Peter Langland-Hassan is an Assistant Professor of Philosophy at the University of Cincinnati.
Abstract: While few in cognitive science hold that humans think in a natural language, a number of theorists have proposed that language use—in the form of inner speech—enables an important kind of self-reflective or “metacognitive” thought. Their core idea is that inner (or “subvocal”) speech, like speech aloud, serves to express one’s thoughts, albeit to an audience of one. By attending to the “inner expressions” of one’s thoughts, one is able to make those thoughts the objects of one’s own attention. A key tenet of these “Expressionist” views (and of common sense) is that inner speech carries the contents of the thoughts it is assumed to express; that, like one’s speech aloud, inner speech episodes means at least roughly what one’s corresponding thoughts mean. However, I will argue that inner speech episodes lack the kind of contents that would be needed for them to have meanings comparable to those of one’s thoughts or to one’s speech aloud. Indeed, inner speech “utterances” lack the sort of contents required for them to constitute expressions of one’s thoughts at all. Empirical results and larger theoretical considerations both suggest the contrary, counterintuitive position that inner speech represents only word sounds and, derivatively, words and sentences. That is, with inner speech we think about sentences, not with sentences. Consequences are considered for “expressionist” views of metacognition, and for recent debates concerning the phenomenology of thought.
