Introduction

Marilyn Deegan
Refugee Studies Programme
University of Oxford

black lineOn this page

Ten years of CTI Textual Studies

The early days of textual computing

The development of textual resources Textual analysis and hypertext

Textual analysis and hypertext

New textual theories and electronic editing

Textual studies and non-textual media

The future of publishing

Computers and teaching

Conclusion

red squarered square

Ten years of CTI Textual Studies

The humanistic disciplines are diverse in terms of materials studied and methodologies adopted, though they have in common the unifying (but somewhat vague) theme of human culture, which includes material culture as well as textual artefacts. There is always much debate about the boundaries we apply to the humanities, and how we determine what should be included or excluded. Some aspects of archaeology, for instance, are within the humanities, while others, such as palaeopathology, come more properly within the sciences; some subjects within history are generally included in the social sciences. Within the humanities, too, the individual subjects and disciplines have some difficulty in deciding their boundaries, and the dividing lines shift according to the historical moment, current fashions, and the interests of individuals. Some practitioners, indeed, are more comfortable when involved in inter- or multi-disciplinary investigations, and much creativity is generated here on the m argins. Unfortunately, this uncertainty of taxonomies becomes a problem when mapped onto the expectations and the understanding of funding bodies when they wish to evaluate research and teaching in the regular (in the UK) Teaching Quality Assessment (TQA) and Research Assessment Exercises (RAE). Within this Guide, and for CTI Textual Studies generally, 'textual studies' is also difficult to define and indeed confine, and this to some extent reflects debates within the subjects which traditionally study texts, and which are themselves continually reexamining and extending the notions of what can be included as 'text'. CTI Textual Studies has always endeavoured to be permissive and inclusive about the definitions, taking the view that most textual teaching and research can be enhanced by new technologies, and that a broad-based approach would best serve the community. For various reasons, the coverage of the Centre has expanded considerably over its ten-year life-span. It began i n 1989 as the CTI Centre for Literature and Linguistic Studies, dealing with literature in all languages and from all periods, and with linguistic studies of texts, in particular the use of linguistic corpora. In 1991, it gained responsibility for philosophy and theology when it took on these subjects from the former CTI Humanities Centre, and in 1994, when the CTI as a whole had its remit extended to include the new universities, it expanded its range to include film, theatre, and media studies. This Guide to Digital Resources represents all these subjects and more, and covers the broadest possible range of resources and methodologies currently available to the textual scholar and student.

Definitions of subjects change, and the boundaries shift, but the main thrust of humanistic endeavour remains the same: the search for meaning and understanding from a wide range of primary and secondary sources. Interpretation of texts and artefacts, from our own contemporary culture or from cultures far removed by time, space, or both, and the sharing of our interpretations have always been (and hopefully will always be) the motivating forces for what we do. Our methods may seem to have changed drastically over the last ten years, but we are, I feel, striving for the same goals to investigate and share with our colleagues and students what it means to be part of human culture. Given that finding, sifting, and interpreting considerable quantities of dispersed evidence from a plethora of locations is a prime task for humanists, the use of communication and information technologies has much to offer, and new methods have been adopted eagerly by many. The opportunities are great, though there remain problems and blockages which are more economic, political, and organizational than they are technical.

The early days of textual computing

Textual computing has a much longer history than the ten years of the CTI Centre for Textual Studies, of course, and indeed goes back as long as computing itself. The development of the stored program computer (originally conceived of by Charles Babbage as the Analytic Engine in the nineteenth century) was a consequence of the code-breaking machines developed in World War II, in particular the Colossus used at Bletchley Park from 1943. The first successful run of the stored program computer was in Manchester in 1948, but some years before this, an American scientist and thinker, Vannaver Bush, had conceptualized of 'thinking machines' which would help the modern world make sense of the then current information explosion. In his seminal article, 'As We May Think', published in the Atlantic Monthly magazine in July 1945, Bush, one of Roosevelt's advisors in World War II, was already turning his mind to peaceful uses of technology, remarking:

This has not been a scientist's war; it has been a war in which all have had a part. The scientists, burying their old professional competition in the demand of a common cause, have shared greatly and learned much. It has been exhilarating to work in effective partnership. What are scientists to do next? (Bush 1945)

Bush was grappling with the problem which has beset humanity since the invention of printing, and possibly before. How, when so much information is available, does one remember what has been read, make connections between facts and ideas, and store and retrieve personal data? Pre-literate humanity was capable of prodigious feats of memory, but the total amount of knowledge it was necessary to absorb was infinitely smaller than it is now. When writing became widespread, personal notebooks and handbooks were used by scholars as aides-mémoires. Today, the scholarly tools are likely to be computers large and small: from workstations to hand-held personal digital assistants.

Bush's own solution was to propose an automated 'memory extender', a device which he called a 'MemEx' in which an individual could store all his or her books, records, and communications. This would be mechanized to enable rapid and flexible consultation. Interestingly, given developments in digital technology at that period, the MemEx was essentially an analogue machine and used microfilm for information storage with a mechanical linking process. The editor of Atlantic Monthly saw Bush's ideas as offering humanity 'access to and control over the inherited knowledge of the ages' and 'a new relationship between thinking man and the sum of our knowledge'. The machine was never built by Bush, but drawings of it exist, and a simulation of it was produced by Paul Kahn and James Nyce for the MIT Symposium in October 1995 which celebrated the fiftieth anniversary of the publication of 'As We May Think' (Nyce and Kahn 1991). This conceptual device may have had no physical manifestat ion, and indeed was probably completely unworkable, but its importance lies in its role as the precursor of all modern hypertext systems.

Despite Bush's notions of 'thinking machines' to help us organize knowledge and store large quantities of data, the first uses of the computer were generally for code-breaking and as an aid to large-scale number-crunching, but its value in processing large volumes of text was perceived early by some pioneers. Father Roberto Busa, the Thomas Aquinas scholar, was one of the first humanists to produce electronic text, laboriously entering texts on punched cards and constructing huge lexical and morphological databases. 'Why', he remarked in 1997 'should one do only what is easy?' (Busa 1998, 9). Busa formulated the idea of automated linguistic analysis in the years 1942-46, and started working with IBM in New York in 1947. He produced more than six million punch cards for his Index Thomisticus, and in 1992 the first edition of his Thomas Aquinas CD-ROM was published.

Other early developments contributed to the field that we now know as humanities computing (see Nielsen 1990 passim, for more on the history of these early technologies). In 1962 Douglas Engelbart (an early follower of Bush) started work on the Augment project, which aimed to produce tools to aid human capabilities and productivity. He also was concerned that the information explosion meant that workers were having problems dealing with all the knowledge they needed to perform even relatively simple tasks, and Augment aimed to increase human capacity by the sharing of knowledge and information. His NLS (oN-Line System) allowed researchers on the project access to all stored working papers in a shared 'journal', which eventually had over 100,000 items in it, and was one of the largest early hypertext systems. Engelbart is also credited with the invention of pointing devices, in particular the mouse in 1968. The mid-1960s saw other humanities computing pioneers conceiving of grand schemes which at the time seemed so innovative as to be impossible, but which now are coming to realization. Ted Nelson in 1965 designed his Xanadu system in which all the books in all the world would be 'deeply intertwingled' (in his words - Nelson incidentally coined the word hypertext). Nelson also tackled the problems of copyrights and payments by proposing that there should be electronic copyright management systems which would keep track of what everyone everywhere was accessing, and charge accordingly through micro-payments. Web interlinking, e-commerce, and digital library developments are bringing into being the systems which Nelson has eloquently argued for over the last thirty years.

The development of textual resources

In the ten years since the establishment of CTI Textual Studies, much has changed in the use of computers for the study of humanities disciplines. In particular, the user population has moved from being a small and dedicated band of technically-minded individuals, prepared to deal with the arcana of mainframe computer command languages and code conversion programs, to embrace almost all scholars working in the fields of studies where texts are the prime objects of investigation and who are using computers now on a daily basis. With the improved facilities available, and the exciting prospects offered, barriers of technical interest, age, or time have been overcome. The movements too are almost exclusively one way - no one surely over the last ten years has tried word-processing and returned to pen or typewriter, or has tried email and returned totally to letters. There are, however, still some economic and organizational barriers to the widespread take-up of IT in the academic worl d. It is technically possible for academics and students in the UK to have access to a wide range of computers and networked resources. The power of desktop and laptop computers grows exponentially, while prices remain stable or even decrease in real terms, and the academic networks (in the UK, JANET and SuperJANET) increase in bandwidth regularly. For domestic use, service providers make access even cheaper, and indeed many now offer this free. Despite the fact that hardware, software, networks, and resources seem relatively cheap, many individuals and institutions still cannot afford them, and a number of complex and interconnected problems exist. Since 1981, we have been in a period of diminishing resource for UK higher and further education. C&IT facilities are an extra burden, and hard decisions sometimes have to be made about whether to spend scarce funds on conventional library resources or new media. Add to this the necessity for increasingly frequent upgrading, and the extra burden of supporting machines and networks, and costs can rapidly spiral upwards. Many now claim that traditional lectures are ineffective, but they are relatively cheap, and the costs do not increase much even if class sizes double. The UK higher and further education sectors are in danger of existing within an environment which is divided into the information poor and the information rich; a dangerous situation which can only be addressed at the higher policy levels of governance.

Despite the above caveats, in personal writing and administrative work, and in much research, there has been a paradigm shift towards electronic resources, and there have indeed been some developments that can be described as revolutionary. Modern linguistics and lexicography rely heavily on large electronic corpora to give accurate information about language usage at different periods of time. Corpus work began in the 1960s with collections such as the Brown Corpus (of American English) and the Lancaster-Oslo-Bergen Corpus (of British English), both of which contain around one million words of English. These sound extensive, and indeed were of enormous value in the early days, but corpora of this size are not adequate for accurate linguistic research, particularly of rare words or constructions, and more recent corpora are orders of magnitude larger that this. The British National Corpus (BNC) contains some 100 million words of written and spoken English, and the COBUILD Bank of English has 330 million words of Modern English. With historical materials, it is possible to build corpora which encompass the totality of the extant writings in a language or of a particular period. For example, the Toronto Corpus of Old English holds the complete texts of literary materials written in English between 450 and 1100, or the Thesaurus Linguae Graecae which contains all literary materials written in ancient Greek between the fifth century BCE and the eighth century CE. These two resources have been invaluable for the study of texts from these periods, as they allow comparisons and precise citations as never before.

Other textual resources which have been made increasingly available are the complete corpora of particular works, authors or genres. There are many versions of the Bible and Bible commentaries, and almost as many complete Shakespeare collections (though perhaps one might pick out the Arden Shakespeare as a particularly good one). Other complete author sets abound, including of course, the very first collection, the works of Thomas Aquinas discussed above. Of genre-focused corpora, perhaps the Chadwyck-Healey English Poetry Full-Text Database, originally produced on CD ROM and now available as part of Literature Online, is the best known, though Chadwyck-Healey and others (both commercial and non-commercial sources) produce many more collections of literary texts. The UK higher education sector was fortunate to have a special purchase deal arranged for the Chadwyck-Healey databases (through CHEST), which has allowed departments to access this for a great deal le ss than the full cost of the resource. Students have therefore been able to use Literature Online with great success in some courses. In St Andrews University, for instance, these databases have been used for several years in Renaissance poetry courses, and some students have achieved outstanding results.

In the first year there was an outstanding piece of work on 'The Phoenix and the Turtle' - outstanding partly because it was done almost entirely on the student's own initiative and without a prototype to follow. In fact, he found his own prototype in the Arden Shakespeare and produced a meticulously word-processed document with double-column notes in the lower half of the page... This year an edition of Marvell's 'The Garden' appeared with text and notes visually illustrated from Ripa, Giovanni Battista Ferrari and Plot's Natural History of Oxfordshire. An edition of 'Lycidas' was so carefully prepared that the student felt entitled to place a copyright mark on the cover, a gesture which would itself have merited a tutorial on the conditions of electronic textuality. (Rhodes 1998, 6)

Electronic textual resources have grown rapidly in the last ten years, but there are still large gaps, and many areas are poorly served. Literature in English and other modern European languages are reasonably well covered, and large numbers of modern classical literary works are available. Given that the Roman character set has become the de facto standard for information interchange, these areas have been the cheapest and easiest to cover. Non-Roman character sets are still more difficult (though infinitely easier than ten years ago), and important though less popular textual resources (legal codes, inscriptions for example) are only partly digitized. Coverage for non-textual materials is even more patchy: film and media studies have enormous problems, though the resources are growing. The materials are more expensive to capture, need greater storage and bandwidth capacity, and if they are commercially produced have copyright restrictions which are almost impossible to ove rcome, even for teaching purposes.

One caveat which we must offer here is in regard to the quality of some electronic texts which are currently available. There are two points to consider: the first concerns the practical issues of how well the text has been captured, how usefully it has been marked up, and how much information comes with it concerning its provenance. The second relates to the quality of the original itself: copyright restrictions (with the recent extensions to UK copyright law) may mean that a text published as much as a hundred years ago is still in copyright. As a result, the electronic texts may not derive from the best or most up-to-date editions. A text which is not the best available may still be useful for search and retrieval, but is unlikely to be a good one for other purposes, as the basis for a new edition, for instance.

Textual analysis and hypertext

The proliferation of electronic textual resources offers many opportunities for new kinds of study, interpretation, and analysis of written works, and analytic tools have been available for many years to assist in this. The nature of the early programs such as OCP (the Oxford Concordance Program), which allowed rapid and easy manipulation of alphanumeric symbols, tended towards a somewhat mechanistic approach, concerned more with scientific analysis than with imaginative reading. Linguistic applications reliant upon the analysis of large quantities of data have been popular, and stylometry, which offers analyses of textual corpora for authorship attribution or dating, was and remains useful, but has largely not been taken up by mainstream textual scholars or literary critics. Of note here is the work of John Burrows on Jane Austen's novels, and numerous analytic articles on the dating and authorship of literary, philosophical, religious, and historical works, many of which a re published in the journals Literary and Linguistic Computing and Computers and the Humanities. Of particular interest and influence is the work of Anthony Kenny (Kenny 1982). More popular recently is the work of such writers as Jay David Bolter (Bolter 1991) and George Landow (see Delany and Landow 1991, Landow 1997) who have used hypertext theories and methods to situate literary works in their cultural contexts, and have encouraged students to engage with new debates and new readings. Landow's 'Victorian Web' is an interesting, early example of this approach, and describes itself as a virtual library with material on just about every facet of the Victorian age. It covers history of culture as well as the social context (economics, religion, philosophy), along with issues as diverse as literature, science and technology, and gender issues. (Online at http://landow.stg.brown.edu/victorian/victov.html)

Landow has also produced webs of individual authors or even poems, for instance, The Dickens Web about the life and times of Charles Dickens and the In Memoriam Web, which analyses the Tennyson poem.

Bolter's groundbreaking work Writing Space: The Computer, Hypertext, and the History of Writing was developed simultaneously as a cybernetic hypertext and a printed book, and offers some profound insights into the reading possibilities offered by the new media (Bolter 1991). Bolter was co-developer of the hypertext software StorySpace, which has been used for a number of hypertext publications exploring notions of new reading. Socrates in the Labyrinth, composed with StorySpace, was one of the first works of hypertext non-fiction to examine and exploit the techniques of hypertext rhetoric discovered in the development of serious hypertext fiction (http://www.eastgate.com/catalog/Socrates.html). Not only has hypertext encouraged new forms of readings, it has also been eagerly embraced by many writers (professional and students) for the liberating environment which it seems to offer. Fiction write rs have experimented with confounding the linearity of the printed codex form for many years (Dickens and other writers of Victorian multi-plot novels in the nineteenth century; Paul Scott, Lawrence Durrell, Charles Palliser, Sebastian Faulkes, and A.S. Byatt more recently). Now cybernetic hypertext allows further experimentation with non-linear fiction. Michael Joyce, co-developer with Bolter of StorySpace, has used the complex inter-linking, analytical and presentational facilities of this program to explore fictional spaces in hypertext novels such as Afternoon (Joyce 1987), and there are many creative writing Web sites where students experiment with non-linear and collaborative writing. See for instance the peer-reviewed hypertext journal of creative writing Kairos (http://english.ttu.edu/kairos/) for some interesting examples of this.

New textual theories and electronic editing

One area which has been of growing interest to theorists is the ontology of electronic text. The written text is fixed; the electronic text is fluid. With the advent of printing, the written text became fixed in multiple copies of essentially the same work and could be widely disseminated. The electronic text can be even more widely disseminated, given that it does not rely on a stable medium to convey it. This has positive and negative consequences. However often an electronic text is copied, it does not degrade. The hundredth copy is exactly the same as the first. As Richard Lanham puts it, 'Unlike print, the electronic text defies conventional wisdom. You can have your cake, give it away, then eat it, and still have it' (Lanham 1993, xii). Electronic text also has the property of simultaneity: one text on my machine in Oxford could be accessed by a thousand machines throughout the world, a thousand virtual texts on a thousand screens. But the minute the screen is switched off, t he text vanishes and has to be recreated from the stored copy for further access. Conversely, my text could be displayed on the same screen as a number of other texts from diverse locations, compared, contrasted, even integrated with them. The boundaries of texts are consequently more permeable. Texts can thus be exchanged, proliferated, transmitted across the world in seconds, integrated with other media, and linked into complex interpretive networks of variants, editions, illustrations and more. As Kathryn Sutherland points out, the dispersion of location, identity, and appearance of electronic text is making a significant contribution to our understanding of textuality. We have become so used to the book as textual mediator that for the most part we scarcely notice its artefactual state and how it imposes its 'machinery' on what we read; we accept a kind of synonymity between text and book. There is some sense in this, since the book (individually as well as generically) has proved a robust machine for text dissemination, while one of the current anxieties about electronic storage media is their rapid obsolescence (Sutherland 1997, introduction passim).

For textual critics, this new textual world is a liberating environment. Some kinds of works still function well in book form; others (dictionaries, encyclopaedias, scholarly editions) have outgrown the bounds of the codex and are enriched by their cybernetic presentation. For other scholars, the theories of cybernetic hypertext prove illuminating in their consideration of multi-layered, multi-plot novels: Patrick Conner's analysis of Huckleberry Finn (Conner 1997) and Kathryn Sutherland's study of Dickens' Little Dorrit (Sutherland 1990) are good examples. The underlying paradigm of hypertext is simple: it is a means of linking together textual materials using what have become known as 'nodes' and 'links'. Its manifestation can soon become highly complicated, however, as nodes and links multiply very quickly, and there are electronic hypertexts available which claim to have more than two million links. The complexity of hypertext systems means that there is an almost limitless number of paths the reader/user can take through a cybernetic system. Each decision to move in a certain direction, while limiting some choices, also offers a vast number of possibilities for the next move. Interactive games, which make the routes to success or failure contingent upon certain right or wrong choices, exploit these ever multiplying links to create rich fictional structures which have been attracting the interest of narratologists (Gibson 1995). For the editor and textual critic, these new possibilities have enabled large-scale projects that can break free of the restrictions of the printed page in trying to represent all the variant stages of a work and its transmission, or the oeuvre of an artist. Computerized methods as applied to textual criticism were initially used to assist the scholar in the production of the conventional end product: a printed critical edition of the text with the base text printed in full and the variants from other texts at the foot of the page or at the end of the work. Other apparata such as commentary, textual notes, explanatory notes, were also arranged either at the foot of the page or in appendices at the back of the book. Over the centuries the critical edition has reached a high level of sophistication in the organizational principles that allow a flat, linear, printed book to present information which is not linear. Now, however, developments in textual presentation software using s tructural markup and hypertext linking mechanisms, mean that critical editions can be published in electronic form, as well as being generated in this way. Many large, collaborative editing projects have been embarked upon which would not have been possible or which would have taken entire scholarly lifetimes of several individuals without computer assistance. The Electronic Beowulf Project, the Canterbury Tales Project, the Wittgenstein Archive, the Women Writers Project, various editions of Yeats and Joyce, would all have been much more difficult without electronic tools and techniques. Some facilities, such as the facsimile presentation of Wittgenstein's complex originals or the inter-linked transcription databases of Beowulf or the Canterbury Tales, might have been impossible.

Textual studies and non-textual media

One of the key properties of the electronic environment is that when units of information are digitized, they are all then in the same form: binary digits, or bits. There is no difference in essence if a unit began as text, image, video, or sound: all translate to bits (though the various media differ in size by orders of magnitude). So while there are already huge advantages in manipulation and analytic possibilities in the use of electronic texts, a whole new dimension of possibilities is added when one incorporates visual and time-based media into these. Of course, the changing of textual materials into other forms is nothing new: film, radio, and television adaptations of textual works of all kinds have been common from the birth of cinema, radio, and television, and have, indeed, brought new generations of readers to classic texts. Who would have predicted that George Eliot's Middlemarch or W.H. Auden's poetry would reach the bestseller lists in the 1990s, which they di d after the TV adaptation of the former and the use of the latter in the film Four Weddings and a Funeral? Parenthetically, it is interesting to note these developments in the light of frequent reports of the death of the book. This prediction is made with every new technology for communication, from sound recordings at the end of the nineteenth century to the Internet revolution. The book as a form seems alive and well, and likely to remain so for the foreseeable future. New technologies for reading may, indeed, change the outward form of books, (and as discussed above, there is also experimentation on inner structures) but they often rely on the same underlying principles of narrative and construction for the production of new reading machines. The latest developments in fact are ones which many readers welcome, rather than deplore: electronic books, accessed through a dedicated reader which simulates as far as possible the printed book, are a revolutionary new step. Book-siz ed and book-shaped, many works can be stored in one reader. A considerable number of current titles are available, some being published simultaneously in print and electronic book form. They cater to some extent to the popular taste: the first offering from Rocket eBooks at the beginning of 1999 being Monica's Story by Andrew Morton. Cost is, unfortunately, a major barrier to the uptake of ebooks. The US bookseller Barnes and Noble advertises the Rocket eBook reader for $329, and this can hold only ten books. While this could be useful in offering access for the visually impaired in a portable form, it is unlikely at the moment to find widespread acceptance. Frederick Kilgour has posited that, in order to take off, the electronic reading device has to score significantly over books in a minimum of six areas: 1) it must be more legible; 2) it should have at least as large a display area; 3) it must be smaller and weigh less than an average novel; 4) it should be manipulated with one hand; 5) it should have a one-time cost of less than an average novel; 6) it should be able to access text from millions of databases anywhere, at any time (Kilgour 1998, 152).

The ability to integrate many different media into complex hypermedia structures where text informs image, image informs sound, is relatively new. 'Hypermedia' says Bolter, 'is the revenge of text upon television... in television, text is absorbed into the video image, but in hypermedia the televised image becomes part of the text' (Bolter 1991, 26). Of course, Bolter's distinctions between television and hypertext become less clear with the advent of digital and interactive television, and facilities which allow Web searching from domestic TV sets. But advanced hypermedia incorporating images, sound, and video has fundamental implications for textuality, and in higher education has fundamental implications for teaching, learning, and research. Boundaries between and among texts become permeable, boundaries between text and other media blur, and disciplinary and even physical distances and differences begin to dissolve. For instance, ten years ago I would have needed to travel all over the world to see the complete manuscripts of Chaucer's Canterbury Tales, but soon I will be able to see them from my desktop. Literary studies and art history begin to converge when a project like the Rosetti Archive integrates both the textual and visual works of the author/painter. Film studies, drama, and literature come together when I can have the Arden Shakespeare, recordings of dramatic performances, and a range of films all on CD-ROM. Some experimental projects are going even further than this. For instance, the Georgian Cities CD ROM (Gallet-Blanchard and Martinet 1999) integrates maps of Edinburgh, Bath, and London with screens and animations on architecture, literature, society and the arts. Various time-based effects alternate texts and pictures in different ways to evoke a distinct sense of narrative progression in key authors or to illustrate musical passages. The Virtual Harlem project at the University of Missouri at Columbia use s immersion technology to create a three-dimensional representation of Harlem at the time of the Negro Renaissance (around 1920-1930) which students can explore at will. In-depth research has been done to create detailed representation of houses, shops, jazz clubs, streets, and people. Students who are studying literary texts or music of the period can immerse themselves in background details of the culture, and can even meet and question leading figures of the period (Carter 1999).

The future of publishing

Given the changes in access to textual resources over the last ten years, the whole nature of the publication process and its economic models have also begun to change. This is particularly true now that some kinds of works are increasingly being made available over the Internet, and is likely to affect teaching and scholarship in textual studies in ways that we cannot yet anticipate. At the time of writing this in the second half of 1999, Oxford University Press has just announced the publication (in March 2000) of the online OED, which will replace both the paper and CD ROM versions, and the Encyclopaedia Britannica is now provided over the Internet in its entirety, free to all, with revenue being generated by advertising on its site. Journals publishing has also been affected greatly, with some reputable scholarly journals now being available in electronic form only, for instance Postmodern Culture and Romanticism on the Net to name but two. In the sciences, the move to electronic journals has been even more marked and rapid, and I would predict a huge growth in this area over the next few years. One project of huge importance in access to journal materials is JSTOR, which is capturing complete back-runs of major journals and making these available on the Internet for browsing, searching, and downloading.

Interestingly, these changes in publishing brought about by electronic media are happening alongside some serious restructuring in the publishing industry. There has been a recent decline in book purchasing in all sectors, though there has been an increase in the number of booksellers, and a move within publishing towards large international conglomerates with smaller presses either being taken over or going to the wall. Paradoxically, ever more titles are being produced, but numbers purchased per title have diminished. In academic publishing, the university presses have in the past been protected from the vicissitudes of commerce by their charitable status and therefore non-profit structures. Previously, they could afford to produce books which would sell only a handful as their raison d'être was service to scholarship and teaching, not profit. We have seen great changes here, with a much more commercial thrust: market forces have come more seriously to the fore, as t he competition grows fiercer. This restricts texts available for research and teaching to those which will sell in larger quantities and therefore to the more mainline works of established authors of past or present. It is possible to sell endless new editions of Shakespeare, whereas lesser-known works are more risky. This makes academic scholarly electronic editing projects of non-canonical writers even more important, though it has to be said that funding bodies can be as conservative as publishers in resourcing better known (and better covered) textual works at the expense of the equally important but lesser known. A good example of a project producing largely non-canonical texts is the Women Writers Project at Brown University, which is creating heavily encoded electronic texts of works written by women in English, in the period 1330-1830. It is encouraging, too, to see that minority languages and cultures, and also now languages written in non-Roman and non-alphabetic char acter sets, receive good treatment on the Internet, with scholars having the opportunity to share resources over wider audiences than hitherto. The excellent Voice of the Shuttle Web site has links to many of these resources.

The latest commercial publishing development, announced in October 1999, is what was cited in a headline in The Independent newspaper of 19 October 1999, as 'Sell First, Print Later'. While this is providing paper rather than electronic resources, it is of importance here as it is a fundamental new move made possible by databanks of electronic textual resources and e-commerce. The Independent article by Jane Robins states

First there was the simple process of buying a book in a bookshop. Then, in 1995, came the Internet bookseller Amazon.com - which quickly sold tens of millions of books online. Now the transformation of book selling is about to enter its next phase: the mass introduction of 'books on demand'. (Robins 1999)

The article goes on to report how the world's largest publisher, Bertelsmann, is planning to use Xerox technology to take an order for a book, then run off a copy. 'Rather than print, then sell, we will be able to sell, then print', they say. The company plans to amass a vast database of out-of-print titles that can be printed on demand, then bound and sent out. This could be good news for teachers, researchers, and students who are suffering from the unavailability of less popular titles that hitherto have not been reprinted because of the cost of producing small runs.

Computers and teaching

The above discussions have ranged around resources for textual studies which can be used for either teaching or research. In the humanities, research very much feeds into teaching, with innovation in the former passing onto the latter, and the resources used are generally much the same. In UK higher education, there is not a 'set textbook' approach to textual studies. Rather, students read widely around a whole range of writers and issues in whatever subject, language, genre, and period they are studying. But over the last ten years, there have of course been some uses of resources which are specifically targeted at teaching and learning, and indeed some projects (STELLA at the University of Glasgow, for instance) which have developed products specifically for student use.

In higher education in the UK, we are increasingly operating in a self-driven learning environment, for reasons which are as much political and economic as they are pedagogic and intellectual. The rapid and under-funded expansion of higher education has meant that staff/student ratios have decreased markedly, and steps have had to be taken to ensure that the quality of education remains as high as before. One huge problem area is library provision, where library budgets have decreased in real terms, while student demand for books has increased rapidly as numbers grow, and as students become less able (or willing) to purchase books from their own dwindling funds. Decreases in the amount of staff/student contact time have meant that self-paced elements in courses, with large amounts of private study, have become the only way to ensure that sufficient work is covered. The Internet is being increasingly used to support traditional teaching activities, with email and electronic discussi on lists established to supplement face-to-face meetings. Documents and information are posted and exchanged on the Web, and virtual learning environments allow entire courses to be delivered online. Digital library provision and computer-based learning materials are two crucial aids to private study, and can add value beyond what has been lost if used well. They also provide access to rare or delicate materials that had previously been available to a limited number of researchers. For instance, one extremely important facility that can be added to the learning experience is access to digital facsimiles of unique source materials. The Virtual Seminars project, to give one example, offers original manuscript reproductions of First World War poetry (in particular that of Wilfred Owen) with edited texts of the poems plus video clips, images, and other background sources on the War to illuminate readings of the literature (Lee and Groves 1999). This illustrates a move in recent yea rs away from a 'packaged' approach, towards a contextualized, explorative approach, which new communication and presentation facilities are making ever more popular and feasible. Another example, one which has become paradigmatic for the development of hypermedia systems for teaching, is the Perseus Project. This is a large multimedia resource for the teaching of classical civilization which incorporates original texts, translations, dictionaries, explanatory essays, image databases of artefacts, video clips, and many other hypermedia objects. Perseus has been in continuous development since the mid-1980s - it started as a HyperCard program before being ported to the Web. From the beginning, it has entranced and seduced both teachers and learners.

There have been some interesting experiments in innovative teaching using the learning environments offered by the Internet. For example, Meg Twycross at Lancaster University has used the Web to teach palaeography, exposing students to a much greater range of manuscript sources in digital form than has even been possible in conventional media (http://www.lancs.ac.uk/users/english/palwork/index/). Students can enlarge the texts, compare them side by side, transcribe them, and then present the results with much greater ease than formerly, and the marking and analysis are also greatly facilitated. There is of course a cost in terms of time to prepare course materials, but this seems amply repaid by the results obtained (Twycross 1999). In the US, teachers of medieval studies at the University of Alaska and Susquehanna University, Pennsylvania, have co-taught a course in history and literature simultaneously to students, using the Internet (Kline and McMillin, 1999). 'Virtual environments', too, is becoming a field of study in its own right, existing at an intersection of interface design, performance studies, literary studies, and cultural studies (Warshauer 1998).

Conclusion

The ten years of CTI Textual Studies have seen great changes in the UK higher education system, in the use of C&IT in all aspects of life, and in attitudes to teaching and learning. The most significant event of this period has of course been the advent of the World Wide Web. Conceived of by Tim Berners-Lee in 1989, this has resulted in a paradigm shift in how we share information and a huge proliferation in resources available (some of which are of variable quality, unfortunately). When even your local hairdresser has a Web site, you know that something significant is happening. Throughout these changing times, the CTI Centre for Textual Studies has attempted to document and disseminate information about innovation, and has had a considerable influence in humanities teaching and the development of humanities computing.

black line