Knowledge Bank

Kakari-musubi (“KM”) is an early Japanese focus construction that highlighted one constituent of a state or event (one term of an open proposition) with one of five emphatic kakari particles, and then concluded the sentence with a specially inflected predicate representing that situation/ proposition. The focused part is the kakari ‘onset’; the final, specially inflected predicate, its musubi ‘closure.’ These features make KM a cleft-like “term focus” (Heine and Reh 1984) construction. While researchers have developed characterizations of KM’s variants based on the semantics and pragmatics of the different kakari particles, similar examination of the musubi’s two differently inflected clause forms—adnominal rentaikei or non-finite izenkei ‘realized form,’ depending on the emphatic particle—has not kept pace (Frellesvig 2010). Semantic and pragmatic similarities between the two musubi forms have been identified (e.g., Quinn 1987), and they have been related derivationally (e.g., Ohno 1955; Martin 1987; Unger 1993; Hayata 1998; Shinzato and Serafim 2013; Whitman 2016). There is some consensus as to why the construction’s agreement rule for four of the five emphatic particles stipulates an adnominal/rentaikei-inflected predicate. However, why the fifth and most emphatic particle, koso, required the izenkei’s non-finite inflection has eluded a convincing explanation. This paper attempts to reconstruct one, by identifying or re-examining evidence that is often indirect, and bolstering it—judiciously, it is hoped—with relevant precedents and principles cited from today’s Japanese and other languages. A variety of evidence indicates that the rentai and the izen inflected clauses were nominals that indicated the speaker’s presupposition of the situation they referred to. From Japan’s earliest documents, rentai-inflected, ad-hoc clause nominals represented a range of individuated discourse entities, from agents, undergoers and experiencers, to events, times, and places, with a deictic ‘the X that [predicate]’ referentiality. It was in this role, we suggest, that rentai-inflected clause nominals were recruited for use as the presupposed part, the musubi, of the dominant type of KM: as a postposed, utterance-final topic, in a syntactic move well attested in early Japanese. This produced statements of the following sort: ‘It’s X—the one that [predicate].’ Thus positioned, in KM the rentai functioned initially as the ad-hoc, entity-referring nominal that it was in other contexts. By contrast, izen-inflected clauses outside KM are nowhere attested as ‘the one that [predicate]’ nominals, and therefore could not have been recruited in that role. All (non-KM) uses of the izen form in OJ served a different but nevertheless topic-like role: they circumstantially grounded other situations, sometimes concessively (‘although’) but most frequently as the presupposed protasis in a ‘given/when(ever) X, Y’ clause-linking construction. So we suggest that the izen-inflected form was recruited in its capacity as a conditional protasis (‘given/provided [predicate] being the case, …’) that stipulated the circumstances in which a second situation (its apodosis) was found. The consequent (apodosis) situation introduced with this provisional (Martin 1975, 1987) type of conditional was commonly non-volitional, “found,” and thus presented, inevitable. Postposed to a designative predication like X ko so ‘X, it’s this!,’ a provisional izenkei would have yielded a biclausal ‘X, it’s this—given P/as far as P is concerned.’ thereby indicating that ko so’s identification of X was categorical for P. Thus, the construction’s category-exhaustive focus was originally guaranteed by a postposed izen-inflected provisional and an emphatically designative apodosis. As it grammaticalized, and the language’s productive provisional became PIZ-ba, PIZ came to be regarded as just another musubi type, referring to the situation presupposed in the process of identifying X with koso. That, in turn, reciprocally, became the construction’s primary indicator of category-exhaustive, epitomizing meaning. On the way to the monoclausal/X koso PIZ/ KM construction, ko so ‘it’s this’ also underwent decategorialization and subjectification. This suggests that the received origin theory for this KM type is anachronistic.