Universalism is the thesis that composition is unrestricted. I'm curious about the following view (suggested to me, but not endorsed, by Dave Liebesman). Universalism is false. There is nothing composed of my nose and the Eiffel Tower. But in the language of the ontology room ("Ontologese"), the sentence 'composition is unrestricted' is true. Further, it is part of the semantics of the quantifiers of English that "they do not utilize the largest possible domain." This last bit doesn't make sense if one understands it in English, since it would then amount to the claim that there are things that are not among the things that there are. Understand it, rather, as a claim in Ontologese.
(N.B. The remainder of this post is in English.)
One hears a lot about Ontologese these days (especially in Boise). I doubt that there really is such a language. Even if there is something interesting for 'there are strange fusions' to mean, other than that there are strange fusions, it is hard to see how the meanings of this and other Ontologese sentences could ever have gotten fixed. It obviously will not do to start in English and announce, with stipulative intent, that the Ontologese 'there is' is to have the absolutely unrestricted domain (even setting aside worries about paradox), for this is just the domain of the ordinary English quantifier. Alternatively, we could announce (with more holistic ambitions) that our words are to have only perfectly eligible meanings, continue our metaphysical debates, and let the magnetic theory of content do its work. This seems to be something like what Sider has in mind. But unless the assignment of meaning is somehow constrained by use, this would evidently result in every word of Ontologese having the same meaning (so that, comically, Ontologese discourse would consist of the endless repetition of the most eligible meaning: The Good The Good The Good The Good...). Yet there seems to be nothing that can serve as the *platitudes* of use needed to constrain the interpretation of a word like 'part', since everything that might play this role (from the axioms of mereology down to judgments about cases) is under dispute in the ontology room.
Moreover, even if we grant that the Ontologese sentence 'composition is unrestricted' is meaningful do we have any reason to think that this sentence is true? If the Ontologese-universalist could tell us, in English, what this sentence means, and it then seemed true to us, that would be a reason. But evidently its meaning cannot be stated in English. (As with Vegas, what happens in the ontology room stays in the ontology room.) Nor can he call upon the standard arguments for universalism in arguing for the truth of this sentence, for the Ontologese-universalist thinks that these arguments are unsound; after all, they have as their conclusion that composition is unrestricted, which he denies (and, ironically, he owes us an explanation of where these plain-English arguments for universalism go wrong). He might instead provide sound-alike arguments in Ontologese, but we have no way of assessing their premises since we do not know what they mean.
I do not mean to suggest that technical vocabulary is intelligible only if it can be translated into one's native tongue. But it is unclear whether there is any way of learning this language, given the aforementioned absence of any undisputed truths of Ontologese to serve as an anchor for figuring out what the rest of the sentences mean. (Imagine trying to radically translate when the tribe cannot even agree among themselves when it is appropriate to say 'gavagai'.) For the same reason, it is unclear how those who claim to speak the language ever learned it.