Building a corpus of the metalanguage of english linguistics 1500-1700: methodological issues

This paper focuses on the methodological issues involved in the selection of sources for a corpus-based study of the English metalanguage that was created to analyse and compare, appraise and classify, teach and learn the vernacular languages of Europe between 1500 and 1700 (MetaLing corpus). Source selection is in fact a critical aspect presenting multiple challenges, since, in a period predating the establishment of comparative philology and linguistics as academic disciplines, language-related discussions are to be found not just in grammar books and language manuals, but in works pertaining to different fields and presenting a large variety of aims. Building on previous research and combining lexicographical analysis and corpus linguistics, the aim of this paper is to explore the potential of (semi-)automated searches of online digital resources for the retrieval of underexplored or non-canonical texts. Integrating what may be conceptualised as bottom-up (lexis to subject) approaches, with top-down (subject to lexis) approaches, the results indicate that onomasiological-oriented approaches have the potential to make terminological blind spots emerge, and that the combination of different onomasiological and semasiological oriented approaches interfacing with the same corpus helps overcome the limits of each individual approach.