The Evolution of Modern Metaphysics: Making Sense of Things (72 page)

Logical positivists appear to have considerable room for manoeuvre then. So why Ayer’s discomfort? Well, if the verification principle is a prescriptive definition, there is still the question of why the prescription should be obeyed. ‘I avoided this awkward question,’ Ayer says, ‘by defying my critics to come up with anything better’ (Ayer (
1992
), p. 149). Very well; but why is the question awkward? After all, insofar as it is an invitation to expound on the significance of the
definiens
, or on the work that might be done by the
definiendum
, there is plenty that Ayer can say in response to it. For instance, he can advert to some of the important differences that there are between matters that can be settled by appeal to some combination of sense experience and conceptual analysis and matters that cannot be settled in that way.

Nonetheless, the question is awkward. And what makes it awkward is something that Michael Dummett forcefully argues in the essay to which Ayer is replying when he makes these remarks: namely, that no answer he gives will be fully satisfactory unless and until it is placed in the context of some general philosophical account of the different ways in which different matters can be settled and of what turns on these differences, an account that is of just the same character as what he (Ayer) wants to cast on the pyre of metaphysics (Dummett (
1992
), pp. 133–134).
⁵⁵
Consider: as a logical positivist, Ayer not only distinguishes between matters that can be settled by appeal to some combination of sense experience and conceptual analysis and matters that cannot be settled in that way; he also insists that matters of the latter kind, which is to say matters that are neither matters of empirical truth nor matters of analytic truth,
are not matters of truth at all
. In particular, he
insists that no evaluative matter is a matter of truth. But this requires some defence.
⁵⁶
And the defence must be a contribution to what, on any remotely orthodox conception of metaphysics, counts as the metaphysics of value. The logical positivist must enter into debate with Spinoza (
Ch. 2
, §3), with Leibniz (
Ch. 3
, §1), with Hume (
Ch. 4
, §5), with Kant (
Ch. 5
, §§7 and 9), with Fichte (
Ch. 6
, §4), with Hegel (
Ch. 8
, passim), and with the early Wittgenstein (
Ch. 9
, §8)
⁵⁷
– debate of just the kind that he is trying to dismiss as spurious. He must try to give an account of value and its place in the world.

To be sure, this enterprise is to some extent an attempt to arrive at what, in the logical positivist’s own terms, counts as analytic truth of the most general kind, or at what, specifically in Carnap’s terms, counts as clarity of linguistic frameworks of the most general kind. But it is not just that. Recall that in §5(b) I mentioned one of the activities which, even by Carnap’s lights, might reasonably attract the label ‘metaphysics’:
clarifying linguistic frameworks of the most general kind
and deciding which to adopt
. The enterprise identified above must include some such decision-making. It cannot be confined to (theoretical) reflection on the relations between different conceptions of evaluation and different conceptions of truth and falsity. It must include (practical) reflection on which of these conceptions are best suited to our various purposes.

Whatever
label attaches to the activity of clarifying linguistic frameworks of the most general kind and deciding which to adopt, decisions of that sort must sooner or later be taken. We must sooner or later reflect, not only on how we make maximally general sense of things, but on how
to
make maximally general sense of things.
⁵⁸
We must consider whether our current sense-making is all that we might have expected it to be; whether such and such alternatives might work better for us; how well we could cope with them; at what cost; with what gain. And in making our decisions – in putting our reflections into practice – we are neither trading in analytic truths nor trading in empirical truths. We are not trading in truths at all. We are, however, practising something at least akin to traditional metaphysics, something that certainly counts as metaphysics on my conception. It seems to me, then, that Carnap and the other logical positivists, like Hume before them, do not so much eliminate metaphysics as put us in mind of its importance.

1
The term ‘logical positivism’ was introduced by Albert Blumberg and Herbert Feigl in Blumberg and Feigl (
1931
). ‘Positivism’ on its own, or ‘positive philosophy’, had earlier been used by Auguste Comte to designate his loosely related system, e.g. in the very title of his masterwork, Comte (
1988
). But I use the expression ‘loosely related’ advisedly. Comte’s use of the term ‘positivism’ and this subsequent use are best regarded as homonymous.

2
Another way to put it, equally caricatural, would be to say that logical positivism was Humeanism transposed from a psychological to a logical key.

3
It is only fair for me to add a reminder that the interpretation of Hume that I am presupposing in this sketch is contested: see
Ch. 4
, §1. Edward Craig, who reads Hume very differently from how I do, concludes the chapter on Hume in Craig (
1987
) with a section entitled ‘Hume’s Heirs?’ in which he argues that logical positivists’ ‘claim to be Hume’s twentieth-century heirs is … in a certain very important respect … the exact reverse of the truth’ (p. 128). The final sentence of Craig’s chapter captures well both its tone and its content: ‘Hume’s heirs indeed!’ (p. 130).

4
See e.g. Ayer (
1971
),
Ch. 4
. In this they were in line with the early Wittgenstein (
Ch. 9
, §4). But the matter is less straightforward than it seems: see further §3(d) below.

5
For discussion of some of the complications that this glides over, see Ayer (
1971
), Introduction. Note that Ayer talks in terms of (non-conclusively) verifying a truth rather than knowing it (see e.g. p. 179). But on a suitably undemanding conception of knowledge these can be regarded as equivalent.

6
Cf. Ayer (
1971
),
Ch. 6
.

7
Carnap was one of the foremost members of the group with which logical positivism was particularly associated, the Vienna Circle. This was a group of between thirty and forty thinkers, from a range of disciplines, who met regularly in Vienna between the wars to discuss philosophy. The group was united by the aim to make philosophy scientific, in a sense of ‘scientific’ broad enough to embrace formal logic. Many members of the group thought that a commitment to logical positivism was indispensable to the realization of that aim. Carnap certainly did.

8
See ‘Ontology’ more generally, esp. §2, for discussion of the idea of a linguistic framework.

Note: throughout this chapter I use the following abbreviations for Carnap’s works:
Aufbau
for Carnap (
1967a
); ‘Autobiography’ for Carnap (
1963a
);
Der Raum
for Carnap (
1922
); ‘Elimination’ for Carnap (
1959
); ‘Goodman’ for Carnap (
1963b
); ‘Introduction’ for Carnap (
1958
);
Logical Syntax
for Carnap (
1937
);
Meaning and Necessity
for Carnap (
1956a
); ‘Ontology’ for Carnap (
1956b
);
Philosophy
for Carnap (
1935
); and
Pseudoproblems
for Carnap (
1967b
).

9
See esp. ‘Ontology’ and
Pseudoproblems
, Pt II.

10
This is an apt point at which to return to the contrast between the history of philosophy and the history of ideas which I mentioned in the Preface, and to emphasize that I am engaged in the former. I am not much concerned with conscious assimilations or repudiations on Carnap’s part.

11
For a beautiful summary, see Quine (
1995a
), pp. 10–14.

12
Quine famously expresses reservations in Quine (
1961b
), §5. For more detailed reservations, see Goodman (
1977
). For a fascinating criticism of a related project in A.J. Ayer, a criticism that identifies some fundamental problems for logical positivism more generally, see Williams (
1981
).

13
Cf. n. 2 above. And see Stroud (
1977
),
Ch. 10
, §I, for a discussion that emphasizes how different these projects are.

14
For discussion, see Goodman (
1963
), which is more sympathetic than Goodman (
1977
), cited in n. 12 above, and to which Carnap replies in ‘Goodman’.

15
Such geometries were developed in the late eighteenth and early nineteenth centuries. It is natural to wonder whether Kant knew of their existence. Unfortunately, we cannot be sure. But probably not: see Potter (
2000
), pp. 36–37.

16
For a beautifully clear and concise account of this conception, see ‘Introduction’. This supersedes Carnap’s doctoral thesis, published as
Der Raum
, in which he still accedes to the synthetic
a priori
.

17
Cf. Stroud (
1984
), pp. 195–196, though for reasons that I am about to sketch I think that Barry Stroud overstates the similarities. The whole of
Chapter
5 of Stroud’s book is concerned with Carnap’s idea of a linguistic framework. He points to some deep problems that he thinks afflict the idea. We shall consider related problems in §§5 and 6.

18
For a thorough comparison of Carnap with Frege, see Gabriel (
2007
). For a clear statement from Carnap himself of how he sees his own insights as amalgamating those of the classical empiricists and Frege, see
Aufbau
, p. vi.

19
Mention should also be made of the influence on Carnap’s semantics of Frege’s distinction between sense and
Bedeutung
: see esp.
Meaning and Necessity
, where in §§28 and 29 Carnap compares and contrasts Frege’s distinction with his own distinction between ‘intension’ and ‘extension’.

20
See e.g.
Logical Syntax
, §17; ‘Ontology’, §5; and ‘Autobiography’, p. 49. Note the ‘Principle of Tolerance’ which Carnap formulates in the first of these passages: ‘It is not our business to set up prohibitions but to arrive at conventions’ (p. 51, emphasis removed). In the second and third passages he positively encourages the adoption of non-standard linguistic frameworks, in case unexpected benefits accrue. If they do not, no harm will be done: ‘the work in the field will sooner or later lead to the elimination of those forms [of expression] which have no useful function’ (‘Ontology’, p. 221).

21
See
Ch. 4
, n. 36.

22
Note that Carnap has an easy way with Frege’s question whether Julius Caesar is a number or not (
Ch. 8
, §3). Accepting the existence of people and accepting the existence of numbers each involves adopting a linguistic framework that makes no reference to the other. So we are free to insist that people and numbers are
sui generis
and that Julius Caesar is therefore not a number: cf. ‘Ontology’, pp. 210–211. (But we are also free, if we so choose, to create and adopt a linguistic framework in which people
are
numbers: cf.
Philosophy
, Lecture III, §3. To that extent Frege was right to view this as a matter of stipulation. Unlike Frege, however, we now have some rationale for so viewing it.)

23
That Carnap overestimates what can be achieved by the arbitrary element in language, and specifically in the language of logic, is a theme of Quine’s well-known attack on his views in Quine (
1966b
). ‘The difficulty,’ Quine writes, ‘is that if logic is to proceed mediately from conventions [i.e. if it is to proceed in accord with Carnap’s understanding of what is arbitrary about it], logic is needed for inferring logic from the conventions’ (p. 98). See also Quine (
1966c
).

24
See
Ch. 9
, n. 20.

25
For a stark drawing of the contrast between logical positivists and Wittgenstein in this respect, see Engelmann (
1967
), pp. 97–98. See also n. 39.

26
They also explain Wittgenstein’s feeling of alienation with respect to the Vienna Circle, despite the popular view, originating with members of the Circle themselves, that he had been one of their primary inspirations. For two full and very instructive accounts of the philosophical relations between Wittgenstein and the Vienna Circle, see Baker (
1988
),
Ch. 6
, §1, and Hacker (
1996
),
Ch. 3
. For insight into their personal relations, see Monk (
1991
), pp. 242–242 and 283ff.