Friday, May 25, 2012

Wittgenstein on clarity

In the Tractatus Wittgenstein says the following things about clarity and elucidation:
One could put the whole sense of the book perhaps in these words: What can be said at all, can be said clearly and whereof one cannot speak, thereof one must be silent.

3.251 A proposition expresses what it expresses in a definite, clearly specifiable way: a proposition is articulated.

3.263 The meanings of primitive signs can be explained through elucidations.  Elucidations [Erläuterungen] are propositions which contain primitive signs.  They can thus only be understood if one is already acquainted with the meanings of these signs.
4.112 The end of philosophy is the logical clarification of thoughts.
Philosophy is not a subject but an activity.
A philosophical work consists essentially of elucidations.
The result of philosophy is not "philosophical propositions" but the clarification of propositions.
Philosophy should make clear and distinct thoughts that, without it, are, as it were, unclear and indistinct.
 4.115 It [i.e. philosophy] will refer to the unsayable in that it presents clearly the sayable.
4.116 Everything that can be thought at all can be thought clearly. Everything that can be said can be said clearly
5.45 If there are primitive signs of logic then a correct logic must make clear their position with regard to one another and justify their being.  The construction of logic out of its primitive signs must be made clear.
5.452 The introduction of a new device in the symbolism of logic must always be an important event.  No new device may be introduced into logic – with, so to speak, a wholly innocent face – in brackets or in a footnote.
(Thus in the Principia Mathematica of Russell and Whitehead there appear definitions and basic laws in words. Why suddenly words here?  This would need a justification. This is missing and must be missing, since the procedure is actually forbidden.)
If, however, the introduction of a new device has proved necessary in one place, then one must ask oneself straightaway: Where must this device now always be used?  Its place in logic must now be made clear.
6.54 My propositions elucidate by whoever understands me perceiving them in the end as nonsensical, when through them – upon them – over them, he has climbed out.  (He must, so to speak, throw away the ladder after he has climbed out upon it.)
He must overcome these propositions, then he sees the world rightly. 
The words "It is clear" and "it is clearly" occur often in the book, about 18 or 19 times in my translation (which aims to be very literal). At first these are references to what is clear in logic, as in 3.251 and 4.115 quoted above, having to do with the clear specification or presentation of what can be said. But later we get into ethics and some strange claims about what is clear, e.g. "6.421 It is clear, that ethics cannot be articulated." To whom is that clear? It isn't clear just on its own but, if at all, then to someone. And not just anyone, but an ethical being.

The inescapability of our humanity, both our physical being, which is affected by sounds when we hear others speak, and our subjective being (is that the word?), which can guess and pick up hints, is made clear enough by Frege (especially by Beaney's Frege Reader and its helpful index). It also comes up in Schopenhauer when he writes that, "The aim of realism is just the object without subject; but it is impossible even to conceive such an object clearly" (The World as Will and Representation, Volume II, p. 12). There is no escaping subjectivity, no as it were absolutely objective communication, no understanding a speaker's sentences without understanding the speaker (although presumably Wittgenstein thinks we can understand an author without understanding his sentences, as indeed does seem to be the case).

Frege and Russell 'cheat' (see, e.g. 5.452 above) because they have to. It is not possible to do what they are trying to do without 'cheating', as Frege, at least, recognizes. And that means there is something wrong with what they are trying to do. It involves a false view of clarity. There is not the clearly sayable and the having-to-be-guessed-at. Or rather, since, after all, some things are clearly sayable and others do have to be guessed at, the distinction between the sayable and the un-sayable can be overstated.

Does this really matter? Wittgenstein seems to think it makes a huge difference to philosophy, to how we conceive of and do philosophy. "Philosophy is not a subject but an activity. A philosophical work consists essentially of elucidations." I take this to mean that philosophy should consist of elucidations, and that these elucidations are acts, not clarificatory statements. The clarification of propositions is not the production of clear propositions. Every proposition that means something, i,.e. every proposition that actually is a proposition, has a clear meaning. It just might not be clear to this or that person. Philosophy, Wittgenstein appears to be saying, is the activity of helping people understand this meaning. Clarity and the lack thereof are psychological matters.

But what about 3.263? It says that: "The meanings of primitive signs can be explained through elucidations.  Elucidations are propositions which contain primitive signs.  They can thus only be understood if one is already acquainted with the meanings of these signs." So elucidations are propositions, after all, not acts of elucidation. But this proposition is sheer nonsense, isn't it? It reads like a joke. The meaning of primitive signs can be explained through elucidations, and these elucidations can only be understood if one already is acquainted with the meanings of these primitive signs. So the point of an elucidation is to explain something in a way that can only be understood by people already acquainted with what is to be explained.

Elsewhere I commented as follows on this passage:
What the…?! This sounds circular and pointless. Don’t know the meaning of a primitive sign? An “elucidation” will help. But you will only understand it if you already know the meaning of the relevant primitive sign! So either knowledge of meaning is not the same thing as understanding when it comes to signs, which seems unlikely (but who knows?). Or explanation of meaning is quite impossible (in the terms presented by the Tractatus up to now). See p. 44 and pp. 49-50 of Joan Weiner’s essay in Future Pasts.

Anscombe (p. 26) suggests that this passage, along with 3.261, provides the best evidence for thinking that the elementary propositions of the Tractatus are simple observation statements, such as “This is a red patch.” Names and only names are primitive signs. Logical signs, as he indicates elsewhere, are not primitive signs. But (see p. 27) what elucidates a name need not be an elementary proposition. And from 6.3751 it follows directly that “This is a red patch” cannot be an elementary proposition.Anscombe concludes that elementary propositions are not simple observation statements, and that this explains why Wittgenstein did not refer to observation in connection with them. What they are he cannot say, but they must exist. See, for instance, 5.5562, 3.23, 2.021, 2.0211, and 4.221.

See also 5.526.
(I had forgotten that I had read Joan Weiner's essay before.) N.N. commented:
Hacker argues (persuasively, it seems to me) that elucidations are ostensive definitions 'seen through a glass darkly.' That is, they are ostensive definitions misconstrued as bipolar propositions (see 'Frege and Wittgenstein on Elucidations,' Mind, Oct., 1975).
The 'smoking gun' in favor of his interpretation is a remark that Wittgenstein made to Waismann in 1932: 'In the Tractatus logical analysis and ostensive definition were unclear to me.' That is, according to Wittgenstein, ostensive definition is addressed in the Tractatus. And even Kenny, who denies that elucidations are ostensive definitions, concedes that the only passage Wittgenstein could be referring to is 3.263 (see Anthony Kenny, 'The Ghost of the Tractatus,' in Legacy of Wittgenstein).  
It isn't clear to me that Wittgenstein means in this remark that the Tractatus says something about ostensive definitions. Perhaps he meant that he had overlooked them. And we only have Waismann's report that Wittgenstein said this, so it's secondhand evidence. I wouldn't put too much weight on the smoking gun therefore.

Nevertheless, Hacker's interpretation has at least a certain plausibility. If we read 3.263 as saying that the meanings of primitive signs can be explained by means of sentences that use those signs to refer to objects with which one is familiar, then it sounds about right. But Anscombe's argument seems good. If a primitive sign is the name of an object and we cannot give any examples of objects then surely an elucidation containing primitive signs is not an ostensive definition. We can give lots of examples of them.

Whatever we make of 3.263, it seems clear to me that the elucidations mentioned in 6.54 and 4.112 are different in kind than those described in 3.263. The latter sound much more like the kind of thing Frege struggles with. The "elucidations" of 3.263 would be philosophical propositions, wouldn't they? But these are rejected in 4.112.

It's still a little hard to see how this is supposed to help anyone see the world rightly. To see Frege's work rightly perhaps, and to see philosophy rightly perhaps also. But the world? For that I think we would have to look more at the remarks on ethics towards the end of the Tractatus


  1. There is an interesting (and possibly relevant) discussion of "elucidation" in The Big Typescript (p. 30)

    So if to explain the word "lilac" I point to a patch and say "This patch is lilac", can this explanation then work in two ways – on the one hand as a definition that uses the patch as a sign, and on the other as an elucidation [Erläuterung]? And how is the latter possible? I would have to assume that the other person is telling the truth and seeing the same thing I'm seeing. A case that really occurs is something like this: In my presence A tells B that a certain object is lilac. I hear this, and have also seen the object, and think to myself: "Now I know for sure what 'lilac' means". That is, I have extracted an explanation of the word from that description. I could say: If what A told B is the truth, then the word "lilac" must have this meaning.

    On Wittgenstein's view of an "elucidation" (in 1932), it is a proposition from which one can extract an (ostensive) definition. No mention is made of the Tractatus, but this could be made to fit 3.263. The meaning of a simple sign is an object. If someone is to extract a ostensive explanation of a name out of a proposition containing that name, then he must be acquainted with the object, i.e., he must see what is being pointed to.

    Perhaps that reading of 3.263 is a stretch, but it seems preferable to "What the . . .?!"

  2. Thanks, Brad. This makes a lot of sense, but aren't there other reasons for thinking that propositions like "This patch is lilac" can't be what TLP 3.263 is talking about? Otherwise 'lilac' would be the name of an object, wouldn't it? And then it would be strange that Wittgenstein couldn't name any objects. But perhaps I'm misremembering that.

    As for whether your suggested interpretation is better than "What the...?!", it would seem to be, yes. But if it doesn't work then it isn't better. (I'm not saying that it doesn't work, but I'm not convinced that it does.) And then there's 6.54, of course. If the author of 3.263 tells us that it is nonsense, then an interpretation that makes sense of it is not necessarily preferable to one that does not. Which is not to say we shouldn't try to make sense of it. But when the stretching gets too painful maybe we should quit.

  3. I can't comment directly on this, but I wonder whether the apparent circularity problem in making sense of what an elucidation is (or how it works)--and the reported remark that logical analysis and ostensive definition were unclear to him when he wrote TLP--is related to the puzzles that LW tries to work out in the rule-following remarks of PI...

    The idea would be that elucidations "acquaint" one with the meanings of signs (the proposed meaning, or how we are to get on with this sign in doing/saying something), but that the elucidation is only later (?) understood, insofar as one gets the hang of using the sign in the proposed way? (That the proof that one has understood the elucidation is that one then proceeds to use the sign/term/etc. in the right/proposed way?) I'm just thinking out loud, and loosely, as it were, so I don't know if this is helpful...

  4. Thanks, Matt. That sounds right as a way to connect thoughts about ostensive definitions with rule-following. And if TLP 3.263 is best read as being about ostensive definitions then the whole thing may well be right. I'm just not sure about 3.263.

    Is it possible that Brad and I are both right about it? That, as it were, what Wittgenstein means is quite right (and is the kind of thing that you and Brad are suggesting), but that what he actually says is nonsense? I think really to settle the matter you would have to go through the whole book and see whether Brad's interpretation holds up or not. It might.

  5. Duncan,

    It's my own view that the other reasons for thinking that a proposition/ostensive definition like "This is lilac" can't be an elucidation can be dealt with. I think, for example, that Anscombe's famous objection fails (this is the topic of an essay that I have under review). Also, while Wittgenstein doesn't give any examples, this (it seems to me) has more to do with the fact that he had not settled on what exactly an elementary color is, and not because he did not think that some colors are elementary.

    Concerning 6.54, I think that a different sort of elucidation is in question (as you suggest). Conant has some interesting comments on this in a footnote to The Method of the Tractatus:

    It is only this species of elucidation that Frege thinks compels us to traffic in nonsense. Within the broader genus of elucidation, elucidations will generally take the form of perfectly meaningful propositions (such as, e.g., elucidations of geometrically primitive terms). It is worth noting, however, that a parallel distinction between a generic and a specific notion of elucidation must also be drawn if one seeks to understand the different occurrences of the term Erläuterung in the Tractatus. In 3.263 what is at issue is a species of the genus pertaining to the elucidation of primitive signs (which I will not explore further here, other than to remark that perfectly meaningful propositions can serve as elucidations of this sort), whereas an understanding of 6.54 is unattainable apart from an understanding of what is peculiar to that species of the genus that aims to elucidate "philosophic matters" (and which proceeds through the employment of Sätze that the reader is to recognize as Unsinn). [pp. 436-7, n. 48]

    It would be interesting to hear what Conant has to say about the species in 3.263, as well as the generic characteristics of elucidation. Concerning the latter, elucidations seem to be a use of signs (as propositions, or as nonsense) that allow us to catch on to something else (definitions, "philosophic matters" respectively). That's a first guess.

  6. Thanks, Brad. Your "first guess" sounds at least about right to me. I'd like to read your essay--I hope it gets published.