Tribal Epistemology
Glover's most uncomfortable contribution to moral psychology was the observation that belief often follows identity rather than leading it. The person who adopts a tribe's position does not weigh evidence and arrive at a conclusion that happens to align with the tribe. The adoption precedes the weighing; evidence is then processed in the light of the already-adopted position. This is not stupidity or bad faith. It is the normal operation of social cognition, and it is why genuine reconsideration is rare even among thoughtful people. Glover documented the mechanism in ethnic conflicts, religious disputes, and political polarization — contexts where the tribal stakes were high and the evidentiary costs of defection were real. He would recognize the AI discourse of 2025–2026 as a textbook case. Triumphalist and elegist positions have hardened into tribal markers; the silent middle, which holds both truths in tension, has no tribe to validate its ambivalence and therefore remains silent.
In the AI Story
The mechanism operates through what Glover called belief-loss aversion. Once a position has been integrated into identity, losing the position is losing part of oneself. The psychological cost is not merely intellectual but existential, and the cost is paid preemptively through the filtering of evidence that might require the loss. The triumphalist who dismisses the Berkeley study's findings as complaints of the unadapted is not evaluating the study and concluding it is flawed. The identity is determining the evaluation.
The mirror operation occurs on the elegist side. The elegist who acknowledges democratization only as a regrettable side effect, never as a genuine gain, is not assessing the evidence. The identity — defender of depth, person who sees what triumphalists cannot — is determining that evidence of gain must be a distraction from the real story of loss.
Glover's framework identifies a specific escalation mechanism: the mutual attribution of bad faith. Once tribal identity has determined the meaning of evidence, the person presenting contrary evidence is not merely wrong. She is suspect. Her motivation becomes the object of analysis rather than her argument. This makes engagement across tribal lines not just difficult but morally contaminating: to take the other side seriously is to risk infection.
The contemporary AI discourse displays every feature Glover identified. The triumphalist dismisses critics as Luddites motivated by fear of displacement. The elegist dismisses advocates as shills motivated by financial interest. Both attributions are sometimes accurate. The accuracy is beside the point: the function of the attribution is not diagnostic but defensive, and the defense operates by closing off the engagement that might have produced genuine understanding.
The remedy, in Glover's framework, is not better arguments. It is longer time horizons and smaller stakes. Tribal epistemology weakens when the immediate identity cost of reconsideration is reduced — when the social environment supports the holding of contradictory truths, when the platform architecture does not punish ambivalence, when the time available for belief formation is long enough to include moments of genuine doubt. None of these conditions are present in the current AI discourse.
Origin
The observation that identity shapes belief is old — Hume noted it; Nietzsche made it central. Its development as a mechanism of institutional failure is Glover's contribution, rooted in his study of how ethnic and ideological groups came to see moral reality in opposite ways and act on the different visions with equal sincerity.
The contemporary application to the AI discourse draws on work in political psychology (Dan Kahan on cultural cognition, Jonathan Haidt on moral tribes) and on the sociology of knowledge production in polarized environments. The distinctive Gloverian contribution is the moral emphasis: tribal epistemology is not merely a cognitive bias to be corrected. It is a moral failure — a failure of the capacity to hold complexity, and therefore a failure of the capacity to respond adequately to situations that are morally complex.
Key Ideas
Identity precedes evidence. The tribe is adopted first; the evidence is then processed through the tribe's interpretive frame.
Belief-loss aversion. The psychological cost of abandoning a tribal position is existential, not merely intellectual. This is why thoughtful people maintain positions against evidence.
Mutual bad-faith attribution. Each tribe accuses the other of being motivated by something other than truth-seeking. The accuracy of the accusations is beside the point; their function is defensive.
Complexity cannot be tribal. The person who holds both truths has no group. Silence is the result, not of conviction's absence but of the absence of a community that rewards its expression.
Remedy is structural. Reducing tribal intensity requires architectures — social, temporal, platform — that reduce the immediate identity costs of reconsideration. The current AI discourse architecture amplifies those costs.
Appears in the Orange Pill Cycle
Further reading
- Jonathan Glover, Humanity: A Moral History of the Twentieth Century (1999)
- Dan Kahan, "Cultural Cognition of Risk" (2007)
- Jonathan Haidt, The Righteous Mind (2012)
- Robert Jay Lifton, The Nazi Doctors (1986)
- Hannah Arendt, The Origins of Totalitarianism (1951)