Epistemology: Hume's Skepticism and Induction Part 2

(intro music) Hi, my name is Daniel Greco. I'm an assistant professor of philosophy at Yale University. Previously, I discussed David Hume's distinction between relations of ideas and matters of fact.

Today, I'd like to discuss how that distinction gets applied by Hume to offer a skeptical argument concerning induction. So how can we come to know about matters of fact? A natural thought is that observation will tell us about matters of fact.

You know that it's sunny because you can go outside and look. Hume is willing to grant, at least for the sake of argument, that observation is a way of knowing about matters of fact. But remember, we started with some examples concerning matters of fact that we haven't yet observed.

Remember? Concerning what will happen in the future, say when the next American presidential election will be, or concerning what animals are like that I, at least, have never directly observed. So how can we know about these matters of fact, if not by observation and also not by the mere operation of thought?

Here's a general strategy that Hume thinks we often use. We project observed regularities, things that have been true in those cases that we have observed, onto unobserved cases. So, for example, all the fires I've observed have been hot.

Whenever I've been exposed to a fire, have gotten a chance to, say, put my hands near it, it's been hot. So I assume the fires that I haven't observed will be hot, too. More generally, we start with some premise that says "all observed Fs have been G" where "F" might be "fire" and "G" might be "heat," but it could be something else, too.

And we draw the conclusion that all unobserved Fs are Gs too. We'll call this pattern of inference "induction. " I think we've gone a bit too quick.

This can't be quite right. We don't always use induction. For instance, here's an example of induction as I've described it.

Every hair of mine that I've observed is black, so in the future all my hairs will continue to be black. Now, I would love it if this were a good, persuasive argument. But it's not, given what's held true for my parents and my aunts and uncles and my grandparents and humanity as a whole.

I have excellent reason to think that at some point in the future some of my hairs will be gray. But even if it were true that every hair of mine that I've observed is black, still that wouldn't give me reason to think that in the future all of my hairs will continue to be black. So we don't always think that inductive arguments, arguments that use induction, are good ones.

I think it would be too quick to dismiss induction as an interesting and important sort of argument on the basis of examples like the one involving my hair. I think we do often rely on something like induction, at least when we're making inferences about very general features of our environment. So in the case of fires being hot, or say in the case of gravity continuing to operate, say, the sun rising tomorrow as it's risen every day in the past, we do take the inductive arguments to be good ones.

So Hume thinks, and I'm inclined to agree, that we're implicitly assuming, when we make arguments like this, something like what's been called the "uniformity of nature. " We're implicitly assuming that the future will resemble the past, at least in its most general respects. Maybe the future won't resemble the past in the respect of my continuing to have black hairs, but it will resemble the past in respect of fire continuing to be hot.

That's more general. It will resemble the past in that people will continue to die, people will be mortal, and in that gravity will continue to operate, all these very general respects in which the future might resemble the past or ways that we think it will. These are cases where we think inductive arguments are good.

You might think we don't need to use induction to come to views about what's gonna happen in unobserved cases, because we can appeal to laws of nature, we can appeal to science, which tells us what the laws of nature are and can let us make predictions about what's going to happen in cases that we haven't yet observed. But Hume thought, and I'm inclined to agree, that really appealing to laws of nature, like say the law of gravity, is just a matter of appealing to inductive arguments. When we appeal to laws of nature, we are at least implicitly assuming that these most general regularities that have held before will continue to hold.

When we say that the Earth will continue to revolve around the sun because gravity says it must, we're implicitly assuming that this regularity that's held in the past concerning how massive bodies interact will continue to hold in the future. That is, we're implicitly assuming that the uniformity of nature, that the future will resemble the past, will continue to hold. Why should we believe that?

Why should we believe that the future will resemble the past, even in its most general respects? It's not a relation of ideas. It's conceivable that the future should fail to resemble the past, even in its most general respects.

Try to imagine that tomorrow the sun doesn't rise, that gravity stops operating. Try to imagine that if you stick your hand into a fire tomorrow, it won't burn but will instead be cold. This is the stuff of science fiction, but it is imaginable.

There's no incoherence in supposing that while some law of nature has held true in all our observation in the past, it will fail to hold true in the future. I imagine you could write some good stories, consistent stories, stories that make sense, premised on the idea that, say, gravity or the laws of chemistry that have been true in the past might fail to be true in the future. So the mere operation of thought isn't going to be enough to convince us that the uniformity of nature holds.

The mere operation of thought, uncovering contradictions, using logic, that's not going to show us that the future must be like the past. Here's another way that we might try and argue that the future will be like the past. We might try to use induction.

After all, we saw that when we're dealing with matters of fact, we generally can't use the mere operation of thought, and when we're dealing with matters of fact that we can't directly observe, we need to use induction. So maybe in arguing that the uniformity of nature will continue to hold, we can use induction. What might that look like?

Here's an argument, an inductive argument. In the past, the future has resembled the past. So in the future, the future will resemble the past.

This looks to be circular. It assumes that because something happened in the past, it will continue in the future. It would only be a good form of argument if we already thought that the future would resemble the past.

Otherwise, the mere fact that something has happened in the past wouldn't be something that we took to suggest at all that in the future, the future will resemble the past. How bad is this? How much should it worry us?

I find it worrying me. We ordinarily think that relying on induction is a good way to form beliefs about the future, it's a good way to form beliefs about what will happen in unobserved cases. In particular, it's a better way than other methods that we might use, say tea leaf reading or astrology or consulting a Magic 8 Ball.

So suppose I'm wondering whether I can fly. I'd really like to fly, and I'm considering jumping out of a tenth story window in the hopes that I'll be able to fly when I do. Will that succeed?

Will I fly or will I die? If we trust induction, induction says I'll die. When things get thrown out of tenth story windows, they fall, at least when people do.

But suppose I take my Magic 8 Ball, shake it up, ask it whether I can fly, and it says, "Without a doubt. " I think most of us would think I should rely on the prediction that induction gives me rather than the prediction that the Magic 8 Ball gives me. Why should I do that?

What's better about induction? We've already shown that we can give a circular argument in favor of induction. But we can do the same thing with the Magic 8 Ball.

Suppose I ask the Magic 8 Ball, "Will you tell me the truth? "Are you a good way for me to form beliefs about the future? " I shake it up and it says, "Without a doubt.

" So when I ask the Magic 8 Ball whether it always tells the truth, it says that it does. When I ask induction whether induction will give me reliable beliefs about the future, it says that it does, too. In both cases, I can give a circular, question-begging argument.

I can assume that some method is a reliable way of forming beliefs about the unobserved, and then, using that method, the method will say, "Good job, stick with me. " But that doesn't distinguish between induction and trusting the Magic 8 Ball. So do we have any reason to think that induction will lead to true beliefs and trusting the Magic 8 Ball won't?

Here's one interpretation of what Hume said, Hume's skeptical solution to the problem of induction. There's no rationally compelling reason to use induction rather than crystal ball gazing or astrology or relying on a Magic 8 Ball. Still, we just can't help but reason inductively.

It's a strong habit, much like birds can't resist flying south in the winter. It's not rational, it's just something we do out of instinct. Because this instinct is so strong, the question of justifying induction doesn't really arise, at least not in practice.

We'll keep using it, whatever philosophical conclusions we come to. So we can set this question aside, recognizing that we're behaving irrationally, but that we're just not able to bring ourselves to behave rationally. Now I can't quite bring myself to accept this position.

And I'm also skeptical that it's really the right interpretation of Hume. But it's very hard to say what's wrong with our Hume-inspired argument, and lots of subtle and interesting philosophy has been done that attempts to do just that. If you're intrigued, that's a topic for another video.

Subtitles by the Amara.