A Priori Workshop: Peter Railton

In his paper, "A Naturalistic Transcendental Argument? (or, Why should Kantians have all the fun?)" Peter Railton argued that a naturalistically acceptable quasi-transcendental argument can be mounted against the sceptic. If I understood it (and I'm far from sure I do), the argument relied on the conditions necessary for a system to be in a position to learn from experience. The condition is, roughly speaking, that the epistemic options must be investigated with an expectation that assigns differential probabilities to each. If each possibility is assigned equal probability, then a system will not be able to rely on the information it gains to adjust the probabilities of further options. Further, it doesn't matter so much exactly which expectations the system begins with as, if there are epistemic reasons for the system to latch onto, then there will be a tendency for systems to converge on those reasons.

What this means for the sceptic is that his demand that his demand that we put aside all prior assignments of probabilities to the sceptical versus the non-sceptical hypotheses, is inconsistent with the supposition (sure made by the sceptic) that we are creatures capable of learning. This doesn't show that the sceptical hypothesis is false (there may be no epistemic reasons), nor that it is meaningless. Rather, it suggests that, contrary to what the sceptic wants to insist, we are justified in making assumptions (embodied in our behaving in accordance with certain norms) concerning those things the sceptic doubts despite not having an apriori 'from-scratch' argument that shows that these assumptions are true or objectively likely to be true.