Figures of Speech
As unlikely as it sounds, literary scholar Professor Hugh Craig has enhanced his appreciation of Shakespeare through statistical analysis.
Renaissance literature expert Professor Hugh Craig is a man of letters. But the computational stylist is equally a man of numbers.
Craig is the director of the Centre for Linguistic and Literary Computing and the Humanities Research institute. He has been an advocate of computer-assisted analysis of language in literature since the controversial field began to emerge in the 1980s.
He has devoted decades of research to proving that statistics can help us analyse and appreciate literary texts.
Craig says computational analysis has two applications in the field of literature: it can help authenticate authorship that is unknown or suspected to have been wrongly attributed and it can be used to build a profile of or define a writer's particular style.
"It is still controversial because people in the literary world don't like numbers, they don't trust numbers, and they don't understand how you can do something as banal as counting things in a literary context," he says.
"That is why it is fun; because it does challenge people and threaten some people. As you can imagine, I get in some pretty heated discussions."
Craig's work is based largely on frequency data and has led to several breakthrough findings in regard to Shakespearean works. Using his computational techniques he found that Shakespeare was the likely author of a number of scenes from the play The Spanish Steps that had previously been attributed to the playwright Ben Jonson. The results are presented in his 2009 co-edited book Shakespeare, Computers and the Mystery of Authorship.
He has also established that Shakespeare did not have the wide vocabulary many credited him with."There was a myth that Shakespeare had an extraordinarily large vocabulary, but our analysis shows that he didn't. His talent was in the way he used regular, ordinary words," Craig explains.
"What we did was look at the words he used and the frequency with which he used them and compared that to what other playwrights of the time were doing. Our research showed the difference in vocabulary was not striking."
Craig's research builds on the work of the centre's founder, Emeritus Professor John Burrows, who was the first to establish that simple function words such as "he", "and", "but" and "if" were rich in stylistic information when analysed using computational techniques.
Ina novel cross-disciplinary exercise, Craig employed the expertise of Professor Pablo Moscato, who heads the University's bioinformatics program, to assist in the analysis of texts. The pair undertook a joint project comparing the structure of language in Shakespeare's plays and poems, which returned interesting evidence of a vast disparity in style between the two literary disciplines.
He has also linked with University speech pathology researchers to study how computational linguistics can be applied in the health sphere.
"We are looking at how people's language changes with ageing but there are other researchers using the techniques to investigate how people's language changes with the onset of Alzheimer's. This could in turn lead to early detection if you could find a way to pick up on those changes in language use," he says.
Craig says computational analysis is not only applicable to the work of great writers. It can be used just as effectively to identify the idiosyncrasies of any individual's language.
"The miracle of language is that we all make something individual out of a common resource. Computer analysis allows us to detect those word patterns more accurately than simply relying on intuition."