Transcendental Arguments and Their Discontents

by Scott F. Aikin and Robert B. Talisse

KantConsider the nihilist who provides us with an argument with the conclusion that nothing exists, or that there are no norms for reason. Take the relativist who contends that all facts are relative to some perspective. Note the skeptic who consistently criticizes not only our claims to knowledge, but our very standards. Call such views Transcendental Pessimism. An appealing and longstanding reply to Transcendental Pessimism is that it is self-defeating in some way. The nihilist nevertheless avows a fact and relies on norms of rationality to run the argument for his own conclusion. The relativist isn't just saying that it's all relative to her perspective, but that it's all relative full stop. The skeptic's conclusion that we have no knowledge or have no reliable means to assess knowledge purports to be a knowledge-like commitment held on purportedly good epistemic grounds. The critical line is this: Transcendental Pessimist views cannot be consistently thought. Such views, to make sense at all, must presuppose precisely what they deny.

So far, this self-defeat maneuver against nihilists, relativists, and skeptics is but an inarticulate hunch. Transcendental arguments are attempts at making that hunch explicit, not only about how the negative views are self-defeating, but also regarding the positive views worth preserving. That is, we deploy transcendental argumentation not only as a critical line against Transcendental Pessimism, but we also (and perhaps thereby) establish some positive conclusion. Call this objective Transcendental Optimism.

Immanuel Kant is widely acknowledged to be the first to overtly use the argument type. The primary example of Kantian transcendental argument comes in the Second Analogy of Kant's Critique of Pure Reason. The rough form of argument runs as follows: One can judge a series of representations is evidence of a series of events only if one holds that the series is asymmetric (it must happen in that order, not in a reverse or other order). One can believe that the representations are asymmetric only if one holds that the events represented are similarly asymmetric. If a series of states is asymmetric, the earlier states are causes of the later states. Therefore: One can take a series of representations as evidence only if one takes them as evidence of a causal order. Experience can be a source of information only if there is a causal order.

In the 20th Century, Donald Davidson employed a transcendental argument in defense of his thesis of radical interpretation. The criterion for identifying anyone as speaking a language is that of taking their utterances as semantically contentful. The condition for identifying semantically contentful utterances is that of interpreting the things people say to be responsive to events in the world around them. In his essay “Radical Interpretation,” Davidson explains the constraint thus: “A theory of interpretation must be supportable by evidence available to interpreters.” And so, we must have our defaults set on interpreting others as saying mostly true things. In his influential essay “On the Very Idea of a Conceptual Scheme,” Davidson writes, “We make maximum sense of the words and thoughts of others when we interpret them in a way that optimizes agreement.” Consequently, we have no intelligible reason to hold that others have different conceptual schemes from us. Radical interpretation is transcendentally dependent on the Principle of Charity.

Now, there are two problems with transcendental arguments; one dialectical, one formal. The dialectical challenge for transcendental arguments is that they seem to either beg the question or are otiose. They, consequently, do not play the rebutting or undercutting role in the critical exchange with the Transcendental Pessimist that the Optimist needs them to. Call this the dialectical dilemma for transcendental arguments.

Consider Davidson's argument. It begins from the requirement that any theory of interpretation must be supportable by evidence of connection between utterances and the world. Such a requirement is widely held to be a form of verificationism – the view that the meaning of a statement is delineated by conditions for its confirmation. This view of meaning does all the heavy lifting in Davidson's argument. But no skeptic or relativist or nihilist (no Pessimist) would accept verificationism. So the argument begs the question. Alternately, note that if the verificationism does all the work, the transcendental argument was, in the end, unnecessary. It is otiose. So if you can't convince the relativist of verificationism, you can't run Davidson's transcendental argument, and if you can sell verificationism to the relativist, you don't need the transcendental argument. As a consequence, either way, the transcendental argument is worthless. That's the dilemma.

The formal problem for transcendental arguments is that their optimistic conclusions are helplessly equivocal. Consider a shortened version of Kant's argument:

P1: It is necessary that: Contentful experience is possible for a subject only if that subject deploys the concepts of cause and effect.

P2: Subjects have contentful experience.

C: There must be cause and effect.

Yet the ambitious transcendentally optimistic conclusion C in fact does not follow. The premises rather support a much more modest result:

C*: Subjects must use the concepts of cause and effect.

As Kant puts it, “Experience itself . . . is thus possible only in so far as we subject the succession of appearances . . . to the law of causality; and as likewise follows, the appearances . . . are themselves possible only in conformity with the law.” Here we can see the difference between the two kinds of conclusion. The same thing happens in many other forms of transcendental argumentation. In order to ask a real question, one must think there are possible answers; in order to interpret others, one must take them to be in broad agreement with you; the condition for expecting an unsupported stone to drop is believing that gravity is real, and so on. What does not follow from any of these holdings, judgings, and believings are the facts of their assertional contents. That, by the way, was what the Pessimist was affirming all along.

We might call transcendental arguments that show just that something substantive must be used, presupposed, or assumed in order to say positive things at all a form of Modest Transcendental Argument. The trouble with modest transcendental arguments, when posed as arguments that show that the use of certain concepts is not optional, or invincible, as Barry Stroud puts it, is that they sound less like justification for these commitments, and more like exculpations. Just because the concepts or commitments are not optional in having our first-order commitments about the world, minds, and morals does not mean they are justified or are good.

The question is whether we can do better than exculpation for our Transcendental Optimism without committing the fallacy of equivocation. We, the authors, think there is a chance of doing better. It looks like this.

If Transcendental Pessimism is self-defeating (you can't consistently believe it), then we have justification in rejecting the view. That justification doesn't guarantee that Pessmism is false, but that we are rational in recognizing that we cannot ever hold the view with positive justification. Notice, now, that Optimism and Pessimism are the only options – if you suspend judgment between the two, you've slipped into Pessimism. It is, to use a term from William James, a forced move. Since we are justified in rejecting Pessimism, we are then justified in accepting Transcendental Optimism. The consequence, of course, is nothing earth-shaking. In fact, the Optimistic thesis was that we were all reasonable in believing that there is a world of causally efficacious things, other minds, and truths all along. The objective with the argument was to make it explicit why.