Contribution
A theoretical chapter that sets close-relationship research and social identity research against each other. Brewer restates her tripartite self (individual, relational, collective), argues that relational and collective bonds serve different adaptive functions and are regulated separately, reviews the evidence that people tell them apart, and extends optimal distinctiveness theory to all three levels. It ends with a prediction about culture: individualism favours the collective self, relational collectivism the relational self.
Key insight
There are two ways to include others in the self, and they are not substitutes. “The relational and collective levels of self … represent two different forms of social identification, i.e., processes by which the individual self is extended to include others as integral to the self-concept” (p. 3). The relational self takes in particular others who are “not interchangeable” (p. 5). The collective self takes in a symbolic group “independent of personal relationships within the group” (p. 3). Humans need both, “belonging with and belonging to” (p. 24).
Assumptions
- Evolutionary functions (pp. 4–5). Close bonds give tailored, privileged mutual aid but are few (about 15 close ties, 150 in a network, citing Dunbar). Symbolic group identities provide “depersonalized” cooperation at scale.
- The self as a set of regulatory systems (pp. 14–15), each with an assessment, a comparator and an activation function (Pickett and Gardner 2005), on the analogy of hunger and thirst.
- Optimal distinctiveness (pp. 12–13): identity needs come in opposed pairs that hold each other in check.
- The evidence is others’: the chapter reports no new data.
Key results
- The tripartite model, restated (pp. 2–3). Brewer and Gardner 1996 replaced the independent/interdependent and personal/social dichotomies with individual, relational and collective self-representations, “not simply three aspects of a single self system, but rather three separate systems with different identity properties, locus of agency, and motivational concerns”.
- People tell the two kinds of belonging apart (pp. 5–7).
- Group attachment (Smith, Murphy and Coats 1999) correlated only .46 (anxiety) and .33 (avoidance) with interpersonal attachment, and uniquely predicted affect and engagement with in-groups.
- The UCLA Loneliness Scale, in more than 2,500 students, has three correlated factors, isolation, relational connectedness and collective connectedness (Hawkley et al. 2005), replicated in older adults.
- Gender (pp. 7–11).
- Women give more relational, men more collective self-descriptions on the Twenty Statements Test, with the same total of interdependent ones (Gabriel and Gardner 1999).
- Men were more likely to choose to “benefit a collective over self-interest”, women “to sacrifice self-interest for the sake of a relationship” (p. 9).
- Primed with “we”, men and women trusted an in-group stranger equally over an out-group one; only women trusted more an out-group stranger with an indirect personal connection (Maddux and Brewer 2005).
- Not substitutes (pp. 11–12). In almost every study relational bonds are more important than collective ones for both sexes; the sexes differ only in collective identity’s weight beyond them.
- Optimal distinctiveness at three levels (pp. 13–15, Table 1): uniqueness against similarity (individual), autonomy against intimacy (relational), separation against inclusion (collective). Satisfying one level can transfer to another, and the levels can conflict.
- Two collectivisms (pp. 17–23).
- Relational collectivism: obligation to specific persons in a network. Group-based collectivism: obligation to the group as a whole, “subordinating individual interests to those of the collective”.
- Because relational obligations are costly, relational collectivists should bound their in-groups tightly. Cited: Chinese and Japanese participants applied equality to friends and equity to strangers more than Americans did.
- Individualists, whose obligations are weak, need large inclusive groups: “individualists are more likely to focus on the collective level of the social self than are relational collectivists” (p. 22). Cited: Americans view their country more favourably than Japanese do; American football fans show more in-group bias than Japanese fans at equal identification.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Relational and collective selves are both extensions of the individual self to include others, one personalized and one depersonalized | moderate as a framework; definitional, with measurement studies showing people distinguish the two | pp. 3–7 |
| C2 | Relational and collective belonging are distinct needs, not substitutable routes to one need | weak to moderate: correlational measurement studies and gender differences; no test of substitution | pp. 5–12 |
| C3 | Each level of the self is regulated by opposed needs whose balance sets the optimal identity | weak: an extension of optimal distinctiveness theory, argued, not tested here | pp. 13–15 |
| C4 | Relational collectivism goes with tight in-group boundaries, individualism with identification with large, inclusive groups | weak to moderate: cross-cultural findings consistent with it, interpreted after the fact | pp. 19–23 |
Concepts
- relational self: the self extended to particular close others and the networks built from those dyads; personalized.
- collective self: the self extended to a symbolic group as a unit; depersonalized.
- optimal distinctiveness: identities chosen to balance a need for inclusion against a need for differentiation.
- relational versus group-based collectivism: obligation to specific persons, or to the group as a whole.
- belonging with and belonging to: personalized and depersonalized interdependence.
Connections
- Brewer and Gardner 1996 (LIT-1757), unread. The chapter’s framework. It is restated, not re-evidenced.
- León’s personalized “we” (THEORY-1115). Brewer gives the same line from psychology. The relational self is a we whose members are not interchangeable, kept apart from the depersonalized collective self. León cites Brewer and Gardner for it (NOTE-1309).
- Self-categorization theory (THEORY-1114). Brewer accepts its account of the collective self but says it relegates relationships to “personal” identity (p. 2). On her model a close relationship is a social identity too.
- Aron et al. (LIT-1706) and Cialdini et al. (LIT-1728): their self–other inclusion is the relational self’s kind of extension.
- Markus and Kitayama (LIT-1703) are among the dichotomies the tripartite model replaces.
- The group entity versus a category (TERM-096). Brewer’s line falls between personalized and depersonalized bonds, not between a whole and a set of like members. Families and teams fall on the relational side.
Bearing on the record
- The owner’s thesis (CLAIM-491). These are inferences; Brewer
does not discuss selflessness.
- Brewer states the thesis’s premise in her own words: social identification is the self “extended to include others as integral to the self-concept” (p. 3). She gives it two forms.
- That splits the thesis. “Collectivism is self-interest on a larger self” reads differently for each collectivism. Group-based collectivism subordinates the individual to the group as a whole; relational collectivism is obligation to particular people. The larger self of the first is a category, of the second a network of individuals.
- Her motivational account fits the thesis. Groups are joined to meet the individual’s needs for inclusion and distinctiveness, so the extension itself is driven by the person’s own needs.
- Her sources still speak of sacrifice: men choosing to “benefit a collective over self-interest”, women to “sacrifice self-interest for the sake of a relationship” (p. 9). Self-interest there means the personal self’s.
- Batson’s objection (CLAIM-494). Empathic helping of a particular person is the relational self’s territory. If the thesis needs an extended self there, it is the relational self, and the oneness measures (Aron’s circles) are its measures. Brewer adds nothing on empathy.
- The triviality objection (CLAIM-492). Brewer’s criteria for the self’s level are self-descriptions (the Twenty Statements Test) and importance ratings (the RIC scale asks what “is most important in life”, p. 9). The first is a criterion apart from caring; the second is caring. The chapter does not test whether they come apart.
- A dated note should be added to CLAIM-491. Proposed text: “Brewer (NOTE-1290) calls both close relationships and group memberships ways the individual self is ‘extended to include others’, but they are different extensions: the relational self takes in particular, non-interchangeable others; the collective self, a symbolic group. She distinguishes relational collectivism, obligation to specific persons, from group-based collectivism, which subordinates the individual to the group as a whole. The thesis needs to say which larger self it means, or that it means both.”
- No THEORY filed.
- No ML instruction.
Limitations
- A draft. The published chapter (2008, pp. 179–196) may differ; it was not seen. The draft’s reference list is incomplete (Brewer and Chen 2007 has no volume or pages).
- No new data; the evidence is cited, often in a sentence.
- Not the 1996 paper. Its priming studies and its levels’ bases of self-worth and frames of reference are not reported.
- Typos.
- “mean produced” for men (p. 8).
- “would defer” for differ (p. 10).
- “different dues” for cues (p. 15).
- “phylogentically and ontogentically” (p. 12); “indivdiual” (p. 9); “Seeley at al.” (p. 12).
- “Smith, Murphy, and Coates” (p. 5) against “Coats” in the references.
- “Cross & Madsen” (p. 3) against “Madson” elsewhere; Cross, Bacon and Morris’s first author is “S. L.” in the references, “S. E.” in the next entry.
- The 1996 title is given as “Who is the ‘we’?” in the references (p. 25).
- “Oyersman”, “Evaluaiton”, “consrual”, “struess”, “cogntion”, “collectiver”, “Press Press” in the references.
Open questions
- Do the relational and collective selves predict different kinds of helping, toward a particular person and toward the group?
- If individualists identify more readily with large groups, do they also help strangers in those groups more than relational collectivists do?