Literary Recollection: The End(s) of Intertextuality
There is a caricature of Marcel Proust in which the despairing writer is consoled by a friend saying, "Aber, aber, mon cher Marcel, nun versuchen Sie sich doch zu erinnern, wo Sie die Zeit verloren haben "
Literature in general, not only in the vein of La Recherche du Temps Perdu, deals with a different form of memory than that of mnemonics, in which the hints of places lead to a retrieval of what has been stored there before. Nevertheless it is difficult to pinpoint the criteria that make this difference. How does literature transcend the technologically limited sense of memory in terms of a storage and retrieval system? In this short paper I can only try to give a short answer to this question.
First of all, we have to realize that the poetics of memory is not an obscure quality soaring above the technical use of classical mnemonics. "Story" and "storage" have the same roots, as Thomas Wägenbaur pointed out. And it was a poet, Simonides of Keos, who is said to have been the inventor of mnemotechniques. The written word Plato later straightens out for once and for all is caught up in the task of data storing per se. Paradoxically, however, Plato's criticism of the hypomnemata which cause the facility of anamnesis to wither away, is passed on to us in writing. Is this a case of a performative self-contradiction? Is the entire history of literary remembering perhaps such a contradiction?
Indeed, it is in a specific way. Literature can be seen as a technique of deconstructing its own mnemonic function, and this is what I call the technique of recollection. What makes it separate from bare mnemonic devices is the opening of the textual spaces in which knowledge is stored, enabling the recipient to "read between the lines". As Renate Lachmann puts it, echoing Julia Kristeva, "The memory of a text is its intertextuality." This can express itself in different ways whose underlying principle is the awakening of an association a hint at something which is absent in the mesh of literary texts. This may or may not be the author's conscious intention. In this manner we can read between the lines of Gibson's Neuromancer an attempted break-out from the deadly confines of techniques of memory (here: in the form of a holographic computer simulation), which is somewhat reminiscent of Hofmannsthal's drama Der Tor und der Tod. Here we have a similar idea that is brought to expression as the cathartic awakening of an aesthete who suddenly realizes that he has forgotten himself in the "holographic book dream" he has been living in, hoping to overcome this at least in death. This, again, is reminiscent of another literary scene, namely the opening of Goethe's Faust in which the misleading "theatre" (V. 454) of a macrocosmic vision is dispensed with in favour of a physical experience of existence that can only be had at the cost of one's own life. Faust's criticism of the macrocosmic theatre is again an intertextual reference to the attempts in Renaissance magic to give a neoplatonic content to popular contemporary theatres of memory, and by this to overcome its mnemonic structure allowing the process ofanamnesis to be experienced. That brings us back to Plato, whom I would like to introduce to you as the first writer to have consciously deployed the anamnetic function of intertextuality avant la lettre. The next step is to draw a line from the dawn of literary remembering techniques to their dusk the loss of intertextuality, technically brought about by the hypertext.
Plato and the Origins of Intertextuality
Our starting point was the question of how Plato could unfold his criticism of writing through writing, i.e. how he could use the medium itself to overcome its own limits. In Phaedrus, Plato clearly states that writing paralyses the activities of remembering by means of prothetic substitution. There is little point in asking psychologists if this is really the case. Their findings are ambiguous as can be seen, for example, in two famous cases of memory malfunction described by Alexander Lurija. These are the "vast memory" of the mnemopath Schereschewski who suffered from the flood of remembered images, and the amnesia of the brain-damaged Sassezki, whose mental aphasia had lacerated the cosmos of his memories. One wrote to forget, the other wrote to remember. Both failed. We therefore have to differentiate and address the question of how writing is used in order to find out under what circumstances it promotes or, as the case may be, hinders, remembering or forgetting.
In Schereschewskis preoccupation with freeing himself from the over-determination of his memory, we can see an attempt to verify the hypothesis of the Phaedrus-dialogue, namely that the process of writing enables us to forget. As Schereschewski writes,
People jot things down so they'll remember them. This seemed ridiculous to me, so I decided to tackle the problem my own way. ... Writing something down means I'll know I won't have to remember it.
What appeared to the mnemonist as salvation is exactly the effect Plato had anticipated. Writing, he wrote,
will produce forgetfulness in the minds of those who learn to use it, because they will not practise their memory. Their trust in writing, produced by external characters which are no part of themselves, will discourage the use of their own memory within them (275a).
This passage is repeatedly cited by those who claim that Plato is a radical opponent of writing who priveleged the oral medium to avoid the side effects of this deceptive pharmakon. I do not think this claim can stand up to criticism. Contrary to the popular belief that Plato's dialogue plays oral speech out against writing in the interests of recollection, I interpret Plato as placing the line of argument crossways to the opposition of both media: Both can either promote or hinder the process of anamnesis depending on how they are used. In a double sense poor Phaedrus takes on the role of negative exemplary. He has just heard a lecture of the rhetorician Lysias and requested this in the written version so that he can learn it off by heart. Now he wishes to try it out on Socrates to see whether he can do as if it emerges as new from his recollection. But the declamation falls flat. The sceptical philosopher who becomes suspicious at the mention of the name Lysias glimpses the notes under Phaedrus's cloak and suggests the amateur mnemonist would be better-off reading the original.
What is problematized here is the purely mechanical reproduction of writing and speech. Plato had already introduced the alternative model in Meno: By clever questioning Socrates managed to decipher the slave's false, mechanical answers until he had reached the mental state of aporia, the precondition of true anamnesis. 
Thus, destroying the false belief that one can speak well is opening the mind for truly speaking well. This is also demonstrated in Phaedrus: Socrates gives a better speech than Lysias because he does not reproduce preconceived rhetorical knowledge. He leaves it up to the muses, the daughters of Mnemosyne, to decide whether he can deliver his theme convincingly or not (237a). That is to say, he faces up to his own aporia. The opening for that which is not already contained in memory sets an anamnetic process in motion. This process is for Socrates a sort of "madness that is heaven-sent" (244a): the soul's undistorted recollection of beauty once beheld (249a-250c).
Analagous to this example of good speech are the utterances that follow on from it about "how to write well" (258d). The criteria remain the same: By means of the respective medium it is necessary to overcome its purely mnemonic usage in order to allow the creative process of recollection to arise.
It is not paradoxical that Plato should have written the Phaedrus-dialogue down. It is instead the logical implementation of the methods it elucidates for producing an aporetic situation. When Plato writes that people's "trust in writing will discourage the use of their own memory within them", the discredited medium appears, due to the self-destructive manner of its use, as the means of its own transcendence.
In the course of the dialogue, this manner of use is subjected to considerable variation. I am indeed convinced that all three models of intertextuality mentioned by Renate Lachmann in Gedächtnis und Literatur participation, tropes and transformation are contained in Plato's dialogue. In this paper, however, these can only be fleetingly alluded to:
The participatory repitition and continuational writing of previous texts is carried out in extraordinarily numerous allusions and reminiscences that are explicit and to an even greater extent, implicit. These reminiscences cite the entire spectrum of the educational canon in antiquity from myth through poetic writing to rhetorical didactics and insodoing make themselves transparent as text. In the centre we have Lysias' speech, imitated meticolously by Plato, while exaggerating "particular stylistic charactistics".
The writing against forerunner texts in the sense of tropes shows itself in the socratic opposing models to Lysias. They undermine the existing text in two steps. Before Socrates can get around to the speech he sets out to achieve namely the countering and surpassing of the older text he delivers a lecture that picks up on Lysias' argumentation and exceeds it. He begins with a stereotypical incitement of the muses, gets underway, interrupts himself with the observation that already "his style is not far from dithyrambic" (238d) to finally conclude that he will indeed have to slow himself down because he is already "breaking out into epic verse" (241e). The apparently spontaneous speech that conforms perfectly to its form propels itself on ad absurdum, turning into its own caricature.
Insodoing it establishes the necessary distance for the intended transformation of the forerunner text. Socrates' second speech now rebuilds the pre-ordained line of argument in order to prove its opposite. However, this upturning of the argument exposes its own construct-character and relativizes itself: At the end Socrates excuses himself to Eros for his overly poetic use of language which had been in order "to please Phaedrus" (257a). The subversive play with its own credibility also affects the didactically blended-in myths, for instance of the Cicadas. Socrates reprimands his eager conversation partner for being unfamiliar with this educational heritage: "Surely it is unbecoming in a devotee of the Muses not to have heard of a thing like that!" (259b). The accused, of course, has had no chance to redeem himself as knowledgeable since the story has just been invented. Phaedrus' suspicions are only raised with the telling of the myth of Theut. Socrates then freely admits that he has made up his report and attempts to justify this with the explanation that he lied in the interests of the truth.
Plato's self-relativization does not spare the representation of his own fundamental principles. Thus Socrates' first speech initially reads as a didactic demonstration of the journey through aporia to anamnesis: Not knowing what to say, "under compulsion" (237a), he starts to speak, to suddenly admit that he is "divinely inspired" (238c). The 'Aha-effect' of the learned Socrates transpires however to be a misconstrual. The impressively demonstrated act of remembering was a parodistic simulation.
I see the specific mark of platonic intertextuality to be such distancing from the mediality of representation via this very means of representation. By contrast to the rigour of the 7th Letter, Plato has achieved a form of writing here about things that cannot be written. In this subversion of written language, Plato opens out an alternative to his central theorem that the mneme and therefore the independent movement of the soul necessary for anamnesis can be offset by the use of hypomnemata. These do not of necessity paralyse this movement, but only on condition that their use is absolutised. This comes about when the things that had previously been known to be mnemonic aids are unconsciously transformed into prothetic memory transplants.
The point of the dialogue is that Lysias, the very master of oral speech, demands of his students the slave-like reproduction of his words while the vitality of the socratic counter-speech lies in the fact that it is constructed according to rules of written composition. The crucial difference is not the choice of the particular medium, but the transparency of the intermedial support-functions: Phaedrus carries his written notes of Lysias' speech "under (his) cloak" (228d), while Plato openly shows Socrates' lecture to be a literary product.
Writing, according to Plato, does therefore not necessarily harbour all the danger of forgetting recollection in the act of storing. Properly used, it can bring about remembering by subverting what is stored in it. Schereschewski's futile efforts at forgetting do not refute this theory. The mnemopath's notes are anti-hypomnemata as is the Phaedrus -dialogue. Much to Schereschewski's chagrin, they bring themselves to memory by the energy of their will to bypass the automations of mnemonic aids. Admittedly what Schereschewski produced was not literature. His notes did not lead to a new quality of remembering. But their underlying dynamic is comparable to that of the literary techniques of recollection.
With its non-topographical and non-linear structure therefore, we can say that literature is a critical alternative to any data storage system that stays within the limits of a mnemonic device. Ironically, however, one of the most advanced data storage systems of our days, the hypertext, is based on this very structure. One would be justified, therefore, in asking whether it is not, in fact, superior even to traditional literature by virtue of its ability to stimulate the process of recollection.
Hypertext: The Termina(liza)tion of Intertextuality
Phaedrus is enjoying a huge wave of popularity among the theorists of new textuality. They praise Plato's dialogue-direction as the transcendence of linear textuality, at the same time criticizing the conversational form which they ascribe to this transcendence for being only pseudo-interactive. In this vein we can approach the following remark of David Bolter's, "The form invites the reader to participate in a conversation and then denies him or her full participation."
Like David Bolter, many theorists of the new medium see the hypertext as the solution to this problem in the auspices of intertextuality. The same applies to David Kolb's Story Space with the title Socrates in the Labyrinth, that circles around the notion of "intermediate structure" and interprets this as the heightened continuation of platonic motifs. He constructs a threefold structure in the history of the media, the crowning synthesis of which is almost automatically the hypertext. At the beginning is the oral dialogue in which Socrates confronts his conversational partners with alternatives that they have to decide on on their own. He brings them "into the position of constant responsibility for their own and others' discourse". The written text prevents this form of interaction from taking place, providing a release, however, from the situative context of the conversation, increasing the variability of the perspectives of reception. The hypertext heightens the advantages of both of its forerunner media while rendering their disadvantages inactive. Through the linking up of texts that are scattered into single isolated passages, the multiplicity of perspectives is exponentially enlarged. At the same time the reader can influence the course of the idea interactively as in an oral conversation.
So much for the theory. What does the practice look like, the reading-situation? Inexperienced readers clicking through hypertexts see themselves initially confronted with a freedom they are unable to deal with. They are put in the position of a theatre director who is totally unfamiliar with the drama he is about to direct. This leads to a situation in which the transition from aporia to anamnesis is pre-determined: After a certain amount of clicking around, the reader will have the déja-vu experience of a text-passage previously beheld.
What I had to condense here is the problem with drawing a parallel between hypertextual interactivity and literary intertextuality. The omnipotence of the author which did not have to wait for modern and postmodern texts to be challenged is not further discredited by the hypertext, but restored. This omnipotence has merely become hard to identify as it has externalized itself in the algorithms of the textual mesh. Insodoing it has mercilessly reduced the free space for imaginative recollecting processes. Every action of the mouse-clicking reader pins his sense for virtuality down to factual contents. Therefore it is a misleading use of language to talk about the new objective medium in terms of virtual reality. The so-called hyperspace reduces precisely those spaces to pre-ordained text-variations in which intertextual structures can develop their dynamic. Every click contributes to turning a potential abundance of association into a desert of dissociation. Polyperspectivity degenerates into patchwork. Thus Sally Jacksons Patchwork Girl is not by chance one of the most authentic exponents of genuine hypertext-literature. As a "clickable map", however, the deconstruction and decentralising of subjectivity loses its aesthetic function; it is demoted to the level of Channel Switching not quite the most original of experiences.
With regard to the "modular reading" associated with hypertexts we can go on in this context to discuss its consequences for memory: These are similar to the symptoms suffered by Sassezki whose amnesia resulted from a fragmented linguistic understanding, i.e. from the loss of the ability to understand "complex links and relationships". He was as well able to identify single features as he had been prior to the injury to his brain, but could not any more "put these together to full pictures". He had difficulties, for example, understanding phrases like "my father's brother". He jumped from one word to the next although both words appeared perfectly clearly to him. He couldn't grasp that the genitive case created a new meaning. Hypertext documents have exactly the same effect. Generally they are constructed in such a way that single expressions are annotated. This becomes a problem when it prevents new meanings from emerging from the combination of annotated expressions within the mind of the reader. Remaining with our present example, while the reader is caused to focus onto the father or the brother as clickable objects, his or her imagination is cut off from the creation of the uncle.
Inversions are as difficult to grasp for the electrified reader as they are for Sassezki. A sentence like "I had breakfast after I had read the newspaper" requires that the first part is still present while the second part is being read. Sassezki always had the feeling that he had walked into a trap the same trap of linearity with which hyperspace awaits us behind the non-linear disguise of the screen it greets us from. The literature of the Gutenberg Galaxy owes a considerable portion of its artistic quality to the turning of succession into simultaneity. The opposite is the case in the hypertext: Here the simultaneous structure is pre-ordained what remains for the reader is its dissolution into monotonous chains of succession.
While normal linguistic understanding is made possible by the ability to "direct the synthesis and organisation of complex associations in one mesh of relations", readers of the hypertext and people with brain-damage in the manner of Sassezki's have lost this ability as a result of war technologies. The result, as in the case of Lurija's tragic client, is mental aphasia.
So there would have been a solution for Schereschewski's problem after all. The hypertext is an ideal instrument of forgetting a forgetting in the sense of Eco's Ars Oblivionalis which functions due to the latent evacuation of achieved recollection that slips away beyond the subject's control.
Conclusion: Towards a New Poetics of Hypermedia
Firstly by way of a brief summary: Techniques of storage cause recollection to degenerate due to the externalization of the act of remembering. Writing is also a storage technique. It becomes a medium of recollection by becoming literary, i.e. by the opening of intertextual spaces that can be filled-in according to the reader's imagination. Hypertexts are structured in such a way that they upturn this literary function: The empty spaces of the intertexts are filled in advance by the positivity of a new textual building-block. With each step through the "web of trails", the hypertext reader flattens out the virtually inspiring dynamics that arise from the experience of the contradiction between a given static texture and its gaps.
If, however, as we can observe in the phenomenon of literary intertextuality, the text is in a position to overcome the inclusions of storage techniques by transcending itself, we may ask whether the hypertext is capable of doing the same. I am convinced that a technique of recollection can be developed in the hyperspace, too. Its precondition is a new theory of intermediality. This should not, however, be confused with multimediality. Intermediality is the deconstruction of multimediality just as intertextuality is the deconstruction of textuality. The hypertext in its current form reverses this deconstruction by technically positivising its virtuality by making it a real object. A new form of using it in the sense of a poetics of remembering has as yet not been attempted. The task is again to transform virtual reality into real virtuality.
 F.W. Bernstein.
 Of course, there has been scepticism as to whether this attribution is really valid. That "a poet should have a tendency towards such matters" is something that Friedrich Georg Jünger, for example, considers "improbable. He is called a poet because he does not follow the learnable techniques of mnemonics that retrieve past thoughts. He follows the Mnemosyne herself, the goddess of recollection and, as such, the mother of the muses" [Jünger, Friedrich Georg: Gedächtnis und Erinnerung; Frankfurt am Main 1987, p.8]. And Stefan Goldmann has in fact been able to show that the story of Simonides as the inventor of the art of narrative, as this is told by the author of the Corpus Ad Herennium, by Cicero and by Quintilian, is an addition made at a later stage. It is a legend invented to legitimate rhetoric [Statt Totenklage Gedächtnis. Zur Erfindung der Mnemotechnik durch Simonides von Keos. In: Poetica 21 (1989), p.43-66]. Even as a legend, however, the story does contain an element of truth. What Mnemosyne and the Muses give to the poet is stored and written down. The life of poetry after this cannot be separated from its medial comportment.
 Lachmann, Renate, Gedächtnis und Literatur; Frankfurt am Main 1990, p.35.
 See Steiner, Uwe C., Die Zeit der Schrift. Die Krise der Schrift und die Vergänglichkeit der Gleichnisse bei Hofmansthal und Rilke, Munich 1996, p.87ff.
 See Matussek, Peter, Goethes Lebens-Erinnerungen, in: Ingensiep, Hans Werner/Hoppe-Sailer, Richard (ed.), Natur Stücke. Zur Kulturgeschichte der Natur; Ostfildern 1996, p. 135-167.
 See Piekara, Frank H. / Ciesinger, Kurt-Georg / Müthig, Klaus Peter, Notizenanfertigen und Behalten, in: Zeitschrift für pädagogische Psychologie 1 (1987), H.4, p.267-280.
 Lurija, Alexander R.: Der Mann, dessen Welt in Scherben ging. Zwei neurologische Geschichten; Reinbek bei Hamburg 1992.
 Lurija, Alexander R.: The Mind of a Mnemonist; Cambridge, London 1987, 69.
 See Assmann, Aleida / Assmann, Jan: Schrift, in: Ritter, Joachim / Gründer, Karlfried (ed.), Historisches Wörterbuch der Philosophie, Vol. 8 R-Sc; Basel Stuttgart 1992, Sp. 1417-1431, here 1424f. Also Assmann, Aleida, Schrift und Gedächtnis - Rivalität oder Allianz? In, Faßler, Manfred / Halbach, Wulf (ed.), Inszenierungen von Information. Motive elektronischer Ordnung, Gießen 1992, p.93-103, here p.95f.
 The slave finally despairs and cries out, "But indeed, Socrates, I don't know." "Do you see", Socrates replied to Meno, "what point this boy has now reached on the path of recollection? ... So in causing him to be at a loss and to be numb, as the electric ray does, ... we have done something useful, as it seems, for his finding out how the matter is; for now he might actually enquire into it gladly, whereas then he thought that he could easily speak well..." (84a-b)
 Lachmann, ibid., p.38 ff.
 Ernst Heitsch's Commentary (Werke Vol III/4, Göttingen 1993, p. 248-252) provides a list of the various authorities referred to by Socrates in addition to approximately 50 quotations which he includes in his speech. Heitsch proves the implicit presence of Isocrates throughout Phaedrus to be the hitherto anonymous addressee of Plato's argumentation (ibid., p. 257-262). This is an early example of a "latent" intertextuality that according to Renate Lachmann's definition "does not disturb the surface of the intratext but nevertheless determines the constitution of meaning" (Lachmann, ibid., p. 57).
 Heitsch, ibid., p. 57 and 77.
 This includes the innumerable word plays that suffuse the principle of literalness for example by interpreting names not etymologically, but by association (See Heitsch, ibid., 242).
 Compare Michel Narcy's hypopthesis "que Platon y donne d'abord un échantillon de l'art de Lysias, qui consiste à écrire comme on parle, puis fait parler Socrate comme un livre. Donner la parole de Socrate comme le modèle de la vrai rhétorique, c'est dire adieu à l'oralité". Narcy, Michel: Platon, l'écriture et les transformations de la rhétorique. In: Rossetti, Livio (ed.): Understanding the Phaedrus: Proceedings of the II. Symposium Platonicum; Sankt Augustin 1992, S. 275279, here S. 279.
 The reference to Plato as the alleged forerunner of hypertextual processes seems to me to be more appropriate than the far more frequent references to Mallarmé, Queneau, Joyce, Borges or Arno Schmidt. The main characteristic that distinguishes the hypertext from traditional texts is its interactivity which can indeed be projected back onto Plato's dialogues. In the following I wish to show what is erased in this process.
 Bolter, J. David, Writing Space, Hillsdale (NJ) 1991, p.111.
 Landow for example writes: "Electronic linking, which provides one of the defining features of hypertext, also embodies Julia Kristeva's notions of intertextuality" [Landow, George P.: What's a Critic to Do? Critical Theory in the Age of Hypertext. In: Landow, George P. (ed.): Hyper/Text/Theory, Baltimore, London 1994, S. 151, hier S. 1]. The glossary in Klepper, Martin / Mayer, Ruth / Schneck, Ernst-Peter (ed.): Hyperkultur. Zur Fiktion des Computerzeitalters; Berlin, New York 1996 explains "Intertextuality" als "Cross-reference structures among texts. Hypertexts possess a sort of direct, open intertextuality since by definition they are made up of many different texts and fragments of texts beside and above each other, all of which can be activated by clicking the mouse" (p. 278).
 Eastgate Systems, Watertown 1995. Since this is a disc there is no need for the customary page references.
 "... there are some generalities that were observed as early as Plato about the effect of writing on communication, memory, and thought. Writing is the first step in thought's both losing control and gaining control. Losing control because, as Plato complains, the thinking process is no longer fully present as it happens, but also because the units of exchange get longer. Gaining control because those units can be frozen, inspected, and structured more intricately, just as Plato wrote and rewrote his dialogues. Losing control because even with that rewriting the words and structures take on lives of their own. They did this in oral discussion too, but written words can find their way into more contexts and more changes."
 This guessing-game can be more or less exciting. Usually it is less exciting, since dramaturgy is not a game controlled by the throwing of a dice. It requires the author's compositional arrangements. It therefore speaks for itself that the famous instances of the genre like David Bolter and Ted Nelson owe their popularity to the printed versions of their hypertext treatises. Kolb also published his essay in book-form - just to be on the safe side - in: Landow, ibid., p.323-344.
 Admittedly this is not the case in collaborative hypertext projects that are written further by the reader's insertions. I am convinced that this involves even more of an enslavement principle which I cannot elucidate any further at this point see my essay: Durch die Maschen. Die Vernetzung des kulturellen Gedächtnisses und ihre Erinnerungslücken. In: Dencker, Klaus Peter (ed.): Labile Ordnungen. Dokumentation des Symposiums INTERFACE 3; Hamburg 1996 (in print).
 Watertown (MA) 1995.
 Ernst, Wolfgang, Geschichte im Rhythmus digitaler Medien; Ts. Cologne 1996, p.2.
 Lurija (1992), ibid., p. 48. If Sassezki was shown a pair of spectacles what he saw was a circle and another circle then a stick, and another stick" (p. 47).
 Lurija, ibid., p.121.
 Ibid., p.113.
 See Eco, Umberto: An Ars Oblivionalis? Forget it! In: PMLA 103 (1988), S. 25461.
 Because the subject thinks it has control over the choice of alternative contexts, while in reality this control remains with the hypertextual algorythm, its ability to stand up to auktorial pressure is reduced. The more alternatives it is confronted with, the more its apparent freedom turns into a prison. These alternatives transform the compositional element of inner ambivalence and multi-tiering into a substantiated objectification the tensions are therefore concretised into a "static" juxtaposition. To this extent the hypertext is the exact opposite of the 'Espace Proustien' (G. Poulet) which is based on the principle of recollected images that arrange themselves in tiers (see Warning, Rainer, Claude Simons Gedächtnisräume: La Route des Flandres. In: Haverkamp, Anselm / Lachman, Renate (ed.): Gedächtniskunst. Raum - Bild - Schrift. Studien zur Mnemotechnik; Frankfurt am Main 1991, p. 356-387, here p.363). Furthermore, in order to offer a large variety of alternatives for interaction, the hypertext has to formulate unspecific transitions that should be as independent as possible from any context so that they are open to being linked in many directions. The criteria for relevance which is responsible for the intensity of recollection therefore have to be made uniform.