I am currently participating in a reading group studying Kant's Critique of Pure Reason. A lot of the conversations have shown that there are a number of topics for beginning readers that need to be clarified in order to properly understand the book, and Kant's project generally. I have found 'necessity' to be one of the most important of these, since it is so closely bound up with the purely a priori.
I am hoping for this post to be very simple and short (at least considering the time it takes me to write). The main intention of it this is to discuss, positively, how 'necessity' functions in Kant's work. If there is feedback, I can help to address further questions and clarify.
Necessity in Kant's Critique of Pure Reason
It makes sense to give Kant's formula of the transcendental principle (more specifically, the postulate of empirical thought) called 'necessity'. In the Critique of Pure Reason (A218, B266) Kant writes:
That which in its connection with the actual is determined in accordance with universal conditions of experience is (that is, exists as) necessary.
This sort of technical formula will not be helpful until it has been properly illustrated, and so I will provide an example which can assist in the exposition.
Either imagine an object falling, or actually drop an object and watch it fall; you are witnessing a necessary connection.
Now, this falling object was experienced as just that - something that fell. It isn't necessary that it fell, it isn't necessary what fell, it isn't necessary how far it fell, it isn't even necessary that it did fall, nor that it could fall again. So what does the necessity relate to?
Considering the formula provided by Kant above, the necessity deals with 1) the connection in the actual, and 2) this connection being determined in accordance with universal conditions of experience; I will discuss these two points in reference to the example.
1) By the actual, we intend here there were actual representations (sensible impression) of the object (whatever fell). Further, there were multiple representations of the object as it fell. The connection in the actual means that these different representations were related to each other as pertaining to a single object. So, the connection in the actual means the relation of the representations of the falling object.
2) By universal conditions of experience we intend laws that are part of every experience, and so the determination of the connection (in the actual) is by laws which apply to every experience. Some may doubt that we can prove such laws, and this is due to a misunderstanding of what is being attempted.
A frequent way of misunderstanding Kant comes along with the following objection: when we find a rule in our experience, it is always due to some regularity that we discover; these regularities - being empirical - can never be known necessarily. The response Kant has for this is to grant the whole argument; it is perfectly accurate to say that we acquire empirical rules from regularities in experience, however, we are granting that there are such things as regularities. The universal conditions of experience are general characterizations of regularity in general - not any particular regularity.
Considering our example, objects that fall are not in every experience, and so 'falling' is not a universal condition of experience, however, a) that we can relate the multiple representations to the same object, and b) that the order of these representations is able to have a significance, are not due to generalizations on witnessing objects fall (or doing anything particular). That is, for a), there is nothing about the representations that show that they are of the same object, and if we did not have the capacity to relate representations, then we would never have objects that take up more than one moment of perception; and, for b), if we could not judge that there was significance in the temporal order of two representations, then there would be no significance to an object falling that would be distinct from an object rising: for an object falling, or rising, we have the same representations, only the order is significant; if we couldn't recognize significance in an order, then we couldn't say something was occurring (falling being a kind of occurrence). The former of these connections Kant refers to as the category of Substance, the later the category of Causality.
Here when we see the falling object, the necessary judgments involved are that the representations relate to the same object, and that the representations are in a definite sequence. This is the same as saying that the actual (representations) are thought under conditions that apply in every experience (universal conditions of experience), which means conditions required in every experience. Here these conditions were that an object experienced over time involves representations related to each other both as a single thing, and in a certain sequence. If something can't be shown to be a part of every experience, and as a form of that experience generally (pertaining to regularity in the broadest sense), then it is also not necessary.
Anything that is empirical is automatically excluded from the necessary, since it is the empirical that is made possible (for us to observe) in an experience that has regularities. If experience had no regularity, generally, then observation itself would be impossible.
If we review Kant's definition again, it should make more sense: "That which in its connection with the actual is determined in accordance with universal conditions of experience is (that is, exists as) necessary."
An additional note can be made here that offers a great deal of added clarity: the categories of Modality (possibility, existence, necessity) do not involve determinations of objects as they are in themselves, but merely different ways in which objects are thought in relation to our way of knowing them. So, by saying we know something be necessity, we are saying that we know it by virtue of (and as) a general sort of pure regularity of experience. We can not claim that we know whatever judgment we made was necessary to make (such as, recognizing something falling), but rather that in experiencing (something falling) there are a regularity in it that could not first be learned from experiences, since experiences already have some degree of regularity.