A Polysynthetic Language in Contact: The Case of Ket
Ket is one of the most enigmatic polysynthetic languages in North Asia. The majority of structural features complicating a clear-cut typological analysis of Ket are due to the long-term contact with the languages of a radically different type that resulted in a peculiar process of structural mimicry (or ‘typological accommodation’ in Vajda’s(2017) terms).The mimicry is most evident in the verbal morphology, which is traditionally regarded as almost exclusively prefixing. While this is true for the oldest layer of verbs with the main lexical root in the final position, Ket’s most productive patterns of verb formation clearly imitate suffixal agglutination typical of the surrounding languages by placing the main lexical root in the initial position with the rest of morphemes following it.
The aim of this paper is to demonstrate that this phenomenon is also attested at the syntactic level. Prototypical polysynthetic languages are largely devoid of overt subordination (cf. Baker 1996). Ket, however, signals adverbial subordination by using postposed relational morphemes attached to fully finite verbs. This pattern is common to adverbial clauses in the neighboring languages, the difference being that they attach relational morphemes to non-finite forms only. This functional-structural parallel is likewise attested in relative clauses. The surrounding languages share a common relativization pattern involving preposed participial relative clauses with a ‘gapped’ relativized noun phrase (Pakendorf 2012). This resembles the major relativization pattern in Ket, in which, however, preposed relative clauses are fully finite.To conclude, formation of adverbial and relative clauses in Ket clearly mimics that of the surrounding languages and does not conform to the expected ‘polysynthetic’ pattern. At the same time, Ket resists accommodating a participle-like morphology, which can be connected with the general tendency among polysynthetic languages not to have truly non-finite forms (cf. Nichols 1992)