Up to twenty years ago or so, it was a common assumption that literature was not a subject one could teach. That was a peculiar assumption, singling out literature from other arts such as painting, architecture and even music, whose schools and workshops, where the master had his pupils working with him, had been a regular institution since the Middle Ages and the Renaissance. The idea that one can't teach literature comes from the new positioning of it apart from the older rhetoric and is a figment of Romantic and Modernist poetics, which saw originality as the goal of art: every artist has "to make it new" as Ezra Pound used to say, so it would be no point in repeating old forms and traditions.

Nowadays classes in creative writing call this assumption into question, thus considering literature again as a special kind of rhetoric. Indeed most professors of literature, and I belong with them, still think that all they can really do is teach some kind of practical and theoretical literary criticism, that is a rationalization of the process of reading, still not properly overlapping the process of writing.

As hypertext is supposed to make it possible to make the author's and the reader's roles overlap, two different interests get mixed: the creative and the re-creative. It may even be an improvement, allowing readers to experience part of the creative excitement. Possibly that's why the role of the teacher in electronic environments and network culture shifts towards creative writing classes.

For, though the reader's and the writer's roles may overlap, the implied author and the empirical reader converge on the text through quite different paths. Writers work hard, inventing, selecting, assembling and shaping. Readers on the contrary may be more or less active, but usually their attitude is simply cooperative. In a sense their pleasure lies in tasting a cake someone else made for them. At most, and Emerson, for instance, blamed Hawthorne for this, readers are called to witness the author making it.

Usually even if readers have to fill in some blanks, they like to be moved and pushed about by the force of the text. They like to follow a path, not to open it.

The author's pleasure is a different kind of enjoyment and comes from the successful effort of giving form to expression and content, of bringing a pattern in view. Even in hypertexts, though readers-as-writers can choose the order of what they read, they still have to play in a field and with rules selected in advance for them by the astute writer. I would say that the role of the reader is to domesticate otherness, the role of the author is to impersonate the other.

Indeed traditional readers cherish the idea of listening to the voice of the author as other. If the new readers become authors as well, they end up in the Narcissistic situation of listening to their own voices. And if what they hear is boring and unpleasant it's their own fault. So hypertext becomes a kind of Rorschach blot and the text with x is reduced to a test with s.

On the other hand if a reader wants to be a writer there are enough computers in the world, not to mention pens and paper, for him or her to write numberless texts without being a writer parasitic on someone else's work. So we understand that the reader as writer must be a new function that peculiarly tries to stay in between and this makes an interesting topic for study.

Then there is the question of author as authority, imposing the hierarchical order of the text on the powerless reader. This is taken as a metaphor of power relationships in society. I think that anything can be used as a vehicle of any tenor in metaphorical linking and thinking. It all depends, pace Althusser, on the interests of the interpreters; not on demonstrable evidence, but on suggestive yet often groundless associations.

The groundless association in this case is suggested by the self-importance intellectuals and artists have magnified for themselves as an autonomous social formation since the middle of the 19th century when the bourgeoisie got rid of them as spokesmen of middle class ideology. Since then artists and intellectuals in the humanities have often liked to believe that their revolutions in and on language and poetics were momentous passages in history.

Misunderstanding interactivity, hypertexts may let the readers as writers indulge in omnipotence dreams, like the handicapped gardener played by Peter Sellers in Hal Ashby's film Being There, the character who tried to change unpleasant reality by switching the TV remote control.

I think on the contrary that if hypertext effectually transforms ways of reading and writing, this will not have any direct, significant effect on power relations in the world. To believe it might would be a reductionism upward, rather than a reductionism downward as economism was.

Of course as a rule, power agencies, like any of us in our own environment, shall have to adjust their survival strategies to new technology, and this may result in some change of hierarchical positions, but it depends on more material competition than on change of metaphorical orders in language. As with the proverbial baggage-train, language as always will follow.

In any case while not refusing new technology as exploratory tools for criticism, most readers and students still feel bound to respect the fixed forms of pre-electronic literary genres, without feeling oppressed for this reason by the ruling class of authors. So we can figure out a range of uses of hypertext from the exploratory for the simple reader, to the creative for the simple author, to the mixed one of the reader as author.

Simple reading itself of course is not a simple process, as it involves perception, understanding and affection. Unexpectedly here the word "process" is a clue in the house of theory, linking literature and technology, old and new. For technology has always been a pre-condition of expression in terms of formulas in oral cultures, or in terms of handwriting and illuminating or printing, ever since writing was invented. But the process here in question is digitalization, the processing that texts undergo by computer programs for the treating of database, which open new horizons to critical insight into literary elaboration.

It's not difficult to imagine that, once mastered and widespread, e-writing tools will appear more or less as irrelevant as the pen and the book are today, and in a political perspective the organizing forms of e-publishing claim prior attention with respect to the transformations literature will undergo when electronically written.

Criticism itself both practical and theoretical implies specific techniques when implemented in the classroom. Beside a number of cultural units, a teacher always teaches himself or herself. As a matter of fact indirect education has a strong impact on students even in the university, especially when objects talked about, such as literary texts, have, inside and outside them, a lot to do with human personalities.

Lionel Trilling, for instance, coming to grips for the first time with contemporary literature in class, suffered with this situation, conceiving of the teacher of literature as a person exposing his or her own values and feelings, and so facing public judgement on matters of private significance.

This may be one of the reasons why sometimes members of creative writing classes discussing their works want to assume and switch pseudonyms, a habit I at first took for a paranoid symptom. Yet the imagery some hypertext theory indulges in: "weave, flow, on-goingness, interstitial", implies of course movement and flexibility, but also, and less obviously, thoroughness and liquidity. A positive perspective on such imagery can point out an ambiance fostering life origin and the principle of fullness; a negative standpoint could stress a regressive retreat into amniotic liquid and psychic inflation. In fact a student abandoned on the waves of the web is usually idealized as a new pedagogical Robinson Crusoe, while we know that, as in real cases of people abandoned to themselves, even if in a flow of information instead of on a desert island, they are likely to regress to a new form of savagery on line.

I've no time now to insist on this educational side of our profession, but here too, even more than elsewhere, new technologies can't help having a formidable impact. Now let us consider a few main paths that the digital treatment of texts opens to the scholar.

When we ask: what is literature?, we are told by present theory that literature is not an essence, but a practice; a practice that changes with time, so that it has no material recognizable features, but only recognizable uses. In turn the use of literature is to be framed along determined behavioural fields called literary genres, that can be viewed as deposits of reading habits which, for example, frame poetry in a perceptive pattern between painting and music, prose between document and dream, and theatre between spectacle and ideas. When we come to hypertext in this perspective, we shall have to see it as a new way of perception and consumption, whose key is provided by the word "multimedia" on one hand, and by the words "link" and "play" on the other.

Of course we would do well to distinguish genres from media. Traditional studies define genres according to either the represented matter inside the text, or to the ideal way the text is presented outside to its audience. So we have an internal and an external criterion to build a theory of genres, and more often than not they are confusedly mixed up. For instance not only should theatre not be considered a literary genre, but not even a genre, because it's a medium in itself, even if we can read dramatic texts in a book. Here lies the root of the principle that allows us to distinguish genres from media. And we also gain an insight into the nature of hypertext, hypermedia and multimedia whose very names betray their hybrid monstrosity.

Expanding Michael Joyce's motto "hypertext is, before anything else, a visual form", I would say that hypertext is a new genre, produced by a new medium, the personal computer, while I am still not able to assess whether software programs like Storyspace and others are sub-genres of hypertext or viceversa.

Anyway what I mean is that hypertext and hypermedia have to be used and appreciated according to new criteria of their own. This is all too evident, but it's also far too easy to say.

As when the novel appeared in the 18th century it had to be compared with epic for lack of new defining terms, so hypertext can't help being judged in a context of literary theories, while it's taking up a new place between textuality and visuality. Textuality and visuality in turn have to come to terms with virtuality. So the new position hypermedia are making their own is crossed by the reading habits of the various media by whose mix it is overrun.

In the last two centuries Western aesthetics and poetics have increasingly privileged ideas and the intellectual or cognitive side of arts, while their sensational, sentimental and ludic aspects have been degraded and exposed as popular, primitive, low-brow and unsofisticated.

As a matter of fact any tool hides a toy in itself. When a tool is no longer used mainly for its end but for sheer pleasure, a shade of perversity is projected onto its operation. Since hypertext is defined mainly in terms of its operational power and more attention is paid to the way it works than to the not yet completely evident purposes of its use, a semantic connotation of fictional, amusing not-seriousness haunts computer performances when applied to humanities in creative environments.

So hypertext, seen as a configuration of literature and other arts prone to a mainly ludic, intuitive and non-intellectual fruition, is blamed as regressive and bound to deadening compulsive repetitions, where moves, shifts, passages, operated through links, marginalize ideas, comprehension and meaning.

In this view -- the complaint goes -- the formal or structural identity of texts is no longer relevant, in fact it is form itself which is to be transcended. One may well insist that the actual form of hypertext is the reading in progress, but interactive, hypertextual re-formation of texts by any new reading makes them more or less infinitely multiform so that one begins to sense that no form is any more relevant than any other, and that relevance must have moved somewhere else, possibly in a vision of impermanence, and/or in a frenzy of linking, to escape a new kind of void horror, where hypertextualists dread to find themselves dropped as soon as they stop connecting anything with everything.

On the opposite side hypertext theorists and practitioners such as the artists and scholars attending our conference today appreciate and highlight hypertextual configuration as the actualization of new visions and perspectives, and of what post-structural theory had only been able to conjecture as mere hypotheses.

Hypertext as a tool for literary criticism operates on many levels with different functions. At its lowest it can simply speed the retrieval of notes and critical comments; but I believe that its quick automatic working foregrounds aspects of criticism usually played down.

Criticism is an addition to the text, but more isn't necessarily better. It may be felt as an imposition on the readers, conditioned by someone else's opinions, or deprived of the pleasure of discovering connections and significance by themselves.

What has been said above about power relations inscribed in the order of language could be repeated for hypertext as a tool for criticism, though hypertexts claim the opposite. Hypertexts try to stretch the cooperative role of the reader to its limits, but in so doing they make these limits more visible. So critical interpretation, more than ever in hypertextual form, instead of emancipating the readers, may seem to tell them what to think and what to enjoy, as if they weren't able to understand, think and enjoy without help. Hypertext links may be felt as marking a kind of abridgement and of facilitated reading for stupid readers. Moreover visual actualization of mental pleasures may be felt as a lowbrow, vulgar use of the text, like an explained piece of irony.

Visual actualization in itself is the diagrammatic projection on space of something, a literary text, made of meaning and pleasure, not only richer and more complex than a graph or a scheme, but absolutely different in nature. For instance links, like positions on diagrams, show potential associations but not the motives of the associations themselves. As intertextuality overwhelms any other interest, psychological, sociological and historical, one may rightly wonder if, in space-articulated structures, specifically spatial link motivations overrule cultural ones. In topografical language visual patterns might impinge on semantic ones much more effectively than in, say, Apollinaire's calligrams or concrete poetry, and literature move thereby near to geometry or informal painting.

Paradoxically through structural visualization hypertext highlights something all of us already know: namely that criticism is always a selective, if not impoverishing, reduction of text and of its experience both as words and as structurally organized words. Though desperately trying, hypertext as criticism can't evade the ontology of all criticism, that is a sign of a sign, but too simple a sign of a more complex one.

Or is it that the text as sign, when it's a literary sign, should be seen as an object, a concrete though universal one, as the new critics ages ago spent their lives to ingrain into the young people we were at that time? If so hypertext makes literal a figure of speech, turning textual objectivity into a visible object, at least producing a blueprint capable of suggesting a model of an object.

On this path we are making for a procedure experimental sciences work with, where simulacra, in the correct scientific sense of the term - not Baudrillard 's - are used for prospecting the results of real situations. Along this path we are also making for an interpretive method where, through visual homology, a text is translated into a picture, the verbal is shorthanded into the visual, with all the advantages and disadvantages of any translation.

When for example Jay Bolter says: "signs and structures on computer screen… have no easy equivalent in speech" (quoted in Joyce 1995, 177) I wonder what the word "equivalent" here means. If by "equivalent" he means the same visual aspect of the signifier, he is right. But if the word "equivalent" is to be taken as the same meaning signified, then he is wrong. So the point is whether in a non-poetic statement the signifier is or is not relevant. I think it is not. So: does hypertext, highlighting visuality, aim at making of any statement a poem?

I guess that what Bolter cares about is the fact that graphs and images may be more impressive than words. Now graphs and images may be emotionally more intense, but they are nonetheless more undetermined intellectually. So they may be allowed to be more effective to persuade a customer, perhaps a lover, but not a student.

Indeed, as in the case with any abstraction, diagrams select some aspects of the represented object and eliminate the rest in order to foreground in the graph the elements they select, elements that otherwise would in the object itself be less visible. Thus they are useful in improving our perception of just those elements when we go back again to experience the object. Diagrams on the other hand are no more sensual and real than any other sign or symbol in verbal communication.

So even more delusive is the claim that, through topographical writing and configuration, hypertext makes up for a proprioceptively and sensually shaped space, "the way the body knows" (M. Joyce, 1995, p.8). I don't know if virtual reality is so ahead in its development, but so far, as all of us know, the chemical formula of water H2O is not wet. Moreover literature is made of words, and has a right to be so. Possibly literature as we know it is about to die, but so far it doesn't want to be mistaken for reality even less the virtual one.

Thus hypertext unites structuralist illusions that pleasurable meanings shaped in words can be fixed into diagrams, with post-structuralist claims that no fixed form can represent the force of the text, not to mention the trace of a meaning (and/or signifier) which fluctuates with every new reading.

Of course hypertext is a quite different object if it's an autonomous creation with no intention of explaining and of devouring pieces of literature. Here hypertexts are not tools but works of a new art, made up of verbal, visual and cinematic techniques. Here the tool can rightly disclose the hidden toy it contains, playing on the intertwined properties of games and art.

In short, art can be seen as a game played with no rules and purposes but the ones it creates itself under the label of style or poetics. In modern art the rule is invention. Games on the contrary have to be played with given rules and with an established aim. But within the rules players have to operate creatively to achieve their purpose.

At this point we realize that, like all arts, hypertext too is the byproduct of its peculiar and specific operative power as a tool. But now the tool power has moved from given utilitarian purposes to other goals, self-exploratory and consequently aesthetic.

Hypertext may thus be a tool for criticism in so far as it employs its technical skills to elucidate the rules that literature gives itself as style and structure. It becomes a new form of art when self-celebrates its power of mixing words and other symbolic forms. It seems to hesitate bewildered betwixt and between when, in translating words into diagrams, stretches literature to the point of graphic parody. And at this early stage it's more than natural that the most typical feature of hypertext seems to be "in-betweenness".

To come to a provisional conclusion, I want to emphasize an achievement hypertext sooner or later will drift towards, an achievement that is implicit in the core of its data processing skills. Indeed the principle of fullness underlying the processability of data of which hypertext is an output may be compared to the chemical process of solution by which the elements of a substance, in this case the text, are dissolved from their previous sense-giving shape and structure, and precipitated into a uniform and formless base, a kind of reference container, where they remain, awaiting the summons of the casual, passer-by reader as author. It's as if Shakespeare were to be deconstructed into a dictionary of the English language, from which of course anybody would be able to write Shakespeare works all over again and many others: besides why should I say many?!, let us say all other possible texts of English literature, past, present and future.