Metaphysical Deduction

What then is a concept for Kant? As opposed to an intuition that is given to the faculty of representation, a concept is an idea or thought that is applied to appearance. The faculty that corresponds to concepts is the understanding.We cannot know anything without the combination of both intuitions and concepts.This, Kant would say, is something that we have to accept, and there is no philosophical explanation why this is so. Unlike God, who knows things in themselves, as we might suppose (what Kant calls ‘intellectual intuition’), we can only know things through the concepts we have of them. Human cognition, Kant says, is ‘discursive (A68/B93). We never just see the tree in our backyard, as it is in itself, rather we see the tree as a ‘tree’, that is to say through the concept ‘tree’.But what is this concept?Like with intuition Kant makes a distinction between pure and empirical concepts. The concept ‘tree’ is an empirical concept. Although the concept itself comes from our own understanding, the material of this concept must come from experience. But a concept is not merely descriptive for Kant; rather he says that all concepts have a ‘function’ (A68/B93). What is the function of a concept? It unifies (or synthesises, which means the same thing for Kant). Thus the empirical concepts organise, shapes or unifies are experience of the world. It is quite wrong therefore to think that I have an immediate perception of a tree, and then I simply apply the concept ‘tree’ to it; rather to be able to see a tree at all I must also have the concept alongside my intuitions. As Sebastian Gardner points out, I do not just have an mass of red sensations, rather I see a red patch as an organised unity, and this unity does not come from sensation itself, for all that sensation gives me is simply this ‘array’ of red sensations.^[1] Kant, however, does not believe that all that the understanding contains are empirical concepts, rather, like with intuitions, there are also pure concepts. Unlike with intuitions, however, these pure concepts are not simply given, rather we have to arrive at them through a process of analysis so that we end up with the most generalised concepts.This is why this section of The Critique of Pure Reason is called the ‘transcendental analytic’. 

Henry Allison demonstrates to us what such an analysis might be.Take the example, he suggests, of the judgement ‘Socrates is a man’.^[2] This judgement about a singular concept, ‘Socrates’ but through analysis of the judgement we can see that it is based on a more general concept ‘man’ and so on as we go back to more and more general concepts. For Kant, however, this process cannot go back forever.We will eventually reach the most generalised concepts, such as ‘entity, property, individual, class, and totality’ that cannot be broken down any further. They are the most fundamental grammar of our judgements about objects. 

Now it is these most general concepts, which Kant calls logical forms and which are the clue for Kant of what he will latter call the categories.In other words Kant is beginning with the most general way in which we talk about things, and from that he is deducing the general categories of objects.It is this deduction that he calls the metaphysical deduction, and which must be distinguished from the transcendental deduction that we will be looking at next week. The key to make sense of this mysterious faculty that Kant calls the understanding is therefore our judgements – what are say about objects. In the most general way, Kant believes, that our judgements about objects can be exhaustively described in 3 possible kinds with for 4 possible elements, which he lays out in ‘the table of judgements’ (A70/B95) and every judgement we use can be shown through analysis to contain these 3 kinds.^[3]

But this is only the clue for pure concepts of the understanding, for we must make a clear distinction between what Kant calls general logic and transcendental logic General logic merely describes the pure form of logical judgement and the rules of thought. It does not demonstrate how these are related to objects. The latter is the function of transcendental logic. The hypothesis of transcendental idealism, the famous Copernican revolution, is that objects must conform to our form of knowledge, and not the other way around.In other words that mode of presentation of the object is determined by our mode of cognitionNow Kant has already argued that human cognition is made of two parts: intuitions and concepts.Thus, he needs to show how concepts themselves determine the object a p.Thus he needs to show not just the logic form of our pure concepts, but show how these pure concepts themselves determine the mode of presentation of objects. It is this that is the movement from the table of judgement to the table of categories. 

The best way to see how this works is to use an example: the category of substance, Kant argues, is deduced from the logical form of categorical judgement. The best way of understanding this is that we can talk about a pure concepts either from the side of the logical judgement or from the side of the picture of the object that it gives us. Thus, categorical judgement always takes the form x is F, as for example Socrates is a man. But when this logical form is applied to an object, we arrive with category of substance. There is an x in which the property ‘man’ inheres. It is very important for Kant to show that substance is a category, and not something ontological, that is to say real. In other words, the category of substance organises our representations, it is itself not a representation.We must be clear that is also not a logical proof. Rather the logic forms are a clue to the categorical forms, and both have their common source in the understanding. In the same way that we have gone from the logical categorical function of judgement, to the category of substance, we can go through all the other parts of the table of judgement, to produce the table of categories (A80/B106). 

Most commentators, however, think that the metaphysical deduction fails to set out to do what it proposes, namely proof conclusive a systematic list of the pure concepts of the understanding. The problem has to do with the table of judgement, which Kant takes for granted to be exhaustive. He does not prove this however, but simply takes these logical functions from tradition. The problem with the table of judgement, however, does not invalidate Kant’s major hypothesis that our experience of the world must be constituted by a priori concepts, rather he needs a better proof of what these concepts might be, rather than just a list of logical functions that he simply accepts are given.^[4] This second proof is given in the transcendental deduction that we shall discuss next week.