David Hume on Induction

The following is an essay I turned in for my British Empiricism class. It is on Hume's Problem of Induction. 

Deductive and inductive reasoning are types of argument structures. In a deductive argument, the truth of the premises guarantees the truth of the conclusion. For example, if 1) All cats are mammals 2) Felix is a cat, then it would follow inescapably that 3) Felix is a mammal. If 1 and 2 are true, there is no possible or conceivable world in which 3 is not true. Inductive arguments do not provide this guarantee. A good inductive argument only makes the conclusion more probable or likely than not. For example, if 1) 90% of people can swim, and 2) Michael is a person, then 3) Michael can swim, though not guaranteed, is likely. It does not follow with logical necessity because Michael may just happen to be part of that 10% of the population that does not swim. 

David Hume’s problem of induction states that inductive justifications that rely on past experiences will fail to yield genuine knowledge about future events. The belief that past events will resemble future events, and this reasoned to inductively, is called the Principle of Uniformity. It states that there is a kind of regularity to nature such that similar causes will produce similar effects under similar circumstances. If, for example, I experience that drinking water quenches my thirst, then I inductively reason that in the future, water will quench my thirst under similar circumstances, similar circumstances being something like the water is similarly cold and has no salt. If my experience tells me there is a constant conjunction between my drinking water and my being quenched, then I say I am rationally justified in believing this will be so in future cases for causal reasons. 

Hume, however, will say No, I am not justified in believing this. There are two ways to justify my reasoning for the Principle of Uniformity, either by it being true in virtue of it being a relation of ideas, or it being true as a matter of fact. Let’s consider whether it is true in virtue of it being a relation of ideas. Relations of ideas are true in virtue of the ideas being related to each other. So for example, a triangle has three sides is true in virtue of the idea of “triangle” and “having three sides” being related or necessarily connected to each other. If the Principle of Uniformity is supposed to be true because it is a relation of idea, then we should be able to deduce a contradiction out of its contrary, just like we are able to deduce a contradiction out of “A triangle does not have three sides”. So what would that contradiction be concerning induction? For Hume, this seems impossible. There doesn’t seem to be any contradiction in the statement, “Similar causes will not produce similar effects under similar circumstances.” After all, can’t I imagine that I will drink water some time in the future and I will not be quenched? And if I can conceive it, it certainly must be at least possible. But if it is possible, it is not deductive or necessary in a way that relations of ideas are. So deduction/relation of ideas cannot justify the Principle of Uniformity. 

Now let’s consider whether the Principle of Uniformity is justified by matters of fact. Matters of fact are statements which happen to be true, but aren’t necessarily true. They are true contingently. That Michael is a swimmer isn’t true in virtue of the ideas of “Michael” and “swimmer” but would be true because Michael just happens to know how to swim because he took lessons. In order to justify the Principle of Uniformity as a matter of fact, I would have to show that it happens to be true. But what does that look like? Well, presumably, I would have to look at past events, that this person knows how to swim and so does that person, and from that fact know that this person in the future, Michael, will know how to swim. But that is precisely the thing I am trying to prove. I cannot use induction to justify induction. That would be circular. 

Summed up, Hume’s argument states that since the Principle of Uniformity cannot be justified by either deduction/relation of ideas or induction/matters of fact, and those are the only two possible modes of justification, then the Principle of Uniformity cannot be justified, and so all inductive reasonings are also unjustified. 

If there is no justification for induction, what at least explains our instances of doing so? For Hume, it is custom and instinct. By custom, he means a pragmatic tool to make experience useful. Without custom, one would be in a kind of paralysis of how to adjust means to ends, or how to use practical reason. Further, any kind of speculation would have to be suspended. But since these are absurd practical consequences, we should be okay with knowing we have no rational basis for induction. Hume also says that acts of induction are produced not by reason but rather by natural instinct. He likens this to feeling loved when something good happens to us, or feeling hated when something bad happens to us. However, it is not a passion. It is an aspect of human nature which interacts with the external world, and since we need to survive, this externally interacting aspect of human nature has thus adapted to survival. We may not know exactly how it does this, but we know that it nonetheless does, just like someone may not know exactly how their car works, but nonetheless knows that the car does work. 

I find Hume’s argument to be wanting. It seems to me that he cannot satisfy his own criteria. If all knowledge is supposed to be true in virtue of it being a relation of idea or matter of fact, then it is difficult to see whether that proposition is supposed to be true because it is a relation of an idea or matter of fact. What in the relation of the ideas “relation of ideas” or “matters of fact” makes it true? I don’t think it does. Further, if we consider that deduction and induction don’t exhaust the types of reasonings we engage in, then it presents us with a false choice. Abduction is an alternative form of reasoning that we can engage in. 

Issue is also to be taken up with the notion that what is conceivable is possible. Like other philosophers of the period, this seems like a premature attempt to bridge the gap between epistemology and ontology. I could be mistaken about what I take to be imaginable. A child may not properly conceive of a math answer, but it doesn’t follow that the answer is not possible. Perhaps Hume can reply that it must at least be possible in principle to conceive it in order for it to be possible, but how can we know what that is without being circular? How can one conceive of a conception? What are conceptions themselves supposed to resemble or be copied from? 

Another issue that can be poked at is that the framing is incorrect. Hume wants to break the connections between causes and effects as a logical connection. But this is in no way uncontroversial. Cause and effect should not be looked at through the eyes of a logician, but a metaphysician. A metaphysician looks at and considers being first, and is the ground for logic and logical rules like the law of non-contradiction, whereas a logician looks at forms without matter or potency, an inappropriate first principle for Aristotelian-Thomists which were still commonplace in Hume’s day. When one comes to the world through the hylemorphic view, with their particular view of substances, causes and effects are pretty common sense. 

Finally, Hume states that even though induction is unjustified, it is still reliable. Without it, we would be ignorant of things that are matters of facts, which implies he believes we do have knowledge matters of fact. He further states that it is necessary that the mind does this and our beliefs properly correspond to the outside world. But why think that the need to survive corresponds to true beliefs? That connection needs justification. For example, I may be, in reality, a rabbit who runs away at the smell of wolves, and this corresponds to the rabbit’s imagination that it is a human person who runs away at the sight of zombies. While there is a correspondence between the two, and perhaps a causal link (the smell of wolves triggers images of zombies) it by no means produces true beliefs. So in what way is this said to be reliable, know matters of facts, and have a necessary corresponding knowledge of the external world? This only seems true if our mental faculties are supposed to be reliable in the first place, which Hume denies.