Inhabiting the virtual city
Judith S. Donath
Designing the virtual environment is a challenging problem. In its newness and abstraction, it presents a nearly blank canvas to the designer. There is neither an inherent form nor much history and shared culture on which to base a design. Its features are unprecedented: millions of people worldwide are able to communicate with each other and share information, transcending the accustomed limitations of time and space. Its infrastructure is in constant flux: these people are communicating through channels that were non-existent five years ago and which are likely to be superseded in the next five.The metaphor of the city is a very useful tool for thinking about the virtual world. The virtual world is abstract and unknown; the city is physical and, though very complex, better known. Here I will look at some of the key parallels between real cities and virtual spaces and between city architects and on-line system designers.
A metaphor is a cognitive tool, a way of structuring thought. In particular, metaphor is useful (and arguably essential) for thinking about abstractions [2]. The cultural and symbolic meaning of ``city'' helps shape our conception of what this new world should be like; using the metaphor transfers the meaning of ``city'' to the abstract virtual environment ( Leiss et al. 1990; Eco 1984). And, while the net is ``fundamentally and profoundly anti-spatial'' (Mitchell 1995), a physical metaphor makes the abstract comprehensible: our constant experience of the spatial world makes such metaphors both powerful and pervasive (Lakoff and Johnson 1980).
There are parallels both between real-world cities and virtual environments as well as between real-world architect and virtual system designers. Both real-world and virtual cities are (or should be) vibrant gathering places of people, centers of commerce and entertainment. The designers of both urban spaces and of virtual system are creating environments that influence the interactions that occur within them and the culture that grows around them. But there are also important differences. In the real world, the designer creates a solid infrastructure of buildings, streets, etc., around which people and objects move. In the virtual world, the designer is working with very different materials and with a much more encompassing task. The whole communication structure and the means by which human beings are represented must be designed; it is as if an architect was responsible not only for making a building, but also the organs of sight by which the inhabitants viewed it.
2.1.1 Parallels: the real and the virtual city
Many points of comparison can be drawn between the physical and the virtual city. In City of Bits (Mitchell 1995), William Mitchell surveys wide range of them, examining the city metaphor in roles ranging from the aesthetic and physical to the economic and political. His analysis highlights the key points of comparison in a series of analogies: ``Muscles / Actuators'', ``Facade / Interface'', ``Territory / Topology'', ``Enclosure / Encryption'', etc.My focus is on the cultural and social parallels: on how the structure of the city affects the life within it and how cultural vocabularies and social institutions evolve in response.
2.1.1.1 The landscape of strangers
One of the strongest parallels between the city and the virtual world is their immense populations. Modern cities are filled with strangers, huge crowds of people one passes on the street but does not know. This is very different from life in a village. It is different even from life in the classical cities - the polis whose agora and civil life are so often cited in discussions of electronic community [3]. This continuous exposure to unknown persons, of diverse and often unfamiliar backgrounds and habits, has been cited by many urban theorists as one of the most salient characteristics of city life ( Canetti 1984; Milgram 1977; Sennett 1974; Simmel 1971; Wirth 1938).The virtual world is similarly populated by millions of people, most of whom are strangers to each other. Examining how this population pattern has shaped real world city culture can help us understand the forces that are shaping the development of on-line cultures. Furthermore, the analogy can point to designs that may help the virtual city avoid some of the real city's problems; it can also suggest directions that may be especially fruitful to explore.
The immense population of the city means that urban social ties are relatively weak. Many of one's daily encounters with others, such as making a purchase at a register, sitting beside someone on a bus, are faceted and limited ( Wirth 1938; Milgram 1977). Surrounded by strangers, the inhabitant of the city is much more anonymous than the small town dweller. This can be liberating - one has more social freedom in the city - but also alienating. Anonymity also diminishes social constraints on behavior: people may be ruder (or worse) among strangers whom they are unlikely to see again than they would be if in the company of acquaintances (Milgram 1977).
Ties in the virtual world are also weak. Here too, many of one's daily encounters with others show only a narrow facet of their lives and personalities, occurring as they do in specialized and limited venues (e.g. motorcyclist newsgroups, Perl programming forums, new parent chat spaces, etc.). As in the real world, the resulting anonymity is both liberating and alienating ( Donath n.d.; Sproull and Kiesler 1991). Anonymous on-line discussion spaces allow people to candidly discuss personal or controversial matters - one can, for example, seek information about a medical problem without making one's condition publicly known or discuss unpopular political beliefs without reprisal ( Froomkin 1995; May 1994). Yet anonymity is also cited as one of the reasons why on-line discussions so often degenerate into angry exchanges. Anonymity hides the offending writer from retribution. Furthermore, in an anonymous setting there are few reminders of the mores of social interaction, which are often cued by the identity of the other (e.g. one speaks - or avoids speaking - in certain ways in front of children or older people or one's boss or clergy, etc.).
Yet alienation need not be the inevitable result of immense population. Many urban theorists cite internal structures, such as close-knit neighborhoods, as constructs that provide an intermediate environment between the complete anonymity of the city at large and the deep relationships of close friends and family ( Jacobs 1992; Milgram 1977). For the designer of virtual environments, the analyses and solutions posed by their real-world counterparts can provide useful insights. Virtual neighborhoods are not a new concept ( Rheingold 1993; Mitchell 1995), but there is much still to be gathered from the metaphor in terms of understanding what the salient features of ``neighborhood'' are and what makes some neighborhoods particularly successful. For instance, both Jacobs (Jacobs 1992) and Whyte (Whyte 1988) emphasize the importance of mixed-use occupancy to strengthen neighborhoods; the combination of work, recreation and residential space, they say, bring a heterogeneous population and results in fuller utilization of the existing resources, especially over time, avoiding the emptiness of the business district at night or the suburban bedroom community by day. The concept of mixed-use can be carried to the virtual realm, inspiring the creation of venues that are accessed by varied groups. An example is the notion of ``surrounding neighborhoods'' on the web (see the discussion in Chapter 4 about multiple links to Portraits in Cyberspace and the wide range of opinions found in the on-line discussions there.)
The characteristics of the city - its density, size, heterogeneity - have elicited cultural responses that are distinctly urban. These responses show the adaptations people have made in order to best benefit from urban conditions. An especially interesting cultural development, from the point of view of the virtual designer, is fashion. The rise of fashion is an urban phenomenon. It has been attributed, in part, to the need to establish one's identity within a population of strangers, while also maintaining an essential privacy ( Ewen 1988; Sennett 1974). Describing how immigrants to the city became acculturated, Ewen wrote:
In such a broad milieu of strangers, style was a dramatic necessity. One was repeatedly made aware of self as other, of one's commodity status within a vast social marketplace, and style provided its user with a powerful medium of encounter and exchange... [S]tyle allowed one to put up a front, to protect one's inner self... True moderns, they were learning to internalize the dictum of Bishop Berkeley, that ``to be is to be perceived''. ( Ewen 1988)The inhabitants of today's virtual environments are similarly faced with competing desires: to be known yet not be exposed. One can see today's Web home page as the cultural equivalent of clothing, a personal display created for public viewing. As I will discuss in Chapter 3, elements of the home page exhibit the social features of fashion: their meaning changes over time as their use shifts to different social groups and these changes may be attributable to conflicting goals of differentiation and imitation (McCracken 1988).
2.1.1.2 The legible city
The design of the city affects not only how well its inhabitants can find their way around, but how well they can make sense of its social and symbolic complexities ( Lynch 1960; Milgram 1977). Kevin Lynch, writing about importance of a ``legible'' environment, said:
Obviously a clear image enables one to move about easily and quickly: to find a friend's house or a policeman or a button store. But an ordered environment can do more than this; it may serve as a broad frame of reference, an organizer of activity or belief or knowledge... Like any good framework, such a structure gives the individual a possibility of choice and a starting-point for the acquisition of further information. A clear image of the surroundings is thus a useful basis for individual growth. ( Lynch 1960)The inhabitants of the virtual city have similar needs. They need to be able to get from one place to another, to know what is available and who are their fellow inhabitants. The abstract nature of the on-line world makes a legible environment especially important: a poorly conceived interface is far more impenetrable than the most twisted alleyways.Both Lynch and Milgram asked inhabitants to draw maps of their city; their goal was to understand how the structure of the city was perceived. Comparisons of these maps showed much individual variation in the features portrayed, based on personal interests and experiences; they also showed group patterns, attributable to differences in class and other affiliations; and they showed architectural and spatial elements that were common to all. An important conclusion from their studies is that legibility in an environment is not at all the same as legibility in, say, a chart, which should have simplicity and clean, grid-like structures. Instead, it is based on memorable structures and symbolic spaces, on the existence of areas with distinctive visual styles. The overly planned space, with its clearly laid paths, may be in effect the least legible because it does not contain the richness of detail that helps one to remember a space. This is an important idea in the design of virtual spaces, where the temptation to create geometrically striking, but indistinguishable spaces is strong.
2.1.2 The architect and the system designer
The city metaphor extends also to the parallels between their respective builders: the urban architect and the designer of virtual spaces. The metaphor of system designer as city planner and architect is the basis of the methodology of this thesis.Design is not a quantifiable science. Though the usability of certain features can be measured (much as ergonomists measure stair-treads for safety and human factors specialists measure menu designs for speed), much design assessment is subjective. One studies things that have been built, to see how they were used and how people changed and adapted them (Brand 1994); from this, one develops an approach to the design of new spaces.
The architect must understand the technology of buildings, the properties of materials and the forces that make buildings stay up (or fall down); the architect must also understand how the buildings and street plans and lighting designs will affect the people who inhabit them.
On one side there is the engineering side of building: a matter of calculating loads and stresses, of making joints watertight and roofs rainproof, of setting down foundations so solidly that the building that stands on them will not crack or sink. But on the other side there is the whole sphere of expression, the attempt to use the constructional forms in such a way as to convey the meaning of the building to the spectator and user, and enable him, with a fuller response on his own side, to participate in its functions - feeling more courtly when he enters a palace... more businesslike and efficient when he enters an office.... and more citizenlike, more cooperative and responsible, more proudly conscious of the community he serves when he goes about his city and participates in its many-sided life. Architecture, in the sense that I here present it to you, is the permanent setting of a culture against which its social drama can be played out with the fullest help to the actors. ( Mumford 1952)Like the architect, the designer of the virtual space must also understand the technology, knowing what is feasible to build today or next year, and what problems, often seemingly simple ones, are still very far from a solution. And, like the architect, the designer's primary role is to create a social environment.
Architecture may not determine human behavior, but... bad design can numb the human spirit and good design can have powerful, positive influences on human being. Of the many values designers seek to build into their designs perhaps none is more important than fostering community and human interaction. ( LeGates and Stout 1996)There are, however, important differences between the role of the architect and of the designer of virtual spaces. In the real world, the built environment is only a small part of the environment as a whole. In the virtual world, the built environment is everything: it is a wholly mediated and synthetic world. The participants in a virtual world are wholly dependent upon the tools supplied by the creators of the space to communicate. The choices made by the designer of a virtual environment thus greatly influence and constrain the means of communication and the ways in which a cultural vocabulary develops.The architect of a virtual space shapes the community in a more profound way than does his or her real-world counterpart. People eat, sleep, and work in buildings; the buildings affect how happily they do these things. But the buildings do not completely control their perception of the world. In the electronic domain, the design of the environment is everything. Whether or not you know that other people are present or privy to a conversation, whether you can connect an on-line identity to a real-world person, whether you have only a faint notion of the personalities of those around you or a vibrant and detailed impression - this is all determined by the design of the environment.
2.1.3 Designing the new
The city metaphor has resonance and depth, for there are many parallels between the city and the virtual world, and the city metaphor provides many striking images with which to think about building on-line environments. It is thus not surprising that the metaphor is in common usage: Digital Cities and Electropolises and Downtown Cyberspaces abound. Yet, too often, the city metaphor is taken literally and the resulting interfaces do not provide the functions of a city; they merely replicate the facade.
Apple's eworld takes the city metaphor quite literally. ``Attention Virtual Shoppers... Now Entering Cyber City!'' is the welcoming banner on the web-page of
alltelga.net
, home thus far of Beckler's Carpet Outlet, Club Atlanta Travel and Kinard Realty, and where a cartoon of the still empty Main Street awaits more tenants. On many sites the city takes the shape of a graphical map, where various functions have been placed in appropriate, clickable buildings: email in the post office; shopping sites in the store; customer service complaints in the courthouse. Here the metaphor is used, but only superficially, the pictures simply substituting for a category label.On other sites, the metaphor draws from rather questionable features of the subject. In GeoCities, ``our Homesteaders set up residency in one of our twenty-four themed communities, based on the content of their home page.'' Here, the creators are replicating the feel and function of carefully monitored, gated neighborhoods and regulated facades, what Sorkin called ``a generic urbanism inflected only by appliqué.'' (Sorkin 1992b)
The electronic city is endlessly mutable. Its structure can grow and change at rates inconceivable in a world of brick and stone; its appearance may be highly subjective, individually tailored to the taste of every visitor. The city metaphor works when it is applied functionally, i.e. when we think about the role of public spaces in social life and the use of landmarks in building collective memories.
As technology advances, it becomes increasingly possible to recreate the experience of ``being there'' while separated by thousands of miles. In this thesis, my emphasis is on understandable, yet novel forms of communication and representation - on interfaces that go ``beyond being there'' (Hollan and Stornetta 1992). While it may be possible soon to have full video/audio in a synthetic space - or to recreate Rome in a 3D walk-thru - my emphasis is with what can be done that is new, that goes beyond the literal re-creation of the physical world.
The fundamental difference between inhabiting the real city and its virtual counterpart is the lack of a body. I will next look at several basic social functions of the body, such as individual recognition, gestural communication, and corporeal discipline. Their absence in the virtual world is a fundamental design issue: the key feature of many designs is either to re-create these functions or to explore new modes of interaction made possible by their absence.Incidents such as ``The Rape in Cyberspace'' and phenomena such as newsgroup invasion and escalated on-line flaming (Donath n.d.) have made evident the difficulty of controlling virtual beings. Social scientists are beginning to examine the involved in maintaining social order without recourse to the physical self: Kollack and Smith, for example, examine the on-line establishment of boundaries and the enforcement of community regulations in the context of the classic `tragedy of the commons' dilemma [6] (Kollock and Smith 1995). A common theme that seems to be emerging from many discussions of this problem is the importance of an established identity: a virtual self whose loss - though in no way equivalent to the body's loss of freedom or life - would be a serious cost to its real-world progenitor ( Donath n.d.; Kollock and Smith 1995; May 1994; Reid 1994). On-line, identity becomes the body.Understanding the social function of these physically-based qualities makes it possible to evaluate the impact, both positive and negative, of their absence in the on-line world. In some cases, their absence may be beneficial, making new forms of modes of communication possible; in others, where the absence of a quality is detrimental, it may be possible to reconstruct its function - possibly in a very different form - on-line.
2.2.1 Expression
Much real-world communication is nonverbal, consisting of physical actions such as gestures, facial expressions, vocal tone, etc. Many social signals, especially, are communicated this way: subtle (or not so subtle) expressions of approval, status recognition, comprehension, etc. Unlike the written word, which is deliberately produced and passes through much conscious filtering, nonverbal expressions are often made subconsciously, often revealing a great deal about their maker's ideas and opinions.
Expression has many social roles. It is an integral part of speech, conveying cues about turn-taking and attention ( Cassell et al 1994); it communicates emotion, both real and feigned ( Goffman 1959; Hauser 1996); it is the fundamental channel for social communication (Landau 1989), signalling agreement, doubt, sympathy, etc.
This sort of expressive gesture is missing from today's text-based on-line environments. A number of researchers have studied the social effects of this lack ( Sproull and Kiesler 1991), as well as some innovative attempts to reintroduce social cues ( Reid 1991). The remedying of this situation is a very active research area, both in the creation of facial and gestural interfaces (e.g. Waters and Terzopoulos 1992, Cassell et al 1994) and in the perception of gestural input (e.g. Thórisson 1995).
A growing number of studies of faces and facial expression in computer environments provide very useful insights into how these reactions are transferred to the computer environment. (Sproull et al. n.d.) looked at viewer's responses to computer interfaces with simulated, human-like faces. They found that people responded to a facial interface in a more socialized way: they attributed personality to the face and presented themselves more positively in interactions with a facial rather than text display. This is consistent with (Nass et al. 1994)'s findings that people very easily and ``incurably'' apply social norms and rules to their interactions with computers.
Facial expression affects the meaning of words.
Eisner's illustration of how variations in facial expressions transform the meaning of the words they accompany demonstrates both the care which needs to be taken in implementing expressive faces in a computer interface and the richness that even (or especially?) a cartoon rendering can bring ( Eisner 1990).These experiments also point out the importance of evaluating all of the effects of a face in the interface. A face, even a simple cartoon face, will convey numerous cues, about its emotional state or attentiveness, etc., even if none were meant. If the goal is to have an interface that can, say, convey a set of emotional states, or turn-taking cues, a facial image may be the solution, but it will also affect the character of the interaction in other, complex ways.
Our perception of facial expressions is both innately hard-wired and culturally learned. We are highly receptive to facial expression. There is evidence that babies several hours old recognized faces: they respond to face-like configurations, but not to random ones ( Hauser 1996). Once we perceive an arrangement to be a face, we ascribe an emotional state to it: ``any configuration which we can interpret as a face, however badly drawn, will ipso facto have such an expression and individuality.'' (Gombrich 1972) Furthermore, there is growing evidence that what emotion we will ascribe to it is also to some degree innate. Cross-cultural studies of emotional expression have shown strong correlations across disparate cultures, both in field studies ( Ekman 1973) and in physiological experiments ( Hauser 1996). These fundamental parameters of expression perception delineate the boundaries of how faces can be used in the interface: face-like configurations will almost invariably be perceived as faces, and there an existing gamut of basic and universal expressions.
Emoticons reflect real world expressions.
And, like physical gestures, they vary between cultures. Japanese emoticons (which are viewed right side up, rather than sideways as is common elsewhere) include a girl's smile with a dot for a mouth, since it is impolite for women to show their teeth:(^.^)and a "banzai" smile with arms upraised in a Japanese gesture.\(^_^)/Among the most commonly used emoticons in Japan are two that have no real equivalent in the Western version a cold sweat:(^^;)and one that says I'm sorry:(_o_)- New York Times, August 12, 1996On the other hand, when and how a particular expression is deployed is culture dependent, as are a great many other communicative gestures, such as bowing in Japan, or shaking the head horizontally for no (as opposed to the vertical nod for the negative which is used in North Africa). Like learning a language, learning when an expression is appropriate and how to use culturally dependent gestures is part of the process of gaining communicative competence (Saville-Troike 1982). Learning expressions and gestures is done through imitation; infants as young as an hour old have been observed imitating facial expressions. And imitating expressions can itself create the portrayed emotion. Experimental subjects, when asked to produce a smile or a frown of anger, showed significant correlated physical responses, such as raised heart rate and skin conductance in the case of anger (Hauser 1996). Thus the face plays a role in empathy (Landau 1989): we see someone express an emotion and respond by imitating the expression which in turn causes us to experience some amount of their emotion.
The presence of a expressive face on the screen is not a passive image, but a highly and often subtly interactive one. A vocabulary of computer-mediated expressions cannot be developed purely by design; rather, it need to evolve within the communicative context. In the real world, learning expression is done in a highly interactive environment: one produces an expression, it is perceived and the perceiver reacts, providing feedback to the producer. Reproducing something of this chain of action and reaction is an important part of developing a virtual expressive vocabulary; however, the sparseness of existing communication channels may prove to be quite problematic.
As will be discussed in greater detail in Chapter 6, there are a great many technical and interface problems still to be solved before a truly expressive face can be part of the a computer mediated discussion. The face is not the only way to portray emotion: a line can be evocative (Arnheim 1974); a bouncing ball can narrate dramatic emotional ranges (Disney). It is far from clear that re-creating the form of real world expression, such as interfaces with faces that smile or frown, is the best or only way to achieve seamless communication of expression on-line. Here, we have introduced some of what is communicated, expanding the possibilities of how to communicate it.
2.2.2 Presence
A key part of what makes a successful public social space is the presence of others (Whyte 1988). In the space-less world of cyberspace, presence is abstracted from its usual meaning of being in a shared physical vicinity. Yet it is clear that people do feel some sense of the presence of others on-line, both in real-time conversations and simply by seeing who is currently logged in.Researchers working with ubiquitous video conferencing systems (see, for instance Isaacs and Tang 1994, Dourish and Bly 1992, Beshers and Feiner 1993) have observed that, even though a variety of conferencing and other functions are available, people used the systems primarily to gain a sense of who else is around. Bly et al record that participants in the Media Spaces project, who had full access to a wide range of teleconferencing functions, often chose to use the system to keep open a window showing the comings and goings of people passing through a central area. ``Although seemingly the most invisible, the use of the media space for peripheral awareness was perhaps its most powerful use.'' (Bly et al. 1993)
Presence, or at least some aspects of it, can be transmitted via media ranging from the most simple and abstract to the most seemingly realistic. It seems to transcend knowledge of real-world physical distances: studies have found that not only are people much more expressive when they are in the company of others than when alone, they are as expressive when they imagine themselves to be with others as when the others are actually present (Hauser 1996). Thus the common phenomenon of people on the telephone gesticulating and making faces, which are of course invisible to the person at the other end.
Seeing the signs of other people's actions is one way of conveying a sense of their presence. Seeing the patterns made by people's activities is fascinating. We are social creatures and like to feel that we are not alone; we are curious and like to know what other people are doing. An interesting example of presence conveyed by patterns can be seen on the Web. One of the search engines has a ``voyeur'' site that puts up, every 20 seconds, an assortment of the currently processed searches. First you see:
new york times · architecture · hogarth print · nude celeberties actresses · playboy · pressedienst · japan spanking · woman having birth pictures · fire proof plants · october · conceptual framework · chat adult · fat acceptance · parasites · college chat · larsen judith · digital and express and provider · streamworks · photomultiplier microstrip porouse · bone thugsAnd a minute later:
denver broncos · free stuff · cindy crawford nude picture · intestinal parasites dogs · shannon tweed · snowcone syrup · erotic free pictures · porsche · nude sex · barcelona · wedding souvenirs · the weather channel · small bikinis · hot tubs · amp schematics · blonde jokes · australian soccer profiles okodak · volkswagen fuel system · novi sadBesides confirming that, whether or not there is a lot of sex on the net, there are certainly a lot of people looking for it, this feature provides one of the strongest impressions of presence on the Web [4]. The ongoing appearance of new searches provides the impression that many others are simultaneously at that site. The topics provide a hint of the individuals behind each search. And the return of some entries, slightly varied (``parasites'' above became ``intestinal parasites dogs'') shows the searcher struggling with results that are too extensive or too small.Change is essential to the perception of presence. If we believe that some immediate change in our environment occurred through the actions of another, we have a sense of their presence, even if not physically co-located. This especially so if the change is interactive, responding to some action of ours. In the virtual world, this interaction is found in the real-time conversation spaces, such as MUDs and chat-rooms, where one's typed comments are immediately transmitted to the screens of all other participants. While the content of chat-room conversations may seem inane, they do successfully communicate a sense of being in a live, inhabited space. Such conversations are more about conveying presence than they are about the text of the words.
There are some aspects of physical presence that cannot be conveyed through media. A noteworthy example is the crowd - physical presence in the extreme ( Canetti 1984; Milgram 1977). Here, the physical power of the group, both as an anonymous force capable of immense and destructive feats and as a force upon the individual in the crowd, is a salient feature that is absent in the virtual world. Is the crowd wholly outside the virtual experience? Is there a design that would make palpable the sensation that one was indeed on-line in the company of millions of other people?
The experience of being on-line is in many ways a solitary one. One sits alone, facing a screen. Yet, connected to that screen is an immense population - and also smaller, more comprehensible groups, one's friends, co-workers, etc. The question here is how to provide a real sense of their presence.
2.2.3 Control
The Panopticon This penitentiary system was originally described by Jeremy Bentham as a circular architecture, designed for continuous and invisible observation of the inhabitants/ inmates from a central point.
Image from ( CFUS 1996).The body is central to social control and discipline: it is hard to catch a disembodied being. Without the body incarceration is impossible and less extreme aspects of social control are difficult. The ease of creating alternate personas in a virtual world raises important questions about the nature of virtual boundaries.
The role of the body in social control is addressed by a number of Foucault's writings, particularly Discipline and Punish, his survey of the progression to increasingly internalized forms of discipline ( Foucault 1979). The book begins with the gruesome torture and dismemberment of an 18th century regicide (punishment centered entirely on the body); traces the birth and development of prisons (control through bodily constraint and, increasingly, through the reformation of the prisoner); describes the panopticon (discipline conducted through surveillance and knowledge); and concludes with what he calls ``the carceral network'' (in which discipline is diffused throughout society).
The extreme point of penal justice under the Ancien Régime was the infinite segmentation of the body of the regicide... The ideal point of penalty today would be an indefinite discipline: an interrogation without end, an investigation that would be extended without limit to a meticulous and ever more analytical observation, a judgement that would at the same time be the constitution of a file that was never closed... ( Foucault 1979)Foucault's history points to a diminishing role of the body as focus for social control, and the rise of greater control through ubiquitous surveillance.The specter of all-seeing, unseeable surveillance provides one common dystopian view of cyberspace: a regimented and monitored society, identity clearly marked, one's history fully scannable - an electronic carceral city (Mitchell 1995). Yet the body's role remains central. If the virtual identity does not connect to the physical self, the meaningfulness of surveillance on an ephemeral population is unclear [5]. In what has become known as the ``The Rape in Cyberspace case'', a player on a MUD committed a series of anti-social acts against other players. Eventually he was ``toaded'' - the virtual character, Mr. Bungle, ceased to exist.
Yet the continued dependence on death as the ultimate keeper of the peace suggests that this new MOO order may not be built on the most solid of foundations. For if life on LambdaMOO began to acquire more coherence in the wake of the toading, death retained all the fuzziness of pre-Bungle days. This truth was rather dramatically borne out, not too many days after Bungle departed, by the arrival of a strange new character named Dr. Jest. There was a forceful eccentricity to the newcomer's manner, but the oddest thing about his style was its striking yet unnameable familiarity. And when he developed the annoying habit of stuffing fellow players into a jar containing a tiny simulacrum of a certain deceased rapist, the source of this familiarity became obvious:Mr. Bungle had risen from the grave. ( Dibbell 1993)
- Sir Thomas Browne. Religio Medici 2.2.4 Recognition
It is the common wonder of all men, how among so many million of faces there should be none alike.
Recognition is not limited to the face. We recognize handwriting, gait and tone of voice; law enforcement agencies find a fingerprint or retinal pattern to be more reliable forms of recognition than the variable and disguiseable face. Still, the face is central to the idea of recognition: ``it is in the face that we recognize each other, and identify ourselves'' (Synnott 1993). In this section on recognition I will focus on facial recognition, to provide a foundation for thinking about the representation of individual identity on-line.
Face recognition has a biological basis. Patients with certain brain injuries are unable to recognize the identity of familiar faces, though they recognize the stimuli as faces and recent work in neurological imaging has shown that specific areas of the brain are active during face (and expression) recognition. (Hauser 1996). Yet the full cognitive process of facial recognition is still not fully understood. The many attempts to reproduce it computationally have made its complexity apparent. Today, most work in computational face recognition depends on constrained datasets: reproducing the human ability to recognized individual faces in widely varying conditions is still beyond computer vision.
Face recognition is a social, as well as biological, task. It involves perceiving what distinguishes a particular person from all others. Perceptual psychologists and other researchers have proposed that we have an internal representation of a normal, or prototype, face and that we process and store other faces in terms of their deviation from this norm ( Brennan 1982; Gombrich 1972; Hochberg 1978). ``It is not really the perception of likeness for which we are originally programmed, but the noticing of unlikeness, the departure from the norm which stands out and sticks in the mind.'' (Gombrich 1972) This theory accounts for several phenomena, including why people have difficulty differentiating and recognizing members of unfamiliar groups; it suggests that the normative face (or faces - different categories may be represented by distinctive norms) is created in a continuous process of categorization and that members of groups for which one has little familiarity will all register as ``outsiders'', rather than as specifically differentiated individuals ( Landau 1989; Gombrich 1972).
Caricature works from a similar principle. Caricatures are drawings in which the subject's most characteristic features are exaggerated ( Gombrich 1972; Brennan 1982). They thus require the existence of a norm, the baseline from which to diverge. Brennan ( Brennan 1982) created an early computer ``Caricature Generator''. A ``normal'' face was created and the subject compared to it, feature by feature. The subject was then redrawn with the differences between it and the norm exaggerated. An interesting feature about this work was that the ``norm'' could be easily changed, making explicit the subjectivity of the caricature, and, by extension, of recognition. Caricatures are often said (though it has not been decisively proven) to be more quickly recognizable in some circumstance than photographs or line drawings ( Hochberg 1972; Brennan 1982); their emphasis of the way a subject differs from the norm provides the perceptual basis for why an ``incorrect'' rendering could be more recognizable than an accurate one.
Recognition is a product of experience, both of people in general (the prototype forming experience) and of one's experience and impression of the individual.
What people experience as likeness throws light on their perceptual categories. Clearly we do not all have the same impression of a person's aria or characteristic face. We do see them differently according to the categories with which we scan our fellow creatures. (
Gombrich 1972)
The face has so far been seen little on-line, though this is rapidly changing with the advent of the Web. Still, it is thus far a very limited face - the still, photographic portrait - and not found in a context where it is used for recognition. As graphics becomes ubiquitous, will the face become the primary means of on-line recognition? Or will other patterns become recognizable as immediate markers of identity? 2.2.5 Remarks
These topics that were discussed here will reappear throughout the thesis, in designs that recreate their function or institute an alternative. Questions about expression and recognition will arise in the discussion of The Illustrated Conversation, when I examine issues concerning the use of graphical representations of people in a conversational interface. The question of recognition will also return, in more abstract form, in the discussions about on-line self-portraiture. How control is maintained in an on-line environment is part of the discussion of public space and providing a sense of presence in a virtual space is a motif in several projects, including Visual Who, WebTalk, and A Day in
the Life of Cyberspace.