Contrastive analysis of compound expressions in Portuguese and French annotated corpora

Abstract

Adjectives are attributive modifiers, and they integrate the Noun Phrase as adjuncts of a nominal head N, sharing with it, in Romance, inflectional features. Some compounding models, however, allow an NP with two adjacent nouns N1N2, in which seems to be an apparent modifying relation. Departing from previous proposals (Noailly 1990, Villalva 2003, Dressler 2006), our analysis implies that endocentric N1N2 sequences in Portuguese and French are in fact modified structures, where N2 is close to an “attributive noun” (Martinho 2007, 2013). Indeed, if a random sequence like * “voiture âge” would be clearly unacceptable, attested expressions like “uma criança modelo” or “une voiture bélier” seem, on the contrary, quite clear about the modifying nature of “modelo” and “bélier” respectively, however raising questions about the category, morphology, syntax and semantics of such a modifier. In these “binominal modified constructions”, frequently classified as lexical or morphosyntactic compounds, we will consider, based on a contrastive study of the concordances extracted from the CETEM and FrWAC corpora, that N2 appears to share in both Portuguese and French the syntax and morphology of the adjectival category. Moreover, since N2 meets allegorical properties but lacks extensional ones (Carnap, 1947), we conclude that, in this kind of stereotypical modification relation, shared by all Romance languages (Fabre 1996), N2 modifiers in fact revert to intensional adjectives with relational or quantified interpretation.