The Evolution of Modern Metaphysics: Making Sense of Things (84 page)

The world is the totality of facts, not things, and facts are true [thoughts];
¹⁷
so the concept of truth is the hinge upon which the door from the philosophy of thought opens into metaphysics, that is, the range of philosophical problems that concern the general character of reality. (‘Beards’, p. 890)

To reflect on the character of truth will in turn be, in Frege’s famous phrase, ‘to discern the laws of truth’ (Frege (
1997l
), p. 325/p. 58 in the original German). And that, as Frege just as famously observed, is the task of logic (ibid.). And so it is that metaphysics will come to have, as Dummett puts it in the title of his most pertinent
book, a
logical basis
.
^18
,
19

2. Realism and Anti-Realism

We were reminded in the previous section of the unsettling effects that reflection on linguistic sense can have. Dummett is well aware of these and of what it takes to come to terms with them. Coming to terms with them is a condition of pursuing metaphysical questions on his conception. But it is not a condition in the sense of a prerequisite, something that has to be satisfied
first
. It is itself part of the metaphysical enterprise. Dummett holds that reflection on linguistic sense forces us to reassess some of our most deeply entrenched convictions
concerning the general character of reality. Determining what to do about these convictions, perhaps determining how to live without them,
is
engaging in metaphysics.

The best known and most elemental example is the conviction that every thought is either true or not true.
²⁰
I say ‘not true’ rather than ‘false’. This is because there are various innocuous ways of challenging the idea that every thought is either true or false. For instance, it is natural to say that a thought about something non-existent, say the thought that Atlantis was ruled by a confederation of kings, is neither true nor false. But, whether or not that is the correct thing to say – perhaps it is not even correct to say that there is a thought involved in such a case
²¹
– the conviction that every thought is either true
or not
remains unassailed. Is it not unassailable? An incontrovertible law of logic?
²²
On what possible grounds could
this
conviction be abandoned? It certainly never crossed Frege’s mind to abandon it (cf. Frege (
1997i
), p. 300/p. 214 in the original German).

Dummett nevertheless urges caution. Reconsider the idea that every thought is either true
or false
. The innocuous ways of challenging this idea concern recognizable relations, in particular recognizable misfits, between our thinking and reality, such as we arguably find in the Atlantis case. Dummett, by contrast, is exercised by the prospect of
un
recognizable relations between our thinking and reality (‘Truth’, p. 23). Thus suppose there is a thought whose truth or non-truth we have no way of settling.
²³
(An example might be a thought concerning some distant and undocumented event in history,
such as the thought that Aristotle sneezed on his first birthday. Another might be a thought concerning a counterfactual that we have no way of negotiating, such as the thought that Descartes would have loved Marmite.) And suppose we take for granted that such a thought is either true or not true. Then, Dummett argues, it is not clear that we can give a satisfactory account of our grasp of the thought.

The argument proceeds very roughly as follows. Our grasp of a thought involves our knowing both how reality must be in order for the thought to be true and, derivatively, how reality must be in order for the thought not to be true. But such a grasp has to admit of public ratification. This is because, if it did not, nobody could ever know whether anybody else grasped the thought. So the thought would be incommunicable. But it is of the very essence of a thought to be communicable. For a thought is what can be expressed by a declarative sentence. And a declarative sentence can express only what it can be perceived to express, only what it can be used to communicate. (Both language learning and language use more generally would be unintelligible otherwise.)
²⁴
However, it is not clear how our grasp of a thought can admit of public ratification if the thought is either true or not true without our being able to tell which. We must not simply take for granted, then, that, even in default of our having some way of telling whether a thought is true or not, it is one or the other.
²⁵

The caution that Dummett is urging here is a caution against a basic realism. It is a caution against the idea that reality outstrips our capacity to know about it. This is what makes his circumspection, on any reasonable conception of metaphysics, a contribution to metaphysics.
²⁶
Where Frege held that the thought expressed by a declarative sentence is a matter of how things must be in order for the sentence to be true, irrespective of us, Dummett urges us to take seriously the ‘anti-realist’ alternative that the
thought expressed by a declarative sentence is a matter of how things must be in order for us to
recognize
that the sentence is true.
^27
,
28

Two things should be emphasized, however, lest his position appear more radical than it really is. First, ‘circumspection’ is the operative word. In developing the line of argument sketched above, Dummett is presenting realists with a challenge; he is not opposing them. Insofar as he is making an anti-realist proposal, it is a proposal
for consideration
. He does not want to preclude an eventual decision in favour of some version of the realism that he is querying. It is just that any such decision must, he thinks, be earned, in full awareness of the problems that afflict it and of the alternatives to it. That is, it is not something to which we can uncritically help ourselves. Thus in his valedictory lecture ‘Realism and Anti-Realism’ Dummett writes:

I viewed [my proposal], and still continue to view it, as a research programme, not the platform of a new philosophical party…. I did not conceive myself as proposing for consideration, let alone sustaining, any precise thesis, to be accepted or rejected. I saw the matter, rather, as the posing of a question how far, and in what contexts, a certain generic line of argument could be pushed. (p. 464; cf. ‘Preface’, p. xxxix)
²⁹

This quotation also highlights the second thing that needs to be emphasized. We must not think of realism and anti-realism as two absolute opposed positions. Each admits of degrees. And each needs to be relativized to an area of discourse. The issue is
how far
one should think realistically or anti-realistically
about
this or that subject matter. It would be entirely reasonable for an extreme anti-realist about mathematics, say, to be a relatively robust realist about history, say. Such a person may think that Dummett’s challenge can be met in the case of history in a way in which it cannot be
met in the case of mathematics, or perhaps that there is a variation on the line of argument sketched above that is compelling in the case of mathematics but that has no analogue in the case of history.
³⁰
It is instructive to see how the quotation from Dummett continues:

I saw the matter … as the posing of a question how far, and in what contexts, a certain generic line of argument could be pushed, where the answers ‘No distance at all’ and ‘In no context whatever’ could not be credibly entertained, and the answers ‘To the bitter end’ and ‘In all conceivable contexts’ were almost as unlikely to be right. (Ibid.)
³¹

Furthermore, the term ‘realism’ has many varied uses within philosophy other than to designate a commitment to the thesis that every thought is either true or not, and Dummett has over the years increasingly come to embrace some of these other uses – even to the extent of suggesting that innocuous challenges to the thesis that every thought is either true
or false
may count as retreats, however pedestrian, from some kind of realism (see
Logical Basis
, p. 325, and ‘Realism and Anti-Realism, p. 468).
³²
The main further ingredient that he now thinks a position needs to have, in order to attract the label ‘realism’ on its most compelling-cum-robust interpretation, is a commitment to the thesis that statements are to be taken at face value (
Logical Basis
, p. 325, and ‘Realism and Anti-Realism’, p. 468). Thus consider the following three arithmetical statements:

7 + 5 = 12.

7 + 5 = 13.

Every even number greater than 2 is the sum of two primes.

A realist about arithmetic, on
this
construal of realism, will still insist that each of these statements is either true or not true.
³³
Such a realist will also, however, insist that ‘7’, which appears to function as a singular term in the first two statements, does function as a singular term there and picks out a particular object; that ‘prime’, which appears to function as a sortal noun in the third statement, does function as a sortal noun there and applies to a particular kind of object; and that ‘every even number greater than 2’, which appears to function as a quantifier in the third statement, does function as a quantifier there and ranges over a particular domain of objects.
³⁴
What does this add? Well, consider the expression ‘Arthur’s sake’, as it occurs in the statement ‘She did it for Arthur’s sake.’ We surely have to say the opposite
in this case. Although that expression appears to function as a singular term in that statement, we surely have to deny that it really does function as a singular term there, or at any rate, as a singular term that picks out a particular object. (She did it so as to benefit Arthur. Only the two of them were involved.) This is a kind of anti-realism with respect to sakes.
³⁵

Be that as it may, the kernel of any realism, in Dummett’s view, remains a commitment to the thesis that every thought is either true or not true, independently of any capacity on our part to tell which. That is the thesis, of all realist theses, that ‘has the greatest metaphysical resonance’ (
Logical Basis
, p. 326).

3. Three Replies to Dummett’s Anti-Realist Challenge

Dummett’s anti-realist challenge to this thesis has met with all sorts of reactions, including all sorts of opposition. I shall consider three replies that are particularly relevant to our enquiry.
³⁶

(a) First Reply

The first reply is as follows. Dummett’s attempt to make metaphysical capital out of considerations about language depends on a broadly Wittgensteinian view of language, whereby both the meaning of an expression and our understanding of the expression should be open to public view in the use that we make of it in communicating with one another.
³⁷
But what if our use of language is not itself metaphysically neutral – as it surely is not?
³⁸
After
all, no one, not even a mathematician, who is unschooled in the philosophy of mathematics is liable to think twice about using the words ‘or’ and ‘not’ in accord with a basic mathematical realism (either every even number greater than 2 is the sum of two primes or it is not). Dummett must think that he is entitled to criticize certain aspects of our use of language. But on what grounds? Surely, it is part of his Wittgensteinian view that our use of language is a datum.

There is a counter-reply to this first reply which runs as follows. It is indeed part of Dummett’s view that our use of language is a datum, but not an unquestionable, indissoluble datum.
³⁹
Where our use of language betrays metaphysical commitments, or where it betrays beliefs of any other kind for that matter, it is vulnerable to criticism: the beliefs betrayed may be illusory. And there is nothing in Dummett’s conception to prevent him from playing the role of critic. Similarly, where our use of language betrays the rules that we take to govern it. There too it is vulnerable to criticism: the rules may conflict with one another. And again there is nothing in Dummett’s conception to prevent him from playing the role of critic. (Whether any given criticism can be sustained or not is another matter. That is to be determined on the merits of the case. This relates back to the point that the anti-realist challenge may be far more powerful in some contexts than in others.)
⁴⁰

This counter-reply seems to me to be perfectly adequate. But does it perhaps leave Dummett in the same difficult position as the one in which we found Wittgenstein in
Chapter 10
, §5? Wittgenstein was able to meet one problem about sense-making only at the price of confronting another: he had no independent leverage for distinguishing between those aspects of our use of language that contribute to successful sense-making and those that do not. Is the same not true of Dummett?