*This text represents a paper that was presented at the Prague Joyce Colloquium, "Joycean Genetics & Hypertext," in September, 2003.

What I say for the first time, as if as testimony, is already a repetition, at least a repeatability; it is already an iterability, more than once at once, more than an instant in one instant, at the same time; and that being the case, the instant is always divided at its very point, at the point of its writing. It is always on the verge [en instance] of becoming divided, whence the problem of idealisation. To the extent that it is repeatable, the singular instant becomes an ideal instant. The root of the testimonial problem of techne is to be found here. The technical reproducability is excluded from testimony, which always calls for the presence of the live voice in the first person. But from the moment that a testimony must be repeated, techne is admitted; it is introduced where it excluded. For this one need not wait for cameras, videos, typewriters, and computers. As soon as the sentence is repeatable, that is, from its origin, the instant it is pronounced and becomes intelligible, thus idealisable, it is already instrumentalisable and affected by technology. And virtuality. It is thus the very instant of the instant that seems to be exemplary: exemplary in the very place where it seems unique and irreplaceable, under the seal of unicity. And it is perhaps here, with the technological both as ideality and prosthetic iterability, that the possibility of fiction and lie, simulacrum and literature, that of the right to literature insinuates itself, at the very origin of truthful testimony, autobiography in good faith, sincere confession, as their essential composability.
[Jacques Derrida, Demeure]1

The question of materiality remains as pressing as ever in current discussions of textual genetics and hypertext, and it is the objective of this paper--although highly speculative in places--to provide something like a notational framework within which we might effectively engage with this question without descending either into classical hermeneutics, epistemology or empiricism, while at the same time enlisting certain aspects of their conceptual infrastructures to the work of textual theory-above all to a theory of hypertextuality.

The significance of James Joyce's work for discussions of textual genetics and hypertext has not always been as straightforward as it has been made to seem--even in the specific context of "Joycean genetics" and "Joycean hypertext." This has to do, to a certain degree, with conflicting theoretical orientations rather than, strictly speaking, with methodologies--at least with working methodologies. In 1995 Geert Lernout mapped out a number of these differences within the field of Joycean genetics,
2 in which praxis is rather foregrounded, delineating between the work of Daniel Ferrer, David Hayman and himself--or between a poststructuralist tendency on the one hand, and a "radical philology" on the other. Without rehearsing the arguments set forth in Lernout's paper, the current discussion may be profitably viewed as a partial response to a number of the points raised in it.


Whatever inheritance has come down to textual genetics and hypertext from the work of the two most prominent of Joyce's "theoretical" exponents, Marshall McLuhan and Jacques Derrida, and however this has been received, the genealogies are far from simple. One aspect, however, may be seen to be more or less consistent, and that is the emphasis upon "technology"; that is, upon the techne of writing. And, in the work of Derrida at least, upon the mechanical, and above all material, basis of techne as such. Elsewhere I have discussed this at length and will not attempt to go back over the same ground here,
3 except to recall to attention certain remarks by Martin Heidegger which, while not directly cited in the writings of either McLuhan or Derrida, provide something like a common motto for their respective projects. In "Die Frage nach der Technik," Heidegger writes: "techne belongs to bringing forth, to poiesis; it is something poetic."4 And later he adds: "not praxis but poiesis may enable us to confront the essential unfolding of technology."5

In short, the question of materiality remains tied to the question of text, to the inherent structurality of language and the
conditions of signification, in advance of any assumption of praxis. For this reason the discussion of hypertext here will focus upon the condition of hypertextuality, rather than upon computer-based electronic writing. It may be instructive to recall that hypertext as it is popularly conceived is fundamentally infrastructural rather than anything like a discrete object or thing. To speak of "a hypertext" in this sense is to engender a fallacy that returns us to such notions as the "literary object." The materiality of hypertext should not be confused with the electro-bibliographical codes that are commonly regarded as standing for it--i.e. the phenomenal aspects of "electronic writing." The materiality of hypertext is rather the condition itself of such phenomena, of writing per se, and thereby underwrites its "technological application."

The implication of this for textual genetics is clear enough, and evidently extends beyond the instrumental or prosthetic function of computerised "hypertext" as a presentational medium for such things as Joyce's notebooks and drafts, or annotated or synoptic editions of his published works. While the efficacy of such "applications" is not in question, the idea that such application
accounts in any way for hypertext, and for the significance of hypertext for textual genetics, most certainly is. As Lernout points out, the "falsifiability" of genetic research lends it to a logistic conception of text: in short, the connection of meaning and verifiability--and this is the question which must firstly be addressed, not, however, as a binary expression of either praxis or poiesis, but rather, in the first instance, of their nexus in a common materiality.

In his discussion of radical philology, Lernout cites Ferrer citing Derrida on the "possibility of disengagement and citational graft which belongs to the structure of every mark, spoken or written, and which constitutes every mark in writing before and outside every horizon of semio-linguistic communication." ('Signature, événement, contexte') Despite the evident irony of this situation, it is worth taking Lernout's objection seriously, that whatever stands "before and outside every horizon of semio-linguistic communication" constitutes--as in Kant's Critique of Practical Reason and Wittgenstein's Tractatus, "that whereof we cannot speak." For Lernout, the anteriority of signification is a matter simply of intuition, and therefore characterises a failure of rigorous methodology.

It is precisely the question of methodology or rather of method, however, which may be regarded as being at stake here. [For Derrida, the anteriority of signification is indicative of the tautological relation of the instantaneousness of the present (posed in the form of the signifier) to techne, and which via the concept of "testimony," devolves in large part upon the impossibility of generalising the instant, while nevertheless confronting the necessarily generalising condition of "iterability" as the structural constraint and pre-condition of its signifier as such (i.e. of "the instant" as "une série de contiguïtés matérielle")].
6 Derrida argues at length in his 1996 collaboration with Bernard Stiegler, Échographies: "que technicité ne soit pas technique, que la pensée de la technique ne soit pas technique, c'est la condition de la pensée."7 And yet in speaking thus we necessarily generalise this concept, as Derrida warns, both as an exemplum and as an ideality (viz. the supplementarity of method). The point for Lernout, however, is that whatever stands as an object of intuitive knowledge is unverifiable; it is not an object of knowledge at all and is therefore irrelevant to the project of textual genetics.

What the project of textual genetics properly is may be debatable, although consensus seems to lean towards a definition based upon "direct treatment of the avant-texte." By "direct treatment" we may assume any number of meanings, from classical hermeneutics or empirical method, to considerations of textual genesis or textual materiality; in whatever context, from historicity to the technics of literary composition, to classical or radical philology, and so forth. In any case the term needs to be qualified, if only for the very practical reason that textual genetics begins with a necessary if apparently contradictory assumption of incompletion, and that at every point it must take this into account, above all in its definition of "verifiability."

Between a conception of semio-linguistic anteriority and of genetic verifiability, there arises the problem of "prediction." If anteriority is purely a matter of intuition, as Lernout argues, then verifiability itself succumbs to the indeterminacy inherent to all forms of predictive modelling. What is significant is not that this indeterminacy arises as a consequence of the "incompletion" of genetics--or from any other limitation of empirical knowledge--but that it is structural and structurally inherent; which is indeed the point of Derrida's statement regarding "possibility" ("the possibility of disengagement and citational graft which belongs to the structure of every mark").

Lernout is obliged to concede that textual genetics, viewed in the context of "radical philology," can never be more than an approximative method or, rather, an approximative system of knowledge, whose tenets must therefore at some point violate the principle of verifiability. Approximation is in this sense not merely a practical necessity with regards to a certain limit implicit to the techne of knowledge, but as a condition bound up with the materiality of "knowledge"--that is, semio-linguistic or signifying materiality. The logical consequences of viewing "knowledge" as an approximative system which will never be varifiable are thus crucial to an understanding of why the argument about the intuitive character of semio-linguistic anteriority does not hold.

With "radical philology," a fictive definitive system of knowledge is established as the basis of epistemological enquiry, with the result that the schematised character of this basis is soon forgotten, and the fictive construction is identified with the actual system. It is with regard to the limits of this construct that semio-linguistic anteriority assumes its "intuitive" character, for Lernout, as that which exceeds "verifiability." The relativism of this system not only contradicts its basic premise of generalisability (something must be generally verifiable, not merely a special instance of verifiability), but it also exposes the system to further logical violations with regard to what we might call "locality" (vis-a-vis Derrida's "disengagement and citational graft") and the systems over-dependence upon context. In short, the predictive limits of textual genetics require that all recourse to context be provisional, and at the same time that the probabilistic feature of this "recourse" NOT be regarded as provisional. Indeed, probability invests the genetic project at every level, consequent upon precisely the "possibility of disengagement and citational graft," as Derrida says, "which belongs to the structure of every mark, spoken or written, and which constitutes every mark in writing before and outside every horizon of semio-linguistic communication."

As Hans Reichenbach has pointed out, "It is one of the elementary laws of approximative procedure that the consequences drawn from a schematised conception do not hold outside the limits of approximation; that in particular no consequences may be drawn from features belonging to the nature of the schematisation only and not to the co-ordinate object." (Experience and Prediction)
8 The question that obtains here is how approximation avails itself in any way of a consequent realisation of its "co-ordinate object," as Reichenbach says.

Before proceeding, it may well be worth going back over one of Reichenbach's assumptions about language and signification in general, before proposing anything like a response to the above question. According to the tenets of logical empiricism, "symbols" are physical bodies or processes like any other, irrespective of their "function." It makes no difference if we consider a symbol to obtain meaning through its correspondence to "facts" or to other "symbols." A symbol itself is a fact. In structural terms it is irrelevant what "class" of fact a symbol "corresponds" to, or why it is taken as corresponding to it. It's significance, and that of language as a whole, resides in the possibility of treating a physical body as a symbol; and symbolization as a function of a possible (meaningful) correspondence between facts.

By treating symbols as facts in this manner contradicts, on the level of semio-linguistic materiality at least, the principle of verifiability. That is, the principle of "truth value," which, as Reichenbach demonstrates, is consequent upon a schematised conception. Moreover, the principle of verifiability is required, in the first instance, to account for the possibility of "correspondence," and subsequently to account for the ultimately approximative nature of correspondence as such. In this way verifiability cedes to probability and is consequent therefore upon prediction rather than upon a determinate "state of affairs."

During a conversation at the 2002 Trieste Joyce Symposium, Fritz Senn made a comment which I intially mis-heard as being: "There is no language in Finnegans Wake"--taking this as one of Senn's typically elliptical and provocative statements. Nevertheless I found myself in qualified agreement, having myself made the claim at the Rome Symposium that in order to begin "reading" Finnegans Wake, we are first of all obliged to make certain assumptions of translation--that we must in fact "translate" the apparent marks on the page into some determinate semio-linguistic system or language-that we must literally read in or into a lapsus of language--or "no-language" (or as Joyce says in Finnegans Wake, "nat language in any sinse of the word"). Hence, we may say that there is "no-language" in Finnegans Wake.

What is at once most striking about Joyce's text, and at the same time most overlooked, is the power of its resemblant quality: the marks on the page resemble (in most instances) the conventional signifiers of the latin alphabet, and appear in part to correspond to particles of any number of languages, both current and extinct. The problem is that this correspondence cannot be generalised, or indeed schematised according to any conventional notion of language, and that it arises in the "first place" out of the fact of this no-language.

The question--and it is a very interesting one--is how, then, we can assume to "read" Finnegans Wake. What makes such a "reading" possible? What, to complicate things, is its "co-ordinate object," as Reichenbach says?

Such questions are evidently not idle, as a vast amount of philological activity has indeed been devoted to enumerating sets of "facts" that correspond, in some way, with Finnegans Wake--whether on a micro- or macro-scale; whether "foreign word lists," place maps, historical narratives, source texts, or whatever. Each of these assume a decipherment of the text; that Finnegans Wake is in fact a type of textual object to be deciphered, dissected, anatomised, classified, and so on. And indeed this too is a fact; is a kind of fact, one among others, that corresponds in some way to an idea of Finnegans Wake. And in and of themselves, each of these facts is "verifiable," to a certain degree, and yet no idea of Finnegans Wake is verifiable. It is because the idea is already a schematisation--the outcome of a set of predictions centred upon a causal arrangement of symbolic "correspondences"--whose actual "co-ordinate object" has effectively been suppressed because (we may assume) it renders the very notion of verifiability nonsensical.

How does it do this? We might say it does this by exposing all presuppositions about language to the broadest implications of semio-linguistic materiality and to the radically probabilistic organisation of language as a whole.

What would it mean to verify the materiality of a "symbol"? One problem is that to verify already entails symbolisation-here, with regard to a measure of "truth value." Another problem is that to assign "truth value" to materiality is tautological. In philosophy the formula S=P provides a simple expression of this effect of semiological "complementarity." It is evident enough that S is not P, and yet the structure of equivalence or correspondence described here is the one which underwrites the entirety of signification: whether it is in the conventional model of the sign (signifier-signified); in the organisation of rhetorical tropes or figures (metaphor, metonymy, allegory, analogy, parataxis, and so on); or in the overarching notion of narrative and schema. We might say, therefore, that language proceeds on the basis of what we might call an "inequality" theorem, and that inequality itself provides the measure of verifiability. It may be that language occurs as such in the "suspension" of verifiability. Or, we might equally characterise language as proceeding from a structural dependence upon a principle of the "arbitrary" (S equals P, where S and P can be any terms whatsoever) which is nevertheless tied to "correspondence" (S does not equal P, where S and P are nevertheless mutually determined and interdependent)--a form of complementarity from which symbolisation emerges as an effect of what I will call "entanglement."


In the section of Gulliver's Travels devoted to the 'Academy of Lagado,' Jonathan Swift describes a form of parodic random text generator "for improving speculative knowledge by practical and mechanical operations." Swift's machine operates on the basis of lexical (or sublexical) combination and recombination, producing random "propositions" which are subsequently analysed for their philosophical content. The analysts, nominally philosophers, are reduced to the role Alan Turing envisaged for computer technicians: that is, as mere attendants. Swift's prototypical computer operates on a basis of non-predictability as the principle, in fact, of what remains a purposive form of textual production. However, the mechanism of recombination or "material variability" is not only arbitrary, it is indifferent to outcomes, no matter how purposive these may appear as "objectives," implying (in more than a satirical manner) that semio-linguistic correspondence is an entirely accidental matter--but more, that it describes something like a statistical outcome; the work of virtually infinite possibility within the (normative) constraints of finite probability.

The history of writing and of forms of encoding or encryption returns constantly to the fact of symbolic substitution, abbreviation, displacement, and so on. Anagrams, acrostics, acronyms, algebraic summarisation, etc, are all commonplace demonstrations of "meaningful" symbolic variability. This is precisely the phenomenon that attracted Ferdinand de Saussure in elaborating his differential approach to general linguistics based upon phonemic and lexical variability. For Saussure, variability underwrites both the semantic identity of a particular symbol in relation to other symbols, at the same time as it opens up the possibility of ambiguity and ultimately of a "limit of meaning."

Like genetics, morphological process can be modified and extended beyond what Lernout terms "a saturable and constraining context." The material indifference of semio-linguistic fragments to the constraints of context reminds us that in fact there is nothing purposeful about language-that language harbours no secret design or intent; that it is not a subject replete with its own psychology or psychological agency. At the same time, to speak of language as such, as of an entity, is to distort the fact of what we might call its "incompleteness." We might say, indeed, that language can only be defined against completion, if not against its possibility, and that it is this incompletion, or condition of possibility, which lends a semantic complexion to these "material fragments." Without such a possibility as this, it would be impossible to claim to read Finnegans Wake, or indeed to read at all.

If we think of the possible combinatory outcomes of a textual machine like Swift's speculative apparatus, we can see that, structurally at least, reading does not begin with "a saturable and constraining context," but rather with the possibility of bringing otherwise disparate, non-contiguous elements into "correspondence." As with the formula S=P, signifiance operates in the possibility of both affecting and arbitrating a "semantic gap"--a disposition with regards to symbolisation that affects itself in a way that may be described as "equally structured." That is, the function of (in)equivalence is mutually affective: there is no purposive or relativistic attribute of tenor or vehicle, for example, but rather a network or simultaneous influence that "determines" the virtual character of a text at any point. And it is here that we may profitably begin to look again at what has come to be called "Joycean hypertext."


Hypertext was defined by Theodor Nelson, who coined the term in the 1960s, as "non-sequential writing"--which may be understood in the post-Mallarméan sense of non-sequentiality, or "simultaneity" as later defined by Apollinaire and Blaise Cendrars. The "hyper" prefix is taken here to refer, not to a meta level of textuality (a form of semantic epiphenomenon), but to an "internal" mechanism of semantic entanglement or linkage, by which any textual element could be brought into "communication" with any other.

Nelson's conception of hypertext was by and large a response to a paper published by Vannevar Bush in 1945, in the Atlantic Monthly, entitled 'As We May Think'--which proposed a primitive form of cybernetic apparatus, in effect a non-sequential information retrieval system. The major feature of Bush's proposal was that it sought to overcome the limitations of existing indexical structures, which required a search thread to be directed along a branching structure whose organisation was rigidly linear. This "tree-diagram" model was to be replaced by what has since come to be called a rhizomatic model, and subsequently-in the computing application of hypertext-any particle of information could conceivably be linked to any other, according to the minimal existing unit of the pixel. In certain respects, Nelson, who was also involved with the development of artificial intelligence, regarded hypertext as at least analogous to a form of mechanical cognition-extending Alan Turing's conception of the "Automatic Computing Engine" as an electronic "brain." In many respects, hypertext and artificial intelligence are seen as analogous, largely upon the basis of symbolic connectivity and the way in which "thought" seems to arise on the basis of what we might call non-local events brought into communication.

An important characteristic of hypertext has since come to be its resemblance to physical systems that possess global properties which continue to evolve globally even as the system becomes spatially (or temporally) separated. In this sense, non-locality and entanglement may be useful terms for understanding the "non-sequential" aspects of signification which constitute hypertext, and the consequent effect of "modification" which attends upon interpretive events. Psycho-linguistics has described the phenomenon of "encincturing," whereby reading is seen as contingent not upon decidability but upon a synthetic and arbitrary mechanism of closure (or semantic "completion"). Significatory possibility is "closed off" for each lexical term within a sequence in order for a propositional structure to obtain; just as the possibilities of sub-lexical recombination must be "closed off" in order for lexical structure to obtain; and so on. At the same time, each event of encincturing brings about a simultaneous modification of the existing signifying structure within which the encinctured term is embedded, as well as modifying the predictive context for those terms which then "follow." *(Derrida, referring to the "hypermnesiac machine" of Finnegan Wake, speaks of "quasi-infinite speed" while Joyce himself talks of "infinite probability" [VI.B.14].)

This process of encincturing has become a commonplace in descriptions of how Joyce's portmanteau words are taken to operate in Finnegans Wake within a more or less conventional linguistic structure. In doing so, it allows for simultaneously divergent possibilities encoded in the portmanteaux, as in the paranomasian character of Joyce's poetics in Finnegans Wake generally. In effect, Joyce's writing exacerbates the question of materiality by reaching constantly beyond the assumes limits of signifying probability, and by situating the "nexus" of semio-linguistic structure in its materiality and not in some form of anima or agency "deposited in it"-as Frederic Jameson, among others, suggests when he states that the "pure temporal movement of signification itself, as it deposits itself in object or letter, is retained, without any ultimate sense of the direction or meaning of that movement." (Marxism & Form, 121)

On the other hand, it may be argued that what hypertext implies--extending the implications of Saussure's sub-lexical approach--is that the phenomenon of signification arises on the basis of "fundamental" material particles in a pre-linguistic state. Particles which, in non-local communication with one another, may assume a schematic or macro-state dimension, as in, for example, the "group triads" H.C.E. and A.L.P., or more radically in regard to the Wakean sigla; the "doodles family." As with the Wake's language taken as a whole, the sigla present non-linguistic particles which also function as figures or tropes whose "co-ordinate object" is, in a certain sense, language itself. It is only by such typogenetic means that we can account for the way in which certain sigla (e.g. "(+)" are able to attain allegorical and schematic dimensions in the Wake. Or, on the other hand, how acrostic or acronymic figures like A.L.P. and H.C.E. attain the status of structural paradigms (or what Danis Rose and John O'Hanlon term "contextual invariants" [Understanding Finnegans Wake. NY: Garland, 1982. xiv]). That Joyce himself provided an index to the sigla in a letter to Harriet Shaw Weaver, is of no real consequence here. Most of the sigla do not in fact appear in the published version of Finnegans Wake, nor do H.C.E. and A.L.P. appear beyond a number of brief mentions. What does appear, instead, is the very mechanism of material recombination which these figures may in some ways be taken to exemplify. H.C.E. and A.L.P. represent, in other words, a matrix of associative structures, whose limits of permutation are arbitrarily determined by the interpretive process-a subject Joyce himself discusses at length in the Wake.

What hypertext shows us, however, is that semio-linguistic materiality, taken by itself as an empirical fact, is equally as meaningless as the generalisation of the text from either a phenomenal or ideological point of view. Signification arises alone upon the possibility of a network of transverse relations--non-local communication between material elements or particles which themselves define anterior transverse relations. Each particle is a set of co-ordinates in n dimensions, a variant point of articulation which is at variance with itself while at the same time producing an effect of structural contiguity. Or, to extend the metaphor, we might say that Joyce's "geometry lesson" provides instructions about how to proceed from Euclidian 3-space (point, line, volume, etc.) to a semio-linguistic multi-demensionality: from an incipient diagrammatics of "marks on the page" to a "system" of language--a movement that I have come to term, as an elaboration on hypertextuality,


But what do these almost Joycean terms mean, "hypertext" and "vortext"? To say hypertext--is it to speak in the same manner as Derrida when he discusses the Wake as a hyper-mnesiac machine? How does this hyper-prefix effectively signify? Could we say hypo-text, perhaps, and thereby arrive at a binary understanding of what hyper-text might signify, at least etymologically or in terms of some measure of relativity; hyper or hypo, which presupposes a norm or normative conception of "text." And by such means might we not also find in the term vortext an echo of the Derridean pro-gramme, of a certain fore-text or avant-texte (which is the concern, as one says, of textual genetics)? Laurent Milesi has pointed to the possible meaning of hyper-text as something over, above, or beyond text; something which exceeds or which is in excess of "text"; something "hyper-linguistic." But in saying so, it is important to keep track of what this could mean: a form of text over, above or beyond "text." As one might say, for example, "poetic language"--as opposed, presumably, to "normal" language. And here we may be reminded of Wittgenstein's contention that there is, strictly speaking, no such thing as poetic language; or as Derrida says, there is no outside of (the) text. The underwriting of poiesis is, and must be, a general condition of language "as a whole": just as "hypertextuality" must remain an inherent possibility of text, of all text, of any text, and so on.

One reason for proposing the term "vortext" (and I discuss this usage in far more detail elsewhere), is to get around the highly distracting business of distinguishing "hypertext" from those meta-textual functions or devices to which it has so often been compared (index, citation, cross-reference, and so on). To some degree the term is intended to structure a way of thinking about language that takes a more general account of programmatics per se, and of material "anteriority" above all. Of course, this is not to summon up a notion of antecedence, or of causal sequentiality, between hypertext and "text," etc., but rather a means of addressing that condition of materiality upon which textuality, broadly speaking, devolves.

In this way it may be more advised to speak of hypertextuality, being a condition, rather than hypertext as some sort of denotative term for a distinct class of text or of textual operations. The term "vortext" may consequently also be taken, beyond its Poundian echoes of Vorticist dynamism and flux, as a form of internally constituted "textual genetics." (And I am conscious here of the fact that the term may suggest an extension of the genetic conception of avant-texte into the realm of avant-garde procedure and "ideality"--as though the vortext were rather an object of some compositional intention encoded within the modernist project.)

The issue here, however, is not so much one of terminology or praxis, but of a general poetics of materiality which must not only be able to account for its models--and increasingly hypertext, in its popular "application" (and despite the term's original coinage) has come to be little more than a "model" for certain apparently mysterious aspects of textuality or of cognition-but to account also for the conditions upon which such models are founded, and the dynamics which cause or allow them to operate or in other ways extend or limit their operations. I do not mean here the instrumentality of hypertext, but its conditionality; vortex would thus suggest a type of on-going event of semio-linguistic instigation or solicitation, rather than of a "simple" technological extension, incrementation or inflation of textuality post facto-as though techne were something applied to language.

One of the problems here, as Daniel Ferrer has pointed out, is one very familiar to textual genetics, and is at base a problem, not of acknowledging or describing, but of accounting for and indeed coming to terms with what it means to speak of "non-sequential writing." It may be that the involutions of discussion about hypertext, and Joycean hypertext in particular, are equally linked to the tautological nature of this formula: non-sequential writing--in which case to evoke hypertext at all is already a redundant gesture; a rhetorical exercise in over-categorisation and schematisation. But at the same time a practical view points to the necessity of re-organising "textual theory" according to the ongoing genesis of techne in its various forms, and the need to "come to terms" with what amounts, institutionally, to a continuing disavowal of a techno-poetics, as Donald Theall says, which has underwritten the project of modernity since at least Mallarmé, and which arguably finds its most comprehensive expression in Finnegans Wake.


© Louis Armand, 2004
volume 4, issue 2, 2003-4