The October issue of The Philosophical Quarterly features an article by Patrick Toner called "Meta-Ontology and Accidental Unity". Toner defends something like Peter van Inwagen's approach to composition: not every two things compose a thing. Artifacts, for example, do not. If one takes the van Inwagen approach to artifacts, one cannot sincerely and truly say 'there are chairs' or 'there are computers'. Or so it seems. However, Toner argues that that is not so. He points out that there are singular predicates that apply to pluralities and which appear to be plural predicates in disguise, for instance, 'family', 'basketball team' and so on. Arguably, Sam, Linda, Amy and Kurt do not compose an object, even if they collectively satisfy the predicate 'family'. On this view, then, 'chair', 'computer', 'telephone' and other similar predicates are (or can be treated as) plural predicates. So, when we say 'there are chairs' or 'chairs exist' we are not committed to there being any single object that satisfies the predicate 'chair'. We are only committed to there being a plurality of entities which collectively satisfy the predicate 'chair'.
This is an interesting argument. However, if the project is descriptive, then I think it fails. For 'family' and 'basketball team' behave just like other singular nominals. They combine with determiners (as in 'every family', 'the basketball team' or 'at most one basketball team'), they are grammatically singular ('every great basketball team is from the East Coast'), and so on. So, the suggestion must be that we ought to treat 'family', 'basketball team', 'chair' and so on as if they were plural.
We can then say 'there are chairs'. This translates as 'there are some Xs that collectively satisfy the predicate 'chair" '. But we should not say 'the chair is soft', 'every chair is broken', 'at most one chair is broken', and so on. Instead we should say something plainly ungrammatical: 'the chair are soft', 'all chair are broken', etc. and we should never say 'at most one chair is broken', for the latter is plainly false. However, we could probably get used to saying things like 'the chair are broken', for people already use plural in combination with singular names and nouns, as in 'Manchester United are playing tonight', 'the government are doing what they can', and so on. My main worry is that since this project requires us to revise English, it is not clear that it fares any better than van Inwagen's original position. After all, van Inwagen's position does not rule out a complete revision of our linguistic practices. For example, we might get used to saying 'there are some particles arranged chair-wise' instead of 'there are chairs'.
Tuesday, October 10, 2006
Meta-Ontology and Singular Plural Predicates
Subscribe to:
Post Comments (Atom)
10 comments:
Hi Brit,
Very interesting. You note that,
"...'family' and 'basketball team' behave just like other singular nominals."
They do so behave, but no one thinks (and you say as much) that, on the one hand, there are all the family members and then, in addition to that, this object called "the family". So Toner's conclusion might be that 'chairs','computers', etc. also *behave* like singular nominals grammatically, though in fact (like 'family') they do not refer to a single object. We know that there is no single object denoted by 'family' and we are not (I think) confused by grammatical usage. Why not assimilate 'chair' to 'family' grammatically? In short, though slightly misleading ontologically, they behave like a singular nominals grammatically.
Hi Mike
There is no single object denoted by 'family' just as there is no single object denoted by 'human'. But I think I can truly say 'there is exactly one family living on 2345 Main Street, St. Louis'. But this translates as 'there is an x that satisfies "family" and for all y, if y satisfies "family", then y = x, and x lives on 2345 Main Street, St. Louis'. So, its truth commits us to there being an object living on 2345 Main Street, etc.
In other words, I think there is an object that is a family. Toner denies this. He thinks there is only a plurality of things there. I think the plurality of things compose an object (albeit a scattered object).
Ok, but none of this affects the suggestion that 'chair' misleadingly behaves like a singular nominative. And this does seem to help PvI. Toner can reject the the translation from 'there is one chair in the room' to (Ex)(Vy)(Cx & Rx & (Cy <-> (x = y))), since that depends on other metaphysical assumptions. You take that translation as proper not simply because 'family' behaves like a singular nominative, but (if I'm reading you right) because you also believe it really is a singular nominative (and not a disguised plural). That is, on your view, there really is a single (scattered) object denoted by 'family'. Right? But then Toner has only to choose some other "singular nominative" for comparison or (more likely) disagree on the metaphysics of families.
Hi Mike,
"Toner can reject the the translation from 'there is one chair in the room' to (Ex)(Vy)(Cx & Rx & (Cy <-> (x = y))), since that depends on other metaphysical assumptions"
But my main point is that his view entails an error theory about natural language, and so, in spite of his proposal being immensely interesting, I am not sure I see how it adds anything to the van Inwagen view.
Also, I do not take 'exactly one chair is broken' to translate the way I suggested because of my metaphysical views. Rather, it translates that way because of the singular 'is'.
'There are some things X that satisfy "chair" and for all things Y, if they satisfy "chair", then they are "proper or improper subsets" of X, and X are broken' is fine. But notice that it contains 'are' rather than 'is'.
Hey Brit,
"But my main point is that his view entails an error theory about natural language, and so, in spite of his proposal being immensely interesting, I am not sure I see how it adds anything to the van Inwagen view."
I don't think so. His view does not seem to entail an error view about natural language. Natural language does not by itself, as far as I can see, entail anything about the metaphysical status of the referents of singular nominals. On the other hand, I guess it might entail an error view about natural or commonsense ontology. I just don't know: is the commonsense view that 'family' refers to a scattered object? My guess is that it isn't. The reference in the singular to a family is just another instance of the familiar practice of potentially misleading nominalization. Toner's observation seems to be this: some singular predicates are (genuinely grammatically singular but oddly) satisfied by a plurality. 'Family' is one. The grammar is ontologically noncommittal. In favor of PvI, one can sincerely and truly say 'there are chairs' or 'there are computers' without being committed to there strictly being such objects and without confessing to speaking shorthand. Those are true since the English singular nominal does not entail that there is some single object called a chair. On the other hand, commonsense ontology might entail that.
"Natural language does not by itself, as far as I can see, entail anything about the metaphysical status of the referents of singular nominals."
But Mike, if English distinguishes between 'is' (singular) and 'are' plural, shouldn't this surface-grammatical distinction be reflected in the semantics? If yes, then we ought to translate 'a table is broken' as 'there is an x such that x is a table and x is broken'. In other words, 'is' indicates that the noun is singular, 'are' that it is plural. An analogy: if 'the government are doing what they can' is grammatical, then 'government' is plural (or perhaps it can function both as plural and as singular). Similarly, if 'a table is broken' is grammatical, then 'table' is singular in English.
Toner might say that just as we say 'students are surrounding the White House', we ought to say 'tables are broken' (when we feel like saying 'a table is broken). But if he is right, then the sentence 'a table is broken' cannot be a sentence in English and true.
"... if English distinguishes between 'is' (singular) and 'are' plural, shouldn't this surface-grammatical distinction be reflected in the semantics?"
But there are interesting exceptions. Isn't is true that 'family' is no ordinary singular nominative? It does seem right, on Toner's behalf, that 'the family' in "the family is well-known" does not refer to single object, but to a plurality or collection, despite the 'is'. As I see it, this feature of the language is what Toner leverages to get 'the chair' in "the chair is nifty" to come out true for PvI. The objection that for PvI there is no single object denoted by 'the chair' seems invidious. There is no single object denoted by 'the family' either.
So maybe our disagreement comes down to this. Very briefly, you are suggesting that since 'the family' takes the singular 'is', that's some (perhaps, good) reason to conclude that the term refers to a single (maybe scattered) object. I want to suggest that sometimes a singular nominal refers to a collection of objects just as it would a single thing.
If there is no one object that is a family, then most of the things we say concerning families are false. Here is a sample:
My family is small
Each family must bring a cake to the party
Mary wants a family
If there is no object that is a chair, then more of the things we say are false.
So, Toner's theory entails an error theory about natural language.
I guess that was my main point. But if Toner's theory entails an error theory about natural language, then why is it better than van Inwagen's view? On van Inwagen's view (as I understand it), we shouldn't sincerely say 'a chair is needed'. Instead, we should say 'some atoms arranged chair-wise are needed' etc.
It might be worth keeping in mind that this isn't the only issue for van Inwagen. It isn't just his denial of mereological sums that leads to his view that there are no chairs. He's got vagueness reasons too. There is no determinate set of simples that makes up the chair, and what could a chair be but a determinate set of simples arranged in a chairlike way? Even if you get around the arrangement of plural simples as a singular thing, you wouldn't thereby avoid his vagueness argument, so more would need to be said (not that there aren't positions that do so, but this isn't enough on its own).
Hi Jeremy,
Yes, I agree. However, I think Toner's view is/would be motivated by the same considerations (as he is denying that the atoms compose one thing). I just don't think his view has an advantage over van Inwagen's.
Post a Comment