Defenses of induction

Notes for October 15

Main points

The problem with inductive inferences, according to Hume, is that they depend on a principle that we cannot establish. Inductive reasoning involves moving from observations in the past to conclusions about the future.Really, the conclusions are about anything that isn’t observed, whether in the past, present, or future. Hume had assumed that in order for this to count as reasoning, we would need to have reason to believe a principle establishing the relevance of observations in the past to predictions about the future: the assumption that the course of nature will continue to be the same in the future as it has been in the past. He argued that we cannot establish this principle without engaging in circular reasoning. So, he concluded, we have no reason for drawing inductive inferences.

Can we do without the principle?

Hume’s conclusion, of course, is insane. We could not survive if we did not make inferences based on past observations. If past observations gave us no reason to draw any particular conclusion about what will happen in the future, this really obvious practice of ours would make no sense. That can’t be right.

So we looked for a way out. What if we got rid of the principle and just said that having a lot of observations is enough reason, all by itself, to draw conclusions about what will happen in the future?

The example of the Club for Men was supposed to show that this won’t work. We could have 499 observations like this: ‘a person comes out the door’ and ‘the person is a man.’ In Hume’s lingo, we have a constant conjunction of coming out the door and being a man. Does that alone give us reason to draw a conclusion like this: ‘the next person to come out the door will be a man’? It does only if we assume that all the members are men. If I had a membership list that showed there was one woman out of 500 members, I would draw the opposite conclusion.

Hume contends that the uniformity principle plays the role of establishing the relevance of observations of the past, much as the assumption about the membership in the Club for Men did.

So are we stuck?

Angela suggested that we are hard wired to make inductive inferences. Franklin added a plausible evolutionary explanation.

This is basically Hume’s view. We can’t help making these inferences. (He uses a mechanism of habit to explain why we do so.) It’s fortunate for us that we do a reasonably good job of it and we don’t really understand why we do so.

Key concepts

  1. The difference between cause and correlation
  2. Why Hume thought we needed to assume the uniformity principle.
This page was written by Michael Green for Problems of Philosophy, Philosophy 1, Fall 2013. It was posted October 24, 2013.
Problems of Philosophy