Swiss Journal of Psychology 63 (1), 2004, 43–51

Original communication

Psychosociological anchoring and structural dynamics in social representations of the heterosexual/homosexual couple Patrick Rateau Université Paul Valéry-Montpellier III, France

This article reports an experimental study on structural variations in a representational field as a function of social regulations activated in symbolic intergroup relations. Two centrality questionnaires were used (refutation technique), one on the representation of the heterosexual couple and one on the representation of the homosexual couple. These questionnaires were administered in the opposite order to two groups of 80 heterosexual men. When there was no intergroup comparison, the participants exhibited the same representation of the two types of couple. In the implicit comparison situation they differentiated the two categories, both by reorganizing the central-system hierarchy of their heterosexual-couple representation, and by over-activating its peripheral elements. Datas are analyzed in terms of the links between the structural dynamics of social representations and the identity dynamics upon which social relations are based. Keywords: Social representation, psychosociological anchoring, categorical differentiation

In his initial work, Moscovici (1961) stressed the role of two major processes in the building of social representations: objectifying, which is a simplifying materialization of a represented object, and anchoring, which is the integration of that object into a network of preexisting knowledge and a set of social and symbolic relations. Later, Doise (1992) distinguished three kinds of anchoring. In psychological anchoring, representations are incorporated into the general beliefs to which individuals adhere to varying degrees. In this framework, interindividual representational variations depend for a large part on the extent of adhesion to the beliefs. Doise and Papastamou (1987) showed, for example, that representations of the causes of delinquency depended on how strongly people adhered to DOI 10.1024/1421-0185.63.1.43

general explanations of the “biologistic” and “hereditarianistic” type. In sociological anchoring, representations are group-based, in accordance with the place the various groups occupy in the set of all social relations. This anchoring mode, which explicitly posits social positions as the essential determinant of representations, follows directly from Bourdieu’s (1977) theory of social fields and its underlying principle of structural homology. A study by Viaux (2000), for instance, showed that individuals build representations of the economy in accordance with whether they occupy a dominating or dominated position in the field of production relations, and that a move from a dominated position to a dominating one is accompanied by the gradual adoption of a representation that fits with Swiss J Psychol 63 (1), 2004, © Verlag Hans Huber, Bern

44

P. Rateau: Social representations and categorical differentation

the new status. Finally, in psychosociological anchoring, representations are embedded in intergroup relations. In this case, the observed representations vary not only with group specificities but also with “positional and categorical divisions specific to a given social field” (Doise, 1992, p. 191). A good example of this is found in Lorenzi-Cioldi’s (1988) work on the effects of between-gender relations on the organization of representations. This last type of anchoring is closely tied to the processes of categorization and social differentiation, a fact which has permitted the application of paradigms developed in these research trends to the experimental study of how the cognitive processes triggered by social regulations function within representations. Such studies have pointed out tight interrelationships between representational and identity-related processes. More specifically, they have shown that representations activate differentiation processes which are regulated by the social context in which the concerned individuals are situated, especially if that context threatens the social identity of the groups in question. Accordingly, when individuals are deluded into believing that an ideologically antagonistic group shares a given representation with them – thereby provoking an identification conflict; see Mugny, Kaiser, and Papastamou, 1983; Mugny and Perez, 1991 – a circumstantial modification of that representation ensues, motivated by the desire for intergroup differentiation (Rateau, 1999). In the same vein, Tafani and Bellon (2001) found that manipulating the evolution of positional asymmetries (strengthening vs. weakening of the dominating vs. dominated group status) leads individuals to change the way they rank the core elements of the representational field. These studies highlight the fundamental role played by categorical differentiation in representational dynamics. More specifically, they show that the structure of a given representational field varies with social regulations rooted in the identitybased positions brought to bear in intergroup relations within that field. Note that these studies are fully in line with recent approaches based on the core theory of social representations (see Abric, 1994). Recall that according to this theory, social representations revolve around two complementary systems: a central system composed of stable cognitions that resist change and guarantee group identity, and a peripheral system composed of conditional and ever-changing elements by means of which representations are adapted to circumstantial constraints and fluctuations in the social context. The adaptive function of the peripheral system allows it to act in defense of a representation whenever it is challenged by the social situation (Flament, 1989). Rateau’s (1999) results mentioned above thus imply that an identity threat triggered by an identification conflict will produce different effects, depending Swiss J Psychol 63 (1), 2004, © Verlag Hans Huber, Bern

on whether the conflict is based on a central or peripheral element of the representation: the will to maintain intergroup differentiation only causes a modulation at the peripheral level. In other words, for the purposes of categorical differentiation, individuals agree to incorporate contradictory information into their representational system only when that information concerns a peripheral cognition. This process prevents any doubt from being cast on their central system, the “guardian” of their identity. The process brought to the fore by Tafani and Bellon (2001) is different, but equally consistent with core theory. They found that individuals faced with a threat to their identity recompose the central system of their representation, not by deleting or adding new cognitions, nor by modifying existing cognitions, but by adjusting the weight assigned to the central dimensions of the representation. In this way, whenever the status of a dominating group is questioned, or the status of a dominated group is accentuated, the identity threat that is generated regulates the hierarchical relations among the central elements of their respective representations: “primary” central cognitions tend to become “secondary”, and vice versa (Rateau, 1995; Abric & Tafani, 1995). Using substitution instructions (where individuals fill in a questionnaire “on their own behalf”, “on behalf of ingroup members”, and “on behalf of outgroup members”), these authors were also able to fully account for their results in terms of categorical differentiation, particularly in terms of Deschamps’s (1982) covariance model: when the dominating group’s status is questioned, that group tends to adopt a differentiation strategy identical to that observed in dominated groups, namely, recourse to outgroup discrimination only, without ingroup differentiation. In sum, these two lines of research have established tight links between the structural functioning of social representations and the identity dynamics activated by the social relations in which individuals are engaged. They stress the merits of studying how categorical differentiation processes affect representational dynamics, and in particular, how these processes show up in the structure of the representational field. The present study attempts to provide insight into this general issue.

Issue and Hypotheses We will focus here on the social representation of “the couple” in a population of heterosexual men. This social object is highly polymorphous, in addition to being strongly linked to identity. It also offers the advantage of having two categorical forms (the heterosexual couple and the homosexual couple) that can be ranked by numerical ma-

P. Rateau: Social representations and categorical differentation

jority/minority. From the numerical standpoint alone, the heterosexual couple is by far the dominating one in French society. One can assume, moreover, that for heterosexual individuals, the representation of the couple is entirely mediated by the heterosexuality of the partners. In other words, references to the couple in general implicitly refer to two heterosexual partners, the different genders of the spouses somehow constituting the dominant norm or prototype of the couple. The instituted relationship between these two types of couple can thus be seen as one of domination, with the heterosexual couple being the accredited, dominant reference within the social body. As such, this object seems to lend itself quite nicely to the study of psychosociological anchoring and its underlying categorical differentiation processes. The idea, then, is to start from the same population, heterosexual men living with a woman for more than five years, and compare the representation of the homosexual couple to the representation of the heterosexual couple, under two conditions: (1) when no comparisons are made between these two types of couple, and (2) when the two types are implicitly compared. In the light of classic categorical differentiation paradigms, two main hypotheses can be set forth. Hypothesis 1. In the absence of an explicit comparison, there should be no major differences between the representation of the heterosexual couple and that of the homosexual couple. Several empirical findings support this hypothesis. Firstly, categorical differentiation is never as strong when the intergroup comparison is not mentioned as when it is (Doise, Deschamps, & Meyer, 1978; Deschamps & Personnaz, 1979). Secondly, crossed-category memberships are at stake here (i.e., the object of study pertains both to the “couple” category and to the “heterosexual” or “homosexual” category). Several studies have found evidence of very little between-category differentiation when the categories are crossed rather than presented singly (Deschamps, 1977; Deschamps & Doise, 1978). A third finding, clearly demonstrated in Lorenzi-Cioldi’s (1988) work, is that intergroup differentiation is more prevalent in dominated groups than in dominating ones. Based on these findings, one can assume that individuals representing the dominating group (heterosexual men) will exhibit little between-category differentiation, especially when the groups are not compared. Hypothesis 2. In cases where the two kinds of couple are compared, one can expect to find significant differences between the two representations. According to classic findings in the categorical-differentiation research, characteristics specific to each category should be accentuated and traits perceived as shared should be played down. In

45

line with core theory and the findings mentioned above, one can expect the intercategorical differentiation process to mainly key on the representation’s peripheral elements. In other words, it is assumed that the peripheral characteristics of the representation of the heterosexual couple will mainly be accentuated via an over-activation process (Flament, 1994; Guimelli, 1996; Rateau, 1999) aimed at protecting the individuals from potential identification with the outgroup.

Method, Materials, and Participants A series of semi-directed interviews conducted on a sample of 20 heterosexual men living with a woman for more than five years pointed out 17 main topics mentioned by participants when talking about “the couple”. These topics became the 17 items of a centrality questionnaire (Moliner, 1996). Recall that this type of questionnaire – a direct outcome of the core theory of representations – is based on a refutation principle: any object exhibiting a feature that negates a central element of a representation cannot be acknowledged as belonging to the field of that representation. It follows that if such an object is presented to individuals in a given group, they will unanimously (100% in theory) refuse or “refute” (Moliner, 1994) that object as belonging to the representation. In line with this procedure, an element will be regarded as central when it gives rise to a refutation rate close to 100%. The questionnaire used here challenged each of the 17 characteristics retained for the study. For example, “Would you say that two persons form a couple if they do not live together?” The participants answered each question on a fourpoint scale: 1 (Yes), 2 (I guess so), 3 (I guess not), and 4 (No). So, for the above question, if nearly 100% of the participants answer “No” or “I guess not”, the fact of “living together” will be included among the central characteristics of their representation of the couple. Two versions of the questionnaire that differed solely by the introductory sentence were generated. In the heterosexual version, the introductory question was “Would you say that two heterosexual persons form a couple if …”. In the homosexual version, the introductory question was “Would you say that two homosexual persons form a couple if …”. After the introductory question, both versions presented the 17 challenged characteristics in the same order (alphabetical order of the first word of the item). The two versions were presented in succession to all participants. Half of the participants had to answer the “hetero” questionnaire first and then the “homo” questionnaire (hetero1/homo2 group) and the other half anSwiss J Psychol 63 (1), 2004, © Verlag Hans Huber, Bern

46

P. Rateau: Social representations and categorical differentation

swered in the opposite order (homo1/hetero2 group). Of course, when the participants were working on the first questionnaire, they did not know they would have to use it again on another target. Because of this, during the first phase, no comparison between heterosexual couples and homosexual couples was explicitly evoked. On the other hand, the second phase induced an implicit comparison between the two kinds of couple. The population tested consisted of 160 men recruited individually in the waiting room of a walk-in garage in a large French city. All were between the ages of 25 and 65 (median age: 35) and had no higher education. They all said they were living with a woman (either their wife, or as man and wife) for more than five years. Eighty were given the questionnaire about the heterosexual couple first (hetero1/homo2) and the other eighty were given the questionnaire about the homosexual couple first (homo1/hetero2).

Results The results will be presented from a descriptive standpoint first in order to bring out the central/peripheral structure of the representations studied. Statistical comparisons will then be made using MANOVAs with one, two-level independent variable (questionnaire order: hetero1/homo2 or homo1/hetero2) and one, two-level repeated-measurement variable (elicited representation: heterosexual couple or homosexual couple).

Centrality Scores The first analysis pertained to the number of negative answers (points 3 and 4 on the scale) obtained for each group. Again, in centrality tests, a characteristic is considered central if it gives rise to a refutation rate close to 100% when challenged. This can be determined using the Kolmogorov-Smirnov test (see Moliner, Rateau, & CohenScali, 2002). Given the number of participants per experimental condition (n = 80), a refutation frequency could be regarded as close to 1.00 (at a significance level of .05) if it was .847 or more (see Kanji, 1993). Table 1 gives the refutation frequencies observed in each group. We can see that four elements turned out to be central in all experimental conditions: Sharing joys and sorrows, Accepting each other as we are, Loving each other and Trusting each other. One can assume that these four elements constitute the common base of the representation of the couple in general, whether homosexual or heterosexual. Note also, in line with core theory, that the representations of the heterosexual and homosexual couples were identical when expressed in first position (hetero1 and homo1), insofar as they were organized around the same central system. This initial result supports Hypothesis 1, since individuals expressed the same representation of the two kinds of couple when an explicit comparison was lacking. Only one condition stood apart from the other three: the case where the participants had to express themselves about the heterosexual couple after having expressed themselves about the homosexual couple (hetero2). Here,

Table 1 Refutation frequency of the heterosexual-couple and homosexual-couple representations, by challenged characteristic.

Seeking stability Establishing a family Making compromises Having the same views on life Sharing the day Sharing joys and sorrows Sharing material possessions Accepting each other as we are Loving each other Being married Building a serious relationship Wanting to have children Living together Trusting each other Making plans together Being faithful Having a sexual relationship

Hetero1

Homo1

Hetero2

Homo2

.787 .537 .837 .487 .670 .937* .526 .887* .862* .175 .787 .437 .687 .962* .675 .787 .750

.837 .562 .825 .787 .775 .912* .662 .850* .862* .500 .775 .375 .675 .900* .775 .700 .787

.837 .850* .837 .700 .700 .912* .862* .937* .887* .437 .875* .625 .737 .937* .775 .837 .837

.762 .350 .775 .587 .625 .850* .500 .850* .887* .237 .700 .300 .600 .900* .662 .775 .800

Note. * Frequencies non-different from 1.00 (p < .05) according to the Kolmogorov-Smirnov Test Swiss J Psychol 63 (1), 2004, © Verlag Hans Huber, Bern

P. Rateau: Social representations and categorical differentation

we find three new central elements: Establishing a family, Sharing material possessions, and Building a serious relationship. In line with Hypothesis 2, when participants were in an implicit comparison situation, they carried out a differentiation process, reflected here by the reinforcement of certain characteristics of their own group. Note, however, that this only happened when the outgroup had to be described first (homo1/hetero2), since in the opposite case, there was no structural difference between the two representations (hetero1/homo2).

Analytic Comparisons Multivariate analysis of variance on the data both confirmed and refined the above results. The table in the Appendix gives the means used in each of the analyses presented below. Representations of the heterosexual couple and the homosexual couple before (hetero1 vs. homo1) and after (hetero2 vs. homo2) social comparison These results amply confirmed the frequency analyses. Contrasting the before-comparison conditions (hetero1 vs. homo1), we find that only 2 of the 17 challenged characteristics, Having the same views on life and Being married, produced significantly more refutations for the homosexual couple than for the heterosexual couple (F(1, 158) = 19.69, p < .01, and F(1, 158) = 20.24, p < .01, respectively). Looking at the structure of these two representations, we can see that the differences are on elements designated as peripheral. It can be concluded, then, that the two representations are highly similar when no explicit comparison is made between the two kinds of couple. However, one can wonder about the two characteristics where the differences were observed: although it seems logical that participants might feel it is more important for a homosexual couple to share views on life due to their same social “peculiarity”, it might seem surprising that half of them considered it important for the couple to be married. A possible interpretation would be that individuals perceive marriage as an institutional union which guarantees that the two partners will form a “serious” couple. In other words, a married homosexual couple may be seen as more legitimate and less “deviant” than a couple without marital ties. In contrast, the refutation scores after the social comparison of the two types of couple (hetero2 vs. homo2) revealed much greater differences on 9 of the 17 challenged characteristics. The F-value (1, 158) was significant for Establishing a family (F = 52.67, p < .01), Having the same views on life (F = 3.65, p < .05), Sharing the day (F = 22.54, p < .01), Sharing material possessions (F = 15.43, p < .01), Accepting each other as we are (F = 4.68, p <

47

.05), Being married (F = 7.74, p < .01), Building a serious relationship (F = 10.37, p < .01), Wanting to have children (F = 19.13, p < .01), and Living together (F = 4.25, p < .05). This analysis clearly illustrates a psychosociological anchoring effect. When the two kinds of couple were not compared to each other, the representation activated was that of the couple in general and was applicable to heterosexual and homosexual couples alike. The only differences were peripheral. On the other hand, the more explicit anchoring of the items in an intergroup comparison process brought out 9 characteristics that qualified the heterosexual couple. The differentiation strategy the participants employed showed up even more decisively in the interactions between the order factor and the repeatedmeasurement factor. Interactions For the 17 items pooled, the interaction was highly significant (Rao’s R (17, 140) = 4.65, p < .01). Taken separately, the individual interactions are quite illustrative of the overall differentiation strategy implemented by the participants (see Table 2). Table 2 separates the elements into three types: 1. Non-interacting, comparison-sensitive elements These elements were the ones designated as central in all four experimental conditions, or as peripheral in all four conditions. This means that they do not differentiate the two kinds of couple but are part of the conception of the couple in general, regardless of the partners’gender. Most of these elements have to do with affection and intimacy. We can see, however, that one of these characteristics stands out against the others: Building a serious relationship. Despite its statistical insensitivity to the order variable, this element took on the status of central element in the hetero2 condition (see Table 1). By virtue of its high refutation rate in the hetero1 condition (.787), one can assume that it is a secondary central element of the heterosexual-couple representation, i.e., an element which, although part of the central system of the representation, happens to be sensitive to certain context effects (Rateau, 1995; Abric & Tafani, 1995). 2. Interacting, comparison-insensitive elements These four elements, Seeking stability (F = 3.91, p < .05), Sharing the day (F = 19.05, p < .01), Living together (F = 3.89, p < .05), and Making plans together (F = 3.87, p < .05), all gave rise to a significant interaction pointing in the same direction, namely, the characteristic tended to be overrated in the representation of the heterosexual couple when it was compared to the homosexual couple (hetero1 vs. hetero2), at the same time as it was underrated in the Swiss J Psychol 63 (1), 2004, © Verlag Hans Huber, Bern

48

P. Rateau: Social representations and categorical differentation

Table 2 Comparison of the means of each representation, by questionnaire order and interaction significance Hetero1

Hetero2

Homo1

Homo2

Interaction

Sharing joys and sorrows Accepting each other as we are Loving each other Trusting each other Being faithful Having a sexual relationship Making compromises Building a serious relationship

3.66 3.51 3.55 3.76 3.25 3.25 3.33 3.25

3.66 3.72 3.61 3.71 3.43 3.33 3.52 3.51

3.56 3.40 3.51 3.57 3.05 3.31 3.41 3.18

3.50 3.42 3.63 3.68 3.23 3.37 3.25 3.00

NS NS NS NS NS NS NS NS

Seeking stability Sharing the day Living together Making plans together

3.21 2.96 2.96 2.97

3.38 3.01 3.16 3.15

3.42 3.17 3.03 3.22

3.12 2.86 2.76 2.92

* ** * *

Wanting to have children Being married Having the same views on life Establishing a family Sharing material possessions

2.30 1.65 2.52 2.56 2.60

< 2.80** < 2.35** < 3.05** < 3.33** < 3.51**

2.27 2.51 3.30 2.67 2.93

> 1.96** > 1.80** > 2.67** > 2.03** > 2.55**

** ** ** ** **

Note. * p < .05, ** p < .01. Central elements are shown in boldface

representation of the homosexual couple when it was compared to the heterosexual couple (homo1 vs. homo2). These characteristics thus typify the heterosexual couple relative to the homosexual couple. We are clearly dealing here with a categorical differentiation process: participants accentuated ingroup specificity by overrating traits perceived as more typical of their own group, and as a corollary, by underrating those same traits in the outgroup. These elements – all peripheral and generally pertaining to the partners’commitment to sharing their daily life – thus appear to be differentiators of the two kinds of couple, although to a relatively weak extent since, while triggering the hypothesized interaction, they remained insensitive to the order effect. They account for the solely-peripheral modulation of the representation under the impact of a categorical differentiation process, but they do not affect its general structure. 3. Interacting, comparison-sensitive elements In addition to giving rise to a significant interaction, the five elements in this category, Wanting to have children (F = 8.89, p < .05), Being married (F = 26.09, p < .01), Having the same views on life (F = 19.05, p < .01), Establishing a family (F = 24.04, p < .01), and Sharing material possessions (F = 8.98, p < .01), also turned out to be significantly sensitive to the presentation order of either or both kinds of couple. These five items were significantly overrated in the hetero2 condition relative to hetero1, and significantly underrated in the homo2 condition relative to homo1. We can see that these characteristics, Swiss J Psychol 63 (1), 2004, © Verlag Hans Huber, Bern

which proved to be very good differentiators of the two kinds of couple, revolve around two major themes: the family (Establishing a family and Wanting to have children) and a formal commitment to a sharing relationship (Being married, Sharing material possessions, Having the same views on life). In the social comparison process, then, these characteristics were very strong markers of the specificity of the heterosexual couple with respect to the homosexual couple. Two of them even moved from the peripheral status in condition hetero1, to the central status in hetero2 (Establishing a family and Sharing material possessions; see Table 1). What has happened here is that, under the impact of a categorical differentiation process, the representation of the heterosexual couple was recomposed via the inclusion of two peripheral elements in the central system.

Discussion To summarize our observations here, when the initial assessment pertained to the homosexual couple (homo1), the subjects exhibited a representation of the heterosexual couple (hetero2) that was more “hetero-specific” than when it was rated first without reference to the homosexual couple (hetero1), i.e., hetero2 > hetero1. By contrast, if the initial assessment pertained to the representation of the heterosexual couple (hetero1), the subjects exhibited a representation of the homosexual couple (homo2) that

P. Rateau: Social representations and categorical differentation

was less “homo-specific” than the one expressed on the first rating, i.e., homo2 < homo1. These findings shed light on how the structural dynamics of a representational field can be linked to the identity dynamics that underlie the social relations in which individuals are engaged. More specifically, they indicate that the structure of this representational field varies as a function of the social regulations activated in the symbolic intergroup relations that occur within that field. This variation showed up in two ways here, both of which have already been observed in earlier research: (1) as a modification in the hierarchy of the central elements of the representation (see Tafani & Bellon, 2001) and (2) as the over-activation of certain peripheral elements (Rateau, 1999). It may seem surprising that this differentiation strategy took effect only when the individuals were implicitly involved in a social comparison process. At the onset, the participants seemed to have the same representation of the heterosexual couple and the homosexual couple. This lack of differentiation can be partly explained by the fact that the categories were crossed: without an intergroup comparison, participants may be more likely to express their representation of the couple in general, irrespective of the partners’gender. In a situation of implicit comparison, on the other hand, they may have focused more on the type of couple presented and revealed their representations specific to that type as they attempted to differentiate the two. While this hypothesis is consistent with studies on categorical differentiation in crossed-category memberships (see above), it does not account for why individuals who were initially asked to talk about the heterosexual couple produced a representation that was identical to the one produced afterwards when they were asked to talk about the homosexual couple. Another hypothesis based on a more recent study by Guimelli and Deschamps (2000; also see Abric, 2003) can be forwarded. According to these authors, social desirability effects lead individuals to fail to reveal or to mask certain areas of their representation field. These so-called “mute” areas are defined by these authors as specific subsets of cognitions which, while remaining available, are not expressed under normal production conditions. All it takes is a change of context for individuals to spontaneously express the corresponding subset. Of course, one can assume that this process will mainly be manifested for objects whose representational field contains cognitions that might cast doubt on accredited social values or norms if expressed in certain situations. In the case of interest here, it is quite possible that participants who had to reveal their representation of the homosexual couple in first place (homo1) adopted a non-differentiating strategy to avoid being subjected to a negative evaluation triggered

49

by discriminating or even stigmatizing responses about homosexuals. When in a social comparison situation, however, these same participants may have felt forced to go against this norm of desirability in order to preserve the identity specific to their group. Thus, for their own category, they overrated the cognitions deemed to best differentiate the two groups: the possibility of establishing a family and sharing possessions, which in French law, are specific to legal marriages between spouses. Inversely, participants who had the opportunity to express themselves about their own group first may have been less inclined to later discriminate against the outgroup, for the purposes of avoiding a negative evaluation likely to follow discriminatory behavior. Although the results obtained here do not allow us to confirm this interpretation, they lay the groundwork for exploring new hypotheses about the mechanisms of categorical differentiation in the structural dynamics of social representation.

References Abric, J. C. (1994). Les représentations sociales: aspects théoriques [Social representations: Theoretical aspects]. In J. C. Abric (Ed.), Pratiques sociales et représentations. Paris: Presses Universitaires de France. Abric, J. C. (2003). La recherche du noyau central et de la zone muette des représentations sociales [The search for the central core and the mute zone of social representations]. In J. C. Abric (Ed.), Méthodes d’étude des représentations sociales. Ramonville Saint Agne, France: Erès. Abric, J. C., & Tafani, E. (1995). Nature et fonctionnement du système central d’une représentation sociale: la représentation de l’entreprise [Nature and functioning of the central system of social representations: the representation of a firm]. Les Cahiers Internationaux de Psychologie Sociale, 28, 22–31. Bourdieu, P. (1977). La production des croyances: contribution à une économie des biens symboliques [The production of beliefs: Contribution to an economy of symbolic possessions]. Actes de la Recherche en Sciences Sociale, 13, 3–43. Deschamps, J. C. (1977). Effect of crossing category memberships on quantitative judgements. European Journal of Social Psychology, 7, 517–521. Deschamps, J. C. (1982). Différenciation entre soi et autrui et entre groupes: recherches sur la “covariation” entre les différenciations interindividuelles et intergroupes [Differentiation between self and others, and between groups: Researches on the covariation of differentiations between individuals and between groups]. In J. P. Codol & J. P. Leyens (Eds.), Cognitive analysis of social behavior. The Hague, Netherlands: Martinus Nijhoff. Swiss J Psychol 63 (1), 2004, © Verlag Hans Huber, Bern

50

P. Rateau: Social representations and categorical differentation

Deschamps, J. C., & Doise, W. (1978). Crossed category memberships in intergroup relations and categorical differentiation. In H. Tajfel (Ed.), Differentiation between social groups. London: Academic Press. Deschamps, J. C., & Personnaz, B. (1979). Etudes entre groupes “dominants” et “dominés”: importance de la présence du hors groupe dans les discriminations évaluatives et comportementales [Studies on “dominant” and “dominated” groups: The importance of the presence of the out group in evaluative and behavioral discriminations]. Information sur les sciences sociales, 18, 269–305. Doise, W. (1992). L’ancrage dans les études sur les représentations sociales [Anchoring in studies on social representations]. Bulletin de Psychologie, 405, 189–195. Doise, W., & Papastamou, S. (1987). Représentations sociales des causes de la délinquance: croyances générales et cas concrets [Social representations of reasons for delinquency: general beliefs and concrete situations]. Déviance et société, 11, 153–162. Doise, W., Deschamps, J.C., & Meyer, G. (1978). The accentuation of intra-category similarities. In H. Tajfel (Ed.), Differentiation between social groups. London: Academic Press. Flament, C. (1989). Structure et dynamique des représentations sociales [Structure and dynamic of social representations]. In D. Jodelet (Ed.), Les représentations sociales. Paris: Presses Universitaires de France. Flament, C. (1994). Aspects périphériques des représentations sociales [Peripheral aspects of social representations]. In C. Guimelli (Ed.), Structures et transformations des représentations sociales. Neuchâtel, Switzerland: Delachaux et Niestlé. Guimelli, C. (1996). Valence et structure des représentations sociales [Valency and structure of social representations]. Bulletin de Psychologie, 422, 58–72. Guimelli, C. & Deschamps, J. C. (2000). Effets de contexte sur la production d’associations verbales. Le cas des représentations sociales des Gitans [Context effects in the production of verbal associations: the case of the social representation of the Gypsies]. Les Cahiers Internationaux de Psychologie Sociale, 47–48, 44–54. Kanji, G. K. (1993). 100 statistical tests. London: Sage. Lorenzi-Cioldi, F. (1988). Individus dominants et groupes dominés [Dominant individuals and dominated groups]. Grenoble, France: Presses Universitaires. Moliner, P. (1994). Les méthodes de repérage et d’identification du noyau des représentations sociales [Methods to locate and to identify the central core of social representations]. In C. Guimelli (Ed.), Structures et transformations des représentations sociales. Neuchâtel, Switzerland: Delachaux et Niestlé.

Swiss J Psychol 63 (1), 2004, © Verlag Hans Huber, Bern

Moliner, P. (1996). Images et représentations sociales [Images and social representations]. Grenoble, France: Presses Universitaires. Moliner, P., Rateau, P., & Cohen-Scali, V. (2002). Les représentations sociales. Pratique des études de terrain [Social representations. Practice of field studies]. Rennes, France: Presses Universitaires. Moscovici, S. (1961). La psychanalyse, son image et son public [Psychoanalysis, its image, and its audience]. Paris: Presses Universitaires de France (reprinted edition, 1976). Mugny, G., Kaiser, C., & Papastamou, S. (1983). Influence minoritaire, identification et relations entre groupes: études expérimentales autour d’une votation [Influence of minorities, identification and relations between groups: Experimental studies in the context of a vote]. Cahiers de Psychologie Sociale, 19, 1–30. Mugny, G., & Pérez, J. A. (1991). The social psychology of minority influence. Cambridge: University Press. Rateau, P. (1995). Le noyau central des représentations sociales comme système hiérarchisé. Une étude sur la représentation du groupe [The central core of social representations as an ordered system. A study about the representation of an group]. Les Cahiers Internationaux de Psychologie Sociale, 26, 29–52. Rateau, P. (1999). Les effets d’un conflit d’identification idéologique sur la structure d’une représentation sociale [The effects of an ideological identification conflict upon the structure of a social representation]. Les Cahiers Internationaux de Psychologie Sociale, 42, 90–101. Tafani, E., & Bellon, S. (2001). Principe d’homologie structurale et dynamique représentationnelle [The principle of structural homology and the representational dynamic]. In P. Moliner (Ed.), La dynamique des représentations sociales. Grenoble, France: Presses Universitaires. Viaux, J. (2000). Principes organisateurs et représentations sociales de l’économie: genèse et dynamique [Organizing principles and social representations of economics: Genesis and dynamic]. Revue Internationale de Psychologie Sociale, 12, 79–105.

Patrick Rateau Laboratoire de Psychologie Sociale Université Paul Valéry-Montpellier III Route de Mende FR-34199 Montpellier Cedex 5 E-mail: [email protected]

P. Rateau: Social representations and categorical differentation

51

Appendix Mean refutation scores in the four experimental conditions Seeking stability Establishing a family Making compromises Having the same views on life Sharing the day Sharing joys and sorrows Sharing material possessions Accepting each other as we are Loving each other Being married Building a serious relationship Wanting to have children Living together Trusting each other Making plans together Being faithful Having a sexual relationship

Hetero1

Homo1

Hetero2

Homo2

3.21 2.56 3.33 2.52 2.96 3.66 2.60 3.51 3.55 1.65 3.25 2.30 2.96 3.76 2.97 3.25 3.25

3.42 2.67 3.41 3.30 3.17 3.56 2.93 3.40 3.51 2.51 3.18 2.27 3.03 3.57 3.22 3.05 3.31

3.38 3.33 3.52 3.05 3.01 3.66 3.51 3.72 3.61 2.35 3.51 2.80 3.16 3.71 3.15 3.43 3.33

3.12 2.03 3.25 2.67 2.86 3.50 2.55 3.42 3.63 1.80 3.00 1.96 2.76 3.68 2.92 3.23 3.37

Swiss J Psychol 63 (1), 2004, © Verlag Hans Huber, Bern