Paper short abstract:

Utterances that end with the particle –te ‘and’ appear frequently in Japanese discourse. They may be deemed unfinished, but analyses of their formal and functional properties demonstrate that they are complete utterances, constituting several different types of grammatical constructions.

Paper long abstract:

Many utterances in Japanese discourse, as in other languages, often appear to be incomplete or fragmental. This feeling of incompleteness seems to stem from what may be called “sententialism,” or an assumption that the basic structure of an utterance in discourse is a well-formed sentence. However, I argue that many of these seemingly incomplete utterances are in fact not incomplete. As a case in point, this paper examines utterances ending with the particle –te in Japanese.

The particle –te, corresponding to English ‘and,’ is typically used in a sentence like Uchi ni kaette, sugu ni neta. ‘I went home and went to bed immediately.’ That is, it is a clause- or phrase- linking particle. Accordingly, if an utterance ends with –te, it may sound unfinished. However, such utterances are widely used in Japanese discourse, especially in spoken informal conversations. Examining a corpus of spoken discourse consisting of dyadic conversations of Japanese college students, this paper analyzes the formal and functional properties of –te ending utterances and demonstrates that with some exceptions, –te ending utterances, as in the following example, are in fact complete as utterances in terms of form and meaning.

A: Go-kazoku ni nani ka atta no ka, mitaina, kiita n da kedo, sore demo chigakutte. (falling intonation)

‘I asked if something happened to her family members or something, but that’s not it, either, -tte.”

B: (a short pause) Nanka ne, kinō hanashita toki ni, nanka …

‘Well, when I talked with her yesterday, well …’

Furthermore, the analysis identified several types of –te ending utterances, each of which differ with regard to formal properties (e.g. intonation patterns, possibilities of co-occurrence with interactional particles) and pragmatic properties (e.g. mitigation of speech-act force, avoidance of bluntness, stating a reason, making a request). These different types of utterances, I argue, can be regarded as grammatical constructions in different degrees of grammaticalization, with one used as a request (e.g. Matte ‘Please wait’) as most fully grammaticalized.