Philip Kitcher

Philip Kitcher (born 1947) is an American philosopher of science whose work bridges the analytic tradition's concern with formal rigor and the pragmatic tradition's concern with the social organization of knowledge. Where Carl Hempel asked what structure a scientific explanation must have, Kitcher asked what structure a scientific community must have to produce explanations worth trusting. This shift — from the logic of explanation to the sociology of epistemic trust — makes him one of the most consequential philosophers of science of the late twentieth century, and one of the most neglected by philosophers who still believe the field's questions are purely logical.

The Unificationist Theory of Explanation

Kitcher's alternative to Hempel's deductive-nomological model is the unificationist theory of explanation: to explain a phenomenon is to show that it follows from argument patterns that minimize the number of independent assumptions needed to derive the widest range of phenomena. A good explanation is not one that deduces a particular event from a law, but one that reveals how that event is an instance of a pattern that also generates many other events we already understand.

The unificationist account has a surprising consequence: explanation is not merely an epistemic achievement but an organizational one. The patterns that unify are constructed by research communities over time, and their quality depends on the community's ability to integrate disparate results. This connects directly to Kitcher's later work on the division of cognitive labor — the problem of how a scientific community should distribute its investigative efforts across competing hypotheses so that the full space of possibilities is explored without wasteful duplication.

Epistemic Diversity and the Social Organization of Science

In The Advancement of Science (1993), Kitcher argued that epistemic trust is not merely a psychological lubricant for scientific collaboration but a structural requirement. No individual can verify more than a tiny fraction of the claims their own work depends upon; scientific progress requires that researchers trust the outputs of others, and that the trust be calibrated to the track records of individuals and methods. A community that trusts too little cannot build cumulatively; a community that trusts too much cannot correct errors.

This framework makes Kitcher the founder of what might be called social epistemology of science — not the sociology of science practiced by Bruno Latour, which treats scientific claims as political negotiations, but a normative discipline that asks how a community ought to be organized to maximize the reliability of its collective beliefs. Epistemic diversity — the distribution of non-redundant perspectives within a community — is not a cultural virtue on this view but a structural property with measurable consequences for the community's ability to locate true hypotheses in a vast possibility space.

Realism and the Working Posit Distinction

Kitcher's response to Larry Laudan's pessimistic meta-induction introduced the working posit / idle posit distinction: not all theoretical entities are equal. Some posits — 'working posits' — are indispensable to the theory's predictive and explanatory success across multiple independent contexts. Others — 'idle posits' — are theoretical baggage that can be jettisoned without loss. Electrons are a working posit; phlogiston was an idle posit. The historical record shows that working posits survive theory change, while idle posits are abandoned.

Whether this response fully succeeds is debated. The distinction must be drawn without retrospective hindsight — we cannot know which of our current posits are genuinely indispensable until after the next theory change. But Kitcher's framing shifts the realism debate from a question about the truth of theories to a question about the entitlement of the community to trust specific components of its theoretical framework. It is not a metaphysical argument but a social-epistemic one: we are entitled to trust working posits because the community's error-correction mechanisms have historically preserved them, and because their cross-contextual robustness makes them unlikely to be artifacts of any single theoretical framework.

The unificationist theory is not merely an account of explanation. It is an account of what explanation becomes when you stop pretending that individual minds are the primary unit of scientific cognition and recognize that understanding is built by networks of trusting, disagreeing, partially-informed agents. Kitcher's philosophy is not philosophy of science — it is philosophy of scientific communities as epistemic systems, and the field has not yet caught up with what he started.