Relational Information Processing and the Implicit Personality Concept: an Updated And

Relational Information Processing and the Implicit Personality Concept: An Updated and Extended Draft

Guido Peeters

Catholic University of Leuven

Center for Social and Cultural Psychology

Technical Report

April 2011

Preliminary Note. The present paper is an extended and updated version of a technical report that appeared in 1991 and of which a reduced version was published in a journal (Peeters, 1991). Until recently neither the original report nor any copies were left. However the content appeared relevant to some recent developments in social cognition research, particularly developments regarding implicit causality in verbs. Hence the paper has been reconstructed and updated using fragments recovered from old floppy disks. The application to implicit causality in verbs will be dealt with in a separate report (forthcoming).

Abstract. A theoretical model is presented dealing with how perceivers form simultaneous impressions of two target persons' (A and B) personalities as a function of how the target persons relate to themselves and to each other. The model implies that perceivers can process relational information following two independent cognitive programs: (a) the SO-program that generates a personalized discourse anchored on the concepts "self" and "other", and (b) the 3P-program that generates a depersonalised discourse anchored on concepts associated with the third pronominal person. Previous research has shown that perceivers are biased to the SO-program. Hence five new experiments were set up in order to test the hypothesis that perceivers may switch to the 3P-program if 3P-related categories are made salient in the stimulus information. The hypothesis was confirmed when participants were required to form impressions about tastes and preferences of the target persons, but the SO-bias persisted in impressions formed of the targets' personalities. The results suggest that the implicit personality concept is semantically connected with the SO-program. The only way to make participants processing information about personality in the way of the depersonalizing 3P-program required the experimenter to explain the 3P-rationale to the participants. However, effects of explaining the 3P-rationale were found to persist over at least one year.

Introduction

Since Asch's (1946) seminal study, there has been a long tradition of social-psychological research on impression formation. The current paradigm requires participants to describe the impressions they have formed of a target or stimulus person on the basis of information provided by the experimenter. For instance, the experimenter offers a brief description telling that stimulus person A is industrious, warm, intelligent, etc., and then asks participants to complete the description using a personality-trait checklist. It has been found that the participants' impressions are underlain by relatively simple implicit personality theories according to which traits are correlated following a limited set of trait-dimensions (Rosenberg & Sedlak, 1972; Abele, Cuddy, Judd, & Yzerbyt, 2008).

One of the criticisms raised with respect to this paradigm was that it overlooked that in everyday life impressions are formed of two or more targets or stimulus persons at the same time, and that these impressions may be interdependent (Peeters, 1967). Hence, an additional impression formation paradigm was proposed that differed from the original one in that the stimulus information provided by the experimenter would not consist of a trait-description of a single stimulus person, but of the description of the interpersonal relationship between two stimulus persons. It was suggested to take interpersonal (dis)liking relations as stimuli because those relations are representative of an important interpersonal relational category. For instance Defares (1963) has demonstrated that (dis)liking relations are representative of so-called personal relations. They also fit Brown's "solidarity relations" (Brown, 1965) as well as a main relational category established by Wish, Deutsch & Kaplan (1976)

Subsequent research provided evidence that interpersonal (dis)liking relations are at least as compelling as trait-stimuli for the induction of impressions of personality but showed also that the bond between relational stimulus and impression could not be dealt with as a simple one-to-one relationship. Rather than to elicit one specific impression, the relational stimulus seemed to set some well-fixed boundaries within which impressions could vary freely (Peeters, 1976).

In search for a theoretical framework suitable to deal with this finding, without an unbridled proliferation of hypothetical ad hoc implicit theories, the Relation Pattern Model (RPM) was designed. In analogy with direct perception theories, it attempted to account for the relationally induced impressions as far as possible by the information that is directly available in the relational stimulus configuration (Peeters, 1983). In order to illustrate this approach we may proceed from the following thought experiment (Peeters, 1986).

Imagine that you observe two pale people (A and B) who suddenly start blushing after they have touched each other accidentally. When asked whether this incident reveals a difference in personality between A and B, you may answer in the negative in that the blushing seems to reveal rather a common personality trait such as "shyness", which means: similarity. However, imagine now that A and B are bottles containing pale liquids and that you are asked whether these liquids differ in chemical composition. In order to test this you pour some of liquid A together with some of liquid B and observe that a chemical reaction takes place changing the initial pale complexion into red. Considering that if the chemical compositions of A and B were alike, the reaction should already have taken place in the bottles without requiring that both are poured together, you conclude that the chemical compositions of A and B must be different.

In either case -- that about the people and that about the bottles -- the same formal relational stimulus configuration is given: two entities, A and B, are said to react upon each other in a way as to change their complexion from pale to red and not to produce this reaction upon themselves when kept in isolation from each other. Nevertheless, this configuration seems to imply similarity in the case of the people, while dissimilarity in the case of the bottles. The RPM shows that, in order to account for this duality, one does not need to call on complex implicit theories regarding personality and chemistry, but that it suffices to analyse the formal patterns of similarities and dissimilarities implied by the relational stimulus configuration. Specifically, (dis)similarities can be derived in at least two ways that can be conceived of as elementary cognitive programs for the processing of relational information. They are referred to as the Self-Other Program (SO-Program) and Third-Person Program (3P-Program) (Peeters, 1983).

According to the SO-Program the entities (A, B, etc.) on which the relations are anchored, are processed as constructs corresponding to respectively the first and the second pronominal person, which means: as 'me and you' or 'self and other'. In this way, the similarity assigned to the people A and B who make blushing each other is derived in that for both holds that they are made to blush by contact with an other while not by contact with the self. According to the 3P-Program, however, the entities are processed as constructs corresponding to the third pronominal person such as: "he, she or it called A, the individual or object labelled B, etc.". In this way the difference between the chemicals in the bottles A and B is derived in that the

one (A) produces a chemical reaction when brought into contact with B and no chemical reaction when in contact with A, while the other (B) shows the opposite pattern, a chemical reaction being produced in contact with A while no reaction in contact with B.

The question arises why the SO-program is used in some cases, such as that about the shy people, while the 3P-Program in other cases, such as that about the bottles with chemicals. There are indications that the content of the inferences (personality versus chemical composition) may be relevant. In an analysis of semiotic implications of the RPM, Peeters (1989) has argued that the SO-Program generates a personalized discourse dealing with each entity -- A, B, etc. -- as a unique "being in itself" of which the most perfect realization is the "person" defined, in agreement with Buber (1970), in terms of "I and Thou". Disregarding the self-other distinction and substituting the impersonal "third pronominal person" for "I and Thou", the 3P-program would generate a depersonalized discourse dealing with entities as mere "sets of attributes" that are apposed, implicitly or explicitly, to an impersonal "he,

she, or it". Hence, in the above thought experiment, the use of the SO-program for making personality inferences may just reflect that the concept of "personality" belongs to a personalized discourse, while the use of the 3P-program for making chemical inferences may reflect that chemistry belongs to a depersonalized discourse1.

Although the above rationale has intuitive appeal, it can be objected that psychologists and lays as well have dealt with the personalities of individuals as "sets of traits" in the same way as chemists deal with chemical objects as "sets of chemical features". The case of the blushing people then may be an exception that is not representative for the way perceivers process information about personality in general. Indeed, it is not difficult to imagine plausible examples of inferences about personality from relational stimuli in a way consistent with the 3P-Program. For instance, consider that A and B feel positively concerning themselves but negatively concerning each other. Analogous to the initial example of the blushing people, A and B relate to each other in a way that is different from the way they relate to themselves. Accordingly, the SO-Program implies similarity, while the 3P-Program dissimilarity between A and B. There seems to be no problem to suggest that, consistent with the 3P-Program, A and B have different personality traits. Indeed, A may like certain traits that he finds in himself but misses in B, and so may B. E.g. A could be a sensitive but slovenly man who likes sensitive people, and B an insensitive but punctual man who likes punctual people and. Finding only himself sensitive, A may feel positively concerning himself but negatively concerning B, and finding only himself punctual, also B may feel positively concerning himself but negatively concerning A.

However plausible the latter rationale may look, experiments failed to provide the slightest evidence that perceivers would use it. Indeed, using an impression-formation paradigm, participants were asked to assign traits to hypothetical individuals A and B. Information about how A and B felt concerning themselves and concerning each other was systematically varied over a number of conditions allowing to determine to which extent traits were assigned in a way consistent with (a) the SO-program, (b) the 3P-program, and (c) the simple common-sense rule that "birds of one feather flock together" implying that the more A and B feel positive concerning each other, the more they share the same traits (liking-similarity rule). The results were overwhelmingly consistent with the SO-program, and to a small but significant extent with the liking-similarity rule, but they failed to show the slightest agreement with the 3P-program (Peeters, 1983). Extensive subsequent research has confirmed that social perceivers are biased to use the SO-program, the 3P-program being limited to particular conditions such as when perceivers draw inferences about impersonal matters such as chemical matters, body sizes, and so forth, rather than about more personalized matters such as personality (for reviews, see: Hendrickx & Peeters, 1997; Peeters, 2004; Peeters & Hendrickx, 2007)

Apparently the example of the blushing people is representative of the way perceivers infer personality traits from relations which is formally different from the way they may derive chemical properties and leads towards a personalized discourse about reality. Hence, the reason why personality inferences involve the SO-program may be a semantic one. The implicit concept of "personality" may be pre-eminently "personalized" by definition and hence be semantically anchored on the notions of "self" and "other" because these notions underlie the personalized discourse. In addition, and in line with the semantic explanation, the results showed that personality inferences mediated by the SO-program were in agreement with the current concept of "implicit personality theory" as it was elaborated by Rosenberg & Sedlak (1972). Specifically, the impressions mediated by the SO-Program involved two

dimensions that reflected Rosenberg's dimensions of intellectual good-bad and social good-bad. The first dimension was represented by traits such as practical, self-confident, etc. versus their opposites, and was found to be associated with (a) liking (versus disliking) the self, (b) being liked (versus disliked) by the other, and (c) behavioral decisions pertaining to the maximization (versus minimization) of the own profit. The second dimension was

represented by traits such as generous, tolerant, etc. versus their opposites and was found to be associated with (a) actively liking versus disliking the other and (b) behavioral decisions pertaining to the maximization (versus minimization) of the profit of the other. In the light of these relational connotations, the former dimension was reinterpreted as self-profitability in

that traits such as practical and self-confident could be considered as profitable for the individual that has those traits rather than for the individual's socius. Analogously, the social good-bad dimension was reinterpreted as other-profitability (Peeters, 1983).

Altogether, the conclusion seems justified that the implicit concept of personality is semantically connected with the SO-Program. Nevertheless, in Peeters (1983) an alternative explanation for the observed dominance of the SO-Program was suggested. Considering that the stimulus information was formulated in very general terms, it was presumed that the concepts self and other might stand out more saliently for being used as cognitive anchors than the vague 3P-constructs he or she labelled A and he or she labelled B. If the

stimulus-information would be specified by additional information such as about A and B's professions, about why and in which respects they like or dislike each other and themselves etc., then A and B may become salient as 3P-constructs and the 3P-Program may be activated without requiring semantic connections between the given 3P-constructs and the content of the inferences that are made. Some indirect support for this explanation for the dominance of

the SO-Program may be provided by the observation that the preference for the SO-Program seems not to be limited to inferences about personality. For instance, it has been found to underlie also inferences about attitudes (Peeters, 1983) and even about apparently impersonal facts such as the outcomes of bargains (Peeters, 1987). Hence, it seems feasible that whichever information would be inferred by the perceivers, they would always use the

SO-Program because of the salience of the concepts self and other as cognitive anchors.

To conclude, we have two possible theories accounting for the differential activation of the SO- and 3P-Program: (a) a semantic theory stressing the role of semantic factors in the differential activation of SO- and 3P-programs, and (b) a construct-salience theory relating the differential activation of SO-and 3P-programs to a difference in salience between the SO-

and 3P-constructs that are present in the stimulus information. However, before proceeding to an experimental test of these theories, we should have a closer look at some formal implications of the RPM.

Some Formal Implications of the RPM

In the examples of the blushing people and the bottles with chemicals, the SO- and

3P- Programs seemed to contradict each other the one resulting in perceived similarity, the other in perceived dissimilarity between A and B. However, this is not necessarily so. Rather the relationship between SO- and 3P-Programs is orthogonal in that each entity that can be dealt with in 3P-related terms such as entity A, entity B, etc. can also be conceived as a self and as an other as well. It follows that the SO- and 3P-Programs are not incompatible, and thus can be applied simultaneously to the same stimulus information. For instance, in our example of the blushing people, it is feasible that applying SO- and 3P-Programs at the same time, a perceiver concludes that there is a certain probability that A and B are two shy people

but that there is also a certain probability that they have different biochemical constitutions that produce the blushing by allergic reaction. There would even be no problem to deal with A and B as two equally shy people (SO-Program) with different biochemical constitutions (3P-Program).

In the latter example, SO- and 3P-Programs were combined in a simple additive way. However, one can think of more complex interactive combinations. For instance, one could consider that blushing in contact with the other does not have the same informational value when the other is A than when the other is B. E.g., imagine that A is a professor and B a student. A perceiver still may interpret A's blushing as a manifestation of shyness in the sense of uncertainty in contact with others, but B's blushing could be interpreted as mere excitement due to the presence of an important other. In the perceiver's rationale, the 3P-Program is implied in that A and B are conceived of as he the student and he the professor. At the same time, however, they are also conceived of as self versus other. For instance, the perceiver does not merely assume that B blushes when in contact with a professor but in addition that this professor should be an "other" person and not the "self". Thus the perceiver would not expect that when B would have become a professor himself he would be blushing all the time because of the permanent presence of a professor (himself).