Support verb constructions in medieval Portuguese: Evidence from the CTA corpus

Abstract

This paper conducts a systematic survey of Support Verb Constructions (SVC) in Medieval Portuguese using the Corpus de Textos Antigos (CTA) corpus. SVC involves verb-noun combinations, where the noun serves as the main predicative element, and the verb conveys grammatical values. Limited historical evidence exists for SVC in earlier stages of Portuguese, with previous studies lacking digital access and Natural Language Processing tools. This study utilizes a subset of the CTA, comprising around 0.5 million tokens, annotated for part-of-speech and lemmata. Out of 175 candidate support verbs in Modern Portuguese, 81 were identified in the corpus, totaling 27,645 occurrences. Manual analysis of concordances revealed a little more than 3,000 SVC instances and more than 900 predicative nouns (types), uncovering several linguistic phenomena. The paper delineates the adopted procedure and explores essential linguistic properties of identified SVC in the CTA corpus, emphasizing the significance of leveraging NLP tools for a comprehensive linguistic description of Old Portuguese.