Basic valency orientation and the middle voice in Hittite

This paper discusses basic valency orientation in Hittite, based on the typology proposed in Nichols et al. (2004). Verb pairs usually employed to test basic valency indicate the clearly transitivizing character of this language; a closer scrutiny of intransitive verbs further reveals the existence of a three-fold distinction featuring two intransitive verbs, a basic stative one (or an adjective), and an overtly marked intransitive change-of-state, in addition to a transitive counterpart overtly marked as causative. The high productivity of causative derivation is shown by the fact that morphologically marked causatives are not only derived from stative verbs, but also from telic intransitives and from transitive verbs. In the case of telic intransitive verbs, a minor pattern is also attested, whereby valency alternation is encoded through voice alternation, with intransitive forms inflected in the middle voice and transitive forms in the active. Since neither voice can be considered to be derived with respect to the other, verbs that display this behavior are indeterminate as to basic valency orientation. In spite of the limited extent to which voice indicates valency alternation, this finding becomes more significant when set into the framework of valency alternation in the early Indo-European languages, and sheds some light (or raises more questions) on the original function of the Hittite and of the Indo-European middle voice, a typologically puzzling category.