Total Knowledge

In the late Fifties, at university in New Zealand, I did the kind of degree in which you are allowed to mix subjects. I spent my first year reading philosophy, geology and English. I never quite got a grip on these subjects, and certainly have not kept up with them, but the memory of what it is like to do philosophy or geology remains; and when I read about debates that are going on in these areas I believe I know, even if I cannot follow it all, what kind of row or celebration is taking place.

Subject-styles, back then, were well reflected in the content of bookshelves. The geologists had journals, off-prints of individual papers and a few large reference books. The philosophers’ shelves were sparsely filled with canonical texts from the Pre-Socratics to Wittgenstein and slimmish books by a few modern, mainly English contemporaries. The English lecturers had lots of books. Not just the texts of their specialty but modern poetry, criticism and so on. Shelf after shelf of them.

The state of the shelves seemed to reflect pretty well the culture of each subject. Geologists went on adding pieces to the jigsaw, paper by paper. Philosophers went on circling around the questions philosophers have always circled around. Academic critics tried to find new ways of-writing about old books, or new ones to write about (‘Why not Commonwealth literature?’).

I was an inveterate picker-up and putterdown of books because I was interested in how they looked. I got to care more about how they were put together and organised than about their content. Indeed, after the first couple of years of university I became a compositor’s apprentice. To follow intellectual history by way of the history of print is rather like getting your political history from fashion magazines. But writers began to turn up – McLuhan, Ong, Eisenstein – who suggested that the shape of knowledge and the shape of books were more than casually connected. As time went by, and electronics took over from mechanical systems for data storage and typesetting, the same was said of computers. George Landow’s Hypertext is about one way of using computers to manipulate texts. He makes high claims for hypertext, a richly interconnected kind of database. ‘Critical theory,’ Landow writes, ‘promises to theorise hypertext and hypertext promises to embody and thereby test aspects of theory, particularly those concerning textuality, narrative, and the roles or functions of reader and writer.’ Landow teaches literature, and his bias is towards the wide ranging linkages which the well-filled shelves of the English department’s offices represent. Before following his line of argument it is as well to distinguish this kind of linkage from the more disciplined nets indicated by the sparser shelves of the geologists and philosophers.

A little while ago an article in Nature, by Donald Hayes, set out to show that science journals are harder to read than they used to be. The test Hayes uses gives a numerical measure of ‘lexical difficulty’. It is crude, but probably pretty accurate. An analysis of the vocabulary of printed and spoken English gives a scale which runs from plus 50 for the hardest science, down through single figures for popular magazines to 0 for newspapers. It bottoms out at minus 60 for dairymen talking to cows. In samples from scientific journals recent issues get higher ratings than those fifty, or even ten years old. The result quantifies an obvious fact: science has become more specialised and it is harder for experts to understand what experis in other fields are saying. A few weeks earlier, the editor of Nature, John Maddox, reported on a conference about scientific journals and electronic publishing. Now that electronic mail (email) allows instant publication, printed journals carry a time penalty. Moreover, as email contributions and commentary appear side by side they mimic a primary function of journal publication: validation by peer review. One physicist is quoted by Maddox as saying: ‘technological issues are already settled ... we have learned to determine by title and abstract whether we wish to read a paper ... the small amount of filtering provided by refereed journals plays no effective role in our research.’ Although general magazines like Nature and Science still function as forums in which cross-fertilisation takes place, Maddox was given pause by a statistic from Anne Ockerson of the Association of Research Libraries in Washington who noted that of the 1.5 million items her member-libraries supplied to each other last year, an estimated two thirds consisted of single copies of articles.

In science the unit of information exchange, the single paper (printed, emailed, circulated as a preprint or whatever) is a node connected by references to other nodes in a continuous net. One can map the growth of science by registering new nodes. Cast an eye over the whole subject and areas of activity show up. The citation index of an article gives the number of connections made to the node which represents it in the net. Nature, touting for subscribers, advertises the fact that ‘latest ISI citation figures show that papers published in Nature occupy seven of the top ten in a list of “Red Hot Papers of 1991”.’ Three of the top four in the list are about a new form of carbon – C60. This is a measure of activity rather than importance, but no matter how original a paper may be, it cannot change the shape of science unless it is well-connected. The active surface of science is thin, like the layer of living tissue below the bark of a tree. It is supported by the wood of old ideas (all wood is, strictly speaking, dead). As time passes, what was the surface is overlaid by new work and itself becomes part of the wood. Distances on the surface get greater, for the whole organism has grown larger. Although this sets a limit (measured perhaps by ‘lexical difficulty’) on direct communication between specialties, they are still joined at lower levels.

An overview of papers in the humanities would also show which areas are active and how ideas pass from one author to another. But the connections, both lateral and back to past texts, are different. Aristotle’s biology was long ago overlaid by better approximations to the truth: his Poetics must still be engaged more or less on its own terms. In the dialogues of the humanities, ideas do not have the power to exclude each other, in the way molecular biology, for example, excluded vitalism. Barthes, writing of strands in French literary criticism, says that ‘Since these different ideological principles can exist simultaneously (and for my part, I can, in a certain sense, accept both simultaneously), we have to conclude that the ideological choice is not the essence of criticism nor ‘truth’ its ultimate test. Criticism is something other than making correct statements in the light of “true” principles.’ One does not even have to accept relativism as extreme as this to acknowledge that the impossibility of a resolution is a typical and even admirable property of debates in the humanities. In this context Landow quotes Rorty’s description of the ‘edifying philosopher’ whose purpose is to keep the conversation going lest people settle on some form of discourse which is seen as final, and which would result in ‘the freezing over of culture’ and the ‘dehumanisation of human beings’.

In the sciences the problem is to get information around efficiently, and to see that specialisation does not involve destructive isolation. In the humanities it is to keep discussion on the rails, however many times the track may bifurcate, and despite more and more material piling in on top of old but not obsolete content. The fear in the humanities is that a time will come when the great universal seminar gets out of hand, when there will be too many members for any one to be heard out, or even for a general understanding of what the issues are to be achieved.

Technology threatens to accelerate this process by offering us the dream/nightmare of a possible ‘docuverse’ – a database made up of the totality of all texts. All writers from all times will be waiting there to leap onto your screen at the touch of a key. A few small parts of the docuverse are already in place and give a taste of what is to come. The catalogues of most libraries are being computerised, and projects like the Eighteenth-Century Short Title Catalogue (ESTC), the computerisation of the Witt Library, and the publication of all printed verse in English on compact disc are complete or under way. Connections which up to now have depended on exceptional memory, serendipity or laborious manual searches will become easy. Traditional dictionaries, chronologies, book indexes – concordances even – allow texts to be searched in a few dimensions: the new ones allow them to be searched in many. The laser-disc version of the OED uses indexes which take up as much space as the text itself, making possible searches by (among other things, and in combination) date, author of quotation, phonetics and philology. But even extended indexing of this sort limits the inquirer to the compiler’s notion of what is significant or the cataloguer’s summary of content.

Which brings us back to Landow. The idea of hypertext began with a paper called ‘As we may think’ by Vannevar Bush, published in the Atlantic Monthly in 1946. An electronic version of the physical solution Bush imagined – a desk with screens on which mechanically retrieved microfilm was to be projetted – is now in place, in the form of computer screens, in most offices, and just waiting, in theory, to be joined up to the docuverse of all recorded texts (and sounds and pictures). Theodor Nelson developed Bush’s ideas. It was he who called the navigable docuverse ‘hypertext’ and defined it as ‘non sequential writing’.

To get some idea of what using hypertext is like imagine yourself in front of a computer screen. It tells you that you have entered a document called Mother Goose. A keyboard and a mouse allow you to interact with what is on the screen. Among the options offered is one labelled ‘Themes Overview’. Click on this and a dictionary of themes appears. Click on ‘Food’, and a list appears which includes ‘bread/loaf’, click on that and a list of rhymes is offered. Choose ‘Ride a cock horse’. (Any of these steps, by the way, can be reversed.) The text of the rhyme is now displayed on the screen. You have in fact accessed the Opies’ Oxford Dictionary of Nursery Rhymes.This includes their footnotes, which you call up by clicking on the little arrows which appear, like asterisks, above words. The note which refers to the ‘pastry cakes of Banbury’ leads you to another hypertext document altogether, this time one on food technology, through which you access the entire texts of several books on baking. You now type in a brief note on Banbury cakes, including a piece of local lore which your own researches have turned up, and extracts from the cookery books. Note that your activities have changed the hypertext document in two ways: you have added to it, and, if you save the information about the path you followed, you have created a record relating various items of information. While you have not altered the Opie-text, you have in a sense added to it, and the next user of the document will be able to draw on your contribution by taking note of the path you followed or by reading your note.

The starting point of Landow’s book (subtitled ‘The Convergence of Contemporary Critical Theory and Technology’) is a coincidence of vocabularies. He observes that ‘in SIZ, Roland Barthes describes an ideal textuality that precisely matches that which has come to be called computer hypertext – text composed of blocks of words (or images) linked electronically by multiple paths, chains or trails in an open-ended, perpetually unfinished textuality described by the terms link, node, network, web and path: “In this ideal text,” says Barthes, “the networks [reseaux] are many and interact, without any one of them being able to surpass the rest; this text is a galaxy of signifiers, not a structure of signifieds; it has no beginning; it is reversible; we gain access to it by several entrances, none of which can be authoritatively declared to be the main one; the codes it mobilises extend as far as the eye can reach,they are indeterminable ... the systems of meaning can take over this absolutely plural text, but their number is never closed, based as it is on the infinity of lan guage.” ’ Scientific papers, which normally deal briefly with one topic, provide manageable chunks of information. To give texts in the humanities (novels, long treatises) the same ‘grain’ when they are included in hypertext they are divided into lexia (a term Landow has borrowed from Barthes): the entities between which cross-references are made.

Hypertext destroys linearity. The order of chapters and sections ceases to be apparent when a text is accessed by way of connections with other texts. Derrida, Landow says, recognises that ‘a form of textuality that goes beyond print “forces us to extend ... the dominant notion of a ‘text’ ”, so that it “is henceforth no longer a finished corpus of writing, some content enclosed in a book or its margins but a differential network, a fabric of traces referring endlessly to something other than itself, to other differential traces”. ’ Hypertext, Landow says, ‘thus creates an open, open bordered text, a text that cannot shut out other texts ... undergoes what Derrida describes as “a sort of overrun [debordement] that spoils all these boundaries and divisions” ... Because hypertext systems permit a reader both to an notate an individual text and to link it to other, perhaps contradictory, texts it destroys one of the most basic characteristics of the printed text, its univocality ... It also makes all the texts connected to a block of text collaborate with that text’.

There is of course a down side, particularly if you wish to colonise textual space, to brand what you write as all yours. Landow quotes Edward Said: ‘the proliferation of information (and what is still more remarkable, a proliferation of the hardware for disseminating and preserving this information) has hopelessly diminished the role apparently played by the individual.’ Michael Heim, among others, fears Babel: ‘Fragments, reused material, the trails and intricate pathways of “hypertext” ... advance the disintegration of the centring voice of contemplative thought.’ Such dissent, Landow says, exemplifies the line of criticism which ‘derives the erosion of the thinking subject directly from electronic information technology’ – the kind of reactions against hypertext one will get from ‘the print author accustomed to the fiction of the autonomous text’.

Hypertext, Landow says, ‘promises to have an effect on cultural and intellectual disciplines as important as those produced by earlier shifts in the technology of cultural memory that followed the invention of writing and printing.’ His enthusiasm for hypertext grew during his use of a teaching system called ‘Intermedia’ (he thinks ‘learning system’ is a better description). To explain its potential, he quotes a distinction Terry Mays makes between implicit and explicit learning. ‘Explicit learning involves the conscious evaluation of hypotheses and the application of rules. Implicit learning is more mysterious: it seems almost like a process of osmosis and becomes increasingly important as tasks or material to be mastered becomes more complex. Much of the learning that occurs with computer systems seems implicit.’

Educational hypertext transfers some of the powers of instructors to students and Landow praises it for practical and organisational reasons as well as deeper ones: more term papers get written on time; more texts can be made available; distance learning is made much more like campus-based instruction. There are psychological advantages, too: students who are backward in discussion are often able to contribute better on a screen, and hypertext aids the unusual or precocious student by making whole contiguous provinces of material available. And, of course, it extends and subverts the canon. Landow is entirely convincing about all this. He gives an attractive picture of collective learning and of the way it breaks down hierarchies among texts and people.

All this is good news for teachers who are not too sensitive about their intellectual authority, and for educational administrators. It is bad news for print culture. Landow writes, that the technology of the printed book ‘engenders certain notions of authorial uniqueness’ that hypertext makes untenable. And that hypertext thus has much in common with ‘some major points of contemporary literary and semiological theory, particularly with Derrida’s emphasis on de-centring and with Barthes’s conception of the readerly versus the writerly text. In fact,’ Landow concludes, ‘hypertext creates an almost embarrassingly literal embodiment of both concepts.’ Why the embarrassment?

Because, I guess, if hypertext, the realisation of a piece of middle-weight Forties futurology, really explains so much of what Derrida, in particular, is about (Landow explicitly rejects the thesis that a complete mapping is possible), the high jinks which Derrida has got up to to liberate readers from falsities of the text seem more like showing-off, less like necessary therapy. Whether or not this is the case, Landow’s picture of what hypertext could do is concrete enough. One can look for the signs of overload in current technologies of information exchange which would give hypertext a competitive advantage. Alternatively Landow challenges one to identify aspects of print culture which will be resistant to hypertextualisation.

First, is it really effective? So far as one can see the simpler tools of electronic text analysis (concordances, catalogues and databases) have speeded up – but not transformed – the more mechanical kinds of scholarship. There just aren’t enough projects in the humanities where getting all the data lined up seems to promise a step forward in understanding – no equivalent to climate studies or genome sequencing. Will hypertext be different? A lot hangs on an asymmetry Landow refers to: in the humanities writers have tended to work alone whereas in science work is usually carried on collectively. The success of the scientific information system stems from this sharing of the workload. Ideas grow in collaborative sessions, large-scale experiments need teams of workers.

Landow suggests that it is not like this in the humanities because funding is different, and research an adjunct of teaching. Hypertext is a door, he thinks, to a world of collaborative works, as effective as the single-voiced ones print culture has supported. This seems unlikely. The strong rhetorical element in the humanities, the need for an identifiable personal voice to carry the argument, precludes it. What carries conviction in a debate in the humanities is work in which ideas and their expression cohere and become a single entity: it is not just Hume’s conclusions, but his locutions, which one returns to. To establish a tone of voice takes time; the single lexia the note to a note, embroiders but cannot give primary shape to, an idea. The authorial voice, even if it can be proved to be an artifact, is a necessary tool.

Science can be successfully condensed. Its results can be inserted into arguments as provisional truths. The equivalent counters of discourse in the humanities (‘Plato’s theory of forms’, ‘Eliot’s Anglo-Catholic conservatism’, ‘Amis’s misogyny’, ‘aleatoric elements in Bacon’s painting’) have meanings which are a function of the reader’s experience. You cannot send anyone out to have the experience of Bacon’s paintings for you. This sets a limit to the territory any exchange in the humanities can cover. The frame of reference cannot be extended by joining your intellectual province to a logical net. The fear that knowledge will become Balkanised is therefore real, and hypertext, although it can offer no solution, does give a picture of what a net might be like that was so large that a challenge at one point would not be felt throughout. If this is how things are, it is unlikely that ideas as pervasive as those of Marx and Freud will occur again, outside the sciences.

The problem is illustrated by the contrast between what was said by the physicist who told Maddox that abstracts were enough and the preface to Michael Dummett’s Frege: Philosophy of Mathematics, in which Dummett launches a passionate diatribe against the corruption of the universities by competing pressures. In particular the pressure to publish. With only one book published in his lifetime. Wittgenstein, he points out, would not have survived; Frege’s citation index would have been very low – his works were hardly referred to by his contemporaries. The pressure to publish, Dummett says, distorts scholarship. He goes on:

  It is not, however [just] that quantity is not the only criterion, but that it is positively harmful ... In the long run it may be impossible to keep pace with the spate of books and of professional journals ... once this happens, their production becomes an irrelevance for the working academic, save for the occasional book or article he happens to stumble on. This applies particularly to philosophy. Historians may be able to ignore much of their colleagues’ work as irrelevant to their periods; but philosophers are seldom so specialised ... they cannot afford [time] to plough through the plethora of not bad, not good books and articles in the hope of hitting on the one that will truly cast light upon the problems with which they are grappling; hence if they are sensible they ignore them altogether.

Hypertext will speed up the processes of looking up, leafing through, and manual copying, but it won’t make it possible for a single mind to grasp all that is there to be grasped in many provinces of knowledge. What hypertext might do is persuade you that the system on which your judgment of ‘not bad, not good’ is based is illusory, that everything is relative, and that any attempt to say ‘this is how the subject stands’ is defeated before you start.

Landow’s description of hypertext, and his equation of its operation with the propositions of critical theory, brings one up against differences in kind between connection in the humanities and in science. What can be shared, in the sense of being made available for inclusion (which is how results are shared in the scientific information net), and what can be shared in the sense of having a representation in the mind of the participant, which is how ideas are shared in the conversation of the humanities, are different things. Against the model of growing interconnection one can set that of secret gardens of thought with nil citation indexes, or of devices which limit and formalise connection, and, to the irritation of many, keep a firm hand on the way the conversation goes – in other words, assert the virtues of print culture’s limitations. Despite exclusions, partiality, blind spots and bad habits (and being open to all kinds of deconstruction), the kind of information control which goes on in magazines like this one provides a context for the exposition and criticism of ideas which is organisationally superior to the fluid structures hypertext promises. Print structure is bought at the cost of less than perfect connection – the price science cannot pay. Hypertext which Landow says is ‘simultaneously an enthusiastic hard sell, a prophecy, a grim warning and a report from the front’, has the effect of making total connectivity seem a delectable toy, the ultimate computer game, but Dummett is probably right – the only way forward in the humanities is to selectively disconnect.