reality" with their audience concerning the target person's characteristics. Communicators who believed that their audience had different information than them about the target person shared their knowledge of the target person by describing his characteristics to the audience. Communicators who believed that their audience had a particular attitude toward the target person produced messages whose evaluative tone matched the audience's attitude. Thus, the goal of the communicators' social action was to achieve "shared reality". The results of our studies show that this type of social action influences and creates meaning.
A dominant objective of social interaction in general is to establish a common "social reality" (see Asch, 1952; Festinger, 1950; Sherif, 1936). Indeed, the concept of "social reality" may capture the essence of the social psychology of cognition better than any other single concept. As described by Festinger (1950), beliefs, attitudes, and opinions vary in the extent to which there is physical evidence for them. It is extremely rare for a belief or opinion to be supported by hard facts of incontrovertible physical reality. The basis for holding most beliefs and opinions is social reality, the fact that others share the belief or opinion:
...where the dependence upon physical reality is low, the dependence upon social reality is correspondingly high. An opinion, a belief, an attitude is "correct", "valid", and "proper" to the extent that it is anchored in a group of people with similar beliefs, opinions, and attitudes (1950, p. 272). Thus, the very nature of most of our reality is social. The reality of our subjective meanings is anchored in the fact that others share the reality. This is especially true of opinion, beliefs, and attitudes about social objects, such as a target person. Therefore, when communicators achieve "shared reality" with their audience about a target person, this "shared reality" is likely to be treated as reality, especially when the representation of the stimulus input, i.e., the physical reality, has decayed or become inaccessible.
By this reasoning, the impact on meaning from the social action of audience tuning largely derives from the fact that it involves achieving a shared reality. It would follow, then, that messages that do not achieve, or even prevent, a shared reality about a target person would not subsequently impact on communicators' memory and impressions of the target person, or at least would impact to a lesser degree. In order to consider this possibility in more detail, it is necessary to distinguish among different types of message modification as social action.
Types of Message Modification as Social Actions
To achieve "shared reality" with their audience, communicators must first identify the characteristics of their audience, such as their knowledge of and attitude toward the subject of the message. Then they must tailor their message to suit these characteristics (e.g., provide "new" information; express a congruent attitude). The cases that have been discussed thus far involve such "basic tuning". For some communicators there are additional characteristics of the audience, such as high power or status, that increase their motivation to tailor their message to suit the audience. Such cases involve "super-tuning". Because super-tuning involves producing a message to achieve "shared reality" with the audience, the message represents social reality and is likely to impact on the communicators' subsequent impressions and memory of the message subject.
Identifying the characteristics of the audience, and even taking those characteristics into account, need not lead to tailoring the message to suit that audience. Social action in communication does not imply attempting to achieve "shared realtiy". Communicators may correctly infer their audience's knowledge of or attitude toward the subject of the message but not be motivated to tailor their message to suit the audience. Indeed, some communicators may even be motivated to resist tailoring their message to suit the audience-- a case of "anti-tuning". Such anti-tuning would still be "social" because the audience is being taken into account, but it would not involve an attempt to achieve "shared reality" with the audience. Will the messages produced by anti-tuning impact on communicators' subsequent memory and impressions? To address this question a couple of further distinctions need to be made.
First, it is necessary to distinguish between anti-tuning that involves just a motivation to diverge from or reject "shared reality" with the audience and anti-tuning that also involves a motivation to converge or achieve "shared reality" with an audience opposed to the immediate audience. An example of the former might be a communicator who dislikes the audience and uses anti-tuning to distance himself or herself from the audience. An example of the latter might be a communicator talking about an in-group member to an out-group member who uses anti-tuning to achieve "shared reality" with other in-group members in opposition to the out-group audience. The message produced in the former case represents simply an absence of "shared reality", whereas the message produced in the latter case represents "shared reality" with an "inner" audience. Thus, the message should have more impact on the communicator's own subsequent memory and impresssions of the message subject in the latter case than in the former case.
Second, it is necessary to distinguish between impact on communicators' subsequent memory and impressions of the message subject and impact on communicators' subsequent memory and impressions of the relationship with the audience. Thus far, studies testing the communication game have been concerned with the former type of impact; in particular, with communicators' memory and impressions of the target person who is the subject or topic of the message. But it would also be possible to examine the effects of tuning, super-tuning, and anti-tuning on communicators' memory and impressions of the audience to whom they communicated. As mentioned earlier, the literature on "speech accomodation" has shown that communicators will shift their speech style to converge with or diverge from the speech style of their audience in order to associate themselves with or disassociate themselves from their audience, respectively (see, for example, Giles & Smith, 1979). When communicators use a convergent speech style they achieve a "shared reality" of association or union with the audience. When they use a divergent speech style they achieve a "shared reality" of disassociation or non-union with the audience. Thus, the production of such messages is likely to impact on the communicators' memory and impressions of their relationship with the audience.
What if communicators not only shifted their speech style but also modified their description or evaluation of the message subject to be similar to or dissimilar from their audience in order to associate themselves with or disassociate themselves from their audience, respectivley? Consider again the two cases described earlier. In the first case, a communicator dislikes the audience and uses anti-tuning simply to distance himself or herself from the audience. As discussed earlier, this anti-tuning is unlikely to impact on the communicator's memory and impressions of the message subject. However, this anti-tuning could impact on the communicator's memory and impressions of his or her relationship with the audience. In the second case, a communicator talks about an in-group member to an out-group member and uses anti-tuning to achieve "shared reality" with other in-group members in opposition to the out-group audience. In this case, the message produced is likely to impact on the communicator's memory and impressions of both the message subject and his or her relationship with the audience. If the communicator was talking about a target person whose group membership was unknown, the communicator might only be motivated to distance himself or herself from the audience. Then this case would be more like the first case and the impact of the message is likely to be restricted to the communicator's memory and impressions of his or her relationship with the audience.
There is a third basic type of social action involving message modification that also needs to be considered-- "non-tuning". Communicators may construct their message to suit their prior beliefs or attitudes rather than to suit either the immediate audience or an "inner" audience. For example, a communicator may believe that people in general have positive traits (the "Pollyana" effect) and thus produce a more positive summary of a target person than is strictly warranted by the target information. To the extent that other people are being taken into account (e.g., beliefs about what other people are generally like), this type of message modification is social action. But this message modification does not involve tailoring to suit some audience-- it is a non-tuning form of social action. Such messages reflect prior beliefs and attitudes but do not construct or achieve new "shared reality". Thus, these messages per se are unlikely to influence (i.e., change) communicators' memory and impressions. That is, they are unlike to create meaning.
In sum, there are four basic types of message modification as social action-- "basic tuning", "super-tuning", "anti-tuning", and "non-tuning". Basic tuning has already been discussed. Let us now review a couple of studies that involved "super-tuning", "anti-tuning", and "non-tuning" and examine how these types of social action impacted on meaning.
Super-tuning, Anti-tuning, and Non-tuning
Interpersonal communication is social interaction and social interaction varies depending on the relation between the interactants. One important relation between interactants is their relative power or status (see French & Raven, 1959; Kelley, 1979). How might the status of the audience relative to the communicator influence communicators' motivation to tailor their message to suit the audience? Doug McCann and I reasoned that communicators' motivation would depend on their personal orientation to status. A personality dimension that specifically distinguishes among people on this basis is "authoritarianism". There is considerable evidence that high authoritarians are more responsive and deferent to a higher status partner than are low authoritarians (e.g., Adorno, Frenkel-Brunswick, Levinson, & Sanford, 1950; Berg & Vidmar, 1975; Harvey & Beverly, 1961; Thibaut & Riecken, 1955). Indeed, there is evidence that low authoritarians respond negatively to a higher status partner (e.g., Epstein, 1965). Thus, when communicating to a higher status audience, high authoritarians might be super-tuners in order to increase their "shared reality" and thereby feel associated with the high status other, whereas low authoritarian communicators might be anti-tuners in order to decrease their "shared reality" and thereby distance themselves from the high status other. From the perspective of speech-accomodation theory (see Giles & Smith, 1979), high authoritarians should converge in order to express similarity, whereas low authoritarians should diverge in order to express dissimilarity. This hypothesis was tested in a study by Higgins and McCann (1984).
The Higgins and McCann (1984) study used the same basic paradigm as the Higgins and Rholes (1978) study discussed earlier. In addition to varying the supposed attitude of the audience toward the subject of the message (i.e., likes or dislikes he target person), the status of the communicators relative to their audience and the authoritarianism of the communicators also varied. Half of the undergraduate subjects communicated to another undergraduate (the Equal Status Audience condition), and the other half of the subjects communicated to a relatively high-status, high-authority senior graduate student (the High Status Audience condition). In each of the four experimental conditions (i.e., type of audience attitude X type of audience status), half of the subjects were selected to be high authoritarians and half were selected to be low authoritarians.
Message tailoring to suit the audience was measured in this study by the extent to which a message evaluatively distorted the original target person information in the direction of the audience's attitude. To obtain this measure, "blind" coders counted the number of times each message either positively or negatively distorted the original target information. When the communicator and the audience had equal status, both high and low authoritarians showed the usual amount of message tailoring to suit the audience; i.e., basic tuning (see top half of Table 4). But when the audience had higher status than the communicator, there was a striking difference in how much high and low authoritarians tailored their message to suit their audience's attitude. As shown in Table 4, the high authoritarians increased the amount they tailored their message, whereas the low authoritarians decreased the amount they tailored their message. Indeed, when the audience supposedly liked the target person, the low authoritarians actually distorted the target information in a predominantly negative direction-- a very strong divergent signal and a remarkable case of anti-tuning! In the high status audience condition, then, there was evidence of both super-tuning (for the high authoritarians) and anti-tuning (for the low authoritarians) in communicative social action.
------------------------------
Insert Table 4 about here
------------------------------
As in our other studies, this study also examined the impact of these different types of social action on the communicators themselves. Again there was evidence that communicators' own memory of the original target information was distorted in the direction of their audience-tailored messages. This effect actually increased over time for the high authoritarian communicators (i.e., increased from the immediate recall measure to the two-week delay measure). Indeed, after two weeks only the high authoritarian communicators had a significant relation between their messages and their recall. A similar pattern of results was also found on a measure of the subjects' personal impressions of the target person ("What sort of person do you think the stimulus person is?") and on a measure of the subjects' own attitude toward the target person (on a scale from "extremely desirable" person to "extremely undesirable" person).
These results are consistent with the proposal that achieving "shared reality" with the audience about the message subject has an increasing impact on communicators' memory and evaluations of the message subject, because overall greater "shared reality" with the audience was acheived by the high than the low authoritarian communicators. Moreover, on the attitude measure especially, there was evidence that the increased effect over time for high authoritarian communicators was greater when they communicated to the high status than the equal status audience. This is also consistent with the proposal because greater "shared reality" with the audience was achieved by the high authoritarians when they communicated to the high status than the equal status audience. Thus, the super-tuning of high authoritarians when communicating to a high status audience led over time to greater self-persuasion. In contrast, the anti-tuning of the low authoritarians when communicating to a high status audience prevented the usual audience effects on social action and constructed meaning.
It should be noted for this study as well that the obtained "saying is believing" effects were not due simply to a direct effect of audience attitude on communicators' own memory, impressions, and attitudes. Figure 2 shows a path analysis for the high authoritarian subjects of the relations among the supposed attitude of the audience toward the target person (like or dislike), the subjects' message about the target person, and the subjects' reproductions of, impressions of, and attitudes toward the target person. Figure 2 clearly shows that there was no direct path from audience attitude to either subjects' reproductions, impressions, or attitudes. Rather, the effect of audience attitude on each of these measures was mediated by its effect on subjects' message descriptions. That is, the results clearly reflect a "saying is believing" effect and not an "audience believing" effect.
------------------------------
Insert Figure 2 about here
------------------------------
The results of this study demonstrate that the social construction of meaning can vary depending on the types of audience tuning that is performed, i.e., "super-tuning" versus "anti-tuning". The results of a study by McCann and Hancock (1983) also support this conclusion. In their study, the audience-tuning behavior of high and low self-monitors were compared. High self-monitors display another kind of super-tuning. Unlike high authoritarians who identify with high status persons and want to be associated with them, high self-monitors are concerned with the situational and interpersonal appropriateness of their behavior and use the characteristics of any interactive partner as clues to regulate their own behavior (see Snyder, 1979). Thus, high self-monitors are motivated to achieve "shared reality" in order to behave appropriately in the situation. In contrast, the behavior of low self-monitors is controlled by their internal states and prior values, beliefs, and attitudes rather than being tailored to fit the social situation (see Snyder, 1979). Thus, low self-monitors are not motivated to achieve "shared reality" with their audience. High self-monitors are likely to be super-tuners whereas low self-monitors are likely to be non-tuners.
This hypothesis was tested in a study by McCann and Hancock (1983), which again used the same basic paradigm as the HIggins and Rholes (1978) study. Consistent with the findings of Higgins and Rholes (1978), there was a general tendency for communicators to label the target person information in their message so as to suit their audience's supposed attitude toward the target person. But as shown in Table 5, this audience-tuning effect was only evident for the high self-monitors. The low self-monitors did not show any significant audience-tuning effect. Instead, as evident in Table 5, low self-monitors showed a clear positivity bias in their message labelling, even when the audience supposedly disliked the target person! This positivity bias could derive from low self-monitors' believing that people generally possess positive traits (the "Pollyanna effect") or that they should give people the benefit of the doubt (the "leniency" effect).
------------------------------
Insert Table 5 about here
------------------------------
This difference between high and low self-monitors in social action produced differences in meaning. Consistent with their message tailoring, the high self-monitors distorted their recall and personal impressions of the target person's behaviors in the direction of the audience's attitude. In contrast, the low self-monitors revealed simply an overall positivity bias in their impressions of the target person that reflected the same positivity bias found in their messages. Thus, as would be expected, a clear "saying is believing" effect was found for the high self-monitors who were "super-tuners", but was not found for the low self-monitors who were "non-tuners".
What Have We Learned, and What's Next?
The studies investigating the communication game have yielded consistent findings that permit some clear conclusions to be drawn. The first set of conclusions concern communication as social action. First, communicators generally take their audience into account and tailor their summary of target information to suit their audience's knowledge of or attitudes toward the target. That is, communicators achieve "shared reality" with their audience and thereby perform a social action. Second, communicators have different motivations to achieve "shared reality" with their audience that influence the extent to which they tailor their message to suit the audience. In addition to the basic level of audience tuning, there is "super-tuning", "anti-tuning", and "non-tuning". That is, communicators perform different types of social action depending on their motivation to achieve "shared reality". Third, when there is a delay between successive messages about a target, a later message will resemble the initial message even if there is a change in the characteristics of the audience to whom the message is addressed. That is, communicators' initial social action can determine their later social action.
The second set of conclusions concern how communication as a form of social action creates meaning. First, communicators generally will use their message summaries about a target as a direct source of information about the target even when the message distorted the original target information in the direction of the audience. As a result, communicators' own memory and personal impressions of a target can become distorted. That is, communication as social action can create meaning. Second, as the delay since exposure to and communication about the original target information increases, communicators' use of their message summaries as a source of target information persists or even increases. That is, the impact of communication on creating meaning persists or even increases over time. Third, when communicators produce detailed descriptions of a target to suit their audience's need for such knowledge, the message production can improve the accuracy of their memory of the target information. That is, communication can have both beneficial and detrimental effects on meaning construction.
It is evident that communication as social action impacts on the construction of meaning-- communicators come to believe what they previously said in their message. What underlies this "saying is believing" effect? One possibility is that the supposed knowledge or attitude of the audience whom the communicator addresses, which in some studies includes the experimenter, directly influences the communicator's own memory and impressions of the subject of the message. This "normative" or "informational" conformity effect (see, for example, Deutsch & Gerard, 1955) is a classic alternative explanation for the "saying is believing" effect. It has been addressed both in the role-playing literature (e.g., Hovland, Janis, & Kelley, 1953) and in the counterattitudinal advocacy literature (e.g., Bem, 1967; Festinger, 1957; Wicklund & Brehm, 1976). Whether the explanation for the "saying is believing" effect is "active participation" (or personal "improvisation"; see Hovland et al., 1953), "dissonance reduction" (see Festinger, 1957, Wicklund & Brehm, 1976), or "self-perception" (see Bem, 1967), normative or informational conformity is dismissed as a sufficient explanation. The role-playing and counterattitudinal literatures report that the "saying is believing" effect occur only if the communicators are induced to believe that they themselves chose what to say in their message. This condition is "active participation" in role-playing (involving personal "improvisation"), "high choice" in dissonance, and "non-manded" in self-perception, each of which is necessary for the "saying is believing" effect to occur. In these literatures, the alternative condition in which the "saying is believing" effect does not occur is forced compliance or non-action (e.g., passive exposure to another's message).
Our studies testing the implications of the communication game were also concerned with ruling out simple normative or informational conformity as a sufficient explanation for our version of the "saying is believing" effect. In the earlier discussions of these studies, various reasons were given for ruling out this alternative explanation. These include: (a) the communicators' knowledge of the audience's characteristics did not impact on meaning (i.e., communicators' own memory or impressions of the target) unless the communicators actually produced a message; (b) the communicators' knowledge of the audience's characteristics had a greater impact on meaning after than before message production; (c) path analyses revealed no direct effect from audience attitude to meaning, only a mediated path through message production; (d) the impact of message production on meaning often increases over time.
The results of our studies, then, support the previous literature on the "saying is believing" effect in dismissing simple normative or informational conformity as a sufficient explanation for the findings. But there is a fundamental difference between our version and explanation of the "saying is believing" effect and those of the previous literature. The previous literature demonstrated self-persuasion effects of "freely" advocating a position on some topic for some future general audience whose characteristic were basically unknown. Audience tuning was not an issue in these studies. In fact, the "saying is believing" effect was thought to occur only to the extent that communicators considered the message to contain their own spontaneous, improvised arguments. The communicators were not expected to modify their messages by taking others into account. They were expected to produce their own personal messages. Thus, the "saying is believing" effect was conceived as deriving from "personal" action rather than from "social" action. Indeed, "social" action that involved taking others into account would have been associated more with the "no choice" or "manded" conditions involving forced compiance that diminishes the "saying is believing" effect.
In contrast, the communication game studies do involve social action--communicators tailoring their message to suit the characterisitics of the audience. And it is precisely this "social"-tuning action that produces the "saying is believing" effect. When communicators did not tailor their message to suit their audience there was no clear"saying is believing" effect-- the "anti-tuning" of the low authoritarians and the "non-tuning" of the low self-monitors. These latter two responses are, if anything, more "personal" than the usual audience-tuning response. Thus, the "saying is believing" effect was less clear when the messages were more "personal". In addition, the "saying is believing" effect found in our "social" action studies tended to persist or even increase over time whereas the "saying is believing" effect found in previous studies of "personal" action tend to decrease over time.
This comparison of the "saying is believing" literatures raises an important issue for future study. In some cases social action, such as direct compliance to the demands of an audience (including the experimenter), diminishes the "saying is believing" effect. But in other cases social action produces a strong "saying is believing" effect, such as tailoring a message to suit the audience's characteristics (i.e., to achieve "shared reality"). Thus social tuning must be distinguished from social compliance as forms of social action. The results of these literatures suggest the following intriguing possibility:
In comparison to "personal" action, social action in the form of social compliance reduces the "saying is believing" effect whereas social action in the form of social tuning increases the effect.
Why might social action in the form of social tuning be an especially powerful force in creating meaning? Social tuning not only involves taking others into account but it also involves achieving a common understanding of the world that facilitates interaction with the audience. Establishing shared realities is basic to social regulatory systems. Socialization involves people learning how others respond to the world, including how others respond to them, and using this knowledge in self-regulation (see, for example, Higgins, 1989). As a basic feature of the self-regulatory system, social tuning is charged with motivational significance. This motivational significance may be the reason that social tuning is such a powerful force in creating meaning.
The proposed distinctions between personal and social action and among forms of social action could be tested in future studies. For example, an experiment might have people produce basically the same message on some subject but under different phenomenological conditions, including "personal" (e.g., freely choosing to write an essay that reflects a particular attitude toward the subject of the message [as in standard dissonance studies]), "social compliance" (e.g., experimenter instructions to tailor a written message to suit an audience with a particular attitude toward the subject of the message), and "social tuning" (e.g., freely writing a message to an audience with a particular attitude toward the subject of the message [as in standard communication-game studies]). The written message should have the greatest impact over time in the "social tuning" condition because only in this condition are they produced for the purpose of achieving "shared reality". Different social tuning conditions could also be manipulated. For example, the in-group versus out-group status [in relation to the communicator] of both the audience and the message subject (i.e., the target person) could be varied. This would vary both the extent of the motivation to achieve "shared reality" and the object of this motivation (i.e., to share reality with the immediate audience versus an alternative "inner" audience). In this way, the issue of how communication as social action creates meaning could be examined within the more general framework of social- and self-regulatory systems-- a framework that considers the "ends" as well as the "means" of interpersonal communication.
The implications of communicators' motivation to achieve "shared reality" for other areas of social psychology could also be examined in future studies. In recent years there has been an increasing interest in the effect of communication processes on social-cognitive phenomena. Much of the interest has focused on the implications of Grice's (1975) maxim of relevance (see, for example, Hilton, 1990; Schwarz, Strack, Hilton & Naderer, 1991; Strack, Schwarz, & Wanke, 1991). But communicators' motivation to achieve "shared reality" could also influence social-psychological phenomena. For example, what inference would observers make about a communicator's attitude on a subject when the communicator produces a negative message about the subject to an audience that the communicator knows has a negative attitude toward the subject? If observers understand that communicators are motivated to achieve "shared reality" with their audience, then they should infer that the communicator's real attitude is probably less negative than indicated by the message. Indeed, there is evidence that observers do make this inferential adjustment (see Newtson & Czerlinsky, 1974).
Why, then, do observers not make a similar adjustment in other cases, such as those reflecting the "fundamental attribution error" (Ross, 1977) or "correspondence bias" (Jones, 1979). In the classic example of this phenomenon (Jones & Harris, 1967), observers inferred a direct correspondence between communicators' pro-Castro speech or anti-Castro speech and the communicators' true personal attitude toward Castro even though they knew that the experimenter had assigned the communicators to the position they advocated as part of a debate. What is notable about this case is that the communicators were not communicating to an audience that they believed held a particular attitude toward Castro. And the observers knew this. Thus, the observers would not infer that the message content derived from a motivation to achieve "shared reality" with an immediate audience. Indeed, the observers might infer that the message content, apart from just the position advocated (i.e., "pro" or "con"), was tuned toward some "internal" or "personal" standard of the communicator. This would produce the "correspondence bias". But this explanation differs from the usual explanation of "correspondence bias" in that it does not postulate that observers are insensitive to situational influences on the actor's behavior (in this case, situational influences on the communicator's message). It is just that observers' major focus regarding situational influences is on audience-tuning. This perspective would predict that the "correspondence bias" would disappear if the observers believed that the debaters were trained to use a strategy of imagining that they were communicating to an audience who held the position to be advocated.
Consider just one other example. In Kahneman and Tversky's (1973) classic study, subjects were asked to judge the probability that a target person was either one of 30 engineers in a sample of 100 engineers and lawyers or was one of 70 engineers in a sample of 100 engineers and lawyers. In addition to one or the other kind of baserate information, the subjects were given a written profile of the target person that was supposedly based on information obtained by a panel of psychologists who had interviewed and tested the entire sample of 100 professionals. The study found that subjects underutilized the baserate information and apparently based their judgments on the profile more than they should have from a Bayesian normative perspective. One reason this might have occurred is that the subjects may have been motivated to communicate a judgment that would achieve "shared reality" with the experimenter and whoever designed the study. Because subjects were likely to perceive the study and the experimenter as being "clinical" (see Zukier & Pepitone, 1984), they could accomplish this best by tuning their judgment in the direction of the "clinical" profile.
By the same logic, then, if the subjects perceived the study and the experimenter as being "scientific", they could achieve "shared reality" best by tuning their judgment in the direction of the "scientific" baserate information. Zukier and Pepitone (1984) report that this is exactly what happens when subjects perceive the study as being "scientific" rather than as "clinical". Our communication perspective also suggests that if the study were framed differently so that subjects did not consider the study to involve interpersonal communication, then the subjects would not be motivated to achieve "shared reality" by tuning their judgments toward the "clinical" profile (in the standard paradigm). Schwarz et al. (1991) report findings that are consistent with this prediction as well.
A final comment. It has been over ten years that I first presented the communication game and discussed its implications for social cognition and persuasion. In the following years, I have been especially intrigued with how the social action of audience tuning can create meaning over time. I am now intrigued with how the impact of social action on creating meaning varies for different forms of social action (e.g., super-tuning vs. anti-tuning vs. non-tuning; social tuning vs. social compliance). Of special interest are the implications of communicators being motivated to achieve "shared reality" with their audience. By taking an increasingly "social" perspective on interpersonal communication, perhaps its role in "social" psychology will be better understood.
Table 1
Mean Pre- to Post-Message Change in Reproductions on "Unchanged" Measure1 and "Polarization" Measure as a Function of an Audience with the "Same" or "Different" Target Person Information
_______________________________________________________________
Pre- to Post-Message Change in Reproductions
"Unchanged" measure "Polarization" measure
Audience
Knowledge
________________________________________________________________
Same -9% + 22%
Different +3% -29%
_______________________________________________________________
1. The "unchanged" measure was the number of original target descriptions that were reproduced without deletion or distortion of the basic information.
Table 2
Mean Number of Positive and Negative Message Labels for Ambiguous and Unambiguous Target Descriptions as a Function of Audience Attitude
_______________________________________________________________
Message Labelling of Target Person Information
Ambiguous Target Information Unambiguous Target Information
Positive label Negative label Positive label Negative label
Audience
Attitude
________________________________________________________________
Like 1.3 0.6 1.8 0.4
Dislike 0.8 1.5 1.3 1.8
________________________________________________________________
Table 3
Mean Percentage of Positive and Negative Reproduction Distortions for Ambiguous and Unambiguous Target Descriptions as a Function of Audience Attitude
_______________________________________________________________
Reproduction Distortions of Target Person Information
Ambiguous Target information Unambiguous Target information
Positive Negative Positive Negative
distortions distortions distortions distortions
Audience
Attitude
________________________________________________________________
Like 25% 19% 6% 2%
Dislike 3% 28% 1% 12%
________________________________________________________________
Table 4
Mean Message Distortion as a Function of Authoritarianism, Audience Status, and Audience Attitude
_______________________________________________________________
High Authoritarians Low Authoritarians
Message
Distortion Positive Negative Positive Negative
_______________________________________________________________
Equal Status Audience
Audience Attitude
Like 1.3 0.7 0.9 0.9
Dislike 0.8 1.2 0.7 1.4
High Status Audience
Audience Attitude
Like 1.7 0.5 0.5 1.1
Dislike 0.7 1.3 0.9 1.1
_______________________________________________________________
Table 5
Mean Number of Positive and Negative Message Labels as a Function of Self-Monitoring and Audience Attitude
______________________________________________________________
High Self-Monitors Low Self-Monitors
Positive labels Negative labels Positive labels Negative labels
Audience
Attitude
_______________________________________________________________
Like 3.1 1.7 3.2 1.5
Dislike 1.7 2.0 3.9 2.6
_______________________________________________________________
Figure 1. Path analysis for the "long inter-message delay" condition of the relations among the attitude of the first audience, the communicators' first and second messages about the target person, and the communicators' impressions of and attitudes toward the target person, plus the relation between the attitude of the second audience and the communicators' second message about the target person.
Figure 2. Path analysis for the high authoritarian communicators of the relations among the attitude of the first audience, the communicators' message about the target person, and the communicators' reproductions of, impressions of, and attitudes toward the target person.
Additional References
Clark, H. H., & Brennan, S.E. (1991). Grounding in communication. In L. B. Resnick, J. M. Levine, and S. D. Teasley (Eds.), Perspectives on socially shared cognition (pp. 127-149). Washington, D. C.: American Psychological Association.
Clark, H.H., & Marshall, C.R. (1981). Definite reference and mutual knowledge. In A. K. Joshi, B. L. Webber, and I. A. Sag (Eds.), Elements of discourse understanding (pp. 10-63). Cambridge: Cambridge University Press.