The Justification of Induction

David Hume presented the classic statement of the problem of induction. In order for induction to be capable of giving us knowledge, we must be able to show that induction is justified. We could do this deductively or inductively, and these are the only options. We cannot do it deductively, because inductive inferences do not guarantee the truth of their conclusions, unlike deductive inferences. We cannot do it inductively, because that assumes what we are trying to prove.

Hence we can have no reason to believe any inductive inferences at all.


The argument is given in Hume 1748, secs IV, V part I.



There is an excellent summary of most of these positions in Lipton n.d..

Bad Non-Solutions

There are a number of tempting responses to the problem which do not touch it in the slightest.

Semantic Solution

Induction is justified because that is part of what justified means.

Pragmatic Solution

No matter how bad induction is, all the alternatives are worse.

Popperian Solution

We do not use induction anyway.

Reliabilist Solution

Hume was using a bad definition of knowledge, and we can justify induction given a better one.