Singular Thoughts

There seem to be good reasons for recognizing singular thought: thoughts that are about particular objects. It seems that singular propositions capture the truth-conditions of such thoughts; that is, propositions individuated by objects and not senses, intensions, descriptions, or even names. But then how do we handle cases where a person regards “Cicero was an orator” as true but regards “Tully was an orator” as false? She seems to believe and disbelieve the same singular proposition. The chapter argues that we need to “unburden” propositions. Beliefs are episodes that have truth-conditions that can be captured by a variety of propositions, and the propositions that “that” clauses refer to don’t capture everything relevant to understanding the belief. I provide some concepts and terminology for implementing these lower expectations for propositions.

Cognitive Focus

The first part of the chapter motivates a unifying alternative to causal theories and description theories of reference and aboutness. The alternative account is built around a claim which the chapter argues brings out the significance for theories of reference and aboutness of the fact that justification is truth conducive: the claim that an aboutness-fixing relation is one which secures what the chapter calls “cognitive focus” on an object of thought. The second part of the chapter argues that a “singular thought” just is a thought made available by a cognitive focus relation, and uses the cognitive focus framework to advance the discussion of descriptive names. An Appendix explains why, though the proposal is a “mental files” proposal, it is better stated without use of this term.

Definite Reports of Indefinites

In certain circumstances, it is possible to report the utterance of an indefinite noun phrase using a definite referring expression, such as a name or pronoun. The referent of the definite must match the individual the original speaker had “in mind” when uttering the indefinite, and the speaker must not have been in a position to use a definite instead (the individual in question must not have been familiar, and so identifiable, to the hearer at the time of the original utterance). To account for this data, the chapter claims that indefinites refer (though truth-conditionally they function like restricted existential quantifiers), but also that coreference is not always sufficient for substitution in an indirect report. (Perhaps surprisingly, the second, less controversial, claim will explain the cases motivating the condition of hearer non-identifiability.) In sum, the data tells us about the semantics of indefinites and the tolerances of indirect reporting.

Varieties of Singularity

What is it for a thought to be singular? The chapter argues that there is no single answer. Singularity in thought is associated with a variety of non-equivalent features. For example, some argue that the object of a singular thought should be something with which we are acquainted, or should involve the exercise of a mental file; or the thought should essentially “involve” its object, or refer to it rigidly. The chapter claims that the main task should be to examine the relations between these various features; there is little interest in trying to determine what a “real” singular thought is.

De Jure Anti-Coreference and Mental Files

The chapter introduces the phenomenon of de jure anti-coreference. Roughly, two representation occurrences are de jure anti-coreferential when they must refer to distinct objects in virtue of meaning. It argues that in contrast to its opposing notion, de jure coreference, it is rarely found in human representational systems. It explains how the Fregean can hope to explain this asymmetry by appealing to senses or mental files. It argues, however, that such approaches, in order to account for dynamic coordination, must ultimately appeal to semantic relationism. This is surprising since semantic relationism is often thought of as an alternative to Fregean semantics.

The Mental Files Theory of Singular Thought

The chapter outlines and evaluates the most ambitious version of the mental files theory of singular thought, according to which mental files are a wide-ranging psychological natural kind, including psychologists’ object-files as a representative subspecies, and underlying all and only singular thinking. It argues that such a theory is unsupported by the available psychological data, and that its defenders may have overestimated the similarities between different notions of “file” used in philosophy and cognitive science. Nevertheless, critical examination of the theory from a psychological perspective opens up promising avenues for research, especially concerning the relationship between our perceptual capacity to individuate and track basic individuals and our higher-level capacities for singular thought.

Content and the Explanatory Role of Experience

It is widely held both that having certain sorts of perceptual experiences can explain one’s ability to have certain sorts of thoughts and that we can use this fact to show that perception and thought differ in certain fundamental ways. Some hold that the explanatory role of experience shows that experiences, unlike thoughts, are not contentful states; others hold that it shows that experiences, unlike thoughts, are nonconceptual. The chapter argues that these theses can’t be established by arguments based on premises requiring experience to play certain explanatory roles. It considers three arguments of this form, which appeal, respectively, to the requirements that experience explain our capacity for singular thought, that it explain the reference of certain demonstrative concepts, and that it explain our ability to learn new concepts. In section 4, it argues that the explanatory role of experience can help decide questions about scope of perceptual representation.

Descriptive Names and Singular Thought

This chapter is concerned with an apparent disagreement between Gareth Evans (1982) and David Kaplan (1989) with regard to the capacities of names introduced into the language by describing their referents. The question at issue is whether such expressions are potential sources of novel singular thought. While Evans’ response is an emphatic “no,” Kaplan’s response is an equally emphatic “yes.” The chapter attempts to resolve (or rather dissolve) this apparent dispute by suggesting that the two philosophers have importantly different phenomena in mind when they talk of “thoughts”—including (and in particular) singular thoughts. Whereas Evans construes thoughts as mental states, Kaplan construes them as semantic contents. Before concluding, it argues for the superiority of the proposed resolution of the Evans/Kaplan debate over that of Francois Recanati (2012), who invokes his Mental File Framework in an attempt to reconcile the seemingly contrary views of Evans and Kaplan.

Singular Thought and Mental Files

This introduction outlines some central challenges to a clear understanding of singular (or de re) thought, and illustrates why the literature has recently turned to the notion of a mental file as a way to theorize it. It sketches three central motivations behind the claim that singular thought is file-based thought. First, file-theorists have stressed that the descriptive information contained in a file does not determine its identity conditions or semantic content. Second, it is sometimes assumed that singular thoughts are devices of de jure coreference in thought and claimed that mental files are the cognitive reality of de jure coreference in thought. Finally, the file-theoretic approach to singular thought may seem to lend empirical respectability to the notion of singular thought. The chapter ends by?introducing some basic questions proponents of this approach must address if it is to fulfill its explanatory aims.

Embedding Language in the World

Direct reference theories hold that nothing beyond reference is carried from speaker to hearer by singular terms. The chapter argues the same is true of common nouns and most other extensional terms such as terms for properties, places, events, and actions. None of these terms carry descriptions, grasp of paradigm property sets, inferential mandates, or anything else to be “loosened” or “tightened” by pragmatic inference. Both thought and language are directly structured by the structure of the world itself, not by peculiarities of the human mind and not by convention. The route from speech to hearer understanding is indirect, passing, typically, through the hearer’s prior grasp of world structure, a structure that hearers may have idiosyncratic ways of grasping. They may have quite different ways of identifying the same thing; that is, different ways of recognizing when new natural or intentional information about the same is arriving at the sensory surfaces.