Pessimistic Meta-Induction

The pessimistic meta-induction is the argument that since the history of science is a graveyard of once-successful theories that were later abandoned — caloric, phlogiston, the luminiferous aether — we have inductive grounds for doubting that our current theories are true, or even approximately true. The argument does not claim that any particular theory is false; it claims that the track record of science as a whole undermines the inference from predictive success to truth.

Formulated most forcefully by Larry Laudan, the argument targets the central realist claim that the success of science is explained by the approximate truth of its theories. If past success did not correlate with truth, then present success does not either — unless one can identify a principled difference between past and present theories that makes the latter immune to the pattern.

The realist response, developed by Philip Kitcher and others, distinguishes between the working posits of a theory (the entities it genuinely needs to make its predictions) and its idle posits (theoretical baggage that can be jettisoned without loss). On this view, electrons survive theory change in a way phlogiston did not, because electrons are doing genuine explanatory work that transcends any particular theory.

Whether this response succeeds depends on whether the working/idle distinction can be drawn non-retrospectively — without the benefit of hindsight that tells us which posits were genuinely indispensable.

_The pessimistic meta-induction is not a skeptic's trick. It is the honest record of a discipline that buries its mistakes and forgets where the graves are._