A Typology of Aspect-Actionality Interactions

要旨

The present work deals with the topic of aspect-actionality interactions from a crosslinguistic and, more specifically, functional-typological perspective. It investigates how the existence of obligatory grammatical aspect shapes the actional classification of languages, and compares such classifications across languages. The interaction is investigated by adopting the concept of aspect-sensitive classes, i.e. actional classes established via interactions of actional meanings and obligatory aspect grams that are available in individual languages. In that sense, the main goals of this crosslinguistic investigation were to establish which actional meanings are expressed by a selection of (inflectional) aspect grams, then to establish what combinations of actional meanings inherent to verb senses are attested through interactions with aspect, and finally to call attention to crosslinguistic variation in membership of thus established actional classes. This investigation adopted a functional-typological approach to. The approach is primarily based on the works of Sergei Tatevosov, Walter Breu, and other authors working within the bidimensional approach to aspect-actionality interactions. Accordingly, it is assumed that the number, structure and membership of actional classes are subject to crosslinguistic variation. Still, this assumption does not preclude the existence of a universal set of semantic parameters that could be used to characterize actionality. These semantic parameters are called actional meanings (building blocks, primitives) and they are the components of attested actional classes. As this is a novel approach, it required a broad reexamination of the existing mainstream approaches, which, for the most part, belong to the formal semantic tradition. This resulted in, on the one hand, calling attention to the language-specific properties of many elements of the dominant mainstream approaches to actionality, and, on the other, incorporating relevant findings of formal semantics into the approach adopted in the present work. In this investigation, five actional meanings are posited: state (φS), process (φP), transition (τ), multiplicative process (m) and quantum of a multiplicative process (q). These meanings, when interacting with grammatical aspect grams of individual languages, give rise to a large number of aspect-sensitive actional classes. However, only some of these classes can be considered crosslinguistically relevant, based on two criteria. Namely, that they consistently occur across languages, and that they have a large verb membership. Based on these criteria, this investigation included eight actional classes: total states, plain activities, achievements, inchoative states, ingressive activities, accomplishments, two-phase verbs, and multiplicative activities. The properties of these classes were investigated based on their interactions with the obligatory aspect grams of 16 typologically diverse languages: Adyghe (Northwest Caucasian), Bagvalal (Nakh-Dagestanian), Belhare (Sino-Tibetan) Cayuga (Iroquoian), Chipewyan (Athabaskan-Eyak-Tlingit), English (Indo-European), French (Indo-European), Japanese (Japonic), Karachay-Balkar (Turkic), Aredesen Laz (Kartvelian), Maltese (Afro-Asiatic), Mari (Uralic), Nyakyusa (Atlantic-Congo), Spanish (Indo-European), Mishar Tatar (Turkic), and Tlachichilco Tepehua (Totonacan). These languages feature a variety of aspect systems. The most common is the perfective-imperfective aspect system, attested in ten languages (Adyghe, Bagvalal, Chipewyan, French, Karachay-Balkar, Aredesen Laz, Mari, Spanish, Mishar Tatar and Tlachichilco Tepehua). One language, English, features the progressive-nonprogressive aspect system. Five languages (Maltese, Belhare, Cayuga, Japanese, and Nyakyusa) each feature an idiosyncratic system, which cannot be reduced to either of the former two systems. Given this, a method of comparison of different aspect systems was needed because such a method i

主題