HOME  |  CONTACT US  |
 

Post Your Opinion
Computing Creativity
by Ian Lancashire

WHAT DOES TACT, AN unprepossessing publicdomain computer program from the University of Toronto, have in common with the US National Supercomputer Center's high-powered software Mosaic? Something rather significant, according to the Getty Foundation, the American Association for Learned Societies, a and the US Coalition for Network Information: TACT, like Mosaic, has been selected as one of the most promising new tools for scholars of the humanities, and will be an integral part of the US National Information Infrastructure called for by President Clinton.

So what exactly is TACT? Most basically, a group of 16 computer programs available on the Internet at no charge from the University of Toronto. These programs incorporate an interactive make-your-own-concordance feature that allows users to do all kinds of interesting things with the words of a text. You can take a word that seems to be important and look at it in all the contexts in which it appears, graph its frequency of distribution in the text, and discover recurring phrases and patterns of usage. TACT also has the ability to make word-lists, find anagrams, compile lists of phrasal combinations, and construct dictionaries for the texts it analyses. It won't yet make coffee, but we're working on it.

Since 1986, the University of Toronto has invested the time of three gifted programmers, a professor of English, and other staff in the development of TACT. During the same period, IBM Canada and the university combined to set up the Centre for Computing in the Humanities, whose mandate is to develop and make available, without charge and open to the public, computing tools and resources. TACT is one of these tools. You may get it from the Centre for Computing in the Humanities World Wide Web front page on the Internet -- its URL is http://www.cch.epas.utoronto.ca:8080/cch/tact.html -- and it won't cost you anything, there are no user fees, and no salesperson will call.

Since TACT costs nothing, anyone can afford to run it, and does: students from high schoolers to doctoral candidates, writers skirting the poverty line, lawyers, journal!sts, and even Bill Gates. TACT also crops up in expected places. You can find it in two scholarly CD- ROMs, one from the International Computer Archive of Modem English located at Bergen, Norway, and the other in a collection of medieval Spanish texts published in Madrid. TACT users study most of the modem European languages and live everywhere from Moscow to Manhattan to Santiago, Chile, which explains why the Modern Language Association of America, an intertional association of more than 30,000 academics, will be publishing TACT's manual in New York this year.

One of the parallel resources spun off from TACT is a big, free collection of the best English poetry from Wyatt to Swinburne, drawn mainly from Representative Poetry, a textbook edited by members of the U of T's English department from 1912 to 1967. Northrop Frye, one of the editors of Representative Poetry, contributed the selections of the verse of William Blake and Christopher Smart. This electronic library is now available on the World Wide Web from the university's Robarts Library, and you can tune in at URL http://library.utoronto.ca/www/utel.

In 1989 in Toronto, Frye gave the keynote address to the largest international humanities computing conference ever held. He was initially reluctant to accept my invitation, saying repeatedly that he knew nothing about computing and had another commitment. In the end, he agreed to come and talk about computing to an audience of more than 400 people at the Medical Sciences auditorium. Everyone loved him. He was the treat of the conference -- and the first to submit his text for publication in the proceedings of the conference, within an hour of his talk.

Frye's willingness to accept this invitation hints at why TACT has proved successful with scholars of the humanities. In his Anatomy of Criticism, Frye worried about the state of literary criticism. He said that if consensus among literary critics about the meaning of the texts they studied was impossible, the whole subject was in danger of becoming a pseudoscience like phrenology. Frye argued that those who studied the humanities had to learn to be like scientists, building on the achievements of their predecessors rather than forever demolishing all previous work and replacing it with entirely new structures and restructures.

When the Centre for Computing in the Humanities began working on TACT, we wanted to make text retrieval and analysis tools to help students of literature check their memories of the works they discuss. TACT's programs serve up texts of which a critic already has a working interpretation in a way that allows us to check our ideas, point by point, against what the author wrote. TACT is a kind of fact-checker of critical thinking. It helps academics like myself to be accountable to the authors whose works we teach in order to make a living.

But it has to be admitted that close readings of poems, plays, and novels are a bit passe in many literature departments today, when theory plays the major role in the writing of literary criticism. Contemporary critics often tend to see themselves as creative writers who take all of culture as their subject: a viewpoint that at its extreme leads to self-delusory assertions that authors no longer exist once their books are published -- in other words, that all authors are dead. Having disposed of the authors, such critics can go on to say anything they like about texts, as long as their arguments are superficially plausible. And in order to increase the probability of their publication, these arguments must be as different as possible from all preceding interpretations.

In 1986, the year when the Centre for Computing in the Humanities was founded, Margaret Atwood described deconstructionism, a philosophically skeptical critical theory, in an interview with Geoff Hancock: "What it ... means is that the text is of no importance. What is of interest is what the critic makes of the text. Alas, alack, pretty soon we'll be getting to pure critical readings with no text at all." Every critical movement tends to generate the seeds of its own contradiction. Like any text-analysis system, TACT appears to deconstruct: it smashes books into the constituent particles of which they are made. Yet in fact it really re-establishes belief in the author, because it finds her living fingerprints in those very same books.

About the time TACT was ready for release, I approached Margaret Atwood's agent for permission to do a computer analysis of The Handmaid's Tale. My reasoning was that it was not enough to say that TACT could make critical thinking better. We had to try to prove that it did. Naturally, since I would be copying the entire book into a new form, I needed Atwood's sayso. Less sensitive to a writer's reliance on the market-place for a living than I should have been, I also asked her to allow me to distribute the electronic text of the novel freely across Canada. I wanted students to have an exemplary text to show how computer analysis could improve reading skills. In retrospect, this was ivory-tower innocence on my part, and it received an appropriate response. Exact quotation aside, her agent replied: Don't create the electronic text, and please don't do the computer analysis either -- Ms. Atwood meant the book to be read by human beings, not machines.

Somehow the paper got done anyway, within the scruples of my conscience, and it was published, and Margaret Atwood last year, reviewing the results, said that it was an intelligent construction, though (let me hasten to add) I neither undermined her publisher's sales by passing out copies of her book nor renewed my request to do so. Fortunately, TACT here accomplished what I had hoped it would. Let me explain how.

The critics whose work I looked at gave confident and contradictory interpretations of The Handmaid's Tale, especially of Offred, the title character whose confessions comprise the novel. Set in a future fascist New England called Gilead, The Handmaid's Tale depicts several months in the life of a woman who has enslaved herself to serve as a surrogate mother for powerful but infertile couples among the ruling e1ite. No critic shared the interpretation of Atwood's own Professor Piexoto, a fictional editor who, in an epilogue to the novel that satirizes academic writing, calls Offred a tart and jokingly puns on her "tail." Taking their cue from Atwood, critics avoid making this mistake, yet sometimes wander into mazes of their own. One feminist critic says that Offred is a feminist hero on the French model: risk-taking, "herstory"-telling, politically rebelling, and joyfully reclaiming her body from the thing that she rents out for use to couples. Another critic sees Offred as a fearful creature who sacrifices her own name, her identity, and her integrity just to survive. Still others stress themes of bestiality and violence, or maternalism and womanly nurturing. All these critical arguments are made plausibly, with apposite references from the novel. My own reading of the text generated yet another interpretation, also bolstered by what seemed to be telling quotations.

Which position was best, and how could one tell? Or were we all -- along with our students -- lost in primal intellectual ooze? I was reminded of my experience of reading Protestant and Romanist doctrinal disputes of the early 16th century, in which everyone quoted the Bible to prove that their position -- which everyone else rejected -- was uniquely right. In the case of The Handmaid's Tale, I found myself, a male like the despised Pieixoto, in the position of disagreeing with critics who were arguing for a more explicitly feminist reading.

I began by transforming The Handmaid's Tale into concordances, lists, and graphs in order to run some tests on what the critics said. First I checked the feminist critic's claim. Whether or not her views were right or wrong, she turned out to have inaccurately remembered the novel. Not only does Offred take no risks, but she says so, repeatedly. Far from rebelling against the patriarchy of Gilead, she surrenders to it at the novel's end "because it can't be helped." As for joy in her own body, Offred's behaviour suggests a more practical attitude; and the word "joy" and its cognates are applied only to Serena Joy, the matriarch who collaborates with the fascist state of Gilead and has the most power to get her way. If there is a feminist in The Handmaid's Tale, she would seem to be Serena Joy or Moira, Offred's rebellious lesbian friend. Does Offred then sacrifice her personal integrity, as another critic says? While the plot gives reason for thinking so, Atwood dedicates many chapters to Offred's night-thoughts, and in these Offred reveals a character that a reader can easily respect and even love. Is violence then the key theme? Or motherhood? Again, simple examination of the text reveals no special emphasis on the words by which these concepts are expressed, and readers using TACT are able to see so.

By this time I had abandoned my first interpretation. TACT displays showed that my memory of the novel was also partial and overemphasized some things. I wondered about how best to define an interpretation that would come from Atwood's writing, not from me. Eventually I decided to look at the novel's high-frequency content words. These include all non-function words. Function words are what we all have to use in talking about anything. They include what linguists call closed-class words: prepositions, articles, and auxiliary verbs such as "be" and "have." This class of vocabulary is closed because no one invents new function words. They form the frames or molds for all our sentences, no matter what we say or write about. The open-class or content words, on the other hand, define our personal vocabulary and our subject. Of about 200 high-frequency content words in Atwood's novel, most refer to a woman's body. Particularly numerous are cognates for "hands" (211) and "eyes" (169). This fact tallies nicely with Atwood's narrative. I had found her subject, an obvious one I suppose, but one that made other interpretations partial and also explained why feminist critics tended to react as they did.

Using a text-analysis system to check critical thinking may seem like overkill. Yet if teachers are responsible for being truthful, they must have more than a method to teach (especially a method that produces criticism which disagrees even on fundamentals). They must have more than a model of thinking that multiplies contradictory critical thoughts and offers no grounds for choosing among them.

Respect for the fights of an author is also at stake here. The humanities should protect writers and represent their stories and ideas to future generations fairly. Academics respond well to TACT, I believe, because, like Frye, they believe that critics and teachers have to try to reach consensus on books, to describe them in ways that can be falsified, if not necessarily verified. Like Atwood, most academics also believe that no one should appropriate a book from its author. It would be better for a critic to write his own book than to remake someone else's book in her own image and pass it off as original thinking.

But TACT does much more than check our memory of what authors have said. It reveals networks of word patterns unique to an author. Such phrasal repetitions survive from the writer's own long-term associative memory at the time he or she first spontaneously generates the firstdraft utterances onto paper or screen, and they seldom exceed seven words in length. This length tallies exactly with the maximum number of words that experiments in cognitive psychology have shown can be stored in our working memory, in what is sometimes called the "phonological loop," at any moment in time. Here literary text analysis and the human sciences meet.

At stake is a method for identifying datestamped fingerprints of an author's original handiwork. I do not mean the conscious expertise that goes into editing drafts and removing repetition (for these use artificial memory, paper or screen, to store copies of the text). These skills can be taught, learned, and applied, step by step. What I do mean are the unconscious cognitive processes that produce our spontaneous speech and many of the often surprising sentences we find ourselves writing. Here the linked thoughts of our associative memory enter the real world. These thoughts seem to be little miracles, because we cannot actually introspectively examine how we form utterances. Language skills call on what neuroscientists call procedural or implicit memory. This kind of memory can only be recalled in performing something. Uttering a sentence on paper or aloud is not unlike tiding a bicycle. Gore Vidal once confessed that when he writes "I never know what's coming next ... That's why I go on, I suppose. To see what the next sentence I write will be."

In an interview with Eleanor Wachtel, Atwood agreed with Vidal: "When people see the nice books with the nice white pages and the nice black writing, what they don't see is the chaos and the complete frenzy and general shambles that the work comes out of '(my italics). Authors know the truth of the maxim, "How can I know what I'm saying until I have edited it?" Margaret Atwood's description of her muse echoes the thoughts of many other writers. It is not that writers are subject to randomness, but rather that minute, apparently trivial, initial conditions of which the author is almost always unaware exert powerful effects on the shape of the writer's work as published years later. Chaos theory, in other words, seems to hold for both human and physical processes.

If we can learn how to distinguish between what authors do as editors, tinkering and fine-tuning, and what they do as utterers, unleashing creative powers that seem chaotic, frenzied, and unpredictable, we may be able to learn some important things about the creative process. There are many possible practical applications as well, such as a scientific basis for the assignment of copyright and the identification of authorship for works whose origins are unknown or problematic. TACT, and other text-analysis programs, unite intellectual and technological endeavour in a way that promises to heal old divisions between the sciences and the humanities, while also sweeping away much of the pretentious and arbitrary gibberish that masquerades as useful literary criticism.

footer

Home First Novel Award Past Winners Subscription Back Issues Timescroll Advertizing Rates
Amazon.ca/Books in Canada Bestsellers List Books in Issue Books in Department About Us