Wittgenstein’s deflationary stance towards both truth and logic goes a long way in explaining his remarkable lack of interest in the formal systematicity of language. As abstractions from complex patterns of linguistic use, recursive rules of the kind that define (or interpret) formal languages cannot explain or even reveal the background against which linguistic moves are made. Primitive games like the builders do more to display this background than the sophisticated constructions of idealizing formal languages. Logic requires the same background that our ordinary linguistic moves require. The problem lies not with rules of logic in particular, but with rule-

following of any sort. Wittgenstein has already suggested that there is this deeper problem when he first identifies the picture that holds us captive with the idea that linguistic mastery is “operating a calculus according to definite rules” (PI §81). The 80s foreshadow the rule-following considerations we are now turning to. In these earlier passages, Wittgenstein briefly introduces three variants of the regress argument, followed by even briefer replies to these arguments. He is, in effect, giving his reader a reminder of the kind of replies we would ordinarily find sufficient although a philosopher would not. In PI §84, he introduces a classical regress argument: “Can’t we imagine a rule determining the application of a rule, and a doubt which it removesand so on?” In PI §86, using a modified version of the builders game, he introduces the problem of the multiplicity of interpretations of how rules are to be applied. And finally in PI §87 he addresses doubts raised about explanations of meaning, “as though an explanation as it were hung in the air unless supported by another one.” All three of these truncated arguments will be examined anew in the rule-

following passages. The regress of applications of a schema is used early in the rule-following discussion to introduce the problem (PI §141). The

multiple interpretations argument becomes shaped as a “paradox of interpretation” (PI §201). We find a repetition of the concern that an interpretation “hangs in the air along with what it interprets, and cannot give it any support” (PI §198). Wittgenstein’s aim is to show that the philosophical picture of language as a calculus with fixed rules is vulnerable to the very philosophical problems it was thought to solve. Wittgenstein’s critical responses are all of the kind that a philosophically

naive person might offer. To the regress of applications engendered by doubt, Wittgenstein answers “But that is not to say that we are in doubt because it is possible for us to imagine a doubt” (PI §84). Imagined doubt is not real doubt. Wittgenstein is looking to the limitations of the philosophical imagination. In the multiple interpretations argument, Wittgenstein appeals to the training one gets in using the schema in one way rather than another: “One learns to look the picture up in the table by receiving a training” (PI §86). How we learn, then, is seen as an appropriate response to certain philosophical questions. As to the fear that our explanation will hang in the air unsupported, Wittgenstein responds that no further explanation is needed “unless we require it to prevent a misunderstanding” (PI §87). These ordinary responses invite the reply that the philosopher requires

something more demanding than what would satisfy the naif. Wittgenstein’s task is to show that the something more demanding is an illusory goal. The philosopher needs to see that the naive responses are fully satisfactory. But these responses can be seen as satisfactory only if the import of the questions shifts in some way. What Wittgenstein must do therefore is to discredit the very problems that make naive answers seem inadequate. How can the appeal to ordinary doubts or to our training or to our actual needs come to seem legitimate philosophically? Aren’t these only pragmatic responses to theoretical problems?