by Akeel Bilgrami
The form of realism that I find most plausible is best elaborated in terms that are a combination of Kant's idea of ‘transcendental idealism' (without any commitment to the ‘ding an sich') and ‘pragmatism' of a rather specific kind that can be found in Charles Sanders Pierce's path-breaking paper, ‘The Fixation of Belief' and developed within a complex account of belief revision by Isaac Levi.
In being Kantian, it is a realism that renounces what Hilary Putnam has called ‘metaphysical realism' or what perhaps in an earlier time might have been called ‘transcendental realism'. And in being pragmatist, it renounces, a fallibilist, Cartesian epistemology.
Let me speak to its pragmatist side first.
Pragmatism, at its most general, says: Something that makes no difference to practice makes no difference to Philosophy. What is yielded when we apply this dictum more specifically to epistemological matters in Philosophy? It yields the following thought. A pragmatist epistemology claims that something that makes no difference to the cognitive practice of inquiry makes no difference to epistemology. And so it finds that a fallibilist form of doubt that is found in Cartesian skepticism makes no difference to inquiry and therefore is not a credible epistemology.
Thus let us take Cartesian skepticism about the external world. It claims both that all our beliefs about the external world could be false and that of any particular such belief, we could never be certain of its truth. These are two distinct claims since the latter does not entail the former. It is the latter claim that pragmatism opposes. The basis of its opposition is that if we can never know of any given belief about the world that it is true, then truth cannot be a goal of inquiry. It makes no good sense to say that truth is a goal of inquiry even though we are never sure in any given case that we have achieved the goal. That would mean inquiry would be like sending a message in a bottle out to sea. What kind of epistemological enterprise is that? We would never have any control over its success, and all success would appear to be a sort of bonus or fluke.
The point can be made more vivid by considering the paradox of the preface. An author often writes: “In the next four hundred pages, there is bound to be something or other that I have said which is false” (and then usually adds, and “I am to blame for it and not those nice people who helped me while I was writing the book.”). But this form of general doubt he expresses makes no difference to his practice as an author. It is not as if it tells him which specific claim of his in the book is false. It just says ‘something or other' that he claims might be false. And such a doubt therefore gives him no instruction about how to go about improving his book. It is the same with Cartesian doubt. If I think that any given belief of mine might be false without any particular evidence offered against any specific belief, it gives me no grounds to reconsider any belief or to help to improve the current state I have arrived at in my inquiry regarding the world.
So, truth is something that, when we achieve it, we must not be blinded as to whether we have achieved it or not. We can know when we have achieved the truth. This does not mean that we cannot revise our beliefs. We can certainly revise beliefs, even beliefs that we currently hold with conviction. But the idea that we can revise a given belief we hold with conviction is not properly expressed by saying, “My belief that p is true, but for all I know it might be false,” or “I know that p, but for all I know, p is false.” This is a point and distinction that was much stressed by Peirce in that great essay and also by philosophers such as Wittgenstein and Austin.
How does this pragmatism relate to Kant's notion of ‘transcendental idealism'? If pragmatism is right, the word ‘belief' can mean two different things for inquiry. First there are beliefs held with conviction and certainty. Second there are beliefs held as hypotheses. In inquiry, the former form a background and the latter are in the foreground of inquiry. And when we test and assess beliefs qua hypotheses, we do so taking for granted without any doubt the truths of the beliefs in the background because it is these background beliefs that provide the lights by which we assess the foreground beliefs or hypotheses. They are simply not, therefore, subject to philosophical and fallibilist forms of doubt. And this could only mean that when we state the conditions which make any sentence true or false, the conditions are always something that are specified by the lights of our own background beliefs. Kant had said that reality is in some sense not entirely independent of our own world-view. This is why he described his view as a ‘transcendental idealism' rather than realism. But, as I will try and show in a moment, transcendental idealism –in the pragmatist elaboration that I have presented — is quite compatible with the realism that demands that truth and reality be independent of our mentality. In the pragmatist idea of inquiry, our settled beliefs, beliefs held with conviction, provide the background lights of assessment, i.e., the worldview from within which we, as inquirers, can assess foreground hypotheses or specify the conditions under which a specific sentence is true.
That is the marriage of pragmatism and transcendental idealism that makes for a sane realism.
But someone might protest: Your notion of realism does not allow truth to be sufficiently independent of belief. I think the only sense in which independence of this kind is required by realism is the independence which says: truth is independent of belief in the sense that a belief is true whether we believe it or not. And the realism I have elaborated in pragmatist and transcendental idealist terms, does allow this. Suppose we believe that p is true. This means that it meets the standards of correctness provided by our background beliefs. But now the standard of correctness provided by our background beliefs in inquiry is such that we also know that, were we to suppose that we judge something to be true which is not dictated by those standards, we would be wrong to do so. Thus were we to suppose that we judge that not-p is true, we would conclude that that judgement was wrong. In other words p is true whether we believe it or we don't believe it and believe its negation. And that is just what is needed to establish the requisite independence that makes truth objective. No greater independence of truth from belief is required to make truth objective.
* * *
Akeel Bilgrami holds the Sidney Morgenbesser Chair in Philosophy at Columbia University and is also the Director of the South Asian Institute there.