Contribution
- A claim that moral judgment cannot be separated from social-relational context: “the moral status of actions cannot be determined independent of the social-relational contexts in which they take place” (p. 57).
- Four moral motives derived from relational models theory, each with its obligations, violations and constituting cues (pp. 61–64).
- Readings of violence, fairness and purity under the four motives, set against Turiel, Hauser, Mikhail and MFT (pp. 65–67).
- Proposals for research on freedom, religion, moral emotions, development and neuroscience (pp. 67–68).
Key insight
The question is not “did the act cause harm, was it unfair, was it impure?” but “what relationship is this, and did the act honour it?” Each of four relationship types obliges something different. The same killing is an atrocity under one model and a duty under another, so moral disagreement is often real disagreement about which relationship applies, not error.
Assumptions
- Function. “Our sense of morality functions to facilitate the generation and maintenance of long-term social-cooperative relationships with others” (p. 59, citing Fiske, Frank and Joyce).
- Relational models theory. Four schemas coordinate nearly all social interaction; they map onto nominal, ordinal, interval and ratio scales (p. 61). The theory’s validation is cited, not re-shown (p. 60–61).
- Scientific, not folk, “moral.” The theory is meant to capture much of what is called moral, not to match any culture’s folk concept (p. 60).
- Motives can be tacit. People need not be able to state the motive or endorse the judgment on reflection (p. 59).
Key results
The four motives (pp. 61–64):
| motive | relational model | requires | characteristic violation |
|---|---|---|---|
| Unity | communal sharing | care for and protect the in-group; give by need; shared responsibility; avoid contamination | defilement of the group, failure to aid a member |
| Hierarchy | authority ranking | subordinates defer and obey; superiors lead, guide and protect | disobedience, disrespect, superiors’ neglect |
| Equality | equality matching | even balance, in-kind reciprocity, turn-taking, equal shares, eye-for-an-eye | failing to reciprocate, unequal treatment |
| Proportionality | market pricing | rewards and punishments proportional to merit, contribution or guilt; cost–benefit trade-offs | cheating, taking benefits not deserved |
- Null morality (p. 64). Where no relationship is constituted, people are morally indifferent (Hopi children hurting birds; Ilongot headhunting).
- Violence (p. 65). Harm can be praiseworthy: against out-group threats (Unity), from superiors to subordinates or under orders (Hierarchy), in retaliation (Equality), or as a necessary cost (Proportionality).
- Fairness (pp. 65–66). Equal treatment is fair only under Equality; it is wrong under Hierarchy or Proportionality, and stealing may be “nonsensical” under Unity or Hierarchy.
- Purity (pp. 66–67). Concerns with purity arise under communal sharing; “the foundations of in-group/loyalty and purity/sanctity are actually variants of the same social-relational regulation motive” (p. 66). Incest is a violation of a culturally specific configuration of kinship; flag desecration offends those who identify with the nation; food taboos constitute clan, caste and religious identity.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Moral judgments depend on the relational model applied, not on features of acts | moderate as a framework | historical and ethnographic cases; cited experiments on relational framing (Clark & Waddell 1985; Fiske & Tetlock 1997 on taboo trade-offs) |
| C2 | There are four universal moral motives, one per relational model | weak–moderate | rests on RMT’s validation, cited; the moral extension is asserted |
| C3 | Violence is often morally motivated, and its acceptability follows the motive | moderate | cross-cultural and vignette findings cited (Cohen et al. 2006; Kelly et al. 2007; Cikara et al. 2010) |
| C4 | Purity and ingroup loyalty are one motive (Unity) | weak | argued from cases; stated as a prediction, untested here |
| C5 | Disgust tracks Unity violations, gratitude Equality, awe and respect Hierarchy | weak | stated as predictions (p. 68) |
| C6 | Where no relationship is constituted, there is moral indifference | weak | two ethnographic anecdotes |
| C7 | Much genuine moral disagreement cannot be resolved by facts or logic | informal argument | follows from C1 if the models are incommensurable |
Concepts
- relationship regulation (RR) — the theory that the core of moral psychology is motives to evaluate and guide behaviour “with reference to prescriptive models for social relationships” (p. 59).
- moral motive — evaluation plus the motivation to act on it, as opposed to a judgment alone (p. 59).
- constituting — establishing a relationship model through cues: bodily similarity and touch for communal sharing, size, height and precedence for authority, concrete matching for equality, symbols such as money for market pricing (Fiske 2004, as summarized).
- null morality — moral indifference where no relational model is active (p. 64).
- virtues — “quasi-moral” traits that make someone a good relationship partner, a penumbra around the motives (p. 59).
Connections
- Morality-as-cooperation (LIT-047, NOTE-034). Shared premise: morality
sustains cooperative relationships. LIT-047 cites this paper (p. 59) for
it. The mapping is rough:
- Unity covers MAC’s family and group, and also purity.
- Hierarchy covers MAC’s hawkish and dovish displays (bravery and respect), and adds superiors’ duty to protect.
- Equality covers reciprocity and equal division.
- Proportionality covers merit-based division and utilitarian trade-offs.
- Property has no motive of its own. The derivations differ: MAC starts from types of game; RR from the scale types of social relations. RR’s key move, that context fixes the morality of an act, matches LIT-047’s footnote 3, where a cooperative act is bad only when it undermines a larger form of cooperation.
- Bloom’s objections in NOTE-034.
- Sex. RR’s answer differs from the Reply’s. Sex wrongs defile a communal relationship (incest, adultery, caste rules), which is relational and, in RR’s sense, cooperative.
- Hitting. RR denies the premise. Battery is not always moralized as wrong: it is often required (honour, revenge, discipline).
- Moral disciplining (LIT-101, NOTE-132). Both say purity is cooperative. LIT-101 says puritanism protects self-control that cooperation needs and excludes contamination; RR says contamination is exactly the Unity concern. They are rival cooperative accounts.
- MFT. RR “adds to MFT … by grounding the foundations in a theory of social relationships” (p. 66). Graham et al. 2013 (LIT-499) accept it as compatible; the MFQ-2 (LIT-500) takes its Equality and Proportionality.
- Intuitionism. Haidt 2001 (LIT-511) used three of the same four models for the primate origins of intuitions. RR cites moral dumbfounding as people lacking words for the motive in use (p. 69).
Bearing on the record
- NOTE-034, C7. RR shows a third way of cutting the domain from a cooperative premise. It counts against the idea that game theory yields a unique list: MAC and RR start from the same function and get different units.
- NOTE-034, C8 (sexual morality is cooperative). RR supplies a worked version of that claim for incest and honour, which the Reply only sketched. It is no better tested.
- No THEORY document cites it. No ML instruction.
Note of 2026-10-16: THEORY-626 now cites LIT-507 as a source.
Limitations
- No new data. The cases are chosen to illustrate, and none is a test that could have come out against the theory.
- Which model is “in force” in a situation is read off after the fact; the paper gives no independent way to fix it before observing the judgment, so C1 risks being unfalsifiable.
- Purity is assigned to Unity by argument. Food taboos with no group boundary (most of the “harmless wrongs” of Haidt, Koller & Dias) fit less easily.
- The ethical implications section mixes description and prescription, though the authors flag the is–ought question (fn. 4).
Open questions
- Does manipulating which relational model is salient change purity judgments as RR predicts, for example disgust at an in-group member’s act versus a stranger’s?
- Do Equality and Proportionality motives, measured independently, predict different judgments of the same violence? The MFQ-2’s split (LIT-500) makes this testable.
Corrections
- none to a seeded skim (there was no seed)