Let us consider first the sources of metaphysical knowledge. The very concept of metaphysics ensures that the sources of metaphysics can’t be empirical. If something could be known through the senses, that would automatically show that it doesn’t belong to metaphysics; that’s an upshot of the meaning of the word ‘metaphysics’. Its basic propositions can never be taken from experience, nor can its basic concepts; for it is not to be physical but metaphysical knowledge, so it must lie beyond experience.

Outer experience is the source of physics properly so-called, and inner experience is the basis for empirical psychology; and metaphysical knowledge can’t come from either of these.

It is thus knowledge a priori—knowledge based on pure understanding and pure reason.

[...]

[W]e cannot rightly start by asking whether synthetic a priori propositions are possible. For there are plenty of them, really given to us with undisputed certainty; and as our present procedure involves starting with what we already know, we shall start from the premise that there is human a priori knowledge of some synthetic propositions. But then we still have to ask how this knowledge is possible, i.e. what makes it possible. When we know this, we can learn how to use such knowledge and can learn what its limits are.

Stated precisely, then, the crucial question is this:

How is it possible to have a priori knowledge of synthetic propositions?

[...]

Metaphysics stands or falls with the solution to this problem. Someone may propound his metaphysical claims as plausibly as he likes, smothering us with conclusions piled on conclusions; but if he hasn’t first answered this question properly, we are entitled to say to him:

‘This is all pointless ungrounded philosophy and false “wisdom”. You purport to be using pure reason to create a priori knowledge, not by merely analysing concepts but by making new connections that don’t rest on the law of contradiction; and you think you have insight into these connections independently of
all experience. But how do you get such insight? How can you justify your claims?’

He can’t answer by appealing to the common sense of mankind, for that isn’t evidence—it’s mere hearsay. . . .

[...]

[O]ur main problem splits into four questions, which will be answered one by one:

(1) How is pure mathematics possible?
(2) How is pure natural science possible?
(3) How is metaphysics possible in general?
(4) How is metaphysics possible as a science?

[...]

(2) How is pure natural science possible?

Nature is the existence of things insofar as it is governed by universal causal laws. If this meant the existence of things in themselves, we couldn’t know nature either a priori or a
posteriori.

One way of knowing things a priori is knowing
them through the analysis of concepts. We couldn’t know nature as it is in itself in that way, because knowledge of what things are like in themselves can never come from analytically dissecting our concepts: we aren’t asking what is contained in our concept of the thing, but rather about what is added to this concept in the reality of the thing itself.

Some synthetic propositions can be known a priori because their truth is assured by the nature of our understanding, somewhat in the way that mathematical truths can be known
a priori because our sensibility assures their truth. But this is also not applicable to the supposed ‘knowledge of nature as it is in itself’, which we are discussing. My understanding has an effect on how things appear to me, but it can’t dictate what things are like in themselves. They don’t have to conform to it; so if I am to know about things in themselves, my understanding must conform to them, not vice versa.

That means that I couldn’t know about them until they had somehow been presented to me; which is to say that I couldn’t know them a priori.

Nor could I have a posteriori knowledge—i.e. knowledge through experience—of the nature of things in themselves.

If I am to bring things under causal laws, these laws must apply to them necessarily, and experience could never show me how things must be—only what there is and how it is. So it can never teach me the nature of things in themselves.
— Kant

(1) How is pure mathematics possible?
(2) How is pure natural science possible?
(3) How is metaphysics possible in general?
(4) How is metaphysics possible as a science? — Kant

Not sure what you are looking for. You probably need to focus the thread more for people to get involved. — shmik

If such questions are appropriate in metaphysics, then it seems completely wrong to exclude empiricism, that which is known through the senses, that which is taken from experience, and that which is physical. We can still question whether anything exists beyond the aforementioned, and if so, how much we can know about it - without making such an exclusion. — Sapientia

When we ask the constitutional question of how objects are disclosed to us, then any object, including any scientific object, must be regarded in its correlation to the mental activity that intends it. This transcendental orientation in no way denies the existence of a real physical world, but rather rejects an objectivist conception of our relation to it. The world is never given to us as a brute fact detachable from our conceptual framework. Rather, it shows up in all the describable ways it does thanks to the structure of our subjectivity and our intentional activities. — Evan Thomson, Mind in Life

But wouldn't we then say that we're doing natural science when we go by empirical objects etc., but when we ask, e.g., about what is beyond them--for which there can be no evidence, for if there were we would just be doing science again--then we are doing metaphysics? — jamalrob

I could comment on 2)3)&4) but I don't think they are possible in the way that Kant thinks they are. Not sure what you are looking for. You probably need to focus the thread more for people to get involved. — shmik

If you want an extremely reductive dismissal of his work, here is how I feel about it: it all consists of taking analytical prejudices and claiming that they're necessary. This is so insofar as the 'common wisdom' of the age is taken to be something that arose necessarily, and so the prejudices that underly it have to be traced backward and declared the ultimate source of all things (since the way we see things now is therefore the way they must always be seen). In short, Kant is the 'mythologizer' par excellence. His business was creating etiological myths about how the world as he knew it came to be, like the story of Jacob and Esau and the red soup, and how what must have happened in order to reach that state therefore had to have a divine or permanent status.

I think there is really little else to say about him. — The Great Whatever

You might be right, but I don't know what exactly you're referring to. Some examples would help. What analytical prejudices? And where does he claim that they're necessary? What's this myth that he allegedly created about how the world as he knew it came to be? — Sapientia

• http://www.uri.edu/personal/szunjic/philos/proleg.htm
• http://sparknotes.com/philosophy/prolegomena/

He indisputably proved that it is wholly impossible for reason to think such a connection a priori and from concepts, because this connection contains necessity; and it is simply not to be seen how it could be, that because something is, something else necessarily must also be, and therefore how the concept of such a connection could be introduced a priori. From this he concluded that reason completely and fully deceives herself with this concept, falsely taking it for her own child, when it is really nothing but a bastard of the imagination, which, impregnated by experience, and having brought certain representations under the law of association, passes off the resulting subjective necessity (i.e., habit) for an objective necessity (from insight). From which he concluded that reason has no power at all to think such connections, not even merely in general, because its concepts would then be bare fictions, and all of its cognitions allegedly established a priori would be nothing but falsely marked ordinary experiences; which is so much as to say that there is no metaphysics at all, and cannot be any.

Metaphysics is often summed up, as it has been on this site, as asking "big questions", regarding things like truth and reality, and including questions such as "What exists?" and "Are objects constituted by the way we see and describe the world?".

If such questions are appropriate in metaphysics, then it seems completely wrong to exclude empiricism, that which is known through the senses, that which is taken from experience, and that which is physical. We can still question whether anything exists beyond the aforementioned, and if so, how much we can know about it - without making such an exclusion. — Sapientia

... there can be no such science unless the requirements expressed here, on which its possibility rests, are met, and, as this has never yet been done, that there is as yet no metaphysics at all — Kant

Since, however, the demand for it can never be exhausted, because the interest of human reason in general is much too intimately interwoven with it, the reader will admit that a complete reform or rather a rebirth of metaphysics, according to a plan completely unknown before now, is inevitably approaching, however much it may be resisted in the meantime

But I fear that the elaboration of the Humean problem in its greatest possible amplification (namely, the Critique of Pure Reason) may well fare just as the problem itself fared when it was first posed

. . .

with regard to a certain obscurity -- arising in part from the expansiveness of the plan, which makes it difficult to survey the main points upon which the investigation depends -- in this respect the complaint is just; and I will redress it through the present Prolegomena.
The previous work, which presents the faculty of pure reason in its entire extent and boundaries, thereby always remains the foundation to which the Prolegomena refer only as preparatory exercises; for this Critique must stand forth as science, systematic and complete to its smallest parts, before one can think of permitting metaphysics to come forward, or even of forming only a distant hope for metaphysics

whosoever undertakes to judge or indeed to construct a metaphysics, must thoroughly satisfy the challenge made here, whether it happens that they accept my solution, or fundamentally reject it and replace it with another

Something that piqued my interest later is when Kant distinguishes between methods to draw the distinction between the Prolegomena and the Critique -- namely, the Prolegomena follows the analytic method, and the Critique follows the synthetic method. Kant insists that the synthetic method is necessary to present all the articulations, whereas the analytic method is good enough -- after accepting the deduction (which, back then, was more akin to legal justification than dedeuctive logical inference) -- for giving the plan in broad strokes. — Moliere

...the synthetic a priori, in short, the incontrovertibly true and valid modes of knowledge that far surpass mere logic, may be described as the roast, the Leibnizian or Cartesian roast, while Hume and English scepticism provide the dialectical salt. — Theodor Adorno

• Analytic: explicative judgements, ones that don’t tell us anything new but just bring out what’s already there.
• Synthetic: ampliative judgements, ones that are informative, i.e., they add something new.

A judgement and its negation are both analytic if and only if one of the pair is self-contradictory, or false by virtue of the definitions of words or its logical form. — Jill Vance Buroker

                  analytic            synthetic
a priori              ✓                   ✓
a posteriori          —                   ✓

One can point to no single book, as for instance one presents a Euclid, and say: this is metaphysics, here you will find the highest aim of this science, knowledge of a supreme being and a future life, proven from principles of pure reason.

...in order to be able to derive, from the principle of the possibility of the given cognition, the possibility of all other synthetic cognition a priori.

...synthetic but pure rational cognition is actual; but we must nonetheless next investigate the ground of this possibility, and ask: how this cognition is possible, so that we put ourselves in a position to determine, from the principles of its possibility, the conditions of its use and the extent and boundaries of the same. Expressed with scholastic precision, the exact problem on which everything hinges is therefore:

How are synthetic propositions a priori possible?

In short, how can there be ampliative or informative judgments that are nevertheless necessarily true? This is the technical problem driving the critical philosophy — Jill Vance Buroker

When Kant asks the question “how are synthetic a priori judgements possible?” he is asking “how is it possible for thought to generate something new?” — Levi Bryant

For how is it possible, asked the acute man, that when I am given one concept I can go beyond it and connect another one to it that is not contained in it, and can indeed do so, as though the latter necessarily belonged to the former? Only experience can provide us with such connections (so he concluded from this difficulty, which he took for an impossibility), and all of this supposed necessity – or, what is the same – this cognition taken for a priori, is nothing but a long-standing habit of finding something to be true and consequently of taking subjective necessity to be objective.

It can be said that the whole of transcendental philosophy, which necessarily precedes all of metaphysics, is itself nothing other than simply the complete solution of the question presented here, but in systematic order and detail, and that until now there has therefore been no transcendental philosophy; for what goes under this name is really a part of metaphysics, but this science is to settle the possibility of metaphysics in the first place, and therefore must precede all metaphysics.

I call all cognition transcendental that is occupied not so much with objects but rather with our mode of cognition of objects insofar as this is to be possible a priori. A system of such concepts would be called transcendental philosophy.

pure natural science, which, a priori and with all of the necessity required for apodictic propositions, propounds laws to which nature is subject.