Graham McFee

Main page  
Baker and Hacker without Baker? Wittgenstein: Text and Context
"On first looking into ...": The placing of the 'slips of paper' ... Wittgenstein, Systematicity and the Use of Philosophy
A Nasty Accident with one' Flies [Inaugural Address] More on the 'Third Wittgenstein' Idea
Everything Goes with Beer

Wittgenstein, Systematicity, and the Use of Philosophy

This work-in-progress began life as a presentation to a conference at Gregynog, in 1996. It responds to some of Dummett's criticisms of Wittgenstein's account of the project of philosophy -- in particular, to the objection that Wittgenstein's account does not offer philosophy a future (because it does not indicate what philosophers should do next); and fails to do so because it lacks systematicity (in the relevant sense). In reply, I lay out some of my account of Wittgenstein's project. While I would probably be less dogmatic today in that respect, it still seems pertinent to highlight some key features of a broadly Wittgensteinian defence of a therapeutic conception of philosophy, against the claims of those who would try to read Wittgenstein as, say, articulating a theory of meaning. (Perhaps it does not give enough weight to the insight, from Gordon Baker, that Wittgenstein's philosophy is therapeutic through-and-through: in this sense, it can be augmented by some ideas from "Baker & Hacker without Baker?".) A revised version of this paper appears in Wittgenstein on mind, meaning and context: Seven essays. (2019 Pageantry Press: pp. 305 (ISBN 978-0-578-43660-9)


This paper defends Wittgenstein's conception of philosophy against some objections raised by Michael Dummett; and then discusses what implications this conception of philosophy might have for the point (in particular, the educational point) of philosophy. For Dummett's ideas imply that this Wittgensteinian conception of philosophy renders its teaching particularly difficult, and its point especially hard to see. So, on this conception, philosophy seems to become pointless, as well as difficult. Further, it seems to offer no clear place for reading the history of philosophy; but none of these 'seemings' is actually so.

           I shall not argue for the Wittgensteinian conception directly, although meeting Dummett's objections may indicate some lines of discussion.


§1       Dummett on Wittgenstein on Systematicity

I begin from Dummett's claims that the lack of systematicity in Wittgenstein's later philosophy is a reason not to 'rate' Wittgenstein: that his work fails to offer a future to analytical philosophy. [Apart from his eminence, Dummett is selected because he would not typically be thought a 'hostile witness' here: Dummett is typically seen as 'for the defence', not the prosecution, of Wittgenstein.]

           As Dummett[1] presents it, at a certain period/time in English-speaking (analytical[2]) philosophy:

... the attempt to be systematic in philosophy was the primal error, founded upon a total misconception of the character of the subject (TOE p. 438).

Again, as Dummett reports it, the villain of the piece was (a caricature of) Carnap, whose Aufbau[3] was taken as an exemplar of the mistakes -- mistakes later (accurately) characterised by Richard Rorty[4] as "... the search for universal commensuration in a final vocabulary". But, Dummett asserts, this rejection of systematicity was based on a misunderstanding of what systematicity requires: in fact, there are "... two distinct senses of the term 'systematic' .." (TOE p. 455):

(1)       "systematic" as "... intended to issue in an articulated system ..." (TOE p. 455): Dummett's examples are the ideas of Spinoza, Kant, possibly Carnap[5].

(2)       "systematic" as "... proceeds according to generally agreed methods of enquiry, and its results are generally accepted or rejected according to commonly agreed criteria" (TOE p. 455)

According to Dummett, Wittgenstein denied both of these to philosophy. Rejection of the first is not contentious (at least, not today!). But the second version, too, was rejected. In particular, the opposition to systematicity of this kind was fuelled by Wittgenstein's idea (as Dummett expresses it, and not exactly badly) that philosophy deals with:

... characteristic kinds of intellectual confusion the only remedy for which is extended or patient treatment, in the sense in which a doctor treats an illness. It is this treatment which is the proper work of the philosopher .... (TOE p. 439)

And, of course, this conception was found in Wittgenstein's thought that his aim in philosophy was "... to shew the fly the way out of the fly bottle" (PI § 309[6])

           Dummett's characterisation of where Wittgenstein goes wrong centrally concerns the nature of the project appropriate for philosophy. For Dummett, Wittgenstein's:

... work does not constitute ... a solid foundation for future work in philosophy. (TOE p. 452)

So, first, Wittgenstein's view is inadequate (among other things) in failing to "... provide a foundation for any future attempt to construct a theory of meaning" (TOE p. 453: my re-ordering). Second, this failure is important because, Dummett says, "... the philosophy of language is the foundation of all other philosophy because it is only by the analysis of language that we can analyse thought" (TOE p. 442). So, for Dummett, Wittgenstein refuses to approach a genuinely fundamental question.

           In another place, Dummett urges that:

... systematisation is not ... motivated solely by a passion for order: like the axiomatic presentation of a mathematical theory, it serves to isolate initial assumptions. (OAP p. 20)

And, later in the same piece, he remarks:

... it is another disadvantage for the repudiation of system that it leaves us with no way of judging, in advance of the attainment of complete success, whether a strategy is likely to be successful. (OAP p. 21)

To bring out this point, Dummett raises an additional issue (see TOE p. 455): can there be progress in philosophy? If philosophy is "... one ... sector in the quest for truth" (TOE p. 455), why has it:

... not yet established a generally accepted methodology, generally accepted criteria of success and, therefore, a body of definitely achieved results (TOE p. 455)?

           One manifestation of this mistaken conception of philosophy, for Dummett, is the impact it has on the philosophy of language, since (in Dummett's view):

... the philosophy of language is the foundation of all other philosophy because it is only by the analysis of language that we can analyse thought. (TOE p. 442)

As he puts it elsewhere,

... a philosophical account of thought can be attained through an account of language, and ... a comprehensive account can only be so attained. (OAP p. 4)

 -- and, we are told, this view was held by Wittgenstein "... in all phases of his career" (OAP p. 4). Here one major issue between Wittgenstein and Dummett's hero, Frege, concerns the degree to which:

... it is essential to our language that its employment is inter-woven with our non-linguistic activities. (TOE p. 447)

Dummett recognises that Wittgenstein's project involves denying:

... that the distinction between sense and force is available to simplify the task of explaining the meanings of sentences. (TOE p. 450)

But this cannot be a suitable project for philosophy if philosophy's task involves articulation of (the possibility of) a systematic theory of meaning, since (Dummett says):

... it is difficult to see how a systematic theory of meaning for a language is possible without acknowledging the distinction between sense and force. (TOE p. 450)

In summary, then, one might think of Dummett's contentions that Wittgenstein's work cannot provide a "solid foundation for future work in philosophy" (TOE p. 452) as involving the following five elements:

  • first, Wittgenstein's unsuitability as a model derives from his failure to aim at a systematic theory of meaning, in the (fundamental) philosophy of language, where,
  • second, such systematisation would be clarificatory by isolating "inital assumptions" (OAP p. 20), and
  • third, such systematisation would permit us to recognise progress in philosophy;
  • further, fourth, a key practical element is Wittgenstein's rejection of a sense/force distinction, while
  • fifth, a key theoretical element is Wittgenstein's 'therapeutic' conception of philosophy.


           Clearly one strategy for a (genuine) defender of Wittgenstein's position would be to urge that, despite appearances to the contrary, Wittgenstein's ideas are potentially systematic in at least the second of Dummett's two senses (the one about methods, procedures and results): consider his involvement with Waismann on what was finally published as The Principles of Linguistic Philosophy[7]. As Gordon Baker[8] has demonstrated, Wittgenstein hoped that this text -- while 'deniable' -- would offer just such a systematic account of his thought. Since he implicitly calls Philosophical Investigations a "bad book" (it has "not come about" that he should produce the "good book" [PI p. x] he was hoping for), some of its inadequacy might consist in its failure to present the systematicity of his views[9]. However, I shall not take that tack. Rather, I shall urge some sense of progress in philosophy.

           First, though, it is worth noting a characteristic (an oddity?) of Dummett's conception of both systematicity and its role: Charles Travis[10] draws a helpful contrast between two attitudes to the philosophy of language, two ways of understanding the enterprise:

One approach to the study of language views its subject as weird and wonderful ... to be discerned by looking very closely; another views language ... as something we probably could have cooked up one day, along with the soup...

On the first approach, the primary goal is exactness ... On the second, it is precision -- describing whatever phenomena in a way that leaves no blurred edges...

Of course, following-through on either attitude might be thought systematic, although the results would typically be radically different. Clearly, Dummett views language, and by implication philosophy, in terms of the second such picture[11] (perhaps this is part of Dummett's 'inheritance' from Quine and Davidson?[12]). But Wittgenstein's view seems closer to the first, on which what is needed is a clearer view of what we do (or might) say about .... whatever; and when we might mislead ourselves by saying this. Yet, if this is correct, might it not suggest that Dummett's arguments against Wittgenstein are misdirected?

           Second, it may be helpful to consider one (irrelevant?) reason why Dummett's view might seem plausible, a reason deriving from the nature of the texts under consideration.


§2       Interlude: The Illusion Created by Wittgenstein's (Published) Texts

Wittgenstein's published texts are misleading here for at least three related reasons. First, they seem to lack targets: thus, for example, Richard Wollheim[13] [typically another Wittgenstein fan] writes of Part Two of Philosophical Investigations as a place "... where whatever structure the book possesses elsewhere is more or less abandoned", with the implication that it never possessed very much! Second (and relatedly), the texts seem to be dogmatic: for instance, in claiming that one will not find anything common to games, in virtue of which they are all games. Third, they seem unsystematic, here meaning something like "random" -- a collection of aphorisms rather than an argument.

           But the reality is quite different. If we look carefully at Wittgenstein's writings, we find, first, that Wittgenstein's writings do contain arguments (when properly understood): as Gordon Baker[14] urges:

... we should proceed on the basis that the texts which Wittgenstein constructed himself consist of carefully thought out arrangements of remarks whose precise wording was of paramount importance. (B p. 66)

On the face of it, treating a body of writing as a work in philosophy seems to presume, as a starting point, that these are carefully wrought arguments, not collections of aperçus. (But, as below, it is hard to find this thought in the editing of many of the key texts.)

           Moreover, although Wittgenstein's remarks are not a string of oracular utterances for interpretation, neither are they transparent to casual scrutiny. Rather, they must be understood as part of an on-going argument, although sometimes with opponents whose views must be reconstructed. For instance, to understand Investigations §67, we may need to see its purpose, which may be clarified somewhat by reflecting on what happens next! For suppose that the term "family resemblance", introduced there, was indeed a new technical expression -- perhaps one of great power: what would we then expect? I would expect to see Wittgenstein utilising this technical term in the remainder of the text, exploiting its power etc. But what do we find instead? The expression "family resemblance" does not re-appear in Investigations -- Wittgenstein does speak of a " family of meaning" (PI §77), a "family of structures" (PI §108), a "family of cases" (PI §164), and a "family of language-games" (PI §179), but these are all standard uses of the familial metaphor!

           Moreover, failure to reconstruct the argumentative purpose can make Wittgenstein's work seem less 'systematic' than it is: perhaps we are no longer prey to the perplexities against which his arguments are directed -- so we do not recognise his polemic against them.

           Second, the editing and/or translating produces (at least some of) the garble. As Baker has lamented:

... this principle [of construing Wittgenstein's writings as emboding carefully constructed arguments] does not apply .. to the texts compiled by editors in various more or less systematic ways from his manuscripts. (B p. 66)

He could add that the translations too are not always transparent[15].

           Further, much turns on Wittgenstein's conception of the point of philosophy: that it aims to show flies the way out of flytraps. But  this too seems to make matters worse! As we have seen, much of Dummett's criticism is explicitly directed at what has been called Wittgenstein's "therapeutic" conception of philosophy -- it seems to support Dummett's objections. If remarks in philosophy are "for a particular purpose" (PI §127) or "for some particular person" (compare BT 406-7), how can they be systematic? And how can they offer a future?

           I turn to a fuller (if still brief) account of Wittgenstein's position later. For now, it is enough that this position be recognised as sufficiently distinctive that failure to grasp it, or its implications, is likely to result in misunderstanding.


§3       A Response to Dummett

Earlier, I summarised Dummett's contentions that Wittgenstein's work cannot provide a "solid foundation for future work in philosophy" (TOE p. 452) as  five elements. I shall now consider each in turn. So:

  • Wittgenstein's unsuitability derives from his failure to aim at a systematic theory of meaning, in the (fundamental) philosophy of language.

           It is clear, I think (although contrast both early Baker and Hacker, and my own early writing[16]), that Wittgenstein would have agreed that he was not in this way aiming at a theory of meaning, systematic or otherwise: and citations to his repudiation of the idea of theory in philosophy [cf PI § 128; BT 419 (PO p. 179); PI § 599], as well as the 'fly-bottle' quotation, illustrate this. For Dummett, though, Wittgenstein ["... in all phases of his career" (OAP p. 4), we are told] held  "... first, that a philosophical account of thought can be attained through an account of language, and, second, that a comprehensive account can only be so attained." (OAP p. 4). Had Wittgenstein held such a view, his failure to pursue this sort of account of language would have been frankly odd.

           What would Wittgenstein have made of the first of Dummett's two points: would he have considered "a philosophical account of thought" something philosophers should, or even could, engage in? After all, what philosophical issues could such an account addresses; what flies does it lead from which fly-bottles? Perhaps he would have rejected this explanation of the importance of the philosophy of language: thought is not all that is crucial, in contrast to the usual, rationalistic accounts -- as Wittgenstein was fond of quoting from Goethe, "In the beginning was the deed" (OC § 402; CV p. 31; PO p. 395).

           Dummett's second point involves a huge confusion, identified by Wittgenstein: for what would a comprehensive philosophical account of thought be, exactly? Surely, as Wittgenstein implies, there could be no such thing -- rather, we must recognise the diversities within what perplexes us and within what, at a particular time, makes sense as a dissolution of a particular person's perplexity. For Wittgenstein did not see the relation of language to understanding as linear, with one form of words (suitably disambiguated) uniquely articulating one meaning only. Yet this did nothing to undermine the clarity of certain verbal expressions: any difficulties here are not somehow generated by language -- if anything, they arise because we seek a uniform 'reading' of forms of words. Wittgenstein (PI §402) asks us to consider a case where:

... we disapprove of the expressions of ordinary language (which are after all performing their office), ... we are tempted to say that our way of speaking does not describe the facts as they really are. As if, for example, the proposition "he has pains" could be false in some other way than by the man's not having pains. As if the form of expression were saying something false even when the proposition, faute de mieux, asserted something true. (PI §402)

Suppose the person really is in pain (what is asserted is true). Still, a 'philosopher' might urge that saying, "He has a pain" implies that one owns or possesses pains, as saying, "He has a gun" would imply possession of a gun: and that pains are not objects, not possessible in this sense. So the form of expression might seem to introduce something false. Taking such a line is not, as Wittgenstein notes, seeing something profound, although it may appear to be. Rather:

... we have got a picture in our heads which conflicts with the picture of our ordinary way of speaking. (PI §402)

And it is that new picture that misleads us. We do not really infer, for example, that the utterance "It is raining" misrepresents the facts because we cannot answer the question, "what is the 'it'?"[17]. We understand actions done "for your sake" and the claim that "It is raining" without postulation of an ontology of sakes or of raining its. And, as we will see, failing to recognise this relation between understanding and language is of fundamental importance: it lets-in the philosophical thesis that what is possibly misleading does in fact mislead.[18]

           For Wittgenstein, the central issue would be whether such a theory of meaning, if discovered, could (in principle) be philosophically revealing -- if it could give us insight that was philosophically relevant. It is this he denies: for example, in the famous sections Investigations §§66-67, the advice "Look and see" is aimed straight atthe heart of the project of Analysis -- the issue is not an empirical one. Rather, even if discovered, the "something in common" cannot be the basis of our understanding, since (a) we do not know it and (b) we do understand...

           Here, it seems to me, lies the interest of Wittgenstein as 'fellow traveller' of Lyotard[19]: despite its obvious absurdities, this account acknowledges that Wittgenstein was not a standard analytical philosopher. In contrast, consider Dummett's rueful comment:

... I could do no more than argue in favour of the analytical side on points where they [the analytical and phenomenological approaches] diverged. A book covering the same ground, written from a phenomenological standpoint, would be a counterweight of the highest interest ... (OAP p. xi)

Surely Dummett is not right here. The book Dummett imagines would (following Ramsey's Maxim[20]) recapitulate the errors shared by both sides (as Dummett's does?), errors Wittgenstein repudiated! [For Wittgenstein knew both 'traditions' (to some degree), and was near enough to the 'separation' to consider it with some detachment: also, despite his 'recruitment' into Cambridge, Wittgenstein was no typical analytical philosopher because one might usefully summarise his later ideas as the repudiation of the Ideal of Analysis (see PI §§ 66-67); and of the misunderstanding of philosophy it embodied.]


  • Systematisation would be clarificatory by isolating "inital assumptions" (OAP p. 20).

Again, this is revealing of Dummett's model; in particular, its mathematical character -- but does it tell us more than that? There are two problems here: first, to what extent are these presuppositions of axiomatisation as such, rather than of the particular theorem? For presuppositions of axiomatisation are relevant only on the assumption that axiomatisation is the right strategy here: and that remains to be demonstrated. Second, can we really identify "assumptions" when what is at issue is what, with Collingwood[21], we call "absolute presuppositions" -- that is, framework considerations[22]? Either of these lines of argument might serve against Dummett's claims.


  • Systematisation would permit us to recognise progress in philosophy.

Recall here that Dummett's claim was that:

... it is another disadvantage for the repudiation of system that it leaves us with no way of judging, in advance of the attainment of complete success, whether a strategy is likely to be successful. (OAP p. 21)

Is this supposed to mean that there is no 'method' here? In one sense this is true: there are no formulae to be recited. In another sense, we can recognise treatments as Wittgensteinian (and also as pseudo-Wittgensteinian) when we see them -- to that extent, there are 'methodological characteristics' here. The difficulty seems to be that these characteristics do not uniquely determine what to do next. But would anything?

           First, and quite generally, we might be suspicious of any claim which either made the application of certain rules uncontentious, or guaranteed what their application would amount to. Second, and more specifically, we might wonder what exactly Dummett is preferring to Wittgenstein's approach. To put that issue another way, how genuinely systematic can any such account be, if the criterion of systematicity is that we be required to be able to judge, in advance of the attainment of complete success, whether a strategy is likely to be successful? Much will turn on how strongly the criterion is applied. Surely we cannot, with good heart, confidently move forward on any strategy which we judge to be likely to be successful merely on the basis of its present/past successes. For instance, 19th century physics judged that its strategies were likely to be successful: that it had 'all the answers' (bar a few trivially anomalous cases). But ten years later the whole edifice was in ruins! Merely judging that one had a method was not enough: and neither is pure inductive argument. Yet how can we make do with less and still satisfy Dummett's requirements?[23]

           To approach from another direction: could a computer follow Dummett's preferred method, supposing he found one? If not (and I assume Dummett does not think that computers could replace philosophers here) isn't there a sense (at least) in which the 'problem' he identifies for Wittgensteinian approaches besets his also? And if Dummett remains undaunted, surely Wittgenstein can too!


  • A key practical element is Wittgenstein's rejection of a sense/force distinction.

As recorded earlier, a major issue between Wittgenstein and Dummett's hero, Frege, concerned the degree to which "... it is essential to our language that its employment is inter-woven with our non-linguistic activities" (TOE p. 447). Wittgenstein was clearly committed to an expansive answer here: as he regularly quoted from Goethe, "In the beginning was the deed" (as above, OC §402; CV p. 31; PO p. 395). But much turns on which activities are thought non-linguistic -- consider what might count as parts of language under certain circumstances: Wittgenstein shows us considering as parts of language (die Sprache: B p. 119) gestures (for example, pointing at someone), drawings (such as those accompanying a geometric proof), or colour-samples (in the language-game in PI §8; PI § 16). Finally, consider the standard metre discussion in Investigations (PI §50): not all remarks about the standard metre are bi-polar ... the standard metre seems part of 'language', for some purposes.

           Clearly, taking a different view of what counts as relevant to the nature of understanding will alter the form of the philosophy of language (which Dummett takes as a fundamental issue): if there is a genuine impact of context on understanding, and/or if the point is either (a) to address our perplexities or (b) to lead us out of temptation, neither can be wholly general -- they must address the specific perplexities, and must begin (cognitively?) where we begin ...

           In contrast, Dummett argues that, if particularism[24] were true, a string of words "... in an imperative sentence must have a meaning of a quite different kind from the same words when they occur in an optative sentence; and that is absurd" (TOE p. 449).

           Perhaps ... but where does the absurdity reside? If sentences as such were not the bearers of meaning (if instead it were utterances, or sentences-in-contexts [sometimes called "statements"]), this absurdity is avoided: of course, the contribution of the sentences to what is meant in uttering them (or using them in context) cannot be ignored -- but it is not all; nor should we think it is. (Making out this thesis is a big problem[25].)

           Should we take a commitment to the sense/force distinction as some kind of indicator here of the possibility of systematicity? Certainly, Dummett gives it centre stage: "... it is difficult to see how a systematic theory of meaning for a language is possible without acknowledging the distinction between sense and force" (TOE p. 450). But Collingwood (and maybe Bradley) rejected that[26] -- so it isn't a direct consequence of commitment to analytical philosophy, much less of Wittgenstein ...

           Dummett recognises Wittgenstein's project: "... to deny that the distinction between sense and force is available to simplify the task of explaining the meanings of sentences" (TOE p. 450). But a key word here is "simplify" -- Wittgenstein's counter-blast would identify simplification of a destructive kind .... perhaps of the kind that leads some writers to imagine that formalisations 'nail down' arguments otherwise left to float free! In the first edition (1972) of his Insight and Illusion, Peter Hacker provides one example of 'destructive' simplification, where "[t]he simplest way of introducing order into the medley ..." of elements said to be related criterially is to "... take the criterial relation to hold between sentences, and to suggest that other uses of 'criteria' can be reduced to this" [p. 286-287] (Dummett makes a similar move[27]). But might this not be the beginning of the garden path? If, in line with the second of the two attitudes to the study of language distinguished earlier (by Travis), the aim is one of exactness, such a craving for simplicity must be out of place.

           One root, then, of Wittgenstein's "... deliberately unsystematic philosophical method" (TOE p. 450) was his rejection of the sense/force distinction. And Dummett treats this issue as whether or not a "systematic theory of meaning is possible" (TOE p. 451). He tries to show that Wittgenstein cannot consistently answer either "yes" or "no" to this question. I imagine Wittgenstein replying in two parts: first (as previously), "Well, say what you like"[28]. And, second, "What has this to do with philosophy?". That is, its relevance needs bringing out (a point to return to).

           As we saw (and to repeat), for Wittgenstein, the issue would be whether such a theory, if discovered, could (in principle) be philosophically revealing -- if it could give us insight that was philosophically relevant. But this issue is not empirical. As we noted from the famous sections Investigations §§66-67, even if discovered, the "something in common" cannot be the basis of our understanding. Hence the project of philosophy cannot be the search for such commonalities. And this is the crucial flaw of the project of Analysis

           In this way, then, Dummett might be seen as imposing an unreasonable requirement: philosophy might have a future without being systematic in his preferred sense. But to see how is to see how Wittgenstein's alternative conception of philosophy functions. So the points discussed thus far really bring us to the nerve of the whole debate: the centrality of Wittgenstein's conception of philosophy. For accepting that it is plausible will permit us to turn the 'not proven' verdicts achieved thus far into reasons for acquittal.


  • A key theoretical element is Wittgenstein's 'therapeutic' conception of philosophy.

As we have seen, Dummett is exactly right to locate his disagreement with Wittgenstein in terms of the very different tasks they alot to philosophy, and (correspondingly) the different possibilities they see for it. But we need to be much clearer (than Dummett is?) of Wittgenstein's views here. The aim, as Wittgenstein describes it, is to provide:

... an order in our knowledge of the use of language: an order with a particular end in view; one out of many possible orders; not the order. (PI §132: 1st emphasis mine)

This Wittgenstein calls "a perspicuous representation" of whatever, a representation which makes the matter perspicuous (clear)[29]. Notice, first, that this is not an order to our use of language, but to our knowledge, to what we recognise: second, that it does not offer new knowledge, really -- rather, it puts in order what we already know; and third, that the order achieved is acknowledged as one ordering among many, selected for some "particular purpose" (PI §127[30]), to answer some puzzlement[31].

           Any representation will mislead, in some circumstances[32]. The right idea here, as Wittgenstein urges, is given by comparing a perspicuous representation with a lamp which, in illuminating the perplexing 'side' of the issue, necessarily throws its other 'side' into shadow[33].

           What we lack is not a (generalised or once-and-for-all) perspicuous representation of our grammar/language[34]; but, rather, a certain representation (for example, the colour-octahedron [PR p. 51ff; WWK p. 42; PR p. 278][35]) to make some part of our grammar, a part which puzzles us, perspicuous (=clear) for us. We are puzzled because what we know, in knowing how to go on in language, is not clear to us; misleading analogies (and such like) suggest themselves[36] -- or someone suggests them to us![37]

           The task of philosophy, on a therapeutic conception, cannot turn on matters in principle unavailable to us -- the perplexities typically arise from looking in the wrong way at what we know, rather than from not knowing enough. [As we have seen, Wittgenstein's characteristic advice, "Look and see", is only plausible if what one is looking for is in plain sight.] So philosophy asks us to re-assess what we already know: the point about the fly-bottle just is that flies do not typically find their own way out, even though there is nothing stopping them!

           This account acknowledges a place for the requirements of surveyability, and can accept temporal points of view -- what perplexes us can change ... and therefore what we need in order to dissolve the perplexity can change too! A key part of the temporality here is the temporality of questions; of what can perplex us -- whether or not it could perplex others. We recognise that, as Baker and Hacker[38] put it:

We cannot return the apples from the Tree of the Knowledge of History ... Though we know the Impressionists were outrageous revolutionaries in the theory and practice of painting, we can no longer see them as outrageous.

But these are facts about what it is to see the Impressionists as part of an on-going tradition, where later events shape how we can make sense of the past.

           So the judgements we make are located historically; some can only be made with hindsight, in two different ways. Thus, historical percipience is required to assert that in such-and-such a house is presently being born the greatest physicist of the twentieth century -- suppose both that I am pointing to Einstein's birth-place, at his birth-time, and that Einstein is the greatest physicist of the twentieth century: hindsight is required, because only the passage of time proves it true (or false). In contrast, the judgement that the creator of the theory of relativity is presently being born in the house makes no sense: the expression "theory of relativity" requires later conceptual events even to be meaningful. More than mere hindsight is required: the concepts which make sense of that assertion are only available to us at a certain time and place. Yet when those concepts are in place, the assertion is true; and earlier persons could not deny the assertion, since to them it would make no sense. In this way, the judgements 'pass one another by': they are incommensurable in the strict sense of being unable to be put into one-to-one correspondence with one another[39]. The conclusion to draw is that what counts as possible now (and hence as true or false) depends in part on the concepts available now -- and we should not under-estimate the epistemological significance of this fact for talk of all possibilities (or of "a finite totality of possibilities")[40].

           Such a fishes'-eye view conception of truth and understanding in the writings of Peter Winch[41] is dismissed as relativism by his critics; and is embraced as relativism (at least sometimes) by Paul Feyerabend.Whatever the view is called, I would defend that view against the criticism that it is just the trivial, self-refuting relativism -- although, as Baker notes, it may still be "... a form of relativism which most [of Wittgenstein's] would-be followers reject" (B p. 43).

           On this conception, philosophy has a different series of tasks from other conceptions. So, if I am right that this project is indeed Wittgenstein's, finding it an inappropriate way of thinking about philosophy will modify our view of Wittgenstein's importance. I will not address that matter here directly. Instead, I will turn, first, to some apparent counter-cases and, then, to some implications of this conception.


§4       Work in Philosophy, on the Therapeutic Conception

As two apparent problems for a therapeutic conception of philosophy, such as Wittgenstein's, consider first the place of the history of philosophy: under the therapeutic conception, what is the point of studying, say, Bradley or Collingwood?[42] Is a concern with the history of philosophy simply a concern with history?

           The second question concerns apparently 'solved' perennial problems: surely the fact that some such matters are solved -- for instance, philosophers are no longer arguing about certain classical proofs of the existence of God (or Gods) -- shows that there are such perennial problems (even if/when we don't know the solution).

           To take these questions in that order: consider the answer of Gordon Baker and Katherine Morris since, first, the impact of their book. Descartes' Dualism[43], is a key motivation here[44]; and, second, the general points might be expected to apply widely to studies in the history of philosophy -- and hence to justify the history of philosophy, if these are the sorts of points that can/should interest us. Baker and Morris highlight two ways in which knowledge of past philosophy might be revealing:

  • "... it might lead us to a see an aspect of concepts to which we have been blind" (DD p. 218)
  • "... making a sustained effort to grasp Descartes' vision might lead us to a greater self-awareness ..." (DD p. 219): it might lead to our coming to "... read his texts with more humility" (DD p. 219) -- surely appropriate for a major thinker -- where the outcome might be "... the possibility of recognising conflicts offrameworks and responding to them intelligently" (DD p. 219).

In general, these might amount to changes in "... nothing less than ways of seeing things or norms for describing them" (DD p. 219). Certainly, this should militate against "...inflexible adherence to such ideas as the idea that there is no rational method except the scientific method, or the idea that science is an instrument for predicting stimulations of our nerve-endings."[45]

           Our position may be clarified by comparison with (one reading of) Collingwood, who urges: "The 'realists' thought that the problems with which philosophy is concerned were unchanging" (A, p.59). Thus they denied the idea that philosophy had a history. For the place of the history of philosophy does seem mysterious on both Ryle's conception of philosophers of the past -- as members of the common-room on extended sabbatical -- and in terms of the Project of Pure Enquiry (as Williams[46] explains Descartes' motives): on either view, suppose we ask the question, "If philosopher X solves this problem, what is the need to read later works?". Here both views agree that the only justification can lie in the trying out of bad answers. For what past philosophers said is either right (and then we just learn from it, treating it as established) or wrong -- and then it is hard to see its value (except perhaps to highlight blind alleys).]

           But, as Collingwood notes, "In a different sense of the word, the 'realists' thought that philosophy had a history" (A, p.59), because various philosophers had given different answers to the eternal questions of philosophy. Yet, as Collingwood recognises, this serves to separate the question of, for example, what Aristotle said about duty (the historical point) from that about whether Aristotle was right about duty, or whether what he said was true (the philosophical questions). Collingwood continues:

... I soon realised that the history of political theory [and by extension of philosophy more generally] is not the history of different answers to one and the same question, but the history of a problem more or less constantly changing, whose solution was changing with it. (A, p.62)

Since the questions change and the reasons which might count towards an answer change, it follows that the answers themselves change -- even if they appear in the same words[47]. So the focus when studying philosophers of the past (for philosophical purposes) must be on what were issues or problems or perplexities for them.

           But it is easy to produce a "gross over-simplification"[48] of Collingwood's position here. For some who have either asserted or denied that Collingwood aims to turn philosophy into a species of history have read him as treating philosophy as the study of (relatively) fixed sets of presuppositions -- just as, in history, it might be thought that Hadian's Wall had a purpose or purposes, and study consisted in ascertaining what these were. So philosophy was to be historical in this sense. (And then the disputants have urged that Collingwood did or did not argue for this.)

           On such a reading, Collingwood's project is not strongly similar to Wittgenstein's. For providing a perspicuous representation cannot, in this way, be thought of as one, fixed task: rather, it is about the addressing of contextual perplexities -- there is no final or absolute account here (Z §447); only the contextual answer to some specific perplexity[49]. So finding what will be revealing here is a difficult matter both because one must search for a revealing image or metaphor or model, where 'revealing-ness' can be hard to determine: and must recognise that it must be revealing to the specific audience for the remark [".. for some specific purpose" (PI §127)]. But doing so will give a place to the history of philosophy.


§5       'Solved' problems?

What are we to make of philosophical problems 'solved'? Can the therapeutic conception countenance this possibility? In fact, an intermediate case will suffice[50]: suppose a bunch of scientists call themselves "Chaos Theorists", and describe certain natural phenomena as "chaotic". (The forecasting of weather is a preferred example.) Now progress against (potential) confusion might be made by bringing an audience to recognise that there is nothing genuinely chaotic about the phenomena. Indeed, one fundamental principle of 'Chaos Theory', the Butterfly Effect (or the Principle of Sensitive Dependence on Initial Conditions) is fully deterministic, since (if we did know both the initial conditions and the relevant laws) we could predict the events in question: the problem with weather forecasting is that we cannot know either the detail of the starting point (the initial conditions) or the (non-linear) equations involved sufficiently precisely. But if we did, weather forecasts could be absolutely accurate -- as the Chaos Theorists' computer-modelling of them demonstrates! And the audience for such a point about determinacy within Chaos Theory would stretch just as far as the influence of the misconception, and also would be just as 'perennial' as that misconception.

           In this way, the need for philosophy should not be expected to diminish:

... fresh puzzles and bafflement may arise as new 'pictures' are invoked by philosophers, frequently from other departments of intellectual endeavour. (B&H 1 p. 477)

Fresh opportunities for misleading ourselves arise, for instance, when we forget that our Chaos Theorists' preferred research tool is the computer -- and that computers genuinely chaotic would be unworkable, for we could not depend on them!

           This case suggests, of course, both the sense in which philosophical problems are perennial and the sense in which they are not. For others might be puzzled about determinacy -- the explanation may need repeating to those who have not heard it (or who have not grasped it!). And there may be no such problem at a later date, because there may be no Chaos Theorists. But whenever there are Chaos Theorists and a bafflement about determinacy, 'the' problem recurs: yet now we need a guarantee that this is one and the same problem -- and our remarks about conceptual change make that a hard request to fulfil except tautologously: that where that perplexity is there, that perplexity is there!

           So philosophical issues might disappear, might become no longer questions puzzling to people -- the concepts to express them might even become unavailable. In such cases, of course, there is no such problem, but in a trivial sense. [We have here a species of the 'paradox of precisification': that we can only make X more precise by turning it into Y -- and then we have not made X more precise!]

           If we consider a problem taken to be solved within philosophy -- for example, the possibility of arguing for God's existence -- we find similar situations still apply: there are those who dispute the so-called solutions, so that only an excessive commitment to rationalism can insist that they must give way here. Equally, there are many arguments destructive of philosophical postions -- I think first of Wittgenstein's argumentative strategy against the possibility of a logically private language -- which have 'sunk without trace' in contemporary philosophy as represented by, say, the collected works of Donald Davidson[51]. Of course, the relative solved-ness (or otherwise) of philosophical problems is just as the therapeutic conception predicts: if your problem is dissolved in such-and-such a way, and if you can internalise that dissolution, then (by definition?) your problem is solved. But satisfying such conditions other than trivially is extremely difficult. We have seen how apparent problems might persist what the discipline considers their solution[52], and how conceptual differences may result in apparent 'solutions' not really dissolving your perplexities: in either of these ways, the therapeutic conception of philosophy has a principled response to this line of criticism.


§6       Teaching and Learning Philosophy

Given that such a (broadly) therapeutic conception of philosophy is to be adopted, what does it mean for the practice -- in particular, the teaching and learning -- of philosophy? We must begin by recognising that, as we have seen Wittgenstein expressing it:

Philosophy is a tool which is useful only against philosophies and against the philosopher in us. (MS 219, 11[53])

Both of these areas of confusion will be 'perennial' in any period: for instance, Descartes' students were permeated both by the general world-picture they were offered throughout their lives (the contribution of philosophies) and by the doubts about such a conception that their conceptual framework permitted, with the philosopher in us representing those doubts that often arise. We are in the same position -- our concepts differ, of course, but both sets are to be understood in terms of the (rather different) conceptual schemes within which we grew up (if sometimes in opposition to such schemes). Further, we may not see what are our needs (for ourselves) -- we may need to be alerted to them[54]. In this situation, how is training in philosophy to be understood?

           It may be helpful to reflect on two standard cases of philosophical puzzlement (of which others can be thought variants):

  • She comes to the philosopher with her perplexity -- not, perhaps, the usual case ... but it might be if the "she" in question were the philosopher, the one perplexed -- this seems OK for one's own case: and works against the accusation of elitism: at worst, there is only "practical elitism"[55] -- the philosopher has actually spent time learning to do what others might have done.
  • He says something which is perplexing him ... but he hasn't yet noticed. Now, the philosopher suggests .... For example, he persists in asking why we are here ... he thinks this is the topic that is perplexing him, but it isn't. Asking this question manifests a perplexity, but not because he can't answer that question: rather, because he thinks this a question/issue. And a perspicuous representation may disabuse him. Equally, he keeps asking why pi is 3.1412 etc. etc. -- for example, he thinks pi could be different in another part of the universe[56]. Again, he thinks his perplexity concerns one thing; and he is perplexed ... but not where he thinks. (And, again, the perspicuous representation may show him that his perplexity is not what he thought ...)

In this second case, one needs to work through issues for oneself: we will typically end-up asking such questions, as the result of standard education -- they represent the philosopher in us (MS 219, 11). Therefore we will typically need to sort them out, at least once. And this is precisely Descartes' position: in order to make a sustained effort to break bad intellectual habits:

[t]he seeker after truth must, once in the course of his life, doubt everything, as far as is possible. (Principles  Part I §1: CSM vol. 1 p. 193 [section title]; H&R vol. 1 p. 219)

The key expression here is "once in the course of his life", for only on the therapeutic conception do we all need this. On the contrary, if Descartes really were after (Williams') 'Project of Pure Enquiry', and if he'd been successful, then his results would have done for us all: we wouldn't need to go through it ourselves. But this is not what Descartes says.

           Nothing in Descartes' position makes this obligation to think through one's foundations uniquely his: so we might take this too as suggesting a general obligation[57]. And what are one's distinctive issues? Some issues might be characteristic of humankind -- at least within some (historical) period[58] -- others might be characteristic of certain cultures or value-systems: in both cases, we can see the philosopher in us as raised by contemporary education or the state of society.

           It may help us if we ask what a philosopher needs to do! What are the training requirements for philosophy? Clearly, these cannot be formalised -- but can we describe the ideal outcomes?

           Then we should see training in philosophy as having at least four related aspects:

  • others' (genuine[59]) perplexities must be recognised: this might be thought an ability not specifically learned.
  • also, the ability not to mistake for perplexities what actually aren't.
  • the ability to find revealing comparisons etc: that is, to find (elements of) representations which will be perspicuous (for us -- that is, for the perplexed)
  • the ability to find representations that will be perspicuous of our grammar[60] when this is relevant factor to know ... to find a magic or redeeming word [erlösende Wort: PO p. 165].

So that there can be no general rules here, but rather a series of matters (a) that we might train by the "mother's method"[61] and (b) that we might be more and less good at.

           This explains, too, one of the oddities of being a philosophy student (say, for a undergraduate degree): for such a student will not typically have philosophical puzzles -- or, if her/she does, it would be surprising if they lined up neatly with the topics undertaken for the course of study (same topics in the same order at the same times seems too much to hope for). So the teacher will, familiarly, have to show both that someone might be puzzled here and how that puzzlement might be removed.


§7       Slow cure?

Whatever one thinks of the philosophical merits of this position, it might seem that there is a straight forward scholarly refutation of it: for did not Wittgenstein insist upon a slow cure? And how is that compatible with my emphasis on the revelatory aspect of perspicuous representation?

           In fact, the matter is claified by considering the key passage (Z §382[62]):

In philosophizing we may not terminate a disease of thought. It must run its natural course, and slow cure is all important.

There are really three key points here: the first, made sharply by Peter Hacker[63], concerns the idea of duration:

... it is rarely only one misguided analogy that bewitches us, but rather a whole range, and as we cast light on one, long shadows distort others. So we must slowly and patiently survey the structures of intellectual illusion from all sides, for they cannot be taken in at a glance.

So the slowness follows partly from the need (typically) to do many things!

           The second point concerns the idea of slow in this context: with what is it being contrasted? As Wittgenstein makes plain, the contrast is with hasty. So that a slow cure here is one that takes its time, that is not hurried. So it must be suitably slow, in respect of the 'working through' of that confusion. And, of course, such aslow cure might be achieved in a flash (perhaps with a "liberating word": PO p. 165) -- if that were the appropriate timing for this cure: what Wittgenstein speaks of as running "... its natural course".

           A third, and related point: as Wittgenstein identifies the problem here, it lies in trying to "terminate" the condition -- to bring it to a slick resolution. So one part of the danger here lies in kinds of 'terminating' of philosophical puzzlement aimed at by ... well, Wittgenstein (Z §382) suggests that "mathematicians" will typify those who make this mistake. So we might expect it to come from the first of the two accounts of understanding language offered by Travis above, the one stressing the ideal of a structure to be uncovered by analysis, such that "... the grammar of natural language often needs substantial revision before it will yield the logical structure of our thoughts."[64] And this is precisely not Wittgenstein's own view.

           In this fashion, then, Wittgenstein identifies ways that philosophy can be done trivially, by (say) paying mere lip-service to alternatives, rather than investigating them. For clearly there is a kind of briskness that would be a major vice: not genuinely addressing the problem. And to think that the simplicity of a system were required might tend towards just such an attitude.



In an odd way, the thesis of this paper is extremely traditional: as we have seen, it is found in Descartes (for example), widely recognised as the founder of modern philosophy. But it isn't a tradition that has actually flourished since (I would argue -- following Baker and Morris) this is not how Descartes (and others) have been read ...

           The thesis: that the odds are that philosophical problems/perplexities will arise for all of us; and that dealing with these perplexities is standardly a matter of exactness. As Wittgenstein put it:

The choice of our words is so important, because the point is to hit upon the physiognomy of the thing exactly, because only the exactly aimed thought can lead to the correct track. The car must be placed on the tracks precisely so, so that it can keep rolling correctly.

One of the most important tasks is to express all false thought processes so characteristically that the reader says, "Yes, that's exactly the way I meant it". (BT pp. 409-410 [PO p. 165])

That is to say, it is systematic, but ... Since it offers a way of understanding the nature of philosophical problems, this account illuminates how to move forward -- although not in terms of a method for the resolution of such problems: and it does so because it recognises the specificity (to persons and contexts) of philosophical problems. Dummett does not see this -- hence his unrealistic and inappropriate account of the systematic. In contrast, Wittgenstein offers an account of philosophy which explains both the difficulty and the importance of its teaching and learning.

           When I speak about philosophy, it often seems to have one of two characters not present as I experience it: it can seem a self-indulgence (just directed at my own soul) or to be extremely patronising, with philosophers solving perplexities others barely see. But both these illusions are generated by taking as general claims what are really context-bound comments.

           Bernard Williams famously asked (I quote from Tom Nagel's account[65]):

What is the point of doing philosophy if you're not extraordinarily good at it? ... If you're not extraordinary, what you do in philosophy will be either unoriginal (and therefore unnecessary) or inadequately supported (and therefore useless). More likely, it will be both unoriginal and wrong.

But this is the wrong question, for all its perceptiveness: for work in philosophy is more contextual than that, more a matter of dealing with specific perplexities -- as Wittgenstein put it:

Work in philosophy is ... actually more of a kind of work on oneself. (BT pp. 406-7; PO p. 161)

But then the achievements can be real (if small) if they answer that perplexity: the 'progress' need be no more abiding than that.



[1]              Michael Dummett Truth and Other Enigmas, London: Duckworth, 1978, cited as "TOE".

[2]              As we will see, Dummett gives an unusual account of the nature of analytical philosophy (for example, in Michael Dummett Origins of Analytical Philosophy, London: Duckworth, 1993, cited as "OAP" p. 4). As Gordon Baker and Peter Hacker (Frege: Logical Excavations, Oxford: Blackwell, 1984 p. 7 note 10) put it:

... the characteristic tenet ... is that the philosophy of language is the foundation of the rest of philosophy

As Baker and Hacker continue, this is not a thesis characteristic of most philosophers normally called "analytical".

[3]              Rudolf Carnap Der logische Aufbau der Welt, trans. G.A. George (as The Logical Structure of the World), London: Routledge & Kegan Paul, 1967.

[4]              Richard Rorty Philosophy and the Mirror of Nature, Princeton: Princeton University Press, 1979 p. 368.

[5]              Carnap had both conceptions: see the Aufbau p. xvii (quoted Stephen Hilmy The Later Wittgenstein Oxford: Blackwell, 1987 p. 315 [n. 494]):

... stone will be carefully joined to stone and a secure building will be erected at which each following generation can continue to work.

[6]              I discuss some elements of this issue in "A nasty accident with one's flies", Inaugural Lecture at the University of Brighton, 7th May 1996.

[7]              F. Waismann The Principles of Linguistic Philosophy, London: Macmillan, 1965.

[8]              Gordon Baker "Verehrung und Verkehrung: Waismann and Wittgenstein" in C.G. Luckhardt (ed) Wittgenstein: Sources and Perspectives, Hassocks, Sussex: Harvester Press, 1979 pp. 243-286.

[9]              Re. this dissatisfaction on Wittgenstein's part, see Hilmy The Later Wittgenstein pp. 20-25: but this might only show that Wittgenstein was dissatisfied with the systematicity of the presentation of his ideas.

[10]             Charles Travis The True and The False: The Domain of the Pragmatic, Amsterdam: John Benjamins B V, 1981 p. 1.

[11]             In contrast to Wittgenstein's own view: see, for example, MS 112 p. 267 (quoted Hilmy The Later Wittgenstein p. 216):

My own view was false ... because I too thought logical analysis must bring to light hidden things (as chemical ans physical analysis does).

               Also MS 213 pp. 100-101 (=PG p. 210; quoted Hilmy The Later Wittgenstein p. 264 [n. 233]), where it begins, "My notion in the Tractatus was wrong ..."

[12]             See here Thomas Nagel Other Minds: Critical Essays 1969-1994, Oxford: Oxford University Press, 1995 p. 6:

... somehow the Quine-Carnap tradition has come to dominate the profession ...

[13]             Richard Wollheim On Art and the Mind, London: Allen Lane, 1973 p. 3.

[14]             Gordon Baker "Some Remarks on 'Language' and 'Grammar'" Grazer Philosophische Studien Vol 42 1992 pp. 107-131, reprinted in his Wittgenstein's Method: Neglected Aspects (Oxford: Blackwell, 2004) cited as "B".

[15]             Four points: first, the translation of often demonstrably wrong -- and this leads to an undue emphasis on, say, the idea of language-games. So that, in PI §116, we find:

When philosophers use a word ... and try to grasp the essence of the thing, one must always ask oneself: is the word ever actually used in this way in the language-game which is its original home?

               [See Gordon Baker and Peter Hacker Wittgenstein: Understanding and Meaning (Volume One of an Analytical Commentary on the Philosophical Investigations, Oxford: Blackwell, 1980, cites as "B&H 1" p. 526.] But when we glance across to the German we do not find the word for language-game, Sprachspiel, but instead the German word usually translated "language", Sprache. So, if we want to know about language-games -- and we do -- this passage. The second point concerns the issue of consistency in translation; and again can be illustrated with reference to a key passage. If one accepts, as I do, that Wittgenstein had contrasted definition (German word: Definition) with something else -- the German word is Erklarung -- one cannot then translate Erklarung as "definition": yet that is just what occurs at PI §§71-72 [B&H 1 p. 359-361]. And that translation is especially suspect since there exists another translation of these passages, made in 1938, in which Wittgenstein himself corrected the word "definition" in the translation to "explanation" [Ts 226: PPI(R) §78]. Now, I accept that translators may need to use different terms to translate a frequently-occurring word -- but they must respect technical terms [compare also B&H 1 p. 140 on need for a uniform translation of Behauptung in PI §§21-24, or B&H 1 p. 639 re. Darstellungsform in PI §158 {PPI(R)§113}]. It is certainly true that attending to the context is very important, that:

... we have no business arguing that a word or phrase which occurs in various remarks must have the same significance in them all. (B pp. 127-128)

               Here, though, the effect of differential translation is to cloak what is being said by running together contrasts (possible and real).

                              As I mentioned before, there is an earlier translation of some of this material, corrected by Wittgenstein. Reviewing it reveals our third and fourth points: for (third) the question of the 'free-ness' of the translation needs to be addressed -- some of Wittgenstein's translations are remarkably free [for example, PI §74: B&H 1 p. 365]. Moreover, and fourth, some of Wittgenstein's own translations contradict those in the Anscombe version [PI §108/PP(R) §115: B&H 1 p. 515; PI §86/PP(R) §93: B&H 1 p. 444]. The net effect is that one starts to distrust the translation: for instance, if there is a question about the consistency of two remarks -- and this is probably the most crucial kind of issue when considering the work of a philosopher -- to find that the 'inconsistency' is the result of poor translation is, frankly, galling.

[16]             See, for example, Gordon Baker "Criteria: A New Foundation for Semantics" Ratio 1974 pp. 156-189; Peter Hacker Insight and Illusion: Wittgenstein on Philosophy and the Metaphysics of Experience Oxford: Clarendon Press, 1972; reference to my work removed for purposes of refereeing.

[17]             Peter Hacker Appearance and Reality, Oxford: Blackwell, 1987 p. 500

[18]             As Gordon Baker ("Philosophy: Simulacrum and Form" [in Greek] in Stuart Shanker (ed) Philosophy in Britain Today, London: Croom Helm, 1986 pp. 1-57) articulates a central Wittgensteinian thesis, the important thing "... is to direct attention away from the form of expressions to their uses" (p. 48). For Wittgenstein (as we have seen) forms of words might mislead -- this partly explains his preference for discussions of use.

                              But this idea too has proved to be misleading: people have misunderstood the use of the word "use" here, having taken it as some sort of definition of "definition". Nothing could be further from the truth: indeed, the whole demand for definiteness this implies is misplaced. For example, MS 136 p. 76 (=RPP II §167: quoted Hilmy The Later Wittgenstein p. 285 [n. 360]):

I now want to say that humans who employ such a concept would not have to be able to describe its use. And were they to try, it is possible that they would give a quite inadequate description. (Like most people, if they tried to describe the use of paper money correctly.)

               Wittgenstein is here making two points: about what we need to be able to do, to count as masters of a concept, and about the impossibility of a 'full' account of the use of a concept.

               The use of the term "use" here is straightforward, if negative: it is just to record that our interest in forms of words is essentially contextual -- that forms of words do not have a specific meaning (perhaps, are not the bearers of meaning: see Alan R. White's idea [Truth, London: Macmillan, 1970 p. 7ff] that sentences are not the bearers of truth; or Travis The True and The False  p. 20ff, especially p. 31ff..).

[19]             Jean-Francois Lyotard The Postmodern Condition, Manchester: University of Manchester Press, 1979.

[20]             The term is Renford Bambrough's: see his Reason, Truth and God, London: Methuen, 1969 p. 10:

... wherever there is a violent and persistent philosophical dispute there is likely to be a false assumption shared by both parties.

               See also F. P. Ramsey The Foundations of Mathematics, London: Routledge & Kegan Paul, 1931 pp. 115-116; Ramsey Foundations, London: Routledge & Kegan Paul, 1978 pp. 20-21.

[21]             R.G. Collingwood Essay on Metaphysics, Oxford: Clarendon Press, 1940, cited as "EM"; also R.G. Collingwood An Autobiography, Oxford: Clarendon Press, 1939, cited as "A" followed by page number. As he introduces the idea:

Whether a given proposition is true or false, significant or meaningless, depends on the question it was meant to answer; and anyone who wishes to know whether a given proposition is true or false, significant or meaningless, must find out what question it was meant to answer (A p. 39).

               If this is, as Collingwood supposes, an appropriate way forward, it follows that understanding of propositions conceived of as forms of words inevitably changes as those forms of words represent the answers to different questions. A key insight of Collingwood's, moreover, was that some of the 'beliefs' held by a particular group at a particular time will not be open to question for that group: they will stand firm. Collingwood introduces the idea in the following way.

A presupposition of one belief may be the answer to another question. [Collingwood would then call it a "relative presupposition".] ... The beliefs which a metaphysician tries to study and codify are presuppositions ... not answers to any questions at , all.  This might be expressed by calling them 'absolute' presuppositions (A, p.66-7).

               Before elucidating this slightly mysterious claim it is worth noting that what Collingwood speaks of as absolute presuppositions are not absolute in the sense of being more ultimate vis-a-vis reality, but simply in being immoveable for those who presuppose them (while they presuppose them).

[22]             For Collingwood, an absolute presupposition is one which one makes in asking other questions, but which is not itself an answer to a question.  We can distinguish broadly two types. First those, perhaps of a rather mundane character, which we build into our thinking as framework conditions.  So that it will not do to question my claim that British pillar boxes are red, insofar as this is used to explain the term "red".  For anything which was not just that colour would not be  red at all (Re. colours, see Hacker Appearance and Reality pp. 120-129). In  this  sense, absolute presuppositions closely approximate to the sorts of framework conditions that Wittgenstein called one's form  of representation. Collingwood also implies that one would stand by one's absolute presuppositions in a way which might allow one to become irritated if they were repeatedly challenged.

                    It is worth having before us the famous footnote (EM, p.48) in which Collingwood speaks about conceptual change in respect of one's absolute presuppositions.

A 'change of fashion' is a superficial change, symptomatic perhaps of deeper more important changes, but itself deep or important. A man adopts it merely because other men do so, or because advertisers, salesmen, etc., suggest it to him ... An absolute presupposition is not a 'dodge', and people who 'start' a new one do not start it because they 'like' to start it. People are not ordinarily aware of their absolute presuppositions, and are not, therefore, thus aware of changes in them. Such a change, therefore, cannot be a matter of choice. Nor is there anything superficial or frivolous about it. It is the most radical change a man can undergo, and entails the abandonment of all of his most firmly established habits and standards of thought and action.

               What Collingwood is stressing here is the sense in which one's absolute presuppositions are not a matter of whim, not merely arbitrary. However, he has to acknowledge that absolute presuppositions do change and, if his theory is not to be an entirely conservative one, this is something he must seek to explain. All he says is as follow:

... the absolute presuppositions of any given society, at any given phase of its history, form a structure which is subject to strains of greater or lesser intensity, which are 'taken up' in various ways but never annihilated. If the strains are too great the structure collapses and is replaced by another, which will be a modification of the old with the destructive strain removed: a modification not consciously derived, but created by a process of unconscious thought (EM, p.

What must be acknowledged here is that Collingwood does not answer the question of why such changes come about.  What is clear, however, is that he does not see such changes as straight-forwardly explicable in some 'rational' manner. For they are changes in the framework conditions within which rationality takes place. Just as in science, where anomalies may be the occasion for change but are not its cause, so in understanding more generally: conceptual structures change, as as a result of changing understanding of human 'beings. But we cannot ascribe a 'cause and effect' relation here. As one might say, both change together.

[23]             I am conscious of only addressing one of Dummett's formulations here: of putting on one side the remark about "generally accepted methodology, generally accepted criteria of success and ... a body of definitely achieved results" (TOE p. 455)

[24]             One of Dummett's objections, therefore, is to such "particularism" (TOE pp. 444-445) -- that it led to superficiality because "... it promoted a conscious disregard for the distinction between semantic and pragmatic aspects" (TOE p. 445). In reply, I might, with Charles Travis [The True and the False], argue that the points turn on the relative valuing of the semantic and pragmatic; or I might, with Charles Travis [The Uses of Sense: Wittgenstein's Philosophy of Language, Oxford: Blackwell, 1989], urge that the distinction is acknowledged, but located differently. Certainly Wittgenstein would have urged, "say what you like": the issue concerns what contributes to meaning and understanding -- and Wittgenstein's point (like Austin's?) would be that meaning is really the province of utterances or of sentences-in-contexts: that is, not with sentences tout court.

[25]             Compare Graham McFee "Bradley, Possibility and a Question-and-Answer Logic" in W. J. Mander (ed) Perspectives on the Logic and Metaphysics of F. H. Bradley, Bristol: Thoemmes Press, 1996 pp. 269-287, especially pp. 278-280.

[26]             See McFee "Bradley ..." pp. 280-281.

[27]             In his "What is a theory of meaning (II)?", reprinted in his The Seas of Language, Oxford: Clarendon Press, 1993 p. 34 note:

... my use of "sentence" and "statement" is very inexact. I do not think that this has vitiated the thought ... if the reader will grant me a little licence.

[28]             As Wittgenstein (PI § 48: translation from Hilmy The Later Wittgenstein p. 174) puts it:

Isn't it irrelevant which we say, so long as we avoid misunderstanding in any particular case!

[29]             Compare Gordon Baker "Philosophical Investigations §122: Neglected Aspects", in R. L. Arrington and H-J. Glock (eds.)Wittgenstein's Philosophical Investigations: Text and Context, London: Routledge, 1991 pp. 35-68; reprinted in B pp. 22-51 -- see especially pp. 42-45.

[30]             Compare BT p. 420; PO p. 181 [B&H 1 p. 557]

[31]             Interestingly, this is not quite the reading from B&H 1 p. 557, where it is pointed out that:

The 'order' the grammarian seeks is of a different kind, determined by different purposes.

               So the other 'orders' envisaged seem to compare that for philosophical purposes with those for other purposes.

[32]             Compare MS 120: "I asked him for a bread knife and he gives me a razor because it is sharper!" [B&H 1 p. 352 Note]

[33]             Compare Descartes [see J. Cottingham, R. Stoothoff, and D. Murdoch (eds) The Philosophical Writings of Descartes, Cambridge: Cambridge University Press, 1985, cited as "CSM", followed by volume and page number]:

... a painter cannot represent all the different sides of a solid body equally well on his flat canvas, and so he choses one of the principal ones, sets it facing the light, and shades the others so as to make them stand out only when viewed from the perspective of the chosen side. (CSM vol. 1 p. 132)

[34]             Compare MS 220 p. 92 ( quoted Hilmy The Later Wittgenstein p. 75):

The regulation of traffic in the streets permits and forbids certain actions on the part of drivers and pedestrians; but it does not attempt to guide the totality of their movements by prescription. And it would be senseless to talk of an 'ideal' ordering of traffic which should do that; in the first place we should have no idea what to imagine as this ideal. If someone wants to make traffic regulations stricter on some point or other, that does not mean that he wants to approximate to such an ideal.

               Again, MS 213 pp. 250-269 (CV p. 14: quoted Hilmy The Later Wittgenstein p. 85):

... we have to be told the object of comparison, the object from which this way of viewing things is derived, otherwise the discussion will constantly be affected by distortion. Because willy-nilly we shall ascribe the properties of the prototype to the objects we are viewing in its light; and we claim, "It must always be ..."

[35]             As Baker and Hacker (Language, Sense and Nonsense, Oxford: Blackwell, 1984) note:

Wittgenstein cannot be credited with any originality in denying that the meaning of a name is its bearer, but he introduced this familiar observation into a context of argument in which it suddenly impressed many philosophers. (p. 242)

               This is just what a perspicuous representation should do!

[36]             Compare MS 220 §99 [quoted B p. 30]:

We then change the aspect by placing side-by-side with one system of expression other systems of expression -- the bondage in which one analogy holds us can be broken by placing it alongside another [analogy] which we consider to be equally well justified.

[37]             Lacking such clarity, we will remain puzzled: but, once we have achieved the clear view, it can seem obvious. As Wittgenstein remarks:

Philosophical problems can be compared with locks on safes, which can be opened by dialling a certain word or number, so that no force can open the door until just this word has been hit upon, and once it has been hit upon, no effort at all is necessary to open the door. (BT p. 417; PO p. 175)

[38]             Baker and Hacker Frege: Logical Excavations, p. 4.

[39]             Thus T. S. Kuhn ("Theory-change as Structure Change: comments on the Sneed Formalism", Erkenntnis, Vol. 10, 1976 pp. 179-199):

In applying the term "incommensurability" to theories, I'd intended only to insist that there was no common language within which both could be fully expressed and which could therefore be used in a point-by-point comparison between them. (pp. 190-191)

[40]             As P. K. Feyerabend (Farewell to Reason, London: Verso, 1987 p. 272 [UD p. 307]) notes, incommensurability is a difficulty for philosophers, not for scientists. Here he is contrasting the perspective of the philosopher with that of the scientist, the practitioner: for any contemporary scientist, looking back at the claims of his scientific forebears, may think them wrong (and correctly) while the detachment of the philosopher allows that these forebears' claims are incommensurable with those of the contemporary scientist.

[41]             See, for example, Peter Winch "Language, Belief and Relativism" in his Trying to Make Sense, Oxford: Blackwell, 1987 pp. 194-207, where he identifies and responds to some such critics.

[42]             Note that there might be a different question if the author were (a) methodologically central and/or (b) right -- as one might urge for, say, Wittgenstein.

[43]             Gordon Baker & Katherine Morris Descartes' Dualism, London: Routledge, 1996, cited as "DD" plus page number.

[44]             In effect, this has two aspects. For we study the relationship of the concepts employed by that historical philosopher (a) to others, and (b) to absolute presuppositions

[45]             Hilary Putnam "Beyond Historicism" in his Realism and Reason (Philosophical Papers Vol. 3) Cambridge: Cambridge University Press, 1983 pp. 287-303: quote from p. 288

[46]             Bernard Williams Descartes: The Project of Pure Enquiry, Harmonsworth: Penguin, 1978.

[47]             It is worth noting in passing the Realist (transcendental realist) threads in -- and outcomes of -- traditional British idealism, and especially Bradley. So that Collingwood (The Idea of History, Oxford: Clarendon Press, 1946) urges:

The effect of Bradley's work on subsequent English philosophy was to induce it, in general, to accept this error [realism] as an axiomatic truth ... in Oxford the result was Cook Wilson and Oxford realism; in Cambridge it was Bertrand Russell and Cambridge realism. (p. 142)

[48]             For discussion, see James Connelly "Metaphysics and Method: A Necessary Unity in the Philosophy of R.G. Collingwood" Storia, Anthropologia e Scienze del Linguaggio 1990 pp. 33-156, cited as "Connelly": quote from p. 34.

[49]             See also the classification that puts Travis' work into 'pragmatics', and thereby devalues it vis-a-vis semantics. Contrast Charles Travis "Pragmatics" in Crispin Wright & Bob Hale (eds.) Blackwell Companion to the Philosophy of Language, Oxford: Blackwell, 1997 pp. 87-107.

[50]             Both the case and the presentation derive from the Inaugural lecture, "A nasty accident with one's flies".

[51]             It is an amusing joke when Dan Dennett's scurrilous Philosophers' Lexicon defines "a Davidsonic Boom" as "the sound a research programme makes when it hits Oxford": what is sad is the aptness of the remark!

[52]             This is just to repeat the point that one cannot guarantee that "logic would seize you by the throat, and force you to do it" (Lewis Carroll: see UD pp. 37-38)

[53]             Quoted Peter Hacker Wittgenstein: Meaning and Mind: Volume 3 of an Analytical Commentary on the Philosophical Investigations, Oxford: Blackwell, 1990 p. 264; and in Kenny p. 48

[54]             And, as Stanley Cavell (Must We Mean What We Say?, New York: Charles Scribners Sons, 1969) remarks, someone

... may even have ... counters to your objections, which for some reason he can't give (perhaps because you've brow beaten him into amnesia). (p. 92)

[55]             In line with Schoenberg's idea that "If it is for all then it is not art, and if it is art then it is not for all." (quoted Joseph Machlis Introduction to Contemporary Music, W. W. Norton & Co., 1961 p. 334).

[56]             NB (a) pi [or e] as different, versus (b) the importance for physical theory of pi [or e] as different. (I actually heard this discussion ...)

[57]             Compare DD p. 87.

[58]             See BT p. 424 [PO pp. 185-186]:

One keeps hearing the remark that philosophy really makes no progress, that the same philosophical problems that occupied the Greeks are still occupying us. But those who say that don't understand the reason it must be so. The reason is that our language has remained the same and still seduces us into asking the same questions over and over. As long as there is a verb "to be" which seems to function like "to eat" and "to drink", as long as there are adjectives like "identical", "true", "false", "possible", as long as one talks about a flow of time and an expanse of space, etc., etc., humans will continue to bump up against the same mysterious difficulties, and stare at something no explanation seems able to remove.

               [For clarity, he might have said "... when these things are true" rather than "... as long as".]

[59]             That is, not the one's, say, Ryle saw -- the perplexities here are not to only those induced by philosophy courses, for example.

[60]             See B. pp. 42-45

[61]             John Wisdom ("The Metamorphosis of Metaphysics" in his Paradox and Discovery Oxford: Blackwell, 1965 cited as "P&D" followed by page number: John Wisdom Proof and Explanation (The Virginia Lectures) [ed. S Barker] Washington, DC: University Press of America, 1991, cited as "VL".) provides a helpful account of understanding, employing concrete cases, by contrasting what he calls "the mother's method" of explanation with "the father's method". Asked what a greyhound is, the father replies with a kind of 'definition', that a greyhound is "a dog of such-and-such a type". [Wisdom comments "Short, conclusive, the father's procedure. That's more what one might call a proof." (VL p. 48).] In contrast, the mother says (pointing):

... that's a greyhound, and you remember your uncle's dog, Entry Badge, well that was a greyhound. But now that [she says, pointing to a Borzoi] is not a greyhound, and even that [she says, pointing to a whippet] is not. (P&D p. 69)

           He concludes that the mother:

.... replies with instances of what is and what is not a greyhound or by comparing greyhounds with what they are not, and these two procedures merge into one. (P&D p. 70)

               It seems to me clear, and entirely appropriate, that we learn to understand and to appreciate value (as most things) by "the mother's method" or a variant of it.

[62]             Origin: TS 232 p. 640 (Ms 137 p. 57 [1948]) -- last sentence (ignored here) in handwriting.

[63]             Peter Hacker Wittgenstein: Meaning and Mind (Volume Three of an Analytical Commentary on the Philosophical Investigations), Oxford: Blackwell, 1990 p. 91

[64]             Michael Woods Conditionals Oxford: Clarendon Press, 1997 p. 97 (from Dorothy Edgington's "Commentary").

[65]             T. Nagel Other Minds p. 10


Return to top of page