Complex verbs, simple alternations: valency and verb classes in Jaminjung.

This chapter investigates valency patterns in Jaminjung, a language of the small Jaminjungan (or Western Mirndi) subgroup of the geographically discontinuous Mirndi language family of Northern Australia. Jaminjung is a typical “Non-Pama-Nyungan” language in that grammatical roles are indicated both by case markers and by pronominal indices (for subject and object) on inflecting verbs. Its most interesting property from the point of view of investigating valency, however, is the prevalence of complex predicates, linked to the existence of two distinct “verbal” parts of speech, Inflecting Verbs (IVs) from a closed class, and Uninflecting Verbs (UVs) from an open class. In complex predicates, IVs and UVs jointly determine the argument structure of the resulting expression. This organisation of the lexicon therefore has far-reaching consequences for the grammatical manifestations of valency in Jaminjung (and in a comparable fashion in other languages in a linguistic area in Northern Australia with a similar part of speech division). It also raises the question of what should be considered as “verbs” for the purposes of a comparative typological investigation of valency and verb classes. Predicate classes based on coding frames can be established only for IVs and complex predicates, since only they can form predicates of independent, finite clauses. However, as demonstrated in some detail in this paper, the main strategy of achieving valency change – in the absence of derivational mechanisms – is a change of IV in complex predicates, with an accompanying change in coding frame, while leaving the UV constant. As a result, covert predicate classes based on alternations (or absence of alternations) can be established mainly for UVs. Remarkably, the classes of UVs established by this diagnostics partly correspond to predicate classes attested cross-linguistically, for example predicates of externally caused change of state, participating in a causative-inchoative alternation, predicates of internal causation (rarely alternating), predicates of manner of motion (exhibiting an S = A alternation), predicates of posture / position (participating in a causative-inchoative-stative alternation), and predicates of manner of manipulation or impact (mostly non-alternating). In addition, the same method also revealed predicate classes corresponding to rather language-specific (possibly culture-specific) lexical fields, such as that of manner of application of heat (participating in a causative-inchoative pattern) or direction of gaze (participating in an S = A alternation). Finally, the particular partition of the verbal lexicon in Jaminjung also leads to a skewed distribution of the major valency classes between the two parts of speech: Overall, there are many more monovalent than bivalent or trivalent UVs, owed to the existence of large classes of UVs of position/configuration, externally caused change of state, change of location, manner of motion, and internal causation, which are all monovalent. Conversely, there are many more transitive (20+) than intransitive (5) or ditransitive (2) IVs. In complex predicates, while the valency of UV and IV can of course also be identical, it is often the IV that contributes an additional Actor (in causative combinations) or Undergoer argument. This can be seen as a manifestation of Economy, since it greatly increases the combinatorial potential of UVs, i.e. of the predicates from an open class. Thus, caution needs to be exerted when determining a position on a “transitivity hierarchy” for Jaminjung since the position will be radically different depending on whether UVs, IVs, or complex predicates are taken as “verb” equivalents.