Being Something: Prospects for a Property-Based Approach to Predicative Quantification
- Author(s): Rieppel, Michael Olivier
- Advisor(s): MacFarlane, John;
- Campbell, John
- et al.
Few questions concerning the character of our talk about the world are more basic than how predicates combine with names to form truth-evaluable sentences. One particularly intriguing fact that any account of predication needs to make room for is that natural language allows for quantification into predicate position, through constructions like `She is everything one might hope to be: healthy, wealthy, and wise'. A natural and initially plausible view has it that predicates function to denote properties, and that such predicative quantifiers accordingly quantify over properties. This approach faces immediate difficulties, however.
After all, properties are objects of a certain sort, quantified over in nominally quantified sentences like `Steel has some properties aluminum lacks'. Indeed, properties seem precisely not to be what predicative quantifiers quantify over: although happy is something Oscar is, the property of being happy surely isn't. Further, if predicative quantifiers quantify over properties, they presumably carry ontological commitment to properties. But for it to be the case that there is something Oscar is, it suffices that he be happy, or morose, or loquacious. And surely the claim that Oscar is, say, loquacious does not commit us to properties. The vehicles for such commitment, it seems, are once again nominally quantified sentences like the one given above.
I begin by asking what a nominalist alternative to the property-based approach might look like, and proceed to develop what I term the Ockhamist Account on the nominalist's behalf. This Ockhamist Account, however, derives its primary appeal from the apparent shortcomings of the property-based view. The remainder of the dissertation investigates whether the problems that initially appear to beset the property-based view do indeed undermine it.
I argue that first worry mentioned above doesn't force us to deny that predicative quantifiers quantify over properties, but should rather lead us to recognize that an expression has semantically relevant features beyond what it denotes. In particular, I argue that we can explain the truth conditional difference between pairs like `Oscar is happy' and `Oscar is the property of being happy' by distinguishing the semantic relations predicative and nominal expressions bear to their semantic values: whereas nominal expressions like `the property of being happy' refer to properties, predicates express or ascribe them. This lets us construe the difference between predicative quantification and nominal quantification over properties not in terms of what is quantified over, but in terms of the semantic relation the bound variables bear to their values.
I next turn to substitution failures involving adjective nominalizations, like the one exhibited by `Wisdom is admirable' and `The property of being wise is admirable'. As I show, once we allow that the semantically relevant features of an expression aren't exhausted by what semantic value it has, we needn't take such cases to demonstrate that `wisdom' fails to have the property of being wise as its semantic value, any more than we need to deny that `happy' has the property of being happy as its semantic value.
The proposal also offers us a compelling response to the problem about ontological commitment. I argue that the objection conflates the question of whether a given quantifier quantifies over certain things with the question of whether the expressions into the position of which it quantifies refer to those things. The view I defend lets us disentangle these two issues, and shows that we can construe predicative quantifiers as quantifying over properties while still acknowledging that the expressions into the position of which they quantify fail to name, or refer to, those properties.
I conclude by considering the problem that Russell's Paradox poses for the property-based view. I canvas various strategies for responding to the threat of the paradox, but also register some reservations about each. I also offer some reasons for thinking that the Ockhamist Account is not entirely free from difficulties in this area, however. The problems raised by Russell's Paradox therefore point to the need for further refinement on both ends, rather than showing that one account must be abandoned in favor of the other.