The Evolution of Modern Metaphysics: Making Sense of Things (34 page)

True, there is nothing here that directly violates any of Kant’s principles, provided that such ‘forcing’ does not issue in knowledge on our part, or, more strictly perhaps, provided that it does not issue in knowledge of the kind to which the discussion hitherto has been tacitly restricted (see the previous section). But in what then does it issue? We might say that it is a variation on the compulsion whereby we believe in our own free will. But let us not pretend that that compulsion is itself completely unmysterious. In all these cases, including the cases of supposed transcendental knowledge that Kant takes himself to propagate in the first
Critique
, Kant is accrediting us with sense-making of a singular kind. On the one hand it is synthetic and
a priori
, which means that we cannot regard it as straightforwardly thin. On the other hand it results from sensitivity to transcendent(al?) features of our own faculties for sense-making, which means that we cannot regard it as straightforwardly thick. The truth is, we do not in the end know how to regard it. We cannot make sense of it.
⁸⁴

10. The Unsatisfactoriness of Kant’s Metaphysics

Kant’s most general attempt to make sense of things is ultimately and profoundly unsatisfactory. In his self-conscious reflections on what it is to make sense of things, he achieves insights of unsurpassed brilliance and gives us greater help than anyone before or since in thinking about what we can and cannot aspire to when we practise metaphysics. But in his attempts
to systematize these insights he appears to violate them and leaves us with something that does not itself, in the end, make sense.

It is as if, even by Kant’s own lights, the only
real
sense that we can make of things is whatever sense we can make of them by looking through our spectacles, which means, in particular, that we cannot make real sense of the claim that the only real sense we can make of things is whatever sense we can make of them by looking through our spectacles. Transcendental idealism appears to foreclose its own acknowledgement. (This is illustrated by a characteristically stubborn propensity, on the part of various transcendental claims that Kant wants to make, to be interpreted in the wrong way. Thus, to revert to an example that I used in §4, there is a natural and compelling sense, which Kant is the first to recognize, in which the sun’s being larger than the moon is quite independent of us. On the other hand, transcendental idealism requires us to recognize a sense in which it is not. In the first edition of the first
Critique
Kant tries to distinguish these senses by appeal to a deep ambiguity in the use of expressions such as ‘independent of us’: the sun’s being larger than the moon is
empirically
independent of us, that is independent of us in terms of the sense we make of things when we look through our spectacles, but not
transcendentally
independent of us, that is not independent of us in terms of the sense we make of things when we reflect on the spectacles themselves (A373). In the later
Prolegomena
Kant laments the fact that, despite having drawn this distinction, he has been interpreted as denying the former independence (§13, Remark III). In other words, he has been interpreted as espousing what, in a slight deviation from Kant’s own usage, is often called empirical idealism.
⁸⁵
In the second edition of the first
Critique
many of Kant’s bolder affirmations of transcendental idealism, along with these efforts to distance it from empirical idealism, are simply excised. It is as if he is engaged in an ongoing struggle to suppress the empirical interpretation of his transcendental claims and, in at least some crucial cases, eventually gives up.)

It is too soon to say where the fault lies. Perhaps Kant has not systematized his insights properly; perhaps they cannot be systematized; perhaps, indeed, they cannot be fully articulated.
⁸⁶
However that may be, the important questions for us, as practising metaphysicians, are ‘How should we react to this?’, ‘How might we use it?’, not ‘Do we accept it?’ For we surely do not. We need some other way of rescuing the
Humean baby.

Appendix: Transcendental Idealism Broadly Construed

I said in §4 that transcendental idealism will play a crucial role in this narrative. But we shall encounter many different versions of it, sometimes only indirectly related to the specific doctrine about space and time that Kant espouses. I therefore need to give some indication of what I mean when I call a doctrine a version of transcendental idealism. That is the purpose of this brief
appendix.

I follow Kant in distinguishing transcendental idealism from another kind of idealism, empirical idealism (see the material in parentheses in §10). And, concomitantly with recognizing versions of transcendental idealism that are only indirectly related to anything that Kant himself has in mind, so too I recognize versions of empirical idealism that are only indirectly related to anything that he has in mind. For my purposes, the crucial distinction between the two kinds of idealism with which Kant himself is concerned – the distinction that I wish to generalize – turns on the following question. Is the dependence of the world of our experience on our experience of it of a piece with, or does it utterly transcend, what we can know about that same world through experience?

To clarify: let
s
be a kind of sense-making. Then idealism with respect to
s
may for these purposes be defined as the view that certain essential features of whatever can be made sense of in accord with
s
depend on features of
s
itself. Empirical idealism, as I intend it, includes the rider that this dependence can itself be made sense of in accord with
s
. Transcendental idealism, as I intend it, includes the rider that it cannot.

1
Throughout this chapter I use the following abbreviations for Kant’s works:
Correspondence
for Kant (
1999
); 2nd
Critique
for Kant (
1996c
); 3rd
Critique
for Kant (
2000
); ‘Enlightenment’ for Kant (
1996a
); first
Critique
for Kant (
1998
);
Groundwork
for Kant (
1996b
);
Lectures
for Kant (
1997
); 1st ‘Logic’ for Kant (
1992a
); 2nd ‘Logic’ for Kant (
1992b
); 3rd ‘Logic’ for Kant (
1992c
);
Prolegomena
for Kant (
2002a
); and
Religion
for Kant (
1996e
). Page references are to the Akademie edition as indicated in the margins of these works, except in the case of the first
Critique
, where page references are to the two original German editions, ‘A’ representing the first edition and ‘B’ the second. All unaccompanied references are to the first
Critique
.

2
It does no great harm to think of Kant as using the term ‘metaphysics’ and its cognates pretty much in accord with my definition, at least in its extension if not in its intension. He himself defines metaphysics in more than one way (see e.g. n. 44). But on what I take to be his most interesting and most considered definition, metaphysics is ‘the investigation of everything that can ever be cognized
a priori
as well as the presentation of that which constitutes a system of pure philosophical cognitions of this kind, but in distinction from all empirical as well as mathematical use of reason’ (A841/B869; see also A850–851/B878–879; and see n. 13 for discussion of Kant’s use of the term ‘cognition’). He insists that this definition is to be preferred to the Aristotelian definition, whereby metaphysics is ‘the science of the first principles of human cognition’ (A843/B871), on the grounds that the Aristotelian definition invokes a difference of degree – at what point does a principle cease to be a ‘first’ principle and become a ‘secondary’ one? – whereas the difference between metaphysics and any other science is, for Kant, a difference of kind. This presumably means that he would take issue with my definition too (‘… most general …’). Nevertheless, what I count as metaphysics can certainly embrace what Kant counts. And even if it can embrace more, this does not affect anything I shall say in this chapter. We can afford to prescind from the differences between Kant’s definition and mine.

3
See ‘Enlightenment’,
passim
. Whether this ideal, in this form, was a defining characteristic of enlightenment, in each of
its
forms, is debatable. Hume, who was a central figure in the Scottish Enlightenment, would have contested it. ‘Reason,’ Hume wrote, ‘is, and ought only to be the slave of the passions, and can never pretend to any other office than to serve and obey them’ (Hume (
1978a
), p.451); see further §5 of the previous chapter. But the ideal was certainly not Kant’s figment. It was another of the bequests that he received. It had several sources. The most notable, from our point of view, was Descartes; see
Ch. 1
, §2.

4
Cf. Deleuze (
2006a
), p. 85.

5
For three excellent overviews of Kant’s relation to metaphysics, each covering much of the territory that I aim to cover in this chapter, see Warnock (1957), pp. 128–136; Gardner (
1999
),
Ch. 1
; and Rescher (
2000
),
Ch. 6
.

6
This is not Kant’s only use of the ‘dogmatic slumber’ metaphor. Later in the
Prolegomena
, in §50, and in ‘Letter to Garve’, dated 21 September 1798, in
Correspondence
, 12:258, he alludes to the antinomies of pure reason – which we shall consider in §6 – in much the same terms. And there is a fourth use of the metaphor, historically the first, at A757/B785.

7
This attitude is well expressed, albeit without reference to Hume, in ‘Letter to Mendelssohn’, dated 8 April 1766, in
Correspondence
, 10:70, and in
Lectures
, 29:957–958.

8
This explains the inferiority complex that metaphysics has long had, as has philosophy more generally, where mathematics is concerned. Ever since the time of Plato, mathematics has been seen as a model for philosophy to aspire to. Plato’s own veneration of mathematics is said to have been given celebrated expression in the inscription which he placed over the door of the academy he founded: ‘Let no one enter here who is ignorant of geometry.’ (The story may however be apocryphal: see Gilbert (
1960
), p. 88.) Cf. also Hume’s urge to privilege mathematics over any other
a priori
endeavour (
Ch. 4
, §5).

9
Note that Kant sometimes uses the word ‘metaphysics’ elliptically for good metaphysics, sometimes elliptically for bad metaphysics, and sometimes neutrally. See respectively A850–851/B878–879, Aviii, and Axii (which is echoed in the title that I have given this chapter) or B21–22. This creates a hazard for anyone quoting him out of context. In context, I think, his intention is always clear.

10
Kant beautifully and concisely summarizes the task at hand, and his proposed way of meeting it, at
Prolegomena
, 4:360. Cf. also, much more extensively, B Introduction, esp. §§III, VI, and VII, and A758–769/B786–792.

11
Cf. B20. For a scathing attack on the whole project, see Nietzsche (
1967c
), §530 (and see further
Ch. 15
, §2). For a very interesting and instructive account of Kant’s relation to Hume, see Stern (
2006
).

12
See n. 28 of the previous chapter and accompanying text.

13
Here and throughout this chapter, except when I am directly quoting from Kant, I put in terms of knowledge what Kant himself typically puts in terms of cognition. ‘Cognition’ is the word used in the Cambridge edition of his works to translate ‘
Erkenntnis
’. The translators are quite right not to use ‘knowledge’ for this purpose, as their predecessors did. For, although it is not abundantly clear what Kant means by ‘
Erkenntnis
’ – either from his various uses of the term or indeed from his explicit definitions of it, which are not always consistent with one another – it
is
clear that he does not mean knowledge. He seems to mean what might best be characterized, in his own terms, as ‘the conscious representation of an object’: see e.g. A320/B376–377 and 2nd ‘Logic’, 24:702. (There are further questions, of course, about his understanding of ‘conscious’, ‘representation’, and ‘object’. In a much looser, non-Kantian formulation he seems to mean the having of something in mind, or better a state that centrally involves the having of something in mind.) This excludes some knowledge and it includes some non-knowledge. The knowledge that it excludes is knowledge that is purely conceptual and makes no reference to any object: I shall have more to say about such knowledge in §8. (But note that this is
not
the same as what will shortly be identified in the main text as analytic knowledge: see A151–152/B190–191 and
Prolegomena
, 4:276.) The non-knowledge that it includes is the conscious representation of an object that contains some error: see 1st ‘Logic’, 24:93–94 and 105, and 3rd ‘Logic’, 9:53–54. My reason for bracketing cognition and talking in terms of knowledge is that this connects better with my broader concerns. My justification for doing so is that, in all the relevant contexts, the questions that Kant raises about cognition, and the answers that he gives, are equally questions and answers about knowledge. When he asks, for example, how cognition of a certain kind is possible (see e.g. B19ff. and
Prolegomena
, §5, and see further §4 below), the kind of cognition in question is likewise a kind of knowledge. Cf. in this connection Bvii–x and
Prolegomena
, 4:371.