Aspect in Russian and Turkish: Semantics and Pragmatics

The multitude of readings of the Russian aspects – particularly the ipf (Table 1) – and the comparison with aspectual usage in the respective English and Turkish equivalents raises the question of how to account for those different readings and the cross-linguistic differences. This paper deals in particular with the following questions: is there an invariant meaning to be found in the Russian pf and ipf aspect, or are we dealing with grammatical polysemy? Are there any regularities in deriving the different readings – language-internally as well as cross-linguistically? What is the common denominator that justifies the subsumption of certain phenomena under the heading of ‘grammatical aspect’? This paper will argue for a so-called ‘selectional theory’ (Bickel 1996; Johanson 1971; 2000) of aspect in combination with well spelled-out pragmatic principles in the sense of Levinson’s heuristics (Levinson 2000). This approach has certain advantages: it captures the crucial assumption that lexical content and aspect (‘situation aspect’ and ‘viewpoint aspect’, Smith 1997) strictly have to be separated, and it provides a cross-linguistic skeleton that allows for systematic language-specific derivation of aspectual readings and for systematic explanation of the cross-linguistic differences. Within a selectional theory of aspect (Figure 1) phases and boundaries constitute the decisive units accessible for aspectual selection by special markers like the English progressive -ing, the Turkish markers iyor(du) and -miş, and the Russian pf aspect. Assumed is a tripartite event structure, consisting of a preparation phase (dynamic phase φdyn), a culmination point (boundary τ) and a consequent state (static phase φstat). Selection is to be understood as making visible a relevant part of an event and making it accessible to truthconditional evaluation at a certain interval of time (‘točka otsčeta’/TO; Padučeva 1996) that may be located after that relevant part of the event or within that part of the event (retrospective/synchronous; Padučeva 1996), thereby distinguishing different groups of aspectual readings (Table 1). Selecting a certain part of an event does not mean dropping or cutting of the rest of the event, rather selection means ‘asserting’ and leaving the other parts of that event for presupposition or implicature. Verbs are classified in terms of combinations of phases and boundaries (Johanson 2000; Bickel 1996), that may be part of the semantic representation of the verb, emerge as the result of VP-composition, or may be pragmatically induced. So the question of class-affiliation of verbs to one of the Vendlerian classes becomes irrelevant, at least with regard to issues of grammatical aspect. The advantage of such an approach is that ‘coercion’ or ‘recategorization’ can be captured by simple composition – semantically or pragmatically achieved. No combinations are excluded a priori, as appears to be the case within Vendlerian approaches that classify verbs according to their internal temporal properties and exclude, e.g., the combination of the progressive with stative verbs. The same holds for the argument structure of the verb that is fixed only during interpretation and not in the lexicon. This assumption avoids the difficulties of Vendlerian approaches with classifying verbs like ‘eat’ and ‘eat something’ (cf. Marten 2002 who argues within the framework of dynamic syntax for an underspecified representation of verbal argument strucuture that is to be fixed only during interpretation). Figure 1 shows only the marked members of the aspectual categories in the respective languages. This does not imply that the unmarked members are semantically vacuous. Rather, they are sense-general (Atlas 1989), i.e. they have a definite meaning, but this meaning is neutral insofar as it is neither identical to one of the readings nor to the sum of them. Accordingly, theses forms are instances of underspecification that receive the appropriate specific reading in the course of interpretation. This serves as an argument against grammatical polysemy and strongly suggests that semantics alone is not enough in deriving aspectual readings, and that some pragmatics is needed, pragmatics being understood as “taking context into account in a principled way” (Bunt 2000: 25).