Ian Lancashire

Semantic drift in Shakespeare, and Early Modern English full-text corpora

How can we detect Early Modern English semantic deviation, which Manfred Gorlach describes as “a difficult and largely unsolved problem for the history of the English lexicon”? Lexical and phrasal neologisms stand out from the simplest word-lists, but readers must understand the meaning of words in context before recognizing semantic drift. New diachronic corpora such as Early English Books Online (EEBO) and Lexicons of Early Modern English (LEME), two web-based corpora, could help, were their vocabulary to be lemmatized and associated with specific senses in the Oxford English Dictionary. Only LEME tries to do so, however, and its more than half a million word-entries are not uniformly analyzed yet. Three instances of semantic deviation in Shakespeare’s plays serve to illustrate the challenge: the “pricking” of a witch’s thumb in Macbeth (a means of torture), the villain’s name “Aron” in Titus Andronicus (a new starch obtained from a weed of that name), and the term “acting” (for “enacting”) as used by Brutus in describing his dream of a conspiracy to assassinate Caesar. Because these innovative senses are undocumented in this period, because no monolingual English dictionaries survive from it, and because these instances of drift disappeared soon afterwards, they are hard to find. Manfred Gorlach’s problem will remain in force for some time to come if we have to rely on literary text analysis to locate semantic deviation.

Corpus linguistics, the humanities, and virtual organizations

Researchers in the Humanities and Social Sciences (HSS) study social groups (of which internets are striking new examples), the new languages and literatures that grow in cybertext culture, and the beliefs and ideas that individuals submerge in their texts. HSS research thus has implications for our understanding of as diverse subjects as digital modes of discourse, mental health, and the ways in which individuals think and access information online. HSS curiosity-based research into the nature of texts, conceived and analyzed computationally, can lead to valuable advances in understanding how virtual organizations can operate effectively. Two examples from Chaucer and Shakespeare research show why. Corpus linguistics, semiotics, and other social- and text-analysis methods can also tell us things about the individual mind that no data-warehousing shopping profile can. Lack of a research infrastructure has prevented HSS, as Marc Renaud, President of the Social Sciences and Humanities Research Council of Canada, has said, from focusing on new technologies to shape the research questions that can be asked, and to supply the data used to answer those questions. New government infrastructure programs, and new openness on the part of the sciences and the private sector to partner with HSS, offers a hope that the current separation of these two research cultures can be bridged.

Literary expert systems

Linguists and historians of language for the past twenty-five years have developed systematic treatments of phonetics, vocabulary and grammar, even as semioticians and literary critics have Pioneered in the study of abstract models of thought. Over the same period, computational linguists have contrived programs analyzing morphology, syntax and semantics to prepare text for similar models that computer scientists call “knowledge representation systems.” Roth humanist and computational linguist do research on the fundamental particles of language. Some engage in experimental work with large quantities of data; others build theoretical models to explain this data; and whereas computer scientists work on NLP, natural-language understanding systems that might diagnose bad writing or provide means for people to “talk” to data bases, humanists discuss harnessing computational techniques to the analysis of style, prosody and meaning.