Verbos de elocução em português: um estudo descritivo com base em grandes corpora e motivado pela linguística computacional

In this article, we present the results as well as the procedures of a wide descriptive, corpus-based study on reporting verbs in Portuguese. The motivation for this research comes from a task carried out in Computational Linguistics – quotation extraction. In order to obtain reporting verbs used in Portuguese, our starting point was a list of translations of the verb “said” in literary contexts. Then, by using large monolingual corpora, we searched for lexical-grammatical patterns that characterize this verb class, in an attempt to broaden the list. Through this methodology, we identified 293 reporting verbs and distributed them amongst eight general patterns in which these verbs are typically found. The comparison between our research and other studies in Portuguese, English and French that investigate reporting verbs attests the success of the methodology that was developed