The term was coined in 1963 by Ted Nelson. Though Nelson was working with mainframe computers at the time, his declaration (a "discovered fact") of hypertext had to do with literature, which, he said, was "an ongoing system of interconnecting documents." A hypertext is such a system, made up of smaller segments of text which are connected (or linked) with each other symbolically, for example via electronic code (like this) or footnote numbers.

Nelson's definition speaks directly to hypertext's primary function in this project: the explicit representation of the relationships inherently present between any written text and its cultural context. These intertextual relationships implicitly connect the writing with other texts (which will be my shorthand for what George Landow refers to as "texts or phenomena that one can treat as texts" (35)). The connection may be stylistic, a direct reference to another text, or perhaps even one not intended by an author, which a reader perceives through his or her own socio-cultural context.

The above is certainly not an exhaustive definition of hypertext, which is also a powerful tool for original creation and in itself deserves much more attention. This project, however, is mainly concerned with representing in a new way what has already existed, in different forms, for centuries. As an instrument for comprehending the intertextual complexity of the Roland canon, a hypertextual representation of it proposes a new reading style, which allows for a perception of the character outside of any one storyline. This, as I argue below, is closer to the way in which Roland was originally perceived by those who participated in the oral tradition than reading a book about him would be.

When a piece of print fiction is written, its author does not usually think of his or her creation being read in this networked context. Would attempting to make intertextual connections explicit interfere with the way authors intend for their works to be read?

Not any more than do footnotes, illustrations and different typesetting, which (though they may be extraneous to the content of the original creation) we find to be perfectly reasonable additions to critical editions of literary works.

The literary legacy of mankind is so vast that inevitably there exist works, fiction and otherwise, which a given modern reader simply cannot fully absorb without a set of footnotes written by a specialist. These notes frequently reference not only other writings, by the same author as well as by others, but also the cultural and historical context of the original work. The usefulness of such notes ranges from absolutely essential, in cases where what the text states is so foreign that without an explanation the reader will feel lost and will eventually completely lose track of what the author is trying to convey, to "merely" necessary, in cases where the original text's subtleties in referencing its particular cultural milieu could go entirely unnoticed by a reader from a different culture or time period or both, who has not specifically studied the context in which the work was written.

Representing a piece of writing in electronic form, therefore, becomes a different way of representing the same thing, rather than something new that is being added to an already-existing work. Certainly, the research which inevitably makes up part of a well-constructed hypertext is new; however, the essence of the original remains untouched.

And yet, hypertext also presents the reader with an opportunity to do something which it has been difficult or impossible to do before, except perhaps in "Choose Your Own Adventure" books: make decisions to follow those narrative paths which are most in accord with the reader's own interests. A writer cannot pretend to write for a universal public: individual interests, and even one individual's interests depending on mood, time of day or other affecting factors, will vary. This hypertextual rendition of a composite Roland legend allows the reader to pursue thematic paths of his or her own liking - or, as another choice, to follow a particular storyline, if that storyline is available and spans several documents. An example of the above is the opening scene in Gianni Celati's Orlando Innamorato. The reader is presented with choices. By following the link which is the word "brother" one can read more, from another work, about Roland's sibling-like relationship with his best friend and comrade Oliver. By clicking on the yellow sun (a symbol used in a consistent manner throughout the project), the reader can "step out" of the fictional world and read a bit of relevant criticism which I have inserted there. (If, however, remaining within the narrative is the desire, one is free not to click on the sun, just as when skipping over a footnote.) Following the link symbolized (again, consistently) by the blue bow and arrow at the bottom of the page results in following the current storyline for a while, if that is what the reader wishes to do. The length of these broken-down storylines varies within the project; reading any one work in its entirety is impossible here, because that would achieve a redundant effect with respect to one a printed book achieves.

This kind of decision-making is to some extent available in print literature, where one is free to read chapters in the order one wishes, skip around and go back to re-read parts. However, it is extremely difficult for an author or editor to efficiently "hide" long footnotes or deviations from the main text. The print's answer to that are endnotes, which take time to find in the back of the book and interrupt the flow of the reading process. Delineating thematic paths one could take even within a single work, let alone between works, is also problematic. The physical process of reading suddenly becomes more complicated: repeatedly putting down one book, picking up another, and finding one's place in it takes time and effort, interrupts the flow of the text, becomes difficult with the necessity to adhere to a particular edition's pagination, and eventually makes the reading process much less enjoyable.

In his seminar on hypertextual creative writing at Brown University, Robert Arellano emphasizes two key characteristics of hypertext: its non-linearity and its interactive nature. The term multi-linear seems more suitable: explicit thematic, chronological, critical and other axes upon which the reading of a hypertext rests do exist, and the decisions as to which one to follow, or when to switch from one to another, are left up to the reader. Making those decisions, as well as implementing them by clicking on links, constitutes the reader's interaction with the hypertext.

Stories about Roland have most definitely been invented multi-linearly: the French Song of Roland (SoR) described Hruodlandus' last battle and subsequent death, about 300 years after the historical character died, basing itself (one has to assume) on the numerous oral versions of the story present at the time. The narrative lines progressed (spreading chronologically as well as geographically) into both re-tellings of the SoR (with different emphases according to the writer's purpose - see Konrad's representation of Roland as a single-minded religious martyr) and - more importantly - new storylines about Roland's life, which take place long before the battle at Roncesvalles (see Pulci, Boiardo, Ariosto).

The "Read RolandHT" part of the project allows the reader to experience the character as a whole, discovering for himself through short passages which of Roland's personality traits are omnipresent, and which on the other hand are peculiarities of the particular culture which produced a given text. It is possible to discover all of this in a linear manner, reading poem after poem after novel about him. However, Roland's status as a legend whose image has been constructed in parallel by several not entirely separate cultures over more than nine centuries calls for a viewing of him in this explicitly interconnected way.

Silvio Gaggi, in his From Text to Hypertext, writes:

"Traditional texts, both visual and verbal, address a localized spectator or reader... A conventional book creates the illusion that at the moment of its being read a single author addresses a single reader. In electronic networks no single author addresses any single reader, or, if one does, their exchange emerges from and immediately reenters a broader context of multiple speakers and listeners. There is a polyphony of voices, and the authority of each of them is continually qualified by their mutually commenting on one another. An active, writerly reader interrogates a conventional text and deos not passively consume it, but interrogation is built into electronic conversations." (111)

Arellano raises an important question: with all due respect to the printing press, how exactly has the advent of the mass-produced book affected our notion of literature? Without a doubt, many of the effects of Gutenberg's invention have been positive: the eventual spread of literacy and the much wider accessibility of ideas are two examples of the book's directly affecting Progress. However, some aspects of how we perceived literature were lost, or at the very least largely forgotten. When the literary tradition had been largely oral, the text was not perceived as something of definite authority, and it was subject to permutation every time it was retold, making the process of passing on information much more active both on the part of the storyteller (who had to remember the story, and to add his own version to the parts he may have forgotten) and on the part of the listener (who was in the storyteller's direct vicinity and in circumstances which would not be repeated, and thus made paying close attention necessary in order for the story to be retained, and possibly retold later). A printed book (just as a manuscript, but the effect is much more widespread), a physically defined and almost-permanent object containing a finite quantity of information, is often viewed as the ultimate authority by the reader, who has much more of an opportunity for being a passive participant, since direct dialogue with the author or narrator is impossible.

As with all myths, there is no ultimately authoritative extant written text about Roland, since the original (oral) myth is something we are extremely unlikely to ever trace down without a time machine. Therefore, as Gaggi points out in a more general context, no one Roland myth should be perceived as a main, or authoritative, piece - and, as we have seen above, there is danger of precisely that happening when the reader gets his Roland fix from a printed book. (Even the French SoR, which could be looked upon as the ultimate authority on Roland partly because it is so often relied upon as such by other works, survives in several differing manuscripts.) Therefore, a hypertext which (among other things) forces one to read passages intermittently from different works about Roland presents a better opportunity to perceive the character in his native - which is to say intertextual - context.

There are understandable reservations about presenting a character based on the combination of storylines from radically different cultures and times. What justifies this method of presentation is the fact that Roland the Hypertext (RolandHT) is not meant to represent any one of the characters that make up his personality. This project re-authors Roland's character, presenting a composite - and this is not that radical a shift from everything that has been done before, both in the oral and in the written tradition. Most of the works in existence, unless they directly reference their source (for example, Gianni Celati's re-telling of Boiardo's Orlando Innamorato), do not make clear their specific source(s) of the Roland lore. All of the extant written versions of the Roland myth must be considered composites of whatever came before them. The twentieth-century corpus of Roland-related works, which almost universally refers explicitly to some Renaissance epic poem or other, must be then considered composite as well, since the ultimate origin is unknown.

I do not attempt to place a higher value on some episodes of the Roland lore while omitting others, endeavoring instead to be eventually completist in my representation of facets of his composite personality. In order to successfully combine the German Roland the Martyr with the Italian Roland the [mad] Lover and with the Norse Roland the Warrior, I found points at which these texts have something in common, be they themes, specific episodes, or mutually comparable ways of presenting an aspect of the character's personality. The episodes could be presented on paper, less successfully because the reader would be tempted to read them linearly and passively, and would depend too much on the author's choice of ordering.

The goal of the project, in part, is a more complete narrative. Each author of a Roland tale utilized the knight to suit his own needs. In order to accomplish that, certain aspects of his personality (different in each work, though some are universally present) were amplified, even exaggerated, often obscuring or completely erasing others, which may instead be prominent in another work. The RolandHT project is aimed at presenting a complete picture of this historical-turned-mythological character, calling the reader's attention to the cultural peculiarities presented in different Roland texts via the utilization of critical blurbs throughout. I also hope to make clear, via the links, the intertextuality (whether intended or not) of the different works in the Rolandiana corpus, inevitable in storytelling which originates from oral tradition.