NOTE-1277: Empathy, Altruism, and Group Identification

Contribution

It gives the self-other merging hypothesis (SMH) a reading that survives May’s conceptual objection. Merging is group identification: forming a self-conception as a member of a group that includes the other. It then draws the consequence for the egoism–altruism dichotomy. Empathy-induced helping, so understood, fits neither side. The paper proposes splitting the distinction into an individual level and a group level.

Key insight

Concern for another’s welfare can be non-instrumental and still be concern for one’s own, when the other is part of a “we” one belongs to. Egoistic helping to relieve one’s own distress uses the other’s welfare as a means: my welfare is causally affected by hers but not made of it. Group-identified helping is the reverse: “Reynie’s first-person plural welfare is constituted by, but is not causally influenced by, Elaine’s third-person singular welfare” (p. 9). Caring for a whole includes caring for its parts for their own sake. So the act is altruistic toward Elaine and egoistic toward “us”.

Assumptions

• Batson’s definitions (p. 2). Altruistic: motivated “with the ultimate goal of increasing another’s welfare” (Batson 2018, p. 22). Empathic concern: “other oriented emotion elicited by and congruent with the perceived welfare of someone in need” (2018, p. 29).
• May’s two conditions (p. 4). An acceptable SMH must be behaviourally adequate, predicting and explaining the helping observed, and psychologically plausible, positing no unrealistic beliefs or desires.
• Salice and Miyazono’s account of identification (p. 5): a Pushmi-Pullyu representation, descriptive and directive at once, not a belief.
• The minimal group paradigm shows identification motivates pro-group behaviour (p. 6), from trivial shared features such as preferring Kandinsky to Klee or over-estimating dots (Tajfel et al. 1971).
• Neutrality about groups (p. 5). Groups may rest on common features or on shared activities; the interpretation does not choose.

Key results

• The debate as the authors set it out (pp. 2–4).
  • Batson’s experiments: Katie Banks (Coke et al. 1978; misattribution of arousal) and Elaine (Batson et al. 1981; similarity and ease of escape). Stocks et al. (2009) answer the psychological-escape reply.
  • Cialdini et al.’s study, as reported (pp. 3–4). Arizona State introductory psychology students imagined a near-stranger, an acquaintance, a good friend or their closest family member (“a sibling if possible”), evicted from their apartment. They chose among seven levels of help, up to letting the person live with them rent-free. They rated “oneness” on a seven-point scale and on the IOS scale, plus empathic concern, sadness and personal distress. Closeness raised empathic concern and oneness, and both predicted helping. “Crucially, when oneness is statistically controlled, empathic concern is not predictive of helping”, while oneness still predicts it with empathic concern controlled (p. 4). No statistics are given.
  • Cialdini et al.’s reading, quoted: empathy is a signal of oneness (1997, p. 491). SMH “seriously undermines the logic of the empathy-altruism hypothesis” by compromising “the distinction between selflessness and selfishness” (1997, p. 481). The helping is called “nonaltruistic”, not egoistic (p. 4).
  • Cialdini et al. deny a physical merging: “What is merged is conceptual, not physical” (1997, p. 482, quoted on p. 4).
  • Batson’s four versions of SMH (2018; note 4, p. 3): self-other identification; including the other in the self (Aron); seeing aspects of the self in the other (Cialdini); self and other as interchangeable exemplars of a shared group identity (Turner 1987). The authors place their reading in the fourth (note 6, p. 5).
  • On Batson et al. (1997): “Batson responded to Cialdini and colleagues by conducting experiments that cast doubt on SMH … and by indicating the methodological problems … (Batson, 2018, Chapter 10). These empirical issues about SMH, however, are outside the scope of this paper” (p. 4).
• May’s challenge (p. 4; May 2011, 2018). Three readings of merging:
  • peculiar belief: X believes she exists in two bodies; unrealistic;
  • indeterminate identity: personal identities blur; empirically unclear, and helping needs Y represented as another person;
  • property: X believes some of her properties are in Y’s body; unrealistic, and still needs an altruistic motive to explain helping.
• The group identification interpretation (pp. 5–6). When X empathizes with Y, “X group identifies with Y and thereby comes to conceive of Y’s welfare as being constitutive of X’s first-person plural (our) welfare” (p. 5).
  • It covers empathy-induced helping only, not all helping.
  • It allows identification without empathy, as in minimal groups.
  • It is adequate because Elaine and Kandinsky studies share a structure: participants learn of a similarity and are then offered a chance to help (p. 6).
  • It is plausible because it posits a shared social identity, not two bodies, and a PPR, not a belief (p. 5).
• Causal versus constitutive (pp. 6–7). Empathy may cause identification, by highlighting commonalities, which fits Cialdini’s signal reading. Or identification may be part of empathy. On the second, after Salice and Montes Sánchez on pride, Elaine’s suffering is the focus of Reynie’s empathy and his social self is its target. The authors are neutral and find the constitutive reading more interesting.
• Two objections (pp. 7–8).
  • Identification needs a desirable group, for self-esteem. Reply: minimal groups give in-group favouritism with no esteem to gain, and a father’s shame at his daughter’s crime is identification that lowers esteem.
  • Empathy reaches strangers outside any shared group: the woman in a burning building, suffering children far away. Two replies. Either we do identify with them, through a salient similarity created by empathic affective mirroring. Or the helping is not empathy-induced but reasoned, as Bloom argues.
• A new taxonomy (pp. 8–11).
  • Empathy-induced, group-identification-driven helping (EGHB) is a kind of group-identification-driven helping (GHB). Reynie’s helping is EGHB; Sticky’s in the Kandinsky study is GHB but probably not EGHB.
  • Not egoistic in the usual sense: the motive is not one’s first-person singular welfare. Self-esteem and reciprocity accounts are rejected. An evolutionary function of reciprocity is not a personal aim.
  • Not altruistic in the usual sense either: Elaine’s welfare matters insofar as it is constitutive of “our” welfare.
  • So: altruistic at the individual level (concern for her welfare, not mine) and egoistic at the group level (concern for our welfare, not theirs). Salice and Satne’s Alba, who does her part in a bridge her community decided on, is GHB of the same shape (p. 10, note 11).

Claims

id	claim	strength	support
C1	Self-other merging is best read as group identification with the other, which makes the other’s welfare constitutive of one’s first-person plural welfare	moderate: conceptual, argued against May’s three readings	pp. 4–6
C2	So read, SMH is behaviourally adequate: it explains Batson’s helping results in the same way as minimal-group favouritism	weak: an analogy of procedure; no new data, and Batson’s own results against SMH are not examined	p. 6
C3	So read, SMH is psychologically plausible, since identification is a non-doxastic representation, not a belief in being two persons	moderate given Salice and Miyazono’s PPR account, itself a conjecture	p. 5
C4	Group identification does not require a desirable group or serve self-esteem	moderate: minimal groups and other-induced shame	p. 7
C5	Empathy for strangers either involves identification via affective mirroring or is not empathy-induced	weak: two alternative replies, neither tested	pp. 7–8
C6	Helping driven by group identification is altruistic at the individual level and egoistic at the group level, because the other’s welfare is a constituent, not a cause, of “our” welfare	moderate: conceptual	pp. 8–11

Concepts

• self-other merging hypothesis (SMH): X helps Y out of concern for X’s own welfare, which through merging includes Y’s.
• group identification: acquiring a self-conception as a group member, shown in first-person plural talk (“we”).
• group identification interpretation: empathizing with Y involves identifying with Y, so that Y’s welfare is part of “our” welfare.
• causal versus constitutive interpretation: empathy causes identification, or identification is part of empathy.
• focus versus target (Salice and Montes Sánchez): what an emotion is directed at versus the self it is ultimately about.
• GHB / EGHB: helping driven by group identification; the subset induced by empathy.
• individual-level versus group-level egoism/altruism: whose welfare motivates (mine or hers), and whose plural welfare it serves (ours or theirs).

Connections

• Salice and Miyazono (NOTE-1266). Miyazono’s earlier paper supplies the account of identification. Its PPR, “you are a member of us; act as such”, is what makes this reading non-doxastic (p. 5). The link has a cost. That paper separated self-transformation (being one of us) from adopting the we-perspective (reasoning about our good), and the team reasoner (“We intend”) from the team benefactor (“I intend that we”). This paper’s Reynie identifies with Elaine but acts for her, not with her. That looks most like a self-transformed team benefactor, a combination the earlier paper did not name. An inference.
• Self-categorization theory (THEORY-1114). The authors place themselves in Batson’s fourth version, Turner’s interchangeable exemplars (note 6), and cite Turner 1982 for identification. But they stay neutral on what grounds groups, and their mechanism is similarity plus affective mirroring, not metacontrast.
• Group entity versus category (TERM-096). The “group” of Reynie and Elaine, and the Kandinsky-lovers, are categories formed by a shared property. The authors also allow groups grounded in shared activity (p. 5). Their argument does not need the distinction, but a test of it would.
• Cialdini et al. 1997 (LIT-1728). The paper is a defence of its conclusion and reports one of its studies (pp. 3–4).
• The SEP entry (read the same day) reports May’s objection and the numerical-identity argument this paper answers.

Bearing on the record

• The owner’s thesis (CLAIM-491). This is the closest statement of the thesis in the literature the record holds, for the case of empathy. It bears on the claim at three points. These are inferences unless marked.
  • It gives the thesis a precise form. Collectivist or empathic helping is concern for a whole one belongs to, of which the other’s welfare is a constituent. That is “self-interest on a larger self-identity” in the authors’ words: egoistic “at the group level”.
  • It splits the thesis’s verdict. At the individual level the same act is altruistic, and concern for the other is non-instrumental (pp. 8–9). So the thesis is not “no action is altruistic”. It is “no altruism without an extended self that it also serves”. Batson’s empathy–altruism hypothesis, read at the individual level, is then compatible with the thesis (the authors say as much, note 3, p. 3).
  • It moves the dispute with CLAIM-494. If empathy constitutively involves identification, no experiment can find empathy with identification absent, and Batson’s controls for merging would have to be failed measures of identification. That rescues the thesis only by making it unfalsifiable for empathic helping.
• The triviality objection (CLAIM-492). The paper meets the objection’s demand in one direction and fails it in the other.
  • Identification can occur without empathy (minimal groups, p. 5), so the criterion is not caring under another name.
  • But the paper gives no measure of identification apart from empathy and helping. Its reply to the strangers objection (p. 8) shows the risk: where empathy reaches a stranger, identification is inferred from a similarity the empathy itself creates. On the constitutive reading, empathic concern always counts as identification. Then “all empathic helping serves an extended self” is true by construction.
• What happened when Batson controlled for merging. Not reported: the authors say Batson et al. “conducted experiments that cast doubt on SMH” and declare the empirical issues out of scope (p. 4).
• A dated note to CLAIM-491 is proposed:

  Note of 2026-10-29: altruistic for her, egoistic for us. Miyazono and Inarimori (NOTE-1277) read Cialdini’s “merging” as group identification. In empathizing with someone, one comes to count her welfare as a constituent of “our” welfare, and helping is then altruistic at the individual level and egoistic at the group level. This is the thesis in a precise form, and it leaves Batson’s individual-level altruism standing. But on their constitutive reading empathy always involves identification. So the thesis gains content only if identification can be measured where empathy is present and found absent.

• No THEORY filed. No ML instruction.

Limitations

• Conceptual only. No data. The empirical rebuttal by Batson et al. (1997) is explicitly not addressed (p. 4).
• The adequacy argument is an analogy of procedure. Elaine’s helpers and Kandinsky-lovers both learn of a similarity and then help. That does not show that the same process explains both. The authors grant that empathy matters in one and not the other (note 7).
• No criterion of identification apart from its effects. The paper never says how to tell, in a given helper, whether identification occurred.
• Cialdini’s study is reported without numbers. The paper’s account of the oneness measure (“a seven-point scale” of “oneness” plus the IOS scale) differs from the SEP’s quotation of Cialdini (1997: 484): the IOS scale plus a 7-point “we” item. The SEP quotes the original, so it is probably the more accurate.
• Typos.
  • “X’s sorry for Y” for “sorrow” (p. 2). Footnote 2’s list numbers “(4)” twice.
  • “motivated them reduce” (p. 3); “the father pride” (p. 7); “One might think is that” (p. 9).
  • Cialdini’s “the distinction between selflessness and selfishness” is cited to p. 481 on p. 4 and to p. 482 on p. 9.
  • In the references, “Doris, J. M., Stitch, S.” for Stich, with the SEP given as published by “Springer”; “Being one of us group identification” lacks its stop.

Open questions

• Can group identification with a specific helped person be measured apart from empathic concern, for instance by first-person plural use or implicit self–other association, and does it carry the empathy–helping link when empathy is held fixed? That is the experiment the paper implies and does not run.
• Is the identified-with “group” here a category (shared property) or a group entity (shared activity)? Would helping differ between the two?