Paper short abstract:

In this paper we compare synchronic and historical theories of Japanese tonology, testing both against actual data and examining exceptions and problem cases. We discuss what it means for a historical account to have explanatory power, and how it can help us understand inter-dialectal relationships.

Paper long abstract:

Standard Tōkyō Japanese is a pitch-accent language; that is, it is tonal, with a restriction of at most one tone per word (called 'the accent'). When nouns are compounded together, the position of the accent in the resulting compound is notoriously tricky to predict, seemingly occurring in patterns that depend on the components, but with plenty of exceptions and variation. Many accounts of this phenomenon restrict themselves to synchronic linguistics, trying to discover patterns from the point of view of the ahistorical language system as acquired by the individual native speaker. These accounts make use of data such as recent loans and nonsense words to uncover the underlying accent-marking tendencies, and postulate any exceptions to be lexicalised entries.

In this paper we explore a historical explanation of compound noun accent, based on Middle Japanese tone materials. We compare the descriptive reach of the historical theory with that of synchronic accent rules, by testing both theories against actual data and giving special attention to cases where the data defies the predicted forms. We discuss briefly what it means for a historical theory to have explanatory adequacy, and how that can be relevant for the synchronic goal of understanding the individually acquired system. Finally, we explore the predictive and explanatory power of the historical theory to provide a single explanation encompassing all Japanese dialects, testing it against novel fieldwork data.

The synchronic rules will be based on proposals by Haruo Kubozono, and the historical reconstruction on the model developed by Ramsey/De Boer.