Written by Andrew on November 5, 2009 – 9:45 pm
I have been reading a couple of densely argued, complex and interesting papers by Paul Boghossian, “The Rule Following Considerations” (Mind 98 (1989): 507-49) and “The Status of Content” (Philosophical Review 99 (1990): 157-84). I thought it might help me to understand the intricacies of Boghossian’s arguments to blog about them and try to lay out the rich vein of argumentation. The main goal of the discussion is pretty clear. Boghossian contends that the sceptical solution leads to some kind of contradiction (on either of the two plausible ways of making sense of it). The sceptical solution replaces the realist truth-conditional picture of language with a sceptical assertion-conditional one. Boghossian argues that the assertion conditional account, considered as a form of nonfactualism about meaning, leads to contradiction in view of what one has to say about truth to maintain a nonfactualist stance about anything, and what one has to say about truth given nonfactualism about meaning in particular. He also argues that the sceptical solution leads to contradiction when considered as a form of error-theory. Thus he challenges the sceptic to put up or shut up: show there is some other way to understand the sceptical solution or withdraw.
The contradiction affecting the nonfactualist version of the sceptical solution is supposed to be that it entails that truth is not a robust property but it presupposes (so also entails?) that truth is a robust property (1989: §16, 1990: 175).
First let me summarise the argument that nonfactualism works only with robust accounts of truth. Nonfactualism about a discourse holds that the predicates of the discourse do not denote properties, nor do the sentences of the discourse express propositions or have genuine truth conditions (1990:160, 1989: §15). Boghossian claims that this much is inconsistent with minimalism about truth, which holds that simply being of the right syntactic sort is sufficient for a sentence to truth-evaluable. Sentences must be significant (‘disciplined by norms of correct usage’ and declarative in form (‘possess an appropriate syntax’) (1990: 163). Now Graham Oppy, Michael Smith and Frank Jackson have challenged this minimal definition of truth-aptness, but the criteria can stand for the present purposes. Given his assumptions about minimalism about truth and nonfactualism, Boghossian draws the obvious inference that nonfactualism is incompatible with minimalism because it is constitutive of nonfactualism that it denies that (some range of) significant declarative sentences have truth conditions. Nonfactualism is therefore committed to a nonminimal or ‘robust’ theory of truth (1990: 165, 1989: §16).
So far so good, but what of the argument that nonfactualism entails that truth is not a robust property? I am having a bit of trouble with this. I think I know what the steps are supposed to be but I can’t tie them to the text and the principles employed in the text even if they deliver the conclusion, look unacceptable to the sceptic to me. So there’s a petitio principi allegation lurking in the wings. I think it is supposed to go like this
1. Nonfactualism says predicates don’t express properties.
2. So it ought to say the predicate ‘true’ doesn’t express a property.
3. The claim that ‘true’ doesn’t express a property is a central commitment of minimalism about truth.
So, nonfactualism implies minimalism about truth.
In ‘the Status of Content’ to get from “the predicate ‘x has truth condition p’ does not express a property” to ” ‘x is true’ does not express a property”, he goes this way:
The truth value of a sentence is fully determined by its truth condition and the relevant worldly facts. There is no way, then, that a sentence’s possessing a truth value could be a thoroughly factual matter (“true” does express a property) if there is nonfactuality in one of its determinants (“has truth condition p” does not express a property (1990: 175)
Isn’t there a hint of use/mention looseness in this passage? Going by the first sentence of the passage it is the truth-condition that matters for the truth value, not the predicate ‘has a truth condition’, but ignore this. ( I suppose the implicit idea is that the predicate would fail to express the property only if the property didn’t exist…. but that is not scepticism. Kripke never denies that addition and quaddition and the other functions exist, only that we have words and concepts to express them)
My worry about this passage is that the first sentence seems to buy in to the truth-conditional picture that the sceptic rejects. The sceptic is supposed to reject the whole picture of sentences having truth conditions. If you reject talk of truth conditions, you likely reject the lore associated with them. If the sceptic says sentences don’t have truth values, where does the argument go from there? Why can’t the sceptic claim that truth is a property of other things besides sentences (propositions say)? For the sake of argument, assume the sceptic says the following about truth: Propositions are sets of possible worlds and a proposition is true if and only if the actual world is a member of the set of worlds that constitutes the proposition. So the truth of a proposition consists in having the actual world as a member. This doctrine seems open to the sceptic. It has nothing particular to do with language. There would still have been set of possible worlds even if intelligent life hadn’t evolved.
The equivalent passage in the Rule-Following Considerations paper is:
Since the truth condition of any sentence S is (in part, anyway) a function of its meaning, a non-factualism about meaning will enjoin a non-factualism about truth conditions: what truth conditions S possesses could hardly be a factual matter if that in virtue of which it has a particular truth-condition is not itself a factual matter (1989: §15).
It is the same move.
I think my argument above is more direct. Nonfactualism says predicates don’t express properties. ‘True’ is a predicate. So the nonfactualist is committed to the claim that ‘true’ doesn’t express a property. This is minimalism about truth.
My worry about this argument is the status of the definitional connection between minimalism about truth and the denial that ‘true’ expresses a property in the context of scepticism. With scepticism about meaning in the picture, we ought to separate out elements of the minimalist thesis. The core element is the denial that truth is a substantive property. This is set alongside a view about what the predicate ‘true’ does given that it doesn’t express a property. The second thesis in the absence of the first would not count as minimalism about truth. Nonfactualism about meaning, if combined with the toy theory of truth described earlier, accepts the second thesis but rejects the first (or at least accepts the denial that the predicate expresses the property without denying that there is a property (of having the actual world as a member)).
Boghossian seems much too quick in ‘the Status of Content’. He equates ”true’ does not refer to a property’ with the denial that truth is a genuine property (1990: 175). And the toy theory would appear to be a counter-example to that move.
Andrew,
First let me reconstruct Boghossian’s argument so that I can be clear about what’s going on there. He argues that non-factualism is inconsistent because (a) although it presupposes a robust conception of truth in the claims it makes about the non-factualism of the given discourse, (b) it can actually be shown to entail a non-robust conception of truth. Hence, Boghossian’s argument against non-factualism depends on being able to demonstrate that it has these contradictory commitments. You focussed on the second issue, (b), and argued that this can’t be demonstrated. But I think the real weakness lies with the first: his claim that non-factualism presupposes a robust notion of truth.
Boghossian argues for this along the following lines: The non-factualist about a particular discourse allows that the sentences of the discourse are declarative sentences, e.g. in the sense that they meet the deflationary criteria of being syntactically well-formed and disciplined (in the way characterised by Wright); but the non-factualist nevertheless denies that such declarative sentences have truth-conditions or are potentially descriptive of facts. So, given that Syntax and Discipline are the deflationary criteria that must be met in order for a discourse to be construed as factual or truth-conditional, the non-factualist’s denial that the sentences of the given discourse possess truth-conditions must be based on a more robust conception of truth-conditions, and hence a more robust conception of truth.
Criticism: Boghossian tries to convince us that non-factualism presupposes a robust notion of truth by formulating the assertion of non-factualism in a particular way: that the declarative sentences of the given discourse do not possess truth-conditions. This formulation presupposes that the sentences are indeed correctly characterised as declarative sentences and hence that they already meet the deflationary criteria for being truth-conditional. And it therefore makes the non-factualist’s denial that they are truth-conditional look like he must be enforcing some extra criteria, and so employing a more robust conception of truth. Boghossian’s formulation, though, is misleading. It obscures the possibility that we can be non-factualists about a particular discourse based on the fact that the sentences of the discourse are not genuine declarative sentences, i.e. they don’t even satisfy the deflationary criteria of being syntactical and/or disciplined. Holding a non-factualism about a particular discourse in that sense does not presuppose a robust concept of truth. It only enforces deflationary criteria and the discourse in question fails to meet them.
I think Boghossian is misled by Kripke’s particular brand of non-factualism. Kripke, on Wittgenstein’s behalf, does seem to hold that the sentences of semantic discourse – meaning ascriptions – are genuine declarative sentences satisfying the ‘Syntax’ and ‘Discipline’ minimal criteria. And so Kripke’s assertion of semantic non-factualism certainly would seem to presuppose a robust concept of truth. Boghossian is wrong to suppose, though, that this is how any non-factualism must be articulated. This character of Kripke’s particular non-factualism is what leads Soames to state that the non-factualism of Kripke’s Wittgenstein’s sceptical solution is not as radical as we might first have thought. It would be far more radical to assert a non-factualism based purely on deflationary criteria.
The really interesting question, then, is this: Given that Kripke’s Wittgenstein’s non-factualism does appear to presuppose non-deflationary criteria for truth-conditionality, is it vulnerable to Boghossian’s overarching criticism that it is actually committed to two incompatible claims about truth? My response to this is to argue that, despite appearances, Kripke’s Wittgenstein is not really committed to the robust notion of truth. This is roughly how I would argue this:
The inflationary conceptions of truth, truth-condition, and fact are part of a broader doctrine; it is broader in the sense that it takes account of more than sentences and facts to include similar claims about, e.g., predicates and properties. This broader doctrine has been called different things, but we can follow George Wilson and call it ‘classical realism’. The core idea is that it operates with an inflationary conception of these key notions. (It is a big question as to what makes these conceptions inflationary, but leave that aside.)
Very roughly, I see Kripke’s Wittgenstein’s sceptical argument (which leads to the non-factualism) proceeding as follows:
(1) Assume ‘classical realism’ (including the inflationary criteria for truth-conditionality);
(2) Derive the sceptical paradox;
(3) Conclude with the negation of classical realism;
(4) Develop the non-factualism of the sceptical solution on the basis of the rejection of classical realism.
Note that the non-factualism of the sceptical solution involves a rejection of classical realism, and hence of the inflationary conceptions of truth and truth-condition.
So, Boghossian is only partly correct: Kripke’s Wittgenstein does presuppose the inflationary criteria for truth-conditionality (as part of the wider doctrine of classical realism), but only to show that it leads to the sceptical paradox and must be abandoned. Boghossian is wrong that Kripke’s Wittgenstein remains committed to it when he defends his particular version of non-factualism. Hence, KWittg’s particular version of non-factualism does not have the contradictory commitments that Boghossian says it does. He is only committed to the deflationary notions.
Thanks for the awesome comment Tom! I can see you’ve been thinking about these issues intensely. But I’m afraid to say that I would support Boghossian against the argument you make (perhaps you would say the same about my strategy… so although we agree that his argument won’t do, we can’t completely agree).
To understand your response better it would help me to know what sentences you think the sceptical solution applies to. Kripke says, p. 95, that it consists in a description of the game of concept attribution, and the concept attribution sentences he explicitly mentions are ones like “Jones means addition by ‘+’” (e.g. p.90). So if the sceptical solution is a form of nonfactualism, then it would appear that sentences like “Jones means addition by ‘+’” are among the ones that are supposedly nonfactual.
Do you think “Jones means addition by ‘+’” fails either the syntax or the discipline condition? On what grounds? It appears to be grammatical and I take it the description on page 90 of when and how it is taken to be assertable is Kripke’s explanation of the norms governing its usage. I think that it is appropriate to call this a matter of ‘norms’ because Kripke uses obviously normative vocabulary when describing the assertion conditions e.g. ” Jones is entitled, subject to correction by others, provisionally to say…” (entitled!).
It seems you walk a tight rope between imposing sufficient restrictions on sentences so that they fail the minimalist and deflationary requirement but without making the restrictions seem inflationary.
Of course, you concede that in places Kripke perhaps does accept the inflationary requirements (perhaps like the above example). In which case, what sort of sentences does the good nonfactualism Kripke should have put forward apply to? I think that it would be contrary to the stated intentions of the sceptical solution (to save the ordinary way we think and talk) not to include sentences like the above example within its compass.
I don’t see how you can get the notion of inflationary criteria for truth-conditionality to do the work you want it to. Do you mean that syntax and discipline come in two varieties inflationary versus deflationary syntax and inflationary versus deflationary discipline? It is a distinction without a difference. But otherwise I don’t see what you mean. If appropriate syntax and being subject to norms of discipline are enough for a sentence to be truth-conditional, then what room is there for a position to grant that and still be nonfactualist?
To answer your first question Andrew, I think the non-factualism of the sceptical solution applies at least to meaning ascriptions (or concept attributions) such as “Jones means addition by ‘+’”. Both Boghossian and Wright have similiar arguments for holding that the non-factualism generalises to all sentences; while Wilson argues that there is no need for this generalisation because the original sceptical argument establishes this global non-factualism directly. So I say that the non-factualism applies ‘at least’ to meaning ascriptions, but I think it can be argued that it also applies to the sentences of other types of discourse; the debate would just be whether or not the latter can be established on the basis of Kripke’s Wittgenstein’s sceptical argument alone.
I might have confused matters a little by talking about the possibility of a non-factualism that fails to meet the Syntax and/or Discipline conditions, i.e. fail to meet the minimal criteria for being truth-conditional. All I wanted to say there is that Boghossian is wrong to generalise and state that non-factualisms necessarily presuppose a robust concept of truth, or inflationary criteria for counting as truth-conditional. The above possibility was merely meant as a counter-example to that because such a non-factualism wouldn’t presuppose inflationary criteria. Note, though, that I did NOT mean that that is how Kripke’s Wittgenstein’s non-factualism should be understood. I think it is obvious that meaning ascriptions such as “Jones means addition by ‘+’” do meet the Syntax and Discipline criteria, and that Kripke thinks that they do.
That should clear up a few of the objections in your last post. But you’re probably wondering now how I can say that KWittg’s non-factualism meets the minimal criteria, on the one hand, and nevertheless hold that Boghossian is wrong about such a non-factualism presupposing a robust notion of truth, on the other.
To respond to that I would need to address what you said in your last paragraph above:
“I don’t see how you can get the notion of inflationary criteria for truth-conditionality to do the work you want it to. Do you mean that syntax and discipline come in two varieties inflationary versus deflationary syntax and inflationary versus deflationary discipline? It is a distinction without a difference. But otherwise I don’t see what you mean. If appropriate syntax and being subject to norms of discipline are enough for a sentence to be truth-conditional, then what room is there for a position to grant that and still be nonfactualist?”
I don’t mean that the Syntax and Discipline criteria come in two varieties: inflationary versus deflationary. Syntax and Discipline are just the deflationary criteria that a discourse has to satisfy in order to be considered truth-conditional in a minimal sense. When I talk about inflationary criteria for truth-conditionality, I have in mind criteria of a different sort. Soames and Kusch (e.g.) have struggled with characterising these inflationary criteria. They both put it down to ‘a contentious explanatory role [that facts, truth-conditions are assigned] in theories of meaning and understanding’ (Soames 1998). Soames puts it like this:
“The facts correlated with sentences play an explanatory role in meaning and understanding: a sentence s is meaningful because it stands for a potential fact the obtaining of which would make s true: a person understands s in virtue of knowing that s is true iff the fact in question obtains” (1998: ‘Facts, Truth-Conditions, and the Skeptical Solution to the Rule-Following Paradox’, p. 321).
The doctrine of ‘classical realism’ (as Wilson understands it) is characterised by assigning this ‘contentious explanatory role’ to truth-conditions; and so, it will only assert that the sentences of a particular discourse are truth-conditional if – beyond meeting the Syntax and Discipline minimal criteria – the corresponding facts can be held to fulfil this explanatory role.
As I said in the last post, KWittg’s non-factualism should be seen as a rejection of the doctrine of classical realism, and with it the inflationary criteria for truth-conditionality. At least as Wilson, Soames, and Kusch view things, this leaves the sceptical solution looking as follows: it will be characterised as non-factualist if the relevant criteria are inflationary because KWittg has shown that semantic discourse cannot meet these criteria; but it will be characterised as factual if the criteria are deflationary because it holds that semantic discourse does satisfy the Syntax and Discourse criteria.
Note also that it is not just that KWittg argues that semantic discourse does not or cannot satisfy the inflationary criteria; he argues that these criteria are illegitimate. KWittg only ever enforces these criteria to show that doing so leads to the sceptical paradox that no-one ever means anything by a word. This is why Boghossian is wrong about KWittg’s non-factualism remaining committed to a robust notion of truth.
Thanks Tom. That does clear things up. In a sense, you actually agree with Boghossian, but you don’t think the position advanced in WRPL is a form non-factualism at all. It’s a form of factualism (given that KW rejects inflationary criteria and does accept that anything that satisfies SYNTAX and DISCIPLINE has truth-conditions of a correspondingly dilute sort).
I agree with you, by the way, that the view in question is supposed to be completely general in scope. Kripke talks about Wittgenstein replacing one picture of language with another. And he kicks off the whole topic with a mathematical example rather than a semantic one. So it seems pretty clear that the view applies to all discourse and not just to the idioms of meaning attribution. I also think that Boghossian’s and Wrights arguments that show the sceptical solution has to ‘go global’ succeed. My only reservation about them is how they are advertised, as if the conclusion had to be wrung forcibly from Kripke when the view was never represented as anything but general.
It might be helpful if you could post more about the factualist interpretation of KW. On this view, he would subscribe to a deflationary conception of truth on which meeting the minimal criteria SYNTAX and DISCIPLINE are sufficient for possessing truth-conditions.
If this is the correct interpretation of KW then Kripke did a poor job of expressing himself. He shouldn’t have described KW as replacing the traditional picture of truth-conditions with an alternative picture based on assertability or justification conditions (e.g ‘replacing’ 73, 77 ‘alternative’ 73, 75) because that is quite misleading since it neglects to mention that, according to your interpretation, the notion of truth-conditions is retained.
I realise you aren’t the originator of the view, so it is unfair to pose these exegetical questions to you, but you are an able defender of it so I would like to know what you think.
I agree that the interpretation I discussed does make Kripke’s insistence that truth-conditions are replaced by assertability-conditions look strange. In order to defend the deflationary factualist interpretation of the sceptical solution, it would perhaps require us to view Kripke’s claim in a slightly different light based on the different senses (inflationary and deflationary) of truth-conditions now in play. Recall that the relevant sense of ‘inflation’ is usually characterised in terms of the ‘contentious explanatory role’ that truth-conditions (and properties and functions) are assigned in accounting for meaning and understanding. The move from truth-conditions to assertability-conditions would then amount to ceasing from appealing to truth-conditions to explain meaning and understanding, and turning instead to assertability-conditions to fill this role. The idea would thus seem to be that we can still posit truth-conditions even after this switch; but that they just don’t do any explanatory work for us.
It is probably fair to say that all this is moving beyond Kripke’s text, but many defenders of this view (especially Kusch) believe they are defending Kripke. The following claim by Kripke could be cited:
“We merely wish to deny the existence of the “superlative fact” that philosophers misleadingly attach to such ordinary forms of words, not the propriety of the forms of words themselves.” (WRPL, p.69)
But this is inconclusive too. Note that he states at the end that he is not opposed to ‘the propriety of the forms of words themselves’; he does not say he is not opposed to meaning (non-superlative) facts, which is what Kusch and others would want him to say. This indicates that Kripke’s text might not be able to help us decide. For example, why would he suggest the distinction between superlative and non-superlative facts, and emphasise that the sceptic only opposes the former?
By maintaining that the sentences of semantic discourse are truth-conditional in some minimal sense, we are stating that these sentences are potentially descriptive of semantic facts and (barring an error theory) that there are semantic facts to be described. That leads me to your question about what we can say about these facts. I don’t have any settled view on this yet. If we can characterise the inflationary conception of semantic facts, we can at least give a negative characterisation of these deflationary facts. As for a positive characterisation, I think it is helpful to turn to the debate between McDowell and Wright.
Both McDowell and Wright posit facts about meaning. But crucially, both of their conceptions of semantic facts are self-consciously opposed to any platonistic or classical realist conception. It is also very interesting that McDowell and Wright disagree with each other over how to characterise these non-platonistic semantic facts: whereas Wright gives a sort of anti-realist characterisation, McDowell gives a characterisation somewhere between the extremes of anti-realism and full-blown platonism. The point I want to make is that this debate about the nature of facts about meaning can continue even after we have rejected classical realism (or the inflationary conception of semantic facts). So, when looking to characterise the deflationary semantic facts that are supposedly left over after KWittg’s argument, this is where I’d look. However, I haven’t yet come to a settled opinion on what the correct view is here. The one thing I am certain of, though, is that McDowell is wrong.
You keep repeating the phrase ‘contentious explanatory role’ but I don’t think this is very helpful. Kripke can’t have anything against explanation as such because the solution makes use of assertion conditions which play an explanatory role. The role of assertion conditions is to licence or entitle speakers to make statements. This is an explanatory role. So Kripke (’s Wittgenstein) cannot be against that. Moreover, assertion conditions are arguably worldly conditions with an explanatory role (they are not out of this world). What is supposed to be wrong with playing an explanatory role? (and is there an objection to it in the text?)
A point of clarification: the claim is not that deflated truth-conditions are identical to assertion conditions, is it? In that case the explanatory role of assertion conditions would be hard to square with the characterisation of inflationary truth-conditions as conditions with an explanatory role.
No, the point is definitely not that the deflationary truth-conditions are identical to the assertability-conditions.
The best way of discussing your other comments is in connection with the following passage from Kripke:
‘Wittgenstein’s theory should not be confused with a theory that, for any m and n, the value of the function we mean by “plus”, is (by definition) the value that (nearly) all the linguistic community would give as the answer. Such a theory would be a theory of the truth conditions of such assertions as “By ‘plus’ we mean such-and-such a function”, or “By ‘plus’ we mean a function, which, when applied to 68 and 57 as arguments, yields 125 as value”… The theory would assert that 125 is the value of the function meant for given arguments, if and only if “125″ is the response nearly everyone would give, given these arguments. Thus the theory would be a social, or community-wide, version of the dispositional theory, and would be open to at least some of the same criticisms as the original form’ (WRPL, 111).
So it can’t be the task of the sceptical solution to specify assertability-conditions in such a way that they come out looking just like a species of truth-condition. Hence, whatever the assertability-conditions that we posit are, they must not be assumed to be necessary and sufficient conditions for someone meaning something by a word. However, the sceptical solution still has a lot to say about ‘the criteria’ for meaning something by a word, as in the following well-known passage:
‘Jones is entitled, subject to correction by others, provisionally to say, “I mean addition by ‘plus’”, whenever he has the feeling of confidence – “now I can go on!” – that he can give “correct” responses in new cases; and he can give “correct” responses in new cases; and he is entitled, again provisionally and subject to correction by others, to judge a new response to be “correct” simply because it is the response he is inclined to give’ (WRPL, 90).
The important notion here is that of the agreement in responses regarding the use of words. In a sense, we could still speak here of ‘explanation’ of what meaning involve. But it does not appeal to facts of any kind (not even about agreement in responses) that would be necessary and sufficient for someone to count as meaning something by a term. When I say that ‘Jones means addition by “plus” because he usually gives the same responses to addition problems that I (and other members of the community) give’, this is only legitimate if it is taken to mean that ‘typically’ I am justified in stating this about Jones given these facts about his responses. But these facts that I do appeal to are not necessary and sufficient conditions for Jones meaning addition by ‘plus’. Hence, to avoid being misled, it would probably be best to replace talk of explanation with talk of clarification (or some such weaker notion).
This relates to a very similar issue in the later Wittgenstein regarding the distinction between explanation and description. Generally speaking, Wittgenstein views philosophy as attempting to explain certain phenomena (e.g. meaning) that aren’t in need of explanation. But with all his talk about language games and forms of life, it might often seem that he himself is giving some sort of explanatory account of how meaning is constituted. Ultimately, though, these notions are only meant to clarify or elucidate, or serve as reminders about the social embeddedness of language.
Returning to Kripke, let’s adopt his neat characterisation of our form of life as ‘The set of responses in which we agree, and the way they interweave with our activities’ (WRPL, 96). Again it might appear that KW is giving some sort of explanatory account by specifying the assertability-conditions of sentences in terms of facts concerning the particular language-user’s agreement in his use and responses with others in the community. But explanation is probably out of place here. So to repeat: one of the main problems that KW finds with the traditional (inflationary) truth-conditional conception of meaning is that it attempts to explain meaning. I was wrong to indicate in the last post that assertability-conditions replace truth-conditions in the sense of filling or taking on the same explanatory role.
I’m in agreement with you Tom, until you say “Hence, to avoid being misled, it would probably be best to replace talk of explanation with talk of clarification (or some such weaker notion).” I don’t think the notion of assertability conditions plays any less of an explanatory role simply because it doesn’t issue in necessary and sufficient conditions for utterances. I think the notion manifestly and obviously does play an explanatory role, so the previous point still applies: Kripke’s Wittgenstein cannot have anything against explanation per se. He’s citing assertion conditions in explanation of the nature and character of actual language use not of meaning.