L1.11 Reflections
“No, no. I just speak as I think. That is why it does not always work.”
pTravis is disappointed and says he would have liked to have read about all the lessons he hadn’t heard. The next morning he finds in his pigeon hole a tape and a tiny digital recorder. There is a note: Old habits!
A week after that and pTravis and Masongill have listened to all the recordings in their den, often joined by Lottie, who sometimes seems embarrassed and is sometimes curious, sometimes proud of what she said at the time.
For some reason a little shy, pTravis sends Hipparchus a note which reads. ‘Could we meet to have an another discussion of concepts. ’Gill and I would like the freedom to ask some stupid questions. We don’t really expect serious answers to them.’
A week after that, they all congregate one evening in the study. Lottie may be present only because there was talk of hot chocolate, chocolate biscuits and tots of calvados, though, after a small sip, she gives her calvados to pTravis in return for his biscuit.
Conceptual fit, nonsense and the ‘resolute reading’ of Wittgenstein
pTravis begins:
“There is a bit on the tape that made me think how little I really understand anything about concepts. Lottie makes an objection to Evans’s Generality Constraint, saying that not every predicate applies to every ‘name’, and you agree but then you say that Evans did qualify it with a footnote: ‘With a proviso about the categorical appropriateness of the predicates to the subjects’ (Evans 1982: 101 fn).
My question is either very simple, or very stupid, or it is both. But what underpins the ‘categorical appropriateness of the predicates to the subjects’? What makes one thing fit and another not? What does fitting – admittedly my choice of word – even mean here?”
“Go on. What ideas do you have?”
“Well here is a very silly analogy. There’s a child’s toy that is a box with a lid with various shaped holes and then some wooden blocks. The blocks only fit some holes. The square peg does not fit in the round hole, for example. But both cylinders and spheres do fit in the round hole. Etc. That notion of fit might be what appropriateness is for concepts. But not really. Concepts are not literally like blocks of wood. So why does one concept fit one ‘slot’ in a thought but not another?”
Hipparchus pauses. “I do not have a simple answer to this. I am going to try to balance two ideas. You will need to think what you should think. First, we do think that our concepts have some independence from us. I asked in the last lecture whether, once mathematical systems were established by convention, one needed further conventions to establish every implication of the first set of conventions. To think so would destroy the objectivity of systems of thought like maths and logic. It would also destroy the objectivity of judgements about whether, for example, the next object I see when I open this cupboard is red or not red. For this to be objective, the word ‘red’ needs to set a standard that is independent of my whim. In other words, we do think that the conceptual order has a kind of objectivity or independence. Even if, in some sense, it is grounded in shared judgements, as Wittgenstein says in §242, still what is grounded is something for which a Platonic description seems apt. So that is one thought. Applying that thought to your question would suggest the following answer: our conceptual system simply rules out the idea that ‘green’ can be applied to ‘ideas’. It is a square peg that does not fit in a round hole.
But now, second, let me suggest a contrary thought. This stems from what is called a ‘resolute reading’ of Wittgenstein, the key idea of which is this. To say that a sentence is nonsense is not to say that its meaning or sense is somehow self-contradictory or impossible. It is rather, simply, that we have given no sense, no meaning and no use to that combination of signs, where as Wittgenstein says: ‘A sign is what is perceptible in a symbol’ (TLP 3.32). Nonsense is a lack of sense – gibberish, if you like – not the possession of an impossible sense. Or, in another slogan, the conceptual realm, the realm of sense, has limits but not limitations.
Applying this thought to your question suggests the following answer. It is not that the concept of ‘green’ has a shape that makes it unable to fit into the slot ‘ideas’ offers. It is just that we have not worked out what use we could make of ‘green ideas’ and it is thus nonsense or gibberish. The sign, the shape or squiggle ‘green’ only becomes a symbol meaning green when deployed in the right context – both the narrow linguistic context of a sentence and but also the right context of application or use – so it cannot be that there are impossible meanings. It is rather that, given a familiar sign, we may attempt to invest it with meaning in a context which is insufficient to confer meaning on it. It is not that we have a sentence with an impossible overall meaning but rather that we have not yet reached as far as a meaningful sentence or use of a sentence or a symbol (depending on what we call a sentence: sign or symbol).
How do these two contradictory thoughts strike you?”
pTravis looks to Masongill and Lottie but neither of them seems to want to say anything. Sometimes philosophical ideas need to sit a while and pTravis has the advantage that he has already done some philosophy before. (Sometimes, some philosophical ideas are best encountered newly. Familiarity with scepticism, for example, tends to corrupt the mind.)
“I am quite pleased that you have put both sides like that because, frankly, I don’t know what to think. If I follow, the resolute reading of Wittgenstein is minimalist about nonsense. Nonsense is just no sense. That seems plausible. But we do not seem to need to decide every case afresh. Concepts seem to have their own momentum. It is no surprise that I do not know what to do with ‘green ideas’. It is obvious that I do not. And, contrary, I think, to the idea that nothing has been said at all, I think that I can say something about why not. ‘Green’ speaks of being green. That is what the sign generally symbolises in English. And then colours are properties of surfaces or volumes. Ideas have neither so could not be coloured. But for me to be able to offer these natural thoughts, I must be able to do something with the phrase ‘green ideas’ simply on the basis of the meanings these words are ascribed in dictionaries. The phrase ‘green ideas’ does not seem to be as nonsensical as a random collection of individual signs from the top line of a qwerty keyboard. So if the resolute reading is right, it needs to be understood through the prism of some conceptual action at a distance and some knowledge of conventional English definitions. That some combinations of symbols have not been given a sense in part follows from their individual rules of use. For ‘green’: to speak of being green. And thus, unlike solving a crossword clue, I know immediately that there’s a problem with the phrase ‘idées vertes’ because in one sense, I know what it means – it means green ideas – even if, in another, I do not know what to do with that.”
“That’s a ‘deep thought’” says Lottie with a cheeky grin, kicking Masongill pointedly.
They lapse into silence.
Is ‘gold’ a name or a predicate?
“Here’s another question!” Masongill volunteers to break what had become a long pause.
“When you were describing Kripke’s non-descriptive accounts of names, you mentioned Putnam’s essentialist account of natural kind types such as gold. ‘Gold’ picks out whatever substance shares the essential structure of some original blob of gold in a baptism ceremony. That sounded like ‘Napoleon’ naming whichever individual was baptised with that name and the name spreads down through time by causal connections, not via a cluster of descriptions. But ‘gold’ isn’t always used as name, is it? If I say that this coin is gold, or that it is a gold coin, or that it is made of gold, that’s a predicate isn’t it? In ‘This coin is gold’, the ‘is’ is not the ‘is’ of identity but of predication. You could not switch it around and say that gold is this coin. So what is going on?”
“You really need someone who specialises in formal semantics to answer this. But, yes, ‘is gold’ is a predicate. As I think I mentioned, even if ‘gold’ is a name, it does not name its instances as these might not have existed. If it is a name, it names a structure to which, Putnam and Kripke argue, it links essentially. There is no possible world in which there is gold but not that structure nor that structure but not gold. But there are worlds in which there is no gold and others where there are different amounts of it, where some actual instances do not exist. The predicate ‘is gold’ picks out, though it does not name, instances of gold ie whatever has the right structure. It does not name the instances because their non-existence would not threaten its content as a concept. The worst that would happen is falsity. If I am right about this – and I admit that I may not be – the function of the word varies depending on how it is articulating conceptual structure in related thoughts.”
Truth and meaning
pTravis waits for a respectful pause and then says:
“I’d like to ask about the relation of truth and meaning. You said that in the task of calculating truth conditions for an object language, Tarski had allowed himself to assume a translation of the object language into the metalanguage. He had assumed meaning, as it were, to elucidate truth in some sense. As I understand, he didn’t really attempt to say what truth, the general concept, is, whether a relation to a fact or whatever, but he showed that there was nothing mysterious in how it attached to object language sentences of whatever complexity.”
“Davidson says that Tarski ‘defined the class of true sentences by giving the extension of the truth predicate, but he did not give the meaning’ (Davidson 1990: 294)” interrupts Hipparchus.
“Davidson, by contrast, aimed to shed light on the very idea of one sentence interpreting another. Light is shed on that by setting out a systematic theory of meaning for an object language re-using Tarski’s formal machinery of satisfaction etc. Not that Davidson actually wrote down any such theory. Nor have we discussed satisfaction but I take it that it is just an internal part of the machinery. Tarski is not trying to reduce truth to reference via satisfaction axioms?”
“Only an idiot would think that” affirms Hipparchus. “Though there are some”.
“But he argued that such a theory would be adequate for providing what we might call ‘meanings’ even though his account spurns what he takes to be an obscure nominalisation. It ‘sweeps away the obscure “means that”’! And, because it would be adequate, that sheds light on the very idea of meaning. It is what is set out in such a theory. And McDowell went further by arguing that such a theory, while stating references, would also be, perhaps show, a theory of sense.”
“Yes: that is how it strikes me.”
“So it is tempting to say that Tarski assumes meaning to shed light on truth while Davidson does the opposite.”
“I agree that that seems a tempting shorthand. And Davidson himself says:
In Tarski’s work, T-sentences are taken to be true because the right branch of the biconditional is assumed to be a translation of the sentence truth conditions for which are being given. But we cannot assume in advance that correct translation can be recognized without pre-empting the point of radical interpretation; in empirical applications, we must abandon the assumption. What I propose is to reverse the direction of explanation: assuming translation, Tarski was able to define truth; the present idea is to take truth as basic and to extract an account of translation or interpretation. The advantages, from the point of view of radical interpretation, are obvious. Truth is a single property which attaches, or fails to attach, to utterances, while each utterance bas its own interpretation; and truth is more apt to connect with fairly simple attitudes of speakers. (Davidson 1984: 134 italics added)”
“That leaves me wondering what Davidson thinks of truth. But let me press on a bit. I know from some extra reading I’ve done that the T schema has also been used to shed light on the nature of truth itself, not just the calculation of truth conditions for object language sentences.”
“Go on.”
“In its purest form, the approach I have in mind suggests that truth is nothing substantial at all, no relation of fit between thoughts and facts, for example, because it is, in fact, redundant. Or almost.
Take ‘s’ is T iff p. Tarski assumes, and Davidson earns the right to the claim, that s and p mean the same. So we can say
‘s’ is T iff s.
And the suggestion is that this shows the only role of truth: a predicate that permits mentioning rather than using a sentence. But it also shows that it is redundant because whenever it is predicated of a mentioned sentence we could strip off the quotation marks and simply use the sentence. No mystery needs attach to truth as a strange property. It is simply an internal tool for describing language use. I must say I like this!
It doesn’t quite work because it might be that one has reason to assert that someone spoke the truth without being in a position to ‘disquote’ what was said. But that still suggests that, even in that circumstance, ‘truth’ is merely a device for compendious endorsement.”
“And…?”
“Well, as I said, I like this approach to truth. But it seems I cannot subscribe to both this and Davidson’s account of meaning. I cannot both presuppose truth to shed light on meaning and then use that theory of meaning to eliminate, almost, truth! As Paul Horwich puts it ‘we would be faced with something like a single equation and two unknowns’ (Horwich 1998: 68).”
“Hmm. Yes. You might be interested in this quotation from John McDowell.” Hipparchus leaves through a volume and reads out:
“The basis of the truth-conditional conception of meaning, as I see it, is the following thought: to specify what would be asserted, in the assertoric utterance of a sentence apt for such use, is to specify a condition under which the sentence (as thus uttered) would be true. The truth-conditional conception of meaning embodies a conception of truth that makes that thought truistic. (I am inclined to think it is the only philosophically hygienic conception of truth there is.) The truism captures what is right about the idea that “… is true”, said of a sentence, functions as a device of disquotation, or, more generally, of cancellation of semantic ascent. (McDowell 1998a: 88–9, emphasis added)
My truism might be held to capture the intuition that finds expression in the redundancy theory of truth. (McDowell 1998a: 90)
So it seems that McDowell does want to hold both the views you also find attractive. How can he do this? I think it follows from how he argues that Davidson’s theory is a theory of sense. If you recall, this does not drop out of the truth-theory itself but rather the interpretative constraints placed on it. Not every true truth-theory for a language is interpretative, is a theory of sense. Only the subset that are arrived at through radical interpretation. In other words, his version of a Davidsonian theory of meaning puts no direct explanatory weight on the concept of truth. It ‘claims no particular conceptual illumination from the notion of truth as such’ (ibid: 43). Thus, there is no need for a substantial independent account of truth.”
“This just seems like a cheat to me. The sort of thing you accused my heroine of!” This is Lottie.
“But the difference between Millikan and McDowell is that Millikan wants to reduce meaning. She thinks it is a purely biological notion. Whereas McDowell, and Davidson too, seek to illuminate these notions but only from within a view of the world that includes meaning and content. It is not a picture of that realm as from outside it. Instead, akin to the articulation of thoughts into concepts construed as abstractions from each whole thought, it is an articulation of meaning, content and truth into merely notionally distinct elements. Each has to be understood in relation to the others. No reduction is possible or attempted.”
“How about Davidson?” asks pTravis. “After all, McDowell offers a particular spin on what Davidson is doing. Does Davidson agree?”
“Well Davidson explicitly rejects an analysis of the meaning or the concept of truth into something more basic. He says ‘Truth is one of the clearest and most basic concepts we have, so it is fruitless to dream of eliminating it in favour of something simpler or more fundamental.’ (Davidson 1990: 314) But he also rejects disquotational approaches
Disquotation cannot, however, pretend to give a complete account of the concept of truth, since it works only in the special case where the metalanguage contains the object language. But neither object language nor metalanguage can contain its own truth predicate. In other words, the very concept we want to explain is explicitly excluded from expression in any consistent language for which disquotation works. To put this another way: if we want to know under what conditions a sentence containing a truth predicate is true, we cannot use that predicate in the disquotational mode. Disquotation does not give the entire content of the concept of truth. At best, then, disquotation gives the extension of a truth predicate for a single language; if we ask what all such predicates have in common, disquotation cannot answer. Something analogous must be said about Tarski’s truth definitions. (2005: 10-11)
And
What is clear is that Tarski did not define the concept of truth, even as applied to sentences. Tarski showed how to define a truth predicate for each of a number of well-behaved languages, but his definitions do not, of course, tell us what these predicates have in common. Put a little differently: he defined various predicates of the form ‘s is true-L’, each applicable to a single language, but he failed to define a predicate of the form ‘s is true in L’ for variable ‘L’. (1990: 285)
In other words, he praises Tarski but takes him not even to be attempting to shed light on the concept of truth. Still, he also rejects redundancy or disquotational approaches. He thinks truth does have a content.
Tarski has shown how the concept of truth can be used to give a clear description of a language. Of course, to give such a description, we must have a grasp of the concept of truth first; but we can have such a grasp without being able to formulate a systematic description of a language. Convention-T connects our untutored grasp of the concept with Tarski’s ingenious machinery; it persuades us that the workings of the machinery accord with the concept as we knew it. (1990: 296)
And the ‘structure and content’ of truth is given by a theory of meaning structured as a Tarskian theory of truth in the way that he, Davidson, has outlined. He says – sorry I seem to be reading a lot of quotations today –:
[T]ruth must somehow be related to the attitudes of rational creatures; this relation is now revealed as springing from the nature of interpersonal understanding. Linguistic communication, the indispensable instrument of fine-grained interpersonal understanding, rests on mutually understood utterances, the contents of which are finally fixed by the patterns and causes of sentences held true. The conceptual underpinning of interpretation is a theory of truth; truth thus rests in the end on belief and, even more ultimately, on the affective attitudes. (ibid: 326)”
“Is that really any different to McDowell’s view?”
“Good question. It seems that they are saying something very similar. Neither explains the nature of truth using meaning, in the way that Tarski does to an extent, though not the general concept of truth. But neither puts much weight on the specific nature of truth except its role within a representation of meaning and mental content in the top-down way we have been discussing. But if this is so, then the earlier slogan that we both mentioned – that Davidson inverts Tarski – is not really true even though Davidson himself seems to support it.”
Millikan, the New Riddle of Induction and grasp of concepts as a primitive ability
There’s a lengthy pause and the others look at Lottie, quizzically. After all, this is her education, really. She looks up from what is left of her hot chocolate and says:
“You tried to spoil what seemed to me to be a proper, modern, scientific account of thought from Ruth Millikan. She says something sensible. Female, too! No surprise there, then! You invoked Wittgenstein, like some god, talking about rules, to suggest that biology provided no way to distinguish thinking of the function of a biological… er… ‘trait’ as perfectly sensible, like an eye being to see, by saying it might do something daft but only from tomorrow. Do you admit that what you said was stupid? That’s one question.
The other is related. If Wittgenstein’s argument, which you never explained, says that the function, or meaning, or content of a biological trait needs interpretation and if that means it cannot explain how meaning or thought reaches sensibly into the future, then how can what we think reach sensibly into the future? I think I know what it is to add 2 to any number you might bowl me, Teach. I can do that but you say I can’t. Answer that one!”
Hipparchus stifles a strong temptation to sigh. “Millikan considers and rejects the criticism I summarised. She imagines two ways to ascribe a function to a trait. For example, we could say:
- The function of the eye is to see objects at any angle in front of its possessor.
- The function of the eye is to see objects at any angle in front of its possessor until some future time t and then do something else.
Past history accords equally well with either. Also, the past history of actual performance diverges from either putative function equally. As we have seen, dispositions and functions are not the same. Oh, and the child’s toy pTravis mentioned is also invoked in ‘teleosemantics’ to stress the idea that matters for biological functions is not what is selected, but what it is selected for. The round slot in the toy allows the insertion of balls or cylinders, for example. It may be that the toy only contains balls not cylinders and also that the actual balls allowed through or ‘selected’ are all green. But they are selected for their round cross section not for being spheres rather than cylinders and not for their green colour.
Millikan argues that, although the two ascriptions of functions above are equally consistent with past biological history – all the many seemingly random births and deaths of rabbits, say – they are not equally good explanations of why eyes have selective fitness. There is a difference in the future for the second option. Millikan rejects it because it is a more complicated explanation. It contains a complication that the simpler ascription does not, and we should always prefer simpler explanations. She says:
I don’t have any particular theory of the nature of explanation up my sleeve. But surely, on any reasonable account, a complexity that can simply be dropped from the explanans without affecting the tightness of the relation of explanans to explanandum is not a functioning part of the explanation. (Millikan 1993: 221)
Now I have two things to say about this. First, it is like Nelson Goldman’s New Riddle of Induction. He argues that past observations of grass and the sky support (many more than) two competing inductive generalisations equally. Either the sky is blue and grass is green. Or the sky is ‘bleen’ and grass is ‘grue’ where:
Bleen = blue until time t and green after.
Grue = green until time t and blue after.
Against the objection that grue and bleen are more complex than green and blue, Goodman points out that simplicity is in the eye of the beholder since:
Green = grue until time t and bleen after.
Blue = bleen until time t and grue after.
Millikan says that first explanation of the function of the eye, above, is simpler. And it looks simpler given the descriptions we find natural. But, and this is my second point, we find some explanations or patterns natural or intuitively simple because we are steeped in a realm of meanings. She, however, is not allowed to presuppose that realm. Her aim is to explain it neutrally by reducing it to biological history. And the history of biological happenings alone cannot do this. Only those seen through a prism of what we find natural explanatory patterns, perceiving a signal under the noise of contingent history ie all the actual performances that did not accord with the function. We see biology as a kind of ‘just so story’ through the lens of meaning. We cannot help using conceptual content to frame explanations so it is no surprise that Millikan’s explanations fit meanings. Meanings are already, illicitly, baked in.”
Lottie looks neither fully persuaded nor fully dismissive. But then that is her constitutional teenager Pyrrhonism. So he continues:
“As to your second question: you are right. We can grasp meanings that reach into the future. The question is whether we can offer an explanation of this. It seems natural to try to explain the fact that we can grasp a meaning, or form an intention, or entertain a thought, all in a flash, by thinking that we have some sort of free-standing mental entity at that time – perhaps, a symbol or image or formula – that determines future applications of words or the reach of thoughts. But such an explanation falls prey to what is called the ‘regress of interpretations’. Any free-standing mental entity that looks to fit a natural but misleading idea of what can come to mind will need interpreting but so too will any symbol encoding the principle of the first interpretation. And the principle of the second interpretation – of the first interpretation – will have to be encoded in another free-standing mental entity also thus needing interpreting.
The alternative? We have a primitive ability to do what we do. We can grasp in a flash the principle of an arithmetic rule or how to apply the concept ‘red’ to future instances. But this is not to be explained except truistically by what can come before the mind’s eye in a flash. In the beginning is the act, or the ability to act.”
“But does that mean there is nothing for philosophy to do?”
“Well, seeing how we are led into philosophical confusion and suggesting routes back is doing something! Though it is piecemeal.
There are philosophers who think that philosophy should be a kind of science in its approach and develop research programmes. One American philosopher, for example, was given a large amount of money – largely for having the biggest beard in academia – to try to show that concepts drop out of non-semantic norms such as institutions where stern tutors beat their students with sticks for not doing their prep! I think that such approaches appeal to a certain sort of petty bourgeoise mind. One that feels the need to count the cost of our rationality. I see no need for that!”
Hipparchus looks slowly round the group with a piecing eye but he looks tired. It occurs to Masongill that their tutor is getting old. Perhaps the worldview they have been exposed to for the last couple of weeks belongs to a fast-fading yester-year. Perhaps it would be enjoyable to be producing knowledge, like widgets or sprockets, in some sort of philosophical factory in a big American city? Making progress! Getting things done!
Then again, if he really thinks that, why is he living in a quiet left-wing commune, hidden in a shabby Victorian out-house behind a crumbling disused stately home in a quiet backwater of England? Nothing ever happens here.