28. Apr 2026
Lieven Danckaert
CNRS (UMR 7023 SFL)/Université de Lille
Voice morphology and clause structure in Latin
Abstract: Latin non-active verb forms like the passive amābatur ‘love.IPFV.NACT.3SG’ and the deponent sequentur ‘follow.FUT.NACT.3PL’ have attracted scholarly attention because of an apparent violation of the universal functional sequence (Calabrese 1985; Zyman & Kalivoda 2021). Specifically, assuming that the so-called ‘-r morphology’ characteristic of non-active imperfective verbs expresses the category Voice, at first sight it appears to be the case that in fact of sitting in their expected position between the lexical root of the verb and Tense, the non-active suffixes appear to the right of those expressing tense. From a comparative point of view, this state of affairs would indeed be surprising (Bybee 1985; Cinque 1999).
In this talk, I defend a variant of an alternative approach, which treats both the active and non-active endings as contextual allomorphs expressing person and number differentially, depending on the voice properties of the host clause. Specifically, I argue that the alternation between the active and the non-active inflection series (i) is governed by a phonologically empty Voice head which occupies its canonical place above vP and (ii) that the distance between the conditioned allomorphs and the conditioning category poses no locality problems (in the sense of Embick 2010). In addition, assuming a correlation between morphology and syntax along lines of Chomsky (2013, 2015), I propose that the fact that Latin voice morphology has no true in situ exponence has consequences for the structure of the Latin clause which go far beyond what first impressions would suggest, both synchronically and diachronically. Topics touched upon include subject placement, the morphological realization of anticausatives, split intransitivity and the structure of past participles occurring in perfective periphrases with the auxiliary esse ‘be’.