Flaming studies often argue that the assumed paucity of social cues in online environments leads to disinhibition and increased flaming online. Other studies empirically challenge the idea that flaming behaviors are an inevitable byproduct of computermediated communication. Nevertheless, both camps share the view that a category of social phenomena called flaming exists and it is possible to determine what constitutes a flame by examining message content. Yet flaming is an oversaturated term that ignores the interactional nature of both flames and flame claims. Importantly, flame claims, in which interlocutors accuse others of flaming are just as important to study as socalled flames. Rather than determining whether a particular message is or is not a true flame, researchers should examine the interplay between flames and flame claims to understand how particular groups establish, negotiate, and challenge both cultural norms and microsocial hierarchies.
Scholarly studies on computermediated communication (CMC) often argue that the assumed paucity of social cues in online environments leads to disinhibition and increased flaming online (Kiesler, et al., 1984; Siegel, et al., 1986; Kiesler and Sproull, 1992; Tannen, 1999). However, many researchers have convincingly challenged the idea that disinhibited behaviors automatically appear and increase in frequency in computermediated environments (Culnan and Markus, 1987; Lea, et al., 1992; Postmes, et al., 1998; Spears and Lea, 1994; Walther, 1992; Walther, et al., 1994; O’Sullivan and Flanagin, 2003). Some scholars often equate disinhibited behavior with the term flaming, which describes a range of phenomena including things like hostility, increased emotionalism, profanity, and criticism (OSullivan and Flanagin, 2003; Lee, 2005).
OSullivan and Flanagin offer a particularly interesting attempt at reconceptualizing the flaming phenomena. They argue that: 1) flaming definitions are imprecise; 2) scholars tend to claim that flaming is primarily or exclusively a CMCbased phenomena; and, 3) researchers often assume that flaming is negative or destructive . They propose a taxonomy that locates flaming within other interactional phenomena such as intentional or miscommunicated violations of local cultural norms. In their view, a flame is a message that violates cultural norms. Their taxonomy considers how several parties, including the sender, the receiver, and a thirdparty observer (such as a researcher or community insider) of a message may or may not perceive a particular message as a flame.
Importantly, many scholars who espouse traditional views of flaming (i.e. they result from disinhibited and anonymous computermediated communication) and many who question those traditional views still tend to accept the fact that a category of online communication called flames exists. According to both perspectives, researchers can objectively discover flames by examining the content of messages. In the CMC literature A general consensus is that flaming consists of aggressive or hostile communication occurring via computermediated channels . Even OSullivan and Flanagin, who call for a rethinking of how scholars have coded and evaluated flames, nevertheless conclude that a category of online phenomena call true flames exist, and they define such flames as [messages] in which the creator/sender intentionally violates interactional norms and is perceived as violating those norms by the receiver as well as by thirdparty observers . Socalled true flames are different from missed flames, failed flames, and inside flames. Missed flames result when the sender intends to flame and a third party sees it as a flame but the receiver does not interpret it as such. In failed flames, the sender intends to flame but neither the receiver nor third parties view it as a flame. Inside flames occur when the sender and receiver see it as a flame but a third party does not understand its flammability . In their model, flames exist as a viable category, albeit one that could use more precise definitions and coding procedures. But if the term is as problematic as OSullivan and Flanagin suggest, why does it continue to receive attention as a viable metric of online interaction?
One answer is that researchers use the term to signal that they understand how online participants characterize certain oppositional aspects of their interaction. However, the term is so oversaturated that it has lost theoretical value (if indeed it ever had any). Worse, retaining the term in scholarly research smuggles in certain theoretical assumptions that may or may not be true. For instance, scholars often use the term to characterize online (as opposed to offline) interaction. But by bracketing off flaming as something automatically different from offline behavior, researchers may miss crucial similarities between online and offline interaction. Further, scholars have often identified flames on the basis of certain predefined criteria in the content of messages. But the flaming process is more complicated and focusing on flames ignores the flame claims that play just as great a role in determining whether or not a socalled flame exists. Flame claims are just as important to study as the socalled flames themselves. By examining the interplay between flame claims and flames researchers may gain crucial insight into how online participants use flames and flame claims to negotiate and challenge key aspects of their online group, such as certain cultural norms.
The unquestioned existence of a category of social phenomena called flames is an example of what sociologist Pierre Bourdieu called doxa or unquestioned beliefs and naturalized ideas about the world that both orthodox and heterodox believers espouse. Believers of orthodoxy seek to maintain a certain set of ideas while believers of heterodoxy seek to challenge those ideas (Bourdieu, 1977). The orthodox view maintains that anonymous persons communicating in a mediated way is likely to produce an increase in socalled flames. In contrast, the heterodox studies that challenge this idea state that a number of factors such as time spent with ones online interlocutors and anticipated future interaction with them greatly influence whether or not flaming will likely be prevalent in online environments.
Nevertheless, scholars from both camps tend to accept a certain unquestioned belief that flames exist. For Bourdieu, doxa represents unquestioned, commonsense beliefs about the world for which interlocutors do not have ready discourse to challenge supposed objective assumptions about how the world works. Even OSullivan and Flanagins reworked taxonomy, which challenges the flaming doxa, assumes that true flames exist, if one considers the perspective not just of the coder, but also of the sender and receiver of the message. OSullivan and Flanagin state that true flames are intentional (whether successful or unsuccessful) negative violations of (negotiated, evolving, and situated) interactional norms .
OSullivan and Flanagins article is important because it opens a space for considering not just how researchers might evaluate message content to determine flames, but also urges an analysis of how senders and receivers (and other third parties) cultural norms may influence the perception of messages. However, their taxonomy is problematic because it does not consider the intentional fallacy and the cultural relativity problems. The first refers to privileging a message authors socalled intention in sending a flame as the most important factor for understanding whether a message is a true flame or was only misperceived as a flame by receivers or outside parties. Yet, as I discuss below, it is not always easy (or even possible) to determine or access a senders intention. The second problem refers to the difficulty in deciding who should be allowed to judge whether or not and which norm a flame supposedly violates. Not everyone agrees that a cultural norm is actually a norm or should be so for everyone. Norms (individual, small group, and large group) conflict and are often in flux, and which norms a researcher chooses to consider in determining whether a message violated a norm depends upon a researchers preferences and cultural predisposition. I will discuss both of these problems in turn, using empirical data to illustrate the difficulties in coding messages as flames. The point here is not to suggest that hurtful messages do not exist online. They most certainly do. Failing to label a message that has hurt someone as a flame is for some an additional violation. The focus of research should not be on determining better coding procedures for socalled flames. Rather, the hope is to encourage less scholarly interest in determining a category of true flames and encourage more scholarly attention on analyzing the interplay between socalled flames and flame claims to answer broader cultural questions, such as how socalled flames and flame claims in different groups have certain social effects, such as to create bids at potential microsocial hierarchies between participants (Lange, 2003).
... it is not always easy (or even possible) to determine or access a senders intention.
The first problem with trying to approach the problem of flames from an intentional perspective is that scholars in several fields long ago problematized the idea that determining and accessing author intentionality is straightforward or even possible. As Duranti points out, a particularly Western perspective in linguistic studies is that a certain intention exists and is fully defined in a speakers mind before the act of speaking . Such a view assumes that wellformed intention exists for every message and that speakers wish to be clear about their intention. Duranti argues, however, that There are many cases in daily life in which the meaning of a given act is not defined until the recipient of that act has replied . Morgan, for example, found that in the speech of African American women, intentionality was not a particularly meaningful or accepted indication of how recipients should interpret messages (Morgan, 1991). Whatever intentionality a speaker claimed, recipients argued that the speaker should take responsibility for how recipients interpreted the speakers messages (Morgan, 1991). For Duranti, Morgan, and other scholars, it is not possible to determine intention prior to an interaction, whereas for OSullivan and Flanagin, intention exists before a sender transmits a message. OSullivan and Flanagin argue that researchers should not be the only ones who code messages, but should also take into account codings of the interactants themselves, who would provide data on the intent of their own messages and their interpretations of messages that they receive .
However, it is unclear whether interactants have welldefined intentions, much less have access to them or wish to share them with researchers and coders. OSullivan and Flanagins analysis assumes that intention exists and is always clear. Yet, as Eisenberg and Philips (1991) convincingly point out, message senders do not always intend to be clear about their messages. Ambiguity is often an important component of messages and may facilitate more effective communication, such as when one person wants to avoid direct confrontation. One may object that a message sender may intend to be ambiguous and thus intention exists even in ambiguous messages. But if a sender intends to be ambiguous and only possibly intends to send a flame, then from a theoretical point of view researchers cannot comfortably code such a message as a definitive flame. From the perspective of the sender of an ambiguous message, the researcher cannot know whether the sender intended to flame.
This point may be illustrated by examining specific empirical data in which one person claims that another has flamed them. In the first example, we will examine implicit claims of over criticism, and in the second example we will examine explicit instances in which someone claims that one person has flamed another person. I observed these exchanges as part of a twoyear ethnographic research project. For the study, I participated as a member of two online communities, which I will call here Community A and Community B. Communities A and B were part of a genre of MUDs (multiuser dimensions) in which participants could roleplay with each other, engage in games, and chat about offline subjects on socalled real life communication channels. In MUDs, people can chat in real time by typing messages to a group of participants at once on public communication channels. Anyone currently logged in to the MUD may see the conversation and join in if they wish. Members of Communities A and B ranged from teenagers to adults who were students and employees from technical companies and academic institutions. I was particularly interested in conversations in which the interlocutors discussed and argued about technology. Many of the participants in both communities were either members of or aspired to join communities and organizations related to computers and networking activities. Much of the tech talk on these groups communication channels often involved mentoring others with respect to correct norms of conversation and behavior within the local community as well as across larger imagined communities of technologists, such as, for instance, the free and open source (F/OSS) community, which promotes open circulation of ideas and technology.
In the following conversation taken from Community A, Jack mentors Brian regarding the correct way to express oneself in cyberspace. Note that Jack and Brian were both believed to be teens, although Jack was said to be a few years older than Brian.
Example (1) from Community A 1 Brian: HEH, I like Rays new title 2 Jack: why do you always do heh in caps? 3 Bert: heh 4 Brian: because it implies that HEH is funnier then just a regular heh. :P 5 Jack: no, its just lame 6 Jack: funnier than a heh is lol 7 Jacob: I still say this thing is...is amazing! 8 Gil: Use false words. 9 Gil: ‘Krunk 10 Brian: fine.. 11 Gil: Thats just krunkin unbelievable! 12 Brian: heh, I like Rays new title 13 Jack: thats knuts! 14 Gil: Krunkin aye! 15 Jack: kanuts, rather :P 16 Brian: how that, mr. I Have To Be Percise On Internet Lingo? 17 Jack: thats good. 18 [Community A] Leo enters the game. 19 Brian: hi leo 20 Leo: hola 21 Jack: heh, you get pissy too easily brian :P 22 Bert: lol ... 23 [Community A] Brian left the game.
In line 1, Brian expresses amusement and admiration for a fellow participants new title for their online character. In line 2 Jack asks Brian why he writes the word HEH in capital letters. In line 4 Brian argues that it implies that HEH is funnier than a heh. In line 5 Jack contradicts Brians explanation and says that it is just lame. In line 6, Jack provides what he displays as the correct elevation of online mirth, which moves from heh to lol (for laughs out loud). In line 10, Brian responds to Jacks critique with a fine and an ellipsis, which indexes forthcoming talk. Brian colludes with Jacks critique by responding with fine and eventually revising his talk to remove the HEH. By criticizing Brian, Jack codes HEH as a nonnormative expression. Although Brian uses the word fine and revises his talk, people often use the word fine in American conversation to signify a begrudging capitulation to end an argument, rather than to show they agree. Brian may not actually agree with the accusation that he violated a communicative norm. In line 12, Brian revises his talk and uses the word heh instead of HEH. In line 16 Brian asks for Jacks approval by saying [hows] that? By asking for Jack to approve of the revision, Brian tacitly confirms that he was in need of correction. In addition, his revised talk socially ratifies Jack as a person capable of providing a correction on Internet talk.
This pattern is similar to one that Jacoby and Gonzales (1991) observed in tech talk among university physicists in facetoface meetings. In their analysis, they found that a person (such as a graduate student) who may not be considered an expert outside of the conversation may correct a socalled expert (such as a professor) within a group. Jacoby and Gonzales pointed out that if someone revises or corrects another persons talk, the revision sets up a kind of bid for a microsocial hierarchy of expertise in which the person doing the correcting displays him or herself as one who is able to correct their interlocutor. The people who revise their talk display themselves to be people who need correction. Thus, accepting a correction signals that one is less expert than the person doing the correcting. This microsocial hierarchy of shifting expertise may not be the same outside of the conversation, or in other conversations. The process of correction is a bid at expertise within a conversation.
The process of correction is a bid at expertise within a conversation.
Although Brian revises his talk and displays himself as one who needs correction, Brian also takes issue with Jacks correction by giving Jack a new title, which is Mr. I Have to Be Precise On Internet Lingo. This title implies that although Brian has revised his talk, he objects to Jacks criticism of Brians online use of HEH. He is implicitly accusing Jack of being too critical and too restrictive about the proper way to express oneself online. In traditional CMC scholarship, one might argue that Brian is accusing Jack of flaming him, if by flaming we mean using hostile or unnecessarily critical language. In line 17 Jack ratifies Brians revision by saying thats good and in line 21 Jack accuses Brian of getting pissy too easily. Jacks accusation in turn displays an objection that Brians critique of Jack has been unduly harsh. In a sense, Jack is counteraccusing Brian of flaming Jack.
If we used OSullivan and Flanagins taxonomy, we would need to accomplish several steps before we could decide whether the interaction contained any true flames. First, we would need to determine each senders intent. Next, we would evaluate the reception of the message, and take into account the opinion of thirdparty observers. But as mentioned above, determining intent is theoretically and methodologically problematic. What is Jacks intent? Was he concerned with maintaining precise communicative standards on the Internet or did he merely wish to taunt Brian? How can we access Jacks intent? What if we asked Jack to code his own talk? How might his answer change if we asked him early in the conversation (say after line 6) rather than after Brian has displayed objection to Jacks criticism (after line 16)? How might Jacks true personal coding change because he is talking to a researcher, whom he may perceive as critical of his criticism?
The problem of intent also brings up the problem of cultural relativity in judging norm violations. Was Jack too critical in the way he corrected Brian? One interpretation of the interaction is that Jack never intended to flame Brian, but merely pointed out a problem with Brians talk. Specifically, Brians HEH did not conform to what Jack displays as normative Internet communication. For Brian, his use of HEH did not violate norms. Norm violation assessments may differ according to the cultural perspective of different participants.
One of netiquettes oftquoted rules is to avoid using capital letters, which is considered shouting online . Instead, to provide emphasis users should use asterisks as you would quotation marks . Obviously, many people do use capital letters when they want to create a certain conversational effect, and I observed many instances of capitalletter use pass by unremarked in the online interaction of Community A. Using lol instead of HEH is not part of netiquette and is nowhere required as a standardized escalation of amusement locally in Community A nor across the Internet. In fact, one might argue that moving from heh to HEH is a logical escalation since netiquette says that a capital letter form of HEH signifies more emphasis than does a lowercase heh.
Deciding who is correctly expressing or violating cultural norms is culturally relative. There are at least two potential norm violations at work here. First, was Brians HEH critizable because it indeed violated cultural norms? Second, was Jacks form of criticism and correction within the bounds of local (and perhaps Internetwide) cultural norms? Is using HEH violating a cultural norm? If so whose? A local norm or an Internet norm? If many instances of capital letter use pass by unremarked why is Brians particular use of capital letters a norm violation? Is it possible to talk about Internet norms given the wide variety of contexts and differences in what constitutes acceptable forms of talk? If Brian did violate a norm by using HEH (one should not use capital letters) then Jack was within his rights to criticize and his criticism did not violate a different communicative norm (one should not unduly criticize another persons form of talk). On the other hand, if Brian did not violate the capital letters norm by using HEH then Jack did violate the norm that says one should not unduly criticize.
If we assume that Jack is operating within community norms (and indeed Brian was out of line), then we might argue that Jack did not intend to flame. If we take Brians position (that his talk violated no norm and that Jack was unduly critical), then Jack has violated community norms and flamed Brian (whether or not Jack intended to flame). Breaking the deadlock requires taking sides in what is essentially a culturally relative interpretation. Not only is determining intent difficult, but in OSullivan and Flanagins taxonomy, deciding which if any cultural norms were violated is not theoretically straightforward, as they acknowledge. The situation is complicated when the researcher considers multiple, possibly contradictory norms between what may be acceptable at the local level such as Community A and what may be acceptable given a larger area of study such Internet talk.
For the moment, let us leave aside the problem of trying to determine the sender, Jacks, intent. It is arguably true that the receiver, Brian, has displayed objection to Jacks criticism. He does revise his talk, but he also accuses Jack of being unduly critical by labeling him, Mr. I Have to Be Precise on Internet Lingo. Using OSullivan and Flanagins formulation, we would next turn to what they call thirdparty observers to the conversation to determine whether an objective flame exists. Yet, OSullivan and Flanagin never account for thirdparty observers disagreement on what constitutes a violation of a cultural norm. In their taxonomy thirdparty observers metaphorically vote as a block and if such observers see a message as a flame, then researchers should code the message one way, and if they do not, researchers should code the message in a different way. But what if thirdparty observers (such as myself as the researcher and Bert who is a participant) disagree on their flame interpretations? As I observed this conversation unfolding, I felt that Jack was being unnecessarily harsh to Brian. I wondered why Jack was singling out Brian when not all capital letter use received criticism. In addition, I had observed Jack and Brian many times and many of the participants (especially older youth) seemed to taunt Brian frequently. That Jack was said to be an older teen and Brian a younger teen (I never met either of them in person) is not surprising given what appeared to me to be a rather adversarial relationship.
Scholars have demonstrated that children tend to orient their talk around age. According to Maynard, it is well know that children display an orientation toward age as an indicator of relative power and status . Children use arguments to create a kind of microsocial hierarchy in which the successful accuser appears to be socially higher than the accused rule breaker. Interestingly, children do not necessarily differ from adults in terms of the things that they argue about. According to Shantz, children and adults tend to argue about similar things, including: valued resources, controlling others behavior, rule violations, facts, and truth . Yet, Maynard states that, it need not be the case that children believe in the exhibited value in any strong sense . Children invoke rules in order to jockey for social position within their group. As Maynard asserts children appear to use normative assertions to promote or maintain an immediate social order .
Jacks criticism of Brian illustrates Maynards point that invoking rules to serve local social ends does not necessarily indicate that the rule invoker uses or believes in that rule (Maynard, 1985). Maynard warns that, care must be exerted not to equate the rule with a social groups own culture. Rather, we should examine how the rule is used in an interactive situation to achieve indigenous social ends which may or may not correspond to the content of the rule . Jack may or may not actually believe that lol is the invariably proper escalation of amusement after heh in online talk. Nevertheless, he criticizes Brian in a way that potentially creates a microsocial hierarchy in which Brian is in a onedown position to Jack. Jacks use of this rule does not mean it is actually a cultural norm in Community A nor for Jack himself. Rather, Jacks use of the rule has the consequence of making a bid to align conversational participants in particular microsocial arrangements.
Jockeying for social position (whether intentional or not) is not necessarily bad for everyone, although some affected participants may feel that these types of interactions are bad. In this assessment, I agree with OSullivan and Flanagin that socalled flames as those potentially contained in the above sample are not necessarily bad. Who are we to decide that males jockeying for social position within a social group is a bad thing? But if participants can use flames for a variety of social purposes both good and bad how do we know what exactly constitutes a flame? If we use norm violation as a metric, then flame candidates will constantly surface from a vast and murky pool. Using a norm violation metric can lead the researcher to code some rather odd things as flames. For instance, what if a norm in a community is to heavily lace ones talk with profanity? What if someone refrains from using profanity and thus violates a local norm? Does this violation mean that the person not using profanity has flamed another person? Much more information is necessary in order to understand what work particular messages are doing in online (and offline) talk. If flames can be good or bad and everything in between, then it becomes complicated to determine what constitutes a socalled true flame. The term flame itself is intimately joined with moral judgments about talk.
Who are we to decide that males jockeying for social position within a social group is a bad thing?
Participants may use talk to portray others and themselves in certain ways, such as experts or nonexperts. Whether or not such portrayals are bad or intentional, they may have certain effects on the interlocutors and others observing the conversation. In her observations of childrens opposition, Goodwin points out that, it is not sufficient to focus exclusively on the talk through which opposition is produced; one must also take into account how actors are portrayed and constituted through that talk . Goodwin discusses how speakers may use certain turn shapes and intonation contours to both build a small effigy of the party being opposed, and display his/her own affective alignment to the actions that such a person performs . Speakers can caricature other speakers by portraying [their] actions as ridiculous or inappropriate . Similarly, Brian caricatures Jack by calling him Mr. I Have to Be Precise on Internet Lingo. By using this term, Brian builds a small effigy of Jack and displays Brians own negative stance against Jacks ridiculous criticism.
OSullivan and Flanagin argue that norm violation defines flames. But it is the process of negotiation between flame claims and the flames that interlocutors use to maintain and challenge cultural norms. For instance, Jack criticizes Brians use of capital letters as nonnormative behavior. In turn Brians implicit flame claim contradicts Jacks bid at claiming that using capital letters violates an online cultural norm. Jacks statement that Brian gets pissy to easily is arguably an attempt to rekey his comment as a joke or teasing rather than as true criticism. If Jack codes his original comment as not true criticism, then Brian has not violated a norm since Jack is not really correcting him, but only teasing him. Brians flame claim challenges the force of the original correction. The revised criticism (downgraded from criticism to teasing) displays a retraction that a cultural norm to capital letter HEH actually exists and deserves strict correction.
To return to OSullivan and Flanagins taxonomy, do thirdparty observers align with Brian or Jack? When I saw this interaction unfolding, I believed Jack was being unduly critical. Whether or not criticism is equal to a flame is another theoretical issue. It has certainly been part of the scholarly tradition on flames to equate hostility in the form of criticism as a flame. But, what is (or should be) the role of thirdparty participants or socalled conversational insiders in determining the existence of a true flame? At that point in time, I had been a participant the community for over a year. I had observed Jack and Brian interact many times and I had observed Brian interact with other people. One could argue I was an insider rather than someone coding conversations taken from a culturally unfamiliar context. Determining who qualifies as a true inside thirdparty observer once again brings up a host of culturally relevant determinations. Who should be coded as an insider? Does an insider always have insight a researcher might not have? Is it possible for a researcher to have insight not available to insiders? What happens when thirdparty observers do not agree on what constitutes a flame, or on which if any cultural norms a speaker has violated?
Let us now examine the reaction of Bert, another thirdparty observer. In line 22 Berts contribution to the conversation is to lol or laugh out loud. His laughter immediately follows Jacks comment that Brian gets pissy too easily. Nowhere does Bert indicate that Jack has stepped out of the bounds of acceptable cultural norms; in fact he laughs at the exchange which displays amusement at Brian and Jacks argument. What was the intent of Berts laughter? A number of interpretations are possible but it is difficult to determine from laughter alone whether or not Bert codes Jacks (or Brians) behavior as violating local or Internetbased communicative norms. One interpretation is that he laughs in support of Jacks countercriticism that Brian has become pissy. As Goodwin and Maynard note, other parties in an interaction may express alignments to a segment of talk as bids to increase their own social positions (Goodwin, 1990; Maynard, 1985). Note that at the end of Jacks countercriticism in line 21 he adds the emoticon :P which signifies a person sticking out ones tongue. Sticking out ones talk in conversation often implies that one is joking or teasing. Jacks use of the :P emoticon tries to code his talk as teasing and jovial. By similarly displaying laughing, Bert expresses alignment to Jacks supposed teasing stance. Berts move arguably places himself in a higher social position along with Jack, who by correcting Brian makes a microsocial bid (that may or may not be intentional or effective) to place Brain as lower on their technosocial hierarchy than Jack (and possibly also Bert).
The above discussion demonstrates that it is quite difficult to determine Jacks true intent in this exchange. Further, differences of opinion in thirdparty observers mean that thirdparty opinions are not reliable, consistent predictors of what constitutes a flame. One interpretation of Jacks comment that Brian has become too pissy is that Jack never intended to be unduly critical and flame, but rather Brian has simply misinterpreted him. According to OSullivan and Flanagins taxonomy this would not constitute a true flame because neither the sender (Jack) nor at least one thirdparty observer (Bert) overtly interpret his comment as a flame. But what are the moral implications of refusing to code Jacks criticism as a flame in ones research if only Brian (and one outside researcher) saw it that way? This interaction would not qualify as a true flame but certainly Brian, the recipient of the flame, displayed emotional effects that indicated it was as hurtful as a flame. However, by not coding it as a flame because an insufficient number of interactants code it as a flame, the researcher in a sense takes sides once again in an essentially culturally relative interpretation. Such a coding scheme ignores what for Brian are very real effects of the message. Three minutes after this conversation, Brian logged off and for the remaining two hours that I was online observing, Brian did not return. Many explanations are possible. Maybe he went to get a sandwich or do homework. Or maybe he was offended about being publicly criticized for something as minor as how he laughs online.
The previous discussion concentrated on Brians implicit claim that Jack had unduly criticized him. Although he never used the word explicitly, in a sense Brian was accusing Jack of flaming him if we invoke traditional flaming definitions. Brians claim was that Jack was being too precise about Internet Lingo. As discussed above, determining whether or not Jack actually flamed him is theoretically illusive. Further, a particular coding involves making a number of culturally relative judgments that may ultimately belittle or ignore the effects of flames on particular participants. Interestingly, it is arguably true that Jack levels a flame claim of his own at Brian. When Jack accuses Brian of getting pissy too easily in line 21, he is claiming that Brians accusation in line 16 that Jack has been unduly critical is itself too critical, and therefore can be interpreted as a flame claim. Yet is Jacks counteraccusation justified? Is Brian flaming Jack in line 16? To answer these questions using any taxonomy or definition brings up the host of problems just discussed. The problem essentially lies in the concept of flaming itself.
OSullivan and Flanagin (2003) correctly state that one of the major problems with the literature on flaming is that it tends to assume that flaming is exclusively an online phenomenon. The history of the term emerged from popular discourse that expressed concern about talking through a computer. OSullivan and Flanagin (2003) state Flaming as a concept emerged from popular discourse surrounding the online community to describe aggressive, hostile, or profanity-laced interactions via email and discussion groups . Although it is true that in certain contexts, some people say they are flaming someone in a facetoface argument, scholars still tend to use the term flaming to refer to online behavior. This has the effect of predisposing a researcher to seeing flaming as a category that is automatically different from the kinds of hostility that take place offline. Whether or not flaming dynamics are similar to offline hostility is an empirical question. But using the term flaming itself predisposes researchers to see difference where there may actually be similarities. Our example suggests that the interlocutors are exhibiting similar conversational patterns to those that scholars (such as Maynard, Goodwin, and Jacoby and Gonzales) of facetoface arguments observed.
Scholars of conversational morality might also argue that the patterns found in the examples above and below are similar to those found in everyday conversation. Turning once again to the argument between Jack and Brian, we see that Brian implicitly criticizes Jack about being overly critical and Jack levels a countercriticism that Brian has been too critical. Scholars of conversational morality have observed that similar morality accusations and counteraccusations are abundant in everyday talk. They argue that morality is so pervasive in everyday talk and in so many contexts that, it is hardly possible to avoid expressions that do not somehow or other carry a moral meaning . Bergman argues that when a person tries to correct another person in a moral way in talk, the recipient of the correction often offers a countercorrection to deflect the implicit accusation that the corrector is more moral than the person being corrected . In our example, Jack corrected Brian, and Brian in turn levied a countercorrection claiming that Jack had not acted correctly in his criticism. They are exhibiting a wellknown pattern of the forms morality takes in everyday talk. Claiming that this particular interaction is a flame risks seeing the interaction as a specifically online phenomenon, rather than one that may be quite similar to everyday conversational patterns imbued with moral positionings.
A particularly interesting area of research in conversational morality has to do with the use of categories in talk. Labeling someone as a drunkard versus a tippler may connote different moral stances of the speaker to the person he or she labels. Merely using one term over another implies moral connotations and judgments. The same is true of the category of flaming. Using flaming as a category engages the researcher in a moral dialogue with the people they study by taking a stance with regard to whether or not one person has flamed another. The problem is that flaming is an oversaturated term that scholars have used to mean many different things. It becomes very difficult to assess whether researchers can adequately label things as culturally relative as hostility and criticism as flames even if they consider a number of different perspectives, such as that of message senders, receivers, and thirdparty observers.
Even if we examine what the participants themselves label flames (rather than trying to assess whether a certain socalled hostile or critical comment is a flame), the same issues of 1) illusivelyidentified intent; 2) culturalrelativity in judging the existence and violation of norms; and, 3) similarity of morality patterns in conversation apply. For example, in Example (2), we see a familiar moral pattern in which a thirdparty observer, Alexander, claims Max has flamed Victor during a discussion about Windows and Linux. Just prior to the segment below, Victor had criticized Maxs preferred computer platform, which was Microsoft Windows. For many in the Free and Open Source Software (F/OSS) community, choosing a computer platform is a moral choice, since many F/OSS participants see Microsoft as engaging in unethical monopolistic practices that bundle expensive products in ways that reduce quality, artificially increase price, and remove consumer choice. After Victor criticizes Windows, Max invokes a common rhetorical device used in technical communities, which is to deflect criticism by exposing the ignorance of the criticizer. If the criticizer is shown to have insufficient knowledge about what they are criticizing, then for many technologists their criticisms are dismissible. At the beginning of the following conversation (which is actually taken from a long discussion), Victor tries to answer Maxs question about how Windows works. If Victor cannot answer the question satisfactorily, then Victor risks demonstrating having insufficient knowledge to criticize Maxs technological choice.
Example (2) from Community B 1 Victor says, You are right, Max. Autoexec and config are out of date. They arent 2 used. If you install and look at Win NT 4.0, you will see that MS was trying to get 3 rid of it back in 96 as well as now. Those things should be in the control panel. 4 Also, more than likely, 16bit is now toast entirely. 5 Victor says, Meant techies channel. 6 Ben uses SunOS, and its faster, more stable, has better memory management, etc... 7 Matthew says, Its probably one of the other *nixs. 8 Max knows he is right. 9 Ben says, There is no 16 bit code in WinME. 10 Sean says, Victor, that doesnt answer the question. 11 Max did not need to to confirm this. :P 12 Max wants you to explain how Windows handles configs now that theyre gone. 13 Ben says, Who, me? 14 Alexander says, Why, Max? 15 Matthew says, Open Linux is getting close to being like Windows. 16 Victor says, Its all in the control panel. The only DOS you will see is an emulation 17 to make it more compatible with th 32bit programs. 18 Terry yawns as you people obviously dont know that much about windows and yet 19 you seem to have a lot of opinions about it. 20 Alexander says, Why do you want him to explain that? 21 Austinsays, sun is not always faster 22 Sean says, *SHADUP* 23 Max says, He claimed to be a Windows expert, Alexander. 24 Austin is not getting into that flamewar. 25 Matthew says, Open Linux, IMHO sucks. 26 Terry grins at Austin 27 Ben says, Not always, but usualy. :P 28 Matthew says, Maybe itll work correctly in the future. 29 Ben uses both, and prefers SunOS/Solaris. 30 Alexander says, So, basically, your purpose is flaming him... taunting him...and/or 31 making hostile remarks and innuendos in his direction? 32 Max says, How is asking him to explain something about windows a flame? 33 Ben hehs. 34 Max believes youre being quite hostile, Alexander. :P 35 Alexander isnt being hostile, Max. 36 Max is asking him to prove his superior windows knowledge by explaining the 37 main feature WinMe adds over Win98. 38 Alexander is simply trying to decide whether or not he wants to register a formal 39 complaint to the admins against you. (: 40 Ben says, The main feature in WinME is rewritten .vxds. :P 41 Max says, Based on what, Alexander? 42 Max says, Because I support Windows and ask people to prove their knowledge 43 that they claim to have? 44 Max says, Or is it because you were offended by being proven wrong? 45 Alexander says, Based on your hostile aggression against him (your intent, not 46 necessarily your words). 47 Victor says, Both of you shut up for a second. 48 Max has no hostile intent. 49 Max says, If I was hostile, Id be threatening people. :P 50 Sean says, Hes not being hostile, Alexander. 51 Max says, Alexander is just trying to find a way to squelch the opposition. 52 Matthew says, They should realy develop DOS emulation for WindowsME so you 53 can play it within Windows. 54 Ben says, Get a boot disk. 55 Victor says, There is not point to arguing over this. Max has not hostile intent from 56 wht Im seting. He is just trying to figure out wheather or not I know what I am 57 saying. Alexander, you need to get off his case since it will only cause problems. 58 Alexander says, Hostility doesnt necessarily have to be violent. It can simply be 59 meanspirited mockery. 60 Robin says, Alexander, I dont think hes being hostile...hes just being as closed 61 minded as the rest of us. ;) 62 Alexander says, Which is what I see coming from you with consistency.
In lines 14 Victor tries to answer Maxs previous question of how Windows handles configs now that they are no longer part of the Windows operating system. In line 20 Alexander asks Max why he wants Victor to explain this. Since Alexander directly asks about Maxs intent, his query provides an interesting opportunity for a researcher to witness how thirdparty observers attempt to access an interlocutors intent and how they judge that intent. In line 23 Max explains that Victor claimed to be a Windows expert. Max is suggesting that if Victor wishes to critique Windows, he should be knowledgeable enough about Windows to criticize it. I have observed such a strategy used many times in technical and academic conversations and in my view falls within the realm of cultural practices in many technical communities. In his explanation about why he queried Victor, Max tacitly claims that he did not intend to flame Victor. Yet in lines 3031 Alexander does not accept Maxs explanation and proceeds to accuse Max of intending to flame Victor. In this example an insider (Alexander) directly queries the message sender (Max) about intent, but does not accept the senders explanation. As Morgan found in her study of AfricanAmerican women, even when a thirdparty observer (who was known to be male) attempts to access the senders intent directly, the answer may not mirror the observers interpretation of events. Notably, Alexander is not alone in his thirdparty interpretation. In line 24 Austin calls the conversation a flame war, and I myself initially felt Max was being unnecessarily aggressive in his style of talk.
Scholars of conversational morality point out that a conversational correction often yields a countercorrection. The original correction invokes doubt about the morality of the speaker under correction. The speaker may defend his or her position by making a counteraccusation that in turn casts doubt on their wouldbe corrector. In line 32 Max offers a countercorrection of Alexanders accusation by asking Alexander to explain how asking Victor technical questions constitutes a flame. Alexander has accused Max of violating a cultural norm, which is to keep conversation on public channels cordial and avoid socalled flaming messages that attack. Alexander argues that Max has violated a local cultural norm of remaining cordial on a public chat line by aggressively asking Victor arcane technical questions. In turn, Max counter accuses Alexander of inappropriately chastizing Max by accusing him of flaming Victor. Once again, Max is claiming that being unduly critical is itself hostile. In line 34 Max levies a counteraccusation that it is Alexander who is hostile and Max ends his comment with the :P emoticon which signifies a person sticking ones tongue out at ones interlocutor, potentially connoting that his comments are joking, teasing, or playful in nature. This pattern conforms to what conversational morality scholars see as a correction which contains an accusation about one persons morality and a countercorrection that claims the original corrector is out of step with local moral norms.
Alexander and Maxs exchange demonstrates once again the problems of accessing intent and the cultural relativity problem of judging whether a message violated any local communicative norms. In lines 3637 Max once again defends his position and explains that he wanted Victor to prove that he knew enough about Windows to criticize it. In lines 3839 Alexander increases the force of his moral accusation by threatening to register a formal complaint to the administrators of the online group. Registering a formal complaint is a more serious threat because it could result in the removal of Maxs speaking privileges on public communication channels in Community B. In lines 4243 Max asks Alexander if he is chastizing Max not because he has violated a cultural norm, but because Max supports Windows, which was an unpopular computer platform among many participants of Community B. In line 44 Max accuses Alexander of threatening to complain because Alexander was embarrassed about being proven wrong about a supposed critique of Windows. In line 51 Max claims that Alexanders intent in chastizing him resulted not from trying to keep communication cordial, but rather as a veiled way to squelch the opposition to Linux. Max claims that Alexander invokes a supposed local cultural norm in order to censor someone who is using immoral, unpopular technology. In lines 4546 Alexander claims that it was not the actual words Max used but rather the intent of the message that prompted Alexander to code Maxs message as a flame. Alexander claims to have access to Maxs intent that lies beyond the actual words Max used. His access also lies beyond what Max himself claims to be his intent. Alexander directly queries Max about his intent, but he does not accept Maxs answer. Alexander displays the belief that he knows Maxs actual intent, despite Maxs denial of that intent in line 48. Similar to the AfricanAmerican women in Morgans study, Maxs stated intent receives little weight in Alexanders social calculus. Alexander displays the belief that Max is responsible for the result of his actions whether or not he intended to flame Victor.
Thirdparty observers do not unanimously see Maxs comments as flames. Some thirdparty observers, including Victor who was the original target of Maxs supposed flame, defend Max to Alexander and deny that his message was inflammatory. In line 50 Sean tells Alexander that Max was not being hostile. In line 6061 Robin tells Alexander that Max was not being hostile but was actually as close minded as the rest of the group. Robins statement implies that Max was acting well within the limits and understanding of what constitutes acceptable local tech talk — which often includes what some people might label aggressive argumentation tactics that do not yield easily to new or contradictory information. In lines 5557, Victor chastizes Alexander and tells him to get off his case since Alexanders threats and moral corrections will only cause problems. Victor displays disagreement to Alexanders implicit claim that Max has violated local communicative norms. After seeing so many participants rally to Maxs defense, I, as a thirdpart observer, began to question my initial interpretation that Max had violated a local communicative norm. My personal coding was different before and after I observed the community reaction to Maxs statement.
Once again, we see how the process of levying flame claims as well as socalled flames play a role in negotiating actual cultural norms. Alexanders criticism tries to maintain a cultural norm of ensuring a certain level of cordiality in online talk. Maxs defenders take issue with Alexanders criticism suggesting that Alexander is not portraying cultural norms correctly. This interaction becomes a kind of negotiation about what is acceptable for this community in terms of cordiality standards in Community B. Further, the exchange shows a potential negotiation process when norms conflict. Although cordiality norms might dictate (in Alexanders view) that Max has flamed Victor, tech talk norms in Community B might accept what some people (but not others) might label aggressive argumentation. As these norms come into conflict participants use the flames and flame claims to determine which cultural norm should win. Similarly, in the first example, the interlocutors encountered conflicting norms in terms of criticism and potential violations of netiquette. Jacks eventual attempt to rekey his netiquette criticism by accusing Brian of becoming too irritated shows that the netiquette criticism was perhaps not as hard and fast a norm for anyone in the conversation. The result of the exchange with Jacks attempt to rekey his criticism perhaps signals that participants in Community A ultimately choose basic cordiality in individual expression over socalled netiquette norms.
Flame claims and flames are not the result of cultural norm violations but instead provide a window into how participants negotiate cultural norms into and out of existence. The interplay between flame claims and flames is the process by which participants maintain, challenge, and negotiate cultural norms that are constantly in flux and potentially at odds. Rather than adjudicating whether a message is or is not a true flame, a more interesting research question involves examining how participants invoke supposed local norms to make flame claims and how these claims play a role in ordering participants microsocial worlds. In Example (1), we saw how Jack invoked supposed Internet norms in ways that resembled how older children try to microsocially orient themselves hierarchically around age. Whether or not an Internet standard of communication exists, and whether or not Jack actually believes that standard, invoking it to correct someone else may have certain interactional effects (whether or not they are bad or intentional). Jacks correction of Brian and Brians subsequent revision of his talk display Jack as more of an expert on Internet talk than is Brian. Again, whether or not Brian actually believes Jack is an expert is a separate question. But Brians public revision of his talk confirms Jacks ability to correct him. It also confirms that Brian was in need of correction. A microsocial hierarchy is momentarily established in which Jack is coded as more of an expert than is Brian on Internet talk. Yet at the same tine, Brian levies a tacit flame claim at Jack and offers a countercorrection that accuses Jack of being inappropriately critical of another persons talk.
The flame claims in Example (2) among adult and young males revolve around a technical and moral debate about the appropriate operating system to use. Alexanders correction of Max yields a counterclaim that Alexander is not interested in cultural norms but rather is attempting censorship to squelch opposition to support of what the community feels is an immoral operating system. Alexanders direct attempt to access Maxs intent was met with disbelief and rejected. Similar to Morgans observations, a thirdparty insider to the conversation did not see stated intent as particularly meaningful for interpreting the interaction and its effects. Similarly, Alexander also held Max responsible for what Alexander perceived were the consequences of his talk despite Maxs protestations that he in fact did not flame Victor. For Max, Alexanders flame claim had little to do with assuring cordiality online but much more to do with displaying allegiance to Linux and opposition to Windows. For Max, Alexanders intent was to censor Max, whether or not Alexander confirms or denies such an intent. Perhaps some researchers reading this may still believe that Max flamed Victor. Such an interpretation is reasonable and potentially interesting and useful, but hardly objective fact. A number of thirdparty inside observers took issue with Alexanders correction and challenged the idea that Alexander had invoked a wellknown and agreed upon cultural norm when he corrected Max. The flame claims as well as the flames themselves offered a way for participants to negotiate conflicting cultural norms of remaining cordial online but also engaging in acceptably aggressive styles of tech talk. For this subgroup in Community B, the flame claims and flames enabled the community to interactively create appropriate communicative norms.
Yet scholars of childrens arguments, technical arguments, and conversational morality might see quite familiar patterns in the types of socalled flames and flame claims that participants routinely levy at one another.
The flaming doxas foundations, which are by definition unquestioned because they remain unseen in the intellectual background, are now shimmering into view. Researchers have started to discuss socalled flaming in ways that challenge the doxa that many orthodox and heterodox researchers take as objectively true about online interaction. Neither group closely considers the history of the term and the intellectual baggage that the term smuggles in to CMC scholarship. As OSullivan and Flanagin correctly state, researchers typically invoke the term to describe exclusively online phenomena. Yet scholars of childrens arguments, technical arguments, and conversational morality might see quite familiar patterns in the types of socalled flames and flame claims that participants routinely levy at one another. These flames and flame claims have various potential effects whether the participants intend them or not. Participants often invoke supposed norms (whether or not such norms exist and whether or not the participants actually believe them) in ways that potentially create a microsocial hierarchy between participants. In this hierarchy, the person making the correction is displayed as more expert or more moral than the person under correction. Determining what constitutes a true flame is less theoretically interesting. More revealing are flame claims and whether they yield microsocial effects (whether intentional or not) such as situating oneself higher on a microsocial hierarchy than ones interlocutor.
Scholars disturb flaming doxa by questioning it in their analyses. Researchers are increasingly problematizing the very concept of flaming by framing it in quotes, which is a scholarly convention that semiotically indicates that there is a problem with the term; indeed there is no agreedupon and straightforward definition. At this point the term flaming is exhaustively oversaturated. Scholars have associated flaming with everything from curse words (which any linguist or anthropologist can quickly demonstrate is culturally relative in terms of their social uses and reception) to vague notions of criticism, emotion, or hostility. As the data examples above show, all of these concepts are culturally and contextually relative.
Contra OSullivan and Flanagin, the solution lies not with finding a more precise definition for flaming, but for extinguishing the term flaming itself. As OSullivan and Flanagin point out, the term flaming emerged from popular discourse that set online hostility or flaming behavior as something apart from similar behavior observed in offline communication. Any determination about similarities or differences between online and offline hostility is an empirical question that researchers can tackle through observation and analysis. But when the term flaming itself is used before researchers examine data, researchers are already predisposed to seek and bracket phenomena that are by definition different from what they might observe offline.
If the term is so problematic and predisposes researchers to see data in a certain way (as automatically different from offline hostility), why does it continue to appear? What is the value of retaining the concept in scholarly research? One explanation is that flaming is an inside term many in the technical community use to characterize certain online dynamics. Social scientists and other online researchers have continued to use this term partially to display expertise and knowledge to technologists that the researchers are aware of certain insider, online conventions of talk. Yet, continuing to use this term unproblematically risks obscuring the actual dynamics of online phenomena and predisposes researchers to see differences between online and offline behavior where they may or may not exist.
Rejecting the term flaming in no way suggests socalled flames or hostile messages have no effects. Rather, the opposite is true. By labeling one thing a flame and another not a flame, scholars may unwittingly engaging in a moral categorization that takes sides with particular participants in ways that might deny real effects (such as a participant displaying emotion over what a researcher would deny was a true flame). The term itself means too many things to be useful at this juncture. A more productive approach is to examine the interplay between specific accusations of hostility, criticism, norm violation or other relevant phenomena under scrutiny and analyze the cultural and social consequences that such moral struggles have for the participants.
About the author
Patricia Lange received her Ph.D. in anthropology from the University of Michigan. She is currently a Postdoctoral Fellow at the Annenberg Center for Communication at the University of Southern California. She is studying how children and youth use digital media, particularly creating, viewing, and exchanging videos on the Internet. Prior to joining the Annenberg Center, Dr. Lange was a visiting scholar in linguistics at Stanford University and a lecturer in anthropology at San José State University. She has also worked as a technology analyst at SRI International where she studied technical and marketing trends in advanced computer interfaces such as virtual environments.
Email: plange [at] annenberg [dot] edu
I would like to thank the participants of Community A and Community B for allowing me to observe and participate in their activities. I very much appreciated and enjoyed learning about them and their communities. I also wish to thank all of the people who agreed to be interviewed for the study. I learned a great deal about community, conversation, and identity by observing two very fascinating and welcoming communities.
J.R. Bergmann, 1998. Introduction: Morality in Discourse, Research on Language and Social Interaction, volume 31, numbers 3 & 4, pp. 279294.
P. Bourdieu, 1977. Outline of a Theory of Practice. Translated by Richard Nice. New York: Cambridge University Press.
M.J. Culnan and M.L. Markus, 1987. Information technologies, In: F.M. Jablin, L.L. Putnam, K.H. Roberts, and L.W. Porter (editors). Handbook of Organizational Communication: An Interdisciplinary Perspective. Newbury Park, Calif.: Sage, pp. 420443.
A. Duranti, 1992. Intentions, Self, and Responsibility: An Essay in Samoan Ethnopragmatics, In: J.H. Hill and J.T. Irvine (editors). Responsibility and Evidence In Oral Discourse. Cambridge: Cambridge University Press, pp. 2447.
M.H. Goodwin, 1990. HeSaidSheSaid: Talk as Social Organization Among Black Children. Bloomington: Indiana University Press.
S. Jacoby and P. Gonzales, 1991. The Constitution of ExpertNovice in Scientific Discourse, Issues in Applied Linguistics, volume 2, number 2 (December), pp. 149181.
S. Kiesler and L. Sproull, 1992. Group Decision Making and Communication technology, Organizational Behavior and Human Decision Processes, volume 52, pp. 96123. http://dx.doi.org/10.1016/0749-5978(92)90047-B
S. Kiesler, J. Siegel, and T. McGuire, 1984. Social Psychological Aspects of ComputerMediated Communication, American Psychologist, volume 39 number 10, pp. 11231134. http://dx.doi.org/10.1037/0003-066X.39.10.1123
P.G. Lange, 2003. Virtual Trouble: Negotiating Access in Online Communities, Ph.D. dissertation, University of Michigan.
M. Lea, T. OShea, P. Fung, and R. Spears, 1992. Flaming in ComputerMediated Communication: Observations, Explanations, Implications, In M. Lea (editor). Contexts of ComputerMediated Communication. London: Harvester Wheatsheaf, pp. 89109.
H. Lee, 2005. Behavioral Strategies for Dealing with Flaming in an Online Forum, Sociological Quarterly, volume 46, pp. 385403. http://dx.doi.org/10.1111/j.1533-8525.2005.00017.x
D.W. Maynard, 1985. How Children Start Arguments, Language in Society, volume 14, pp. 130. http://dx.doi.org/10.1017/S0047404500010915
M.H. Morgan, 1991. Indirectness and Interpretation in African American Womens Discourse, Pragmatics, volume 1, number 4, pp. 421451.
P.B. OSullivan and A.J. Flanagin, 2003. Reconceptualizing Flaming and Other Problematic Messages, New Media & Society, volume 5, number 1, pp. 6994. http://dx.doi.org/10.1177/1461444803005001908
B. Pfaffenberger, 1997, Websters New World Dictionary of Computer Terms. Sixth edition. New York: Que.
C.U. Shantz, 1987. Conflicts Between Children, Child Development, volume 58, pp. 283305. http://dx.doi.org/10.2307/1130507
J. Siegel, V. Dubrovsky, S. Kiesler, and T.W. McGuire, 1986. Group Processes in ComputerMediated Communication, Organizational Behavior and Human Decision Processes, volume 37, number 2, pp. 15787l. http://dx.doi.org/10.1016/0749-5978(86)90050-6
R. Spears and M. Lea, 1994. Panacea or Panopticon? The Hidden Power in ComputerMediated Communication, Communication Research, volume 21, number 4 (August), pp. 427459. http://dx.doi.org/10.1177/009365094021004001
D. Tannen, 1999. The Argument Culture: Stopping Americas War of Words. New York: Ballantine Books.
Paper received 13 July 2006; accepted 20 August 2006.
Copyright ©2006, First Monday.
Copyright ©2006, Patricia G. Lange, All Rights Reserved.
What is your claim to flame? by Patricia G. Lange
First Monday, volume 11, number 9 (September 2006),
A Great Cities Initiative of the University of Illinois at Chicago University Library.
© First Monday, 1995-2017. ISSN 1396-0466.