One of Hume’s most famous positions, defended in both A Treatise of Human Nature and An Enquiry Concerning Human Understanding, is that we cannot use reason to gain knowledge of any causal relation or any regularity in the world. Hume argues that causation — at least insofar as we can know it — is merely a habit of the human mind, not a concept we can ascribe to nature through reasoning. He also argues that there is no such thing as inductive reasoning; that is, Hume argues that claims that a pattern we have observed will also hold across the parts of the world that are so far unobserved may be subjectively unavoidable, but that they are never the result of a reasonable argument. The connection between these two ideas is evidently very strong.

Now Kant technically agrees with Hume that reason cannot give us causal and inductive knowledge, but this is indeed only a technicality. Unlike Hume, Kant distinguishes three faculties of human thought: the understanding (Verstand), reason (Vernunft), and the power of judgement (Urteilskraft). And strictly speaking it is the understanding that gives us causal knowledge.

Let’s discuss the distinction between these three powers. The understanding is the faculty of concepts. Unlike sensibility, in which individual objects are given to us, the understanding uses general rules. Indeed, for Kant any concept is a general rule. The concept of ‘cow’, for instance, is a rule for classifying things into the group of cows and the group of non-cows. Sensibility may give us a direct connection to a concrete individual; but this connection is useless for thought until we apply concepts to it; and these conceptual rules are grasped by the understanding.

By the time of the Third Critique, the Critique of the Power of Judgement, Kant had started to make a clear distinction between the understanding and the power of judgement, which was far from clear in the Critique of Pure Reason. With this new distinction in mind, we can say the following. When I come across an object and classify it as a cow, it is the power of judgement that is active; because the power of judgement, in its determining use (more about that later), is simply the faculty for classifying individuals under general rules. The role of the understanding is to connect concepts together into judgements, e.g., in the judgement that all cows are animals, or the judgement that no cows are divine, or the judgement that if something is a cow, it is not a dog. The role of the power of judgement is to judge that that thing, the thing I’m seeing right now, is a cow.

What is the role of reason? Reason is the faculty that seeks to grasp the totality of the judgements of the understanding. Here’s an example I like to use. About six months ago, my daughter was at a stage where she could count indefinitely. If you gave her a number, she could tell you the number that was one more, and she could continue this more or less without limit. But my son, who is two years older, was at the stage where he could then overlook that activity and exclaim that it is an infinite activity; that it will never end; that there are infinitely many numbers. My daughter was using her understanding, while my son was using what Kant calls his reason.

In the Critique of Pure Reason, Kant rather famously argues that reason cannot give is us a priori knowledge of the world. When it claims that time or space are finite, or infinite; when it claims that matter is or is not infinitely divisible; when it claims that there is or is not any freedom; when it claims that the soul is or is not a substance; when it claims that God exists or does not exist; then it always going beyond the boundaries of what it is allowed to say. Less well known is Kant’s positive story about reason in the part of the book called the Doctrine of Method. Here Kant argues that reason plays an essential methodological role in pushing us to seek for completeness of our understanding of the world. Thus, it is reason that pushes us to seek for missing elements in our classifications, for laws that are as general as possible, and so on. Without reason, we would be able to think about and draw conclusions about the things we experience, but we would not be pushed in the direction of truly systematic inquiry.

The way I currently understand Kant’s development — but it will be interesting for me to learn more about this from the secondary literature — is that some of these methodological roles of reason are shifted to the power of judgement. And we can see why. There is a fundamental difference between two things that reason is supposed to do in the First Critique. When the understanding posits that every event has a cause, reason asks: what about the totality of causes? Is it finite or infinite? That is one thing. It is, as Kant says, to seek the unconditioned in a sequence of conditions. When the understanding has given us several laws of nature, and then to ask: isn’t there some more general law under which all of these fall? That would seem to be something else. The second type of question is typical for a type of question that assumes that nature is simple, or elegant, or purposeful, or fit for our understanding. And there is nothing in the Kantian account of reason that really brings us to those kinds of terms. So it’s perhaps not too surprising that in the Second Critique, reason turns out to find its own in practical affairs — that is, moral affairs, not affairs having to do with practicalities — and in the Third Critique it is the power of judgement that gets connected with notions of simplicity, elegance, and purpose. When physicist Steven Weinberg tells us that “we would not accept any theory as final unless it were beautiful,” he is saying something that belongs not in the First, but in the Third Critique.

So how does the power of judgement get connected to these notions? In its determining use, the power of judgement classifies individuals under more abstract concepts. But in its reflective use, Kant says, it does what is in a sense the opposite: given an individual, it seeks for some abstract concept under which it could be classified. Of course this is the same power: it is the power to connect the individual with the abstract. But in the first case, the abstract is given, and in the second case it is lacking. And so when we are confronted with the massive multiplicity that is experience, it is the power of judgement that pushes us to seek a simplicity behind the multiplicity, and, as Kant says, to proceed as if the world were designed by a divine understanding. Why? Because we have no alternative. This is the only possible way to make sense of things. It is subjectively necessary.

This is very different from the transcendental truths we found in the First Critique. The understanding has a constitutive role with respect to the world of appearances. The world with which we, as finite beings, engage has to be such that every event has a cause, that all events stand in thoroughgoing connection, that there is permanent substance, that everything has quantity and quality, and so on. We know this a priori. These are objective necessities. By contrast, we do not know that the world will be simple and purposeful. The power of judgement is merely regulative with respect to itself, rather than constitutive with respect to the world. We must assume that Nature is purposeful and simple, but we cannot know this to be the case, and we have no idea how far our investigations will bear fruit. (Though perhaps we are never free to desist from them, having always to hope that they can be pushed further? We would not accept any theory as final, unless…)

All of this turns out to be linked to an extremely important objection against Kant’s criticism of Hume. According to Hume, we cannot reasonably conclude to the truth of any regularity or necessary connection. Kant argues quite forcefully in the Critique of Pure Reason that all events in the world fall under universal causal laws; that it is indeed constitutive of being an event in the world that it falls under a universal causal law, since this is constitutive of having a place in an objective time order. Let’s suppose that he is right. Then there is an important sense in which Hume is wrong about the nature and role of causation. But what isn’t clear at all is that Hume is wrong about any particular judgement about a causal relationship or an inductive pattern. For nothing in Kant’s theory tells us which events are causally related and what the universal rules actually are. Nothing even guarantees that they are simple and few enough that we could gain any knowledge of them. Kant says nothing about this in the First Critique, but here he is in the introduction to the Critique of the Power of Judgement:

For it may certainly be thought that, in spite of all the uniformity of things in nature in accordance with the universal laws, without which the form of an experiential cognition in general would not obtain at all, the specific diversity of the empirical laws of nature together with their effects could nevertheless be so great that it would be impossible for our understanding to discover in them an order that we can grasp, to divide its products into genera and species in order to use the principles for the explanation and the understanding of one for the explanation and comprehension of the other as well, and to make an interconnected experience out of material that is for us so confused (strictly speaking, only infinitely manifold and not fitted for our power of comprehension).

Kant, Critique of the Power of Judgement, 5:185

So it turns out that Kant agrees that while the First Critique shows that that which we are after in inductive causal reasoning (causal regularity) exists, it has given us no guarantee that we will actually be able to succeed at such reasoning. And therefore we need an additional a priori principle, even if it is only subjectively and not objectively necessary, which is given to us by the power of judgement. And this principle tells us to investigate the world as if it were the product of an intelligence; and thus, to seek for the simplest, for the most comprehensive, for the systematic, for the complete, for, as Weinberg says but Kant perhaps does not, the beautiful.

What is, at this level of discussion, the difference with Hume? Hume tells us, as a psychological fact about human nature, that we are so constituted that we always seek simple patterns. This is the force of habit. Kant denies that his story is in any sense psychological. The point is not the we happen to seek simple patterns. The point is that finite beings such as we have to seek simple patterns (and completeness, and so on), if the experience of Nature is to be so much as possible. We know a priori that being structured by causal laws is constitutive of Nature. Now unless we seek simple patterns and trust that we will be able to find them, we cannot synthesise the confusing multiplicity of sensation into a structure of causal laws; and thus, no such thing as experience (in the strict sense) would even be possible. We have no guarantee of success. But pursuing the project of making inductive sense of the world is part and parcel of what it is to experience a world in the first place.

That’s a far cry from Hume’s inductive scepticism. What’s more, I think it’s true.