This dissertation centers on the assertion that Ezra Pound’s Cantos are a hypertext. It locates itself at the intersection of Pound scholarship, hypertext theory and media studies. My goals then are multiple. I wish to address certain problems in Pound scholarship that have thus far resisted critical understanding. I believe the emerging paradigm of hypertext provides a means of understanding the structure of the Cantos and a means of answering the persistent complaints of the poem’s formlessness. At the same time, by actually applying the concept of hypertext to such a problem, I hope to clarify the idea of hypertextual literacy. From learning to read the Cantos hypertextually, I will be able to generalize the skills required of such reading and their relation to traditional concepts of literacy. Such work is important, for if hypertext is indeed the next wave in communication paradigms, it is imperative that we become aware of its capabilities, its demands and its biases.
This dissertation will consist of three sections. Part one, "What Is Hypertext?," will cover the general history and theory of hypertext and its relation to previous communication paradigms. Part two, "Ezra Pound in the Late Age of Print," will place Pound within the history of communication technology and show his affinity with hypertext theory as an outgrowth of his historical circumstances. Part three, "Surfing the Cantos," will present the practical application of hypertext theory to the problem of reading Pound’s later Cantos. What follows is a synopsis of the issues to be explored in each of those sections.
This most basic question requires a fairly complicated answer, one that will occupy the first section of my dissertation. In the course of answering it, I will review not only hypertext theory as it has developed, but media theory in general. I differ from some theorists in that I assert that hypertext has been an available textual paradigm for centuries, but that it has been neglected because the technologies of writing have obscured it. It is into such an extended hypertext tradition that I wish to place Pound and the Cantos.
The term "hypertext" was coined in the 1960s by computer scientist Ted Nelson to describe an alternative form of textuality enabled by the rise of computers. Nelson offers the following definition:
Well, by "hypertext" I mean nonsequential writing -- text that branches and allows choices to the reader, best read at an interactive screen.
As popularly conceived, this is a series of text chunks connected by links which offer the reader different pathways.
Subsequent discussions of hypertext have offered various refinements and elaborations of this explanation, but it remains a useful starting point. Nelson establishes the two basic concepts of hypertext: nodes (his "text chunks"), which constitute the content, and links, which connect the nodes so as to allow multiple reading paths. It is easy to see how computers, with their random-access digital storage and almost instantaneous retrieval would be well suited to presenting text in such a manner. The current World Wide Web is a hypertext in this sense, with each "page" or file being a node in which links may be encoded to transfer to other pages.
Yet hypertext is more than simply a new way to present texts using computers. Media theorists since Marshall McLuhan remind us that communication technologies are not neutral containers of information. Different technologies carry with them different biases and establish their own relationships between the reader, author, and text. When the technology undergoes a major change, those inherent biases are transformed as well.
The technology of print, for example, suggests a certain textual paradigm through its physical construction. A fixed sequence of words bound into a fixed sequence of pages emphasizes linear, sequential reading. The job of the reader is to follow a single path that has been prescribed by the author. The text is stable (it will appear much the same every time the reader returns to it) and clearly bounded (it consists of the words on the pages and nothing else). Electronic media change the physical status of the text and, thus, change the textual paradigm. Because the text is no longer embodied by physical marks it can be both fluid and open-ended. By breaking down the text into smaller units (nodes), which are then joined not by consecutive pagination but by links, the text changes. The text no longer resembles beads on a string but a network of points in space.
Hypertext theorists like George Landow and J. David Bolter assert that hypertext represents a "paradigm shift" in the history of communication, along the lines of the introduction of writing or the invention of print. The linking that is the essential feature of hypertext disrupts the paradigm of print. Links offer multiple paths through a text, which contraverts the linear sequentiality of print. Multiple pathways require choice on the part of the reader, giving him or her a creative role in the construction of the text, a role traditionally assigned only to authors of print texts. The text itself becomes less stable, since the reading experience will differ from reader to reader or from reading to reading, depending on the choices made. Even the notion of completion in such a text loses meaning, since there need be no definitive endpoint to such reading other than the reader’s decision to stop. Finally, linking erases the traditional boundaries between texts. What we might think of as a "single" text could be only a node or set of nodes linked together in a larger hypertextual network. Landow explains: "Electronic linking, which emphasizes making connections, immediately expands a text by providing large numbers of points to which other texts can attach themselves." Nelson speaks of a "docuverse" in which all texts and their links "unite into what is essentially a swirling complex of equi-accessible unity, a single great universal text and data grid." In short, the most basic terminology of the print paradigm ceases to be relevant in an electronic environment.
I agree that the significance of this development is staggering, but I wish to complicate the issue by suggesting that the hypertext paradigm is considerably older than has usually been assumed, indeed that it has existed almost as long as writing has but has been overshadowed by the linear paradigms most natural to the technologies that have thus far been available.
While it is true that electronic media, and computers in particular, are particularly well suited to presenting the hypertext paradigm, that paradigm in itself is not bound to any particular technology. One of Nelson’s sources for the hypertext concept was Vannevar Bush, whose 1945 article, "As We May Think," proposed the "memex" system as a means of dealing with the ever-expanding body of scientific knowledge. The memex would be "a sort of mechanized private file and library," which would be capable of most of the features we associate with electronic hypertext using only technology no more advanced than microfilm. More generally, hypertext is really a structural concept, the text as a network of associated nodes, which can be applied to texts of all sorts. Indeed, as Landow discusses, this network model has already been applied in various forms to the central texts of the print tradition by critical theorists such as Barthes and Derrida.
Nelson goes so far as to assert that "hypertext is fundamentally traditional and in the mainstream of literature." Nelson’s point is that while print encourages, almost demands, linearity, there have always been other options and impulses. It is only because "nothing else has been practical" that sequential textuality has persisted as long as it has. Within the tradition of print, numerous mechanisms, some quite conventional and established, have developed to facilitate nonsequential approaches to that sequential medium. The table of contents or index of a printed book enables, albeit clumsily, nonsequential access to the work and can therefore be thought of as a hypertextual device. Similarly, cross-references, bibliographies, sidebars and other common features of printed texts are all more or less hypertextual.
If one looks closer at the print tradition, one can find whole genres of texts that are ill-suited to sequential, printed presentation. Dictionaries, encyclopedias and reference manuals are all collections of related nodes of information for which sequential reading makes little sense. Not surprisingly, these have been among the first texts to be translated to electronic, hypertextual format.
Even those texts most associated with the technology of print -- literary and expository writing -- can be conceived of hypertextually. Nelson states: "Literature is an ongoing series of interconnecting documents. In any ongoing literature, there is perpetual interpretation and reinterpretation, and links between documents help us follow these connections." Literary allusions, quotations, borrowings and imitations can all be seen as implicit "links" to other documents in the larger body of literature. In expository prose, links can be even more explicit. Landow reminds us that "[t]he standard scholarly article in the humanities or physical sciences perfectly embodies the underlying notions of hypertexts multisequentially read text." Footnotes, he argues, constitute side-paths linked from the "main" text, and may in turn point to secondary material quite outside the article itself.
Thus, it is possible to conceive of a hypertextual tradition that has existed side-by-side with the dominant sequential paradigm throughout the history of print (and perhaps earlier). This tradition has been difficult to identify because of the power of the print paradigm. The physical distinctness of print eclipses any other conception of the text. Nelson explains:
Within bodies of writing, everywhere, there are linkages we tend not to see. The individual document, at hand, is what we deal with; we do not see the total linked collection of them all at once. But they are there, the documents not present as well as those that are, and the grand cat’s-cradle among them all.
My hypertextual approach seeks to render visible these unseen links, not only the individual ones, but also the "grand cat’s cradle among them all," to reconsider literary productions as complex networks situated within even larger webs of association.
In the second part of the dissertation I will show Pound’s connection to the issues raised in the first. This connection is both historical and theoretical. My main focus here will be Pound’s critical prose, in which he develops the poetic theories that underlie the Cantos. Though varied and sometimes inconsistent, Pound’s criticism shows him searching for a new vocabulary to describe a textuality that did not fit the paradigms of his time. To an extent virtually unmatched by Pound’s predecessors or contemporaries, this search is conscious and fundamental to his poetic project. His poetics eventually comes to closely resemble hypertext theory in all but name. Indeed, it might be argued that hypertext theory owes its rapid ascendance in part to the questioning of the dominant print paradigm inherent in Pound’s poetry and criticism.
Pound’s interest in such issues should not be considered anachronistic or surprising. He was remarkably engaged with the publishing institutions of his day. His activities as an editor for various "little magazines," his participation in the two numbers of Blast, which proclaimed its violently avant-garde agenda down to the level of typography, his publication of deluxe semi-illuminated editions of some of his poems, and his proposal of the "square $ series," to name a few of his activities, gave him a particular awareness of the social and technological forces mediating literary production. As Jerome McGann puts it, "the history of modernist writing could be written as a history of the modernist book. Were one to write that history, Ezra Pound would appear once again the crucial point of departure." Having no obvious alternative, Pound could not reject print completely, but he seems to possess the sensibility of what Bolter calls "the late age of print": an increasing awareness of the limitations of print and a groping forward for new possibilities for expression in an obsolescing technology.
It is in such an environment that Pound develops his poetics. Through his many phases, one can find a continuing affinity with the basic theories of hypertext. In the early essay series, "I Gather the Limbs of Osiris" (1911), he proclaims "the method of the Luminous Detail": "Any fact is, in a sense, ‘significant’. Any fact may be ‘symptomatic’. But certain facts give one a sudden insight into circumadjacent conditions, into their causes, their effects, into sequence, and law" (SP 22). In his Imagist period he offers the definition of the image as "a radiant node or cluster; it is what I can, and must perforce, call a VORTEX, from which, and through which, and into which, ideas are constantly rushing" (GB 92). Both statements already show an interest in the evocative power of small details in relation, that is, of nodes and their links. Pound’s definition of logopoeia in "How to Read" (1929) goes even further in conceiving of poetry as a network of associations:
Logopoeia, ‘the dance of the intellect among words,’ that is to say, it employs words not only for their direct meaning, but it takes count in a special way of habits of usage, of the context we expect to find with the word, its usual concomitants, of its known acceptances, and of ironical play. (LE 25)
From these statements, which establish crucial concepts for his later poetics, to Nelson’s hypertextual definition of literature as "an ongoing series of interconnecting documents" it is but a short jump.
Pound’s critical and poetic theories eventually culminate in the "ideogrammic method," which is most fully expounded in the "textbooks" ABC of Reading (1934) and Guide to Kulchur (1938). Pound makes the declaration that the "proper METHOD for studying poetry and good letters is the method of contemporary biologists, that is careful first-hand examination of the matter and continual COMPARISON of one ‘slide’ or specimen with another" (ABCR 17). Pound took this statement to heart; much of his later criticism, like his later poetry, consists almost entirely of presenting "slides" in the form of literary and historical exempla for comparison.
Leaving aside the question of how much the ideogrammic method resembles either the practice of the biologist or of Chinese ideogram formation, the importance of this concept for Pound is its faith in the power of associative linking, which is likewise the underlying premise of hypertext theory. The early theorists of hypertext took their inspiration from what they perceived as the patterns of thought. Bush’s seminal article asserts: "The human mind. . . . operates by association. With one item in its grasp, it snaps instantly to the next that is suggested by the association of thoughts. . . ." His memex was meant to reproduce this process of instantaneous associative linking. Nelson also offers hypertext as a superior alternative to print because of its closer resemblance to thought:
The structure of ideas is never sequential; indeed, our thought processes are not very sequential either. . . . as you consider a thing, your thoughts crisscross it constantly, reviewing first one connection, then another. Each new idea is compared with many parts of the whole picture, or with some mental visualization of the whole picture itself.
Hypertextual reading, as theorized, would be almost identical to thinking; in following links, the hypertext reader forges mental associations between the linked nodes. Such reading is more strenuous that the usual conception, for the reader assumes some of the responsibility for the cohesion of the text.
Pound makes similar demands on his readers. His ideogrammic method often requires the reader to figure out for him- or herself what the connection between two "slides" might be. The "Great Bass: Part One" chapter of Guide to Kulchur, for example, consists of four short passages and an explanation: "These disjunct paragraphs belong together, Gaudier, Great Bass, Leibniz, Erigena, are parts of one ideogram, they are not merely separate subjects" (75). Pound has provided the material and pointed out the existence of links, but the formation of the ideogram is left to the reader, who must forge the links between the "disjunct paragraphs."
This insistence on reader engagement carries over to the Cantos. Similarly composed by linking together nodes drawn from disparate sources, the Cantos demand that the reader make those links meaningful. The hypertextual character of this work can perhaps be seen most clearly in the ways it has baffled and frustrated readers who have approached it with the typical expectations of print.
The problem is illustrated by the many hostile responses to the Cantos by contemporary critics. For example, in 1956, Yvor Winters offered a particularly sharp attack on the Cantos:
The work has no narrative structure, such as that of The Iliad; it has no expository structure, such as that of The Divine Comedy . . . . The structure appears to be that of more or less free association, or progression through reverie. . . . The details, especially in the early Cantos, are frequently very lovely, but since there is no structure nor very much in the way of meaning, the details remain details and nothing more, and what we have is the ghost of poetry. . . .
Winters’s critique is typical of many commentaries on the late Cantos in its fundamental charge of formlessness. What makes it useful is that it shows that reading habit and not form is the stumbling block. Winters correctly identifies the primary structural device of the Cantos -- the associative link -- but rejects it out of hand. To him, bound to a sequential, print-bound paradigm of textuality, the only valid structures are linear: narration and exposition. Unwilling or unable to consider linking as a legitimate practice, Winters cannot see anything more than a heap of disconnected "details" in the Cantos.
Sympathetic critics, seeking to defend Pound against critiques like Winters’s, have sought to explain the associative structure of the Cantos through metaphors from other arts. The two most frequent metaphors, collage and fugue, do indeed help illustrate certain principles of Pound’s poetry, but they do not help with the problem of reading the Cantos. The visual metaphor of collage suggests the possibility of creating a whole out of discrete, heterogeneous parts that achieve "meaning" through relation with each other. However, works of visual art are static. They can be viewed and absorbed as complete wholes in a way that no literary work can be. The Cantos in particular cannot be viewed from above or outside, only from within. Thus, while the metaphor seems to explain Pound’s technique, it fails to really provide a method of reading these juxtapositions.
The metaphor of fugue was suggested by Pound himself. In a letter to his father in 1927, Pound offered the following explanation:
Have I ever given you outline of main scheme ::: or whatever it is?
1. Rather like, or unlike subject and response and counter subject in fugue.
A. A. Live man goes down into world of Dead
C. B. The "repeat in history"
B. C. The "magic moment" or moment of metamorphosis, bust through from quotidien into "divine or permanent world." Gods, etc. (L 210)
The usefulness of this concept is that it draws attention to Pound’s use of recurring motifs throughout the Cantos. His frequent return to key figures or events gives them emphasis and triggers the reader’s memory. In this sense he can be understood as creating variations on a limited set of themes. However, any strict application of the fugal metaphor also fails. First, it tends to reduce the poem to a three-part form that oversimplifies the complexities of the Cantos. More significantly, as with visual metaphors, fugue does not describe the reader’s experience of the text. Music (at least performed music) is necessarily sequential. The auditor of a fugue does not have a reader’s ability to slow down, back up, or jump within the work. In hypertext and in the Cantos, this kind of movement is seen as the fundamental activity of reading.
While both mataphors do provide useful concepts for understanding Pound’s novel poetic structure, there remain important experiential differences between reading and either listening or viewing that make them inadequate models for the Cantos. I would suggest that only hypertext offers an accurate model of the nonlinear structure of the Cantos. Hypertext can teach us to read the Cantos.
The question remains, however: how do we read hypertextually? That is the subject of the third and final section of the dissertation. Because hypertext represents such a radical shift from the paradigm of print textuality, hypertextual reading requires nothing short of a new literacy. I will demonstrate the skills involved on Pound’s later Cantos: Section: Rock-Drill (1956), Thrones (1959), and Drafts and Fragments (1968). While the problem of literacy exists for all of the Cantos, it is most pronounced among these later sections; they have drawn some of the harshest criticisms and have perplexed even sympathetic readers. Thus they should prove to be a sufficiently rigorous test of the validity and usefulness of such methods.
Hypertextual literacy requires us to give up the habits we have learned from print, particularly the notion of reading as mastery. Because one of the essential features of hypertext is the offering of choices to the reader, the goal of hypertextual reading cannot be a definitive or "final" interpretation of the text. Choice means that in any given reading we experience only one of many possibilities. Landow cautions: "Quantity removes mastery and authority, for one can only sample, not master a text."
At the center of my approach is the concept of text as space proposed by Bolter. One way of imagining a hypertext structure is as a three-dimensional network, with reading being a form of travel or "navigation" among the connected nodes. Such an image seems particularly appropriate for a poem whose central figure is Odysseus and which begins by setting "forth on the godly sea" (1/3). To read the Cantos, even sequentially, is to undertake a journey. One "moves" from source to source or era to era. Following the sequence of the printed text is rather like following a charted course; the choices have been made for one. However, the Cantos offer other possibilities, opportunities for the reader to make other choices and follow different paths than the linear sequence of the printed pages.
Pound’s use of repetition offers one such opportunity. The recurrence of specific figures, events, and verbatim phrases functions in print like the multiply linked node in an electronic hypertext. These repetitions provide the recurring element with multiple contexts, multiple associations. When the reader encounters such occurrences, he or she is invited to travel mentally, if not physically to other occurrences in the text. One example is the Chinese ideogram, ling2 (85/543), which opens Rock-Drill. Deeply evocative in its own right, ling2 is repeated three more times within Canto 85 (85/551, 552, 555), again in Canto 86 (86/560), and later in Thrones (97/675; 104/738, 740). In Cantos 85 and 86 it serves as a kind of refrain punctuating the generally sequential movement through Chinese history as recorded in the Shu King. George Kearns comments:
At each appearance in Rock-Drill and Thrones, the character brings with it a new context to modify or extend the implications of ling2 in the lives of Elizabeth I, Coke, Senator Benton, the Irish revolutionary and minister Desmond Fitzgerald, and others whose sense of responsibility is ‘rooted’ in awareness of natural order.
Put in hypertextual terms, the ling2 node is linked to several other nodes and the reader follows a looping path away from and back to the ling2 node that develops for it in an increasingly rich context of other associations. Such recurrence implicitly suggests an alternate reading sequence, skipping from one occurrence of ling2 to the next to examine only these various immediate contexts and thus circumventing the linear sequence imposed by the printed page. These reading paths also recall the hypertextual model of thought discussed previously. The increasing aggregations of association around particular nodes are a form of knowledge or meaning in the poem.
A second opportunity for non-sequential reading is offered by Pound’s use of quotation and allusion. Every intertextual reference provides a potential exit point from the Cantos into the larger "docuverse." These links out become quite explicit in the late Cantos, as in this passage from Canto 88:
Use of foreign coin until 1819.
Exception Spanish milled dollars,
every dealer occupied in exporting them, page 446
their exclusion an unconstitutional fraud...
A currency of intrinsic value FOR WHICH
They paid interest to NOBODY
page 446
column two
("Thirty Years", Benton)
Is suppressed in favor of fluctuation,
this country a thoroughfare.
(88/582-83, Pound’s ellipsis)
Here, while excerpting Senator Thomas Hart Benton’s memoir, Thirty Years’ View, Pound gives explicit direction to his source by author and title and even page number. Such references outraged contemporaries, because they violate the traditional sense of a printed text as a complete and self-contained whole. In hypertextual reading, however, it is perfectly natural for one to follow links "out" of a given work. From such a standpoint, it is perfectly permissible to actually follow Pound’s direction to page 446 of Benton’s memoir and from Benton in any number of directions: to other contemporary accounts of the coinage crisis Benton mentions, to biographical material on Benton and his career, to modern economic theories on coinage and monetary value. Any of these paths might eventually circle back to Pound, or they might leave his orbit entirely. The destinations are bounded only by the interest and resources of the individual reader.
The point of these examples is that the text becomes a field of possibilities and reading an exploration of those possibilities. Reading in the age of print has become equated with consumption. To read a book "cover to cover" is to consume and thus master it. Only rarely are books accorded such esteem that re-reading becomes a norm, and even then, such works are often grumbled at for being "difficult." The Cantos foil such expectations of reading. They can only be read exploratorily. Or to put it another way, they can only be surfed.
This surfing metaphor has become popularized in the phrase "surfing the web," describing a non-goal-oriented following of hypertext links on the World Wide Web. I find the metaphor apt for a general theory of hypertext reading, for it points to a fusion of freedom and constraint that I find inherent in the act of reading. The surfer must follow lines of force already present in the waves, but with skill he or she can move with great freedom within and across those lines. So, too, with the hypertext reader. Nodes and links set up certain lines of force; the reader as surfer rides those lines through the work. Each "run" will be different. They can be longer or shorter, yield more or less interesting results. In surfing the Cantos, I do not attempt to master them into a comprehensive trope or plot. Rather, I hope to find interesting lines of force and ride them to find the exhilaration and the pleasure of the text.
I do not know where these explorations will finally lead, but I do expect them to contribute to the fields of both Pound scholarship and hypertext theory. As regards Pound, I hope to finally remove the unnecessary and obstructive print-bound expectations of unity, closure and sequence that continue to hinder the reading of the Cantos. By replacing these inappropriate expectations with the idea of a reader-determined, creative, exploratory reading, I believe the Cantos will generate the appreciation and the enjoyment that they deserve. In addition, by providing a practical test of hypertextual reading, I hope fill an existing void in hypertext theory. While there is already much commentary on what hypertext is not -- its difference from and subversion of other textual paradigms -- there is far less consensus on what it is, on what hypertextual reading can possibly offer in terms of positive meaning. A hypertextual reading of the Cantos should provide a starting point for serious discussion of the skills of hypertextual literacy. As digital media, computers and the World Wide Web continue their meteoric rise within the communication scene, such a discussion seems all the more urgent.
Back to Research Interests || VWC Home