In Wittgenstein's Metaphilosophy, Paul Horwich attempts to offer an analytic articulation and defence of Wittgenstein's so-called metaphilosophy in order to exhibit and exemplify its relevance for mainstream problems in philosophy of language and mind. The author assumes the existence of a ‘polar split between the great majority of philosophers and Wittgensteinians’ (xii). These philosophers do not think that Wittgenstein's ‘ideas are relevant to their work’, while Wittgensteinians are described as ‘engaged in feuds with one another that no one else cares about’. Horwich's self-appointed task is the enterprise of putting an end to this so-called ghettoization of Wittgensteinians.

We agree with Horwich that Wittgenstein is far, far too important a philosopher to be ‘ghettoized’. And we agree with him about the fundamental importance, if one is to appreciate Wittgenstein's thought, of comparing his philosophical approach with so-called ‘metaphilosophical’ enterprises, for Wittgenstein's conception of and methods/practice of philosophy are undoubtedly a fundamental contribution of his.

Unfortunately though, we believe that to attribute to Wittgenstein a conception of ‘metaphilosophy’ – as Horwich's monograph does – is not only a move which lacks textual evidence but a move that expresses a deep misunderstanding of the philosophical landscape it is meant to capture. We shall seek to explain why.

Three moments can be distinguished in Horwich's essay:

The first one can be described as an attempt to provide an outline of Wittgenstein's so called metaphilosophy (chapters 1 and 2). It is based on the assumption that a statable, free-standing philosophical methodology can and is to be extracted from his works in order to make its relevance for mainstream contemporary philosophical debates obvious.

The second one can be described as an attempt to ground exegetically the relevance of this metaphilosophical account in Wittgenstein's two most famous works: chapter 3 is focused on the Tractatus (herefrom TLP) and chapter 4 on Philosophical Investigations (herefrom PI). It is based on the assumption that Wittgenstein's philosophical evolution can be satisfactorily captured solely under the terms of a metaphilosophical account.

The third one can be described as an application of this metaphilosophical account: chapter 5 exposes and criticizes Kripke's reading of the Investigations and chapter 6 proposes an application of Wittgenstein's ideas to contemporary debates about the nature of experience and consciousness.

Horwich starts by presenting Wittgenstein's remark that ‘Philosophy is a battle against the bewitchment of our intelligence by means of language’ (PI 109) as his fundamental idea whose ‘shocking implication is that philosophy cannot deliver the sort of knowledge that is usually regarded as its raison d'être’ (1). By contrast with traditional philosophy which promises ‘to provide fundamental insights into the human condition and the ultimate character of the universe’ (2), Wittgenstein's philosophy would enable one to see that there are ‘no realms of phenomena whose study is the special business of a philosopher’ about which one could formulate ‘profound a priori theories and sophisticated arguments in support of them’ (1). Traditional philosophical questions concerning the nature of numbers, time, knowledge, truth, justice, beauty, free will, etc. would derive their fascination from paradoxes understood as conceptual tensions stemming from perverse exaggerations of linguistic analogies (19). Therefore the ‘important revelations’ typically promised within philosophy would be impossible as, in the end, for Wittgenstein, ‘the aim of a reasonable philosophical methodology should be to dissolve such puzzles – to expose the irrational over-generalizations from which they emerge’ (19). This conception is presented as a ‘revolutionary metaphilosophical perspective’ (20).

One could wonder about the conception of revolution attributed to an author (Wittgenstein) who continuously appealed to us not to overestimate ‘progress’ and who correlatively worked to exhibit the nonsensicality of totalizing points-of-view. In other words: the danger is that Horwich, in the name of Wittgenstein's conception of philosophy, offers in fact just one more totalizing account, one more dogmatic position.

Horwich assumes that Wittgenstein was reacting against ‘loosely related forms of intellectual pursuit to which the label « philosophy » has…been applied’ (21). Therefrom the label ‘T-philosophy’ is proposed by Horwich to capture simultaneously the « traditional » and « theoretical » character of these forms of intellectual pursuit characterized by their aim of building and defending philosophical theories. Such productions would count as theories because of their claim for generality and depth, as they would be meant to organize, unify and explain common-sense commitments and eventually correct them; and they would be philosophical because of their a priori character, as they cannot be supported by perceptual observations, by contrast with scientific theories (21). Epistemic tensions arising at the initial stages of T-philosophical investigations are then presented to lay the groundwork for different sorts of theoretical resolution (29) among which are: (i) sceptical theories which imply that our assertoric practice with respect to the phenomena at issue ought to be abandoned (30), (ii) revisionist theories which relinquish certain normally held views about phenomena (30), (iii) mysterian theories which hold that certain phenomena are essentially bizarre, and (iv) conservative-systematic theories (such as ‘Gettierology’ and ‘Trolleyology’ (33)). These kinds of theories are then criticized for diverse reasons. Revisionist theories ‘fly in face of our data’ as the revisions they effect exhibit less their groundedness than the ungroundedness of the frame in which they appear (45). Sceptical and mysterian theories involve confusing assimilations between distinct linguo-conceptual practices which require the intelligibility of the described phenomena at least in order to establish their intrinsic unintelligibility (48). Conservative systematization projects preclude the interest of the results they may lead to because of the complexity of intuitive ‘data’ (49).

Yet if criticized for diverse reasons, these theoretical strategies are described as sharing a similar ‘general form of a philosophical issue’ (50). Horwich describes its supposed Wittgensteinian dissolution with a procedure in eight steps that is then applied to philosophical worries concerning the notions of numbers, time, truth and good (50–61). Each time the worries are presented as grounded in broadly-scientistic expectations concerning the reaching of a priori principles concerning phenomena, which one attempts to reach/satisfy on the basis of an undue linguistic analogy, resulting in a paradoxal tension due to the confused assimilation, and finally ‘therapeutically’ overcome through the bringing out of the initially employed yet forgotten analogy. Horwich declares that ‘Wittgenstein thinks that each distinctively philosophical problematic exemplifies something like the structure’ that he (Horwich) has articulated for dealing with the ‘general form of a philosophical issue’. We wonder why Horwich did not ponder Wittgenstein's self-conscious exhibition of nonsensicality at the close of TLP which thoroughly undermines his earlier apparent-commitment to a ‘general form’-style doctrine, and why Horwich further did not ponder Wittgenstein's dismissal, throughout his later work, of philosophical projects which seek to articulate the ‘general form’ of things (e.g. of a proposition; but not just of that. See also PI 131, which is deeply apposite at this point, we fear). It seems to us that Horwich is simply giving us a new form of T-philosophy while allegedly rejecting T-philosophy. This is (simply) a recipe for confusion.

Horwich expressly argues that theorizing is to be shunned: linguistic or conceptual activity is not to be explained; its epistemological foundations are not to be provided; idealized forms of a priori knowledge are not to be given, reformed, streamlined or completed (65). Theorizing in these forms is presented as prohibited by Wittgenstein while he is taken to be realizing metaphilosophical investigations which are themselves – for ungiven reasons – allowed (65). But of course Wittgenstein offers no prohibitions (see e.g. PI 305 & 308). Horwich suggests that Wittgenstein is banning something, but Wittgenstein is – rather – seeking to make something unattractive; this is the point of his great remarks about seeking to suggest to philosophers et al a novel way of seeing the ‘place’ that Cantor opened up in such a way that they will no longer take it to be a paradise, and no longer desire to go/stay there. Whereas Horwich's Wittgenstein would presumably bar the entrance to Cantor's paradise (which would just make people keener than ever to get in…).

Moreover, Wittgenstein explicitly considers and dismisses in PI §121 the idea that philosophy and metaphilosophy would be different from one another in the way that Horwich proposes. The very notion of metaphilosophy which is said by Horwich to be Wittgenstein's is, crucially, absent in 121. It is also hard to see how overcoming the distinction between philosophy and metaphilosophy (as Wittgenstein does) could possibly be consistent with attempting to ground the latter as a distinctive alleged locus of truths. All this then may lead one to wonder deeply about Horwich's core assumptions concerning Wittgenstein's philosophy.

The bottom line is this: it is misleading really even to talk about Wittgenstein having a ‘metaphilosophy’. This can already be seen in PI 121. Let us look at it in full:

Horwich writes as if he is telling us about Wittgenstein's alleged ‘second-order philosophy’. But if PI 121 indeed deals with the problem of the supposed need to obtain a ‘second-order’ philosophy, as if what philosophy ‘itself’ cannot offer could be only offered by a meta-philosophy, it is palpably not in order to assume such a need. On the contrary it is in order to make explicit and dismiss the presupposition according to which philosophy would be somehow necessarily lacking an account of itself if not practised from a second-order point of view. If there is any metaphilosophy in Wittgenstein then the least that can be expected from a presentation of it is not to presuppose its supposed exteriority to practices. Philosophy is also an activity among others. It may be practised in order to obtain a synoptic or perspicuous (re)presentation of aspects of the ordinary but it does not start from a position pre-designed for such a result (or, for that matter, for any other) result. Such a series of propositions as Horwich envisages about philosophy, propositions assumed to be true solely on the basis of a supposed second-order grasp of their truth, is ipso facto not Wittgenstein's ‘metaphilosophy’.

We have suggested in this review that it is absolutely right to think Wittgenstein's so-called ‘metaphilosophical’ remarks of real importance. But Horwich, who claims to explicitate for us what Wittgenstein's metaphilosophy ‘is’, cannot help one to see this aright.

Hutchinson and Read have argued previously that the tendency toward ‘doctrinal’ readings of Wittgenstein – readings that present Wittgenstein as the purveyor, against what he says in PI 108–133 and elsewhere, of doctrines – is most drastically caught short by the failure of such readings to take seriously Wittgenstein's use of modal terms. ‘Doctrinal’ readings generally tend not to worry too much about so-called ‘metaphilosophical’ considerations. They simply extract doctrines from Wittgenstein's work and get on with using such doctrines in their philosophical arguments. So in a way it is a good thing that Horwich has written this book; at least he has tried to offer a ‘full-length’ doctrinal reading (See page 17 for his explicit endorsement of the claim that Wittgenstein's work contains doctrines) which actually does address matters ‘metaphilosophical’, and which (superficially at least) engages with 108–133, etc. Unfortunately however, his efforts offer no illumination, and in fact make matters worse. For the waters have now been muddied: Horwich offers an approach that is in fact doctrinal, while being at times in form would-be therapeutic. Supposedly rejecting ‘T-philosophy’, Horwich nevertheless offers theory-in-weak-disguise: a theory about the nature of philosophy as it ought to be, a theory of ‘Meaning = Use’, and so on. Nowhere does he take on board the deep nuance – the true rejection of theoryism – implicit in Wittgenstein's extremely careful use of modal terms (this care has been written about with sophistication by the later Gordon Baker and especially by Katherine Morris – writers of whom Horwich seems entirely ignorant, just as he seems ignorant of the ‘New Wittgensteinian’ readings of later Wittgenstein).

This can perhaps be seen most clearly in Horwich's chapter on ‘Meaning’, part of the ‘second moment’ in his book as we called it above. Horwich, like other doctrinal readers, thinks that Wittgenstein in PI 43 is offering an all-purpose definition of meaning as use (116–7). And he extracts a dispositionalist theory (one of ‘actual use-dispositions’ – see pages 167–9) from Wittgenstein's ‘rule-following considerations’ (141).

This is no way to free Wittgenstein from a ‘ghetto’. It amounts to taking Wittgenstein from his natural habitat and reconstructing him as a robot in an artificial terrain constructed on the basis of theories that are imposed upon him from without.

Truly great philosophers are virtually always ahead of their readers. In order to have any chance of standing on their shoulders, one has first not to crush them under the weight of one's prejudices. It is pointless to seek to take Wittgenstein into the terrain of mainstream philosophy, if what is dragged there is a weak caricature which soon enough will turn into some sort of meta-strawman. And that is so, with regard to what Horwich has produced.

Let us seek rather to return again to the radicalism of Wittgenstein's practice, free from any theory or dogma. Let us, unlike Horwich, consider seriously the possibility that Wittgenstein wrote in the way he did, not because he was a terrible writer, but because the style and method and content of what he was doing are all tied into one.

Let us find a way (PI 129) of noticing at last what is before our eyes. In this task, we don't need yet another theory, an articulated ‘metaphilosophy’ which Wittgenstein's endlessly subtle remarks then get crudely shoehorned into.