Synopsis

Refereed articles

Information articles

Notes on contributors

Print friendly version

Digital memories, analogues of affect

Robert Payne

In Michael Ignatieff?s 1992 novel Scar Tissue, the narrator, a philosophy professor, attempts to come to terms with his mother?s increasing debilitation of mental capacity and memory. While taking the "unrepeatable opportunity" of her condition to "observe the relation between selfhood and memory" (52-3) ? ultimately his own self as much as his mother?s ? the narrator recalls watching a documentary about the neuropsychological experiments conducted by Wilder Penfield in the 1940s in which Penfield claimed to discover the localisation of memory function in the human brain. Penfield chronicled the surprising responses given by conscious neurosurgery patients to electrical stimulation of particular areas of their brain ? notably, the vivid evocation of buried memories, re-experienced as if by the flick of a switch. He writes:

It was evident at once that these were not dreams. They were electrical activations of the sequential record of consciousness, a record that had been laid down during the patient?s earlier experience. The patient "re-lived" all that he had been aware of in that earlier period of time as in a moving-picture "flashback." (Penfield, 1975: 21)

Despite the sample?s low proportion of successful memory recovery and the unverifiable nature of the memories themselves, the image of a memory switch specifically located in the brain is nonetheless compelling. As the novel?s narrator continues:

Localisation of function ? implied that the memory images stored in different parts of the brain were activated every time recollection occurred. This theory, which Penfield?s experiments appeared to confirm, was surely the origin of my belief that my mother?s memories were still intact, like a butterfly collection left behind in the attic of an abandoned house. (Ignatieff, 1992: 52)

The narrator?s metaphor of fragile beauty is apt for many people?s understanding of their own long-term memories, safely stored but forgotten. For others, such memories may be purposely discarded in the attic, necessary in order to move forward into a new psychological house. Either way, the image accords with the belief held by Penfield that the human memory constitutes a kind of permanent but only temporarily accessible archive ? what he calls a "laid down" record. Increasingly disillusioned by medical discourse, the narrator of Scar Tissue comes to realise that his mother?s memory failure is not just a symptom of neurological illness, of the biological failure of particular localised memory zones in her brain, but a crisis of subjectivity:

My mistake had been to suppose that a memory image could subsist apart from an image of the self, that memories could persist apart from the act of speaking or thinking about them from a given standpoint. It was this junction between past and present that she was losing. She was wondering who the ?I? was in her own sentences. (53)

Unanchored by a central subject position ("a given standpoint"), memory floats free. The narrator observes his mother to be drifting in an open narrative space, without linearity or grammar, moving towards a state of complete linguistic and discursive deconstruction. In these observations, biology precedes and determines construction of the mother?s discursive self, in that the degradation of her biology is seen to cause a larger loss of self through memory failure. But if subjectivity need not be restricted by biology, does one?s discursive self-construction guard against memory loss and subjectivity crisis by generating the self specifically through memory gain?

This article will analyse the role played by memory in producing subjectivity, but where the identity produced is itself an effect of the construction or reconstruction of particular memories. By engaging with the interface between human and computer memory, the focus of my examination will be on how the construction of memory in an online context hopes to flesh out the spectral, post-biological self, inscribed as a digital assemblage of context-specific contingencies. In a space where self-construction can leave human biology redundant rather than failing, how does the digital self remember who it is? And where the individually bounded has given way to the multiply and prosthetically connected, whose self produces and is produced by whose memories? What is the "I" in one?s own sentences?

*

Random Access Memory is an Internet-based memory project, a self-described "experiment in collective recollection" (www.randomaccessmemory.org). Visitors to the website may randomly browse through the database of memory narratives contributed by other users and also contribute their own. Each memory lodged must be categorised by user name and the year to which the memory refers. Browsing may therefore involve searching by user name or year, or by any keywords which may appear in databased memories, or following a user-determined itinerary by means of user name or year hyperlinks. On each log-in, a different and apparently random memory is featured on the site?s homepage, and this may act as a point of departure for browsing either through the hyperlinked year or author, or via independently chosen criteria (perhaps a particular word or idea from the narrative). Even though navigation through archived memories is mostly indexical, the project is also a kind of hypertext for two main reasons. First, memory authors and readers undertake an interactive, collaborative, non-linear narrative process, as I will outline further. And second, this collaborative textual space allows readers to jump from narrative to narrative according to their own choice of mental image or keyword, even if that image or keyword is not actually hyperlinked. Memories lodged may take any written form, with the only author information available being the pseudonym provided. Currently, the website claims over twenty thousand memories in its archive, contributed by over six thousand "rememberers".

In computer terminology, "random access memory" (RAM) refers to a computer?s short-term memory, which stores recent and frequently used information. Because the information required for a programme to operate is specific and controlled in type and location, RAM does not strictly involve random access but direct access, only giving the impression of random availability. Borrowing this common-usage computer term for its title, the web project Random Access Memory aims to provide "a backup archive" for easy storage and retrieval of memories. Spontaneous contribution is encouraged: " If you suddenly remember the sled you played with as a child, or the sweet, eggy cake your mother fed you dipped in tea, sign in and record the memory at that moment." It aims to preserve what we know to be fleeting and slippery, as might a dream journal kept beside one?s bed. The primary function seems most to favour and assist individual "rememberers" with a more reliable storage facility than human memory, even if the latter must of course be active in the first instance. For those with poor memories, the site can provide a prosthetic memory to enhance or replace one?s own failing neurological functioning. In this specific sense, it merely updates the technology of the journal, the photo album or the dictaphone by appropriating newer informatic systems as memory augmentation. But in the place of non-verbal representations of memory, literacy is privileged. And despite its multimedia basis and further potential, the site hearkens to the dominance of print culture.

Adopting a function not conventionally related to computer RAM, but more closely reminiscent of human thought processing and earlier prostheses like the TV remote control, browsing the Random Access Memory archive (which the site implies to be a secondary function) suggests a greater possibility of randomness, even if search terms are limited to specific criteria. However, the browser also experiences a randomness effect in that the choice of any one search criterion (the year "1977", the word "cake") produces a random-appearing list of heterogeneous memory narratives, which might only be connected, at least evidently, by the database?s code of indexing and hyperlinks.

In a recent essay, José van Dijck (2004) makes a strong argument for the indistinction of memory and its mediation, where the various media in which memories may be stored or through which they may be activated should not be seen as external to the processes of "memory work" (Annette Kuhn, cited in van Dijck, 2004: 263). Van Dijck also refers to the aide-mémoire function of a technological medium (photograph, archive, journal) in terms of "artificial prosthesis", suggesting that such media have been "paradoxically defined as invaluable yet insidious tools for memory" (272). Prosthesis proves to be an apposite concept, drawing on the extension of human function that is invested in online practices and ontologies, including the associative workings of hypertext believed to be analogous to modes of human cognition. As Alison Landsberg (2000: 191) points out, because mass media "fundamentally alter our notion of what counts as experience," perhaps memory "might always have been prosthetic". By co-implicating the functioning of human memory and technological media, both writers register an implicit understanding that prosthesis must be recognised as transforming the constitutive boundaries of the body to which it is added and with which it then merges: a cyborgian conception of the body as incorporating prosthetic otherness, and thus dissolving organic/artificial distinctions.

However, not to risk a monologic or colonising dynamic of incorporating otherness, the merging of body and prosthesis might equally be figured inversely as "going online", an expression which in this context best refers to the ontology of digital network connectivity. Van Dijck?s intertwining of memory and mediating technologies here allows a more specific inference of prosthetic engagement. Where electronic and digital memory tools are increasing in volume and potential, she writes, memory "may become less a process of recalling than a topological skill, the ability to locate and identify pieces of culture that identify the place of the self in relation to others" (272). In other words, perhaps not exclusively but especially in the context of digital technology, memory production reflects the subject?s desire to constitute itself in terms of its social, cultural and virtual connectedness, and in this sense is a means of localisation. Where the narrator of Scar Tissue notices his mother?s inability to locate herself grammatically, narratively and ontologically as a subject, van Dijck draws attention to memory as cultural contextualisation. In this sense, localisation of memory function is not so much a fact of neurobiology as it is a process generated by cultural negotiation ? particularly with various mediating devices, but also for them.

The function of each mediating form thereby exemplifies a more general principle which, I will argue, equally governs the "memory work" undertaken in Random Access Memory: that produced memories add up to a version of the self which is local to the discursive context of their production. Eviatar Zerubavel (2003: 5) describes "norms of remembrance" which dictate both the form and content of memory production within communities. Individuals acquire a sense of what and how it is appropriate to remember through what he calls "mnemonic socialization", whether "implicitly encoded" or as "explicit normative prescriptions". Participation in that community relies on the subject?s internalisation of norms of collective memory ? not just what is remembered (a religious holiday, for example) but how (through family gatherings and rituals). Here community and its structures mediate memory production. In a similar vein, but aside from memory?s sociological function, van Dijck argues that the construction and reconstruction of memories is not just mediated by common cultural formats but dictated by habitual knowledge of them, even if one has no clear projection of what the memory product may be used for:

In spite of the indeterminacy of a memory object?s final reification [?] existing cultural forms nevertheless frame or even generate their production. A range of cultural forms, such as diaries, personal photographs and so on, prefigure people?s choices of what and how to capture memories. (265)

Adapting Zerubavel?s vocabulary, we might call this "mnemonic enculturation", and I am arguing that engagement with Random Access Memory, as a cultural space and as a prosthetic memory device, entails exactly this kind of productivity.

*

The first time I logged on to Random Access Memory, I had already decided to contribute a particular memory to the site?s archive. Having heard of the website and been intrigued by the concept, I chose and mentally constructed the narrative of a significant memory, one still recently deposited in and easily retrievable from my short-term memory.

In the hospital room where my dad was dying. He got out of bed to take a piss but thought, in his morphine-addled state, that the hand basin was a toilet. With his hands against the wall for support, he leaned over the basin thinking he needed to piss and set off the automatic tap sensor every time he leaned in, but he had no idea what was going on. We didn?t know whether to laugh or not. I helped him with his underwear as he fumbled to get his penis out and I wondered if he felt any humiliation.

What I will call the contagious efficacy of Random Access Memory seemed already to have taken hold, in no small part due to the vigour of social and electronic networks through which news of websites and other popular cultural objects is continuously transmitted. After signing in, uploading my pre-planned memory and browsing through the memories contributed by other users, the contagious effect was to intensify. Soon after this first time, I found myself making three more unplanned contributions to the archive. These subsequent memories, all retrieved from much deeper in my long-term memory store, seemed to have been spontaneously re-evoked in response to now-unknown elements of other users? memories that I had been reading on the site. In other words, my memories re-appeared to me like mental hypertexts, as if programmed to open from now-erased hyperlinks; as if stimulated by a localised neurosurgical probe.

My parents had come back from a holiday in Fiji and brought me a souvenir doll of a policeman in a traditional skirt. I took it to school for show and tell and in the playground at lunchtime a boy in my class called Christian Tyson grabbed it from me and ran around mocking me for having a male doll in a skirt. I think he pulled the skirt off to show it naked. He must have dropped it on the ground and it cracked. I was really upset.

My after-impression of this second interaction with Random Access Memory was one of having come into transmissible contact with a contagion: that something embedded in the memory narratives I had browsed communicated with my own inner narratives of self. But particularly it was that they seemed to communicate interaffectively; as I later realised, all three of these unplanned memories reconstructed experiences of childhood shame, which I had never before categorised in these terms. The affective constitution of personal memories is well recognised, as are the affective triggers that particular memories may contain. Less expectedly, the narrated memories of others may also resonate affectively for the listener or reader and activate his or her own recall. But to what degree can one claim subjective and narrative agency over the production of these memories and related affects?

Anna Gibbs (2001) analyses the phenomenon of affective contagion by drawing on the theories of psychologist Silvan Tomkins, according to whom affects are "neuro-physiological events" (cited in Gibbs, 2001: 3) which exist in complex interactions with each other and with thought to contribute to the emotional formation of the subject. As Gibbs continues to summarise, Tomkins?s theory identifies a finite number of such affects,

each of which acts to amplify the gradient and intensity of a neural firing, producing a positive feedback loop in which more of the same affect will be evoked in both the person experiencing the affect and in the observer (a phenomenon known as ?affective resonance?) (3).

It is exactly this relationship of "positive feedback" between affective responses and resonances and the concept of "neural firing" which is pertinent to my own perceptions of the affects which seemed to resonate from memory narratives archived on Random Access Memory. In particular, the digital image of "neural firing" is, I will go on to outline more fully, at the very least reminiscent of the programmed function of hypertext, which itself enacts the principles of feedback. This correspondence is anything but coincidental, taking into account Eve Kosofsky Sedgwick and Adam Frank?s (2003: 105) reading of Tomkins which identifies his contribution to what they label "the cybernetic fold": "a rich intellectual ecology" during the 1940s-60s in which the development of systems theory came to project the potential for productive overlaps between information systems and biological function.

One ultimate outcome of this intellectual moment was hypertext. As George Landow (1997) outlines, the conception of hypertext in its present form builds on the realisation that information technology should supplement the inevitable failings of brain function, particularly memory, rather than working against the mind in artificial systems of indexing. This is the direction signposted by engineer Vannevar Bush, whose 1945 essay "As We May Think" included his now commonly cited belief that the human mind best retains and retrieves information not by strict, mechanical classification but by "association of thoughts, in accordance with some intricate web of trails carried by cells of the brain" (cited in Landow, 1997: 5-6). Bush?s design for a "memex" device applied these principles in terms of current analogue technology, which in turn influenced IT pioneer Ted Nelson to conceive a digital system of information storage and retrieval based on associative linkage ? in short, hypertext. By applying Bush?s ideas to emerging digital technology, Nelson?s innovations with hypertext bridged the conceptual divide between the quantitative efficiency of digital computation and the qualitative multiplicity of the human mind.

Echoing a fundamental aspect of Bush?s proto-hypertext project, Sedgwick and Frank?s (2003: 100) discussion of Tomkins underlines "the highly complex, highly explicit layering of biological and machine models in his understanding of the human being". But while they go on to describe Tomkins? seemingly parallel "habit of layering digital (on/off) with analog (graduated and/or multiply differentiated) representational models", they are adamant to disrupt the "tacit homology" between digital and machine, analogue and biological (101). In this sense, the "neural firing" that leads to affective responses is a less reductive description than the term might first read, and is less successfully characterised mechanically as a digital switch. Just as the digital switching of a hyperlink leads to a multiple yet finite number of other texts in various representational forms, Sedgwick and Frank emphasise that despite Tomkins?s apparent modelling of affect activation as digital (on/off), his theory nonetheless

ramifies toward a many-valued (and in that sense analogic) understanding of affect: if the on/off of "neural firing" is qualitatively undifferentiated, the on/off of affect activation is qualitatively highly differentiated ? among no fewer than seven affects. (103)

That my own layering of body/machine interaction with Random Access Memory appeared to activate an affect-memory switch is therefore a perception that requires deeper investigation. The systems theory model of a positive feedback loop that describes the dynamics of both affect in Tomkins?s terms and contagions more generally is also applicable to the textuality of hypertext, rather than just its computational function, such that the productive layering of hypertextual participation roles offers a useful analogy to the concept of resonance. Of particular correspondence is Gibbs?s (2001: 5) acknowledgment that affective resonance might also include situations where "one is drawn to the affect already felt", and that one might "seek out occasions" to express a particular affective response.

Playing catch and kiss in the playground at school. I caught a boy in my class and asked him if he wanted a boy kiss or a girl kiss. I remember the strange unidentifiable sensation of pre-sexual desire as I considered kissing another boy, maybe not aware that this was forbidden. He said he wanted a girl kiss. He must have been aware.

As has been widely discussed, Landow?s work on hypertext builds strongly on a base of post-structuralist literary theory, describing hypertext as a kind of idealised actualisation of textual deconstruction. In particular for Landow, hypertext disrupts a conventional model of authorship by which the text makes available a corresponding autonomous and unitary author-subject for the reader to uncover. Instead, he identifies a collaborative process where "author" and "reader" roles overlap productively, strongly echoing the (ironically now-canonical) theories of authorship produced by Roland Barthes and Michel Foucault. As an effect of this collaboration, hypertext knows no boundaries, no predetermined linearity or sequence, and therefore no distinct authorial control over the creation of textual meaning. In Barthes?s terms, the interaction is producerly: writers of individual hypertext contributions or "lexia" (cited Landow, 1997: 3) are equally readers of other lexia, who in turn become writers of further lexia, and so on. In the terminology of systems theory, hypertext operates in a positive feedback loop, where input progressively causes an exponential change in the system?s output, and in turn its input, and so on. Textual influence and autonomy can scarcely be distinguished or attributed in what becomes a network of co-production.

Mapping Random Access Memory against these theoretical principles, what I have provisionally called the primary and secondary modes of use of the site ? memory archiving and memory browsing ? exist in an equally dialogic relationship. On the surface, the first of these makes use of a relatively uncomplicated and functional service: reliable storage space for one?s memories, which may remain as private (that is, as pseudonymous) as the user wishes. That the archive is also necessarily open to public view suggests that a merely depository use of the site ? to leave one?s memories digitally "backed up" ? is if not impossible to ensure, then certainly impossible to distinguish. Archiving one?s memories takes place with the understanding, at the very least, that they are likely to be read by others. By extension, it would be reasonable to assume that the desire for these memories to be read by others is of equal or greater intention. In other words, memory archiving and memory browsing may be mutually generating uses of the site. Rephrased in literary terms, authorship and readership can no longer be held as distinct roles in the construction of textual meaning.

For browsers or readers of Random Access Memory, these dynamics challenge the idea of a centred self that might temptingly be imagined to reside behind or within the postings attributed to any one user name or author. This is particularly so when the writing of memories may be strongly subject to the textual and contextual influences of other users? narratives. As my own experience suggests, precisely how the project?s archive expands is through a consciously or unconsciously dialogic production of intertextuality. By the same token, this version of hypertextual co-productivity must also challenge the autonomy of users or authors in their own perception. Whether archiving their believed authentic memories as a genuine attempt at narrative self-certainty or more explicitly fictionalising memories as a means of acknowledged self-construction, all authors write themselves into their memory archive through and as fiction. Even without the direct channels of influence that Random Access Memory?s hypertextual space allows, any memory loses authenticity with any attempt to mediate it ? to explain, narrate or otherwise represent it. As van Dijck indicates, any memory is always already discursively mediated: arguably at the point of its mental cognition, and certainly by the point of its narrative reconstruction. It cannot therefore be read as entirely pure in its authorship.

In its literalised interface between human and machine capacities for memory storage and retrieval, the Random Access Memory project might appear to apply the theory that human memory operates analogously to hypertext: a specific stimulus (word, song, smell, image) marks the launching point for memory recall, just as if the stimulus were a hyperlink in an electronic document, directing the reader to an associated image or text. The randomness effect of Freudian word association offers a fruitful prototype here. A hyperlink enacts a programmed shift in visual and textual focus and may only appear to be unexpectedly or randomly associative because its processes are occluded from view by the surface simulations of the contemporary computer?s graphic user interface (Turkle, 1995). As psychoanalysis has taught us, human memory recall is perhaps no less dependent on code, only less calculably.

Hyperlinks thereby constitute finitely controlled simulations of what Vannevar Bush took to be the mind?s associative leaps. Michelle Kendrick (2001: 233) argues that this presumed homology is often cited to naturalise a claim for hypertext?s immediacy to human cognition, and by extension that hypertext "truly reveals the subject". In this way the "cognitive model", as Kendrick labels it, is an argument for hypertext as brain prosthesis. At its extreme, this is an image of telerobotic thought control and performative agency which serves to erase the mediating technology (direct upload from mind to hypertext). But at the same time, Kendrick points out, the common description of hypertext as a rarefied, collaborative space which best serves the mind?s cognitive freedom celebrates the mediating technology for its special capacities (238ff).

Equally, on a larger scale, what is marketed as the Internet?s unique provision of users? unrestricted freedom to "explore" the uncharted hypertextual linkages of a worldwide network is a programmed illusion. A digital map already exists for every possible itinerary. As Mark Nunes (1995) writes,

the connections between nodes precede the attempt to explore this terrain, meaning that every "journey" in cyberspace is a repetition and a retracing of steps. Internet has no provision for "undiscovered country", only the simulation of such, like a planned treasure hunt.

While the multiplicity of links available in hypertexts also suggests a magnification of choice for the reader ? a wide range of associated visual or narrative directions to choose to follow ? it is important to remember that in a basic sense this effect is simply the sum of several binary options. Each hyperlink offers precisely two possibilities: to open another text or not. Certainly, the newly opened text may itself offer various visual or narrative possibilities, including further linkages to choose. But at a fundamental level, hypertext?s renown for textual openness and freedom is dependent on the deceptive metonymy of digital symbols which offer no more than a yes/no, on/off, open/closed choice.

Christopher J. Keep (1999: 174-5) questions the politics of choice within hypertext?s rehearsed ethos of the reader?s "unrestricted ?freedom?", which he calls "one of the dominant myths of electronic textuality". As Keep and Kendrick both note, Landow and others have identified freedom in the ability of the hypertext reader to move through the text in an individually chosen and non-sequential fashion, maintaining that this movement implies the reader simultaneously "writes" the text ? their text. In response, Keep argues that this seeming investiture of textual authority is underlaid by "a consumerist ideology that allows for the reappearance of the self-knowing subject of laissez-faire capitalism in the world of ?de-centered? hypertextuality." The textual choices that the reader feels "free" to make, he continues, are inaccurately figured as those of the supermarket shopper, afforded multiple options to choose between at leisure. But "just as surely as the supermarket forces its desires on the shopper," Keep writes, the restriction to the hypertext reader?s complete autonomy and coherence lies in the "organizational structure" which "necessarily mediates between the reader and his or her desires" (175). Choice, openness and freedom are then idealised, rhetorical qualities justifying technological innovation rather than being actual effects of it.

Accordingly, Kendrick (2001: 242) critiques what she sees as frequent attempts to theorise hypertext in narrowly formalist terms which tend to overlook content as a whole, particularly in theories where the mere fact of "technology enabling ?choice?" counts as a celebration of neo-authorial autonomy. She summarises:

Hypertext, reductio ad absurdum, is a technological mechanism: click here, it is hypertext; no links, it is not. Hypertext is thus equated with a keystroke choice? I click and drag; therefore I am (244).

So while on the one hand, Kendrick argues, hypertext?s digital technology is seen to allow the dispersal and disembodiment of the author function, on the other hand this same foregrounding of technology restores agency to the reader through his or her active textual choices. The myth of the autonomous textual subject is only rhetorically erased by hypertext?s deconstruction of authorship and reinscribed in the role of unencumbered readership. In response, Keep ( 1999: 165-6) writes:

Contrary to those who claim that a hypertext allows the reader to become its "author", the reader has no a priori subjectivity that he or she brings to the screen. Subjectivity, in this sense, is formed in relation to our bodily sense of occupying a specific space and time: it is constituted and reconstituted from moment to moment as the reader responds to the various challenges and opportunities of reading an electronic narrative.

I am suggesting that the online personae inscribed by Random Access Memory users? narratives (including my own) come into being in exactly this way, "authored" in conjunction with users? parallel participation as readers, which is itself a multi-levelled practice of negotiating choice within the constraints of "organizational structure" that appear to enable choice more completely, if abstractly. As I have suggested, the "organizational structure" of Random Access Memory differs from that of other hypertexts (particularly hypertext fictions) in that the possibilities of actual linkage from one memory text to another are limited to the archive?s search criteria, namely user name, year and keyword. Nonetheless, innumerable possibilities for mental and affective linkage exist in addition to programmed archive navigation, and it is particularly in the space between programmed links that ?memory work? might be most productive.

What initially appeared as a discrepancy between my perception of Random Access Memory as activating a digital affect-memory switch and a critical understanding of mnemonic enculturation ? the former a representation of external agency, the latter a mode of iterative identification ? can now be reconciled. In short, these representations are not in fact distinct. Prosthetic memory is a form of mnemonic enculturation, of mediating subjective memory production through cultural engagement. To isolate who or what is responsible for the "origin" of memory production ? to attribute "authorship" ? is an impossible task. Rather, we have Keep?s notion of "our bodily sense of occupying a specific space and time" as a contingent process of becoming, very much akin to what van Dijck calls the "topological skill" of self-formation. What counts as the bodily in the digital media context includes, as I have argued, the digital medium itself as prosthesis, where the performance of self within and for the specific space and time of the digital medium recontours the conceptual limits of the body. Kendrick is right to refocus attention on the elided content of hypertext. The subject inscribed by the overlapping roles of hypertextual production is not the bland result of production merely being possible and enacted (that is, the simple, digital fact of technological engagement) but a continually evolving interactant in the narrative and affective material of lexia; self-inscription as narrative participation. And rather than effecting a wholesale merging of human body and digital medium such that the hypertext subject is conceived merely as a digital effect, the mutual interaction between body and prosthesis I am describing ? their reciprocal layering in a positive feedback loop ? is best understood as a relationship of virtuality and the subject-in-becoming as a virtual body.

In his discussion of affect in digital artworks, Mark Hansen (2004) also privileges the virtual as a successful philosophical model for the interactivity and interaffectivity of such work. Drawing on the ideas of Timothy Murray, Hansen writes that:

by materializing the virtual elements of the object or work in a form that opens them to interaction with the user-participant, new media artworks detach these elements from the work as a static object [?] and resituate them within the interactional process through which the user-participant becomes dynamically coupled with the work. (145)

The specific artworks that promote Hansen?s enquiry are what he calls "digital facial images", that is, digitally generated close-up images of a human face, with which the viewer-participant?s encounter ? "and specifically the affective correlate it generates" in the viewer ? "comes to function as the medium for the interface between the domain of digital information and the embodied human" (128). While otherwise independent of Tomkins?s theory of affect, Hansen?s description of face as affective interface corresponds with Tomkins? specific localisation of affective register on the face. As Gibbs (2001: 3) writes, this is because the face "communicates our affective states to others, but also, importantly to ourselves, via feedback." Gibbs goes on to discuss how the device of televisual close-up may function to transmit more efficiently and widely and to amplify such affective states, through the "vector" of a "talking head" (4). For both Hansen and Gibbs, the face localises the relationship of interaffectivity between "talking head" and viewer which, as Hansen states above and Gibbs implies, is therefore a relationship of virtuality; what Margaret Morse (1998: 17) describes in terms of "fictions of presence".

In the absence of the face as focus for interlocution, Random Access Memory is not so much a site for the register of actual affects, which might then be communicated perceptually to the browsing subject as resonance, but the materialisation of the possibility of interaffectivity. Where hypertext as a digital construction actualises the virtual feedback loop that is intertextuality, through the vectors of hyperlinks, the hypertext interface opens a space in which "fictions of presence" with other subjects might be more concretely literalised. Random Access Memory?s version of this space specifically allows its "rememberers" to "seek out occasions" for the self-affirming affective resonance of memory co-production.

*

Pissing my pants in my kindergarten class. The teacher had left the room and I really needed to piss but I knew that no one was allowed to leave the room without asking, so I guess I took her at her word.

Hoping to experiment with the dynamics of "collective recollection", the producers of Random Access Memory have exploited one of the most common and useful functions of online technologies ? to most efficiently create a collective that draws from geographic if not demographic diversity. But as with many technological innovations which become integrated into everyday cultural practice, utility is often partly or entirely superseded by the unanticipated potential of secondary functions. An online archive is clearly not the only device for the intertextual or interaffective co-reproduction of personal memories ? sitting around in a group and reminiscing will do that effectively enough. And the space formed by members of that group is no less one of virtual relations within a fiction of shared presence, indeed a fiction of a shared past as presence. But what Random Access Memory does add to the formula, in addition to fictions of telepresence between participants, is to provide through a literal interface a set of metonymised and layered interfaces between body and prosthesis, digital and analogue, memory and affect, intertext and hypertext, self and other.

The personal experiences of Random Access Memory that I have narrated took place some years ago at a time when I began to perceive an epidemic of academic shame, by which I mostly mean an epidemic of academic writing about shame. Should it have been any surprise, then, that the memories I lodged on the website, just as I began to locate myself topologically within the academic loop, all spoke of shame? What now seems clear is that the virtual co-productivity of hypertext offers an analogy to the academic enterprise, itself dynamised by what Sedgwick and Frank (2003: 95) call "cross-disciplinary transmission". Like academic writing on and of shame and other affects, Random Access Memory attempts to localise and legitimate through print culture what may otherwise float free from subjective autonomy. But both means of expression draw attention to the instability of "a given standpoint", in the words of the narrator of Scar Tissue, and instead invite us to participate in virtually becoming the "I" in our own sentences.

References

Dijck, J. van. (2004) "Mediated Memories: Personal Cultural Memory as Object of Cultural Analysis" in Continuum 18:2 pp. 261-277.

Gibbs, A. (2001) "Contagious Feelings: Pauline Hanson and the Epidemiology of Affect" in Australian Humanities Review 24 http://www.lib.latrobe.edu.au/AHR/archive/Issue-December-2001/gibbs.html

Hansen, M. (2004) New Philosophy for New Media, MIT Press: Cambridge, Mass.

Ignatieff, M. (1992) Scar Tissue, Vintage: London.

Keep, C. (1999) "The Disturbing Liveliness of Machines: Rethinking the Body in Hypertext Theory and Fiction" in M-L. Ryan (ed.), Cyberspace Textuality: Computer Technology and Literary Theory, Indiana University Press: Bloomington, pp. 164-181.

Kendrick, M. (2001) "Interactive Technology and the Remediation of the Subject of Writing" in Configurations 9, pp. 231-251.

Landow, G. (1997) Hypertext 2.0, Johns Hopkins University Press: Baltimore.

Landsberg, A. (2000) "Prosthetic Memory: Total Recall and Blade Runner" in D. Bell and B. Kennedy (eds.), The Cybercultures Reader, Routledge: London, pp.190-201.

Morse, M. (1998) Virtualities: Television, Media Art, and Cyberculture, Indiana University Press: Bloomington.

Nunes, M. (1995) "Baudrillard in Cyberspace: Internet, Virtuality, and Postmodernity". http://www.dc.peachnet.edu/~mnunes/jbnet.html

Penfield, W. (1975) The Mystery of the Mind: A Critical Study of Consciousness and the Human Brain, Princeton University Press: Princeton.

Sedgwick, E. (2003) Touching Feeling: Affect, Pedagogy, Performativity, Duke University Press: Durham

Turkle, S. (1995) Life on the Screen: Identity in the Age of the Internet, Simon and Schuster: New York.

Zerubavel, E. (2003) Time Maps: Collective Memory and the Social Shape of the Past, University of Chicago Press: Chicago.

Note

My thanks to Cristyn Davies for her insightful comments and suggestions on this article, and to Sara Knox for telling me about Random Access Memory in the first place.