READING HYPERTEXTTHEORETICAL AMBITIONS AND EMPIRICAL STUDIES

Abstract

Recent discussions of hypertext fictions have at times suggested that conventional print literature is now superseded, that a computational approach to understanding is required, or that the principal source of influence on the reader is the material embodiment of the literary text whether its medium is print or digital. Such arguments overlook the sophisticated array of reading processes that help direct a reader's engagement with any fiction text. Empirical and theoretical studies of reading show that these include narrative expectations about character, plot, and setting, and four types of feelings: evaluative feelings, feelings about narrative aspects, aesthetic feelings, and self-modifying feelings. Yet readings of hyperfictions so far have been limited mainly to aspects of plot. In an analysis of one section of Caitlin Fisher's hyperfiction Waves of Girls, the effect of the design of links, graphics, sounds, and other elements on reader's feelings and narrative expectations is examined. This analysis suggests that readers of hyperfictions can be both immersed and interactive, but that a better understanding of the reading processes that facilitate this is required.

In the Electronic Book Review two years ago, Marku Eskelinen referred to »the dead ends of hypertext theory and its posthuman derivatives«, echoing a common assumption that earlier debates about the problems of literary hypertext are now irrelevant. The founding arguments of authors such as George Landow, J. David Bolter, and Michael Joyce have been superseded. Eskelinen's own mandate is to demonstrate the combinatorial properties of cybertext. This, he claims, will show »how the textual medium works«. He will achieve »this goal by approaching computers as computers, and not [...] as something completely different, be that theatre, cinema, comics or (poorly read) continental philosophy«.[1] While it is true that earlier attempts to situate hypertext in relation to poststructuralist text theorists such as Barthes or Derrida now seem misplaced, even naïve, I will suggest that the questions these accounts raised about the status of hypertext have not been superseded so much as abandoned.

Hypertext fiction, or hyperfiction as I will call it, continues to be written, with numerous websites devoted to either fictions or critical discussion; it has even broached the pages of the Norton anthology of postmodern American fiction (with excerpts from Michael Joyce and J. Yellowlees Douglas). Whether this signals its arrival on the stage of mainstream literature remains in doubt. As Stuart Moulthrop has recently complained, »Why does no corporate publisher offer a line of hypertexts? [...] Electronic texts have indeed been marginalized, excluded, and misrepresented by leading institutions of print culture.«[2] Yet the active and energetic presence of hyperfiction poses a challenge to our conceptions of what text is and how we might relate to it. In particular, it provides a significant new perspective on narrative techniques that have long been familiar to us from traditional literature: hyperfiction defamiliarizes questions about narrator, focalization, temporal and spatial setting, and other elements. Thus the surprise for me, contemplating the critical literature on hyperfiction, is how few systematic accounts have been provided of the experience of reading it.[3] By reading, I don't mean proposals such as those of Jim Rosenberg, who elaborates a valuable theoretical description of readerly functions,[4] but phenomenological accounts of the flows and disruptions of reading as these unfold in relation to a specific hyperfiction.

This seems a central issue if we are to resolve some of the outstanding questions about hyperfiction. But we are unlikely to find any help in Eskelinen's approach, where two typical strategies forestall attention to reading. First, he confines text to a set of formal operations: he urges us »to see a text as a concrete (and not metaphorical) machine consisting of the medium, the operator, and the strings of signs«.[5] Katherine Hayles remarks, in a comment on Eskelinen's paper, that in this approach »cybertext theory elides materiality in order to create a template based on function, generally casting a blind eye to how these functions are instantiated in particular media«.[6]

Second, Eskelinen disparages the texts of traditional or ›linear‹ literature, which he characterizes as »literary objects that are static, intransient, determinate, impersonal, random access, solely interpretative and without links«.[7] In these respects, Eskelinen's version of cybertext theory imposes a theoretical, computer-derived grid on its objects of study; moreover, it sacrifices the contribution that study of the actual richness, indeterminacy, and personal variability of literary reading would provide. Like a modern-day Plato shaping his digital Republic in the image of the computer, literary experience as we have previously known it is to be banished. He is not alone: Friedrich Kittler also engages in the argument that literature as we have formerly known it is superseded. Electronic media, he remarks, ensure that »the hallucinatory power of reading and writing has become obsolete«.[8] But cutting hyperfiction off from the literary tradition seems shortsighted: as Hayles remarks, »To think of hypertexts [...] as depending primarily on computation for their effects is to render them virtually unintelligible as works capable of making readers care about the stories they tell.« What we require now, Hayles adds, is a way of understanding »how computational operations work together with linking structures and literary devices to create richly textured works that are something like computer games and something like literature«.[9]

Hayles's own more sophisticated approach, however, elides the specifically literary components of reading in favour of a theory that foregrounds materiality as the basis of reading. For her, »the physical form of the literary artifact always affects what the words (and other semiotic components) mean«.[10] It is clear that the material or technical media of hyperfictions play a significant role since, unlike the incidental process of turning the pages of a linear narrative in a printed book, the reader must interact deliberately with the medium itself in order to continue reading. But hyperfictions, whatever computational or game-like processes they contain, are also narratives. Whatever the medium, readers bring to narrative a range of expectations and capacities drawn from their experience with the various forms of narrative (plot, character, focalization et cetera), as well as experience of their own stories in life.[11] They are also likely to bring a rich understanding of poetic language ranging from early childhood verbal play to the work of Dickens. It seems unlikely that this experience is left behind when the reader enters the hyperfictional world. But Hayles's move to privilege the material basis of textuality forecloses the possibility that reading processes may operate independently of »the materiality of inscription« (p. 130), providing a separate source of influence on how readers respond to and construe the texts they read. Thus my topic in this chapter will be how hyperfiction engages the reader's narrative and poetic capacities, how far the hypertext machinery modifies them, and what effects this has on the processes of reading. These issues are best raised not in the abstract, but in relation to a specific hyperfiction, hence the importance of stories of reading.[12]

While many kinds of reading will be required, those currently available focus mainly on story events and the machinery by which they are assembled. For example, in the readings of Joyce's Afternoon, Jill Walker (Piecing) focuses much of her attention on the characters, their relationships, and reconstructions of what may have occurred;[13] J. Yellowlees Douglas, in the most extensive account available, focuses principally on how her four readings eventually enabled her to reach closure on the main questions raised by the plot.[14] While it may be true, as J. David Bolter puts it, that in hyperfiction the reader's struggle is »to make the story happen and to make sense of what happens« (p. 126),[15] the story-driven reading is only one of several possible approaches to narrative. A reader can also read for the pleasure of the language of the text itself, its sensory and imagistic qualities; she can read for the insights the text might offer into her personal situation or the culture she inhabits; or read for signs of intertextuality, the presence of quotations and allusions to other works of literature or media. Whichever perspective is taken there are specific pleasures and frustrations involved in reading narrative, thus one productive way of situating story-driven reading and going beyond it is to consider the affective dimensions of reading. What kinds of affective response are invoked and what are their implications?

A preliminary proposal in this respect has recently been made by J. Yellowlees Douglas and Andrew Hargadon in a paper that foregrounds two affective states, immersion and engagement.[16] These states are broadly similar to the immersion and interactivity discussed by Marie-Laure Ryan.[17] Douglas and Hargadon anticipate that hyperfiction, or »interactives«, may eventually enable what Csikszentmihalyi calls »flow«, »a condition where self-consciousness disappears« (p. 163). For Ryan, on the other hand, immersion and interactivity remain irreconcilable. The problem is the hypertext link, she argues, »because every time the reader is asked to make a choice she assumes an external perspective on the worlds of the textual universe« (p. 20). Thus Ryan's discussion of feeling is limited to states of immersion, where she suggests that we experience suspense, defined as »the reader's desire for the knowledge that awaits her at the end of narrative time« (p. 140). She notes that it is (paradoxically) possible for us to experience real emotions during our reading about fictional characters and events. In the multiple worlds of hyperfiction, however, the offering of alternative fates to characters militates against the experience of real emotion: »Emotional immersion requires a sense of the inexorable character of fate« (p. 263).

Douglas and Hargadon, in contrast, appeal to schema theory to help explain the feelings of the hyperfiction reader. They suggest that »the predictability afforded by genre schemas makes them ideal fodder for the trance-like reading [that] is the hallmark of the immersive reading experience«; in contrast, the disruption of schemas obliges the reader »to assume an extra-textual perspective on the text itself, as well as on the schemas that have shaped it and the scripts operating within it« (p. 156). In other words, reader's feelings are contingent on the instantiation or interruption of schemas during reading. The engaged reader is said to represent a development of earlier modernist practices, for example, the reader who follows up the footnotes to T. S. Eliot's The Wasteland, or turns to a critical explanation of Joyce's Ulysses. »Readers of hypertext fiction«, they suggest, »like Joyce's and Eliot's audiences, are more likely to seek out secondary sources to supplement their array of schemas for understanding the text« (p. 161). To read in this multiple, critically aware way would appear to preclude the single-minded, absorbed state of immersion.

The reliance of Douglas and Hargadon on schema theory points to a difficulty in their account. Schema theory has not been successful in explaining feeling during reading. As Rand Spiro argued some time ago, schema theory provides an inadequate basis for understanding experiential (feeling-based) responses to texts.[18] First, as Spiro showed in the case of a short story by James Joyce, we can understand a story (experiential understanding) without knowing what it is about (schema-based understanding). Secondly, during reading the informational, situational aspects of a text can become ›overlearned‹ and relegated to the background; what captures attention is our feelings about the text. Spiro's account implies that schemas and feelings are thus separate sources of meaning. He proposes that reading involves two levels: first, the comprehension process of assigning events to types, second, the invocation of personal meaning (p. 82);[19] the two levels may often occur concurrently, a suggestion that would allow us to infer that feeling as well as schemas contributes to textual coherence. In fact, the relation of feelings and schemas may often be the reverse of that proposed by Douglas and Hargadon.

The Russian critic Victor Shklovsky argued with the view of Herbert Spencer, who saw style as organized to make reading efficient – what one might see as the smooth instantiation and unfolding of schemata. On the contrary, said Shklovsky, in literature style is designed to create a special, unusual perception of an object. It works to deautomate perception,

that one may recover the sensation of life; it exists to make one feel things, to make the stone stony. The purpose of art is to impart the sensation of things as they are perceived and not as they are known. The technique of art is to make objects ›unfamiliar‹, to make forms difficult, to increase the difficulty and length of perception because the process of perception is an aesthetic end in itself and must be prolonged.[20]

Shklovsky insisted that literature makes »one feel things«; stylistic devices in literary texts »emphasize the emotional effect of an expression«.[21] This has largely been overlooked in discussions of the literary ›devices‹ that he analysed. His account is consistent with an earlier generation of British romantic writers. For example, in praising Wordsworth's poetry Coleridge referred to the poet's ability »to combine the child's sense of wonder and novelty with the appearances, which every day for perhaps forty years had rendered familiar«.[22] The textual devices that achieve such effects were described as »foregrounding« by Czech theorist Mukarovský:[23] they include such features as metre, assonance, ellipsis, metaphor, and irony.

In this perspective, foregrounding arouses feeling in the reader, and has the power to subvert schemas. This suggests, indeed, that one of the aims of literature is to challenge our familiar schemas.[24] As our research with readers has shown, response to foregrounding calls into question the positing of a divide between immersion and interactivity and the ways this has preoccupied hyperfiction critics. Readers of literary texts are capable of both immersion and self-awareness as readers: they both respond with feeling to a significant detail of a text, yet at the same time they can savour the power of the language or consider its implications for themselves.[25] We have found this, for example, in some of the responses we collected to Coleridge's narrative poem The Rime of the Ancient Mariner, when readers were asked to think aloud about passages they had found particularly striking (Kuiken, Miall, and Sikora, in press).[26] If interactive reading means adopting, in the words of Douglas and Hargadon, »an extra-textual perspective on the text«, then our readers were, at such moments, considering multiple perspectives on the same text – although not through the agency of alternatives provided through links to other lexias; and when foregrounding challenges or disables existing schemas, it is the emergence of alternative perspectives that, in particular characterizes literary reading. We can understand such reading as interactive, although the reader is thoroughly absorbed by the multiple perspectives in question.

Considering the role of feeling in reading traditional literary narratives can give us a better purchase on what may be distinctive in reading hyperfiction. Feeling is at issue in more than one way, however, not only in response to foregrounding. Following earlier work by Kneepens and Zwaan (1994),[27] we have recently proposed a four level typology of feelings in literary response.[28]

First, evaluative feelings towards the text: the pleasure or frustration experienced during reading, or felt in retrospect towards the text as a whole. Readers often turn to the same genre time after time (e.g. romance fictions) because they anticipate the kind of feeling that reading another text will induce. It seems likely that readers of literary narratives or hyperfictions, which vary so much one from another, are less likely to be in search of a standard feeling or set of feelings, but readers undoubtedly evaluate hyperfictions in the light of their expectations and whatever satisfactions they experience.

Second, narrative feelings in response to specific aspects of the fictional events, such as empathy with a character, intrigue over a turn in the plot, or resonance with the mood of a setting. The basis of such feeling lies in our social skills, our everyday experience in understanding and situating the lives of others. In reading fiction we play out a simulation, as Keith Oatley has put it,[29] running the action plans of the characters on our own planning mechanism, and experiencing the feelings consequent on their actions. We may need to distinguish feelings that occur in response to a character (e.g., pity) from feelings that the reader shares with a character (e.g., disliking another character). Since hyperfictions are narratives, a dependency on narrative feelings is inevitable, although a hyperfiction may be designed to subvert or undercut them.

Third, aesthetic feelings of the kind outlined above in the response to foregrounding, that is, feelings evoked by stylistic moments that are unusual or striking. These are moments that may challenge reader's assumptions (or schemas), leading them to revise their framework for interpretation, sometimes with consequent implications for their understanding beyond the text – which touches on the fourth level.

Fourth, what we have termed self-modifying feelings that restructure the reader's interpretation, prompting the reader to new insights into herself or her world. Some transformation in understanding is brought about, through feeling, in the self of the reader. Here is where readers may become simultaneously immersed and interactive. Whether this is likely to occur in the context of reading a hyperfiction is a matter for empirical study. If Hayles's proposal[30] that hyperfictions are part literature part gaming is correct, the satisfactions of solving a game may be incompatible with self-modifying feelings.

Empirical study of readers underlines the inadequacy of Eskelinen's characterization of literary texts as »static, intransient, determinate«[31]: readers show not only considerable variation between their readings (pointing to the indeterminacy of the literary text), but also much flexibility within readings in the perspectives taken from one episode to the next. The multivariate nature of literary reading can, of course, be captured only very partially by such a typology of feelings. But, as noted earlier, we need not suppose that the processes of reading change fundamentally when directed towards hyperfiction. Thus we can propose that these four sources of feeling will be present during hyperfiction reading. If there is a significant departure from this range of occasions for feeling, that in itself may provide us with an indication of what distinguishes hyperfiction from traditional reading. »If there is a defining flaw of the cybertext debate,« Scott Rettberg has noted, »it is a failure to take into account the ›non-trivial effort‹ of ›mere‹ interpretation that even lowly works of linear literature require.«[32] The difficulty in understanding hyperfiction lies in the additional complications that the digital medium places on the work of interpretation. While these may not constitute »constraints«, as Rettberg puts it, they represent additional sources for feeling beyond the textual ones outlined above. They may include not only the much-discussed link as well as other navigational features, but also the inability of the reader to judge the size of the textual collection she is reading and her place within it, the often idiosyncratic role of images and such graphic elements as changes in type font and size, machine-driven changes or choices, the role of sound, and the use of multiple windows.

To give the discussion a concrete form, I turn now to an analysis of one specific hyperfiction. I will consider both the narrative features it presents and the complications of the hypertext medium in relation to the typology of feelings outlined above. I focus on These Waves of Girls by Caitlin Fisher, a web-based hyperfiction[33] which received the fiction prize in the Electronic Literature Organization's 2001 awards. In comparison with some other entries on the shortlist, Waves is not technically sophisticated; it employs the hypertext structure of lexia and links familiar from the earliest hyperfictions, such as Michael Joyce's Afternoon (1987).[34] Included within it, however, on almost every screen, are a range of graphics that surround the text or provide a background for it, intimating by their proximity a certain timbre to the text, or a comment on it that may support or cut across its apparent meaning. Fisher also provides a number of sound files: beginning with the sound of girls' laughter that accompanies the opening screen, other files provide occasional sound effects, or offer an oral reading of the text currently on screen in (I assume) Fisher's own voice. The author herself describes her work as

a hypermedia novella exploring memory, girlhoods, cruelty, childhood play and sexuality. The piece is composed as a series of small stories, artifacts, interconnections and meditations from the point of view of a four year old, a ten-year old, a twenty year old.[35]

Larry McCaffery, the fiction judge, whose comments are provided on the ELO website, refers explicitly to the feelings that Waves arouses. Its elements, he says, are »[l]inked in often surprising ways that establish hidden connections that often seem to be operating on the basis of emotional, associational logic«; it »is by turns, tender, terrifying, erotic, lyrical, witty, surprising«.[36]

No single narrative lies behind the numerous lexias, although thematically many of the lexias are related by explorations of the lesbian feelings of the characters. It offers a highly varied set of narrative fragments illustrating the lives of girls, involving a wide cast of characters and a range of different settings and incidents. The narrative voice may be that of a single character, whose identity from childhood to adulthood is revealed in a series of snapshots; we hear a participant narrator in every lexia. The work is organized into eight sections; each section provides its own menu of links in a column on the left. Within lexias links vary in direction: some continue a narrative thread across several lexias; others jump the reader to a quite different scene, both temporally and spatially. Most lexias provide several text-based links (usually signalled by the familiar underlined blue font of the standard web browser); links from graphics also occur, although these are unpredictable, since on a given screen some graphics will provide links while some will not. A few lexias consist only of a graphic with one forward link. Overall, the design avoids the reader traps set by Joyce's use of Storyspace in Afternoon, in which the reader periodically becomes stuck in a loop. After reading for a while, however, links return to lexias seen before, and as this occurs with increasing frequency there is a sense of eddying around the same set of preoccupations. The texts of the lexias vary considerably in length. The shorter lexias consist of just a few words. The longest appears to be waves/mr_anderson.htm,[37] which consists of nearly 1700 words – as long as some complete short stories – which requires repeated scrolling down the screen. It describes three years of high school, the narrator's interest in the science teacher Mr. Anderson, and a sexual fantasy she has about him some years later after she learns that he is dead – her fantasy is poetically mirrored by the spacing of the text on screen and the graphics.

To consider the feelings invoked by Fisher's work, I will focus on just a few lexias and their settings within the network of lexias that surround them. I will, of course, be unable to do justice to the graphics that the work contains. Here is one short lexia, showing the links as underlined text (waves/farmsky.htm):

The text is embedded between two graphics: below is an indistinct black and white image including a square shape that appears to be a box; above is the Flash-based image of a blue and grey moving sky repeated (in smaller form) from the opening screen; the sound of girls' laughter which accompanied the opening screen also plays, which may indicate that this lexia is intended to be central in some way.

I have approached this lexia along a pathway of three related lexias describing memories of a five-year old girl on a farm (falling out of an apple tree »one hot July day«; the cows; eating asparagus;[38] note that it is possible to arrive at this lexia via other pathways; I will not consider the effects these create in the present discussion). Each prior lexia instantiates the schema of a small girl playing on a farm, her attention being caught first by one thing then another; no narrative line is developed, as no causal relations are developed between the successive lexia. But the farm setting for the current lexia has been established: no shift in schema or conflict of schemata is at issue when we read it. The reader's attention is thus likely to fall on the sensory qualities of the lexia, which is a frequent feature of Fisher's writing – that lyrical quality mentioned by McCaffery. I will single out three foregrounded features: (a) the physical pricking of the heat is emphasised by the /t/ alliteration in the first few words; (b) the unexpected deictic implications of »my sky« and »the field«, since no previous referents have prepared for the assumptions made here; and (c) the enactive metaphor of »clouds turning to fierce animals«. Each has a specific affective charge, enlivening our sense of this five-year old girl's situation: her bodily awareness, her appropriation of the environment above and around her, and her propensity for fantasy.

As our previous research with readers has shown,[39] and in support of Shklovsky's argument, foregrounded passages are defamiliarizing, calling into question the adequacy of the reader's current schemata while, at the moment of reading, not offering an alternative (this may emerge later[40]); thus feeling forms a significant component of the reader's response at such moments. The issue here, then, is to understand what role feelings in response to foregrounding are likely to play: what influence will they have on the reader's perspective and understanding, and, more specifically, in what ways does the hypertext medium interact with such feelings?

First, if we consider the lexia itself, the role of the graphics is unclear: the moving sky above the text contains clouds (layers of stratus forms), but these are not amenable to being seen as »fierce animals«, so this cannot illustrate the sky seen by the girl in the text. The lower graphic seems unrelated to the scene. The sounds of laughter also seem unrelated to the girl's situation, at least at a first reading: she is alone in the field, yet there is a chorus of girls laughing. Possibly the multiple or fractured selves of the girl are represented in this way, the self throughout this hyperfiction, especially in lexia depicting young girls, often being a conflux of mismatched desires and fantasies. Here I can reason with myself that, by wearing a two-piece bathing suit the girl is mimicking the appearance of older girls, yet her behaviour in fantasizing about the clouds is that of a young girl. This, however, is a reading that I make on intellectual rather than sensory grounds: on the face of it, the sounds of laughter are unconnected with the sensory resonance of the text before me; it has no direct connection with the three foregrounded features I have identified.

Second, since several links are provided, we must ask whether the affective charge of the foregrounded passages is developed in subsequent lexia. We might infer, since the links are visible in the text, that Fisher has in mind a reader pressing (as it were) on words that »yield«, following the approach of Michael Joyce in Afternoon (to quote the information leaflet accompanying his software): two of the three foregrounded passages provide links. If immersion is to become a possibility for the reader, it is here that we should expect it, with a subsequent lexia that takes up the affective implications of the word or words in the first lexia that provides the link. Considered in this light, however, the effects are inconsistent.

There are five links. Tabulated in order, the opening texts of the subsequent lexia are:

1. a two-piece bathing suit links to: »Summer of '76 covered in baby oil, a bit of bikini, wet cotton swabs over my closed eyelids.«

2. my sky links to: »I am growing up but not out of my grandmother's bed. As a small child, I breathe with difficulty and in the middle of the night the utter silence in the room wakens my grandmother.«

3. the field links to: graphic of woodland, text at foot: »At night the fireflies come out.«

4. lower graphic, links to the same lexia as #1: »Summer of '76 covered in baby oil, a bit of bikini.«

5. the arrow links to: »Hot hand on my stomach, it's easy to bridge this distance. Days like this I still smell apple trees.«

In the first, the effects of the tickling heat are dissipated by the smooth baby oil, and the image which then supersedes it of the cotton swabs over the eyes soaked by blood instead of oil. The narrator in this lexia also seems more knowing, suggesting that she may be several years older than the five-year old of the previous lexia. The second link from »my sky« seems unrelated, other than focusing (as usual) on the memory of a young girl. Similarly, the third link jumps us from the field to an unrelated memory of woodland and fireflies. The fourth link, from the lower graphic, calls up the same lexia as #1; no logical or poetic rationale seems evident to justify this. The only link that clearly demonstrates continuity is, oddly, the last, which is offered by the arrow at the foot of the screen: this, »Hot hand on my stomach,« seems to be a memory in the present, evoked by bodily experience similar to the one we saw in the first foregrounded passage, and it reminds us of the first lexia in the chain in which the narrator fell out of an apple tree. This lexia, on the face of it, would have been more effective if linked to the opening phrase <http://www.yorku.ca/caitlin/waves/summer_of_76.htm>; in its first few words it even echoes the /t/ alliteration. It should also be noticed that the implications of the third foregrounded passage, »clouds turning to fierce animals in the distance«, are not taken up in any of the four lexias (unless the false echo of »it's easy to bridge this distance« is to count).

Thus the links from this lexia, except in one case, cut across the affective responses of the reader elicited by its foregrounding. How, then, are these links to be justified? It is evident that a reader must set aside the affective implications that have been evoked in order to respond fully to the next lexia. We can then retrospectively attempt to construct a relationship between the present and the previous lexia. In the case of my sky, for example, the young child's appropriation of the sky seems an instance of omnipotence of thought,[41] and is reflected in the games and stories of the grandmother and the child described in the remainder of this lexia. I can similarly construct relationships between the other lexias and the phrases that link to them, but this is a rational not an affective exercise. The power of the foregrounded passages is that their affective resonance is anticipatory, preparing us to experience relationships that we have not seen before. This is the immersive experience, quite different from the interactive role that we play (or that is forced upon us) when we reason out such relationships retrospectively.

In the light of this analysis of the links provided in one lexia, we are now in a position to evaluate the effectiveness of this portion of the hyperfiction in relation to the different types of feeling experienced in narrative. The four types, it will be recalled, are evaluative, narrative, aesthetic, and self-modifying. First, the evaluative feelings of satisfaction, pleasure, or frustration, will depend in part on readers' specific expectations as they embark on a given text, thus experienced readers of hyperfiction will be prepared for the formal leaps and disjunctions that the linking structure provides. The satisfactions this affords, however, are likely to be interactive rather than immersive; readers will take pleasure in puzzling out a relationship between successive lexias, and inferring retrospectively what themes (such as childhood omnipotence of thought) are addressed.

Second, narrative feelings are likely to be short-lived. Linked sequences of lexias provide only a brief focus on a given character or setting. This affords little time for the reader's feelings to become committed to any particular stance towards the narrative – such as the point of view of the five-year old girl or a resonance with the apples trees, fields, and cows of her farm. Fisher partly overcomes this limitation by offering several longer lexia: the narrative of Mr. Anderson, for example, provides a rich and coherent narrative extending over a number of years, and this has the benefit of providing a context for otherwise less explicable short narrative passages in other lexia. But in general narrative feelings are frustrated by the disconnection that often obtains between one lexia and the next in terms of characters, settings, and other narrative constituents.

Third, aesthetic feelings are undoubtedly aroused by Fisher's lyrical writing, as we have suggested in the case of the short lexia with three foregrounded features. But the feelings are in most cases balked by the jump to a subsequent lexia that fails to develop their implications. The risk here, then, is that readers will cease to invest their feelings in the text being read, since the result of reading one lexia after another is a kaleidoscope of aesthetic feelings rather than an emerging and meaningful pattern. As Ryan suggests, commenting on her reading of Joyce's Twelve Blue,[42] »The effect is that of an amnesic mind that desperately tries to grasp some chains of association but cannot hold on to them long enough to recapture a coherent picture of the past.«[43]

Fourth, self-modifying feelings in the hyperfiction context are thus, necessarily, unlikely. The inability of the reader to sustain a particular focus, to experience a modification of feelings over a series of lexia, suggests that any transformation in understanding beyond the superficial is unlikely to occur. Fisher herself has told us that »writing for a cyber universe has a whole new kind of grammar«.[44] But, as I have tried to show, the expectations we have of narrative and the implications of the different types of feelings called up in us during reading, suggest that a new grammar confronts an array of inherent and deeply influential psychological processes. Hyperfiction is misconceived to the extent that it ignores or confounds these processes. Avant garde pronouncements of the kind I cited from Eskelinen omit considering in what ways the processes of reading are not computational, and that reading is not a function like operating a machine.

Whether hyperfiction has a future outside the small domain of enthusiasts and academic followers seems quite uncertain. Fisher's work contains much writing that is suggestive, lyrical, and evocative; yet it is, I would suggest, only partly successful, because the structure in which it is embedded often dissipates its power to affect us and impose its own perspectives on us. As a reader I remain outside its small narrative circles, witnesses to fragments of experience and apparently random shifts in focus. Yet Fisher's work can also show the way, I believe, to a more effective use of the medium. In the short lexia I examined, where the arrow link provided access to an aesthetically related lexia, we jump a gap not only in the hypertext machinery but also in narrative time, yet in a way that satisfies the feelings that have been evoked. This is by no means the only example available in Fisher's hyperfiction where a linked lexia is both unexpected in some respect, yet aesthetically satisfying. My argument in this paper, then, resolves onto this final point: in order to understand hyperfiction and what it may promise, writers and critics should be attentive to the empirical literature on reading. The processes of reading have a structure and a resilience that hyperfiction does not put into abeyance, and that will confound the effectiveness of hyperfiction as readily as they will support it. Hyperfiction can very likely be both immersive and interactive, as Douglas and Harnadon have imagined, but this depends on hyperfiction writers knowing better what reading processes underlie the experience of being immersed and interactive.