Sanity regained but charisma lost

Susan Hockey's informative handbook about what computers do and might do for analysers of literature and language is an engaging cross between a travelogue, a manufacturer's manual and a cookbook. It is also a kind of ghost story, haunted by the wry shades of all those scholarly foot-sloggers of yore: the melancholy men and women BC (Before Computers), whom Robert Burton described as suffering from rheums and piles as they (literally) took pains over heavy pre-electronic intellectual labour; people who went mad and blind or simply up the wall counting words, collating tropes and phrases, making concordances entirely by hand, or just vainly scanning some fat collected works for the quotation they could have sworn they thought they remembered.

It is a happiness you cannot help feeling as you rummage through these pages, that the Reverend Alexander Cruden, whose toils compiling his great 18th-century Bible concordance tipped him into delusions and insanity, is not alive to contemplate the ease with which anyone with a few GCSEs can knock off computerised concordances of incredible complexity. Those of Hockey's readers who earned PhDs assembling concordances that our machines can now compile even as we sleep are surely entitled to their wryness - leave alone those more recent heroes of early computation arduously typing out their great hoards of cards to feed into the Moloch maws of old mainframes - a sentence per card: 60,000 of them in the case of the early Chaucer project.

That old Chaucer effort is "inconceivable now" says Hockey, with feeling. It is one of the rare thoughts her book spares for toilings past. "Where next?" is her last chapter. Onward and upward is its gist. And the tone is rightly celebratory as we are shown the fabulous possibilities for literary analysis, for putting factual meat on critical hunches, for stylistics and attribution studies, for grammar analysis, for linguistic and socio-linguistic surveys, and for mere information about what is in our mouths and on our minds all of which are made handily available by electronic texts, concordances, language corpora, dictionaries and databases galore.

A brave new world has arrived, and one to be marvelled over daily when you can read from your desk a foreign library catalogue, or check the Oxford English Dictionary or consult a colleague even as you write, not to mention when you can find out at the press of a few keys that there are 784 occurrences of the name Emma in Emma, and see at a glance how often Portia talks of love in The Merchant of Venice or readily find what precedes or follows a word in a literary text or in language use.

Hockey has a plethora of such lovely examples. And gone are the days, as she points out, when grammarians need rely on "toy sentences" for their conclusions. The great electronic language corpora, such as the British National Corpus, are in effect the most representative speakers of our languages, with the largest possible ears to hear all the variants and subtleties of usage, and the most retentive and accessible memory for what is said, far surpassing the most knowledgeable and recollective of any of us.

The accessibility of our machine word-hoards does, of course, do away with the old charisma attached to those teachers who impressed by having so much literature at their fingertips, or to editors who astonished lesser mortals by their ready recognitions of quotations and parallels and analogues simply from their great memory stores (that one small head could carry all he knew...). Bravura annotations, even ordinary annotations, of scholarly texts lose their glamour when machines can do so much of our remembering for us.

What the machine cannot manage, though, is interpretation. And this is Hockey's most important caution. Computers can assemble mind-boggling heaps of facts, can draw out and put together almost anything in a text or a language-set and in more or less any order or grouping we demand, but what is then done with this material relies on old-fashioned critical intelligence and hermeneutic skill. This is not a new thought (and much of Hockey's value is in relaying not only other people's electronic methods and results, but also in passing on well-deserved doubts and scepticisms).

Some of these objections are, I think, a theoretical step too far - such as C. Huitfeld's claim that "there are no facts about a text which are objective in the sense of not being interpretational", so you can never have anything wholly reliable in the electronic texts you set up. Which is evidently, I would suggest, false; and one knows from which hyper-deconstructionist neck of the woods it floats in from. But the constant reminders here that encoding texts and linguistic items for electronic consumption involves much interpretation even before the critic or other analytical user of the programme arrives on the scene is most salutary. Hockey rightly cautions that tagging (how you mark the items, such as sentences of words or names) is crucial and inevitably contentious. There is no standard language or code; the difficulties are endless. If there is one big lesson to learn here it is "don't attempt this at home or on your own". Buy other people's texts and databases and always go in groups. Above all, as Hockey keeps pressing, one must resist the dream (which has hovered dementingly over all literary number-crunching) of final concrete critical factuality, the delusion that facts are by themselves sufficient in a reading, that mere compilations of evidence, of the kind now so readily available through computers, can finally solve problems of stylistics, and attribution and so on. Computers are not, as Hockey keeps saying, the ultimate magic black box.

Pleasing wisdoms. As pleasing as the belief professed here that books will survive - not least because they are more portable and easier to read than screens. They are cheaper too. And here is Hockey's biggest gap, the commercial side of our brave new scenarios. Electronic materials come very expensive. I cannot afford a fraction of the CD-Roms I use daily. My institution, happily for me, provides them. The great electronifications of Greek and Latin literature, women's writings and so on, exist only courtesy of universities and publishers with deep pockets. Making a living out of electronic publications is hard. The problem of reward for the producers of electronic editions of texts is still not solved. The scholarly CD-Rom market is too tiny for serious profit, which is why Bible CD-Roms, for which there is a certain popular demand, tend not to carry scholarly editions, and why Chadwyck-Healey, the great producer of electronic texts, set up its online subscription service. And still its main subscribers are institutions. So the implied critic in these certainly brave new times is even less of a common reader than ever before. Which is something deeply bad amid so much that is absolutely to the good.

Valentine Cunningham is professor of English, University of Oxford.