An Alliance Theory of Antisemitism

Here’s a new paper, “Justifying Antisemitism: Political Liberalism and Perceptions of Prejudices.”
The paper has pre-registration, three experiments, quota samples, open data, and a discussion section that reports the findings that cut against the paper’s news hook as prominently as the ones that make it. The most publicized claim available to the authors was “liberals tolerate antisemitism,” and they lead instead with the finding that complicates it: absent justification, liberals dislike the antisemite more than conservatives do, across every target group tested. They also report the Experiment 3 failure of the crucial interaction rather than burying it, and they use the Lakens–Etz logic correctly, treating a nonsignificant estimate inside the prior confidence interval as weak corroboration rather than as refutation. The theoretical contrast with the black sheep effect is smart: existing literature might have predicted liberals would punish an ingroup-marked antisemite extra hard, and they found the opposite, which means the result was capable of surprising them.
The soft spot is the manipulation. “[Name] doesn’t like Jews, because [Name] strongly disapproves of Israel and its war in Gaza” is meant to hold the prejudice constant and vary the justification. But the justification changes what the first clause means. A liberal reader can parse the sentence as political anger spilling into loose talk, a person whose real attitude object is a state and whose “doesn’t like Jews” is careless shorthand, the way “I hate Russians” in March 2022 often meant “I hate what Russia is doing.” If participants charitably reinterpret the attitude rather than tolerate the prejudice, the liking boost measures forgiveness of imprecision, and the antisemitism has partly dissolved in the reader’s construal before any licensing occurs. The bigotry mediation is fully consistent with this reading: the justification reduced the inference that the person hates Jews as Jews. The authors would answer that this is exactly the point, that real-world antisemitism travels under this construal and benefits from this charity, and they would be right that the ambiguity is ecologically valid. But it means the paper cannot cleanly distinguish “liberals tolerate antisemitism when justified” from “liberals reclassify justified statements as not antisemitism.” Those are different findings with different remedies, and the abstract sells the first while the data permit the second. One more limitation: the dependent measure is liking a one-sentence stranger, which sits a long way from hiring, defending, marching beside, or excusing. And the most unsettling result in the paper gets the least attention: in Experiment 3 the conspiracy justification, Jewish power over markets, governments, and media, licensed roughly as well as the Israel justification at the sample mean, and liberals rated the conspiracist as ideologically closer to themselves. The authors call it unexpected and move on. It deserved a page.
Now the Alliance Theory fit. The paper wires into David Pinsof and cites “Strange Bedfellows” as the frame for its second explanation, and names the explanation “alliance politics.” In a previous post I said the mapping-and-prediction program had not been run and that verdicts on the theory would have to arrive as pre-registered results that cost some coalition something to accept. This paper is an early installment of that, and it pays the theory in three currencies. First, it confirms the badge logic experimentally: the identical prejudice plus a left-coded justification reads as ingroup membership, and ideological-distance mediation carries part of the liking effect. The justification functions as a coalition marker, which is Pinsof’s central claim caught in the act. Second, and this is the paper’s theoretical addition, it explains something Alliance Theory needed explained: why bigotry as such is penalized at all. If beliefs are just badges, why does anyone punish indiscriminate hostility? Answer: modern coalitions are heterogeneous alliances of subgroups, so a member who hates without targeting is a hazard to the roster itself; he might hate inward. Aversion to bigots is alliance maintenance. The justification then works by narrowing the inferred hostility to a licensed target outside the coalition, converting a roster hazard into a roster asset. Third, the asymmetry in the results, left-coded justifications licensing for liberals while right-coded ones mostly failed for conservatives, looks at first like trouble for Pinsof’s symmetric-machinery commitment, and turns out to be its vindication. The machinery is symmetric; the rosters are not. The conservative coalition currently holds Israel and, through evangelical philosemitism, Jews in its ally column, so an antisemite waving a traditional-values flag presents conservatives with a contradiction: ingroup badge, attack on an ingroup-adjacent target. The mediation data show the contradiction directly, ideological closeness pulling liking up while inferred bigotry pulls it down, canceling. The liberal roster, having moved Israel into the rival column since 2023, presents no such contradiction. Same engine, different maps, different outputs. That is Alliance Theory’s structure-versus-machinery distinction earning its keep, and it also stages, in miniature, the Brandt crux from earlier in this thread: coalition cues beat content cues when the two conflict, at least for perceivers evaluating strangers.
The neutralization theory of hatred paper fits here. Gresham Sykes (1922-2010) and David Matza (1930-2018) argued in 1957 that delinquents mostly share conventional morality and act against it by deploying techniques of neutralization: denial of responsibility, denial of injury, denial of the victim, condemnation of the condemners, appeal to higher loyalties. Two of the five map onto the Israel justification with no forcing. “Because of Israel’s war” is denial of the victim, the target class rendered blameworthy and thus not a victim at all, via the collective-responsibility move that assigns Gaza to a Jewish American who may never have set foot in Israel. “Because of the human rights of Palestinians” is appeal to higher loyalties, hostility recast as the overflow of a superior moral commitment; the paper’s own imagined confession, “I care so much about the rights of Palestinians that I can’t help but feel some disdain toward Jews,” is a textbook specimen. Neutralizations must be drawn from the surrounding value system, which is why they work. That single sentence explains the paper’s entire pattern of which justifications licensed and which backfired. The race-biology justification failed with liberals and backfired at scale because it draws on a value system the culture has anathematized; the crucifixion justification moved only the highly religious, the residue of the vocabulary’s former hosts; the human-rights justification succeeded with liberals because human rights is the hegemonic moral vocabulary of their coalition. And it explains the historical sequence the paper gestures at through Lewis and Feldman: theological antisemitism in the age of faith, racial antisemitism in the age of science, anti-Zionist antisemitism in the age of human rights. Each generation’s antisemitism speaks the language of that generation’s virtue, which is precisely why each generation has trouble recognizing it.
Sykes and Matza built a first-person theory: neutralization quiets the actor’s own conscience so the act can proceed. Crandall and Eshleman’s justification-suppression model, the paper’s proximate ancestor, kept that first-person focus. Moon, Barlev, and Neuberg move the whole apparatus to the third person and show that the techniques neutralize the audience: the justification’s job is not only to let the speaker live with his hostility but to let observers keep liking him, retain him on the roster, and spare themselves the cost of policing an ally. Neutralization theory supplies the content constraint, which vocabularies can license, namely those drawn from the perceiver’s sacred values. Alliance Theory supplies the audience and the stakes, whose values must be invoked and why the license is granted, namely coalition maintenance and the narrowing of inferred threat. The synthesis makes one prediction the paper does not test and someone should: condemnation of the condemners, the fifth technique, is the next neutralization in the sequence, already visible in the wild as the claim that antisemitism accusations are bad-faith weapons to silence criticism of Israel. On the joint account, that move should license further hostility for liberal perceivers, since it simultaneously invokes a coalition value, resisting powerful silencers, and reclassifies the anti-antisemite as the aggressor. If it works, the licensing loop closes: the justification excuses the hostility, and the meta-justification excuses ignoring anyone who objects.