By Pollet Samvelian
This pattern is reminiscent of split ergativity, and has led Mackenzie (1961), among to analyze the transitive construction in the past tenses as an ‘agential construction’. Mackenzie claims that in this construction, the noun phrase referring to the Agent argument of the verb ‘is in no way equivalent to a Subject, in concord with the verbal form’ (p. 107). If the Agent is not the subject, it follows that the clitic cannot 268 / Pollet Samvelian realize subject-verb agreement, but is in fact the Agent argument of the verb.
That is the reason why Mackenzie uses the term ‘agential suffix’ to designate the clitic in this case. Though it is not explicitly claimed by the author, such an analysis implies that the grammatical subject of a transitive verb in the past tenses is in fact its Patient, or its ‘Direct Affectee’ in Mackenzie’s terms. This would explain the use of verbal personal endings, which realize subjet-verb agreement, in a similar way as with the present tenses.
The main advantage of Mackenzie’s view is that it provides a unified account for each set of personal bound morphemes: the forms in Table (3) are always regarded as inflectional verbal affixes and function as agreement-markers, while clitics realize an argument of the verb and constitute one of the options for argument realization, together with noun phrases (or prepositional phrases) and independent pronouns. In other words, they are bound pronouns, and not agreement markers. Appealing though it is, and supported by historical facts, this analysis faces serious problems.
Texas Linguistic Society IX:
The Morphosyntax of
and Jessica White
(vol. eds.) 2007
and Stephen Wechsler (series ed.)
CENTER FOR THE STUDY OF LANGUAGE AND INFORMATION
What Sorani Kurdish Absolute Prepositions Tell Us about Cliticization
Please read more in the attached PDF file.