One of the fundamental debates over how to understand Wittgenstein’s Philosophical Investigations concerns whether he is developing theories of language and mind that are alternatives to traditional theories. His explicit metaphilosophical remarks state otherwise. Famously Wittgenstein insists that he offers no theories, provides no explanations of the phenomena with which philosophers have been preoccupied. Still less can philosophy reform language. Rather “philosophy may in no way interfere with the actual use of language; it can in the end only describe it” (PI §124). Yet much in Wittgenstein’s later writings can plausibly be taken to offer an alternative conception of language and mind. So, on the one hand, we find constructive interpretations that can only downplay the significance of the metaphilosophical remarks. And, on the other hand, we find “the new Wittgenstein” of the resolute reader, who privileges the metaphilosophical remarks.1 The constructive reader tends to view Wittgenstein’s style of philosophizing as something more idiosyncratic to the man and so not essential to pursuing his critical and positive insights. Wittgenstein’s thought is seen as continuous with the philosophical tradition, though offering radically new theories. The quietist, on the other hand, takes the style of arguing to be the essential feature of Wittgenstein’s task, which is to change in a profound way our entire attitude towards the philosophical project. We are to replace the theoretical drive of traditional philosophy with a piecemeal struggle to overcome the urge to theorize. The struggle ends only for the individual philosopher who comes to recognize how misplaced the drive to theory really is. The constructive reader tends to ignore the structure of the Investigations

in favor of finding whatever passages support the theory of language or mind

that he finds in the text. The quietist is sensitive to the order of the passages and the importance of accommodating them as steps in the dissolution of philosophical puzzlement, but he misses the content of Wittgenstein’s later thought. So it seems that we have either a carefully structured sequence of passages designed to lead the philosopher out of the fly-bottle or we have an alternative contentful theory that renders the structure of the work of a decidedly secondary importance. Yet we need not be forced to choose between these two interpretations. A therapeutic methodology of the sort emphasized by Wittgenstein’s metaphilosophical remarks does lie at the heart of his project. But the way in which the quietist privileges such remarks obscures the content of Wittgenstein’s theoretical diagnosis of philosophical illusion, a diagnosis that reveals very different pictures of language and mind. The Philosophical Investigations (especially Part I) is a carefully structured diagnostic work that defends, as part of that diagnostic work, an alternative positive picture of our situation in the world. Structure alone leads to a gnostic reading of Wittgenstein. Content alone leads to submitting to the very philosophical illusions Wittgenstein seeks to dispel. The quietist correctly suggests that Wittgenstein is at his best when confronting an enemy. The constructivist is correct in finding positive views insinuated in and through the diagnostic critique. Wittgenstein, like the great philosophers of the philosophical canon,

introduces a new philosophical problem and a new argumentative strategy. Indeed new modes of argument create recognition of problems that could not be seen before. The Socratic method both reveals the inadequacies of common-sense reliance on particular examples and shows what is required to address this defect, namely, to identify the essences that unite particulars. Descartes’ method of doubt shows the singular importance of answering the opening question of the Meditations-What are the firm and constant foundations of knowledge?—as well as providing the key for answering this question. Locke’s “plain and historical method” is intimately tied to his project of identifying the scope and limits of human understanding. Kant’s new question-What are the conditions necessary for the possibility of experience?—itself opens space for the transcendental deduction as a solution. Each new philosophical question is tied to a potent new form of argumentation. Wittgenstein has also achieved this. Wittgenstein’s targets are the representationalist theory of language and its

close ally the Cartesian theory of mind. Wittgenstein’s arguments are not only therapeutic, they are diagnostic. He aims to show how philosophers are in the grip of certain pictures of language and mind and the consequences this has for their theories. At the foundation of philosophical theorizing are confusions about the role normative similarity plays in our language games. Whether asserting that the ball is red or continuing with the natural number sequence, the philosopher is blind to his assumptions of sameness that make such assertions and continuations transparent. The Cartesian theory of mind

suffers from a similar blindness, whether we are expecting someone to tea or feeling pain. Wittgenstein’s diagnostic method is tied to his use of two powerful argu-

ments. The first are what I call his “conflation arguments.” The representational picture and the Cartesian picture are ushered in through grammatical errors. The most important of these errors is that of conflating the means of representation with the object of representation. These arguments pair, diagnostically, with “paradox arguments.” Paradox arguments reveal the hidden contradictions that result from the confusions. Wittgenstein uses four pairs of these arguments in his examination of representationalist and Cartesian theories. The structure of Part I is thus more complex than has been appreciated. In the process of diagnosing and refuting representationalist theories,

Wittgenstein removes these theories from view, revealing an alternative picture of how we are situated in the world. In the clearing, many things that had a secondary significance, if any at all, are now seen to be vital and important. Three important positive themes will be developed. First, Wittgenstein’s critique is not directed solely against representationalist theories of language and mind. He is critical of overly intellectualized theories in general. So, Davidson’s interpretation-based theory of meaning is one that succumbs to Wittgenstein’s critique of interpretation. Brandom’s inferentialist semantics, though not representational, nonetheless is a highly intellectualized picture of our use of language. Secondly, Wittgenstein takes the situation of initiate learning as having central methodological and explanatory importance. Every major problem and line of argument begins with a description of the child learning, whether it is the acquisition of a first language or initiation into arithmetic. The learning situation provides a window onto the structure of normativity, particularly the normativity of our very basic judgments of world or mind. Recognizing this entails recognition of the social character of language and mind. This leads to the third main theme of the book. Language and mind are visible only against a background that cannot itself be representational. Wittgenstein identifies three dimensions of this background: stage-setting, mastery of techniques, and certainty. Representational uses of language, indeed any use of language, requires a context in which a particular use of language occurs, the exercise of techniques for going on in the same way, and a shared sense of the obvious and the certain.

Since I think that the Philosophical Investigations is a highly structured argumentative text directed to pursuing a fundamental new problem in philosophy, let me begin with Wittgenstein’s metaphilosophical remarks. These remarks PI §§89-133 occur in the interlude between the discussion of representationalist theory of meaning and rule-following passages.2 It is my

view that most commentators misinterpret Wittgenstein on this issue, but that is not a culpable misinterpretation so much as one that results from a certain failure of self-understanding on Wittgenstein’s own part. I am going to pursue this issue by way of two questions: one interpretative and the other textual. The interpretative question asks whether Wittgenstein’s explicit metaphilosophy coheres with his actual method. I shall make two points here. First, I shall bring out the ways in which Wittgenstein’s metaphilosophical remarks are more complicated than they are often presented as being. Second, I shall argue that some of the most emphasized passages do conflict with Wittgenstein’s actual methods of argument. Wittgenstein’s arguments are in the service of a theoretical diagnosis of philosophical error that relies on the careful structuring of Part I of the Investigations. This structure, I shall argue, provides an answer to the textual question: Why do the metaphilosophical remarks occur where they do, that is, after a full examination of the relation of words to objects and before the rule-following considerations? Neither the constructive reader nor the quietist reading has an answer to this question. Most sketches of Wittgenstein’s later method draw heavily on these pas-

sages. Let me highlight the elements of this familiar approach.3 Wittgenstein’s new philosophy is an activity, not a set of theses, doctrines or theories. It is an activity that aims at showing that traditional philosophical theory-building is “a house of cards,” constructed with illusory materials in pursuit of ends that are themselves the result of confusion and mistaken understandings of the grammar of our language. The aim of the new philosophy is to bring philosophy to an end by disclosing these confusions and mistakes. The primary method for achieving this end is a method of description. Through the perspicuous display of timely reminders and descriptions of language in use, the theoretical and explanatory aspirations of traditional philosophy are revealed to overshoot the limits of language and so say nothing. The philosophical point of the descriptions and reminders is given by their relation to purported philosophical problems. Philosophy for the later Wittgenstein is thus wholly critical and reactive. It gets its point from the entanglement in confusion and illusion that results when individuals make reflective use of language to misguided ends. The metaphor of disease or mental illness is a crucial component of this reading. As with neurosis, the only way to free oneself from the trap is by coming to realize that its problems and solutions are illusory, grounded in a willful and intellectualized imagination. There is nothing positive for philosophy to do, even though it will remain an ineradicable part of human life.4

There is no question that the target of this dismal assessment is the scientistic optimism of the logical positivist movement in Anglo-American philosophy, the twentieth-century embodiment of perennial philosophical error. Much in this sketch is certainly to be found in Wittgenstein, both in his

metaphilosophy and in his actual method. But if we have only this, much of the structure of the Investigations is missed and the content of the work

remains unassimilated. So, let’s look more closely at these passages. PI §§89133 has three major stages. It opens with an initial contrast between the logical method of the Frege-Russell-Tractatus approach and Wittgenstein’s new grammatical method (§§89-91). The second stage characterizes the new method at work (PI §§92-120). This grammatical investigation aims at providing a theoretical diagnosis of how traditional philosophy goes wrong. These passages describe a dialectical structure, in which an initial diagnosis of mistake (§§92-6) is followed by a closer examination of the phenomenology of philosophical puzzlement (§§97-103), which then permits a deeper diagnosis of philosophical error, one showing how the mistake and phenomenology of puzzlement connect (§§104-20). The final stage is a summing-up of the tools by which error can be removed (PI §§121-33). Stage 1. Preliminaries: logical investigation vs grammatical investigation

(PI §§89-91). The opening passages contrast the “logical investigation” of contemporary philosophy that “explores the nature of all things” (PI §89) with the new “grammatical investigation” that “sheds light on our problem by clearing misunderstanding away” (PI §90).5 A logical investigation sees “logic at the bottom of all the sciences” (PI §89); it is the science, as it were, of “the ‘possibilities’ of phenomena” (PI §90). As such, no empirical matters bear upon its success or failure. “It is of the essence of our [logical] investigation,” Wittgenstein tells us, “that we do not seek to learn anything new by it” (PI §89). A logical investigation seeks to give an account of the possibilities of phenomena in terms of logic. A grammatical investigation aims to show that this very goal is a mistake, one that can be cleared away by reminding ourselves “of the kind of statement that we make about phenomena” (PI §90). Demonstrating grammatical confusion is one of the motivations that

inform logical investigation as well as Wittgenstein’s new method. Central to the philosophical project of the Tractatus and logical positivism is the elimination of metaphysics as nonsense, by revealing misunderstandings concerning the use of words. Some misunderstandings, Wittgenstein agrees, “can be removed by substituting one form of expression for another; this may be called an ‘analysis’ of our forms of expression” (PI §90). Wittgenstein must distinguish his own new project, a grammatical investigation, from the analytic project of a logical investigation. Both purport to undermine certain forms of philosophical theorizing by showing how such theorizing is grounded in a misunderstanding of the grammar of our ordinary language. The difference, Wittgenstein tells us, is that in using analysis as the means for displaying misunderstandings, “it may come to look as if there were something like a final analysis of our forms of language, and so a single completely resolved form of every expression” (PI §91). In this way, the method for revealing misunderstandings itself becomes the key component of the very philosophical project under attack. At the end of these passages, Wittgenstein will express the difference between logical and grammatical

investigations by replacing the goal of “complete exactness” (PI §91) with one “aiming at … complete clarity” (PI §133). We see that in the opening passages Wittgenstein distinguishes his new

grammatical investigation from the logical investigation exemplified in the Tractatus while acknowledging similarities that might obscure the key differences. For both methodologies, no new empirical facts are to be uncovered (PI §89). Both seek to clear misunderstandings away by means of a philosophical investigation of language (PI §90). And both “are trying to understand the essence of language” (PI §92). But, and here is the difference that makes all the difference, a grammatical investigation aims to show that the logical (or analytic) investigation itself is ensnared in the very misunderstandings that it purports to clarify. To show this requires considerable diagnostic work. Yet it is the key to Wittgenstein’s critique of the Tractatus. Where the Tractatus sees confusion in metaphysical statements, the Investigations sees the same confusion at work in the method of analysis.6 This means that Wittgenstein’s explanation of philosophical mistake is quite different from the explanation offered by the positivists including the author of the Tractatus. Stage 2. diagnosis: misunderstandings and bewitching pictures (PI §§92-

120). The logical investigation of the Tractatus identifies the mistake of metaphysical theorizing with its trying to say what cannot be said. Its method of logical analysis “eliminates misunderstandings by making our expressions more exact” (PI §91). The success of this method supports the idea that we can achieve “a state of complete exactness,” that will reveal the limits of thought. This eliding of the limited success of logical analysis with an ideal of complete exactness for any possible system of representation is the initiating mistake. It leads to seeing “in the essence [of language], not something that already lies open to view and that becomes surveyable by a rearrangement, but something that lies beneath the surface” (PI §92). This misunderstanding of the scope of logical analysis, combined with the “enormous importance attaching” to language, leads to a search for “a pure intermediary between the propositional signs and the facts” (PI §94)—the proposition.7 This “seduces us into thinking that something extraordinary, something unique, must be achieved by propositions” (PI §93). The analysis of language, the means by which illusory theory building was to be eliminated, is taken to require a hidden logical structure that ensures that “[t] hought, language, now appears to us as the unique correlate, picture, of the world” (§96). This illusion can only be dispelled by recognizing that “when we say, and mean, that such-and-such is the case, we-and our meaning-do not stop anywhere short of the fact; but we mean: this-is-so” (PI §95). This requires turning the apparent inadequacy of this remark into disillusionment with the method of logical analysis. This explains why we find the first of Wittgenstein’s distinctive paradox arguments occurring in these metaphilosophical reflections. The method of logical analysis incurs paradox of its own.