Looking for concepts in Early Modern English

Abstract

The idea that conceptual meaning in discourse could be identified in constellations of lexical co-occurrences in a particular “universe” of discourse was key in guiding the computational historical semantic–pragmatic work conducted in the Linguistic dna project. The project mapped prominent lexical co-occurrences across the two hundred years of publications in Early English Books Online (eebo-tcp; Text Creation Partnership edition), yielding concept models – constellations of non-adjacent lemmas that consistently co-occur across spans of up to 100 tokens. The goal was to map meaning onto concept models as “discursive concepts”, using encyclopaedic knowledge, pragmatic analysis and context. The first question concerns the effectiveness of making early hypotheses about the discursive meaning of concept models based on the inferred connections between the lemmas in a quad constellation. The second question is whether the meaning of frequent, apparently stable concept models changes upon their closer scrutiny in the discourses they lead us into. A reader familiar with the particular universe of discourse in which these quads occur, and with the social, historical, literary and philosophical traditions, and the context that they occupy, might be effectively primed by their encyclopaedic knowledge to hypothesise this discursive meaning. This paper demonstrates the efficacy of hypothesis building using encyclopaedic knowledge and pragmatic analysis to interpret optimally relevant concept models.