Chapter 6

Wittgenstein, Idealism and philosophy

The Limits of Language in the Later Method

For this is what disputes between Idealist, Solipsists and Realists look like.  The one party attack the normal form of expression as if they were attacking a statement; the others defend it, as if they were stating facts recognized by every reasonable human being.[1]

 

According to the early Wittgenstein, philosophy comes about through man’s ‘impulse to run up against the limits of language’.  It stems from our natural amazement at life: amazement that the world exists and that we can think and speak.  We try to express this amazement in the form of a question; but to do so is to utter nonsense.  Amazement is not a question and therefore has no answer.  Rather, it is the manifestation of the limit of what makes sense.  On the other hand, these limits make manifest the unspeakable truth in idealism: the limits of language are the limits of the world.

A concern with the limits of language remained central to Wittgenstein’s thought throughout his work, surviving the various changes from the Tractatus to the Philosophical Investigations.  In both of his principle works he makes a conscious effort to sketch the bounds of what makes sense, and to do it from the inside.  For example, in the Tractatus he says of philosophy that ‘It must set limits to what cannot be thought by working outwards through what can be thought.’[2] In Philosophical Investigations he states:

The results of philosophy are the uncovering of one or another piece of plain nonsense and the bumps that the understanding has got by running its head up against the limits of language.  These bumps make us see the value of the discovery.[3]

There are, of course, important differences in methodology between the two approaches, not least the shift in emphasis from more general considerations to a concern with detail and the particular case.  Furthermore, the later work criticises the tractarian conception of meaning, and makes a more concerted effort to reject philosophical theories in general.  Does this mean that Wittgenstein left behind the transcendental idealism of the Tractatus?[4]  In particular, to what extent can the Philosophical Investigations be said to present a kind of idealism?  Clearly the later work is not solipsistic — indeed, it is motivated to a large extent by an attempt to exorcise the solipsism of the Tractatus — but does this involve a rejection (as Hacker seems to think[5]) of idealism in general?  It does seem that Wittgenstein rejected the transcendental idealism that supported the metaphysics of the Tractatus.  In the Philosophical Investigations he claims that the concepts involved in this doctrine lacked content:

Thought, language, now appear to us as the unique correlate, picture, of the world.  These concepts: proposition, language, thought, world, stand in line one behind the other, each equivalent to each.  (But what are these words to be used for now?  The language-game in which they are to be applied is missing.)[6]

 

But despite this apparent rejection of the Tractarian connection between language and the world, one influential interpretation does not see the later Wittgenstein as rejecting transcendental idealism completely.  Bernard Williams has suggested that the rejection of transcendental solipsism, with the move from ‘I’ to ‘we’ in Wittgenstein’s later philosophy, occurs within the transcendental level[7].

The comparison of Wittgenstein and Kant has developed into something of an industry.  Contributions to the debate have been made by Norman Malcolm[8], Derek Bolton[9], Jonathan Lear[10], Thomas Nagel[11], and Daniel Hutto[12], to name but a few.  Unfortunately, the issue is somewhat confused by the fact that there is no clear agreement on some fundamental questions that underlie the debate.  Most importantly, there seems no general agreement on the question of what constitutes an idealist doctrine, let alone a Kantian one.  Hutto defends Wittgenstein from being “tarred with the brush of idealism” by making a favourable comparison between Wittgenstein’s philosophy and the work of Donald Davidson.  Davidson, he claims, is “a self-styled realist”, and if the label of idealism will not stick to him, it will not stick to Wittgenstein.  This argument ignores the fact that idealism comes in many different forms, and applies at different levels of analysis.  Kant, as we have seen, used his transcendental idealism to attack the ‘dogmatic idealism’ of Berkeley and the ‘sceptical idealism’ of Descartes.  He argued that his transcendental idealism implied empirical realism.  Kant, too, was a ‘self-styled realist’.  Nor is Kant’s idealism restricted to his anti-realist views of time and space.  As the discussion of idealism in part one of this thesis makes clear, the term ‘idealism’ need not be restricted to traditional antirealist doctrines.  The underlying notion identified there was the idea that the world is limited in some way by our ability to conceive it.  This is the notion of idealism used by Nagel[13], who characterises both Davidson[14] and Wittgenstein as idealist.

A standard objection to the Kantian interpretation of Wittgenstein is that there is no room in his philosophy for the crucial Kantian distinction between the world of appearance and the world as it is in itself.  There is no sense in which the later Wittgenstein can be seen as accepting the idea that we can filter out the mind’s contribution to experience and consider it independently.  Transcendental inquiry was for Kant an a priori investigation of the application of concepts to objects that ultimately reveals necessary truths about the formal structure imposed by mind.  Wittgenstein, of course, had no interest in such structures.  Lear replies[15] by claiming that a rejection of the transcendental distinction does not mean a rejection of transcendental philosophy.  Instead, we might be able to take a transcendental stance towards ordinary activities like language use.  But the use of such tactics may stretch the comparison between Wittgenstein and Kant to the point of vacuity.  And the problem is compounded by the fact that there is no agreed interpretation of either of their views, or even which aspects of their views are the most essential.  The question of whether or not particular aspects are essential to their respective philosophies will only be biased by the attempt to assimilate Wittgenstein’s views to Kant’s.

There is another reason why we should not paste over the later Wittgenstein’s apparent lack of regard for the transcendental distinction.  Denying that distinction is precisely what brought Davidson to reject the dualism of conceptual scheme and content (given), and hence to embrace a kind of conceptual idealism.

There is something to be said for comparing Wittgenstein’s and Davidson’s views in this respect.  Wittgenstein’s views on the conditions of meaning seem to imply that nothing can make sense which purports to reach beyond the outer bounds of human experience and life.  Similarly Davidson’s views on truth imply that nothing can make sense that purports to reach beyond the outer bounds of our language and languages like ours.  For Wittgenstein, it is only within a custom or practice that there exists the possibility of agreement and disagreement on the application of a rule, and thus the possibility of getting the application of a rule right or wrong.  Since language use is a matter of rule following, we cannot use language to refer to that which we cannot make any judgements about.  Wittgenstein would seem to agree with Davidson that we cannot make sense of a radically different conceptual scheme to our own.

On the surface, though, there also seems to be a major point of difference between Wittgenstein and Davidson: in some respects Wittgenstein does seem to be at least encouraging conceptual relativism.  Wittgenstein, for instance, is fond of giving ‘anthropomorphic’ examples in which he confronts us with a different form of life from our own.  But the acceptance of different forms of life does not necessarily lead us to conceptual relativism, just as the acceptance of different languages does not lead us there.  What would commit us to such relativism is an acceptance of different forms of life that are incommensurable with ours.  Bernard Williams has argued that on Wittgenstein’s theory of meaning we would not be able to consider such forms of life[16].  Far from being a relativist, Wittgenstein presents us with different ways of applying concepts from our point of view.  We understand them precisely because they are not so alien to us as to be incommensurable with our way of acting.

Of course, comparing Wittgenstein’s view’s with Davidson’s can fall prey to the same dangers as the comparison with Kant’s views.  Ultimately, the comparison would not do justice to either philosopher.  It would not be difficult to find important differences in their philosophies.  (The fact that Davidson believes one can have a theory of meaning is one salient difference.)  The suggestion is, rather, that we replace the vexed question of ‘In what sense was Wittgenstein a transcendental idealist?’ with the question ‘In what sense was Wittgenstein a conceptual idealist?’  In any case, this question seems to be more in the spirit of Williams’ original article than the subsequent comparisons with Kant.  According to Williams, Wittgenstein was concerned to show the limits of sense by ‘moving around reflexively inside our view of things and sensing when one began to be near the edge by the increasing incomprehensibility of things regarded from whatever way-out point of view one had moved to’[17].  The idea is that Wittgenstein is an idealist in subscribing to the following principle: what the world is for us is shown by the fact that some things and not others make sense[18].

Further evidence for the claim that Wittgenstein is a conceptual idealist can be found in Wittgenstein’s work.  We have already seen (in chapter 4) that Wittgenstein rejects a certain strong form of realism (Platonism) and that this rejection can justifiably be regarded as a rejection of conceptual realism.  In chapter 5 we made sense of Wittgenstein’s rejection of the given along roughly conceptual idealist lines: it is not possible in language to refer to something that is not individuated by a public concept.  Our concepts are embedded in our practices, and cannot be thought of as being derived purely from an independent world.  In short, Wittgenstein seems to hold a central claim of conceptual idealism that the nature of the objects, properties, and facts to which our concepts correspond is not fixed independently of the nature of the concepts that correspond to them.

That this claim could be attributed to the writer of Philosophical Investigations is supported by section 371:

Essence is expressed by grammar.

And by 373:

Grammar tells what kind of object anything is.  (Theology as grammar).

As well as Zettel, section 55:

Like everything metaphysical the harmony between thought and reality is to be found in the grammar of the language.

All of these quotes suggest that Wittgenstein held the view that the nature of the world we inhabit is determined by our linguistic practices[19].  What is apt to strike one as most odd about attributing such a view to Wittgenstein is that it amounts to a substantial metaphysical thesis, and Wittgenstein infamously rejects the validity of all philosophical theses.  Indeed, all the claims that Wittgenstein held this or that kind of idealism (or that he held an opposing realism) share one assumption: that despite his insistence to the contrary, Wittgenstein’s philosophy can be assimilated to some form of metaphysical doctrine.  The common assumption, rarely stated explicitly, is that it is not possible to consistently avoid substantial metaphysical claims when doing philosophy.  This view is further encouraged by the fact that Wittgenstein’s remarks on philosophy have proved to be the most perplexing of his work, and are often dismissed as incoherent.

I do not think that Wittgenstein can be regarded as an idealist.  Although he shares some of the assumptions of the conceptual idealist, he does not draw the metaphysical consequences that would make his philosophy a form of idealism.  Not only does he think that drawing such consequences is nonsensical, but he has a conception of language that precisely rules out the (tractarian) idea that the limits of language are the limits of the world.  This conception of language also throws light on his otherwise obscure remarks on philosophy, and shows that these remarks are not at all incoherent.  In the following section I will examine Wittgenstein’s conception of language and its relation to his rejection of philosophical theories.  We will then be in a position to finally put to rest the idea that Wittgenstein was some kind of idealist by stating precisely his relation to conceptual idealism.

6.1 Wittgenstein’s Remarks on Philosophy

In the Tractatus Wittgenstein notoriously goes beyond the limits that his conclusion draws, and he explicitly acknowledges this, asking us to reject his premises as nonsense once we have recognised their significance.  In Philosophical Investigations in contrast, he attacked much of the metaphysical and semantic theory of the earlier work, and made an effort to bring the concepts that are employed – such as ‘language’ and ‘meaning’ – down to earth.  His remarks on philosophical method, however, seem to have the same kind of paradoxical nature as is manifest in the earlier work.  Just as the Tractatus offers a theory of what can and cannot be said (both in general and in philosophy), so too Philosophical Investigations seems to offer a theory of what can and cannot be said in philosophy.  He makes the apparently unsupported claim that one cannot offer theses in philosophy.  In what follows, I will discuss those remarks and attempt to make sense of them.  I will argue that the inconsistency is more apparent than real.  Wittgenstein’s views on philosophical method are derived from a certain conception of language and its relation to philosophy, and that this conception does not amount to a ‘theory’ in the sense that Wittgenstein rules out.

In the Philosophical Investigations, a philosophical problem is said to have the form: “I don’t know my way about.”[20]  Conceptual confusion arises when we become entangled in our own language.  It is the task of the philosopher to untie the knots and remove the puzzlement that arises “when language is like an engine idling”, when it is without genuine content and understanding.  Thus philosophy has the roughly negative role of sweeping away misunderstandings.  It leaves ordinary language and real understanding as it is, only no longer obscured by confusions.

There are several ways in which confusions can arise, but one error that Wittgenstein repeatedly draws attention to is the tendency to over generalise, and find similarity were there is diversity.  For example, when we see a law in the way a word is used, and we try to apply it consistently.  The results can be paradoxical, and we are left in philosophical bewilderment.  Or for example, the question ‘What is time?’ gives the impression that a definition is being asked for, but giving one does not provide understanding but creates misunderstanding.  In the search for a law that we can apply consistently, we may first say, for example, that “time is the movement of heavenly bodies”.  But applying this consistently soon leads to paradoxical results.  Realising the definition is unsatisfactory we discard it and look around for another.  This very process of searching for definitions that will be more satisfactory is what convinces us that there must be some such correct answer[21].

There is also the danger of false simplifications and assimilation, or of applying an analogy without regard for how far it will stretch.  There are misunderstandings concerning the use of words, for instance, caused by seeing “analogies between the forms of expression in different regions of our language”[22].  Thus mistakes are made by “philosophising mathematicians” who fail to see that there are different uses of the word ‘kind’ when they talk about kinds of numbers or kinds of proof – as if it had the same meaning as in the context “kinds of apples”[23].

Then there are analogies that create false pictures of the way language might work.  For example, seeing that we can refer to an object in conversation by pointing to it, and knowing that we can point to a thing by looking at it, we mistakenly imagine that we can refer to a sensation by a mental act of directing one’s attention to it.[24]

One of the reasons all these confusions get a grip on philosophers is because of a mistaken tendency to apply the methods of science, and a temptation to answer philosophical questions in the way science does.  This is what Wittgenstein claims to be behind much of the false simplification and assimilation (a criticism aimed at the Tractatus as much as other philosophy).  But his criticism of the scientific method goes far deeper than this, and his reaction is a complete rejection of most of what was previously called philosophy: ‘It can never be our job to reduce anything.’[25] Philosophy, as Wittgenstein prescribes it, is ‘purely descriptive’:

It was true to say that our considerations could not be scientific ones.  It was not of any possible interest to us to find out empirically ‘that, contrary to our preconceived ideas, it is possible to think such-and-such’–whatever that may mean.  (The conception of thought as a gaseous medium.)  And we may not advance any kind of theory.  There must not be anything hypothetical in our considerations.  We must do away with all explanation, and description alone must take its place.  And this description gets it light, that is to say its purpose, from the philosophical problems.  These are, of course, not empirical problems; they are solved, rather, by looking into the workings of our language, and in such a way as to make us recognize those workings: in despite of an urge to misunderstand them.  The problems are solved, not by giving new information, but by arranging what we have always known.  Philosophy is a battle against the bewitchment of our intelligence by means of language.[26]

There are two points I would like to note about this passage.  The first is that Wittgenstein considers the practice of offering theories in philosophy as a mistake, brought about by misapplying the method of science.  The concepts of ‘theory’ and ‘explanation’ are grouped together, along with the notion of advancing something ‘hypothetical’.  This suggests the following connection between the concepts: scientific method proposes theories by postulating the hidden nature of the world.  It is hidden in the sense that it is not observable, and requires the conjecture of hypotheses that must be tested against reality.  This method relates to Wittgenstein’s characterisation of philosophy in that this proposal and evaluation of theories is done on the basis of their explanatory power.  A good theory is not one that seems self-evidently true, for that would be a mere observation.  It is evaluated, rather, on the basis of what it explains, and how it explains it.  The explanation it affords is thus offered as the justification for accepting the theory as true.  This does seem to be the method employed by most philosophers.  A philosophical problem is outlined, and a theory is proposed of the underlying ‘philosophical’ facts.  It is then argued that this theory explains the existence of the facts that constituted the problem, or shows them to be illusory, and thereby offers a solution.  This solution is offered as a justification of the theory, and how well it deals with various philosophical problems (and how well it fits with other accepted theories) is the measure by which it is compared with competing theories[27].

The second point is that Wittgenstein sees philosophy as being primarily concerned with language.  His primary target in the early sections of Philosophical Investigations is the theory of language of the Tractatus, and the influence of Russell and Frege contained therein.  The central idea that is criticised is that philosophy should produce theories of the world that are offered on the basis of their ability to explain language.  He also criticises the Russellian idea, expressed in the Tractatus (3.325), that the first task of philosophy is to reform language using a symbolism that more correctly mirrors the logical form of the world:

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.[28]

Philosophy simply puts everything before us, and neither explains nor deduces anything.–Since everything lies open to view there is nothing to explain.  For what is hidden, for example, is of no interest to us.[29]

Wittgenstein’s use of the words ‘explains’ and ‘deduces’ requires some scrutiny.  They are perhaps a little strong or general for his purposes, for was it not Wittgenstein that explained why there could be no private language?  Did he not deduce that such a language was not possible?  And Wittgenstein elsewhere accepts a form of ‘explanation’ that is not the one that science employs, but is a way of demonstrating ‘connections’ between phenomena[30].  It should also be noted that Wittgenstein’s use of the word ‘theory’ is also specialised.  Theories in mathematics were considered by him to be part of the game of mathematics, and as such were not considered subject to his criticism.  Similarly, “the so-called ‘theory of chess’ is itself just another game: the demonstration that I can get there in eight moves consists in my actually getting there in the symbolism, hence in doing in signs what, on a chess-board, I do with chessmen”[31].  So we might restrict the meaning of the terms ‘explanation’ and ‘deduction’ to their uses associated with constructive and predictive theories that attempt to probe beyond phenomena.  Certain observations, which might be considered ‘philosophical theories’ by others, are not considered by Wittgenstein to be ‘theories’ as he uses the term.  The fact that the meaning of a word can (in a large class of cases) be defined as the use we make of it is not presented as a theory that requires some justification in terms of its explanatory power.  It is presented as an observation, and if it has a justification, it consists in giving examples, or in presenting the facts that make this particular observation easier to make.  Of course, Wittgenstein then goes on to defend his observation against philosophical theories that contradict it.  And he does that by unearthing the peculiarly philosophical intuitions that underlie those theories – intuitions that are not based on the observation of actual language use.  Once an observation is made, there is no reason that we cannot adduce other related facts.  We can even deduce that certain philosophical notions (‘theories’) are ruled out by the implications of an observation.  For example, given certain observations about the nature of rule following, one can deduce that a private language is impossible.  What Wittgenstein claims cannot be deduced is the hidden nature of the world that is supposedly required to support the surface phenomena of linguistic practice.  The philosopher gains no special insight into reality with his investigations.  He can merely make clearer and plainer that which is already in full view.

So we could summarise Wittgenstein’s notion of a theory by pointing to the cases where we postulate something hidden beyond the actual use we make (or could make) of words.  Consider, for example, the beetle in a box analogy with respect to sensation language.  Theories ‘cancel out’: they are what could be otherwise given the observable phenomena.  This may leave the term ‘theory’ a little vague – in the sense that it doesn’t give a precise definition that covers all cases – but this is in line with Wittgenstein’s own remarks on the nature of language and philosophy.  Trying to give a more precise definition runs the risk of contradicting Wittgenstein’s own methodological principles.  Wittgenstein can be said to be teaching the use of the term ‘theory’ through the examples of the philosophical notions he attacks.  And this is precisely the way he says such concepts can be taught.[32]

But does Wittgenstein’s method really proceed by means of mere description?  If so, how does it solve philosophical problems?  Reading through the Philosophical Investigations one can see the following general approach being repeated: Wittgenstein takes a question that results in philosophical confusion and considers various answers.  Throughout he examines how we actually use the terms involved in the question, and rejects the problematic answers that one is tempted to give.  For example, when considering the question, ‘What is meaning?’  he looks at various answers that suggest that ‘meaning’ refers to some peculiar process in the mind, or something external that the mind ‘grasps’.  In making tacit conceptions of meaning clear and explicit we can see clearly how they are mistaken.  The conception we are left with when these mistaken ideas are cleared away consists of a series of examples of the way we actually proceed in ‘language-games’.  Our understanding of these games is not augmented with a new theory of what underlies them.  It remains as it did before, but cleared of the obscuring misunderstandings we harboured before our investigation.  Thus the work of the philosopher is to bring us back to our ordinary and unproblematic understanding, and one does this by “assembling reminders for a particular purpose.”[33]

Philosophy then, is supposed to proceed without metaphysical theories, or the explanations these theories offer, and we are told to be content with ‘mere description’.  This conception of philosophy is not, of course, a mere description of the way philosophy is actually done.  It does not characterise what went by the name of philosophy before Wittgenstein, nor does it describe most of what we would recognise as philosophy today.  So is Wittgenstein not offering a theory of philosophy, and thereby contradicting his own demands?  In particular, is he not giving us a theory of the nature of philosophical problems?  The idea that all philosophical problems are nothing but linguistic muddles certainly appears to be a theory of philosophical problems.  The apparent contradiction (or at least much of it) can be removed if we carefully examine Wittgenstein’s reasons for holding it.  For what Wittgenstein has banished is theories of the way language works in terms of the purported hidden nature of reality that supports language.  But Wittgenstein makes several claims about language that imply that this kind of support cannot be found.  These claims, I think, can be considered to be observations rather than speculations, at least if one considers the area of thought that Wittgenstein was most interested in.  I will argue that, at most, Wittgenstein was guilty of over-generalising these ideas in a way that may have prevented him from seeing certain philosophic problems, for example in areas such as ethics and aesthetics.

In Philosophical Investigations Wittgenstein’s primary focus is the kind of theory of language that he himself was previously inclined to give (as discussed in chapter 4).  But throughout his later work he held views about language that imply his further views about the nature of philosophical problems.  Thus, these latter views are not merely presented as a thesis.  They are part of a continuing commentary on the nature of language, not a theory of the reality that supports language.  In the next section I will examine the later Wittgenstein’s conception of language, and try to judge if they make his view of philosophy inevitable.  In the process we should not only be able to settle the question of whether this philosophical method is consistent or not, but we should also be able to pick up some clues as to Wittgenstein’s relation to idealism.[34] 

6.2 The Autonomy of Grammar

“Grammar consists of conventions”[35], and these conventions are in a certain sense arbitrary.  This is at odds with the conceptions of language preferred by both the conceptual idealist and the conceptual realist.  The conceptual realist holds some form of the picture of language that Wittgenstein repeatedly attacks.  This picture has it that language has a definite purpose, which is to communicate thoughts about objects.  If we construe all of language on ‘the model of object and designation’, grammar is not arbitrary, but reflects the nature of those objects.  The grammar of a word is derived from the nature of the object that the word refers to.  Hence, for the conceptual realist, the concepts embedded in our grammar aspire to reflect the independent nature of reality. 

For the conceptual idealist, the connection between language and the world is even stronger.  On the conceptual idealist conception of meaning, certain general considerations about the nature of representation or judgement in part determine (necessarily reflect) the nature of the objects, properties and facts that our concepts correspond to.  Thus the conceptual idealist proposes that the analysis of those concepts can provide insight into the ‘true nature of reality’.  Language contains metaphysical truths.  To say that grammar is arbitrary is to deny that there is any such link between our concepts and substantial metaphysical truth.  This is true, according to Wittgenstein, since language does not have a definite purpose that involves the nature of reality in the relevant respect.  Instead, it merely reflects contingent facts about human nature.

Wittgenstein first rejects the conception of language as something transcendentally isomorphic with reality[36], and instead wishes to use quite ordinary phenomena as the raw data of philosophical thinking:

Language is not defined for us as an arrangement fulfilling a definite purpose.  Rather “language” is for me a name for a collection and I understand it as including German, English, and so on, and further various systems of signs which have more or less affinity with these languages.

Language is of interest to me as a phenomenon and not as a means to a particular end.[37]

Languages, of course, can be invented for a particular purpose, in which case that purpose will be reflected in the language.  But language in general does not come about like this.  This is one of the essential points to emerge from comparing language with the playing of a game.

To invent a language could mean to invent an instrument for a particular purpose on the basis of the laws of nature (or consistently with them); but it also has the other sense, analogous to that in which we speak of the invention of a game.

Here I am stating something about the grammar of the word “language” by connecting it with the grammar of the word “invent”.

Are the rules of chess arbitrary?  Imagine that it turned out that only chess entertained and satisfied people.  Then the rules aren’t arbitrary if the purpose of the game is to be achieved.

“The rules of a game are arbitrary” means: the concept ‘game’ is not defined by the effect the game is supposed to have on us.[38]

Grammar is arbitrary in the sense that the rules of chess are arbitrary.  Nothing justifies the rule that one may move the king only one square at a time.  That is just the way the game is played.  The way a game is played may reflect nothing whatsoever.  If this comparison with language holds true, it follows that a certain approach to philosophy, which proposes theories of reality on the basis of their potential to explain linguistic phenomena, is misguided.  We cannot look to metaphysics to explain language, since grammar affords no explanation:

Grammar is not accountable to any reality.  It is grammatical rules that determine meaning (constitute it) and so they themselves are not answerable to any meaning and to that extent are arbitrary.

There cannot be a question whether these or other rules are the correct ones for the use of “not” (that is, whether they accord with its meaning).  For without these rules the word has as yet no meaning; and if we change the rules, it now has another meaning (or none), and in that case we may just as well change the word too.[39]

Wittgenstein believed that even the grammar of the logical constants was arbitrary: rather than reflecting some necessary fact about the world, they reflect contingent facts about human nature.  In a footnote to p. 147 of Philosophical Investigations he refers to the tendency to ‘invent a myth of meaning’[40].  We feel as if the fact that three negatives yield a negative again is somehow contained in the single negative.  And in a sense it is: that is the convention for use the word ‘not’.  But we could agree to use the word ‘not’ differently.  For example, such that two negatives yield a negative (as in the intended – and understood – meaning of “I ain’t done nothin’.”)  What makes some rules seem more necessary than others is our form of life.  We need only imagine a different way of living and communicating (involving, perhaps, different laws of nature) in order to make the use of some other rules intelligible[41].

This contradicts the conceptual realist intuition that the concept ‘not’ is independent of how we decide to use it.  A Platonist theory of logic, for example, holds that ‘not’ refers to an abstract mind-independent object that embodies the nature of negation.  Wittgenstein’s point is not that we cannot consider rules in such an abstract form, but that we need not use one rule rather than another.  His point is that there can be no philosophical justification of using one rather than another.  The reason we find the choice of one so necessary is because it is bound up with our nature[42].  That is how we normally think.  Of course, there is a strong philosophical tendency to the conceptual realism that Wittgenstein is rejecting here.  This can make Wittgenstein’s view seem inconsistent.  To the conceptual realist it may seem very much like Wittgenstein is offering an anti-realist theory of concepts.  But Wittgenstein is only rejecting the philosophical conception of meaning that the conceptual realist is proposing[43].  This is in perfectly consistent with his view of philosophy.  “Philosophizing is: rejecting false arguments.”[44] As we saw in chapter 4, the sceptical requirements that the conceptual realists theory of meaning aims to meet are ill conceived.  In general, the justification that the conceptual realist wants can never be achieved.

To understand Wittgenstein’s reasons for holding this, one must consider the background against which he was writing.  In the philosophy of language prevalent at the time, the task set by Russell and Frege, as well as his younger self, was to look for a theory of reality that accounted for certain regularities in language.  It was assumed that these regularities must reflect metaphysical necessities.  But for the later Wittgenstein, this assumption does not stand up to close scrutiny.  This is what Wittgenstein means when he says that language is ‘arbitrary’:

“The only correlate in language to an intrinsic necessity is an arbitrary rule.  It is the only thing which one can milk out of this intrinsic necessity into a proposition.”

Why don’t I call cookery rules arbitrary, and why am I tempted to call the rules of grammar arbitrary?  Because I think of the concept “cookery” as defined by the end of cookery, and I don’t think of the concept “language” as defined by the end of language.  You cook badly if you are guided in your cooking by the rules other than the right ones; but if you follow other rules than those of chess you are playing another game; and if you follow grammatical rules other than such and such ones, that does not mean you say something wrong, no, you are speaking of something else.[45]

The conception of language in Wittgenstein’s mature thought derives from the simple observation that the rules of grammar could be quite other than they are.  They would not thereby by ‘wrong’; they would simply be different rules.  To assume that the rules of grammar reflect the nature of reality is to assume that there is only one possible grammar, or set of grammars, that corresponds in some way to reality.  But this confuses questions of content with questions of form.  A grammar is not true or false.  The judgements that it allows us to make are:

The rules of grammar are arbitrary in the same sense as the choice of a unit of measurement.  But that means no more than that the choice is independent of the length of the objects to be measured and that the choice of one unit is not ‘true’ and of another ‘false’ in the way that a statement of length is true or false.  Of course that is only a remark on the grammar of the word “unit of length”.[46]

But why say that the rules of grammar are not in general justifiable by reference to reality?  Why can we not specify a grammar that is justifiable by reality, which therefore reflects the nature of reality?  After all, Wittgenstein concedes that certain ‘rules of representation’ are not conventions, since they can be justified:

I do not call the rules of representation conventions if they can be justified by the fact that a representation made in accordance with them will agree with reality.  For instance the rule “paint the sky brighter than anything that receives its light from it” is not a convention.[47]

The answer is that the philosophical project of specifying the reality that justifies these rules of representation cannot be completed in general, for in the end all we can describe is reality as it is conceived by a particular grammar.  The reality that we call upon to justify our grammar must be described, and this description is made possible by a certain grammar.  But then it too stands in need of justification:

The rules of grammar cannot be justified by showing that their application makes a representation agree with reality.  For this justification would itself have to describe what is represented.  And if something can be said in the justification and is permitted by its grammar – why shouldn’t it also be permitted by the grammar that I am trying to justify?  Why shouldn’t both forms of expression have the same freedom?  And how could what the one says restrict what the other can say?[48]

Wittgenstein’s remarks on philosophy in Philosophical Investigations are not unsupported speculations.  They arise from a picture of language that he developed over many years.  This picture results, moreover, from a persistent attempt to see and represent language without the obscuring apparatus of a philosophical doctrine.  It crucially involves rejecting the dualism of the surface structure of language and an underlying deep structure that supposedly reflects the essence of the world.  Without this dualism we cannot solve philosophical problems in the method supposed, for example, by Russell.

Of course, this brief description of Wittgenstein’s views on language is not enough to prove the validity of his controversial views on philosophy.  But I think we have done enough to show that they are not evidently and straightforwardly self-contradictory.  We have seen that Wittgenstein rejection of ‘theory’ in philosophy is a rejection of applying the method of science to philosophical problems.  He rejects the idea that we can postulate a hidden nature of the world that underlies our ordinary use of words on the basis that doing so has ‘explanatory’ power.  This follows in part from his views on language and grammar, which crucially involve rejecting the distinction between the observable surface phenomena of language and an underlying hidden structure.  A ‘deep’ grammar is always offered in philosophy as a kind of justification of our use of words, but such a justification is never possible.  For what is the point in offering a justification with something that is just in need of a justification itself?  Grammar can never be justified (and does not need be) because it arbitrary.  Grammatical rules are only answerable to more rules, not to the nature of a pre-conceptualised (pre-grammatical) reality.

None of this is evidently self-contradictory, although it does involve some major assumptions about the nature of philosophy.  I will mention three that strike me as important.  Firstly, in the view that philosophy should not proceed on a model of science, there is the assumption that philosophy must never advance uncertainly with hypotheses.  It is not a matter of assessing the best theory, but of proceeding slowly and certainly.  This is not itself a hypothesis – for it is too fundamental to compare it with the alternative – but it is an assumption.  Secondly, Wittgenstein also inherited a certain conception of philosophy from Frege and Russell as being primarily concerned with language.  Again this involves some major assumptions, but again they are on the level where one must make assumptions in order to begin at all.  Finally, it was Wittgenstein’s view that all philosophical problems can be solved by ‘untangling the knots’ of language.  Where does this optimism come from?    On Wittgenstein’s view of language and its connection with the world, removing confusions seems to be the only way of solving philosophical problems.  But why does he assume that all philosophical problems are soluble?  It seems to me that it comes from the conviction, which Wittgenstein developed early in life, that all philosophical problems and really ‘pseudo-problems’ – albeit pseudo-problems that are often difficult to dissolve.  It comes from the principal, expressed explicitly in the Tractatus, and implicitly in the treatment of the ‘sceptical paradox’, that a question without an answer is not a real question at all – that we cannot meaningfully express a question that has no possible answer.  This is also a substantial and fundamental assumption in Wittgenstein’s method.

I would also reserve judgement on whether Wittgenstein completed the task of revealing the nature of language to be always as he describes it to be.  Wittgenstein simply refused to consider whole areas of discourse such as ethics and aesthetics.  So we might put the matter this way: given certain general assumptions about the nature of language, Wittgenstein’s views on philosophy follow.  We might doubt that Wittgenstein’s remarks on philosophy have the general validity that he presented them as having.  But even if that were the case, that would not by itself commit Wittgenstein to some metaphysical views, such as the isomorphism between the world and language implied by conceptual idealism, that he himself explicitly denied.

The single element of idealism that survived from the Tractatus is this: that, as philosophers, we are confined within a language that cannot be anything but ordinary.  There is no way to specify a language that more deeply or accurately reflects the nature of reality, for such a language would just as much be without justification.  No justification (such as the given) can be specified without bringing it within ordinary language use, robbing it of its independent status, and thereby its justificatory role.

6.3 Wittgenstein and Idealism

I said the sole remnant of the earlier idealism is a certain concern with the limits of language.  It is inherited from the idea that we found in Wittgenstein’s ‘Transcendental Solipsism’ of §5.6 of the Tractatus: The limits of my language mean the limits of my world.  So does the later Wittgenstein nevertheless hold a kind of idealism?  One that is based on the claim that the limits of our world are the limits of our language?  This would just be the so-called move from the ‘I’ to the ‘We’ that Williams claimed occurred within the transcendental level.

The mistake here is to forget that Wittgenstein has rejected the transcendental scaffolding that made the earlier statement of the limits of language into a form of transcendental idealism.  As we have already discussed, Wittgenstein rejected the transcendental isomorphism between language and world. The later Wittgenstein rejects the distinction between the surface structure of language and an underlying structure that more accurately reflects the Way Things Really Are.  So while it is true to say that we cannot speak beyond the limits of language, that does not tell us what things there could or could not be.  The limits of language are merely the limits of what can be said (not what there is).  Finding the limits of language (by running up against them – finding inconsistencies and confusions) has only a therapeutic value.  The limit of language in the later work is merely the point where our attempts to philosophise bring us to utter nonsense, where we are tempted to point insignificantly towards the world in an attempt to justify language.

The best way to make this point is to state the difference between Wittgenstein’s philosophical outlook and that of conceptual idealism.  Conceptual idealism is an attempt to answer the question ‘How is metaphysics possible?’  In chapter 3, we discussed ways in which the conceptual idealist might argue for metaphysical truths.  These methods are ruled out on Wittgenstein’s view of language.

For Morris, a metaphysical result is the establishment of a metaphysical reduction.  For this to take place, there must be some (but not too great a) distance between our concepts and the objects they pick out.  This allows for the possibility that two concepts may nevertheless be of the same thing, thus allowing that one might be reduced to the other.  But for Wittgenstein, grammatical rules, and the concepts contained within them, are inseparable from the kind of things they apply to.  A different concept would just be a concept of a different thing.  Recall the point that Wittgenstein makes in Philosophical Grammar about there being no “question whether these or other rules are the correct ones for the use of ‘not’”[49].  More generally, there is no question as to whether these or other rules are correct for a particular meaning, since the way we use a word determines what it means, not the other way round.

Now, to this rather general claim, the metaphysician will no doubt object, and with good reason.  It is clear that we sometimes have different concepts for the same thing, and we may discover this fact.  The classic case of Hesperus and Phosphorus is undeniable[50].  Or, to take another well-worn example, the word ‘water’ as applied by a competent speaker of English before the advent for modern chemistry, has a meaning given by its use.  And that use may well have been best described in terms of the distinguishable macro properties of water.  The word ‘water’ was used to name that clear liquid that quenched thirst, filed our lakes and oceans, and so on.  When it was discovered by science that such fluids shared some underlying nature or structure that was specifiable in a new scientific vocabulary, the grammatical rules for correctly applying the word ‘water’ were reformed by the scientific community.  A new grammatical rule was added in the form of the proposition “water = H₂O”.  According to Wittgenstein’s maxim, then, that the meaning of a word is given by its use, it would seem that the meaning has changed.  At first blush, Wittgenstein’s views on meaning seem to bring him into line with the internalist, and in direct contradiction with the intuitions of the externalist.  The former holds that the only facts that are relevant for determining the intensional meaning of a speaker’s utterance are facts about the speaker.  The externalist claims that the “meaning” has remained constant through the discovery that “water = H₂O”, and draws the moral that meaning is determined in part by facts external to the speaker.  On his account, the discovery that water is H₂O is not only an empirical discovery, but also the discovery of a necessary proposition a posteriori.  But note that if the externalist wants to maintain that the meaning has remained constant, but that the use has changed, he owes us an account of what this constant “meaning” is, given that is neither the referent nor the use of the term.

Yet we have to concede to the externalist that there does seem to be some constancy in intentional meaning, and the discussion of Wittgenstein’s views on meaning in the previous two chapters do show some affinities with externalism[51].  To this we can add that Wittgenstein’s notion of grammar actually makes sense of the idea that the use changes, while the meaning somehow remains constant.  For in such a case we do not have a radical change of grammar, but an anticipated refinement.  A newly discovered empirical proposition (that water molecules consist of two hydrogen atoms bound to an oxygen atom) is hardened into a grammatical proposition.  A symptom becomes a criterion.  That hardening may take place almost immediately in the case of scientific discoveries, of course, where the scientific theory (modern chemistry) already prescribes that natural kinds have “essences”.  That is to say, it is already written into the grammar of natural kind terms that some such refinement of the grammar will take place[52].  Given this process of predetermined refinement, it certainly does make some sense to say that the concept involved has remained the same.  The grammar of the word “water”, both before and after the refinement, can be described by pointing at some water and saying, “that stuff”.  This, as Putnam originally pointed out[53], is how we are inclined to describe and teach our use of natural kind terms.  And there is a sense in which someone brought up in an environment identical to ours in every respect, except that the stuff referred to with the term “water” has a different chemical composition, has a different “use” for that term.  This is to take the externalists intuitions into account - both that we are using the same concept before and after the discovery, and that the twin earthlings have a different use.  It is also true that we defer “correct use” to future scientific discoveries, in particular when there is a well-entrenched conceptual framework already in place (e.g. chemistry) - just as we defer correct use to the community or a group of experts.  One could talk about “wide vs. narrow use”, where the former, but not the latter, includes facts about the environment and the linguistic community.  Furthermore, describing the use of the term is such a general way, we can make sense of the idea that what we discover when we find out that “water is H₂O” is in part a discovery about the way we have already been using the term “water”.  The idea that we can “discover” a grammatical proposition in this sense does explain externalist’s intuitions, and the feeling that we have discovered something necessary.[54]

This raises the issue of how “necessary a posteriori” propositions fit with Wittgenstein’s views, particularly his views on the nature of necessity.  According to Wittgenstein, empirical propositions can be said to describe possible states of affairs, but necessary propositions cannot be said to describe necessary states of affairs.  Their role is normative rather than descriptive: they function as, or are linked to, “grammatical propositions”[55].  But what of so-called necessary a posteriori propositions, which are purported to be both necessary, and about the world?  As we have seen, grammar is a matter of convention.  How can a “proposition” such as “water = H₂0” be both a matter of convention and have empirical content?

To understand this, we need to consider again Wittgenstein’s view of grammar and the role convention in language.  It is true that the claim that grammar is based on convention is central to Wittgenstein’s philosophy, but it must be recalled that this applies to grammar as a whole.  In fact, grammar incorporates conceptual frameworks (such as modern chemistry) that make demands on how we decide further grammatical rules.[56]  If these demands are followed, then the arbitrariness of a grammatical rule becomes partly removed from its direct application[57].  An example of the following kind will illustrate the point well.  Let us suppose that we agree to use the term “pepple” to designate that thing I currently have in my pocket, and suppose we take “pepple” to, in Kripke’s terminology, “rigidly designate” that item[58].  So, the pepple might exist in counterfactual situations, and indeed it is possible that the pepple was never in my pocket at all.  Nevertheless, it remains true that the pepple is, in fact, in my pocket, and we use that fact to name it, despite the fact that we do not yet know what the pepple is.  (We could imagine this naming as part of a game, where the aim is to guess the nature of the pepple).  While the pepple remains in my pocket, we have a perfectly adequate handle on how to use our new word.  And that is because we already have a grammar in which we can talk about things in my pocket.  And despite the quite arbitrary way in which this item was named, we are now in a position to empirically investigate the nature of peppleness.  Should it turn out that my investigation reveals that the pepple is, in fact, a coin, we would proudly declare that, not only is the pepple this coin, but, since identity is a necessary property, we will have discovered that the pepple necessarily is this coin.

So is this not an example of a necessary a posteriori proposition?  For Wittgenstein, necessary propositions are grammatical, so on his account it is necessary in so far as it has a normative role in how the terms involved are used.  Nevertheless, when we named the pepple, the full use of the word, while already prescribed by our naming ceremony, remained in some ways unknown.  Although it has already been arbitrated, it was not known, even by the namer, that this object was a pepple.  Thus we can learn, a posteriori, what a pepple is, even though this is a matter of a grammar that is ultimately arbitrary and based on convention.  A grammatical a posteriori proposition, we might say, is a proposition that is part convention, part empirical.  (If we again imagine the game of guessing the nature of peppleness, we can imagine that the judge in the game gets to peek at the object.  In doing so he discovers both an empirical fact and a grammatical rule that can both be expressed with “the pepple is this coin”).

In a similar way, we can see that “water = H₂O” is also partly grammatical.  While it was discovered empirically that water has such a structure, the idea that some such structure was there to be found belongs to a certain conceptual framework that constitutes a grammar and embodies rules that are ultimately arbitrary.  This allows us to perform reductions that reveal something of the grammar in which they are applicable, and thus provide a means of conceptual analysis.  While this grants a great deal to the externalist, and even hints in the direction of essentialism[59], given Wittgenstein’s full account of the nature of language, externalism cannot offer us truly metaphysical reductions.  Against this idea, Wittgenstein offers two decisive considerations.

First we have the by now familiar point that offering externalism as a theory that explains language use on the basis of the underlying structure of the world has the overwhelming problems emphasised in Philosophical Investigations.  Any such theory assumes that reality has a structure that can be discerned prior to our linguist practices.  But discerning structure is a matter of language use, and thus presupposes a grammar.  This grammar, if it is to be considered more fundamental, must be justified as such.  These justifications will inevitably come to an end and we will just say that these are the sorts of discriminations that we tend to make – this is what we do.  We cannot have a theory of language that makes reference to the world considered prior to that which we are trying to give a theory of: the employment of concepts.

There is an obvious counter-objection to this line of thinking that may be called ‘scientism’.  The language of science, it will be argued, is not at all arbitrary, since it is designed for a very particular purpose: to discover the true nature of the universe.  The grammar it embodies, it will be further argued, is justified by its utility and the predictive power of its theories.  This point ignores that even in scientific language games we bring to those games natural and cultural dispositions that could be otherwise.  This brings us to the second point, which marks the most salient difference with conceptual idealism.  While it is true that we defer correct usage of natural kind terms to the scientific community, we need not use those particular concepts.  Wittgenstein, far from holding the view that certain concepts are essential to understanding the universe, argues that no concepts are absolutely essential:

[I]f anyone believes that certain concepts are absolutely the correct ones, and that having different ones would mean not realizing something that we realize—then let him imagine certain very general facts of nature to be different from what we are used to, and the formation of concepts different from the usual ones will become intelligible to him.[60]

This is also the crucial point to understanding the difference between Wittgenstein’s views and Davidson’s.  The Davidsonian method is to look for the ‘large features of reality’ in the ‘large features of language’.  But we have already seen that the later Wittgenstein argued explicitly that the large features of language that he discerned in his earlier work could not be used meaningfully as super-concepts.  If concepts such as ‘proposition’, ‘language’ and ‘thought’ have meanings, they must be thoroughly ordinary.[61]  In any case, these concepts alone do little more than set the problem of metaphysics.  According to Davidson, even once we have established the possibility of a theory of truth (with which Wittgenstein would not concur), there remains the task of settling the ontological categories.  Davidson, for example, argues that a large number of sentences we hold to be true could not be if there were not events, but concedes that a theory of truth would not specify that events exist.  Since Wittgenstein holds that our conception of the world is not absolutely necessary, we[62] might well countenance a conceptual scheme in which there was no category corresponding to events.  In contrast to conceptual realism however, the question of whether one scheme or another is more correct does not arise for Wittgenstein.  They are simply different.  They are embedded in different forms of life.  So, like the conceptual realist, Wittgenstein does seem to hold that it makes sense to consider (in some general way) that a conceptual scheme may be so different that it would be incommensurable with our own.  This will be true if the interests and nature of some creature made its form of life radically different from ours.  ‘If a lion could talk,’ says Wittgenstein ‘we could not understand him’[63].

6.4 Wittgenstein and Kant

I earlier put aside the question of whether the later Wittgenstein was a ‘Transcendental’ idealist, on the grounds that the question is too ambiguous.  The philosophies of Kant and Wittgenstein are too complex, and too subtle, to lend themselves to any systematic comparison.  Those who have argued for the merit in comparing their views have more often provided little in the way of evidence from the corresponding texts[64].  In order to advance the debate, Hans-Johann Glock has distinguished four different ways in which the philosophies of these two thinkers might be compared: questions of actual influence; parallels at the methodological level; substantive similarities in philosophical logic; and substantive similarities in the philosophy of mind[65].  For the most part, this thesis has concerned itself with the first two of these, particularly in chapter 2, where we considered Kant’s influence on Wittgenstein’s early thought.  There we found both points of influence and similarities at a very general level of methodology[66].  But when we turn to the later Wittgenstein, this kind of comparison becomes too tenuous to warrant detailed claims, for the connections become too far removed.  For the most part, that which links Wittgenstein with Kant are just those aspects of Wittgenstein’s early work that the later Wittgenstein went to so much trouble to refute.  Thus we have just as much reason to consider Wittgenstein’s mature thought as a rejection of the Kantian justification of (a restricted) metaphysics.

Perhaps the best level on which to compare these very different philosophical approaches is on the very general level sketched in chapter 3: as approaches to metaphysics.  Both avoid the dichotomy of conceptual idealism and conceptual realism by restricting metaphysical claims, though Wittgenstein does so more fervently.  Both acknowledge the limits of our conceptual scheme, and accept (at least implicitly) the possibility of other, incommensurable conceptual schemes[67].  Against the conceptual realist, both Wittgenstein and Kant are keen to prescribe strict boundaries on talk about that which is beyond our conceptual scheme.

One salient difference is that, even taken in its most negative sense, Kant’s concept of ‘noumena’ has no real parallel in Wittgenstein’s later thought.  And while both accept the possibility of alternative conceptual schemes, there is no sense, however limited, of a ‘thing-in-itself’ in Wittgenstein’s mature thought.  We get a grip on considering different points of view by considering different interests and forms of life – not by considering, in however an abstract form, a common reality that differing schemes are applied to.

The most profound link between their philosophical approaches is their shared commitment to restricting metaphysics.  But as we saw in chapter 1, Kant’s task was ultimately to save metaphysics by placing it on firm foundations.  This aspect of Kant’s thought is precisely that which developed into modern day conceptual idealism.  Wittgenstein took the restrictive aspect of Kant’s project to greater extremes – even in the Tractatus – and ended up with philosophical quietism.

The quietism discussed in this chapter is the main point of continuity with the ‘transcendental idealism’ of the Tractatus.  The critique of meaning and philosophy in Philosophical Investigations can be related to the quietism of the Tractatus in the following way: the later critique shows how we cannot express the transcendental conditions of meaning and the world.  While the younger Wittgenstein thought that they made themselves manifest, but cannot be said, the later Wittgenstein shows how the quest to discover them is misguided in the first place.  What is beyond language is merely nonsense.  This shows that the differences with Kant are more significant than the similarities.  Kant wanted to save metaphysics; Wittgenstein set out to destroy it[68].

Part of the quietism of the Tractatus was a rejection of the notion of synthetic a priori truths.  In section 6, Wittgenstein argues that the only a priori truths are analytic.  The propositions of mathematics and the fundamental principles of mechanics are analytic in the sense that they say nothing about the world, but show something about the nature of representation. “The logical propositions describe the scaffolding of the world, or rather they present it. They ‘treat’ of nothing.”[69]

Wittgenstein later reconsidered this anti-Kantian aspect of the Tractatus[70].  His earlier view was that the elementary propositions are independent of each other.  But this presents a problem, since the proposition ‘this object is red all over’ excludes the proposition ‘this object is green’, and Wittgenstein failed to find an analysis of such statements in terms of common elementary propositions.  He therefore gave up on the thesis that the elementary propositions are independent, and recognised that statements like ‘Nothing can be red and green all over’ are logically necessary without being strict tautologies.  He contemplated calling such propositions ‘synthetic a priori’[71].  As Glock points out, this amounts to an admission that some propositions seem to anticipate reality without being based on experience of that reality.  Neither the Platonist view that these statements are about some reality beyond empirical reality, nor the positivist view that they say nothing, were plausible alternatives for the later Wittgenstein[72]: both the Platonism and the positivism of the Tractatus were rejected.

But neither could Wittgenstein reasonably turn to Kant on this issue.  Kant’s theory of the synthetic a priori assumes too much of the theoretical architecture of Kant’s transcendental enquiry.  It is worth saying something, therefore, about Wittgenstein’s reassessment of necessary propositions.  For given the anti-theoretical stance described above, it is not clear what account of the so-called synthetic a priori Wittgenstein is entitled to give.

The first thing to note is that the later Wittgenstein did reject the notion of synthetic a priori truths.  According to Glock, for Wittgenstein necessary propositions are a priori[73] because they are not about anything, and therefore not synthetic in Kant’s sense[74].  As we have already seen, for Wittgenstein necessary propositions are normative rather than descriptive.  They function as ‘grammatical propositions’, which are used to express grammatical rules or ‘norms of representation’.[75]

The account of necessary propositions that Glock finds in Wittgenstein then goes as follows.  The apparent ‘hardness’ of necessary propositions is explained by the fact that a grammatical proposition “antecedes experience in an innocuous sense.”[76]  The statement “Black is darker than white” is not corrigible to experience since the putative statement “This white object is darker than that black object” is a nonsensical combination of signs.  Nothing can count as both deserving the attribution of being ‘white’ and being ‘darker than black’.  Necessary connections are connections in grammar, and thus, in a broad sense[77], conceptual.

Glock goes on to say of Wittgenstein’s later work that:

The bounds of sense are drawn not by synthetic principles a priori, but by linguistic rules which exclude certain moves within a language game… Accordingly, while not all logical necessity is analytic, it is all conceptual, and hence linguistic, that is to say, it is determined by what we treat as a meaningful employment of words.[78]

This linguistic account is just what we would expect from Wittgenstein given the characterisation of his philosophy provided in this chapter.  For Wittgenstein, we cannot give philosophical theories in terms of the hidden structure of the world.  As philosophers, we are restricted to making observations about language.

Glock goes on to argue, however, that this view of necessary propositions will not adequately explain the a priori status of some simple ‘metaphysical’ propositions, such as ‘Every event must have a cause’, and that this amounts to a ‘lacuna’ in Wittgenstein’s account.  He points out that, “on a straightforward interpretation”[79], Wittgenstein’s treatment of the law of causality rests on the claim that the phrase ‘uncaused event’ is ruled out as nonsensical on grammatical grounds.  But that is shown to be false if we consider the following case: image that one morning dinosaur footprints appeared on the ceiling.  It is true that we would not, in the absence of any plausible candidate for a cause of this event, simply shrug our shoulders and say “Just one of those things!”  But Glock maintains that eventually we would be prepared to give up in our quest for a suitable cause – “for instance because the laws of nature not only fail to provide one, but suggest that none is to be had”.  Thus the idea of an ‘uncaused event’ cannot be ruled out on conceptual grounds.

But Glock has evidently brushed over a distinction between the kind of ‘conceptual grounds’ that feature in Wittgenstein’s thought, and a simple account based on analyticity: that the concept of a cause is implied by the concept of an event.  In any case, our suspicions are aroused by Glock’s apparent willingness to countenance the idea of an uncaused event.  He gives no indication of how the laws of nature might lead us to believe that the dinosaur prints were uncaused, and has smoothed over the fact that accepting such an idea would be met with almost insurmountable conceptual resistance.  He has only made clear two unsatisfactory alternatives: either we find the notion of an uncaused event straightforwardly contradictory, or we find it perfectly acceptable.  He rejects the former, but if the latter is accepted as true, we seem to be thereby giving up on the a priori status of the law of causality.  For in the situation where we accept the existence of an uncaused event, we have given up on the truth of the statement ‘every event must have a cause’, let alone it’s a priori status.

What has gone wrong here?  Consider the difficulty we would have accepting the possibility of the dinosaur prints being uncaused.  Glock is right to claim that it is not that the phrase ‘uncaused event’ is straightforwardly nonsensical, but it may be appropriate to object that, “that is not how we normally think about events!”  I think that Wittgenstein might express this confusion by saying, “I don’t know what game we’re playing now!”

We can explain both the oddness of the phrase ‘uncaused event’ and its apparent coherence if we consider this phrase as being a new use of the word ‘event’.  The phrase has been taken out of the context of its usual language games, and although this is possible – we can invent new games – it makes us feel uneasy because the rules of this new game have not yet been made clear or decided upon.  Ordinarily, the concept of an event is used in the context of classical (or folk) mechanics.  It belongs to a family of games that have as their purpose the formulation of explanations.  Without further understanding of how a language game that gave up on the law of causality could provide us with explanations, the new use of the verbal sign ‘event’ remains detached from its previous meaning.  The term ‘event’ has its place in that family of language games that are connected with giving explanations of the world in a particular form.  According to Wittgenstein, the sentence ‘every event must have a cause’ expresses one of the common grammatical rules of these games.  Of course we can choose to give up on that rule – the rules of grammar are in a certain sense arbitrary. But they are not arbitrary when one considers them with respect to the nature of the games we play.  To use other rules is to play different games.

Glock claims that what is missing from Wittgenstein’s account can be found in Kant.  What links the idea of an event with the idea of a cause must be a third concept.  In a Kantian spirit, Glock suggests that this third concept is the notion of experience: ‘Events must be caused not because random and chaotic changes do not qualify as events, but because persistently chaotic events are not plausible objects of self-conscious experience’[80].  But that will not fill any lacuna in Wittgenstein’s later thought, because it does not fit with his thought at all.  The subject matter of Wittgenstein’s later philosophy is grammar, not ‘self-conscious experience’.  Wittgenstein would not have couched his thoughts using such an ill-defined term, which smacks of a super-concept:

We are under the illusion that what is peculiar, profound, essential, in our investigation, resides in its trying to grasp the incomparable essence of language.  That is, the order existing between the concepts of proposition, word, proof, truth, experience, and so on.  This order is a super-order between–so to speak–super-concepts.[81]

The third concept that links ‘event’ and ‘cause’ for Wittgenstein is not the concept of experience, but the concept of an explanation, or more generally the language games we play when we formulate and consider explanations.

This leaves us one remaining question about the consistency of Wittgenstein’s pronouncements on philosophy.  For is not the concept of a language game a super-concept?  One thing that can be said in favour of the coherence of Wittgenstein’s method is that the concept of a language game at least remains, in a broad sense, a linguistic concept.  Its employment as a device to solve philosophical problems is in keeping with the heart of Wittgenstein’s method: to assemble reminders from observations of language, and to do so in a way that does justice to the richness of linguistic activity.  Wittgenstein is not an idealist because he is not offering a theory of reality.  He is engaged in the task of forming a coherent and perspicuous representation of language.  Philosophy tells us nothing, but simply reminds us of the kinds of creatures we are and the way we communicate.