Two kinds of verbal roots in Latin: evidence from thematic vowels and word-formation processes

This paper aims at tackling under a new perspective a crucial asymmetry of Latin verb morphology, that is the distinction between stem and root in different kind of word formations. We do not concentrate here on the inflectional domain directly, but we address the issue taking in consideration two different ways of encoding agentivity in nominalizations, namely we will compare the -tor formations with the so called synthetic or root-compounds, in order to identify which rules select either the bare root (ponti-fek-s) or the perfect participle stem (cap-tor, lauda-tor) in different morphological environments. The theoretical framework we adopt is the Nanosyntactic one, which will allow us to account for the morphological operations according to a constrained set of principled rules. In § 2 we present the crucial properties of agentive nominalizations realized with -tor as well as those of root compounds, focusing on the different conditions which trigger the use of the root morpheme or the perfect participle stem; in § 3 we introduce the basic tenets of the Nanosyntactic framework; in § 4 the morphological properties of Latin roots and thematic stems are accounted for under the nanosyntactic principles; finally in § 5 we draw the general conclusions and point to the open issues.