Development of verb compounding in Tariana and loan translation
Verb compounds in Tariana and in East-Tucano languages show striking similarity in all ways but productivity. While verb compounding is extremely widespread in East-Tucano, in Tariana it is limited to a few cases discussed here, and in §4.
Verb compounding in Tariana appears to be expanding, and to involve more and more verbs as the result of grammatical calquing (loan translations). Verb roots get spontaneously used as second components of verb compounds where they follow the fully inflected verb. The loss of prefixes follows the general tendency to lose non-Tucano categories in Tariana and to conform to a general suffixing tendency of the Tucano type.
An example of such an expansion is the way-yena ‘little by little’ was used by my consultants. This enclitic (loan translation of Tucano tiha ‘do little by little’) occurs with a very limited group of verbs (see §4.3). It was used spontaneously with the verb-nawa ‘stop raining’ (40); then the speakers remembered that it can be used with ‘burn’, ‘go out’ (fire) and ‘speak’ (of a baby). Its use with other verbs — ‘go down’ (of a river in dry season), and ‘eat’ (real food: of a baby: 13) — were accepted only after one speaker, Olívia Brito, conscientious of the necessity of teaching me as much as possible, reminded the others of a corresponding construction with Tucano tiha ‘do little by little’ (14).
Note that parallel grammaticalization patterns for the development of enclitics are not found in other Arawak languages of the region.
This suggests the following scenario for the development and grammaticalization of verb compounded structures in Tariana.
The second verb in two-verb sequences became fused with the first one, losing its cross-referencing prefixes under the influence of East-Tucano patterns. These sequences were of different kinds, and they probably gave rise to different classes of verbal enclitics.
Reanalysis of verb sequences containing verbs with more ‘basic’ semantics such as-sita ‘manage; be able to’ and whyume ‘be last’ resulted in the creation of aspect and switch-reference markers. This reanalysis helps to fill in the gaps arising from structural differences between Tariana and East-Tucano verb structure and to express categories absent from North Arawak but important in Tucano — e.g. aspect and switch-reference. For instance,-sita was used to match East-Tucano completive (Tucano-toha: West 1980:56; see below §4.3.2 on-yena ‘little by little’).
Other verb sequences consisted of quite idiosyncratic combinations where a given verb could be used with just a couple of verbs, e.g. ‘cover the whole surface of something’, or ‘greet wagging one’s tail (for an animal)’, or ‘fall by being unstuck’. These idiosyncratic combinations are comparable to the way incorporation operates in, say, Boumaa Fijian (Dixon 1988:227). Most verbs can incorporate only their prototypical objects, and some object-incorporated forms even acquire an idiomatic sense.
The second component gradually started losing its verbal status and got reanalyzed as an enclitic. Its combinatory possibilities increased (which agrees with the idea of ‘gain’, and not necessary ‘loss’ in grammaticalization; cf. Heine, Claudi and Hünnemeyer 1991:110–11, and also with the generally accepted correlation between productivity of a morpheme and its grammaticalization). It was still used as a modifier because it still retained its independent stress and did not undergo any phonological reduction. The original verb was often quite rarely used anyway; with the development of an enclitic and in the context of general obsolescence of the language it tended to fall into disuse. This is a scenario for the development of Aktionsart enclitics out of rather rarely used verbs with ‘obscure’ semantics. Possibly, many if not all of the semantically idiosyncratic enclitics (discussed in the next section) could have been calqued directly from Tucano compounds — see §§4.1 and 4.4.
The difference between language change in ‘healthy’ and in endangered or obsolescent languages very often lies not in the sorts of change, which are typically the same (Campbell and Munztel 1989). Rather, it lies in the quantity of change, and in the speed with which the obsolescent language changes. As Schmidt (1985:213) pointed out, “one distinguishing feature of the Dyirbal death situation is that vast amounts of change are compressed into a short timespan of about 25 years”. In the case of Tariana, the processes of grammaticalization of compounded verbal roots — which presumably had been taking place for a while during East-Tucano and Tariana coexistence in the same area before Tariana actually became endangered — are sped up by increasing loan-translations from Tucano in the context of obsolescence of Tariana and the reduction of spheres where it is used.
‘Direct calquing’ or loan translation and subsequent grammaticalization of the resulting structures has brought about an ‘unexpected’ phenomenon: rather than losing morphological markers as the result of language obsolescence, Tariana appears to be gaining new morphology (see the discussion on ‘gain’ rather than ‘loss’ of morphological complexity in obsolescent languages in Dorian forthcoming).
