Friday, 30 September 2011

Wittgenstein’s New Philosophy: A “No Theory” Theory?

According to Wittgenstein, the bewitching allure of metaphysics has proved empty (§§89-108). It has led to a collection of a priori theories concerning how the world must be, but such theories are founded upon illusions brought about by misunderstandings of, and misrepresentations of, our forms of expression. The result has not been to create a body of philosophical knowledge but to generate a series of seemingly intractable problems – problems that trap us between what must be the case (according to the theories) and what is the case, so far as our everyday lives are concerned.

Thus philosophy is left with the task of revealing the illusionary nature of these problems. They are not to be solved but dissolved: “The results of philosophy are the discovery of some piece of plain nonsense and the bumps that the understanding has got by running up against the limits of language” (§119). Obviously, the proper technique for achieving this cannot itself be theoretical – that would just repeat the process that caused the problem in the first place. Instead, Wittgenstein proposes an altogether different method:

[…] we may not advance any kind of theory. There must not be anything hypothetical in our considerations. All explanation must disappear, and description alone must take its place.
Philosophical Investigations §109

Such description will be tailored according to the problem at hand; we will be “marshalling recollections for a particular purpose” (§127). And the results will not be new facts about the world but a clearer understanding of what we already know, in the same way that a map might give us a clearer understanding of our home town.

This, basically, is Wittgenstein’s New Deal for Philosophy, as set out in §§109-133. It was a radical and controversial idea when the Investigations was published in 1953, and it remains so today. It’s an idea that even some who admire his philosophy find a step too far (John Searle, for instance). And although many of the arguments in the Investigations have been recognised as hugely important, I think it’s fair to say that the philosophical community as a whole has declined to take Wittgenstein up on his offer.

I’ll discuss their reluctance in a future post. For now, however, I want to address an obvious question raised by the rejection of theory, namely: how the hell can you do philosophy without theories?

It seems that such an approach would make it impossible to state any conclusions, because that (surely) involves stating a theory. When, for example, Wittgenstein says “the meaning of a word is its use in the language” (§43), isn’t that a theory? Indeed, isn’t it a counter-theory to “A name means an object. The object is its meaning” (TLP 3.203)? Might we not go further and say that any general statement (eg, “cats are more intelligent than dogs”) makes a theoretical claim about the world? And if that’s true, it’s hard to see how theories can be avoided.

The first thing to point out is that actually not every general statement can meaningfully be called theoretical. A theory operates in an area of contention – there has to be something that’s up for grabs. So “cats are more intelligent than dogs” could be considered theoretical because it is by no means obviously true. There is something to be tested here. On the other hand, “giraffes have longer necks than swans” is not (for us) theoretical. It merely states an established truth. Likewise, it is not theoretical to point out that “losing my mind” is unlike “losing my hat”. We would all admit that I cannot look for my mind or offer a reward for the person who finds it; someone can’t have taken my mind because he mistook it for his own, and so on. Such statements are not contentious and therefore not theoretical.

There is, however, an important distinction between the above two examples. The non-theoretical status of the giraffe/swan statement is a posteriori; it is a contingent truth that we have established about the world. We could imagine a situation where this truth was still up for grabs, and there the statement would be theoretical. We would need more facts about giraffes or swans to settle the issue.

The mind/hat statement, on the other hand, works differently. It does not remind us of an established empirical truth, but a grammatical one. We didn’t discover it by encountering hats and minds; we learnt it when we learnt our language, and the truth it expresses partly constitutes what minds and hats are. The only way it could be “up for grabs” would be if someone didn’t know how we use the words “mind” or “hat”. And that person would require linguistic instruction rather than new empirical facts. In this sense, the mind/hat statement is a priori. I don't mean that it's "true for all possible worlds" or anything like that, but that it is about the concepts we need to master in order to make empirical statements such as “your hat is bigger than my hat” or “your mind is sharper than my mind” (perhaps it would be better to call it a grammatical truth rather than an a priori one). And part of mastering such concepts involves realising that it’s nonsense to say “I’ve lost my mind, stop what you’re doing and help me look for it”.

Since these grammatical observations are not theoretical, it is possible to derive general statements from them which are also not theoretical. Such statements are summaries. They do not rely on deduction and do not express hypotheses. They can be verified, not by experiment, but simply by looking and seeing whether they correctly reflect the established facts.

This highlights the important distinction between TLP 3.203 and §43. “A name means an object” is a dogmatic expression of an a priori theory – it must be so, given the requirements of the Picture Theory of Meaning. By contrast, “The meaning of a word is its use in the language” asserts what we will all admit to be true if we look carefully at our forms of expression and, in particular, at the way the word “meaning” is generally explained (be warned, however, that this is a highly controversial claim). Wittgenstein does not say it must be so (indeed, he explicitly says it is not always so), merely that – most of the time – it is so. And it is precisely this sort of statement he has in mind when he says, “If someone were to advance theses in philosophy, it would never be possible to debate them, because everyone would agree to them” (§129).

Here we might admit that we can draw non-theoretical conclusions, yet still wonder why theories must be ruled out altogether. After all, couldn’t we use our new-found linguistic insights to construct better theories? No. The suggestion misunderstands the flaw at the heart of philosophical theories.

A philosophical theory looks like it’s making an a priori claim about how the world must be, but actually its a priori nature comes from its use of conceptual rules. Determinism, for example, flows from reflections on the fact that every event has a cause. But “every event has a cause” is not an empirical fact (unlike, say “every child has a penny”); it is a conceptual precondition for certain types of activity – scientific investigations, for example. It provides a rule legitimising these activities and guarantees that, given an event, you can always ask “what caused this?” That is the nature of its “must”. It does not, however, guarantee anything about what is or isn’t the case.

It is illicit, therefore, to move from “every event has a cause” to “free actions do not exist” because that is making an unsanctioned existential claim about the world. What you can do, however, is examine the conceptual underpinning of “free will” together with the concept of causation to see how they relate to each other. This involves no illicit move because everything remains at an a priori level. It does not, however, save the theory by getting rid of its mistakes – it gets rid of the theory. That is because we are no longer deducing what must be the case, but consulting the rule-book to see how things are. It’s as if we had a theory that castling is impossible in chess because the king can only move one square at a time. Then we look up the rules and see that under certain circumstances it is perfectly legitimate. We cannot now seriously have a theory that the rules of castling don’t exist, nor that castling is invalidated by the game’s other rules. It is part of the game, and that’s that.

Wittgenstein’s rejection of philosophical theories is not based on the notion that they’re unlikely to yield results. His argument is that they cannot yield results because they are conceptually incoherent. They attempt to deduce a priori truths about the world based on rules that provide no justification for such deductions. Dispensing with them is not (pace Searle) an unnecessary piece of philosophical extravagance; the need flows directly from Wittgenstein’s ideas about meaning as use, language-games, the nature of rules, and family resemblance concepts. If those are accepted then ditching theory is mandatory, not optional.

Finally (and at the risk of stating the bleeding obvious) it is also not itself a theory. It is a proposal offered as the only way of avoiding the endlessly repeated mistakes of the past and providing us with a way to see the world aright when we become entangled in conceptual confusion. The price to pay consists in renouncing philosophy as a heroic endeavour – one where the next great mind might finally hit upon the correct theory and explain things to everyone’s satisfaction. (Two and a half thousand years and still waiting….) Instead it would be a more humble matter of “marshalling recollections for a particular purpose”. It would require patience and skill rather than god-like genius.

For many that price is too high. 


  1. This is nicely put, and I certainly agree with most of it. But what would you say to someone who objected to this: “The meaning of a word is its use in the language” asserts what we will all admit to be true if we look carefully at our forms of expression and, in particular, at the way the word “meaning” is generally explained.?

    They might object, I think, in two ways, which are possibly related. The first is that, as you say, careful looking is required here in a way that it doesn't seem to be in the mind/hat case. That one's just obvious. How important is this difference? Does it just show, as you also say, that patience and skill are needed?

    Secondly, perhaps because people don't always have this patience and skill, isn't it a fact that we don't all admit it to be true that the meaning of a word is usually its use in the language? Many philosophers seem to deny this, after all. I don't mean that Wittgensteinians can't respond in any way to these philosophers, but I'm not sure what the right response is.

  2. There's certainly something a bit disingenuous about the claim that we'd all readily admit to "meaning is use" and I probably should've flagged that up rather than just towing the "party line", so to speak. (It's linked in my mind to W's rather cheeky habit of starting a sentence "Of course,..." and then saying something extremely controversial and/or baffling.) Indeed, I suspect most people would readily admit to TLP 3.203 but would be a bit dubious of PI 43.

    However, I don't think this hamstrings W for two reasons.

    First, it is an example of the general difficulty of carrying out W's descriptive method. Being a competent language-user doesn't require competence in surveying the uses of words - noting subtle distinctions and similarities, etc. Most people couldn't tell you the relationship between "know" and "understand" in different areas of language even though they correctly make the necessary adjustments day-in, day-out. They need to be reminded of what they already do and even then it is far from easy because habitual ways of looking at language - ingrained by years of practice - reassert themselves at every opportunity. "I see it, I see it - damn! It's gone!" is one of the characteristic experiences of attempting the descriptive method. But if they are taken through it step by step then (it is hoped) they WILL agree. If they don't then you take them through it again.

    And (point two) the important thing is that their agreement is not based on new facts. You are not trying to reform their language use but getting them to see how they already use it. So the process is still descriptive rather than theoretical.

    As for PHILOSOPHERS who disagree, I would say (sweepingly) that they haven't been able to shake off a theoretical approach to language ("rigid designators" etc). I'm reluctant to engage too fully with this at the moment because I'm more intent on getting W's position straight in my own mind. But I hope in due course to turn my attention to Quine, Davidson, Kripke et al.

  3. Yes, that sounds right. Makes it hard to persuade philosophers, but perhaps that's their fault. And it doesn't mean it's impossible. I look forward to seeing what you have to say about Quine and co.

  4. It sounds right, but I am aware that (so far) I've just been trotting out the Wittgensteinian line as best I can. That's the value of this blog and its major drawback, both to me personally and to anyone else who reads it.

    In fact it should have huge quotation-marks round it and the subtitle "If Wittgenstein was right on all points this is how he would sound to me".

    I can't see how far he was actually right until I fully confront his critics. But one thing at a time!

  5. Hi Phillip - how can you even raise the possibility that Wittgenstein was not right on everything :-) I think this is a difficult area to get right and unusually I had some hesitations about some of the things you say. I don't find the theory/non-theory contrast that helpful and I am a bit worried about suggesting that some propositions are not theories for us but might be theories for other people. I also worry a bit about suggesting that grammatical points are a priori. I am pretty confident that your use of the phrase is not confusing you, but it certainly seems to me that it might confuse others.
    So what alternative is there - I have always gone for a grammatical/substantive contrast where grammatical propositions set the rules of language and substantive propositions say something about the world. Of course, when you push on this distinction things do get pretty difficult and although I am sure Wittgenstein would have been capable to getting to the bottom of this issue, I am not sure he ever really did (and therefore of course neither have I).
    I have some sympathy with DR's point and it does seem a bit disingenuous to say that what Wittgenstein says is very different from what other philosophers have said because no one could disagree with it and then not give much weight to the fact that lots of people seem to disagree with it!
    Again the way I would try to avoid that problem would be simply to say that while it would certainly be a criticism of a traditional philosopher if one reacted to him by saying: that does not really give me any new insights into the world or increase my wisdom, it would not be a criticism of Wittgenstein since explicitly he is not trying to tell you something new about the world only remove your confusions.
    So a lot of words from me and probably stuff you are very familiar with, but its what occurred to me after reading your post :-)

  6. Hi Paul. To be honest, I'm not entirely happy with "a priori" myself. I was struggling to make a distinction and wasn't totally sure of the way to put it. That's why I used the qualifier "In this sense". I was really contrasting "a priori" with "empirical", but is that a legitimate classification? I've been away from academic philosophy for a long while and my terminology gets a bit sloppy sometimes.

    However, by "a priori" I certainly didn't mean "must be true for all possible worlds" or anything like that. I'd probably better point that out in the post.

    RE: the move from theory to non-theory in empirical assertions, it does seem to me to be what happens. Is "The Earth goes round the Sun" a theory for us? 400 years ago it was. Now I'd say not. However, I wouldn't die in a ditch over this because the important distinction is between empirical/grammatical assertions, not really between empirical theories/facts.

    Many thanks for your (and DR's) comments, btw. I tend to sound much more confident about all this than I really am and so it's helpful to get counter-suggestions. Really I'm just trying to figure it all out, even if I write as if I've got it down cold!

  7. I wanted to say hello also! I've enjoyed reading several posts on this blog.

    On the theory/non-theory debate, I think, it's not so much about the content in what you say - but rather, what claim you want to make by saying it. If you claim that giraffes really have longer necks than swans, then that is a theory. However if you just speak of it as an assumption or a 'reminder', it is not a theory.
    Similar "The earth goes round the sun" is a theory if you claim that it's true, however no theory when you just assume it. (I think Wittgenstein makes similar point when he speaks about Moore's hand in On Certainty)

    The most important thing is to keep track of the empirical/logical distinction, I think. The philosophy of Wittgenstein circles very much around this.


    1. Hi, and thanks!

      I agree with your point. There's also some interesting stuff on theories/experiments in W's Remarks on the Foundation of Mathematics. In fact, I've been thinking quite a bit recently about what is or isn't a theory, because I've been considering the implications of W's approach for religion and reading The God Delusion. I intend to blog about that at some stage - either here or elsewhere.

  8. "But “every event has a cause” is not an empirical fact (unlike, say “every child has a penny”)..."
    Could I ask you to you flesh this out just a wee bit, Philip? It baffles me. If I am a witness to a crime, is it not the empirical facts that prove decisive? What am I missing?

  9. Hi Jim. I think it's like this: if someone says "Every child has a penny" this is an assertion about how things are. It can be true or false and is, at least in theory, capable of being established. We might, for example, round up all the children and search them for pennies.

    But suppose I say "Every event has a cause" and someone replies "Prove it" - what on earth am I to do now? I might show that this event and that event had causes, but what warrants the move from "these events had causes" to "every even has a cause"? On the other hand, how might I refute it? How might I prove that this event has no cause (as opposed to a cause that I don't know about)?

    "Every event has a cause" does not express a fact so much as an attitude towards the world. More exactly, it is a procedural rule we adopt when investigating why things happened. And it is warranted because it has proved its worth - so much so that we tend to take it as an obvious truth. But it is still a rule for all that, and therefore neither true nor false.

    It is not, however, a grammatical rule. It is not like "All bachelors are unmarried". To say "There could not be an uncaused event" is not to express a logical rule, but to affirm one's allegiance to the procedural rule. We could well say "This event has no cause" without distorting the concepts "event" and "cause" - indeed, something remarkably like this happens in quantum physics, where probability takes over from cause as the basic explanation of why things happen as they do.

  10. Crikey Philip, that's really illuminating. It clears up a whole bunch of loose ends in one go. Thanks so much.

    1. Many thanks, Jim. That last point about "Every event has a cause" not being grammatical is not something I make clear in the post itself. In fact, I think I only grasped that later. If I ever get round to revising my material (and I plan to) I ought to make it clear because it has important implications.

      And speaking more generally, I would add that it's a very difficult point to keep hold of. There are some sections in Zettel (ie #608-613) touching on this topic which are truly hair-raising when you first read them. They just seem mad. But I would gloss them as follows: think of the rule of causality as a kind of crude but extremely useful tool. Why should we expect, however, that nature *must* function in this manner at all levels and in all its manifestations? Given what we already know about the strangeness and astonishing complexity of the world's structure, wouldn't it be an almost absurd coincidence if the causal rule turned out to be the first and last word on the matter?

      To put it less poetically: suppose at some stage physics reaches a point where it cannot find any clue as to where to go next. What should we say about what lies beyond that point? That there must be causes which are beyond our finite understanding? Or that here the very notion of "cause" peters out? Either choice seems to me more like an act of faith than a rational theory.

  11. Fascinating. I really like what you say about “Every event has a cause” being a procedural rule we adopt towards the world. But perhaps it goes much deeper than allegiance (at least, if by “allegiance” we mean something that we can sometimes abandon without incurring dire consequences). It’s implicit in our every action, even the mistaken ones.

    "Giving grounds, however, justifying the evidence, comes to an end; - but the end is not certain propositions' striking us immediately as true, i.e. it is not a kind of seeing on our part; it is our acting, which lies at the bottom of the language-game." Wittgenstein §204

  12. Oops, it's from "On Certainty" of course.

  13. I'd certainly agree that we couldn't live in anything like the way we do without the concept of causation. However, that's not the same as "every event has a cause". Indeed, there is a hugely important area of our lives where causation takes a back seat and perhaps disappears altogether: the sphere of human action. Here our attitude is precisely of a cause (the human being) which is not itself caused. And this attitude (which Wittgenstein calls "an attitude towards a soul") is an intrinsic part of the framework within which language takes place.