In a hypertext environment a lack of linearity does not destroy narrative.
George Landow: Hypertext, 117.
[E]very interactive fiction depends upon a fiction of interaction. Stuart Moulthrop: “Traveling in the Breakdown lane”, 61.
When reading about hyperfiction Texts, two things occur frequently[1]. First, the special character of hyperfiction as fiction is ignored, and it is treated as any hypertext. Second, usually people write about hyperfiction Texts as promises of what is to come; analyses strictly dealing with characteristics of existing hyperfictions are rare.
I try to relate to the latter complaint in this paper by concentrating on three existing hyperfiction Texts: Michael Joyce’s Afternoon. a story, Stuart Moulthrop’s Victory Garden, and Shelley Jackson’s Patchwork Girl . All these three Texts have been written with Story Space hypertext program and I am discussing the Apple Macintosh versions of the Texts (the Windows versions being slightly different). The complaint that hyperfiction as fiction is ignored, however, requires more scrutiny. There are many implications for the mixing of ‘hypertext’ and ‘hyperfiction’. One of the main implications is the exaggeration of the reader’s interactivity. ‘Interactivity’ is usually interpreted as the possibility - which almost automatically turns to a responsibility - to take ‘the role of the author,’ for the reader to create his/her own story. The hypertext theoreticians have made a distinction between ‘interactive’ and ‘proactive’ reader, but in practice this distinction seems to be forgotten (terms from Moulthrop 1991, 130). The existing hyperfictions largely maintain the distinct role of the author as the ultimate creator of the Text (cf. Moulthrop 1995, 64: “but on the whole, literary [eg. fictional] hypertext keeps the roles of author and reader distinct”, or Landow 1992, 114: “In a hypertext narrative it is the author who provides multiple possibilities,...” )[2] The reader has to some extent the possibility of ‘interacting’ with the (making of the) story, but not of taking the role of a ‘proactive author’.
It is also essential to highlight the nature of hyperfiction stories as narratives. The MUDs (Multi-User Domains, Multi-User Dungeons) are an example of collaborative fiction, but their narrative element is usually weak. Or, if the narrative part is strengthened, as is the case with the adventure-MUDs, the collaborative potential is diminished to mere ‘interactivity’.
The problem is then, mainly, to define ‘interactivity’ with respect to narrative hyperfiction. The lexias of hyperfiction (more or less independent passages of text, which are linked to each other) are pre-defined, as are the links between them. This means that the Text as such must be seen as given to the reader. The reader’s interactivity, then, has an effect on the unfolding or construction of the narrative, but not on the Text per se. A phrase much used with hypertext, “reader-as-author” (cf. Landow & Delany 1994, 29), should be replaced with the more accurate “reader-as-(co)narrator” - but even this is somewhat misleading, since interaction is limited only to certain parts of the narrator’s activity.
This refinement has some practical implications for the developing of hyperfiction interfaces, but as stated earlier, in this paper I am not dealing with what hyperfiction could be. Instead, I will concentrate on the question of how the existing hyperfictions use visual devices to structure their narration. I have chosen examples on the basis of their reliance on alphanumeric text, so that they use illustrations, animation or other such devices only little or not at all. That is, I’m not dealing with primarily visual narratives, but with the visual structuring of primarily written narratives.
First, we must take a more specific look at the concept of ‘interactive narration’. Narrative fiction in general can be described with a three-level model: 1) the Text describes how 2) the narrator tells, what 3) the characters do/perceive (Tammi 1992, 10-11). The notion of ‘text’ with regard to hypertext is problematic, as is demonstrated in numerous treatises. The main argument in (almost) all of them is that there are no fixed boundaries for hypertext. This is an argument which works in two different modes: as Roland Barthes - and many after him - has shown, the reader’s activity in any reading causes the Text to diffuse to an infinite number of other (cultural) Texts; on another level, hypertext links ‘concretize’ the former model, bringing hypertext into relation with numerous other Texts. While the first sense of the openness of Texts is true of hyperfiction Texts also, the latter one is valid only when dealing with ‘real hypertexts’, that is, with hypertexts using external links in addition to internal links. But the hyperfictions I’m dealing here with DO NOT use external links, which radically undermines the claim of unfixed borders for these Texts. To say Afternoon, Victory Garden, or Patchwork Girl are limitless and thus infinite, is acceptable only in the sense that all Texts (printed as well as electronic) are unfixed. So in this paper I do not find the notion of ‘text’ any more problematic than it is with printed fiction. And because these hypertexts as a whole are created by their authors, who in addition to providing the reader with text lexias and links between them, can also program different kinds of restrictions to links, thus controlling the readers’ possibilities, it is obvious that interaction cannot occur at the level of ‘Text’ .
The lexias provide the reader with characters and actions, that is, the level of “what characters do/ perceive.” It is true that these actions can be interpreted in quite different ways, depending on the order of reading the lexias - a phenomenon described by Jane Y. Douglas in her introduction to Victory Garden. Thus the reader’s choices do affect the level of the fictional, represented world, but only in a limited manner. Using the distinction made by Russian formalists, the effects on the level of fabula (the story as a chronological chain of events) are caused by choices made on the level of sjuzet (the story as told).This distinction between fabula and sjuzet has seen many redefinitions - the most significant of which is the introduction of a third level, narration (the things connected to the narrator and to the narration as an act) - and also criticism. The most powerful criticism against this distinction is the claim that there is no pre-existing fabula (represented world), but it is created only after, and depending on, the sjuzet/ narration[3]. This last remark is most important when dealing with hyperfiction; Jay David Bolter for example has claimed of Afternoon: “We could say that there is no story at all; there are only readings” (Bolter 1990, 124). In effect, what Bolter is saying about hyperfiction is that there is no story (fabula), but only plots (sjuzets).
Since the structure of hyperfiction Text is pre-defined, the reader’s choices have the effect of 1) determining what lexias (parts of story/ stories) are read - which, since the existing hyperfictions are limited to a moderate number of lexias, in practice collapses into 2) determining the order in which the lexias are read. So if we use the three-level model - story, plot, narration - in which plot is, somewhat simplistically, understood as the order of telling the events, the reader’s interaction occurs at the level of plot, but in addition he/ she has a significant task in making decisions about the narrative voice. And as mentioned earlier, the decisions made at the level of plot do affect the level of story. To sum up, the reader’s choices have effects on all three levels, story, plot, narration, but the concrete interaction occurs at the level of plot.[4]
Each lexia has a narrative voice, and actually, each lexia in isolation functions according to the general narratological models: “Although conventional reading habits apply within each lexia, once one leaves the shadowy bounds of any text unit, new rules and new experience apply.” (Landow 1992, 4) With hyperfiction narratives, the most prominent problem is identifying the narrator or focaliser in each lexia; closely linked to this is the problem of the temporal dimension. There are at least three temporal levels involved when reading fiction: story time, narration time, and reading time. Trying to establish the place of each lexia in relation to story time, and also in relation to narration time, as well as the identity of narrator/ focaliser, functions in hyperfiction reading very much the same way as in printed fiction - the main difference being the intentional vagueness of many hyperfiction lexias (a practice not unfamiliar to modern and postmodern printed narratives as well), thus ‘empowering’ the reader.
The order of story events is the main thing a reader of interactive hyperfiction has control over. There are different ways to establish this possibility, but usually the main decision the reader can make is whether he/ she follows the ‘default’ story line, or chooses an alternative one. The intentionality of choice is usually strongly restricted, making the choosing seem quite random to the reader. This randomness is heavily argued against by Moulthrop (or whoever has written the instructions “Reading Victory Garden”), who claims that “links are complex and subtle, but never random .” From the viewpoint of the hyperfiction author this is obviously true, but for the reader it is not as obvious. George Landow and Paul Delany have described the rhetorical devices of hypertexts, stating that each link should provide the reader with adequate information about the link (eg. departure and arrival information; Landow & Delany 1993, 19). This of course cannot be applied as such to hyperfiction, but it has some relevance for it, too. For as long as the reader doesn’t know the effects of his/ her choices, the interaction seems random. Randomness naturally has its own role in aesthetics, (and can be used intentionally to some degree[5]), but if there are merely no choices in hand for the reader, with foreseeable consequences, then the interaction collapses to mere guessing. One of the possible solutions to this problem is the use of visual navigating devices for the reader - which Landow sees as one of the most efficient devices for the intellectual mapping of hypertext (Landow 1994, 85; cf. Slatin 1994, 165). These navigating tools are my main focus in what follows, and the ‘visual’ in my title should be restricted to this meaning - the visuality of illustrations or writing as such (typography) get only limited attention .
Each computer program meant to be used by humans needs some kind of interface. Brenda Laurel has developed a theory of interfaces based on the model of theatre (Computers as Theatre, 1993), in which she states that the aim for interface designs should be to create a representational context where people can participate as agents, stripped of the “metacontext” of the interface as a discrete concern (Laurel 1993, 9). One of Laurel’s theses is that using an interface should already establish a shift from the ‘user as audience’ to the ‘user as actor on stage’ (Laurel 1993, 17). There is one small but significant part in the hyperfiction that establishes the user’s (reader’s) presence in the representational world: the cursor-arrow, with which he/ she controls the hypertext. But Laurel’s account in itself is not enough to handle hyperfiction, since for her the ‘representational world’ is the interface itself, the level of virtual text on the screen in the case of hyperfiction[6]. What makes things complicated is the fact that with hyperfiction this level functions to produce another (possibly multilevelled) representational world, that is, the fictional world; and the cursor does not mark the reader’s presence in that world. Because the textual-fictional world is the center of interest for hyperfiction authors as well as readers, it is understandable that hyperfiction interfaces are constructed to draw as little attention to themselves (to the metacontext as Laurel says) as possible. On the other hand, the more interactivity (or even proactivity) the hyperfiction aspires to, the greater the distance from the fictional world; in other words, high interactivity makes the “willing suspension of disbelief” required in experiencing fiction more difficult. It is in this sense that I understand Michael Joyce’s comment that one of the stories of hyperfiction is always the story of its own telling (Joyce 1996b, 160). There are, however, ways to use interfaces so that the fictional illusion is not badly disturbed.[7]
Michael Joyce’s Afternoon. a story, arguably the first hyperfiction Text proper, uses only a minimal amount of visual devices or any kind of navigational tools - which can be seen as one answer to the problem of maintaining the fictional illusion. There is no overview of the conceptual space of the hypertext, and there is no visible signing of the words that ‘yield’, that is, take the reader to an alternative story line. That leaves the ‘tool bar’ as the only distinctive visual means used in addition to text. The tool bar has four ‘buttons’ and a type-in area (see Fig. 1). The first of the buttons is an arrow pointing left: pressing it takes the reader back to the previously read lexias. The second button has an icon of an open book on it: pressing it shows the reader a list of the links possible to choose from the lexias currently read. The third button is ‘yes/no’, which lets the reader answer possible questions posed in the Text, while the fourth button is ‘print’, letting the reader make a hard copy of the lexia currently read. There is also the type-in area, in which the reader can either answer possible questions in the Text, or type any word he/she considers as possibly activating some link. In addition to the tool bar there are two possible ways to navigate through Afternoon: the reader can either press ‘return’ after reading a lexia, which takes him/ her to the next lexia in the default story line (there are numerous default stories depending on decisions made earlier in the reading), or he/ she can double-click any word on the current lexia which he/ she thinks may be a yielding word.
Fig.1. The Toolbar from Afternoon
The icons on the buttons are, then, the only visual devices in Afternoon, but to some degree they do affect the mental structuring of the Text. First of all, the use of the left arrow strongly implies that there is a directional line, with respect to which it is possible to define ‘back’, but also ‘forth’ (the ‘return’ key has a close relation to linear print, suggesting forward movement). In addition, especially when the directions follow the traditional (Western) way of depicting time lines, the button also suggests temporal control, referring to ‘back’ and ‘forth’ in time, too. This temporality can be understood in a trivial way, as the lexia which was read before (in reading time). But it also has strong implications about the temporality of the story line. At least today, when all readers have learned to read with printed literature, the expectation of temporal succession is strong: the lexia read first describes a story-time earlier than the lexia read next.
There are, however, situations where these two temporalities - reading time and story time - do not coincide. When a reader has chosen an alternative link (either clicking a word that yields, or choosing from the list) instead of the default one, and then presses the back-button, he/she will go back only on the level of reading time; with regard to the alternative story line, he/she just abandons it. This has the effect of strenghtening the central role of the default story line, and the alternative status of other story lines: the backtracking takes the reader always (ultimately) back to the default storyline, instead of probing further in the history of the alternative storyline.
Another narrative device heavily used in Afternoon is that of leaving out specific indications of persons, places and times. The lexias themselves don’t create strong feeling of temporal succession or causality among themselves. On the other hand, as Landow has pointed out: “The very existence of links in hypermedia conditions the reader to expect purposeful, important relationship between linked materials.” (Landow 1993, 83)
To sum up, the left-pointing arrow and ‘return’ suggest a (story)line, which reinforce the reader’s expectation of linear temporality, to establish a concept of stable story. The existence of links, expressed with the open book icon, tells the reader that there are different alternative storylines. But because of the linear model evoked, the reader most probably will treat these storylines as intersecting lines or as flashbacks/ prolepses according to the default line, instead of the truly ‘net-like’ or ‘rhizomatic’ structure cherished by most hypertext theoreticians. George Landow has argued that Afternoon “produces an experience very similar to that provided by reading the unified plot described by narratologists,” but only if we consider the plot as something “created by the reader-author [...], rather than a phenomenon belonging solely to the text” (Landow 1992, 116). I think that Afternoon creates the feeling of an underlying stable storyline in much the same way as modernist printed narratives, though there does exist more than one version of the story. This feeling is also strenghtened by the use of open book as the icon for the links: after all, the reading of Afternoon does not differ so much from leafing through a printed book...
The print button has one interesting characteristic: using it produces the hard copy of the text inside the text window only. That is, the title bar (each text window has a title bar), the buttons, or possible background documents, are not printed. Thus it defines in its own very concrete way what is considered as ‘text proper’ - this means for example that the lexia-titles work more like technical properties of links than chapter titles in printed fiction; however, it is difficult to define who is responsible for this, the author, the programmer, the program used?.
Stuart Moulthrop’s Victory Garden has a basically similar, if somewhat more elaborate visual navigation system than Afternoon. The tool bar is the same, the functioning of the ‘return’ key is the same, and there is the possibility of double-clicking on yielding words. With regard to the latter option, there is the difference that in Victory Garden the yielding words can be made visible by alt-option keys (pressing them shows yielding words framed). Also, Victory Garden employs more typographical possibilities of the screen and also includes some pictures (see Fig. 2 and 3).
Fig. 2. A screen from Victory Garden
Fig. 3. A screen from Victory Garden
The most important difference, though, is Victory Garden’s use of a navigational ‘map’ - a visualization of the Text’s cognitive space (Fig. 4 and 5). There is an overview of the map, from which the reader can choose either the ‘Northern’, ‘Central’, or ‘Southern’ part of the map to see in more detail. The map shows names of some ‘important’ lexias, and links between them. The reader can make his/ her entry to the Text by choosing one of the lexias named in the map. There are other possibilities, too, for example a list of different stories with short descriptions, of which the reader can choose one. The map is not just a visualization of the structure of lexias, but in a highly symbolic way, a representational map (the use of qualifiers like “Northern” and “Southern” relate to this) - for example, it can be seen as “a map of a garden with paths and benches” (Coover 1992).
In fact the map offers only very limited possibilities for actual navigating. Only a few of the more than 900 lexias are shown, and the links between those shown in the map are so numerous and complicated that they tell the reader very little. Intentional navigating is not possible, but the map offers one way to choose one’s place inside the story space. The use of a representational map is, of course, one way of making the interface ‘fuse’ to the represented fictional world (even though in this case it is quite a symbolic representation - it isn’t anything like a geographical map of the places of action).
Fig. 4. The navigational map from Victory Garden
Fig. 5. A close view of the southern part of map (Victory Garden)
The map as such, however, has strong implications about the nature of this narrative. The author himself has written extensively about the relation of story and map in interactive fiction, and I will quote him here at some length:
they [the students in a hyperfiction class] had given up hope that the metonymic flow of language in any given node would take them coherently to a conclusion. Instead they were plotting their own readings through a cartographic space, hoping to discover a design which [...] might prove to be buried or scattered in the text. The map, which represents the text as totality or metaphor, was not something to be reached through the devious paths of discursive metonymy, rather it was a primary conceptual framework, providing the essential categories of “right”, “left”, “up” and “down” by which these readers oriented themselves. Metaphor here was not identified with finality or revelation, but with the initial incitement to hypertextual reading, the sense of being precipitated into an unexplored space. (Moulthrop 1991, 128)
Moulthrop seems to have made use of this observation in his own hyperfiction (the Text to which the citation refers was Moulthrop’s interactive adaptation of Borges’ “Garden of Forking Paths” - the Garden in Moulthrop’s own Text is obviously an allusion to Borges). Acknowledging the fact that many links, especially ‘associative’ ones, often leave the reader perplexed, he provides the reader of Victory Garden with a map. Note that the Text to which the quotation above refers did not allow the readers to see any kind of map - they constructed the underlying virtual map solely on the grounds of the navigational arrow-buttons! The map in Victory Garden, however, doesn’t allow the reader the possibility of going systematically through the ‘cartographic space’ - as a procedure for reading fiction this would be, indeed, quite artificial[8]. But otherwise it has mainly the same effects as the students’ imagined map in the quotation: it shows that the Text is limited, there is a limited number of storylines, and the links between the lexias are fixed (this, however, does not mean that the lexias could not be read in many alternative orders). Because of this, the reading of Victory Garden turns easily - at least after some time - into a process of spotting ‘blank areas’ on the map, that is, looking for those parts of the map which are not yet familiar to the reader. This can be seen as a mode of ‘reading as plot’ instead of the more traditional ‘reading for the plot’ (as described by Peter Brooks in Reading for the Plot, 1984), but as Moulthrop himself has noted, the ultimate goal still is to ‘fill in’ the underlying story (Moulthrop 1991, 128).
The fixedness of the Text as whole does not mean that there were not numerous different potential stories. Or to quote Moulthrop again: “To conceive of a text as a navigable space is not the same as seeing it in terms of a single, predetermined course of reading” (Moulthrop 1991, 129). On the other hand, there are limitations:
As we have noticed, however, no hypertextual product can realize the “strictly infinite labyrinth” of Borges’ fantasy. [...] No matter how numerous the branches, some reader [...] could work out all the permutations of a given structure and reduce the hypertext to a finite series of linear narratives, the whole of which could be treated as an eccentric but still quite conventional fiction. (Moulthrop 1991, 129)
What I find most important here is not the actual ‘working out’ of all the permutations, but the realisation that there is this ‘finite series of linear narratives’. If Afternoon’s tool bar already evoked this idea, Victory Garden’s map makes it evident. This can be seen as a parallel to (although somewhat weaker than) the way “the fixity of printed text as an object in physical space makes the text as an object in mental space seem equally stable and fixed.” (Slatin 1994, 155) The fixedness of the representational object (the map of Victory Garden) makes this hyperfiction Text also seem stable and fixed[9].
There is also an interesting dimension to Victory Garden’s map. As Robert Coover has pointed out, the map resembles a garden, the rectangles of lexias being like benches by the garden paths. This interpretation preserves the metaphoric identity of the Text. At the same time as it serves as closure for the Textual level, it also opens up different metaphoric interpretations. It can be read as metatextual comment on reading hyperfiction: it is like strolling around in a garden, sitting down awhile now and then, choosing the interesting looking paths etc. It can also be read against the background of the Persian Gulf War: either cynically, fiction is a closed play ground, where even war can be contemplated without anxiety, or, as parodic comment:”the most important thing is to take care of gardening”. This latter interpretation also bears heavy ironic undertones because of the allusion to Voltaire’s Candide [10].
I will briefly describe Shelley Jackson’s hypertext Patchwork Girl here, even if its status as narrative fiction is not at all self evident. In any case, at least it contains sequences that can best be described as stories, and it employs a visual navigating system (and an underlying structure as a whole; see Fig. 7) quite different compared to Afternoon or Victory Garden. Patchwork Girl also uses illustrations, and colors, in its ‘patchwork’ of texts.
Fig. 6. The toolbar from Patchwork Girl
The tool bar in Patchwork Girl differs from the one described above. It has four arrows as well as a question mark, a two way arrow and three dots. The functioning of this tool bar, though, cannot be understood without the underlying visual presentation of the Textual space (the visual mapping works in three different modes, but I will only concentrate on one of these). The first map - the map here being more like ‘a cognitive map’ instead of a representational one - shows named ‘boxes’ and lines between them[11]. The first level contains the front picture, title page, and the main ‘chapters’. Each box (at least possibly) contains other text constellations. That is, the structure of Patchwork Girl is multilevelled, and this is explicated in a fashion quite a different from Afternoon or Victory Garden. The three dots button allows the reader to choose whether the text-screen or the map-screen is to be shown as uppermost. The down arrow takes the reader one level down in the structure, while the up arrow takes him/ her one level up in the structure. The arrows left and right takes the reader to the lexia either to the left or to the right in the current level, without taking into account the linking. The left and right arrows, then, make possible the kind of going through the navigational space as described by Moulthrop above. The two way arrow, in turn, takes the reader to the next lexia in the default line (with the alt-key pressed down, to the lexia previously read), while the question mark button shows the reader the links leading to the lexia currently read, and the links leading from the lexia currently read, as well as some other information.
Fig. 7. The cognitive map of Patchwork Girl
Without pursuing the implications of multilevelled structure as such, which is, in fact the very model of embedded narration - and underlies (though it is well hidden) Afternoon and Victory Garden as well - I’ll give a couple of examples of the use of visual mapping of the story space employed by Jackson[12]. The first is a section titled “A Crazy Quilt”. In this interweaving of lexias the signification occurs also on the level of visual structure (Fig. 8). The structure of visualization is the crazy quilt, and this structure, instead of any inherent logic of the lexias themselves, governs the whole. The map in this case does not simply represent the order of text (lexias), but provides that order.
Fig. 8. ‘Crazy Quilt’ from Patchwork Girl
There is then not only a quantitative, but also a significant qualitative difference between the visual devices used by Afternoon and Victory Garden on one hand, and Patchwork Girl, on the other. Whereas in the former hyperfictions the visual devices have only a minor role, in the latter one the visual devices partly dominate the narrative.
Although hypertext theory has strongly emphasized the importance of visual elements to hypertext - “the defining characteristics of hypertext [include] particularly its tendency to marry the visual and the verbal.” (Landow 1992, 103) - they are not at all central for the hyperfiction Texts I have discussed here. There are a couple of trivial resolutions to this apparent contradiction. First, Joyce and Moulthrop are both writers, not AD’s or visual artists, because of which they probably feel much more at home with text than with images; Jackson, on the other hand, has previously written an illustrated children’s book, and her familiarity with visual aspects of narrative may be reflected in her more radical approach to visual mapping in hypertext.
The other extratextual reason is, most probably, the need to demarcate ‘serious hypertexts’ (a slogan for Eastgate Systems which has published all the three Texts) from ‘games’. This is quite interesting, keeping in mind the central role to hypertextualists of the poststructuralist notions of writing/reading as something playful - obviously “playing” is more acceptable than “gaming”... I would argue, however, that the technological appeal is quite an important aspect of hyperfiction, too. The narrative techniques used by Julio Cortazar, Vladimir Nabokov, Raymond Federman, Italo Calvino, etc., which for many readers were mere tricks, and annoying distractions, gain now totally different responses because of their adaptation to the electronic environment. Whereas jumping from page to page when reading Hopscotch might have been annoying, it is much more enjoyable to explore the possibilities of hyperfiction interfaces. The aspect of mastering a computer environment is an essential part of the hyperfiction reading experience, an aspect common with playing computer games.[13]
There is no inherent reason for not using visual devices in hyperfiction narratives, and in fact I am inclined to claim that most interactive fiction IS dominantly visual - the most developed adventure games being nothing but interactive fiction. I am quite willing to state that the future of interactive fiction is dominantly visual (and most probably, based on virtual reality devices).[14] The role of writing is, though, quite different in that kind of fiction; in fact it is more convenient to approach it from the theory of film. But that is a different story. Printed fiction did not perish because of film, so there is no reason for text based hyperfiction to perish because of audio-visual interactive fiction. Language has, after all, its own unique properties.
It seems to me that these three hyperfiction Texts, in their text-dominance, provide aesthetic experiences quite similar to that of printed fiction. In addition, I would argue that the use of visual navigation devices just strengthens this effect. Victory Garden resembles realistic conventions mostly because of its visual-representational map. The map helps the reader to conceive of the Text as a coherent whole, with some ultimate metaphoric meaning.
When reading Afternoon, one cannot but notice the many parallels between the ‘secondary orality’ Walter J. Ong ascribes to electronic writing; William Dickey has very similar arguments, though his views of oral techniques differ somewhat from Ong’s (Ong 1995, 143-144; Dickey 1994, 150). The most notable technique of oral literature used in Afternoon is episodic structure. Episodic structure naturally can be ascribed to all three of the Texts discussed here, but with Afternoon it is a central narrative device used in quite a distinctive manner. Ong makes a comparison to ‘primary oral’ literature:
Homer had a huge repertoire of episodes to string together but, without writing, absolutely no way to organize them in strict chronological order [...] organization of the Iliad suggests boxes within boxes created by thematic recurrences,... (Ong 1995, 143)
Repetition with slight variation is maybe Joyce’s most salient technique of Joyce’s in Afternoon, and there is a strong tendency to replace linear narration with episodic structure[15]. As Ong describes:”The poet will get caught up with the description [...] and completely lose the narrative track.” (Ong 1995, 147) With hyperfiction we must naturally replace the ‘poet’ with the ‘reader’, but otherwise this is - potentially - an accurate picture of hyperfiction experience. But the difference lies in the fact that hyperfiction reader has the navigation system available, and he/ she will never actually lose ‘the narrative track’, unless intentionally. And that is the difference between primary and secondary orality: whereas there was no alternative to episodic structure for a pre-print poet, in the case of secondary orality these techniques are self-consciously used as devices. “They are impressionistic and imagistic variations on the plotted stories that preceded them.” (Ong 1995, 150)
Ong’s account of ‘deplotted stories’ is somewhat narrow - he mentions works like Alain Robbe-Grillet’s Marienbad and Julio Cortazar’s Hopscotch, clearly postmodernist Texts. But modernist Texts already went quite a long way towards ‘deplotted’ stories, at least when compared to realistic conventions. I think that the closest precursors to Afternoon are to be found in modernist Texts from James Joyce (sic!) to Virginia Woolf, etc[16]. This also corresponds quite well to Brian McHale’s definition of modernism, as narratives dominated by epistemological questions (McHale 1987, 9). Afternoon is wholly built on the question of whether the car wreck bodies the narrator saw were his ex-wife’s and son’s. The possibility of alternative stories explaining this initial problem is opened up because of the narrator’s unwillingness to get accurate information about the situation (‘the riddle’ of Dickey’s techniques of oral literature, 1994, 150). It is thus the epistemological uncertainty which causes different stories, and their existence depends on informational lack - they are epistemological (fictional) worlds in Umberto Eco’s sense (Eco 1979). Thus the potentiality of ontologically multiple parallel worlds will never be really evoked.
The foregrounding of ontological questions - a defining element of postmodernist fiction (McHale 1987, 10) - can be done in many different ways, the laying bare of the narrative structure behind the story being one of these. The way Patchwork Girl is partially built from the premises of its visual structuring resembles closely some postmodernist techniques. For example, ‘A Crazy Quilt’ shown above, has a striking similarity with Italo Calvino’s The Castle of Crossed Destinies , a story collection written according to a tarot-card deck: the cards are assembled in columns and rows, which intersect each other, and the stories are based each on one row or one column of cards. The effect is then that the stories also intersect each other; they use the same materials, but the meaning of each card depends on the current story. The collection of stories as such weren’t particularly postmodernist (though the intent reader might notice the ‘art in the closed field’ feeling[17]), but the tarot cards accompany each story as illustrations, and in addition, there is one page with an illustration of all the 78 cards, that is, the overall structure of the story collection (to be exact, there are two different tarot decks, and two different story collections in Castle of the Crossed Destinies). It is this open declaration of the way the stories are generated according to a strict procedure which makes Calvino’s book markedly postmodernist.[18]
‘A Crazy Quilt’ also takes as its starting point the graphical layout of the lexias, and through the visual map explicates its own premises. The whole section is thematically and metaphorically linked to the theme of Patchwork Girl, since one of the themes of the Text is that identities are patched together of various textual parts, but also that stories can be seen as constructed this way. ‘A Crazy Quilt’, though, differs from Calvino’s story collection in that its stories are not as tight or coherent as Calvino’s. But as Landow has noted, the mere existence of a link between lexias implies a meaningful, if not causal, relation between them. So the reader is tempted to construct a story from the parts of the quilt. Here is one actual instance where the hyperfiction reader works as a bricoleur, using the materials available and constructing his/ her own story out of them - in fact, in much the same way as Calvino allegedly did when writing his stories. And as Calvino’s example suggests, each reader may take a tarot deck and try to construct stories him/herself, and Patchwork Girl, through its multiple visual navigating devices offers the reader a tempting opportunity to search for stories hidden in the lexias of ‘A Crazy Quilt’.
What I have found about the visual aspects of these three hyperfiction narratives can be summed up in three points: 1) the three Texts I have examined make very little use of visual devices; 2) the visual structuring is mainly done with navigational tools; and 3) the presence or absence of visual means has an impact on the overall aesthetics of the Texts, so that the (virtual) absence of visual effects in Afternoon makes it appear episodic and epistemologically unstable, the use of a representational map in Victory Garden restores the coherence of the Text and makes it possible for the reader relate to it in the same way as to realistic narration, and the use of the conceptual map of Patchwork Girl as one site of signification lends it a postmodernist feel.
The question of the bricoleur, though, is a crucial one and calls for some remarks. There may be strong opposition to my account of hyperfiction, because I rely too heavily on the narratology of printed fiction, and neglect the role of the hyperfiction reader as a bricoleur (cf. Landow 1992, 115, speaking about Afternoon:”This construction of an evanescent entity or wholeness always occurs in reading, but in reading hypertext it takes the additional form of constructing, however provisionally, one’s own Text out of fragments, out of separate lexias. It is a case, in other words, of Levi-Strauss’s bricolage, for every hypertext reader-author is inevitably a bricoleur”). I have concentrated here solely on three existing hyperfiction Texts, which means that my conclusions are not automatically to be generalized - hypertext as such is such a rapidly developing area that the nature of hyperfiction most probably will change drastically in the near future (a possibly perpetual state, as claimed by John Slatin 1994, 153), and there most probably already are hyperfiction Texts which operate in totally different ways than the three Texts described here[19].
With these three Texts I cannot see how the reader could actually work as a bricoleur - the mostly random selection of links barely makes one a bricoleur, and random it is, since as many have noted, the associative links seldom seem evident for the reader (cf. Jane Y. Douglas’ Introduction to Victory Garden:”’What comes next,’ and what I expect to see come next can be two entirely different quantities.”). John Slatin, for example, when writing about interactive poetry, has made this quite clear:”The difficulty here, of course, is that what are to me self-evident associations may not be even faintly apparent to you, and vice versa.” (Slatin 1993, 162) Only Patchwork Girl of these three Texts has some potential in this direction, mainly thanks to its use of conceptual mapping. I think that to make the reader more interactive, a real bricoleur, requires more information about underlying structure. It is just on this metatextual level that the structuring of hypertext, as well as hyperfiction, occurs and J. David Bolter has quite fittingly cited Ricoeur writing about the ‘followability’ and second order writing of hypertext. Slatin has a similar argument to make:
I think of hypertext coherence as appearing at the metatextual level - that is, at the level where the reader perceives [...] “the pattern which connects.” [...] This metatextual level is perhaps best represented by a visual map of some kind,...
Hyperpoet Jim Rosenberg’s ideas about the ‘external syntax’ - how visual means can be used as a ‘channel for syntax’ allowing, for example, a kind of simultaneity even in textual representation - of hypertext seem to be accurate with hypertext fiction, too. If narrative as a whole is seen - in structuralist fashion - as ‘an extended sentence’ - the visual navigation tools can be seen as providing the syntax for that sentence, the “pattern which connects”. (Rosenberg 1996a)
Even if these three Texts do not offer the possibilies for a true bricolage, this is not to say that they have somehow failed; quite the opposite -- they are excellent works in their own right, and make very efficient use of their limited interactivity. It is necessary to develop already a vision of future interactive fiction, and to think about the consequences it has for narratological theory (or reader-response theory, or the theory of fictionality), and not to take existing hyperfictions as the final statement[20]. But on the other hand, I find it also necessary to study these Texts as they currently are, not only as pre-texts for what is to come. Since reading habits change slowly, it is useful to note how efficiently the traditional views of narratology and fictionality can be made to work with hyperfiction. And although the traditional concepts inevitably function as obstacles to seeing the truly new aspects of hyperfiction, they may also reveal some inherent characteristics of narrative fiction which are valid also with respect to interactive narrative fiction.
[1] I use Text, with capital ‘T’, in the sense of ‘textual work’ to differentiate it from text, with which I mean alphanumeric text. All the citations naturally appear in their original form, thus not maintaining this distinction.
[2] Some scholars have even argued for increased authorial control, cf. Jane Y. Douglas 1989, or, Bolter 1994, 116: “On the other hand, the author has a unique opportunity to control the procedure of reading, because he or she can program restrictions into the text itself.”
[3] One of the best criticisms of fabula/ sjuzet distinction is Barbara Herrnstein-Smiths article “Narrative Versions, Narrative Theories” (1981). However, I think Herrnstein-Smith fails in her aim to totally get rid of this distinction, but she very convincingly shows the invalidity of the idea of a pre-existing, ‘actual’ story.
[4] For an excellent study of the relationship between Gérard Genette’s narratological model and hyperfiction, see Liestol’s “Wittgenstein, Genette, and the Reader’s Narrative in Hypertext” (1994). He sees hyperfiction narratives basically as four-levelled: 1. Discourse as discoursed, 2. Discourse as stored, 3. Story as discoursed, 4. Stories as stored (potential story lines) (Liestol 1994, 97). The story as discoursed can be seen as parallel to my idea of the reader’s choosing the order of events told, discourse as discoursed being parallel to the reader’s activity as co-narrator (who is restricted by the ‘discourse as stored’, that is the hypertext as whole, including not only the text lexias, but also the linking structure).
[5] This is the case in some hypertextual poetry as described by William Dickey 1994, 147:”The sense of chance, of an aleatory element affecting the viewer’s understanding of the work, has been incorporated as a fundamental element of the poem.”
[6] For Laurel, a text seen on the computer screen is always ‘virtual text’, since it is a representation of code stored in the computer memory.
[7] Because of this, Landow & Delany’s remark on the possibly necessary restriction of the reader’s freedom to interact with hypertext seems even more accurate with hyperfiction (1994, 21).
[8] It is to be mentioned that all hypertexts written by Story Space can be read as stand alone applications or with the Story Space program installed. In the latter case the reader can see the conceptual map of the hypertext structure even with Afternoon orVictory Garden. This conceptual map is discussed in more detail later in relation to Patchwork Girl, since it allows the reader to see the conceptual map also in stand alone version.
[9] There is another, less fixed way to see Victory Garden, stressing the multitude of its materials and oscillating between several possible interpretations (SEE: "Reading Victory Garden").
[10] The mention of Candide here is not random, quite the opposite, since it evokes the Leibnizian theory of possible worlds (the corner-stone of Leibniz’s theodicy “this is the best of all the possible worlds” being wildly abused in the novel). I believe the theories of fictionality based on possible worlds semantics (see especially Ronen (1994) and Ryan (1992)) are most appropriate when dealing with hyperfiction.
[11] I borrow the term ‘cognitive mapping’ from Fredric Jameson’s article with the same title (1988). There is, however, no room here to discuss the relation of Jameson’s cognitive mapping to hypertext mapping.
[12] While the idea of multilevelled, embedded narration can be already found in such classics of narratology as Wayne C. Booth’s The Rhetoric of Fiction (1961), and Gérard Genette’s Narrative Discourse (1980), more thorough treatise of emebedded narration and its use in fiction- and very useful in connection with hyperfiction - is Marie-Laure Ryan’s Possible Worlds, Artificial Intelligence, and Narrative Theory (1992), (cf. McHale 1987, 112-130).
[13] In fact, Stuart Moulthrop (1991, 130) uses the game comparison when writing about interactive and proactive readers:”Readers could very easily learn to approach hypertexts not interactively but proactively, not as players but as masters of the game.” For a discussion about the role of ‘play’ and ‘games’ in fiction, see Waugh (1984, 34-47).
[14] The increasing visuality of hyperfiction is strongly argued for by Michael Joyce:”We are headed to what Don Byrd calls the post-alphabetic age. The image impinges upon the word so much as to imperil its hegemony and maybe even its meaningfulness. Storyspace will become more visual even as it insists upon the sensuality and visual qualities of the word.” (from the interview with Shady Cosgrove, quoted in the user’s manual for Afternoon). It is also worth mentioning that visual representation, especially if animated, requires a lot of memory and programming time: the writing of a computer game today strongly resembles film producing, involving tens if not hundreds of designers and programmers; also it is the games which require the most of performance of home computers.
[15] This is closely related to hypertext’s general tendency towards fragmentation, because of which “reading units take on a life of their own as they become more self-contained because less dependent on what comes before or after in a linear succession.”, Landow & Delany 1994, 10.
[16] Jane Yellowlees Douglas makes the same comparison with respect to WOE - Or a Memory of What Will Be - another hyperfiction by Michael Joyce.
[17] About postmodernist strategy of artificially restricting the materials available to the author, see McHale (1987, 64; 70) and Waugh (1984, 47-48).
[18] The poetics of the (mostly) French literary group Oulipo is largely based on different kinds of combinatorial and permutational procedures (for example, Raymond Queneau’s Excercises in Style, which includes about one hundred variations of one story); not surprisingly, Italo Calvino (even though Italian), was one of the most prominent members of Oulipo.
[19] It can be argued that Story Space (with which all these three texts are written) is a program especially designed for producing hypertexts dominated by alphabetic text. It it thus an interesting question whether we should treat hyperfiction texts written with different programs as representing different genres...
[20] For example, Deena Larsen’s hyperfiction Marble Springs (Eastgate Systems) includes ‘margins’, blank spaces where the reader can write her comments or own stories - Larsen encourages the readers to send their contributions to her, and some of them may be inserted to the forthcoming upgraded versions of the Marble Springs. That is, the interactivity may take some radically new forms in future.