Hugo Heagren

Many philosophers have presumed something like: the nature of the significance of names, perceptual experiences and singular thoughts is the same. This shared significance is often called ‘content’. Call this presumption 'Unity'. This presumption has gone largely undefended, though it has been relied on in a variety of philosophical work. The purpose of this dissertation is to provide good reasons to think it is true, and to draw out its immediate and significant consequences. This latter goal requires three separate theories: of what has content, of what content is, and of how it is determined what content is had by which things which have it. I devote a chapter to each. Chapter 1 introduces the terms of the project, and situates it with respect to some recent figures. Chapter 2 offers a good reason to believe Unity and also functions as a historical survey. Summarising the literature on these issues since Frege, I divide the examples and cases which arise in these discussions into four groups, and show that each kind of case arises in discussions of all three phenomena. On this basis I offer a methodological argument for Unity. I believe this is the first such argument. In Chapter 3 I argue for a view of what things have content. This is important because prima facie, names, thoughts and perceptual experiences are very different kinds of thing, and it would be surprising if they bore identical relations to the world. In fact I show that names, thoughts and experiences are not the things which really bear content and that the things which do are not so different after all: utterances of names, cognitive tokens of concepts, and ‘object files’. Chapter 4 describes my theory of the nature of the ‘significance’ in Unity: something’s significance is a combination of its referent (if any), and the properties an agent judges to hold of that referent. I show that when properly understood, the account handles well all the existing concerns across the literature. This kind of account is not new. Those who have held similar views (and even those who have argued against them) have all assumed that on such a view, what referent something has is determined by the satisfaction of the properties believed to hold of that referent. This principle has been the basis of many arguments against views like mine and I reject it. But this leaves a gap in my account: if reference is not determined by satisfaction, how is it determined? In Chapter 5 I close with a discussion of what fixes each aspect of the significance I discussed in the previous chapter. My theory is an elaboration of the ideas of ‘causal’ theorists, including Kripke, Evans and Sainsbury.

Suppose a causal theory of reference for utterances of proper names is correct. For an utterance of a name to refer to an object: (1) there must be a ‘baptism’, an event where the name is bestowed on the object; (2) the utterance must be suitably causally related to the baptism. The latter condition has seen significant research, but the act and conditions of baptising are surprisingly under-explored. I offer an account of baptisms. Most previous work has assumed one of two simple pictures: that baptisms are explicit ceremonies or that they are uses of a name with certain relational properties, each motivated by certain paradigmatic cases. In the first two sections of this essay I describe both pictures and explain how they fail. The paradigm cases are telling though, and in the third section I develop an account of baptisms which explains all the cases so far considered. On my account, a baptism of an object with a name is an act of instituting a rule to the effect that later uses of that name will refer to that object.

Practice theories are a genus of causal theories of reference. They claim that the semantic referent of an utterance of a name is determined by features of a practice of using that name to speaker-refer to, or coordinate actions around, a certain object. Practices might extend beyond the utterance, so the reference of some utterances is determined by future events. This entails no commitment on when facts about these future events are themselves determined; one might say they are determined as they happen (and are indeterminate beforehand) or that they are determinate at all times (including beforehand). The practice theory entails unacceptable consequences on either view. On the first, some utterances will have their referent determined once after some future events, then determined again, differently after further events: one utterance may have different semantic referents at different times. This entails similar, unacceptable, consequences for truth and epistemic status. On the second, there are cases of systematic and consistent semantic error which are indistinguishable from cases of systematic success. These are common and imperceptible enough that for all we know, this phenomenon may be very widespread, suggesting we should be skeptical of our knowledge of the semantics of our utterances.

Vague names, like “Everest” and “Belle Epoque” seem to refer to objects without clear boundaries. Supervaluationism claims that this vagueness is a feature of language, not of the objects referred to; vagueness in names is just ambiguity between many possible referents. This general idea admits of two more specific versions. Both give similar treatments of standard uses of vague names, but have very different results for other cases, such as reference achieved by descriptions including mentioned names. Considering two examples, I show neither variant of supervaluationism can account for the truth of all types of sentences about those names themselves. If I am right that these types exhaust supervaluationism, the theory is shown to be false. This problem closely resembles others in the super-valuationist literature about disquotation failure for truth. Treatments of vague truth and vague reference come apart though, and I show that the two problems are different enough that none of the popular solutions will succeed for reference. I consider—and reject—several specific objections, and two more general ways to recover supervaluationism following my arguments. I conclude that supervaluationism is at best a useful formalism for some kinds of vagueness reference, but fails as a general account.