The reception of Alliance Theory followed the format built into the venue and then followed the theory’s own logic, which is the fun part.
Psychological Inquiry runs target articles with peer commentaries and an author reply, and “Strange Bedfellows” appeared in volume 34, number 3 (2023), pages 139-160, followed by the authors’ reply, “The Strange Epicycles of Political Psychology: A Response to Commentaries,” pages 226-238. The reply’s title tells you the temperature. Calling your critics’ frameworks “epicycles” compares mainstream political psychology to pre-Copernican astronomy, patch upon patch to save a failing model.
The journal published commentaries including “Strange Bedfellows and Their Irrational Pillow Talk”, “The Alliance Theory: A Strategic Model of Moral Judgments?”, “Seven Grand Challenges for Evolutionary Political Psychology or: Political Ideologies as Ad-Hoc Alliances…So What?”, “Distinguishing Between Worldview Conflict and Shared Alliances: Commentary on Pinsof, Sears, and Haselton”, “Speech Repression and Outrage from Orthodox Activists as Attempts at Facilitating Mobilization and Gaining Status among Allies”, “Culture, Partisanship, and Signaling: The Social Nature of Political Belief Systems”, “Political Belief Systems Are Not Singularly Rooted in Alliance Psychology”, “The Dangers of Alliances Caused the Evolution of Moral Principles”, and “Political Ideology is Not Meaningfully Explained by Alliances and is Not Inconsistent with Attitudinal Inconsistencies”. Read as a set, the titles map the battlefield: some commentators accept the machinery and dispute its scope (“not singularly rooted,” “so what?”), some defend the constructs Pinsof attacked (the ideology-is-real title), and some extend the frame to new targets (the speech-repression piece applies it rather than contests it).
Two commentaries are worth knowing by author. Mark Brandt, with Abigail Cassario, wrote “Distinguishing Between Worldview Conflict and Shared Alliances: Commentary on Pinsof, Sears, and Haselton”, arguing that Alliance Theory’s claim that group alliances determine belief contents needs separating from Brandt’s own worldview-conflict account. This one carries extra weight because David Pinsof leaned on Brandt’s data throughout the target article; the man whose studies supplied the ammunition showed up to say the ammunition fits a different gun. The second is Nichola Raihani and Quentin Atkinson, “It’s More Complicated Than That: Alliances Are One of Many Factors Shaping Political Belief Systems”, which grants that alliance machinery contributes but argues the account fails as a whole story. They ask why people pick a side and stick with it rather than defecting to the winning team when convenient, call the chimpanzee analogy a loose fit, and charge the target article with a biased reading of the evidence. Their sharpest analytic point: evidence that egalitarian commitments shift with context does not rule out stable dispositions, any more than the fact that anyone can be provoked to anger rules out dispositional differences in aggression. Pure alliance logic predicts more defection than the world contains.
Beyond the journal, the reception splits along lines the theory predicts, and nobody involved seems to enjoy the irony. The paper descends from the coalitional-psychology lineage of Tooby, Cosmides, Kurzban, DeScioli, and Petersen, and that community, plus the heterodox and rationalist blogosphere, adopted it enthusiastically. Pinsof turned the argument into a popular Substack, Everything Is Bullshit, and kept giving talks; UCLA hosted him presenting the theory as late as May 2025, billing his current research as covering political psychology and the nature of social status. On the other side sits the ideological-asymmetry camp around John Jost, whose system-justification program is among the paper’s explicit targets. Jost’s 2024 Journal of Social Issues article “Both-Sideology Endangers Democracy and Social Science” attacks the symmetry research program, Brandt, Crawford, the ideological-conflict hypothesis, on which Pinsof’s empirical case rests, arguing left-right equivalence claims are both empirically wrong and politically dangerous. The fight over Alliance Theory is thus a proxy front in the older asymmetry-versus-symmetry war, and each camp’s verdict on the paper tracks its coalition membership with a fidelity Pinsof could cite as data.
The paper has entered the working literature rather than conquered it. It gets used in social and personality psychology as a live alternative account, for instance in Woitzel and Koch’s 2025 Personality and Social Psychology Bulletin study of ideological favoritism, and in review pieces on left-right psychology such as Reyna’s 2024 Social and Personality Psychology Compass article. Political science mostly ignores it, partly because its core claim, that mass belief systems lack ideological coherence, restates Converse’s 1964 finding that the discipline already metabolized, with the provocative addition that elites are no more coherent than the masses, just more loyal. That addition is the part political scientists have least absorbed and the part most worth watching.
The paper is absorbed by its natal coalition, resisted by the coalition it attacked, granted partial credit by the empirical middle (alliances matter, the monocausal version overreaches), and not yet tested by the cross-cultural mapping program the authors proposed.
The Theory That Eats Its Readers: Alliance Theory and the Structure of Its Own Reception
A theory of political belief faces a hazard no theory of, say, protein folding ever meets: its subject matter includes the people who will judge it. When David Pinsof, David O. Sears, and Martie Haselton published “Strange Bedfellows: The Alliance Theory of Political Belief Systems” in Psychological Inquiry in 2023, they argued that political beliefs are badges of coalition membership rather than derivations from values. Belief systems, on their account, are patchworks of ad hoc justification assembled to support allies and damage rivals, and the machinery that assembles them, perpetrator biases, victim biases, attributional biases, runs symmetrically in every human head. The paper drew nine commentaries and an authors’ reply. It also drew a reception whose shape the theory predicts, a fact nobody in the exchange examined, perhaps because examining it costs every participant something.
Begin with the two objections that survive contact with the target article, because the reflexive argument only earns its place after the serious criticism has been faced.
The first comes from Nichola Raihani and Quentin Atkinson, and it concerns stickiness. If political beliefs are alliance badges, allegiance should track advantage. Coalition members should defect when the coalition fails them, drift toward winners, reprice their loyalties the way markets reprice assets. Some people do. Most do not. Partisanship in the United States behaves less like a portfolio and more like a birthmark; it survives lost elections, lost jobs, lost wars, and decades of betrayed material interest. Raihani and Atkinson press the point with an analogy: showing that egalitarian judgments shift with context no more eliminates stable dispositions than showing that anyone can be provoked to anger eliminates dispositional differences in aggression. Context-sensitivity and character can both be real. The target article, they argue, treats evidence of flexibility as evidence against stability, which does not follow.
The objection has more reach than its authors claim for it. Stickiness is not merely a residue Alliance Theory fails to explain; it points at the phenomenon the theory is weakest on, which is cost. Alliance machinery, as Pinsof and colleagues describe it, explains cheap beliefs beautifully, the poll answers, the double standards, the flexible outrage. It explains expensive beliefs badly. The partisan who stays loyal through forty years of defeat, the convert who breaks with his family, the believer who takes a loyalty to the grave, all of them are paying prices that a badge-logic accounts for only by adding auxiliary hypotheses about costly signaling, and auxiliary hypotheses are what the authors’ own reply, “The Strange Epicycles of Political Psychology,” accuses the other side of collecting. A theory that mocks epicycles must travel light.
The second live objection comes from Mark Brandt, writing with Abigail Cassario, and it carries a special sting because the target article is built substantially on Brandt’s data. Brandt’s worldview-conflict research established much of the symmetry evidence Pinsof deploys: liberals and conservatives dislike their ideological opponents at equal rates, discriminate at equal rates, and direct their hostility at whichever groups they perceive as ideologically distant. Pinsof reads these findings as showing that alliance, not worldview, drives evaluation. Brandt and Cassario answer that the findings show perceived worldview conflict driving evaluation, which is a different engine. On the worldview-conflict account, people hold something like actual beliefs, perceive others as sharing or opposing them, and respond to the perceived disagreement. On the alliance account, the beliefs are downstream of the roster. The two models often predict the same behavior, which is why Pinsof could borrow the data, but they part company at a testable joint: worldview conflict predicts that manipulating perceived belief similarity changes evaluation even when coalition membership is held constant, while Alliance Theory predicts that coalition cues dominate belief cues when the two conflict. The man who ran the studies says the studies belong to the first model. The exchange is a property dispute over an empirical estate, and it remains unsettled.
These two objections mark the theory’s honest frontier. Now the reflexive point.
Sort the published and public reactions to “Strange Bedfellows” by verdict, then sort the reactors by intellectual lineage, and the two sorts produce nearly the same list. The theory descends from the coalitional-psychology tradition of John Tooby (1952-2023) and Leda Cosmides (b. 1957), through Robert Kurzban, Peter DeScioli, Michael Bang Petersen, and the Weeden-Kurzban self-interest program. That community received the paper as a consolidation of things it already held. The adjacent heterodox ecosystem, the rationalist blogs, the evolutionary podcasts, the readers primed by a decade of replication-crisis skepticism toward social psychology’s ideology research, amplified it; Pinsof’s Substack built an audience on the argument. On the other side, the paper’s explicit targets, the system-justification program of John Jost, the moral-foundations program, the authoritarianism literature descending from Adorno through Altemeyer, either ignored it or answered it as part of a larger counterattack. Jost’s 2024 “Both-Sideology Endangers Democracy and Social Science” prosecutes the entire symmetry literature on which Pinsof’s case rests, and prosecutes it in a journal of social issues, on the announced ground that the science is wrong and that its wrongness endangers democracy, a coupling of empirical and coalitional stakes that could serve as an exhibit in the target article’s table of propagandistic biases. Verdicts tracked rosters. The commentaries in the journal itself, refereed and collegial, spread across the middle, but the middle too is legible: the commentators who accept the machinery and dispute its monopoly are, by and large, researchers whose own programs the machinery leaves standing, and the commentators who defend ideology’s reality are researchers whose programs require it.
The lazy version of this observation is a gotcha, and the gotcha is worthless. Every theory’s reception is somewhat sorted by school; that is what schools are. The interesting version asks what follows when the sorted theory is a theory of sorting, and the answer comes in three steps.
First, the reception constitutes weak evidence for the theory. Alliance Theory predicts that evaluations of politically consequential claims will track the evaluator’s coalition position more than the claim’s evidential merits. Its own reception conforms. Conformity of one case proves little, but a theory whose reception had cut cleanly across lineage, with system-justification researchers persuaded and coalitional psychologists dismissive, would have presented a small anomaly. No anomaly appeared.
Second, and cutting the other way, the theory poisons its own well. If beliefs are badges, then the belief that beliefs are badges is a badge, worn by a coalition of evolutionary psychologists, symmetry researchers, and heterodox commentators who profit reputationally from mainstream social psychology’s embarrassment. Pinsof’s community has its rivals, its grievances, its market. The theory supplies its enemies a fully general dismissal: you would say that, your roster requires it. But the dismissal is symmetric, which is the trap. Jost’s camp can wave away Alliance Theory as coalition propaganda only by invoking the very machinery, motivated evaluation in service of group position, that the theory posits and Jost’s camp minimizes when the symmetric version is pointed at the left. Each side’s cheapest weapon against the other is a concession to the other. The debate is a room where every gun fires backward.
Third, the loop is escapable, and specifying the exits is where the essay stops being clever and starts being useful. Self-referential taint does not distinguish true theories from false ones; a theory of gravity formulated by falling men is not thereby refuted. What distinguishes them is prediction risked in advance. Alliance Theory’s authors proposed a program: map a society’s alliance structure first, then predict the contents of its belief systems, including the double standards, from the roster plus the bias catalogue. The program has not been run at scale, and it is the only verdict that will not itself be a badge. Three tests would carry real weight. Run the mapping program in societies whose alliance structures differ sharply from the American one, pre-registering the predicted belief patchworks; the theory’s own stochasticity claim, that alliance structures are historically arbitrary, guarantees the test set exists. Stage the Brandt crux directly: pit coalition cues against belief cues experimentally and measure which dominates evaluation when they conflict, with both labs party to the design. And price the expensive beliefs: if loyalty that survives decades of loss can be shown to yield offsetting coalition returns, the stickiness objection dissolves; if it cannot, Raihani and Atkinson have found the theory’s boundary, and the boundary is where dispositions, or doctrines, live.
There is a fourth exit, older than any of them. Adversarial collaboration, the practice of rival camps designing studies together, exists precisely because science is staffed by coalition animals, and it works, when it works, by making the alliance machinery fight itself. A Pinsof-Jost collaboration is difficult to imagine, which is roughly the point; the difficulty measures how much of the dispute is roster.
What the reflexive reading finally yields is not a verdict on Alliance Theory but a sharpened sense of what a verdict would have to look like. The theory says the war of ideas is mostly a war of teams wearing ideas. Its reception was a war of teams wearing ideas. The demonstration is either the theory confirming itself or the theory devouring itself, and no one inside the war can say which, because saying is joining. The only tribunal left standing is the one science built for exactly this predicament: prediction, pre-registration, and the slow accumulation of results that cost the winning coalition something to accept. Converse showed sixty years ago that mass publics lack ideological constraint, and the discipline absorbed it because the finding kept arriving no matter who ran the survey. If Alliance Theory is right, its evidence will have to arrive the same way, over the objections of its enemies and, harder, over the applause of its friends.
Alliance Theory is situationism transposed into political psychology, the inferential moves match almost line for line, and the history of the person-situation debate tells you roughly how this one ends. But the transposition changes two things that matter, and one of them cuts in Pinsof’s favor.
Take the parallel first.Walter Mischel (1930-2018) argued in Personality and Assessment (1968) that cross-situational consistency in behavior was embarrassingly low, correlations around .30, and that “traits” were largely attributions observers project onto noisy conduct. Social psychology supplied the flagship demonstrations: mood and hurry and ambient noise pushing helping behavior around while character sat idle. Gilbert Harman (1938-2021) and then John Doris, in Lack of Character (2002), carried the result into philosophy: virtue ethics presupposes robust traits, robust traits don’t exist, therefore the whole edifice of character talk rests on a systematic attribution error. Now read Pinsof: cross-target consistency in political values is embarrassingly low, liberals who find CEO pay unfair find movie-star pay fine, conservatives who revere authority defect from the FBI in eighteen months, and “values” are largely attributions, by observers and by the believers themselves, projected onto conduct that alliance machinery is driving. Same structure: within-person inconsistency wielded as proof that the underlying disposition is a ghost. Even the debunking psychology matches, with Doris invoking the fundamental attribution error where Pinsof invokes the moralistic mask over coalition interest.
And the counterattack that beat situationism is the counterattack Raihani and Atkinson mount, nearly verbatim. Seymour Epstein (1924-2016) showed that single behaviors are unreliable indicators and that aggregation across occasions restores strong trait prediction. William Fleeson’s density-distribution work then formalized the peace: a person is a distribution of states, highly variable moment to moment, with a stable mean and a stable spread. The introvert has extraverted hours; the introversion is the mean, and the mean barely moves across decades. Raihani and Atkinson’s anger analogy is this exact argument, that anyone can be provoked, and dispositional differences in aggression exist anyway. Context-sensitivity and character are compatible because character just is the shape of one’s context-sensitivity. Meanwhile the situationist canon fared badly in the replication era, the Stanford prison study discredited as theater, the priming literature collapsing, while the trait side accumulated longitudinal stability data, behavioral-genetic evidence, and predictive validity for mortality, divorce, and career on par with class and IQ. Doris’s strong claim lost ground because its evidence rotted while the other side’s compounded.
So does Alliance Theory await the same fate? Here the transposition matters.
First difference, favoring David Pinsof: his evidence base is, so far, sturdier than situationism’s was. The core symmetry findings, Brandt, Crawford, Chambers, the Ditto meta-analysis of partisan bias, have replicated reasonably well, and the rapid mass flips are not lab curiosities but public polling: Republican support for Putin tripling as Trump embraced him, the FBI reversal, the COVID sortings. A disposition cannot reverse sign in eighteen months. A roster can. Nothing in the situationist canon was this strong, because a dime in a phone booth is a trivial situation and a realigned coalition is not.
Second difference, favoring the dispositionists: political psychology already possesses the stability evidence that personality psychology had to build, and some of it was built by Pinsof’s own coauthor. David O. Sears (b. 1935) spent his career documenting symbolic politics, the finding that party identification and core political predispositions crystallize early and persist across the lifespan with a stability that rivals any Big Five trait. Twin studies since Alford, Funk, and Hibbing (2005) put substantial heritability on political attitudes. The stickiness that Raihani and Atkinson wave at the theory is not a promissory note; it is forty years of data, a chunk of it bearing Sears’s name. “Strange Bedfellows” is, among other things, a man’s late-career argument with his own archive.
How, then, do the two accounts fit together rather than collide? The paper contains the treaty terms, in the section most readers skip. Pinsof and colleagues allow that individual differences may shape beliefs through allegiances: sexual restrictedness breeding enmity toward promiscuous groups and thence policy positions, formidability breeding military allegiance and thence hawkishness. Dispositions explain recruitment and the between-person variance: who is drawn to which coalitions, who bonds tightly and who sits loose, why the same roster contains zealots and tourists. Alliance machinery explains contents and the within-person variance: which beliefs a member expresses, the double standards, the patchwork, the flips when the roster reshuffles. Disposition picks the team and sets the grip strength; the team writes the catechism. Fleeson’s formalism translates directly: a citizen’s expressed positions are a distribution whose spread the coalitional situation drives and whose mean the disposition anchors. Goren’s longitudinal finding, that party identification predicts later egalitarianism and not the reverse, settles the direction for contents while leaving recruitment untouched, since something upstream still chose the party.
The contradiction survives only at the strong poles. Strong Alliance Theory predicts that controlling for allegiance eliminates trait-belief correlations, a claim the paper states and the field has barely tested. Strong dispositionism predicts that values drive both allegiance and belief, which Goren and the flip data already wound. Both poles will lose, if the person-situation precedent holds, and the precedent held everywhere else: that war ended in Kurt Lewin’s (1890-1947) old formula, behavior as a function of person and situation, with the interesting science relocated to the interaction. Expressed belief as a function of disposition and roster is where this one lands.
One last twist. What is a standing alliance, held for forty years, transmitted to one’s children, woven into marriage and neighborhood and self-description? It is a disposition. The vocabulary war conceals an ontological merger: loyalty sustained across decades stops being a move in a coalition game and becomes character, exactly as a virtue, on the best post-Doris accounts, is not a situation-proof essence but a stabilized pattern of situation-response. The Sunni-Shia tag holding for fourteen centuries, the Democrat who buries three losing candidates and knocks doors for a fourth, the quietist in his cell: at that time-scale the distinction between “his coalition” and “his character” has no cash value. Pinsof dissolves values into alliances; time re-precipitates alliances into values. The theory is right about the solvent and silent about the sediment, and the sediment is what Doris’s critics keep pointing at.
