The Anatolian languages are unique among the Indo-European languages in having a suffix for neuter nouns in the agent position, hereafter referred to as the agentive suffix. There exist several theories concerning the grammatical analysis of this suffix (Melchert 2007). In this article I expand on research by Goedebuure (2013) by testing these theories for all languages in which this construction is attested. It turns out that the agentive was originally a personifying suffix *-ont-, a function still present in Old Hittite and Luwian. This suffix was grammaticalised into a grammatical suffix already in Proto-Anatolian. This suffix could only occur in the common gender nominative. In Neo-Hittite, the construction -ant-s/-ant-es was reanalysed as case endings -anza/-anteš of a new ergative case appearing only in the neuter gender. A similar reanalysis was happening in Lycian. The suffix *-ont- was grammaticalised in order to be able to form neuter agents, which was impossible in Proto-Indo-European. The non-Anatolian Indo-European languages filled this gap by extending the function of the neutral subject/patient ending *-Ø and *-om to the agent function. This shared innovation constitutes an argument in favour of the Indo-Anatolian hypothesis.
|Title of host publication||The Precursors of Proto-Indo-European|
|Editors||Alwin Kloekhorst, Tijmen Pronk|
|Publication status||Published - 25 Sep 2019|
|Name||Leiden studies in Indo-European|