Friday, July 28, 2006


Don't Mess With Texas

Back in June I blogged about Mark Sainsbury's talk on fiction at the Arche Audit. In the comments, Ocham pointed me in the direction of this review by Peter Hank of Mark's most recent book Reference Without Referents. (Does anyone actually follow up all these links in sentences in blog posts?). I said I'd try to post something on why I thought Mark's case against the Millian was stronger than Hank made out, and I think it's about time I got on with it.

Mark basically wants to steer a middle course between descriptivism and Millianism. The descriptivist runs into to trouble for reasons familiar from Kripke, while the Millian has trouble accounting for apparently successful communication involving empty names, and with its committment to Shakespearianism. Mark's central point is that these two positions don't exhaust logical space; there's a stable resting point between them which avoids the shortfalls of each.

Firstly, Mark wants to develop a proposal made by John McDowell in 'On the Sense and Reference of a Proper Name', whereby we draw the sense/reference distinction within a Davidson truth-theoretic semantic theory. McDowell's point was that the distinction so drawn gave us resources enough to avoid Frege's puzzle (and other problematic inferences validated by Shakespearianism) without conceding anything about the semantics of proper names to the descriptivist. Secondly, Mark proposes, following Tyler Burge, that the correct logic in which to set semantic theory is not classical, but rather a negative free logic. This allows empty proper names to be treated within semantic theory just as referring proper names are; by homophonic axioms assigning a reference-condition, the condition an arbitrary object would have to satisfy in order to be the bearer.

Mark motivates the latter proposal to a large extent by giving examples to show that empty proper names can be meaningful, as can the sentences in which they feature. They are thus as firmly the concern of semantic theory as referring proper names are, and should be treated in a uniform manner. Hank's principle objection to Mark's position is he has conflated meaningfulness with intelligibility; that all the data Mark gives shows is that empty proper names can be intelligible, but the Millian can accept this without conceding that they are meaningful and hence to be treated by semantics. Hank suggests that:

'. . . a Millian could hold that when speakers utter sentences containing empty names they succeed in conveying propositions, even though the sentences they utter do not themselves semantically express any proposition.' (372)

Hanks offers the following example. The Millian can hold that an utterance of

(10) Santa Claus will come down the chimney tonight.

'can communicate the proposition that the jolly fat man from the North Pole who drives a flying sleigh will come down the chimney tonight, even thought (10) itself is strictly speaking meaningless' (372). It's enough to vindicate the intelligibility of empty names that they can be 'reliably used' (373) in conveying these sorts of descriptive propositions, but that isn't yet enough to vindicate Mark's claim that they must be considered meaningful:

'Empty names may be intelligible, in the sense that they can be reliably and successfully used in communication, without being meaningful.' (373)

Likewise, a belief report like an utterance of

(11) The child believes that Santa Claus will come down the chimney tonight.

can convey that the child has some descriptive belief, even though 'semantically, no belief is expressed by the embedded clause in (11) and hence at the semantic level no belief is attributed to the child'. (373) Hank notes that Mark tries to dismiss this kind of proposal by noting that intuitively 'children all have the same belief when they believe that Santa Claus will come down the chimney tonight, even though their parents have given them slightly different versions of the Santa Claus story...'. (RWR: 89), but he suggests that this intuition is too easily accomodated by the Millian to decide the issue.

I'm guessing Mark will be particularly unhappy about the suggestion that believe reports like (11) convey that the child has some descriptive belief without attributing any belief at the semantic level. I think that Mark, like McDowell, takes very seriously Davidson's insight that a theory of meaning is in part a tool to allow us to make accurate belief-attributions. It's partly this that drives the thought that homophony is an ideal we should aspire to in semantic theory; homophony ensures there's a certain transparency to language which allows our interpretation of another's utterances to be a useful guide to their beliefs. It looks like Hank's proposal doesn't allow semantic theory to play this role anymore, and that seems like a real cost.

I hope Mark will be sympathetic to that response, but it probably invokes too substantial a commitment to be a fully satisfying response to Hank's charge that Mark hasn't motivated his own position on empty names over the Millian's. I think the real problem with Hank's suggestion is that it offers does nothing towards settling the explanatory debt incurred by recognition that '[e]mpty names may be intelligible, in the sense that they can be reliably and successfully used in communication' (373). In offering examples that show that empty proper names may be intelligible in this sense, Mark has highlighted a feature of out language practices which calls out for explanation.

Mark's explanation of the reliability and success of our communication with empty proper names is appealingly simple, for it just appeals to the compositional mechanicisms which it is hoped explain reliable, successful communication with other subsentential linguistic items. What is the rival explanation Hank offers of how an utterance of (10) reliably and successfully communicates some descriptive proposition (despite lacking meaning)? He doesn't give one, he just points out that there is logical space for such a view. I just don't think that meets Mark's challenge to the Millian, for it just fails to engage with the explanatory task which Mark's examples motivate.

A similar problem faces Hank's proposal regarding belief reports. Hank fails to note that Mark does not simply claim the intuition that children who believe that Santa Claus will come down the chimney tonight might all share the same belief despite hearing different versions of the story from their parents. He also writes, immediately following the remark quoted by Hank:

'The direction taken by our envisaged critic involves all the difficulties involved in seeing non-empty names as expressing descriptive thoughts: the whole point of a name (we may add: empty or non-empty) is to enable communication among those who do not coincide in the information they associate with the expression.' (RWR: 89)

Again, there's a challenge here to explain successful communication (which surely takes place) between children and adults who don't all associate the same description with the expression 'Santa Claus'. Mark's point is that his envisaged opponent faces all the problems highlighted by Kripke's Feynman examples and the like in trying to meet this challenge. And again, Hank hasn't even gestured at how an explanation which avoids these worries might go.

In sum, then, merely pointing out position in logical space for a Millianism that concedes the intelligibility of empty names without conceding their meaningfulness is totally unsatisfying as a response to Mark's challenge to Millianism; it just leaves the major explanatory tasks untouched.

Labels: ,

Hey Aidan,

Just a quick question for the uninitiated: What do you mean by "Shakespearianism"?
Hey Alex, how's things?

By Shakespeareanism, I mean the thesis that coreferential expressions are intersubstitutable salva veritate (ie. without change in truth value). The familiar problems come from intensional contexts like belief ascriptions, modal contexts, and the like. For example, Shakespearianism commits us to the validity of the following argument:

1. Lois Lane believes Superman can fly.

2. Clark Kent is Superman.


3. Lois Lane believes Superman can fly.

The name comes from Peter Geach - here's a quote from David Sosa's 'The Import of the Puzzle About Belief':

'Geach introduced the term 'Shakespearean' after the line "A rose / By any other name / would smell as sweet."' (374)
Thanks for posting this. But let me put the two positions in a way that makes it less easy 'to steer a middle course'.

* Vulcan does not have an atmosphere

If that sentence is true, then something (Vulcan) does not have an atmosphere. If it is not true, then something (Vulcan) does have an atmosphere. Either way, we have something such as Vulcan. But is there such a thing?

There are only two positions here, and no middle course. Either deny that the sentence above implies the existence of something that has no atmosphere. This is Sainsbury's way out, and also the way of Russellian descriptivism, with its distinction between wide and narrow scope negation, even for proper name sentences. (Interestingly Abelard in the 12th century makes the same distinction, between Socrates non est albus and non Socrates est albus).

Or accept that the proper name refers in both cases. That is the direct reference view. If you also accept that there are empty names, you then have to distinguish the semantics of empty name sentences from the semantics of 'real' sentences. I.e. some such distinction as Hank draws between intelligible and meaningful.

In summary: there is no middle course. Either we reject particularisation for negative proper name sentences. Or we distinguish the semantics of full names from empty names. Hank gives no argument for this. But the argument, essentially, is as above.
Thanks Ocham. But I have to admit, I'm missing something here. Why is it a bad thing that Sainsbury's solution to problems like negative existentials looks like the descriptivist's? I thought the whole point was to show that the theoretical advantages of descriptivism could be had without committment to the kind of thesis about the semantics of proper names that runs into Kripke's objections. I'm not seeing why observing that both positions deal with negative existentials by rejecting particularisation shows that Sainsbury's position isn't distinct from descriptivism, since I don't see what's been said so far to suggest his position faces the same difficulties that descriptivism faces.

I know from your previous comments that you're not happy with rejecting particularism to handle negative existentials, but I take it that's a somewhat tangential point here.
>>> Why is it a bad thing that Sainsbury's solution to problems like negative existentials looks like the descriptivist's?

Because of particularisation. He admits there is a bit of a problem about sentences like 'Vulcan has no atmosphere', and if you read his book carefully you see he fudges the problem somewhat.

>>> I'm not seeing why observing that both positions deal with negative existentials by rejecting particularisation shows that Sainsbury's position isn't distinct from descriptivism

Of course it is distinct. But what they both have in common is the need to deal with our intuition is that empty names are semantically like 'real' names, and thus that Hanks' distinction between 'intelligible' and 'meaningful' is artificial.

>>> I thought the whole point was to show that the theoretical advantages of descriptivism could be had without committment to the kind of thesis about the semantics of proper names that runs into Kripke's objections.

I read this sentence again and realised I didn't understand it. What are you referring to as 'the kind of thesis about the semantics of proper names that runs into Kripke's objections'? You mean descriptivist semantics, yes?
Yes, sorry, I meant descriptivist semantics in that sentence.
What are we calling 'descriptivism' here? I had a look at Kripke (N+N) again, and he is actually very vague about what it is. He seems to characterise a sort of non-rigid descriptivism, and then proves that names cannot be non-rigid descriptors, because they aren't.

Some of the medieval logicians held that proper names are descriptive of a particular feature that only the very individual possessing it, and no other, can have. 'Platonitas solius unius est hominis et hoc non cuiuslibet sed solius Platonis, humanitas vero et Platonis et caeterorum quicumque hoc vocabulo continentur'. Non-rigid descriptors, if you like.

The distinction between referring expressions and predicates (whether rigid or not) is different, of course.
Yes, Kripke's unfortunately talks rather loosely about the Frege/Russell view of proper names. I had in mind roughly Mark's characterisation: 'according to descriptivist theories, the meaning of a name is given by or is equivalent to some body of associated information, and the referent of the name is whatever this information is true of, if anything'. (RWR: 2)

The medieval view you describe sounds close to the view Kripke ascribed to Frege and Russell. I take it the talk of a 'body of associated information' in Mark's characterisation is to accomodate also the Searle-style 'cluster' view that Kripke discusses and rejects.
Not quite. The medieval idea, though not clear, is closer to some idea of 'haecceity', a single property, rather like a McDowellian property, that just applies to that individual.

There was a lot of discussion of names like 'Alexander', which can mean Alexander (of Troy) or Alexander (of Macedonia). But this is not like the way that 'man' is true of many individuals. For 'Alexander' has a different meaning when it applies to different people. (Mill makes the same point in System of Logic - in fact most logicians in every period do - progress in logic largely depends on forgetting its history).

Aquinas writes "Now it can be objected that the name Socrates or Plato is naturally predicated of many, since nothing prevents there being many who are called by this name. But the reply to this is clear, if the words of Aristotle are attended to. For he himself does not divide names into universal and particular, but things. And for that reason it is to be understood that 'universal' is said when not only the name can be said of many things, but what is signified by the name, is naturally found in many things. This however does not happen with the names above, for the name Socrates or Plato signifies human nature according as it is in this material. But if this name is imposed upon other men it will signify human nature in another material, and thus there will be another signification of it, wherefore it will not be universal, but equivocal."
Apologies, but I completely missed the fact that you are actually at Texas. The introductory bit of your blog mentions Scotland, and I got no further. Does this mean you see Mark regularly? If so, give him my regards. I haven't been in touch for a long time. He will know me as d e a n b u c k n e r.
Yup, I'm one of his students. I'm in Glasgow just now, but I'll be sure to send your regards when I'm back in Austin in a couple of weeks.
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?