|Computers & Texts No. 12
One of the qualities by which Gibbon was most impressed in Erasmus was the fact that 'his learning was all real, and founded on the accurate perusal of the ancient authors. The numerous Editions he published sufficiently evince it; and besides, those convenient compilations of all sorts, where a modern author can learn to be a profound scholar at a very small expence, did not then exist.' Gibbon is referring to the dictionaries, glossaries and thesauri so characteristic of late Renaissance classical scholarship: works such as Charles du Fresne Ducange's Glossaria of middle and late Greek and Latin, and Robert Estienne's Thesaurus Linguæ Latinæ. The view that the learning which derived from, or was dependent upon, such massively impressive but also pragmatic codifications of the literary work was in someway underhand proved tenacious. In the squabble over the Epistles of Phalaris, Atterbury and Boyle sneered genteelly at Bentley's tastelessly minute (although damagingly accurate) 'dictionary learning'. In Hardy's A Pair of Blue Eyes, Elfride Swancourt's father, having discovered that Stephen Smith is the son of a stonemason, disparages as 'very cut-and-dried' the familiarity with Horace which had allowed the young visitor to delight his host by capping one line of Horace with the following line. Such learning in one of lowly station could not have been come by honestly. It was a kind of cheating, to which gentlemen were unable to descend. They read classical literature to imbibe its ethical spirit, and therefore were eager to see through or past the detail of the material, verbal container which conveyed to them that precious essence.
But what would Gibbon, or Boyle, or even the Reverend Swancourt, have said of the modern descendants of the Renaissance's 'convenient compilations', namely the accumulating mass of electronic versions of literary texts? Surely the appearance of prodigious scholarship has never come cheaper (at least in terms of effort: the financial cost of these electronic editions, although declining, is still sufficient to deter all but the most affluent). When the whole of Greek and Latin literature can be searched for strings of characters in a matter of minutes, all classical allusions can be identified by anyone who has sat through a Windows tutorial. Even Bentley or Scaliger might have thought twice before declaring categorically that a certain phrase did not occur anywhere in the writings of Greece and Rome. Such dogmatic certainty is now within the reach of those who would be hard put to decline mensa.
Chadwyck-Healey have a dominant position in the application of this technology to English literature: the English Poetry Full-Text Database, English Verse Drama, and the Editions and Adaptations of Shakespeare database under consideration here will shortly be joined by parallel publications devoted to prose drama, early English prose fiction and eighteenth- century fiction. The electronic subjugation of our literature has developed an unstoppable momentum.
But what kind of activity are these publications intended to foster? This depends on two factors: the scope of what is included, and the design of the software which controls access to it. In the case of the Editions and Adaptations of Shakespeare, the decisions about what to include have been sensible and even generous. The most imaginative aspect of the project is the inclusion of over one hundred adaptations and re-writings of Shakespeare. These are unfamiliar to all but the specialist scholar, and it is both exciting and enlightening to browse in this section of the database. Aside from this, we are given a generous number of complete editions, from the First Folio to the nineteenth- century Cambridge Shakespeare, separate printings of individual plays, and apocryphal and related works. The very fullness of what is provided prompts some regrets: it would be good to have access to later folios, and Hanmer's edition is a surprising omission from the list of eighteenth-century editions. And if re-writings and adaptations are to be included, why not also spuria such as Ireland's Vortigern and Rowena? But, as Sir Charles Chadwyck-Healey himself pointed out in a recent letter to the Times Literary Supplement, one of the great advantages of electronic editions is that they are genuinely open-ended. Once the structure has been established, new materials can be added. Moreover, because the texts have been encoded using SGML, the transcribed materials are not permanently wedded to the DynaText software currently employed (excellent though it is, because of its thorough implementation of SGML). Once better software is developed, provided it too is based on SGML, then the laborious business of transcription and encoding should not need to be repeated.
The software which controls access to the transcribed texts will be familiar to those who have used the English Poetry Full-Text Database. The screen has the same, perhaps slightly spare, appearance, but the searches work very quickly, and can be customised in helpful ways. A particularly fascinating feature is the ability to search for proximity, a command which could be known colloquially as 'the Spurgeon': the program will search for two words occurring within a variable number of words, which the user can specify. We can expect the verbal patterning of Shakespeare's plays to become once more fashionable.
The selection of material is generous, and the software is the best currently available: but if you look at the how these materials are used in the computer suites where they are networked, they seem to support only sporadic and low-level activity. It would appear that nobody reads whole texts in this form, although it is possible to do so. Perhaps more surprisingly, people access these packages for very short sessions. Scholars still frame their hypotheses when reading traditional books, and then test them in short, concentrated periods of activity with the electronic databases. Of course, the knowledge that the electronic databases are there to be consulted arguably encourages people to frame more ambitious and perhaps more interesting hypotheses than they otherwise would. Nevertheless, for the moment it seems to be the case that these electronic resources are supplements to, and not replacements for, books. It is therefore slightly disingenuous of the publishers to lay such emphasis on the alleged increase in holdings that the purchase of these databases would represent. To own the Editions and Adaptations of Shakespeare database is not the same as owning the books from which it is derived. It is both more and (more importantly) less than that.
Given that these packages seem to be used in this narrowly instrumental way, it is a major shortcoming that digitized images of the pages of the original books are not included. A few images are presently provided, mainly of title pages and illustrations, though these are of such coarse resolution that, even if they were of bibliographically-significant material, they would be of only limited usefulness. Their purpose in the present release seems to be merely that of providing a certain flavour of the physical object. However, the need for the database eventually to be completed by the addition of a complete set of page images scanned at a serviceable level of resolution can be appreciated if one considers how a scholarly user would respond if a search were to yield an interesting or non-trivial result. Although we are assured that the texts have been double-keyed and then proofed, such a user would still go back to the original book to ensure that what the program had thrown up was not just a freak coincidence of error between keystrokers then freakishly compounded by inattention on the part of the proof-reader. And one would do this whether or not one ever discovered an actual instance of error. It is only when digitized images of the original pages are made available within the database that such recourse to the book will be unnecessary. Electronic editions currently being published by CUP, such as the editions of Johnson's Dictionary and of the Canterbury Tales, have this facility, and allow the user to scroll synchronously with two windows, one containing the transcription, the other the digitized image. Such images are not particularly costly to prepare. But they occupy large amounts of memory, even when monochrome and compressed.
All of which suggests that the days of real learning are not numbered. The value of these large databases is entirely consequential: that is to say, it depends completely on the intelligence of the questions they are set to answer, and the wit of the broader projects into which the results will be inserted. For as far ahead as we can see, these projects will still be imagined, these questions will still be framed, by meditative readers who look up from the page, surprised and delighted by thought.
This review was first published in The Times Higher Education Supplement,14th June 1996.
[Table of Contents] [Letter to the Editor]
Computers & Texts 12 (1996), 25. Not to be republished in any form
without the author's permission.
HTML Author: Michael Fraser (firstname.lastname@example.org)
Document Created: 22 August 1996
The URL of this document is http://info.ox.ac.uk/ctitext/publish/comtxt/ct12/womers.html