The Reliabilist Justification of Induction

Hume analysed 'knowledge' as justified true belief, and argued that that induction could not produce justification even if it produced truth. 'Knowledge' is better analysed as 'reliably-produced true belief'. Induction could be a reliable process, and thus could produce knowledge.

We can form the belief that induction is a reliable process on the basis of induction, because induction has been reliable in the past. If induction is a reliable process, this is a true belief, formed by a reliable process. Thus, if induction is a reliable process, we know that it is.


The best presentation of this position is in Lipton n.d..


The reliabilist account of knowledge is controversial.

The inductive justification seems viciously circular: don't we need to know that we know induction is reliable before we can use that knowledge to guide our decisions?