[A lecture given in 1990 to the Bibliographical Society of Canada held at the University of Prince Edward Island. An enlarged version of this was given to the Bibliographical Society of London.]
It is not inappropriate that I am back again in Canada at your kind invitation to say something about the ways in which my long-standing concern with bibliography and with computers has altered or modified my views of the nature and purpose of bibliographical enquiry. It was after all at UBC in 1952 that my interest in bibliography was aroused by Reginald Watters, and I worked as his research assistant on Canadian bibliography for two summers as well as part-time during academic sessions. My honours thesis was a bibliography of Swift (how inconsiderate of danger are the young!) and it is still on the reference shelves in the UBC library. A few years later, when I was studying to be a medievalist at Toronto, I became interested in computers and started with Jess Bessinger the concordance to Old English poetry. At Fredericton, in 1958, I began my Bibliography of the English Language: first as a revision of Kennedy but within a year I realised that nothing short of a fresh start would suffice. Other developments in my career have, of course, occurred in England but Canada certainly provided me with enough experience to realise that bibliography is a tedious enough discipline on its own: when you combine it with computers you had better keep a clear head at all times!
That we are living at a time of great change in the ways in which knowledge is produced, disseminated and acquired no one could doubt. Whether the many changes which affect our understanding of how books have been produced - and therefore how to describe them accurately and meaningfully - and which affect equally our perspective of the nature and extent of the bibliographical universe are entirely beneficial forms the substance of what one would like you to consider today. The changes affect virtually every aspect of that discipline which we call bibliography and the challenge now is to accommodate some of our most cherished principles to the new circumstances. What is the basis for understanding the transmission of text? Is authorial intention now recoverable? If not, does it matter?
Are the principles as we have received them from theorists like Bowers and Tanselle any longer relevant? What do we mean by the sociology of text in the contemporary environment which permits an author to take entire responsibility for the preparation, correction and presentation of his work? It seems clear enough to me that some of the vocabulary of bibliography - which works well enough with books produced by hand or by mechanical presses and metal type - needs re-defining. That the concept of issue, state, impression or edition may still apply I do not doubt. But how are they discoverable? The evidence, I suspect, is not easily accessible and in many cases (for example, the tapes in some publishers' electronic archives) has been deliberately destroyed.
And what of our dependence on bibliography to enable us to carry out our researches? With libraries, archives and record centres inexorably driven to catalogue their collections in machine-readable form it is already difficult in some institutions to get to the basis from which an item has been described. In the British Library, for example, we have disgracefully withdrawn from access the Title Room which contains the manual record of every item acquired by the library since 1757 so that it is no longer possible to reconstruct the changes which have occurred in professional and expert opinion on exactly what an item really is - these were all documented on the title slips as a matter of course. The published General Catalogue of the British Museum, which everyone knows, reveals only a small part of the intellectual effort which went into its creation. The same story can be told in many a library I have used in the past five years. But not only is it often not possible to see what lies behind a bibliographical description; we are, in addition, at the mercy of the software which is used to retrieve information from a machine-readable catalogue. But no software can retrieve what has not been encoded and I find it dispiriting that with all the energy consumed by international bodies concerned with standardisation there is so little understanding of, and even less concern for how researchers actually use bibliographical information. If our society is in thrall to the concept of the supermarket then I fear the worst for the future of bibliography, for that philosophy is predicated on market forces almost exclusively. And the motto is a simple one: 1f you can find it you can buy it!"
These matters are wide and complex enough to justify a full-length study so I cannot promise to furnish you with solutions in the space of forty minutes!
Let us consider first the primary obligation of a bibliographer which is, as de Bure put it long ago, 1he history and description of books." Between them the two substantives sustain most of the activities of the bibliographer: authorship; printing history; subject matter (the sociology of the book); the physical properties and location of individual specimens and the families to which they belong (the archaeology of the book). By following accepted principles it is just about possible to reconstruct the printing history of a text produced before 1964, and the likelihood of success may well depend on whether or not the researcher has access to the publisher's archives (if they survive). Why 1964? For a number of reasons. It was the year in which I started Scolar Press; but more important it was a year in which a number of developments which very few perceived at the time as all that significant:
By 1966 it had become clear to me that the world of book production would never be the same again and that if bibliographers in the future wanted to understand how books were produced after 1964 they would have to become familiar with a variety of techniques not described in Gaskell who maintains a dignified silence on developments after 1950. It is, I suggest, quite reasonable to articulate principles for the analysis of texts up to 1950 predicated on what the eye perceives. After 1964 the eye can be an unreliable aid. For example, one of bibliography's traditional descriptors - format - is becoming increasingly difficult to determine, due to the movement of the signature to the fold and the practice (in some libraries) of re-binding paperbacks using over-stitching and notched glued bindings. The abandonment of watermarks in reel-fed paper (as used on multi-station offset and rotogravure presses) means that yet another aid in determining format has all but disappeared.
By 1980 printing and binding companies had been transformed out of recognition, and I will tell you a story which illustrates that. I placed an advertisement in several national and trade papers inviting companies to submit specifications of their plant for what would be a massive publishing programme of reprints. In those depressing days the possibility of new business prompted virtually every printing works in Britain, Holland, France and Germany to respond. The sifting of what turned out to be several hundredweight of literature convinced me that the bibliographer of the future will need to have access to a reference library of printing and binding literature. No such library exists. I know, for example, that the first edition of Peter Porter's The Cost of Seriousness was printed by the Bowering Press in Plymouth in 1978, and that the 1987 reprint by Arrowsmith in Bristol is an offset replica. But how was the 1978 edition produced? I suspect film-setting because of the placement of punctuation marks and the inflexible letter-spacing.
A feature of modem books which will certainly confuse the bibliographer of the future is the fact that that while publishers are careful (on the whole) about which company prints their books they seldom divulge details about origination of the copy. There is some doubt about the identity of the first computer-produced book, and for a simple reason: in the 1960s publishers were not anxious to advertise their use of computers lest it got them into trouble with the trade unions. Andrew Garve's The Long Short Cut, published by Harper and Row in 1968 claims to be the first. It was set in 10 pt Videocomp Janson on an RCA machine running at 600 characters per second, and was printed from film positives made by The Haddon Craftsmen. Dent's edition of Dylan Thomas clearly has priority over Harper and Row, but there probably lurk earlier examples which only access to publishers' archives will reveal.
Between 1964 and 1980 the quarrels between the print and photographic unions and publishers led many companies to evade detection by simply being silent about production methods. I was involved in all this at Scolar Press and so can speak on the matter with some authority. The technological developments during this period are easy to chronicle:
Today we have the capability to assemble both text and images in colour on the office PC producing quite extraordinary results with 600 dpi colour laser printers. I recently visited a company near University College where I saw a 400 page book with 220 anatomical colour plates being assembled on a Macintosh machine with Quark Express. The fully composed film-pages were then sent to Taiwan for printing. Nowhere does the book reveal all this - for the simple reason that if medical students knew how it had been produced they would feel cheated at having to pay £ 120 to purchase a copy! How often do we find publishers taking advantage of technology to produce books more cheaply actually reducing their prices?
An early example of a book for which Camera Ready Copy (CRC) is the second edition of Principles of Interactive Computer Graphics by William Newman and Robert Sproull, published by McGraw-Hill in 1979. The authors explain in the Preface:
The Xerox Palo Alto Research Center very generously made its facilities available for what turned into a somewhat ambitious project to produce camera-ready copy directly from the edited manuscript. The line illustrations were created with the aid of an interactive graphics system, and the text was edited on-line, formatted, combined with the line illustrations and printed on a xerographic matrix printer.
This, as I have discovered, is the first instance of what was to become WINDOWS, the remarkable graphic capabilities of the Apple Mac, and VENTURA which brings typesetting into the home.
John Walsh, Production Manager at the Harvard University Press, assures me that the first book produced from an author's floppy disk was Ithiel de Sola Pool's Technologies of Freedom, published in 1983, the same year in which my Introduction to the first edition of ESTC was set from a Sirius disk which I supplied. Pool's book was produced at M.I.T. on a Honeywell 6800 word-processor. Although Pool acknowledges the contribution of various individuals there is not a mention anywhere in the book of the method of its production. Now, since it is unlikely that Pool's book will ever be the subject of bibliographical scrutiny that may not matter. But what of the Cambridge edition of Lawrence? Here, at least, we have a major twentieth-century writer.
Well, as I have discovered - and so might a future bibliographer with access to the Cambridge University Press archives - Cambridge struck a deal with Granada to publish the definitive text of Lawrence in cheap format without the notes. Cambridge agreed to supply Granada with tapes which they could format as they wished. The first book was Apocalypse - his last work. Knowing these facts, what possible reason could there be for doubting that the two editions are identical, except for typographical conventions such as the use of double or single quotes? Unless, that is, Granada received a tape which belonged to a different generation than that which produced the Cambridge text. And that is exactly what happened. Final corrections never reached Granada!
One final example concerns that cause cél.èbre The Satanic Verses. The English edition declares itself to have been "printed in Great Britain and filmset in Monophoto Bembo." The American edition states that it was "printed in the United States of America in Bembo." Except for preliminaries and divisional titles they are identical and could only have been printed from film supplied from a single source. Penguin/Viking are, understandably, reluctant to discuss the matter. But of one thing you can be certain: the tape exists which can, when circumstances allow, produce the copy for a paperback edition in a matter of days.
It seems clear to me that bibliography must take account of the fact that before long most writers will be composing their work on personal computers, and it will be commonplace to find poets taking charge of the entire page layout of their work. Now that Postscript printers are cheap and can produce positive film masters this seems not only logical but, for those poets who take an interest in the appearance of words on the page (like Yeats) appropriate. But what will happen to the physical evidence for all those transformations from the bare text to finished layout? It will, I am sure, disappear without trace. If publishers' archives can have a precarious fate, what of authors' papers? I know of many instances where publishers have either discarded their tape archive; or, what is more usual, find that there is no longer any equipment available to read them! That is one reason why all attempts to form a magnetic archive for the history of the book have failed. A second reason is the understandable reluctance of publishers to surrender what is essentially a substantial piece of capital.
Since machine-readable bibliography is now pervasive in most research libraries there are other uncertainties which bibliography will have to take into account in the years ahead. One such has to do with collaborative databases, or bibliographic utilities, such as RLIN and OCLC as distinct from those like ESTC and ISTC which are edited and maintained to as high a standard as is practicable. In the latter, for example, copy notes supplied by contributing libraries are retained and often shed light on the possible complexity of an edition or sequence of editions. With the exception of AMC (Archive/Manuscript Control) records on OCLC copy notes are infrequently found unless they are sufficiently distinctive to warrant a separate record. For national libraries which are responsible for producing the national bibliographic record copy notes are, almost by definition, ruled out. The copyright copy possesses, so to speak, an abstract and uncorrupt status. In the case of books in the British Library it is necessary to check not only the appropriate BNB file for the book but also the current catalogue where you may well find another copy which has annotations or corrections by the author. 'This is a factor about which our readers are entirely innocent.
Warren Haas indicated some of the changes affecting research libraries in a thoughtful essay in Scholarly Publishing in 1980:
That statement is as true in 1992 as it was in 1980; but we are no nearer a solution to the dilemma of balancing the needs and expectations of advanced research. Designing an OPAC for a library concerned, for the most part, with servicing pre-determined and well-understood needs is not very difficult. Success depends, generally, on a clear perception by the host of the needs of the user. For a large encyclopedic library that perception is crucially important, and more often than not conspicuously absent. A large complex research collection, such as the British Library, demands a collegial perception, resident not in any one mind, but shared amongst a skilled and scholarly staff who understand the collections for which they are responsible, and conscious of the defects of the various tools at their disposal. Richard de Gennaro predicted in 1979 (Library Journal, 15 November, 1979) that what was needed was
Haas went a step further and added:
Neither de Gennaro nor Haas specified precisely what would emerge from all this, but beneath the persuasive prose their lurks, I think, the very real possibility that what the prophets intended was that libraries should become information stores - or information supermarkets staffed by the familiar low-paid worker in a white uniform whose answer to all questions is "Don't ask me - I just work here!" And as the destiny of our research collections now rests more with supermanagers than with librarians who had some understanding of the needs of researchers the prospects seem bleak.
One model proposed to bring about the transformation of which de Gennaro and Haas predicted is the network, in which all libraries are interconnected and it becomes theoretically possible to discover from any node what is available everywhere else. It is technically feasible, of course, providing that there is a NASA-size budget available for the purpose. But there are a few details to be resolved first. One such will be the development of transparent software which can translate a search formulated according to a familiar protocol into one understood by a remote computer using an incompatible one, and containing data differently structured. Another will be the development of local systems capable of handling hundreds of simultaneous interrogations, many of which require connection to remote hosts. And, of course, there is the logistical problem of providing the hardware and the space for it, because the possibilities provided by automation typically yield results which take much longer to assimilate than consulting a linear catalogue. Access to a diversity of alternative sources, such as is envisaged in the planning of the new British Library at St Pancras, will serve only to lengthen the time spent at a terminal.
My experience at the British Library where I have providing research "surgeries" for readers for over two years using on-line technology convinces me that simply putting terminals in the reading rooms of research libraries will produce frustration and annoyance. What is just beginning to be understood is the fact that the consequences of the transformation predicted by Haas go far beyond what was ever imagined by 1ibrary managers". Electronic information cannot, by its very nature, respond to the supermarket model for selling food, household furniture or plants. In the model familiar to all of us we see what we want and either accept or reject it. But readers have no way of "seeing" electronic databases and little or no understanding of the rules and standards used to encode the data. Nor have they any interest in these things, being far too pre-occupied with their understanding of the information at their disposal. So I see only two alternatives: either they must learn about MARC (in all its myriad variants) and AACR2 (and all its immense corpus of rules and rule-exceptions); or libraries must supply them with a highly trained cadre of interpreters and navigators to get them the information they need. If the former solution should occur then we can expect a boom in library school registration! I think this unlikely. But the second solution will mean that curatorial staff in libraries (many of whom belong to a generation before computers began to dominate library administration) will have to undergo re-training both in the technology and in the different forms which machine-readable files can take. ESTC is not like ISTC, just as BLAISE is not like RLIN. And if we stir in Europe - a European Research Libraries Group is now being actively promoted - there is going to be a great deal to learn. And if Brussels ever gets a say in a European RLG then the future for bibliography might be even grimmer than I have suggested for we can confidently expect a flood of rules and regulations concerning the transnational EC record every bit as absurd as those which define the egg or the apple.
So, were the prophets wrong when, twenty years ago, they promised us a bright future in which information about everything would be available everywhere? As put by Joseph Raben in 1979 (Scholarly Publishing, April, 1979):
Whatever the historians of the future make of our period, they will undoubtedly record that it corresponded in its potential to the Renaissance, to the opening of men's minds to all kinds of new and old knowledge as that knowledge was made accessible through the technology of the printing press. If we can recognize the revolutionary nature of the latest technological advance, we may also contribute to another birth of learning.
I must say, I am unsure about living in a Renaissance, though I am sure about living through a revolution.