Realism and antirealism

2. Ontological realism/antirealism

The primary form of the definition deals directly in terms of what really exists. A realist about Xs, for example, maintains that Xs (or facts or states of affairs involving them) exist independently of how anyone thinks or feels about them; whereas an antirealist holds that they are so dependent. We are not speaking here of causal (in)dependence: the fact that there would be no houses if people had not had certain thoughts should not force us into antirealism about houses. So the point of the definition is better brought out by saying that what it is for an X to exist does not involve any such factors (whatever their causal role in the production of Xs may be). Nor does the definition entail an antirealist stance towards the mental. Realism about mental states is a prima facie plausible option, holding that our mental states are what they are whatever we think they are, or whatever we would come to think they were if we investigated.

Where philosophers have argued for realism about some particular subject matter (for example, universals, ethical value, the entities of scientific theory), one particular argument is repeatedly found. For the subject matter in question, it is claimed, we find that everyone’s opinion is the same, or tends to become the same if they investigate, or that (in science) theory seems to ‘converge’, later theories appearing to account for the partial success of their predecessors. Why should this be, unless it is the effect of a reality independent of us, our opinions and our theorizing? (See Universals; Scientific realism and antirealism.)

In consequence, there are two broad antirealist strategies, both common. One is to argue that the supposed conformity of opinion, actual or potential, does not exist – so we hear of the diversity of ethical or aesthetic judgements, for instance, or the extent to which judgements of colour depend on viewing conditions and the state of the observer. The other is to accept the conformity, but explain it as arising from a uniformity of our nature rather than the independent nature of things. Thus it is argued that moral ‘objectivity’ is really ‘inter-subjectivity’ – that is, a result of shared human psychological responses rather than of independent moral properties in the world – or that the similarity between different languages’ schemes of classification is a product of shared basic human interests, not something forced on us by ‘real’ universals.

In modern times nobody has made a more radical use of this method of explaining conformity of judgement in terms of intersubjectivity than Kant. He argued that even the experience of our environment as extended in space and time was a human reaction to things that were in themselves not of a spatiotemporal nature, and to which other beings might just as legitimately react altogether differently. In the face of this it may be felt that the argument from conformity is better used to establish a very abstract realism, namely that there must be something independent of us, rather than that any specific property or type of thing must be so.

Two other objections have been used against certain forms of realism. One is that the realist provides no account of how the supposed real things or properties can actually have an effect on our experience. What sense do we have, it is asked, that is affected by the ethical properties of the moral realist, or by the real properties of necessity and possibility that the modal realist posits? The common realist practice of speaking of ‘intuition’ in these contexts is rejected as providing only a word, not an answer. The second type of objection (christened the ‘argument from queerness’ by John Mackie (1977), who used it in the moral context) claims that the things or properties in which the realist believes would need to be too strange to be credible (see Moral realism; Modal logic, philosophical issues in).

A closely related definition of the realism–antirealism distinction focuses not on the independence of things but on the truth of judgements about them: realism takes truth to be correspondence with fact and our knowledge of truth to be a separate matter, whereas antirealism defines truth ‘in epistemic terms’, that is to say as what human beings would believe after the best possible application of their cognitive faculties. This is much more a change of perspective than of substance. It is natural to think that if some object exists independently of us, then judging truly must consist in getting our judgement to match the way the object is; while if the object is determined by (perhaps a projection of) our cognitive and/or affective faculties, judging truly can only mean judging as those very faculties lead us to judge.

Harder to assess is the position of reductive analysis in the debate. A reductive analysis exists where what makes statements about one kind of thing, A, true or false are the facts about another kind B. (As are then said to be reducible to Bs.) Classically, phenomenalism claims that statements about physical objects are thus reducible to statements about sensory experiences; behaviourism holds that propositions about mental states are reducible to ones about dispositions to physical behaviour. Does accepting such a reduction mean accepting antirealism about the As, while rejecting reduction of A-statements mean accepting realism? Some philosophers speak in this way, and there is a clear point to doing so: if a reduction is possible, then a complete statement of everything there ‘really is’ would not need to mention As – it could speak of Bs instead. Besides, reductive analyses have usually been offered in opposition to a different conception of what As are, and in relation to that (rejected) conception of an A the reducer is certainly saying that there are no As. But it is not thereby said that As and facts about them are dependent upon us – only that they are really certain sorts of fact about Bs; our attitude to their independence is therefore a question of whatever we think about the latter (see Reduction, problems of).