When to prefer split-self conceptions

Abstract Self-referential solitude speech invokes diverse conceptions of its speakers in different languages. In unvocalized solitude speech, the speakers are conceptualized as hearing their other selves in Ainu, as directing the speech to their other selves in English, and as holding the speech content in mind rather than directing the speech to themselves in Japanese and Korean. These four languages further differ in the range of pronominal reference to the thinking and speaking self in solitude speech. Ainu prefers second-person self-reference in unvocalized speech, not in vocalized speech, English and Korean encourage or tolerate second-person self-reference in both vocalized and unvocalized speech, and Japanese disfavors second-person self-reference in both types of speech. These cross-linguistic similarities and differences can reflect socio-cultural assumptions and worldviews of the relevant linguistic communities. We explore some relations between each language’s (dis)preference for solitude speakers’ split-self conception and assumptions/worldviews that encourage or discourage that conception.