Grammaticalization of actional values in Archaic Latin

This paper aims at showing how prefixation does not modify systematically the actional value of Latin verbs. Scholars of historical grammars of Indoeuropean languages and traditional studies about prefixation agree with the idea that in Latin, and in all Indoeuropean languages, preverbation results on actional value of predicates: Delbruck 1897 argued that prefixes perfectivize verbal meaning since they indicate its accomplishment; Meillet and Vendryes 1924 claimed that the prefixe focuses a specific point of the process described by the non prefixed verb; van der Heide 1934 believed that Latin preverbs express the accomplishment of the process denotated by the non prefixed verbs. More recently Lehmann 1995 supposed that the semantics of Latin preverbs is thoroughly irregular and unpredictable; Haverling 2003 claimed that Latin (early Latin in particular) has a rich and complex system of verbal prefixes which are used to grammaticalize the opposition between telic and atelic actionality; Romagno 2003 goes beyond: she has examined the overall set of Latin preverbs which do not have a syntactic counterpart in corresponding prepositions (ad, cum, per, ex, etc.) dropping out prefixes like re- and dis- that do not have any corresponding preposition. The working hypothesis this paper is actually formulating arises from the semantic and contextual exploration of re- prefixation in archaic Latin. A text corpus of comedies by Plautus (c. 254–184 BC) has been investigated and 84 re- prefixed verbs have been selected. Bases are mainly represented by accomplishment predicates; nevertheless there is no lack of activity predicates, achievement predicates, state predicates. Re- prefixation does not entail a systematic modification in the actional value of the prefixed verb: cito [ACTIVITY] → recito [ACTIVITY], iacio [ACHIEVEMENT] → reicio [ACHIEVEMENT], ambulo [ACTIVITY] → redambulo, [ACTIVITY], sto [STATE] → resto [STATE]. This fact redefines the previous hypothesis: if re- does not entail a modification in the actional value but it works on lexical semantics of the base, it seems likely that the grammaticalization path of the prefix has not been accomplished yet and that, at the level of the present analysis, it could preserve its own full lexical meaning. I will try to demonstrate that this full lexical meaning is widely predictable.