HANS‐JOHANN GLOCK
This chapter canvasses the main features of Wittgenstein’s conception of philosophy, both early and late. It also assesses these features for their merits, partly with a view to current debates. I shall argue that his radical position is more than a whimsical manifestation of an anti‐scientific ideology: it is supported by arguments deriving from astute observations about the peculiar character of philosophical problems on the one hand, and logico‐semantic ideas on the other. In particular, I shall defend Wittgenstein’s claim that the distinctive task of theoretical philosophy is a priori and hence conceptual. The chapter also diagnoses three tensions in Wittgenstein’s account of conceptual elucidation: (i) treating it as a kind of (psycho‐) therapy or propaganda for a particular point of view vs. regarding it as a type of dialectic argument; (ii) insisting on it having a purely critical purpose in dissolving philosophical puzzles vs. allowing for a more positive project of conceptual self‐understanding; (iii) rejecting systematic theories vs. envisaging systematic surveys of our conceptual scheme. It urges, moreover, that these tensions should be resolved in favor of the second members of these pairs of alternatives. Finally, in line with my contention that Wittgenstein’s metaphilosophical views are both supported by and in turn support certain of his philosophical views, the chapter casts aspersions on an ambition that Wittgenstein and many of his followers share with other metaphilosophers (notably Descartes, phenomenologists, and logical positivists). According to the “myth of mere method,” one can metaphilosophically reform philosophy by devising procedures for the resolution of philosophical problems that do not in turn depend on contestable philosophical views derived by way of equally contestable methods. Abandoning the myth leads to a more sober and modest conception of philosophy’s “metaphilosophical” reflection on its own nature.
Wittgenstein’s interest goes back to 1912, when he gave a paper “What is Philosophy?” to the Moral Sciences Club at Cambridge. In the preface of the Tractatus he boasted of having provided a “definitive solution” to “the problems of philosophy.” In 1930 he maintained that his “new method” of philosophizing constituted a “kink” in the “development of human thought” comparable to the Galilean revolution in science. To the end of his career he insisted that what mattered most about his work was not its specific results but its new way of philosophizing, a method or skill which would enable us to fend for ourselves (M 322; MWL 5.2; AWL 97; MS 155, pp.73ff; Letter to Moore, 17 June 1941; PPF §202).
Wittgenstein was right to regard his methodological views as novel and radical. They run up not just against the scientific spirit of the twentieth century (CV 6–7), but against a predominant tendency within Western philosophy. Ever since its inception, philosophy has been regarded as a cognitive discipline, one that aspires to knowledge about reality. For Platonists philosophy is an a priori endeavor that scrutinizes not empirical reality, but a world of abstract “ideas,” and grounds our knowledge by deducing all truths from ultimate principles about these entities. For Aristotelians it is continuous with the special sciences because it describes more general and fundamental features of reality – it is the “queen” of the sciences. And according to Locke it is continuous with the sciences because it is their “underlaborer,” removing obstacles in their path. Finally, radical empiricists contend that all disciplines, including philosophy, mathematics, and logic, describe reality on the basis of empirical evidence.
By contrast, Kant conceived his transcendental philosophy as a second‐order discipline. While science describes reality, philosophy is not directly concerned with objects of any kind, whether physical, mental, or abstract. Instead, it reflects on the preconditions of our knowing or experiencing the objects of the material world. In spite of this reflective turn, however, Kant remained within the cognitivist orthodoxy. He insisted that philosophy can deliver “synthetic a priori” truths which express necessary preconditions of experience, e.g., that any experience must be a potential object of self‐consciousness, that substances persist through qualitative change, and that every event has a cause.
Wittgenstein’s conception of philosophy stands in the tradition of Kant. First, for both philosophy is primarily a critical activity which curbs the excesses of metaphysics and clarifies nonphilosophical thoughts (TLP 4.112, 6.53; Kant, [1781/1787] 1998, A 11/B 24–5, A 735/B 763, A 851/B 879). Secondly, inspired by Schopenhauer and Hertz, Wittgenstein draws a Kantian contrast between science, which pictures or represents the world through a posteriori propositions, and philosophy, which reflects on the nature and preconditions of this representation (TLP 4.11ff). This diverges from Frege and Russell, to whom his deliberations are indebted in other respects. Frege never propounded a general conception of philosophy. But he intimated a hierarchy of disciplines in which logic grounds metaphysics and psychology (1893/1966, p.XIX). Russell, throughout his numerous evolutions, held fast to a “scientific conception of philosophy,” according to which it shares the tasks of science and should emulate its methods. And during his middle period he identified philosophy with logic (1914/1993, pp.216–17; 1918, pp.75, 95–119).
In his earliest recorded reflections, Wittgenstein (echoing Frege) claimed that philosophy consists of logic – its basis – and metaphysics, and (echoing Russell) that it differs from science in being the “doctrine of the logical form of scientific propositions” (NL 106). Later he reserves the label “metaphysics” for the illegitimate philosophy of the past. Legitimate philosophy is a “critique of language.” “Most of the propositions and questions to be found in philosophical works are not false but nonsensical” (TLP 4.003; see preface; 3.323–5, 6.51; NB 22.8.14, 1.5.15). They stem from failure to understand the logic of language, a failure that results in asking pseudo‐questions that admit of no answer. The task of philosophy is not to answer these questions, but to show that they violate the bounds of sense.
The Tractatus takes over Frege’s anti‐psychologistic separation of logic from psychology (4.1121, 6.3631, 6.423) and accepts Russell’s identification of philosophy with logic (4.003–1) (see also Chapter 4, WITTGENSTEIN AND FREGE and Chapter 5, WITTGENSTEIN AND RUSSELL). But Wittgenstein’s philosophy of logic departs radically from his predecessors. With considerable chutzpah, he included their avant‐garde formal systems under the label “the old logic,” and castigated them for having failed to clarify the nature of logic (4.003f, 4.1121, 4.126).
At the turn of the century, there were four positions concerning this topic. According to Mill’s radical empiricism, logic consists of well‐corroborated inductive generalizations. According to psychologism, logical truths or “laws of thought” describe how human beings (by and large) think, their basic mental operations, and are determined by the makeup of the human mind. Against both positions Platonists like Frege protested that logical truths are both necessary and objective, and that this special status can only be secured by assuming that their subject matter – logical objects, concepts, thoughts – are abstract entities inhabiting a “third realm” beyond space and time, rather than material objects or private ideas in the minds of individuals. Finally, Russell held that the propositions of logic are supremely general truths about the most pervasive traits of reality.
Wittgenstein eschews all four alternatives (see Chapter 17, LOGIC AND THE TRACTATUS). Necessary propositions are neither inductive generalizations about the world nor statements about the way people actually think. Nor are they about a Platonist hinterworld or the most pervasive features of reality. Philosophy qua logic is a second‐order discipline. “Logic is transcendental” (6.13). Unlike science, it does not itself represent any kind of reality. Instead, it concerns the preconditions of representing reality, just as Kant’s philosophy reflects on the “transcendental” preconditions of experiencing reality. Philosophy is the “logical clarification of thought.” It investigates the nature and limits of thought, because it is in thought that we represent reality. Echoing Kant’s critical philosophy, the Tractatus aims to draw the bounds between legitimate discourse, which represents reality, and illegitimate speculation – notably metaphysics (4.11ff). At the same time, it gives a linguistic twist to the Kantian tale.
Thus the aim of the book is to draw a limit to thought, or rather – not to thought but to the expression of thoughts: for in order to be able to draw the limits of thought, we should have to find both sides of the limit thinkable (i.e. we should have to be able to think what cannot be thought). It will therefore only be in language that the limit can be drawn, and what lies on the other side of the limit will simply be nonsense.
(TLP, preface)
Language is more than a secondary manifestation of something pre‐ or nonlinguistic. For a thought is neither a mental process nor an abstract entity, but itself a proposition (Satz), a sentential sign (Satzzeichen) that has been projected onto reality (3.5f). Thoughts can be completely expressed in language, and philosophy can establish the limits and preconditions of thought by establishing the limits and preconditions of the linguistic expression of thought.
Indeed, these limits must be drawn in language. They cannot be drawn by propositions talking about both sides of the limit. By definition, such propositions would have to be about things that cannot be thought about and thereby transcend the bounds of sense. The limits of thought can only be drawn from the inside, namely by delineating the “rules of logical grammar” or “logical syntax” (3.32–3.325). These rules determine whether a combination of signs is meaningful, that is, capable of representing reality either truly or falsely. What lies beyond these limits is not unknowable things in themselves, as in Kant, but only nonsensical combinations of signs, e.g., “The concert‐tone A is red.” The special status of necessary propositions is due not to the fact that they describe a peculiar reality, but to the fact that they reflect “rules of symbolism” (6.12ff). These cannot be overturned by empirical propositions, since nothing contravening them counts as a meaningful proposition. Logical syntax antecedes questions of truth and falsity, and thereby matters of fact.
Wittgenstein’s “logic of representation” (4.015) comprises the most general preconditions for the possibility of symbolic representation. Consequently, there is no such thing as a logically defective language. Any language, any sign‐system capable of representing reality, must conform to the rules of logical syntax. Natural languages are capable of “expressing every sense.” Therefore their propositions must be “in perfect logical order” just as they are.
They are not in any way logically less correct or less exact or more confused than propositions written down […] in Russell’s symbolism or any other “Begriffsschrift” (Only it is easier for us to gather their logical form when they are expressed in an appropriate symbolism.).
(Letter to Ogden, 10 May 1922; see TLP 4.002, 5.5563)
Ordinary language allows the formulation of nonsensical pseudo‐propositions because it conceals the logical form of propositions: quantifiers look like proper names (“nobody”) or predicates (“exists”), ambiguities lead to philosophical confusions (“is” functions as copula, sign of identity, and existential quantifier), and “formal concepts” like object look like genuine concepts employed in empirical classification, such as apple. To guard against such deception, however, we require not an ideal language capable of expressing things natural languages cannot express, but an ideal notation (Zeichensprache). Such a notation is “governed by logical grammar – by logical syntax” (3.325); it displays the hidden logical form that ordinary propositions possessed all along.
The idea is to express in an appropriate symbolism what in ordinary language leads to endless misunderstandings […] where ordinary language disguises logical structure, where it allows the formation of pseudo‐propositions, where it uses one term in an infinity of different meanings, we must replace it by a symbolism which gives a clear picture of the logical structure, excludes pseudo‐propositions, and uses its terms unambiguously.
(RLF 163)
Persistent misinterpretations notwithstanding, therefore, the Tractatus is not a contribution to an ideal language philosophy in the vein of Frege, Russell, Carnap, and Quine; instead, it is a precursor of the project of a formal theory of meaning for natural languages launched by Davidson.
Wittgenstein’s account of the logico‐semantic system that makes representation possible, his “theory of symbolism,” revolves around his picture theory (see Chapter 8, THE PICTURE THEORY, and Chapter 32, WITTGENSTEIN ON INTENTIONALITY). It distinguishes three types of sentential sign‐combinations. The only strictly meaningful propositions are empirical statements of the kind advanced by science. They have a “sense” by virtue of depicting possible states of affairs. This implies that they are bipolar, capable of being true (if the depicted state of affairs obtains), yet equally capable of being false (if it does not).
The propositions of logic are “tautologies” and “contradictions.” Their necessity simply reflects the fact that through truth‐functional operations they combine bipolar elementary propositions in such a way that all information cancels out. They exclude and hence say nothing, which means (in Wittgenstein’s terminology) that they are “senseless,” i.e., have zero sense (factual content). “It is raining” says something true or false, and so does “It is not raining.” By contrast, “Either it is raining or it is not raining” says nothing about the weather, nor about anything else.
Finally, the pronouncements of metaphysics are nonsensical “pseudo‐propositions.” They try to say what could not be otherwise, e.g., that red is a color, or 1 a number. What they seem to exclude – e.g., red being a sound – contravenes logic, and is hence nonsensical. More importantly still, the attempt to refer to something nonsensical, if only to exclude it (as in Russell’s theory of types), is itself nonsensical. For we cannot refer to something illogical like the class of lions being a lion by means of a meaningful expression. What such philosophical pseudo‐propositions try to say is shown by the structure of genuine propositions – e.g., that in bipolar propositions “red” can combine only with names of points in the visual field, not with names of musical tones.
The distinction between what can be said by meaningful propositions and what can only be shown pervades the Tractatus from the preface to the famous final admonition “Whereof one cannot speak, thereof one must remain silent.” The Tractatus pronouncements themselves are in the end condemned as nonsensical, because they try to articulate metaphysical truths about the essence of language. Such truths, by Wittgenstein’s own lights, cannot be expressed in philosophical propositions, yet they manifest themselves in nonphilosophical propositions properly analyzed.
My propositions serve as elucidations in the following way: anyone who understands me eventually recognizes them as nonsensical, when he has used them – as steps – to climb up beyond them. (He must, so to speak, throw away the ladder after he has climbed up it.) He must transcend these propositions, and then he will see the world aright.
(TLP 6.54)
Readers as diverse as Russell, Ramsey, Carnap, and Neurath have shrunk back from this paradoxical conclusion. Some recent commentators retort that the auto‐da‐fé at the end must be taken literally (see Chapter 10, RESOLUTE READINGS OF THE TRACTATUS). The Tractatus does not consist of “illuminating nonsense,” nonsense that vainly tries to hint at ineffable truths, but of “plain nonsense,” nonsense in the same drastic sense as gibberish like “ab sur ah” or “piggly tiggle wiggle.” The purpose of the exercise is therapeutic. By producing such sheer nonsense, Wittgenstein tries to unmask the idea of metaphysical truths (effable or ineffable) as absurd and to wean us off the temptation to engage in philosophy.
These “resolute” readings are susceptible to both exegetical and substantive criticism (see Chapter 11, INEFFABILITY AND NONSENSE IN THE TRACTATUS). At the same time, we need not rest content with lumbering the Tractatus with the idea of ineffable truths. It is crucial to take seriously its propadeutic aim, explicit in 6.53. The book is self‐defeating, because, in delineating the essential preconditions of representation it violates its own restrictions on what it makes sense to say – the principle of bipolarity. This is a pitfall for any a priori attempt to draw the bounds of knowledge or sense in such a way as to exclude nonempirical knowledge or propositions – witness Kant and the logical positivists. Wittgenstein heroically tried to brave it by violating his self‐imposed prohibitions solely to attain a “correct logical point of view” (4.1213), an insight into the essence and structure of language that would allow one to engage in critical logical analysis, without committing further violations. Once we have devised an ideal notation that displays the logical structure of meaningful propositions, we can throw away the ladder on which we have climbed up, namely the pronouncements on the essence of meaningful propositions that we needed to construct the notation.
From this perspective, Russell’s aspiration to introduce scientific method into philosophy is misguided. Proper philosophy cannot be a doctrine, since there are no philosophical propositions. It is an activity, not of deliberately uttering nonsense with the therapeutic aim of debunking philosophy, mind you, but of logical analysis. Without propounding any propositions of its own, analysis clarifies the logical form of meaningful propositions – i.e., of bipolar empirical propositions – by translating them into the ideal notation. This positive task is complemented by the negative task of demonstrating that the would‐be propositions of metaphysics violate the rules of logical syntax since they resist such translation.
Wittgenstein later took this linguistic turn in a different direction. The core of his method remained the “transition from the question of truth to the question of meaning” (MS 106, p.46). The connection between philosophy and language is twofold. First, philosophy is interested in language because of the latter’s “paramount role in human life” (BT 194–5, 413). There is an internal connection between thought and its linguistic expression. Qua rational beings, humans are therefore at the same time essentially language‐using animals. The second connection is that the a priori nature of philosophical problems and the necessary status of philosophical propositions are rooted in linguistic rules: “Philosophy is the grammar of the words ‘must’ and ‘can’, for that is how it shows what is a priori and what a posteriori” (CE 411). Philosophy is an activity striving for clarity rather than a cognitive discipline (LWL 1; AWL 225; RPP I §115). At the same time Wittgenstein drops the ineffable metaphysics, and he replaces the mere promise of critical analysis by a dialectic practice: philosophy dissolves the conceptual confusions to which philosophical problems are alleged to owe their existence.
This noncognitivist picture chimes to varying degrees with other analytic critiques of metaphysics, notably those of some logical positivists (Schlick, Waismann, Carnap) and of some conceptual analysts (Wisdom, Ryle) (see Chapter 46, WITTGENSTEIN AND THE VIENNA CIRCLE and Chapter 47, WITTGENSTEIN AND ORDINARY LANGUAGE PHILOSOPHY). But it appears to impoverish philosophy, and is generally considered to be the weakest part of Wittgenstein’s later work – slogans unsupported by argument and belied by his own positive “theory construction” which can be isolated from the rest (e.g., Dummett, 1978, p.434). Wittgenstein’s methodological views must ultimately be judged by their results – the proof of the pudding is in the eating. But the impression that they are unsupported by argument arises from failure to recognize the connections which hold, firstly, between various aspects of his conception and, secondly, between his conception of philosophy as a whole and other parts of his later philosophy, notably his account of logical necessity and his understanding of language or grammar. Indeed, they arise from a coherent line of thought that can be reconstructed along the following steps:
In what follows, I shall examine these specific aspects of Wittgenstein’s conception of philosophy (sections 3.1–3.6). The final sections (3.7–3.8) are devoted to two wider issues.
What links Wittgenstein’s philosophizing with the metaphysical tradition is that both aim to resolve the problems that constitute the subject matter of philosophy (PG 193; BT 416, 431; Z §447; PLP 6). Wittgenstein suggests his “new method” as a new way of dealing with these problems. That method is superior because it is based on Wittgenstein’s phenomenology of philosophical puzzlement, which furnishes a better understanding of the character of philosophical problems (see LWL 1; PG 193; AWL 27–8; M 113–14). In the main, the problems concerned are those of theoretical philosophy – logic, metaphysics, epistemology, philosophy of mind (BB 62; RW 160; M 314–5; CV 25). Wittgenstein illustrates their peculiar nature by reference to Augustine’s question “What is time?.” Such problems concern not arcane phenomena but notions we are familiar with in nonphilosophical (everyday and specialized) discourse. Nonetheless they have proven intractable, as is evident from philosophy’s dismal failure to make progress in its ambition to fathom the nature of reality (see BT 424 and below).
Wittgenstein explains this enigmatic combination by insisting, against empiricism and naturalism, that philosophy is a priori (LWL 79–80; AWL 3; see 97, 205). Philosophical problems cannot be solved by empirical observation or scientific experiment, since they concern concepts mastery of which is a precondition of establishing new empirical facts; at the same time, such mastery does not guarantee the kind of comprehension required to avoid philosophical puzzlement (PI §89; see §§95, 428; BB 30–1; BT 435; RPP II §289; Z §452; CV 4).
One might retort with Russell (1912/1967, p.90; 1956, p.281) that philosophy is a proto‐science, dealing with questions not yet amenable to empirical methods. “Kicking upstairs” (Austin, 1961/1970, pp.231–2) topics like infinity, matter in motion, types of learning, forms of theoretical and practical reasoning, or linguistic universals by passing them on to specialized disciplines is a role that philosophy as an academic discipline has fulfilled admirably. Whether or not it is their queen, philosophy is the mother of all (non‐applied) sciences, even though its children are rarely grateful to their parent. Nevertheless, the fact that the special sciences developed out of philosophy does not entail that the questions that exercise philosophy are after all invariably empirical. For while the topics may be shared, they can give rise to distinct kinds of problems.
Some problems have remained within the purview of philosophy ever since its inception. Among them are problems that concern topics investigated by independent academic disciplines. Accordingly, disciplinary secession from philosophy is no panacea for philosophical perplexities (see Hacker, 2009, pp.131–3). Finally, at present philosophy as a distinctive intellectual pursuit is constituted at least in large part by such problems. These include questions such as “What is truth?,” “Is knowledge possible?,” “How is the mind related to the body?,” and “Are there universally binding moral principles?.” These puzzles are of a peculiar kind. They continue to defy the otherwise highly successful methods of empirical science. What is more, in many cases at least there are principled reasons for this failure. For instance, there is a difference between the questions “What is true?” about a particular topic and “What is truth?,” i.e., how is that notion to be explained; one cannot without circularity allay skeptical doubts about the possibility of empirical knowledge by appeal to empirical scientific findings; one cannot on pain of a naturalistic fallacy deduce the (non)existence of normative principles from the (non)existence of a moral consensus among human beings; and so on.
Such considerations provide at least a prima facie case for regarding philosophy as a priori in a minimal sense: the philosophical questions and disputes concern not the empirical findings themselves, but at most the relevance the latter have for such problems. Thus the discoveries of the neurosciences, impressive though they are, have not simply solved either the mind–body problem or the problem of free will. Instead, they have provoked fresh disputes about the relation between mental and neurophysiological phenomena and about notions like decision, liberty, responsibility, and rationality.
Finally, Wittgenstein argues powerfully against attempts to reduce the necessary propositions of logic, mathematics, and metaphysics to empirical generalizations; in the course of doing so he even managed to anticipate weaknesses in Quine’s renowned attacks on analyticity, apriority, and necessity (see Glock, 1996, pp.129–35). Wittgenstein has been accused of indulging in armchair science, yet he would respond that it is scientistic philosophers who engage in an incoherent discipline – empirical metaphysics, the attempt to invoke empirical data to resolve conceptual issues that are prior to experience.
Wittgenstein was hostile to the scientific spirit of the twentieth century, which he deplored in Russell and the logical positivists. He rejected the belief in progress, and abhorred the “idol worship” of science as both a symptom and a cause of cultural decline (CV 6–7, 49, 56, 63; RW 112).
However, it is imperative to distinguish between Wittgenstein’s personal ideology and his philosophical methodology. The latter rejects not science but scientism, the imperialist tendencies of scientific thinking that result from the idea that science is the measure of all things. Wittgenstein insists that philosophy cannot adopt the tasks and methods of science. There should be a division of labor between science and philosophy’s reflection on our conceptual apparatus (CV 16), a division that is difficult to uphold given the twentieth‐century obsession with science (PR 7; BB 17–18). Accordingly, Wittgenstein’s notorious prohibition of theories, hypotheses, and explanations in philosophy (PI §§109, 126, 496; RFM VI §31) does not evince a general irrationalism.
Causal explanations of empirical phenomena are of course legitimate, yet their place is in the nomological sciences (see Chapter 37, WITTGENSTEIN ON CAUSATION AND INDUCTION). They are banned from philosophy, on the grounds that they are irrelevant to the solution of problems that are conceptual rather than factual (Z §458; CV 79). Furthermore, Wittgenstein’s animadversions against explanation in philosophy must be interpreted with care, since his own philosophizing features two types of non‐nomological explanations. First, it provides etiological explanations that pinpoint the sources of philosophical confusions (see below). More importantly, it revolves around explanations of meaning. These are not explanations of why we use a certain term, or of what the (perlocutionary) effects of using it are on particular hearers, but of how we use it correctly, i.e., they specify “grammatical” rules for its established use (PI §§120, 491–8). Such explanations are not, therefore, incompatible with the idea that philosophy is descriptive, in the sense of articulating the rules that guide our linguistic practice.
But why should philosophical investigations be conceptual to begin with? Wittgenstein’s general answer derives from his conventionalist account of apriority and necessity. The a priori and necessary status of the propositions sought by metaphysics is due to the fact that they reflect concepts and conceptual connections. Furthermore, our conceptual scheme is embodied in our language. What Wittgenstein (misleadingly) calls the “grammar” of a language is not confined to morphology and syntax; it is the overall system of logico‐semantic rules, of the constitutive rules that determine what it makes sense to say in that language (PR 51; LWL 46–59; PG 60, 133, 143; PI §496). Therefore, explicating our concepts and thereby our “form of representation” takes the form of articulating linguistic rules through “grammatical propositions” (see Chapter 14, GRAMMAR AND GRAMMATICAL STATEMENTS).
Grammatical propositions antecede experience in an innocuous sense (PR 143; LWL 12; AWL 90). They can neither be confirmed nor confuted by experience. “Black is darker than white,” for instance, cannot be overthrown by the putative statement “This white object is darker than that black object,” since in established usage nothing counts as being both white and darker than black. This antecedence to experience renders intelligible the apparently mysterious “hardness” of necessary propositions (PI §437; RFM I §121; PG 126–7). To say that it is logically impossible for a white object to be darker than a black one is to say that given our semantic rules, it makes no sense to apply “white” and “darker than black” to one and the same object.
Defending this position is beyond the remit of this chapter (see Chapter 21, NECESSITY AND APRIORITY; cf. Kalhat, 2008). In any event, the idea that philosophy seeks to analyze or define concepts rather than to decide what they actually apply to on the basis of experience has a venerable pedigree. Ever since Socrates, philosophers have been concerned with “What is X?” and “What are Xs?” questions, e.g., “What is justice?,” “What is knowledge?,” “What is truth?.” In response to these questions, they have traditionally sought analytic definitions of X(s). Such definitions specify conditions or features which are individually necessary and jointly sufficient for being X. Furthermore, these features should not just in fact be possessed by all and only things that are X; rather, only things possessing all of the defining features can be X, and anything possessing them all is ipso facto X.
Admittedly, questions of the “What is (are) X(s)?” form can be requests for empirical information about contingent features of X(s). But as posed in philosophical reflection they are not directed at contingent features of X(s), features that X(s) may or may not possess and that need to be established empirically by looking at these instances. They are directed instead at the nature or essence of X, at what makes something an X in the first place. That kind of question is properly answered by an explanation of what “X” means. For such an explanation specifies what counts as X, or what it is to be an X, independently of features that X(s) may or may not possess. Similarly for questions of the form “What makes something (an) X?”: these can be requests for a causal explanation that specifies how come that certain objects are X. But they can also be requests for a semantic explanation that specifies what constitutes being an X, i.e., conditions by virtue of which something qualifies as X in the first place. As regards their explanatory role in actual practice, there is no difference between what Carnap called, respectively, the “material” and the “formal mode.” It matters little, for instance, whether we answer the question “What is a drake?” by saying that a drake is a male duck or by saying that “drake” means male duck. (The essentialism of Kripke and Putnam creates a gap between nature and meaning; but it is subject to Wittgensteinian objections, e.g., Hanfling, 2000, ch.12; Glock, 2003, ch.3.)
At the same time, the Witttgensteinian idea that philosophy is simply “grammatical investigation,” i.e., conceptual analysis, stands in need of qualification and modification. First, many questions that are of supreme philosophical relevance are not simply of the definitional “What is X?” type. Consider the question of whether nonhuman animals possess mental powers. In tackling it we must pay heed to the conditions for the applicability of mental terms. And our knowledge of what these conditions are is a priori in that it is independent of empirical knowledge whether these conditions are fulfilled in particular cases. At the same time, it is obvious that the question to which creatures these terms actually apply also hinges on contingent facts about these creatures to be established empirically.
But couldn’t one extract exclusively conceptual questions? For instance: Does it make linguistic sense to apply mental expressions to animals? Would anything count as an animal thinking that something is the case? Such an approach might be feasible in principle. Yet it remains difficult to see how one might make progress with these questions without at least considering ethological findings as a heuristic device. The complexity and flexibility of some animal behavior alerts us to the possibility that relatively advanced mental phenomena can be manifested without linguistic expression, contrary to the convictions of many Wittgensteinians. Furthermore, a purist isolation of conceptual issues is barren in at least one respect. Even philosophers are interested in the question of what mental capacities animals actually have. Unsurprisingly, since it is that issue which has so many important implications inside and outside of philosophy.
Finally, we must avoid the Socratic mistake of thinking that a cast‐iron definition of “X” is needed in advance of building empirical theories about X. What is required is a preliminary understanding of “X”; and this understanding is subject to critical elucidation in philosophical reflection and modification in the course of scientific theory‐building. Wittgenstein was fully aware of the pitfalls of the Socratic stance (TS 302, p.14; PG 121–2; see Chapter 25, VAGUENESS AND FAMILY RESEMBLANCE). But he was reluctant to acknowledge that philosophical reflection can and must be sensitive to scientific claims, provided that it is to contribute to a full understanding of topics that raise both scientific and philosophical problems. The interaction of conceptual and factual aspects implies that the division of labor between the conceptual clarification of philosophy on the one hand and the factual discoveries and theory formation in science on the other requires dynamic interaction rather than splendid isolation.
According to the Tractatus there are metaphysical truths, albeit ineffable, about the logical structures shared between language and reality. By contrast, the later Wittgenstein demythologizes metaphysics (LWL 21; MS 157b, p.4; see Chapter 12, METAPHYICS: FROM INEFFABILITY TO NORMATIVITY). It is constitutive of metaphysics that it conflates factual (sachliche) and conceptual issues, as well as scientific theories (or hypotheses) and norms of representation (Z §458). Metaphysics claims to establish true propositions about the essence of reality. Its propositions have the form of statements of fact. Science teaches us that no human can run faster than 40 kilometers per hour, or that there is no intra‐mercurial planet; metaphysics that no human can have the pains of another, or that there is no uncaused event. According to Strawson (1959), such pronouncements belong to “descriptive metaphysics,” since they explicate our conceptual framework. According to Wittgenstein, they are grammatical rules – often distorted – in propositional disguise (BB 18, 35; AWL 18, 65–9; WVC 67). By these lights, “Every event has a cause” is a rule that partly determines what counts as an “event.” But such a diagnosis is inaccurate. Our conceptual scheme does not simply rule out as nonsensical the expression “uncaused event.” Let’s assume that one morning we find dinosaur footprints on the ceiling. Let’s further assume that we have a reason to abandon the search for an explanation of the footprints, such as that the laws of nature not only fail to provide one but also suggest that none is to be had (the example of quantum mechanics shows that this is at any rate a possibility). Even in that case, we would not cease to call the appearance of the footprints an event. A physical change would be an event, even if a causal explanation of it could be ruled out ab initio. Consequently, being caused is not part of our explanation of the term “event,” or of the linguistic rules governing its use.
Kant was right, therefore, to deny that the law of causality simply explicates the concept of an event. Insofar as it is less susceptible to falsification than common‐or‐garden empirical generalizations it is by virtue of being a regulative principle. A qualified version of it may also possess a constitutive role in our conceptual scheme (see Glock, 2012). First, we need to acknowledge – contra Kant and Wittgenstein – that not all loosely speaking conceptual truths are trivial. Next, some conceptual truths are nontrivial because they are not definitional. The connection between the constituent concepts of such propositions is provided by their complex interplay with other concepts that do not themselves occur in the proposition. Thus Strawson has argued powerfully that most events must be caused, not because random changes do not qualify as events, but because persistently chaotic events are not possible objects of self‐conscious experience.
What about “revisionary metaphysics”? A specimen Wittgenstein considers is the solipsist who insists “Only my present experiences are real!” (see Chapter 9, WITTGENSTEIN ON SOLIPSISM). Sentences of this kind are not disguised grammar, but either nonsense or “expressions of discontent with our grammar” (BB 55–7). Yet the attempt to improve our conceptual scheme by aligning it to reality is misguided. Grammar is “autonomous,” not responsible to putative essences (see Chapter 15, THE AUTONOMY OF GRAMMAR). Consequently, there are no metaphysical grounds for either defending or reforming our conceptual scheme.
Empirical science operates with concepts. It decides whether they in fact apply to certain phenomena; it also provides causal explanations of how phenomena come to satisfy these concepts. By contrast, philosophy clarifies concepts, if and when the need arises, notably by investigating their conditions of application, the conditions that something must fulfill to satisfy these concepts.
In addition to the application and the elucidation of concepts, there is also conceptual construction or concept formation, the devising of novel conceptual structures. This activity is one of the hallmarks of mathematics, which invents novel formal tools for describing and explaining empirical phenomena. Of course, there is concept formation outside mathematics as well. It plays an indispensable role in the empirical sciences, notably when these develop new paradigms during scientific revolutions. Concept formation also features in other forms of discourse, ranging from religion through morality to the historical and social sciences. It occurs whenever new ways of classifying or explaining phenomena, of thinking about or making sense of them are introduced. Philosophy is no exception. The concepts of family resemblance and language‐game, for instance, are no less philosophical innovations than analyticity and apriority. The moot question is whether concept formation has the same purpose and importance in philosophy as in empirical science, namely of furnishing novel tools for classifying and explaining phenomena in the world.
The Tractatus insisted that language must be governed by a complex system of exact rules if it is to represent reality. After his return to philosophy Wittgenstein came to reject this imputation as “dogmatic” (PI §§81, 92–7, 108–9, 131). Speaking a language is not operating a calculus of arcane rules (WVC 77; LWL 16–17; PG 114–15; PI §§126–9). By this token, both the Tractatus and contemporary formal semantics go wrong in seeking “deep” or “unheard of” discoveries through logical analysis. There are no “surprises” in grammar (WVC 77; LWL 16–17; BT 418–19, 435–6; PG 114–15, 210; PI §133; MS 109, p.212; MS 116, pp.80–2). “What is hidden, is of no interest to us” (PI §126). As competent speakers we are already familiar with the grammar of our language. Alas, we can be tempted into ignoring or distorting grammar. One cause of such lapses is misleading analogies and pictures suggested by the surface‐grammar of language; a whole “mythology” is laid down in our language (GB 133; BT 433–5; PI §§422–6; OC §90; MS 110, p.184). Another cause is general intellectual tendencies such as a science‐induced “craving for generality” (BB 17–18) or an urge to seek further explanations and justifications when “we ought to look at what happens as a ‘proto‐phenomenon’” and simply to note “this language‐game is played” (PI §§654–5). The antidote, however, consists neither in replacing ordinary language by an ideal one nor in uncovering the logical syntax disguised by the school‐grammatical surface with the aid of logical analysis. Instead, what is needed is grammatical reminders of how we use words outside philosophy “It makes sense to say ‘I know that she has toothache”’ or “A dog cannot be said to believe that its master will return in a week.” These are articulations of rules that we have been following all along. Their point is to draw attention to the violation of grammar by philosophers. They are part of a dialectical critique of sense, an “undogmatic procedure” that contrasts with the dogmatic insistence of the Tractatus that only certain combinations of signs can make sense because of the constraints imposed by the picture theory (WVC 183–6; see PR 54–5; PI §§89–90, 127; BT 419, 424–5 and below). Wittgenstein tries to show that his interlocutors use words according to conflicting rules, without relying on contentious views of his own.
Wittgenstein and his followers have provoked the complaint of setting themselves up as “guardians of semantic inertia” (Gregory, 1987, pp.242–3) who criticize philosophical and scientific theories as confused simply because they diverge from ordinary use. These complaints ignore that by “ordinary use” they do not necessarily mean everyday use; instead, they mean established use, whether it be in common parlance or in technical forms of discourse with a tightly regimented vocabulary (see Ryle, 1971b, ch.23). Thus, Wittgenstein does not extol the virtues of everyday over technical language, or of the mundane everyday over the sophisticated specialized employment of a term. Nor does he prohibit the introduction of technical terminology in either science or philosophy. He refrains explicitly from criticizing philosophical positions merely for violating “common sense” (BB 48–9, 58–9) or for employing novel terms or familiar words in ways that differ from the established patterns of use (see PI §254; RPP I §548; RPP II §289; LPP 270).
Rather, Wittgenstein insists that such novel terms or uses need to be adequately explained by laying down clear rules. He further alleges that metaphysical questions and theories – no matter whether propounded inside or outside of the academic discipline of philosophy – get off the ground only because they employ terms in a way which is at odds with their official explanations, and that they trade on deviant rules along with the ordinary ones. In effect, Wittgenstein tries to confront metaphysicians with a trilemma: either their novel uses of terms remain unexplained (unintelligibility), or it is revealed that they use expressions according to incompatible rules (inconsistency), or their consistent employment of new concepts simply passes by the ordinary use – including the standard use of technical terms – and hence the concepts in terms of which the philosophical problems were phrased (ignoratio elenchi).
On his return to philosophy Wittgenstein castigated as “hellish” Moore’s idea that logical analysis is required to establish what, if anything, we mean by our propositions (WVC 129–30). A “correct logical point of view” is achieved not through a quasi‐geological excavation, but through a quasi‐geographical “overview” (Übersicht; see Chapter 16, SURVEYABILITY). Such an overview displays, in a synoptic fashion, features of our linguistic practice that lie open to view. It is vain to hope for a decomposition of propositions into ultimate components, or even for detecting a single definite structure in them. Insofar as analysis is legitimate, it either amounts to the description of grammar, or to the substitution of one kind of notation by another, less misleading one (PR 51; WVC 45–7; BT 418; PI §§90–2). But Wittgenstein’s only example of the latter method is a notation which paraphrases “is” by either “=” or “∈” (TS 220, §§98–9). And the former method is a version of “connective analysis” in Strawson’s sense (1992, ch.2): the explanation of concepts and the description of conceptual connections by way of implication, presupposition, and exclusion.
Wittgenstein’s case for connective analysis and against philosophical discoveries and surprises in grammar rests on his critique of the “calculus model” of linguistic meaning and understanding. He attacks the view – widespread within contemporary theories of language – that normal speakers follow complex logical and semantic rules of which they have tacit knowledge even though they can never become aware of these rules (see also Baker and Hacker, 1984, chs 8–9; Searle, 1997). Its central shortcoming is that it ignores the difference between following a rule and merely acting in accordance with it. The former presupposes that the rule provides the agent’s reason for acting as she does. But while the physiological causes of speech and understanding may be completely unknown to speakers, this cannot hold of their reasons.
Wittgenstein knew from personal experience that the dissolution of conceptual confusions can be as complex as the knots it unties (Z §452). He sometimes maintains, however, that the articulations of grammatical rules that play a decisive role in this process are platitudes and trivialities (see below). Yet he himself, not to mention Ryle, taught us that people often follow rules without explicitly consulting them, as in the case of proficient chess players. In most cases, they will be able to specify these rules when prompted. There are exceptions, however. For example, competent speakers may be incapable of explaining the difference between pairs like “automatically” and “inadvertently,” “bottle” and “jar,” “almost” and “nearly” (Rundle, 1990, ch.4). More to the philosophical point, they may be at a loss to explain their use of the definite article, and of the subjunctive, or the sequence of tenses in conditionals. Wittgenstein would accept examples as adequate explanations. But in some cases even they may not be forthcoming. What he would rightly insist on is that speakers must nevertheless be capable of recognizing the correct formulations of the relevant rules, if only with a little help from their friends. Even this potentiality is absent in the case of many of the rules featuring in formal theories of syntax and semantics. Indeed, many speakers of natural languages are incapable of as much as learning the recherché rules thereby imputed to them, even when these are expressed in a less formal way. This means that there is not even a minimal sense in which such rules guide their linguistic behavior (Glock, 2003, pp.244–9).
The bottom line is this: insofar as language is governed by grammatical rules, these are not simply open to view. As Wittgenstein himself remarked: “The aspects of things that are most important to us are hidden because of their simplicity and familiarity” (PI §129). Grammatical rules need to be made explicit. This is not a matter of gathering new information about reality – as competent speakers we have all the information we need. But it is a matter of spelling out the “participatory knowledge” we have by virtue of having mastered a natural language (see Hare, 1960; Hanfling, 2000, pp.52–5). It requires elicitation, reflection, and articulation; and it may involve trial and error, as the history of conceptual analysis amply demonstrates.
According to step (C) (section 3.3), philosophy is concerned with grammatical rules that it can neither justify nor reform but only describe; according to step (D) (section 3.4), these descriptions are reminders and may be downright trivial. This raises the question: what are grammatical reminders good for? One disparaging suspicion is: very little!
Philosophy may in no way interfere with the actual use of language; it can in the end only describe it. For it cannot give it any foundation either. It leaves everything as it is.
(PI §124)
This dictum has been read as licensing a type of intellectual quietism (Blackburn, 1984, p.146). However, Wittgenstein does not deny that language changes (PI §18). And there are nonphilosophical grounds for conceptual change, e.g., in science. His point is that it is not philosophy’s business to bring about such reform by introducing an ideal language. More importantly, Wittgenstein does not purport to leave philosophy as it is. Instead, he tries to reveal it as “plain nonsense” and “houses of cards” (PI §§ 118–19; BT 413, 425). This leaves the possibility that philosophy exclusively serves a critical purpose, the unmasking of conceptual confusions. And sure enough, Wittgenstein often states the aim of philosophy in purely negative terms, namely “to show the fly the way out of the fly‐bottle” by making philosophical problems “completely disappear” (PI §§309, 133; see AWL 21; BT 425; CV 43). But why should one engage in philosophy at all, if it only gets rid of errors it itself has created? One answer is that philosophy is of value to “the philosopher in us” (TS 219, p.11): the temptation to conceptual confusion is not confined to professional philosophers. Yet that still leaves unanswered Ryle’s astute question of what a fly would miss that never got into the fly‐bottle (1971a, p.114). Here we must appreciate that philosophy should not dissolve our urge to ask philosophical questions by any old means, e.g., a knock on the head, but through an understanding of their nature and sources. A fly which never got into the bottle will not only lack the ability to extricate itself, a kind of know‐how, but also the conceptual clarity which Wittgenstein regarded as an end in itself (PR, preface; CV 7).
What is more, Wittgenstein attached “fundamental significance” to the concept of a perspicuous representation (übersichtliche Darstellung) since it affords “an understanding which consists in ‘seeing connections’” between different parts of grammar (PI §122; see GB 133; BT 417).
Whether one regards a successful overview merely as a tool of philosophical critique or as interesting in its own right (as Strawson does on behalf of descriptive metaphysics) is a matter of intellectual temperament. One way of putting this point is that philosophy contributes to human understanding though not to human knowledge (Glock, 1996, p.283; Hacker, 2009). Even this undersells the positive side of philosophy, however. After all, it makes perfectly good sense to know the meaning of expressions, or to “know one’s way around” (RPP I §§303, 1054) the conceptual landscape, e.g., by knowing that “A knows that p” entails “p.”
Even confining oneself to philosophy’s critical task, one may hold – with Kant – that it can be systematic and make progress. Rejecting such aspirations appears to be a hallmark of Wittgenstein’s later conception. The hostility to systematic investigations seems to be borne out by Wittgenstein’s notorious prohibition of hypotheses and theories. This appearance is partly deceptive, however. Wittgenstein had an overly restrictive conception of theory, confining it either to the hypothetico‐deductive theories of empirical science (PI §109; see Hanfling, 2004) or to the attempt to provide analytic definitions of what he regarded as family‐resemblance concepts (e.g. PG 119–20; RPP I §633). Neither proscription rules out dealing with philosophical questions in a sustained and orderly fashion.
It is two other claims that militate against systematic aspirations. One is Wittgenstein’s stress on the motley of “language‐games” and the disorderly and dynamic character of natural languages, as opposed to the artificial formal calculi (BB 16–17; PI §§18, 23, 108). The other is his acknowledgment that an overview of grammar establishes “an order” in our understanding of language which is purpose‐relative – namely to the resolution of specific problems – not “the order” (PI §132; TS 220, §107). There are different potentially helpful articulations of the same grammatical rules.
Nevertheless, the very notion of an overview suggests that there is a sense in which Wittgensteinian philosophy can be systematic. Indeed, Wittgenstein provided two different “classifications of psychological concepts” (RPP I §895; RPP II §§63, 148; Z §472). He also envisaged a “genealogical tree” (Stammbaum) for them, as for number concepts, presumably a way of showing how e.g., the system of natural numbers can be extended into that of signed integers (RPP I §722). These overviews do not aspire to “precision.” But Wittgenstein envisaged a “complete overview of everything which can create unclarity” (Z §§273, 464). This need not mean that there is a “totality” or “complete list of rules” for our language: the notion of “all rules” is dubious even for a single term, since clear criteria of identity exist only for codified rules, e.g., those of chess (MS 157a, p.108; TS 220, p.92). But it suggests that overviews of particular segments of grammar can be as comprehensive as one pleases.
Accordingly, there can be progress in mapping conceptual landscapes and resolving particular problems. But this is compatible with Wittgenstein’s claim that philosophy is open‐ended (Z §447; BB 44). Like the expansion of π, philosophy can get better, without ever getting nearer to completion. The reason is that even a global overview of grammar cannot provide a once‐and‐for‐all panacea for philosophical puzzles. First, the language in which they are rooted changes, thereby creating new problems, as happened with the development of the new physics, of formal logic, or of computers; secondly, there is no definite number of ways of getting confused. By a similar token, “there is not a philosophical method, though there are indeed methods, like different therapies” (PI §133). Philosophy cannot terminate if, as Kant and Wittgenstein suggest, the fascination with philosophical problems is part of the human condition (BT 422–4). Some passages intimate that this tendency might be eradicated by cultural change (RFM II §23; CV 68–9). Yet unlike postmodern prophets of the demise of philosophy such as Rorty, Wittgenstein provides no clues as to what such a change would amount to.
There have always been “irrationalist” interpretations, which distance both Wittgenstein’s philosophical practice and his metaphilosophical views from the ideal of rational argument (for a critical survey see Glock, 2007; cp. Chapter 2, WITTGENSTEIN’S TEXTS AND STYLE). One of them is Pyrrhonian. Wittgenstein does not just aim to overcome traditional, metaphysical philosophizing by a better “critical” variety, the story goes; he seeks to bring philosophy as such to an end. Just as the Pyrrhonian skeptic studiously eschews taking a stance even on the possibility of knowledge, Wittgenstein decries the very notion of having a philosophical view, while also avoiding having a metaphilosophical view about having philosophical views (see Rorty, 1979, p.371; Fogelin, 1987, ch.15; Stern, 2004).
There is no gainsaying that some passages lend succor to this interpretation.
If one tried to advance theses in philosophy, it would never be possible to debate them, because everyone would agree to them.
(PI §128; see §599)
Philosophical remarks are “homely,” “stale truisms” (TS 213, p.412; MS 109, p.212; TS 220, §§89–90; TS 219, p.6). Indeed, Wittgenstein professes not to rely on “opinions” anybody could dispute (AWL 97; LFM 22; RFM 160; LC 72). But as argued above, this is part of a dialectic procedure. Philosophy provides “reminders” (PI §127) of patterns of linguistic use that competent speakers are perfectly familiar with.
To be sure, Wittgenstein opines that philosophy is “flat” (PI §§126, 599). Unlike the deductive‐nomological sciences and formal disciplines like mathematics or logic it does not revolve around deductive reasoning (even though the latter can feature in its exposition). Deduction establishes the consequences of premises, but a dialectical critique of sense scrutinizes the meaning of those premises and the intelligibility of the questions.
Accordingly, Wittgenstein’s style of argument is elenctic rather than demonstrative, yet this is no bar to it being perfectly rational. Furthermore, the Pyrhonnian insistence that Wittgenstein refuses to advance claims of any kind is blatantly incompatible with the descriptions of the “actual use of language” and “the quiet weighing of linguistic facts” he explicitly propagates (e.g., PI §124; Z §447). Finally, if, textual evidence notwithstanding, Wittgenstein had indeed adopted a Pyrrhonian “no position”‐position, he would confront a fatal dilemma. Either his remarks conform to his “no position”‐methodology, in which case they cannot amount to a genuine contribution to philosophical or metaphilosophical debate. Or they do not, in which case his practice belies his stated methodological views. Furthermore, he would be propounding the (non‐obvious) thesis that there are no (non‐obvious) philosophical theses. In either case – incommensurability and inconsistency – his attacks on traditional philosophy would be self‐contradictory and his conception of philosophy incoherent.
It is true that the Investigations feature few explicit answers to Wittgenstein’s numerous self‐posed questions (Kenny, 2004, p.78). But many of the questions are rhetorical. Again, Wittgenstein did not take sides in traditional philosophical disputes like realism vs. idealism, monism vs. dualism, nominalism vs. Platonism, etc.; instead, he tried to undermine the dubious assumptions common to the participants. But once again this is a perfectly rational strategy pioneered by Kant’s transcendental dialectic and embraced by Ramsey (1990, pp.11–12). Wittgenstein also tried to dissolve questions that lead to such misguided alternatives. But the strategy of questioning vexing questions is once again prudent and was recommended by the supremely sober G.E. Moore (1903, preface). Moreover, in questioning a question Wittgenstein sought the “right question” (see PI §§133, 189, 321; RPP I §600; MS 130, p.107; WAM 27–8). And he did provide answers to Socratic questions like “What is understanding?,” since doing so is a prerequisite of dissolving misguided questions and theories. What he rejects with respect to such Socratic questions is merely the insistence that they can only be answered by analytic definitions (BB 17–20; PI §64).
Even where Wittgenstein rejects a traditional question as phrased, his remarks must nevertheless address an underlying problem. Otherwise he simply would not have anything to say on the topics at issue and his rejection would be no more than an expression of lack of interest, something those pursuing the question can ignore. Thus, when Wittgenstein dismisses questions like “What is the ground of necessary truth?” he still addresses the philosophical problem of necessity by other questions like “What is it for a proposition to be necessary?.” Questioning a question in a philosophically relevant sense must involve taking up an underlying common problem in a more adequate way.
Wittgenstein suggested that philosophical illumination may arise from a book featuring nothing but jokes and questions and that we should respond to all philosophical questions not by giving an answer, but by asking a new question (RFM III §5; WAM 27–8). In that very remark, however, he willy‐nilly provides an answer to the question of what role questions play in philosophy. This rejoinder sounds bloody‐minded only because it has to match the obstinacy of seriously adopting a “no position”‐position.
As these occasional remarks show, Wittgenstein was not entirely immune to Pyrrhonian urges. He was also fond of comparing his philosophical critique to a kind of psychotherapy that disabuses us of the inclination to raise philosophical questions (PI §§133, 254–5; BT 407–10; Z §382; see Chapter 43, WITTGENSTEIN ON PSYCHOANALYSIS). Moreover, in a lecture he confessed to “making propaganda for one style of thinking as opposed to another” (LC 28). Unfortunately, both procedures are philosophically immaterial. For their only criterion of success appears to be the suppression of a certain intellectual urge. They cannot distinguish between achieving this goal by extrinsic means, such as hypnosis, drugs, or a knock on the head, and achieving it in the only way that can be philosophically pertinent, namely through showing that something goes wrong in a philosophical question or view.
It is fortunate, therefore, that Wittgenstein himself acknowledged that philosophical critique ought to be a rational rather than medical or rhetorical enterprise. He insisted that philosophy should provide arguments that are “absolutely conclusive,” and he described his own thought as the “rejection of wrong arguments,” an avenue open to those feeling a need for “transparency of their own argumentation” (MS 161, p.3; BT 408, 421). Wittgenstein’s work contains or at least intimates plenty of powerful and profound arguments of an elenctic type. It is just that, because of his idiosyncratic style, these arguments need to be spelled out by painstaking exegesis. The rational line for interpreters is to acknowledge that Wittgenstein’s work combines rationalist and irrationalist elements. The rational line for philosophers is to explore the arguments, insights, and instructive errors it has to offer. This exhortation presupposes, of course, that philosophy is a rational enterprise based on arguments. But one cannot argue against this presupposition without engaging in precisely such an enterprise. And in the absence of such arguments there is no reason to abandon the ideal of rational philosophizing. Therefore it is a presupposition to which we should commit.
At the same time, let us be mindful of the limits of rational argument. Wittgenstein himself insisted on this point, e.g., in his discussion of the autonomy of grammar and of the foundations of certainty and knowledge. As regards the nature of philosophy, however, he failed to draw one important lesson. In his early work, at least, he succumbed to what one might call the myth of mere method. This is the illusion that one can fashion philosophical methods in a presuppositionless manner, one which does not in turn draw on philosophical views, e.g., about logical necessity, linguistic meaning, or the nature of philosophical problems. In the Tractatus the method, in particular an ideal notation for the analysis of propositions, is supposed to be put in place by propaedeutic claims about the essence of representation that are then disowned as nonsensical. In the Investigations it seems that the method and the metaphilosophical remarks describing it are supposed to emerge automatically as a spin‐off from reflections on specific philosophical problems. But the Tractatus procedure is self‐refuting; and the philosophical problems discussed in the Investigations only cry out for Wittgenstein’s treatment on a certain understanding of their nature, an understanding which itself is philosophically contentious. Consider just one looming circularity. Wittgenstein entreats us to address philosophical problems by looking at nonphilosophical linguistic use. That recommendation depends on two ideas: the content of these problems (their constitutive concepts) is determined by the linguistic meaning of the expressions involved; the meaning of those expressions is a function of their use. These ideas are in turn supported by observations concerning the nonphilosophical use of “meaning,” its cognates and related notions like that of what is said or asked. We have reason to condone this particular circle, since it is difficult to envisage a better starting point for clarifying or even modifying a notion like meaning than its established use. Nevertheless, in urging that point one cannot bootstrap oneself onto a privileged methodological plane.
The nature of philosophy is itself a contested philosophical issue, and views about this issue are philosophically controversial. The label “metaphilosophy” notwithstanding it is not a distinct higher‐order discipline, but an integral part of philosophy itself. By contrast to therapeutic followers like Lazerowitz (1964/2004), who theorized about philosophy from the external vantage point of psychoanalysis, Wittgenstein himself was aware of this point (PI §121). Once it is acknowledged that one cannot engage in metaphilosophy without doing philosophy, however, the myth of mere method collapses. One cannot swim without venturing into the water. And one cannot address philosophical problems, the nature of philosophy included, without doing philosophy, and hence without philosophical arguments and commitments of one’s own. What one can do is to ensure consistency between philosophical methods, metaphilosophical and substantive views, and to argue for all of them in as plausible and unassuming a way as possible. Such a modest procedure must treat the actual character of philosophical investigations as one starting point. In that respect, normative metaphilosophy refers back to descriptive methodology. And here Wittgenstein’s phenomenology of philosophical puzzlement can make a lasting contribution, even to those who disagree vehemently with his overall conception of our subject.