JCMC 5 (1) September 1999
Collab-U CMC Play E-Commerce Symposium Net Law InfoSpaces Usenet
NetStudy VEs VOs O-Journ HigherEd Conversation
Impression Formation in Cyberspace: Online Expectations and Offline Experiences in Text-based Virtual Communities
Department of Anthropology
Table of Contents
- Categories, Prototypes, and Prototype Effects
- Prototype Theory
- Virtual Communities
- Textual and Paralinguistic Cues in Virtual Communities
- Prototypes in Impression Formation
- Images and Realities
- Messages, Models, and Meanings
- About the Author
AbstractHow do people in cyberspace picture one another? More specifically, how do individuals engaged in text-based computer-mediated communication (CMC), with its paucity of visual and auditory cues, form impressions of those with whom they interact? And how do expectations formed online compare with offline experiences? Researchers have begun to answer these questions, drawing primarily on theories of stereotyping. This paper uses prototype theory and related models to extend previous research and to account for discrepancies between online image and offline reality. It draws on interviews with individuals who first met others online and subsequently moved to face-to-face interaction; it also utilizes comparisons between text-based impressions formed online and photographs displayed on web pages.
IntroductionIt is not uncommon for people who meet in the text-based environments of cyberspace--asynchronous news groups and bulletin boards and synchronous chat rooms and virtual communities--to be mistaken, and sometimes wildly so, when they imagine one another's offline appearances. For example, in an article about online dating (A. Hamilton 1999), one man complains "It's draining when you realize how different people are from what they project online," and another story (J. Hamilton 1999) about the mainstreaming of online romances describes a pathway to disappointment: "The correspondents finally meet, but the chemistry crashes like a warped hard drive. Her extra five pounds is actually 50. His definition of a full head of hair proves to be a bit thin." The discrepancy between image and reality is also captured in cartoons. One depicts a sophisticated, thirty-something woman, sitting at a table for two in an upscale restaurant, saying "I loved your E-mail, but I thought you'd be older." Her dinner companion is a little boy (Weber 1998).
In face-to-face interaction, physical appearance, vocabulary, grammar, other linguistic markers (including tone and accent), and nonverbal cues ordinarily influence the ways in which people initially form impressions of one another (Ekman & Keltner, 1997; Goffman, 1959; Lea & Spears, 1995; Zebrowitz, 1996). In cyberspace, many of these indicators are absent or strongly attenuated. The paucity of these features raises two questions: in the absence of visual and auditory cues, how do individuals engaged in text-based computer-mediated communication (CMC) form impressions of those with whom they interact, and how do expectations formed online compare with offline experiences? Researchers have begun to address these questions (Lea & Spears, 1992, 1995; Walther, 1993, 1996, 1997). Yet studies of computer-mediated communication have paid scant attention to the specific assumptions, categories, and norms that constitute the cognitive models within which impressions are formed or to the fit between imagination and reality.
This paper applies prototype theory to identify types of cognitive models people use in forming impressions in cyberspace and to examine how different types of cognitive models are related to the fit between online image and offline reality. It focuses on the ways in which different models generate different interpretations of the same message. Although this point is implicit in theories of the dynamics of computer-mediated communication, it has not been developed in them. Walther (1996), for example, notes the ways in which senders of messages optimize self-presentation and the ways in which receivers idealize senders. From the sender's perspective, a significant feature of CMC is the opportunity to construct messages carefully, thereby enhancing the representation of self. From this viewpoint, the emphasis is on the meaning of the message to the sender. From the receiver's perspective, the circumstances of text-based CMC are conducive to an idealization of the sender. From this point of view, the emphasis is on the meaning of the message to its recipient. Most research on computer-mediated communication has focused on senders' messages; much less attention has been given to assessing the ways in which receivers interpret messages and especially to the differences in the ways in which messages are interpreted. Whereas social identity theory (Walther, 1996, 1997) illuminates the process of relationship development in CMC, including the formation of positive and negative impressions, prototype theory explains why these impressions are often at odds with offline experiences.
The paper proceeds as follows. First, I review theories of categorization and cognitive models, especially prototype theory or variants of it, that bear upon impression formation in general and on discrepancies between expectation and experience in particular. Next, I present a short account of the kind of virtual community in which fieldwork was conducted and the kinds of cues present in it on which people base their impressions of one another. Following that, I describe the procedures I followed in gathering the data I present here regarding impression formation. Then I describe the ways in which informants' online expectations are shaped by the cues they use and the cognitive models they employ in interpreting them. Next, I compare people's online expectations and offline experiences, concentrating on the fit between them. Finally, I suggest the implications of the data for further research on impression formation in computer-mediated communication.
Categories, Prototypes, and Prototype EffectsThere are different theories of categorization. In "classical" theories, categories have clear boundaries and membership in them is defined by necessary and sufficient conditions (cf. Cantor & Mischel, 1979; Lakoff, 1987a, 1987b; Medin and Wattenmaker, 1987; Rosch, 1978; Smith & Medin, 1981). In the classical view, category members share common attributes. Although there are categories that meet these criteria, there are others that do not, making classical theories incomplete. Rosch (1978) developed a different approach to understanding categories. She reasoned, as Lakoff noted (1987a, p. 7), that "if categories are defined only by properties that all members share, then no members should be better examples of the category than any other members" and "if categories are defined only by properties inherent in the members, then categories should be independent of the peculiarities of any beings doing the categorizing." However, Rosch (1978, p. 35) noted that "Most, if not all, categories do not have clear-cut boundaries." Rather, they are defined in terms of their clear cases, not in terms of boundaries or necessary and sufficient conditions of membership.
Rosch labeled the clearest cases of category membership "prototypes." Prototypes are "defined operationally by people's judgments of goodness of fit in the category" (Rosch, 1978, p. 36; cf. Cantor & Mischel, 1979, p. 29; Lakoff, 1987a, p. 7, 41; Lakoff, 1987b, p. 63; Neisser, 1987, p. 2). As Lakoff notes (1987a, p. 15), "subjects' ratings of how good an example of a category a member is judged to be…are effects." He describes these judgments about the differences among the category members as "prototype effects" (Lakoff, 1987a, pp. 40-41, 1987b, p. 63). The concept of prototype effects bears directly upon discrepancies between online expectations and offline experiences, not surprising since the aim of the early research on prototype effects was "showing asymmetries among category members and asymmetric structures within categories" (Lakoff, 1987a, p. 40). Prototype effects are especially important in person perception, not only because of the difficulty in identifying "a set of necessary and sufficient features shared by all members of any particular person category," but also because people employ many person categories, different people use different categories, and because, within person categories, relations of inclusion among categories and sub-categories are fuzzy, e.g., a salesman could be a "PR type" or a "comic joker type" (Cantor & Mischel, 1979, pp. 10-11; cf. Murphy & Lassaline, 1997, p. 117).
Prototype TheoryPrototype theory explains the development of prototype effects (Lakoff, 1987a, 1987b). Prototype effects result from different sources. One is a lack of fit between a classical category and its background assumptions--one meaning of the term "context" is such background assumptions (Jacobson, 1996) The concept of "bachelor" exemplifies this source of prototype effects (cf. Lakoff, 1987a, 1987b; Quinn & Holland, 1987). Ordinarily, the word refers to an unmarried adult male: such a view assumes a society in which men marry by a typical age. Yet there are circumstances in which those assumptions do not apply, and, therefore, men who do not clearly fit the category. Lakoff's examples include priests, men in long-term cohabiting couples, and homosexuals, among others (1987a, 1987b). Another source is the nature of the cognitive models (and related concepts such as "frames," "scripts," and "schema") within which categories are embedded, models that are viewed as "theories" of the world (Lakoff, 1987a). Such theories entail socially constructed and culturally varied assumptions and knowledge (cf. D'Andrade, 1995; Fiske & Taylor, 1991; Quinn & Holland, 1987; Strauss & Quinn, 1997).
Prototype effects are generated by different kinds of cognitive models. Lakoff examines several such models, one of which is a metonymic model. A metonymic model is one in which one member of a category is used to understand the membership as a whole (Lakoff 1987a, 1987b). Social stereotypes and exemplars are two types of metonymic models that are prevalent in impression formation in cyberspace.
A social stereotype is a model in which an individual or a subcategory is culturally recognized as representing the category as a whole. As Lakoff notes (1987b, pp. 76-77), "Social stereotypes are usually conscious and are often the subject of public discussion… [and] we often have names for stereotypes, for example, Uncle Tom, Jewish Princess, stud." Stereotyping has been the focus of much research on impression formation in computer-mediated communication (Lea & Spears, 1992, 1995; Walther, 1996, 1997), as well as in studies of offline behavior (cf. Brewer, 1988, 1996; Goffman 1959; Mackie et al., 1996; Stangor & Schaller, 1996) Although researchers have found that stereotypes may function to reduce information overload, they may also operate to augment an "information-impoverished environment" (Stangor & Schaller 1996, p. 21; cf. Mackie et al. 1996, pp. 44-45; Brewer 1996, p. 257), a condition that characterizes much text-based CMC.
In the case of person perception, exemplars (or "salient examples" as Lakoff terms them) are specific individuals whom people have encountered and who are taken as representative of others who are thought to be members of the same category (1987a, 1987b; Cantor & Mischel, 1979; Goffman, 1959; Medin & Wattenmaker, 1987; Stangor & Schaller, 1996) Exemplars give rise to prototype effects because different people have experiences with different individuals who serve as the reference points of a category and because, as new instances of a category are added to it, exemplars are subject to change. As Smith and Medin (1981, p. 10) put it, "frequent experience with a new exemplar…can alter one's concept….Similarly, if one person has extensive experience with a particular exemplar while another person does not, they may end up with very different concepts…."
Other kinds of metonymic models include typical examples and ideal types. A model of "typical examples" is based on knowledge about features common to a number of individuals that distinguish them as an identifiable class. Prototype effects arise when people generalize from typical to atypical cases (Lakoff, 1987a; cf. Lakoff, 1987b; Goffman, 1959). By contrast, ideal types are based on "abstract cases which may be neither typical or stereotypical" (Lakoff, 1987a, p. 87). As Lakoff notes, an ideal husband is a "good provider, faithful, strong, respected, attractive" while the stereotypical husband is "bumbling, dull, pot-bellied" (Lakoff, 1987b, p. 78), neither of which may reflect the typical husband. Ideal types are culturally informed, vary culturally, and often constitute standards in terms of which people evaluate one another. The use of such ideal types leads to prototype effects, since there is an "asymmetry between ideal and non-ideal cases" (Lakoff, 1987b, p. 78).
Although differing from metonymic models, a radial model is another source of prototype effects. It has "a central case and conventionalized variations on it which cannot be predicted by general rules" (Lakoff, 1987a, pp. 84, 91; Lakoff, 1987b. pp. 74, 82; cf. D'Andrade, 1995, p. 39). Lakoff (1987b) uses the concept of "mother" to illustrate a radial structure. The central case is that in which a cluster of models converge: mothers are females "who gave birth to the child, supplied her half of the child's genes, nurtured the child, is married to the child's father, is one generation older than the child, and is the child's legal guardian." The variant subcategories include stepmothers, adoptive mothers, birth mothers, foster mothers, surrogate mothers, genetic mothers, and unwed mothers. Each variant is related to a different cognitive model. Prototype effects occur when one model is taken (mistaken) for another.
Although cognitive models give rise to prototype effects, they also may generate images that are a better fit with realities. A model based on typical examples could lead to correct predictions about category membership, when knowledge from known typical cases is applied to cases that are unknown but not atypical. The principles that inform radial structures can also generate expectations that are congruent with experiences. Lakoff enumerates several such principles, including "chaining" and "motivation" (1987a, pp. 95-96). Chaining refers to ways in which subcategories are linked to one another in a network (cf. Strauss & Quinn, 1997), and motivation refers to the general principle (or set of principles), external to a radial structure's subcategories and their related cognitive models, that "makes sense" of their relationships (cf. D'Andrade, 1995, p. 39). In short, the principles of chaining and motivation entail theories of the world that account for coherence within categories.
Other analysts have contributed to the development of theory-based views of categorization that reinforce Lakoff's. Hampton (1997, p. 135), for example, states that a "theory view of concepts" holds that "people represent concepts through a deeper understanding of the causal connections between their observed characteristics," adding that a concept "is embedded in a wider set of interlocking theoretical structures " in which such characteristics are "linked to others with explanatory, causal, or goal-directed links." And Hahn and Chater (1997, pp. 49-50) contrast models based on similarity (e.g., prototypes and exemplars) with those that are "theory- or explanation-based views" which "focus on the relationships between concepts and our knowledge of, and theories about, the world." In their view (1997, p. 50), theory refers to a "body of knowledge that may include scientific principles, stereotypes and informal observations of past experiences….[in which] properties of objects are not independent and thus not independently assessed in categorization but are embedded within networks of inter-property relationships which organize and link them." They note (1997, p. 50) that "Lakoff's (1987) "idealized cognitive models" are another expression of the same idea."
Virtual CommunitiesI examine the role of cognitive models in impression formation in a type of text-based virtual community known as a MOO (or Mud Object Oriented). (For further discussion of MOOs and MUDs, see Curtis, 1997; Jacobson, 1996; Marvin, 1995; Reid, 1995). A MOO is a type of software program that permits multiple users (also called participants or players) to simultaneously access a shared database and to communicate and interact synchronously and asynchronously in a virtual environment. The environment is typically characterized by a spatial metaphor and an architectural motif. The database consists of rooms, entrances and exits, and other objects, and users can manipulate and extend it. Objects, including participants themselves, are representations of discrete things. Each object has associated properties that store the data that describe it and verbs that define and implement its behavior.
In addition to creating and programming other objects, participants communicate with one another in various ways. They "talk" by typing a command and a message that appears as text on their computer monitors. For example, if Tower types ‘say Hello’, he would see on his monitor You say, "Hello" and others would see on their monitors Tower says, "Hello." They can also "act" or emote by typing a command and a message that displays nonverbal actions. For example, if Tower types ‘emote jumps across the stream,’ he and others would see on their screens Tower jumps across the stream. In addition to synchronous communication, MOO participants can also interact with one another using asynchronous modes of messaging analogous to those found elsewhere in cyberspace. They can send and receive MOO-mail, the equivalent of e-mail; they can post messages to special interest groups, a system comparable to the newsgroups found on USENET; and they can, using MOO-mail, distribute messages to groups of people, replicating mailing lists found elsewhere on the Internet.
Textual and Paralinguistic Cues in Virtual CommunitiesIn the context of text-based virtual communities, the cues that shape impressions are participants' screen names, descriptions, and the discourse in which they engage. Curtis (1997, pp. 126-127) noted that names are commonly drawn "from or inspired by myth, fantasy, or other literature, common names from real life, names of concepts, animals, and everyday objects that have representative connotations." Others have noted the importance of "linguistic style" (Lea & Spears, 1992, p. 323) in computer-mediated communication, including "language intensity, verbal immediacy, and lexical diversity" (Walther, 1993, p. 386). In addition to the words people choose, paralinguistic cues also influence the ways in which participants see each other. These cues include "typographical marks and other textual features," including the use of capital and lowercase letters, ellipsis, exclamation marks, typing errors, and emoticons (Lea & Spears, 1992, p. 324), the last of which are exemplified by a variety of "smiley faces."
Any of these cues, singly or in combination, may be used to from impressions in cyberspace. For example, a participant who uses the pseudonym "Tower" may be thought to be tall, especially if described as "Steel girders and tinted glass skin. Sways in the wind, on occasion." A person who talks about "dons," "roundabouts," and "lorries" may be classified as English; someone who mentions spending days in a library may be construed as scholarly. Moreover, once placed in a particular category, other inferences may be drawn: a tall person may be thought to be thin; an English person may be thought to prefer warm beer; a scholar may be thought to wear glasses. Yet, different people may attach different attributes to each of these categories; one person's "Englishman" might be different from another's.
MethodsTo understand the ways in which participants in virtual communities envisioned one another’s offline characteristics-their looks and mannerisms in "real life"-on the basis of information available online, fieldwork was conducted in four different virtual communities. The MOOs varied in size, ranging in the average number of participants connected at any one time from a dozen or so to a couple of hundred. The number of MOOs selected was limited by the time available to do fieldwork. For reasons of confidentiality, I do not name the particular communities, although they are all "social" MOOs rather than "adventure" or game-playing ones (cf. Bartle, 1996 and Curtis, 1997 for discussions of differences among MOOs, MUDs, and other virtual communities). I assign pseudonyms to the participants interviewed and/or otherwise observed (as well as those they mentioned), except if they have given permission to employ the names they use online. In logs of interviews, I substitute R for the researcher's name. An advantage of assigning pseudonyms to informants, even if they are pseudonyms for the pseudonyms used for online characters, is that they disguise the digital identities of participants; a disadvantage is that they distract from the pseudonyms participants assign to themselves, names that often evoke particular images. In choosing fictitious names for informants, I have relied on a panel of observers to suggest names that are consonant with the imagery of those whose identities I disguise.
The data come from two groups of informants. The first group consists of MOO participants whom I interviewed online. To recruit informants, I posted the following note to mailing lists on each of the four MOOs that served as fieldwork sites.
For research on the ways in which MOOers (and other participants in CMC) form impressions of one another, I’m interested in talking with people about their experiences in meetings MOOers offline. I’m particularly interested in the "fit" (or lack thereof) between expectations based on MOO interactions and "rl" [real life or offline] experiences. Are people what you expected them to be? If not, why not? In what ways do their online and offline characteristics differ? If you are interested and willing to take time to talk, please moomail me.
Within two weeks of posting the message, fifteen participants responded. I conducted interviews ranging in length between an hour and two hours with these volunteers, logging the conversations (with their permission); in some instances, they provided additional information via moo-mail (a form of email internal to a virtual community). I interviewed them in the capacity of my online character, informing them that I was a social anthropologist, and explaining the nature and conduct of the research to them. I did not ask them to provide information that would enable me to identify them offline. I asked them to recall how they imagined people before they met them offline (and before seeing photos of them prior to their offline encounters, if they saw such photos). I also asked them to describe, on the basis of those offline meetings, the appearances and mannerisms of those they had imagined. These informants described 36 people they had met both online and offline.
For the purpose of this project, I took informants' offline meetings as a given. I did not inquire systematically into where or why these meetings occurred, although most mentioned something about the circumstances under which they had met, including social gatherings organized by groups of MOO participants. I did not systematically ask about the outcome of these offline meetings or about the nature of relationships formed, if any, although informants did provide some information on these issues. The focus of the interviews was on the fit between online expectation and offline experience, and that was the topic informants addressed.
The second group of informants consists of 23 students in a course on social relations in cyberspace who, observing and interacting with people in the four fieldwork sites, formed impressions of 27 MOOers. The people described were selected from a list of 35 MOOers who, I had determined through preliminary fieldwork, had photographs online (on various web pages). The students were not given the URLs of these photographs and were explicitly instructed not to search for or look at online photos of people they were going to observe. In addition to the list of MOOers with online photographs, students were given the following directions:
After the students recorded their impressions, they were shown online photographs of those they had described. All together, 38 informants, including the 15 MOOers I interviewed and the 23 students who observed MOOers, formed impressions of 63 individuals. Of the 27 MOO participants described by student-informants, 15 were described by more than one person, permitting comparisons of the ways in which different observers imagined the same individual. Thus, the data include (1) impressions formed by MOOers who had met other MOOers online and their accounts of subsequent offline meetings and (2) impressions formed by students via their online observations of MOOers and the online photographs of individuals described by the students. In addition, I interviewed several of the MOOers who were the foci of student observations, providing data that confirmed and/or supplemented the information evident in the online photographs. Describe how you imagine any three of the virtual characters [included in the list], including their physical features, their mannerisms, and any other personal attributes that come to mind, and analyze how you arrived at your views, focusing on the reasons underlying your interpretations. Base your impressions on their names, their descriptions, the ways in which they talk and/or act (with you and/or with others), and any other information available online to you, except photographs.
Prototypes in Impression FormationParticipants in virtual communities often base their images of players on stereotypes. In many instances, the stereotypes are conjured up by screen names. For example, one woman described the way she imagined three MOO participants.
Another informant, when imagining a man she met online, drew on a fairy tale hero. On the basis of the way he spoke and acted online, she had expected him to be handsome and thin; in fact, he was heavier than she had expected. As she put it, "I guess my stereotype of handsome is somewhat slimmer…At some level, I think I must have pictured him as Prince Charming." JoshSamBob's name sounds like a white southerner. He fits my idea of a stereotypical frat boy. He's just under 6 feet tall, medium build, blonde hair, blue eyes, his lips are a bit thin. He's particular about what he wears and is in general very neat. He tends to fidget a lot and would much rather be out playing sports than sitting in a lecture hall or library. He enjoys horseplay and being where the action is.
demon is short and has short dark brown hair and wears dark colored (but not usually black) comfortable clothing. I have no idea what a demon is supposed to look like, but they have an unsavory reputation. She doesn't describe her character any more thoroughly than that, and my mental image of the character is of a small, dark, occasionally humanoid shaped cloud. The small and dark carry over into my idea of her offline image and are reinforced by the lowercase spelling of her name.
timberwolf is a broad shouldered man who wears cotton flannel shirts with the sleeves rolled up and likes to spend a lot of time outdoors. He wears gold wire-frame glasses and his somewhat unruly brown hair is starting to go gray. He's from the Pacific Northwest so that explains the flannel shirt, pseudo-lumberjack outdoorsman part. The graying hair goes with his being a professor, as do the glasses, although the fact that they are gold wire-rim has more to do with his name and the fact that he doesn't capitalize it, a sort of understated elegance.
Participants also often use exemplars in imagining others. Several saw people they met online in terms of people they knew offline. One woman said:
A man wrote: Chloe says, "I pictured him to be Scottish…average height, thin and sandy haired"
R says, "what led you to imagine him those characteristics?"
Chloe says, "I knew he was Scottish because he had told me so"
R says, "How about height, weight, coloring?"
Chloe says, "I think I had him pictured as a composite identity- of Scottish people I knew, someone else with the same RL [real life or offline] name..."
R says, "How about his voice?
Chloe says, "I pictured it to sound more like my friend’s voice-she’s from
A woman, in describing what she thought a man she was going to meet offline would look like, based her expectation on men she had dated before: Whenever I think I know enough about a MOOer's mannerisms to form an image, I tend to map em into someone I already know IRL [in real life or offline]. Perhaps a mixture of several people that I know. When I first got on here, [one] MOOer's writing style - specifically, his debating style - WAS extremely similar to a coworker's, so much so that when I read the MOOer's words I imagined my coworker sitting in his chair, staring at his screen, typing and squinting and whatever. Still another informant commented on the way she imagined others' heights: "I imagined he would be short. I’ve not had much to do with tall men before. They always seem to be short. I also thought he would be clean shaven. I’ve always gone out with clean shaven men before too. He had described himself as tall, but I just couldn't picture it." She was consistent in her expectations, imagining women online to be like those she knew offline. Nightfall says, "He was taller than I thought he would be."
R says, "What led you to think he might have been shorter than he really is?"
Nightfall says, "I dunno, I guess I think of men being 5’10 on average."
R says, "Why do you think men will be 5’10"?"
Nightfall says, "My father and husband are about 5’10"
Nightfall says, "I expect men to be the same height as the significant males in my life"
Typical examples and ideal types also inform people's expectations of offline appearances. One American woman said she supposes people will be white: "I almost always assume someone is white until they tell me otherwise, and even then sometimes it's hard to "see" in my mind's eye." She also imagined people would be slender, an ideal characteristic in her view: "I think we all like to assume the people we like are attractive, which usually means thinner in this culture" R says, "How tall are you?"
Nightfall says, "5'2"
R says, "When you imagine female MOOers, do you see them as 5'2"?"
Nightfall says, "No they are about 5'5"
R says, "Why do you give them those extra inches?"
Nightfall says, "Probably most women I know are about that height."
Images and RealitiesFor most informants, offline experiences did not match online expectations. Talkativeness was one area in which discrepancies occurred. One woman described the difference between her expectation of a man she had encountered online and her offline meeting with him in this way:
Several informants remarked that people were "chattier" in cyberspace than in real life, a finding that is not surprising, since all actions in virtual communities involve one form or another of computer-mediated communication. One informant said: Mystery wasn't quite what I expected. He was much quieter irl [in real life] than I expected. He's very chatty online. He also doesn't move the way I imagined him to. He's more stolid, less fluid…stiffer. In contrast to offline interaction, where the meaning of "silence" may be interpreted with reference to non-verbal cues (e.g., facial expressions, posture, gestures, proximity),"response latency" (Lea & Spears, 1995, p. 218) in computer-mediated communication may give rise to ambiguity and alarm. The absence of a "verbal" response may be construed as agreement, disagreement, or indifference; it may also be taken as a sign of technical trouble (e.g., computer malfunction or "lag" due to network overload). Participants manage the uncertainty by typing; they signal their presence and their attention by "talking." the line-by-line nature of MOO communication smoothes out most minor lags in communication that can be a major part of the impression you form of somebody. A MOO friend whom I've known face-to-face for a number of years tends to be very soft-spoken and thoughtful in real life, but on MOO this is invisible; there's no volume control and, given typing speeds, multitasking and other factors, the pauses for thought are buried amidst the delays of the medium. On occasion I've been on conferencing systems (like UNIX "talk") that are character-by-character. There's definitely a different impression in this sort of communication; the lags in your typing are visible and apparent, and if you're a decently fast typist you have no opportunity to reflect on your words before sending them, no opportunity to see them on the screen and reconsider them.
On the other hand, some informants spoke of others' online conciseness, which contrasted with their offline expansiveness. For example, one man described his expectation and experience of meeting a MOOer in this way:
Several features of cyberspace may contribute to this sort of discrepancy. It takes about 4 to 5 times longer to exchange messages in computer-mediated communication than in face-to-face communication (Walther, 1996), and people, especially experienced participants in virtual communities (in contrast to "newbies"), attempt to sustain the flow of interaction by typing briefly and quickly taking turns in communicating. Another factor may be that people are "multitasking" (i.e., working in different computer applications and/or engaged in both online and offline activities) and manage diverse demands on their time and attention by responding briefly and at irregular intervals. Buford says, "Cardinal surprised me; he is softer and gentler than I expected."
R says, "What had led you to expect otherwise?"
Buford says, "He is terse online. Since I was asking mostly dumb questions [about how to do things in the virtual community], I felt I annoyed him or he made me feel as if I were annoying him, when it was just his online directness. He always answered, but sometimes quite briefly... like 'try help' [for a specific object or verb]."
Body imagery was another area in which participants reported discrepancies between online expectation and offline experience. For example, one informant said:
Of another woman, he said: I had no idea what to expect with Katya. From her descriptions I got the impression she would be overweight, kinda hackerish, but when we met, I found her very attractive. Normal sized, nice hair, not at all the stereotypical programmer. Another informant described his image of a woman with whom he had interacted online: She wasn't quite what I was expecting. Her descriptions have her with lots of long hair, so I had some expectations about that. She's not got long luxurious hair down to her feet. She's thinner, she's got a somewhat masculine build and mannerisms. I didn't expect her to have masculine mannerisms. I just didn't get that. The reality was not what he expected: Her words were huge. Her rapier wit was killing. Alice can slice and dice with ASCII text but do it in ways that are wonderfully funny. I was sure she was gigantic, physically. If her words were any indication of her size…she had to be 6 ft tall and/or weigh at least 300 lbs. My impression of her, based solely on what she wrote, was one of towering everything -- guts, strength, physical dimension. I was sure she was fearless. There certainly didn't seem to be any topic she was afraid to tackle or any disaster she wouldn't confront head on.
Another woman also pictured people larger online than they were offline. Her images of them were based in part on their volubility and in part on the absence of reality checks. That is, in the absence of disconfirming data, she imaginatively filled in the gaps with details they did not possess, and she did not see the flaws they did: As it turned out, I was far from wrong and absolutely wrong all at the same time. When we did meet in person, a year or so into our relationship, I was astounded that she was so tiny -- barely 5 ft, 100 lbs soaking wet (wearing wool). The physical Alice was an amazement. She was, however, just as fearless, gutsy, and literate in person as she was [online]. No topic was sacrosanct; no allusion to literacy slipped by her. I was totally surprised at how small she was and completely delighted at how sharp, how funny. She slices and dices as easily with the spoken word as she does with the written. One student's image of a woman he observed online combines typical examples, stereotypes, and exemplars. This is his description of Malaika: On MOO everything can seem larger than life--it can be quite a surprise to realize the people are ordinary. People here can seem more witty and amusing and clever and sexy than the people one knows irl [in real life]. People project a persona here sometimes, and when you meet them they are shyer or whatever. A friend of mine said he was disappointed when he first started meeting people from the MOO. He had the impression they were demi-gods. From information and photos available on this woman's web page, it is apparent that she is in her late teens, but is a petite brunette with long hair. I imagined that Malaika was high-school to early college age. This I inferred from something she said about her parents "letting" her go to Canada (no one keeps their 24-year-old home at gunpoint, at least no one I know of). Another thing she said made me think she was a lesbian and I have an image of short blond hair and a lot of leather clothing for that image. This is based on the fact that of the two gay people I've known well, they've both been blond. In addition, the media stereotypes influenced the length portion; after seeing multiple "stereotypical" lesbian characters, the word always conjures in my mind an image of shorter hair. Having no information to support these images, I assumed her to be fairly "average" looking--a typical person one might encounter while walking down the street. In other words, I did not imagine her to be exceptionally thin or obese, exceptionally tall or short, or have any large body piercings. I imagine her to have a deep voice, which is heavily rooted in stereotypes I've seen; a high voice doesn't, for me, fit with the images I've seen of people portrayed as lesbians (or the two I know personally).
One informant's picture of a MOOer combined typical examples and stereotyping. The former led to a better fit between image and reality, the latter to a mistaken one. He described a woman participant named deLaMer this way:
From an online interview I conducted with deLaMer and from online photographs of her, it is evident she is a 27 year old Asian woman. She is 5'6" and 175 lbs. She described herself to me as "chunky." She has dark eyes, short black hair, and a bright smile. My first judgement was based on name alone. Two things struck me immediately -- the name itself, and its spelling. In the vernacular of cyberspace, "lamer" is a derogatory term for someone that is judged "uncool" or "stupid," and often applied to newbies. The spelling, with its somewhat random combination of upper and lower-case letters, is widely recognized as being in the style of persons who try to act "elite" but who are actually lamers. From these two clues I concluded that the character was likely named in jest, quite sarcastically in fact. This self-deprecating sarcasm reminded me of one person in particular -- myself. And so my first impression of deLaMer was as a sarcastic 19-year-old white male. However, deLaMer's description indicates that she is a 27 year old who presents as female. It also shows that deLaMer's timezone is Philippine Standard Time, and that her home is a place named Laguna Bleu. Even though the time zone does not necessarily indicate location, I now picture her as a Filipino woman -- I get a stereotypical image which includes long dark hair and short stature. I also picture her as thin (for me the Laguna is evocative of swimming and diving, and therefore physical fitness). Finally, I imagine her with a friendly grin on her face, a result of her personality.
Messages, Models, and MeaningsThe meaning of a message (or of a cue) is shaped by the model in terms of which it is interpreted. Differences in the ways in which messages are construed are illustrated in accounts of the same person produced by different informants. Three cases depict this point. Several student-informants provided descriptions of MOO participants named "aldon," "CrashLander," and "Melon." I present their MOO aliases and descriptions and then informants' images of them. The images are based upon combinations of metonymic models and theories of the world.
Case 1. aldon has several aliases, including nardil, squat, amh, Mr.Right, purr-fessur, zoloft, mattress_meat, alson, ghoti-wife, beer_bottle, and morphfiend. His description of his MOO character is "aldon is metaphysical vagabond working as a baggage handler the airport of the emotionally challenged. He has a keen interest in technology, group dynamics, philosophy, literature, and finance. He has dirt under his fingernails from too much gardening. aldon is starting to learn to MOOprogram"
Six informants provided the following images of aldon.
From information I gathered in an online interview with aldon and from an online photograph of him, he is 39, 5'8", 185 lbs., and has brown hair (with a receding hairline), a light beard, and hazel eyes. He wears glasses. He has a broad smile. He said this about his choice of his character's name: "I tend not to think a lot about choice of name. aldon is my real first name. Over time, I've picked up a few different names along the way. I don't put a lot into the name. The description is probably much more important. And, if the place where your name is used allows for easily changing names, then so what? You can be much more creative with the description." (1) The fact that aldon used his real name as his alias is revealing. Most people enjoy the creative freedom associated with creating your own name, but he was content to stick with his real name. This made me think of him as an uptight, rather boring and uncreative person. I imagine him throwing on the closest thing in the morning, barely looking in the mirror before going out. aldon’s self-description portrays him as a metaphysical vagabond, a phrase which caused me to think of him as a man with an artsy goatee and moustache. I also pictured him as older, because his interests seemed to reflect a more mature mindset (how many college boys like gardening?) I visualized aldon as overweight due to the amount of time he spends online, just sitting in a chair. Playing off of the stereotypical computer nerd image, I picture him with horrible vision, a receding hairline, and small, focused eyes. But that other side of him, the side that likes gardening and is sensitive and artsy works against the stereotype by refining it down to a softer version. His smile is where this artsy, sensitive edge reveals itself; I picture it stretching across his entire face.
(2) I imagine aldon as a cross between my father and my high school music teacher. Mr. Polansky, my teacher, has always been very willing to educate others, and will spend as much time as is needed. aldon shares this quality in good measure, and has taught me to morph, program, find helpful features, emote from afar, and has given me much useful information. Therefore, I imagine that aldon has Mr. Polansky's unique nose, and similarly tinted hair. aldon and my father have many things in common. My father is also very willing to teach, and he enjoys it very much. For these reasons, I imagine aldon looking like a somewhat younger version of my father; dark, curly hair with some grey, tall, brown eyes, and broad shoulders. aldon has a small beer belly, the result of his indoor, desk job. However, he has muscular arms and broad shoulders, and likes to go outdoors with his children, doing activities such as bike riding. aldon wears glasses, the product of many hours in front of a computer screen. His clothing is good quality, but is slightly disheveled.
(3) aldon's name had no role in my impression formation. My impression of him was based on my conversation with him and his conversations with others. Immediately after I entered the MOO, aldon waved to me. This simple action made an impression on me. Right then, I saw aldon as very nice guy, someone who is very warm, and very social. These observations further supported my "nice guy" image I had of him. What do I mean by "nice guy"? I guess you could say I have a "nice guy" stereotype. I'm not sure exactly how it was formed, but it is probably due to the fact that whenever I meet a "nice guy", (a male who is extremely warm, engaging, and social), they tend to have the same characteristics. Based on this, I see aldon as tall, dark-haired, fairly handsome, and somewhat on the thin side. At first, I thought he was about my age, but later I imagined him as older because of a piece of information he gave me. aldon told me that he was "in the middle of a divorce" and that he has children. Based on this, I then thought that aldon was in his mid 40's. The reason I thought this is because people usually get married when they are in their late 20's, maybe early 30's. They usually have their kids by about five years into their marriage, so this brings them to about 35. People usually stay married for about ten years before they decide to get divorced. It takes that long for the honeymoon to stop, for reality to sink in, and for them to realize that they've changed and they shouldn't be married anymore. That's why I think aldon is in his mid 40's. My reasoning is a combination of my "man in the middle of a divorce" stereotype, and a logical model based on the average age of divorce.
(4) The name "aldon" struck me as oddly British, maybe an old British surname, so I logically pictured him with a British accent. aldon helped me learn to program and giving me pointers on how to morph, add features, and all sorts of helpful information. Because he was so helpful and uncritical, I pictured him to be a caring old man, just like my grandfather was. He was ready to help with anything, and unwilling to criticize even if I messed things up completely. I pictured him to have thinning gray hair like my grandfather, and the same short stocky build. I also got the impression that he would wear glasses because he spends so much time on the computer that he might be somewhat near-sighted.
(5) aldon’s physical features did not immediately form in my mind, but a few images and impressions have developed. His use of sexual terminology led me to picture him in his late twenties. Perhaps this was due to the immature and "dirty" presentation of information that someone in their late twenties, who was not quite ready to reach thirty and not ready to let go of his youth might say. I also pictured him as very clean cut because of his verbal precision, with a sleek and maybe even sneaky look. I imagined him with a solid build and perfectly combed hair, but with a few wrinkles in his appearance because as he stated in his description, he had the dirt under his fingernails so he was probably not meticulously kept. I imagine that he likes to wear bright colors like red, orange and yellow due to his use of profanity in conversation as well as what I considered to be "boldness" in his choice of words.
(6) aldon's name was not at all evocative to me -- in fact, I still have no idea as to what exactly it means. It did strike me as a male name (perhaps because it is a combination of the male names Al and Don) however, and so I assumed from the beginning that he presented male (I was correct in this assumption). aldon was idling in the room when I arrived, and he returned in this manner: aldon decides he is back. I assumed this was an attempt at dry humor (something very common in all forms of text-based CMC so far as I can tell). He then proceeded to introduce himself to a newbie. He conversed with her genially, which I found very striking -- in most of my text-based CMC experience (whether MOO, IRC, AOL chat rooms, or even newsgroups) it has been very rare that a person will welcome a newbie openly and in a friendly manner. Generally newbies are flamed or, at best, ignored. I thought that perhaps aldon was a relative newbie himself -- I later discovered that he has in fact been on [this MOO] for over four years. There was, at that point, only one other character on the MOO who I knew to be so openly friendly to newbies -- Tower, who I also knew in real life. For this reason I began to picture him as a younger version of Tower-- a tall white male, perhaps with dark brown hair. According to his profile "he has a keen interest in technology, group dynamics, philosophy, literature, finance." This struck me as rather academic in nature, and for some reason that connection caused me to think of him as wearing glasses. Finally, his profile included "he has dirt under his finger nails from too much gardening." Though consciously I doubt that this was the image he was going for, subconsciously the "dirt" part resulted in my picturing him wearing a dirt-stained white tee-shirt.
Case 2. CrashLander has several aliases, including clr, mom, Trish, parody, crash, widdershins, crashie, Someday, vw, Crazy, Red_Rash_Clan, irisheyes, Dr._Ackly, and clargh. Her description is: "Brrruuuummmmm, brrrummmmmmmmmmmmmm, brummmmmmmmmm, screeeeeeeeeeeeeetch ************* CRASH *************
Man, I hate it when that happens."
From an online interview I conducted with CrashLander and from online photographs of her, it is evident that she is 33, 5'10", and has medium-length blonde hair and blue eyes. Her face is oval in shape and her smile is broad and bright. She describes herself in reality (i.e., her offline self) in this way: (1) Crash’s name has an androgynous edge to it, playing on the "ungenderedness" of a plane crashing. Because of this, I visualized her as a strong woman with a determined, pointed chin, and clear blue eyes. In addition, the androgyny of her name led me to imagine her with a short yet feminine haircut. Her name suggests excitement, explosions, fire, and energy, which created an image of an energetic individual. The visual image of a fire (from the "crash") served to create my image of her body and personality. I pictured her wearing bright colors, with lots of reds and oranges. The attractiveness of a fire was shifted onto my image of her as an attractive woman. The flames of the fire, dancing crazily in every direction, translated onto the energy I sensed she had, and made me picture her body as fit from so much exercise. With Crash my methodology did not involve moving from cues to stereotypes, but rather correlating and combining the cues together to see where they fit and where they do not. For instance, when I discovered that she was a mother, I didn’t create an image of her based on my stereotypical notion of what a mother is. Instead, I correlated her energy and explosiveness with her motherhood, and started to picture her as a creative, energetic mother who loved being with children and playing with them. Then adding on my image of her as attractive and witty, I formed an image of her as a young mother who still has an edge. I came to analyze Crash horizontally, adding information to the existing correlated image. Crash’s communication style fit right in with my conceptualization of her name. She hardly ever ended her sentences with punctuation and often made typing mistakes, which made me think that she was impatient and energetic, without the time to carefully end the sentence or check the spelling. This energetic air implied a body that is fit and thin. She spoke with such humor and bravado that I got the impression that she was attractive, since confidence often correlates with an attractive physical appearance. Her sarcastic replies created an image of dark raised eyebrows and jet-black hair. Sarcasm always brings up images of blackness in my mind, perhaps because sarcasm is also referred to as "dark humor." In addition, the people I know with dark hair have tended to be sarcastic and bitter. Since I find both sarcasm and black hair attractive, this also strengthened my image of Crash as a good-looking woman. Like a chain link, each aspect of Crash I pictured inspired another aspect. These links were often related not by general stereotypes, but by personal correlations I have (such as sarcasm correlating with attractiveness).
(2) I formed an impression of CrashLander's looks and some of her personality from her character descriptions, conversations with her, and observations of her conversations with others. Crash is female, thirty-three years old, married with two children, a nine year old son and a three year old daughter, has a dog, and was one of eleven children. From this fairly straightforward information, I have made many inferences to fill in the blanks of my knowledge of her. I picture Crash with light, very white skin, dark hair, either bright blue or green eyes, and a dimple in her chin. I believe that she looks as if she might be Irish, because of her stories of growing up in a Catholic family with eleven children. Her appearance comes from a very good friend of mine who is Irish and comes from a huge family like CrashLander's. From her writing, I feel that CrashLander is a fairly informal person. Therefore, her clothing is fairly casual, perhaps jeans and a sweater or T-shirt.
(3) CrashLander's name alone immediately gave me the impression that this individual’s mannerisms were aggressive, succinct and to the point. The name also led me to believe that this person was impatient and maybe even disappointed about various situations in life. The image that guided me to these conclusions based on the name was that of a pilot flying a plane, and only being able to land the plane by crashing. It was an image of an individual who was supposed to fulfill certain roles and responsibilities (landing the plane safely) and failed in this attempt. This led me to believe that CrashLander was a younger person, someone with little experience in both decision making and in expectations and subsequent disappointments. My impression of CrashLander’s age was also confirmed through dialog that I observed. I witnessed CrashLander’s interest in music on multiple occasions and due to the fact that at one time, she (as I pictured her) mentioned a web-site about music on the Internet, I immediately pictured a teenager with lots of extra time on his or her hands surfing the web for "cool" music sites. Consequently, my impression of CrashLander became one of a young female in her late teens or early twenties. I pictured her as a husky and tough girl due to the aggressive nature of her name with her hair pulled back away from her face to perhaps have a more masculine appearance.
Case 3. Melon has several aliases, including melonhead, Lemon, men, Melonardo_da_Vinci, Melonangelo, MeIon, Melonious_Thunk, Tallulah_Melonhead, Ack!, Mucking_Felon, Melon_Guest, Ponce_de_Melon, Muave, Maudlin_Melon, Enigmelon, Bugsy_Melone, Melooney_Toon, Merry_Melonies, Venus_de_Melon, Norma_Desmelond, because, Melon_Curie, me1on, Vitameatavegemelon, Meloncholy_Baby, ScArY_MeLoN_HaiD, m/, Melon_Spice, and Hurricane_Melon. Her self-description is: "You see nothing special." I'm 33 yrs old and I've been married for 13 years. I have two children, a dog and a house in the suburbs. I am 5' 10" tall. My eyes are a greyish shade of blue. I'm of Irish American descent, and this shows in my skin tone (so white I'm nearly blue), the shape of my face (looks like a potato, mainly) and, I suppose, in my 'gift of gab.' I have blonde hair that's almost shoulder length. It's worn in an out-grown 'bob' style. It's pin straight and I haven't bothered to curl it for several years. My body shape is what I guess you'd call 'large-frame', I am overweight, but not grossly so. Just in the average-American-is-overweight sort of way. CrashLander was my brother's nickname…but I made it my own nickname in senior year of high school when I survived my first of five total collisions. In real life (i.e., offline), based on Melon's appearance in online photographs and on her statements to me in an interview I conducted with her online, she is a 40 year old artist. Her online photographs show a woman who has dark brown eyes, dark shoulder length hair, and a broad and brilliant smile. This is how Melon described herself to me: (1) When I saw the name "Melon" I immediately pictured the roundness, softness, and sweetness of the melon, forming an image of her body. The word "melon" also has an underlying sexual and female tone to it, with its possible reference to breasts. The resultant image created was of a rounded, curvaceous woman, sensual in her largeness. Her aliases often reminded me of something my 14-year-old sister would choose, signaling her playful immaturity. Immaturity is correlated in my mind with dirty blonde hair cut in that super-bobsie look, drawing on the stereotype of the silly blonde. But despite the fact that the aliases were playful, they hinted at her age; a 14 year old would not enjoy puns like "Venus_de_Melon" as much as an older wacky woman. I was also struck by the fact that one of her aliases, "ScArY_MeLoN_hAiD," utilizes the combined upper case and lower-case letters so prevalent in alterna-teen aliases. This type of linguistic style brings up ideas of nonconformity, implying that the person is trying very hard to be alternative and different. I consequently formed an image of a wacky woman who loved wearing outrageous clothing. The objects Melon was holding reminded me of something my mom would have made - a yellow rose, a flower, another rose, a teddy bear, and a long string of seaweed. In addition, she once mentioned "karma’ in a conversation I was listening to. This sounded to me like the remnants of a hippie of the sixties, which parallels my mother. Using the exemplar of my mother’s hippie personality, I transferred that onto Melon, and began to perceive her as a crystal-wearing, colorfully dressed woman. Melon’s interaction style in the context of the MOO created my image of her as a bubbly, blonde, curvy woman, full of red-lipped smiles and flirtatious innuendoes. This then led to my image of her as a secretary, since the descriptions seemed to revolve around stereotypical notions of what a secretary is like. She was extremely talkative, and since verbose people tend to use their hands a lot in conversation, I began to see her as an extremely expressive speaker. Her linguistic style was very informal, so much so I sometimes took her to be uneducated.
(2) The very first thing that I noticed about this person (I did not yet know his/her gender) was the name. I noticed that it can be pronounced in one of two ways. It can be pronounced with the first syllable stressed, and it sounds like a citrus fruit. My initial inclination, however, was to pronounce it with the second syllable stressed, and it sounds like a city in Italy. I liked this way better, because it was different, and is very pleasing to the ear. My own preference on how to pronounce the name had an effect on how I thought this person looked. Melon (the city in Italy) sounds very "artsy" and sophisticated to me, so, at the very sound of the name, an image is created in my mind. When I say "artsy" I mostly mean that I picture this person to dress in a certain manner. If Melon was female, I pictured her to wear puffy shirts and long, flowing dresses. If Melon was male, I pictured him to wear eccentric button-down shirts with tight-fitting pants. Basically, this is my stereotypical view of how a starving artist would dress. I observed Melon chatting with others and I also spoke with her myself. (Obviously, I soon learned her gender.) It was through what she said to others and to me that I formed most of my impression of her, although I still pictured her as "artsy" and sophisticated. At first, I thought that Melon was in her 20's, most likely because that is how old I am. I tend to project my age onto a lot of MOO characters before I know their real age. However, my impression of Melon's age soon changed when she told me that she liked gardening and flowers. Immediately, I pictured her as much older, like in her late 60's. I realize that the reason I did this is because I relied on the image of my grandmother. My 70-year old grandmother is the only person I know that has a passion for gardening and flowers, so I immediately thought of her when Melon mentioned them. I see Melon as short (around 5'4" as a guess). She is thin and is very small-boned. She is the type of person that would shake hands very softly with someone.
(3) Melon's mention of her "ex-husband" led me to picture her as a somewhat burnt-out middle-aged divorcee. My physical image of her is a somewhat short and overweight woman with chin-length hair, glasses, and a wry but world-weary smile. She sits a bit slouched over, and generally appears tired. I suspect that the basic template for this image came from my elementary school bus driver, which is somewhat odd because my bus driver was a frightening, nasty person, whereas I rather liked Melon. However, they both strike me as examples of weary middle-aged people; Melon just manages to handle her weariness with a bit more grace and humor. As Melon’s pictures had recently been posted on the net, there was much discussion as to what she does look like. One person described her as a "Persian Kate Winslett". From this I imagine Melon to have a dark complexion, a round face with big brown eyes, and wavy brown hair. I imagine her teeth to be very white and for her to have full lips and a big smile. I imagine that she is plump (not fat, just chubby) and approximately 5’4" (I think the height is a reflection of her being a mother- my mom is that height). I imagine that she wears brighter colors like red and green (is that what one wears in Persia? - stereotypes again).
(4) Melon is a complicated character. From remarks she makes, she seems to have hit a rough patch irl. Her description is minimal: it's not clear whether the "nothing" you see refers to the online character or the offline person (or both), whether it's a commentary on virtual life or a depressive's self-doubt, or whether she thinks that character names are more interesting and evocative than character descriptions. Her various aliases, with their wry inversions (Melonious_Thunk, Mucking_Felon) and artistic references (Melonardo_da_Vinci, Melonangelo), and her repartee reveal a quick wit and a sharp sense of humor, which make me think she must have twinkling eyes and a bright smile. Her sometimes bawdy comments suggest a zest for life, despite her current troubles. I imagine Melon to be as nimble on her feet as she is in her banter. This makes me think she is lithe; however, I also picture her to be as ample in person as she is expansive in her language. So I see her as being 5'8" or so. I don't know if she is Middle Eastern or southern European in background--her ex-husband suggests the former, her daughter's name suggests the latter--but, in either case, I imagine her to have dark hair and dark eyes. Given her daughter's age, I imagine Melon to be in her middle to late 30s. All in all, she would seem to be an attractive woman.
5'9" tall, normally size 12. Head slightly larger than average (I can tell by trying on "one size fits all" hats). Earlobes triple-pierced, but tastefully. Hair, dark brown. "Cinnaberry" at the moment, and just above my shoulders. Wear my hair up a lot. Glasses yes, but contacts when I go anywhere where I want to look good. People who know me never fail to comment that my glasses make me "look smart". I always wonder what I looked like without them. Skin, starting to lose its youthful dewiness, but still a nicely-put-together face. I mentioned that I'm tall. I'm statuesque; in fact I've modeled as a statue for a drawing of Justice. So anyway, I guess I'm proud of that, and my curvaceousness. I wave my arms around a lot and I have a very animated face. When I was younger I tried theater, I was good at dance, at acting, at sports. The name Melon was given to me by a guy who ran a caucus-type group I was briefly on before I started mooing. Not a terribly interesting story behind it. I love the name though, because it could mean so many different things
ConclusionsAlthough researchers have emphasized the difficulties involved in impression formation in computer-mediated communication, people in the text-based virtual communities of cyberspace do develop images of one another. These impressions are based not only on cues provided, but also on the conceptual categories and cognitive models people use in interpreting those cues. Prototypical effects associated with different models, including but not limited to the stereotypes noted in previous research on computer-mediated communication (Lea and Spears, 1992, 1995; Walther 1996), contribute to discrepancies between online images and offline realities. Moreover, models and categories vary across individuals, further contributing to differences between online expectations and offline experiences. Prototype theory advances the findings of social identity models of information processing in CMC (Lea and Spears 1995, Walther 1996), which focus on the development of online impressions, by analyzing sources of variation in the specific impressions formed and by accounting for the mismatches between online impressions and offline experiences.
Differences in online impression formation, and in the stereotypes, exemplars, and other theory-based views of the world on which they are grounded, provide further evidence that cues gain their significance in terms of the cognitive models or contexts within which they are viewed. When participants in interaction employ different conceptual frameworks, different meanings are attributed to the same message. Of course, this pattern is not peculiar to cyberspace. Geertz (1973) portrayed this scenario as a "confusion of tongues." It is also evident in other studies of offline life, including Suttles' ethnography (1968) of ethnic and racial miscommunication, Tannen's account (1990) of gendered misconceptions, and Jacobson and Ziegler's (1996) analysis of misinterpretation and mistrust between scientists and non-scientists. Paradigms, including prototype theory and related models of "theories of the world", developed to understand offline behavior, are applicable to online behavior, and studies of the latter can contribute to assessing and expanding the former.
These findings suggest a number of topics and questions to be addressed in future research. It is important to further study variation in the content of categories and cognitive models. The interpretive frameworks of different groups (e.g., techies and non-techies) appear to differ significantly, as social identity theory would suggest (cf. Abrams & Hogg, 1990; Robinson, 1996; Walther, 1997). Examining them, as well as those of still other groups, should illuminate our understanding of the cognitive models used in online (and offline) communication and the sources of miscommunication.
Attention should also be given to factors that affect the accuracy of impressions in text-based CMC. Although metonymic and other cognitive models contribute to misperceptions, other factors, in particular individuating information (see Brewer, 1996), may bear upon the fit between expectation and experience. Does the amount of individuating information correlate with duration of an online relationship, and, correspondingly, with clarity and/or accuracy of impressions formed? Some data would suggest not. For example, Reid describes a couple who met online in a MOO, spent a "few months" getting to know one another, and then met offline. The woman recounting the story said, "He was different from what I'd expected, mostly in the way he looked…." (quoted in Lea & Spears, 1995, p. 205). In interviews I conducted for this study, informants mentioned interacting with people online for as long as several years before meeting face-to-face and still being surprised by one another's appearances. Although I was not able to select informants in terms of the length of time they spent online with other people and/or the amount and kinds of information they provided one another, further studies should focus on these factors in naturally occurring situations of the sort found in virtual communities.
Finally, research can also address the issue of how the particular context of interaction affects the process of impression formation and the fit between online expectations and offline experiences. Social context affects the activation of the categories in terms of which people interpret one another (Barsalou,1987; Brewer, 1988, 1996; Danet et al., 1998; Jacobson, 1996; Kunda & Thagard, 1996; Lea & Spears, 1992, 1995; Spears & Lea, 1994). Participants' assumptions about MOOs (as a kind of virtual community) frame the ways in which they interpret signals others send. For example, some participants assume they are fictional and what happens in them is not to be taken seriously; others take them very seriously, arguing that online worlds are as real as those experienced offline. Some participants assume that others misrepresent their offline gender and other personal characteristics; other participants assume that people are like them, sharing attributes of age, race, and socioeconomic status. In virtual communities of the sort reported here, impression formation is influenced by choice of pseudonym and description of virtual character as well as the content of messages exchanged. How does this process work in other modes of computer-mediated communication such as email and newsgroups, in which participants may not use pseudonyms or employ digital characters described in more or less detail? Answers to this question await comparative analyses across different modes of computer-mediated communication.
AcknowledgementsI want to extend my appreciation to several people who helped with the research on which this paper is based. First, I thank the participants of the virtual communities who gave me their time and their accounts of online expectations and offline experiences in meeting other MOOers. They remain anonymous, but they know who they are. I also thank the students in my course who did fieldwork and who contributed their impressions of the MOOers they encountered. I am especially grateful to Michelle Amato, Ksenia Babich, Jennifer Bunk, Michelle Carter, Michael Cohen, Shelley Coughlin, Alex Darrow, Abbie Moscovich, Allyson Pitt, Jonathan Pratt, Jesse Richman, Beth Sanders, Kerri Schoonover, and Sarah Soslow . For discussions of prototype theory, I am indebted to my colleague, Benson Saler. I also want to thank several people who discussed with me issues in impression formation and/or who commented on earlier drafts of this paper: Samantha Herdman, Elizabeth Hess, Mac Parks, Kaisa Vahahyyppa, Joe Walther, and the editor and two anonymous reviewers of this journal.
ReferencesAbrams, D., & Hogg, M. A. (Ed.). (1990). Social identity theory: Constructive and critical advances, New York: Springer-Verlag.
Barsalou, L. W. (1987). The instability of graded structure: Implications for the nature of concepts. In U. Neisser (Ed.), Concepts and conceptual development: Ecological and intellectual factors in Categorization. (pp. 101-140). MA: Cambridge UP.
Bartle, R. (1996). Hearts, clubs, diamonds, spades: Players who suit MUDS, Journal of MUD Research [online], 1(1), Available: http://journal.tinymush.org/v1n1/bartle.html.
Brewer, M. B. (1988). A dual process model of impression formation. In T. K. Srull and R. S.Wyer, Jr. (Eds.), Advances in Social Cognition. NJ: Lawrence Erlbaum Associates. (1) 1-36.
Brewer, M. B. (1996). When stereotypes lead to stereotyping: The use of stereotypes in person perception. In C. N. Macrae, C. Stangor, and M. Hewstone (Eds.), Culture of the Internet. (pp. 254-275). New York: The Guilford Press.
Cantor, N., & Mischel, W. (1979). Prototypes in person perception. Advances in Experimental Social Psychology, (12) 3-52.
Curtis, P. (1197). Mudding: Social phenomena in text-based virtual realities. In S. Kiesler (Ed.), Culture of the Internet. (pp. 121-142). NJ: Lawrence Erlbaum Associates.
D'Andrade, R. (1995). The development of cognitive anthropology. MA: Cambridge University Press.
Danet, B, Ruedenberg, B. L., & Rosenbaum-Tamari, Y. (1998). Hmmm…where's that smoke coming from?: Writing, play, and performance on internet relay chat. In F. Sudweeks, M.McLaughlin, and S. Rafaeli (Eds.), Network and Netplay:Virtual Groups on the Internet. (pp. 41-76). AAAI Press/The MIT Press.
Ekman, P., & Keltner, D. (1996). Universal facial expressions of emotion: An old controversy and new findings. In U. Segerstrale and P. Molnar (Eds.), Nonverbal Communication: Where Nature Meets Culture. (pp. 27-46) NJ: Lawrence Erlbaum Associates.
Fiske, S. T., & Taylor, S. E. (1991). Social Cognition (Second Edition). NY: McGraw-Hill.
Geertz, C. (1973). Thick description: Toward an interpretive theory of culture. The Interpretation of Cultures. (pp. 3-30) New York: Basic Books.
Goffman, E. (1959). The presentation of self in everyday life. Garden City, New York: Doubleday Anchor Books.
Hahn, U., & Chater, N. (1997). Concepts and similarity. In K. Lamberts and D. Shanks (Eds), Knowledge, Concepts, and Categories. (pp. 43-92). Cambridge, MA: The MIT Press.
Hamilton, A. (1999, February 15). You've got male!. Time, 83.
Hamilton, J. O'C. (1999, February 22). When cupid uses cursor, BusinessWeek, 26B, 26D.
Hampton, J. (1997). Conceptual combination. In K. Lamberts and D. Shanks (Eds.), Knowledge, Concepts, and Categories. (pp. 133-159). Cambridge, MA: The MIT Press.
Jacobson, D. (1996). Contexts and cues in cyberspace: The Pragmatics of Naming in Text-Based Virtual Realities. Journal of Anthropological Research. ( 52) 461-479.
Jacobson, D., & Ziegler, C. (1996). Popular delusions and scientific beliefs: Con-flicting belief systems of scientists and administrators in the creation of a covert technological surveillance system., InL. Nader (Ed.), Naked Science. (pp. 228-238). New York: Routledge.
Kunda, Z., & Thagard, P. (1996). Forming impressions from stereotypes, traits, and behaviors: A parallel-constraint-satisfaction theory. Psychological Review. 103 (2) 284-308.
Lakoff, G. (1987a). Women, fire, and dangerous things: What categories reveal about the mind. Chicago: University of Chicago Press.
Lakoff, G. (1987b). Cognitive models and prototype theory. In U. Neisser (Ed.), Concepts and conceptual development: Ecological and intellectual factors in categorization. (pp. 63-100) Cambridge: Cambridge University Press.
Lea, M., & Spears, R. (1992). Paralanguage and social perception in computer-mediated communication. Journal of Organizational Computing. (2,3,4) 321-341.
Lea, M., & Spears, R. (1995). Love at first byte? building person-al relationships Over computer networks. In J. T. Wood and S. Duck (Eds.), Under-studied Relation-ships: Off the Beaten Track. (pp. 197-233) Thousand Oaks, CA: Sage Publications.
Mackie, D. M., Hamilton, D. L., Susskind, J., & Rosselli, F. (1996). Social psychological foundations of stereotype formation. In C. N. Macrae, C. Stangor, and M. Hewstone (Eds.), Stereotypes and Stereotyping. (pp. 41-78) New York: The Guilford Press.
Marvin, L. (1995). Spoof, spam, lurk and lag: The aesthetics of text-based virtual realities. Journal of Computer-Mediated Communication [Online] 1,(2). Available: www.ascusc.org/jcmc/vol1/issue2/marvin.html.
Medin, D. L., & Wattenmaker, W. D. (1987). Category cohesiveness, theories, and cognitive archaeology. In U. Neisser (Ed.), Concepts and conceptual development: Ecological and intellectual Factors in Categorization. (pp. 25-62). Cambridge: Cambridge University Press.
Neisser, U. (1987). Introduction. In U. Neisser (Ed.), Concepts and conceptual development: Ecological and intellectual factors in categorization. (pp. 1-10). Cambridge: Cambridge University Press.
Quinn, N., & Holland, D. (1987). Culture and cognition. In D. Holland and N. Quinn (Eds.), Cultural models in language and thought. (pp. 3-40). Cambridge: Cambridge University Press.
Reid, E. (1995). Virtual worlds: Culture and imagination. In S. Jones (Ed.), Cybersociety:Computer-Mediated Communication. (pp.164-83). Thousand Oaks, CA: Sage Publications.
Robinson, W. P. (Ed.). (1996). Social groups and identities: Developing the legacy of Henri Tajfel, Butterworth-Heinemann, Oxford.
Rosch, E. (1978). Principles of categorization. In E. Rosch and B. B. Lloyd (Eds.), Cognition and categorization. (pp. 27-48). Hillsdale, NJ: Lawrence Erlbaum Associates.
Smith, E. E., & Medin, D. L. (1979). Categories and concepts. Cambridge: Harvard University Press.
Spears, R., & Lea, M. (1994). Panacea or panopticon? The hidden power in computer-mediated communication. Communication Research, 21(4) 427-459.
Stangor, C., & Schaller, M. (1996). Stereotypes as individual and collective representations. In C. N. Macrae, C. Stangor, and M. Hewstone (Eds.), Stereotypes and stereotyping. (pp. 3-37) New York: The Guilford Press.
Strauss, C., & Quinn, N. (1997). A cognitive theory of cultural meaning. Cambridge: Cambridge University Press.
Suttles, G. (1968). The social order of the slum: Ethnicity and territory in the inner city. Chicago: University of Chicago Press.
Tannen, D. (1990). You just don't understand: Women and men in conversation. New York: Morrow.
Walther, J. B. (1993). Impression development in computer-mediated interaction. Western Journal of Communication, (57) 381-398.
Walther, J. B. (1996). Computer-mediated communication: Impersonal, interpersonal, and hyperpersonal interaction. Communication Research, 23(1) 3-43.
Walther, J. B. (1997). Group and interpersonal effects in international computer-mediated collaboration. Human Communication Research, 23 (3) 342-369.
Weber, R. (1998, September 14). New Yorker. 71.
Zebrowitz, L. A. (1996). Physical appearance as a basis of stereotyping. In C. N. Macrae, C. Stangor, and M. Hewstone (Eds.), Stereotypes and Stereotyping. (pp. 79-120) New York: The Guilford Press.
About the AuthorDavid Jacobson is Associate Professor of Anthropology at Brandeis University. He has conducted fieldwork in East Africa, Boston, Los Alamos, and cyberspace. His publications include Itinerant Townsmen: Friendship and Social Order in Uganda (Waveland, 1986), Reading Ethnography (SUNY Press, 1991), and Spying Without Spies (with Charles Ziegler, Praeger, 1995). He has also published "Contexts and Cues in Cyberspace: The Pragmatics of Naming in Text-based Virtual Realities" (Journal of Anthropological Research, 1996) and "Doing Research in Cyberspace" (Field Methods, 1999).
Address: Department of Anthropology, Brandeis University Waltham, MA 02454-9110 USA