|Latin Passive Morphology Revisited
|Article dans des actes
|Année de la conférence
|Nom de la conférence
|Meeting of the Linguistic Association of Great Britain
|Date de publication
|C-COM, Deponency, Derivation, Inflection, Latin, Morphology, Passive
Latin uses distinct morphological active and passive endings. Active endings attach to verbs asso- ciated with active syntax, e.g. transitive verbs. Passive endings attach not only to verbs with passive syntax, but also to so called deponent verbs with active syntax. In addition, deponents display forms usually restricted to verbs using active paradigms, namely present and future participles and gerunds. The deponents’ mismatch between morphological form and syntactic value (Baerman, 2007) shows that the relation between morphological forms and the associated syntax is not straightforward. The traditional approach considers it to be an inflectional process and is followed by recent studies such as (Stump, 2006; Hippisley, 2007). On the other hand, (Bresnan, 1982) states that passives may serve as an input for derivational processes. If one accepts that inflected forms cannot function as an input for derivational processes, this inevitably entails a derivational treatment of passivisation. Other ap- proaches in favour of the derivational analysis argue for lexically specified argument-structure (Sadler and Spencer, 1998; Sag et al., 2003) or propose lexeme formation rules in the sense of (Blevins, 2003).
We propose a novel solution to the active/passive interface and in particular the question of de- ponency using independently motivated approaches. We argue from an inferential realisational point of view that morphological passive is best analysed as an independent inflection class. We claim that, at least for Latin, voice change should be treated as derivation within morphology and propose three arguments supporting this claim. (1) As opposed to inflectional processes, deriva- tion is usually said to operate semantic change in an unpredictable way. (Kiparsky, 2005) shows that Latin passive forms may convey up to five distinct values. These values are lexically specified, i.e. not directly predictable. Thus, the variety of semantic changes observed in Latin data and their resulting un-predictablility argues towards a derivational approach to Latin passives. (2) Treat- ing voice change as a derivational operation also significantly simplifies the morphology-argument structure interface. Lexically specified voice allows for lexically specified argument structures. No separate mechanism is needed to account for passivisation induced argument structure change. (3) We show that satisfyingly modelling the deponents’ behaviour requires the existence of two types of inflection classes (A and B) corresponding to active and passive. Thus passivisation is best repre- sented as the derivational process assigning Class B to a base originally belonging to Class A. The so called Latin “deponent” verbs retrieve most of their forms from subparadigms (zones) of Class B and the additional forms from zones of Class A (Tables 1–3) and are thus heteroclites rather than deponents. However, this analysis makes no predictions as to the existence of deponency in general.
The combination of these arguments favours a derivational account of Latin passives. We provide a PFM fragment (Stump, 2001) for all forms of all types of Latin verbs, incl. (semi-)deponents.