Flame wars onlineCyberspace has captured the attention of America. The latest issue of Time magazine is devoted to this topic. Very little in the way of academic research on the social dynamics of cyberspace communications has been done. One of the things that is known, but not fully explained, is the preponderance of argumentative notes. “Flame wars” is the term used for this, and all newcomers to cyberspace are warned about the inflammatory nature of notes they are likely to see. People who have never seen or heard each other take the anonymity of cyberspace as an excuse to be rude in ways that one does often not see in real life.

Recent research on the cognitive and affective bases of behavior may shed some light on this phenomena. Theoretical models that describe the relationships between mood and behavior are increasingly being used to conceptualize such issues as one’s social judgment of self and others. Research findings, although not unequivocal, tend to support the notion that there is an interaction effect between one’s mood and the way in which social judgments are made. Stated in its simplest terms, happy people make more positive judgments, and sad people make more negative ones. This paper will examine these research findings, in the light of the theoretical conceptual ANT model. The phenomena of flame wars will be addressed, according to these findings. The likelihood of flame wars being a result of sad, self-focused people using the anonymity of cyberspace for such uninhibited mood outlets will be examined.

The Associative Network Theory (ANT) is a set of assumptions that conceptualizes the mind and its memory functioning as a systematic network of interlinked nodes. It states that emotion provides the fundamental way in which memory is organized. Information that is affectively congruent to the contents of a memory node are more accessible to it. In this model, what is available to become conscious thought is dependent on the links between nodes that are currently activated above a threshold limit, and other nodes that are affectively congruent. A node is activated above threshold by input from other nodes. This spreading activation of nodes is conceived of as being responsible for the coding and later retrieval of new information. That emotion nodes are so strongly interlinked with memory nodes is used to account for the fact that cognition, and resultant behavior, are often found to be congruent with induced or measured mood.

Mood priming is the term used to describe the selective likelihood of a mood influencing cognition. Those memory nodes that are linked to a currently active emotion node are primed, and are more easily accessed than others. Experimental research into the relationship between mood and evaluation of others support the idea that these links are based on affective congruence. One such experiment induced a mood in subjects by manipulating feedback about their performance on a bogus test. The happy people were significantly more likely than the sad people to make positive judgments of people when presented with descriptions of people to evaluate. These descriptions were balanced in their content of positive and negative descriptors (Forgas, 1987). The ANT conceptualizes this as an association bias. The memory nodes available for cognitive judgments (evaluative nodes) are primed by their being affectively congruent (thus linked) to the current mood being experienced. If the target of an association is ambiguous (no strong bias for a judgment for or against) as is the case with the evaluation of the balanced written description of an unknown stranger, the mood congruent memory nodes will be activated above threshold first.

Flame wars most often erupt among strangers. Newcomers to Internet discussion groups (newbies), and new members of a particular group are often the source of, or the target of, inflammatory messages. Some research findings have suggested that the congruence observed between mood and judgments is due to the current affective mood of the subject taking up memory capacity, leaving less room for consideration of, or attention to, salient details (Ellis, 1991). The ANT model accounts for this in its descriptions of emotional valence as the fundamental way memories are stored. It is a given that when people are presented with a new person or situation, and are called to judge it, they have only their memory of similar events or people to go on. It is apparent then that sad people who read notes from strangers on their computer screens are the ones most primed to make, and act upon, negative judgments of the authors of the messages.

To fully understand the phenomena of flame wars, we must account for the inhibition of normal social norms. The prohibitions against being rude, unkind and argumentative when in a social situation are strong. One of the moderator variables that have been proposed to account for inconsistent findings of congruency in the mood and judgment research is that these effects are not found when such things as established social norms dictate how judgments should be made (Clark, 1991). The act of public typing as a social interaction is a very new phenomena, one people face for the first time by the thousands every day. Established social norms are not a factor for newcomers. It is newbies who act out most often, and express disrespect for others and show disregard for the feelings of others in cyberspace. This explanation fits with the ANT model and mood congruent research, for flame wars (started by sad people, new to the environment) make judgments of others in a mood congruent way, unencumbered by knowledge of social norms.

The lack of the normal social clues, such as tone of voice and facial expressions, that moderate and guide much of social interaction, and would form the bases of social judgments, are completely lacking in cyberspace. One consistent finding from the mood and judgment research for the congruent effects of a sad mood on social judgments is that congruency exists when the sad mood is a general one (not a transient emotion) and when the situation to be judged contains lots of ambiguity (Forgas, 1990). The ANT model accounts for this finding because of the strengthening of links between nodes that occur with use. A dysphoric mood, as opposed to a state of anger, will have more well established links to affectively congruent memory nodes, causing negative judgments to be more likely to occur. Since the lack of established social norms and the ambiguity of the guides to behavior in cyberspace can help account for flame wars, then it remains to predict the type of person most susceptible to this behavior. We see now that a dysphoric mood is a good candidate.

Mood has been shown to impact the attention process. What one choose to pay attention to, the selection of what is salient, is often found to be mood congruent. In a study that induced happy and sad mood by exposure to music, subjects demonstrated a marked difference in the amount of time they took to recognize a word that described their mood, as opposed to words that did not (Niedenthal, 1994) The ANT model clearly indicates that such a preference for mood congruent input to become silent should exist. The way that cognition is dependent on the currently active emotion nodes dictates this. In terms of explaining flame wars, this factor takes on added significance, due to the frequent complaint of may cyberspace participants that they are overloaded with information to sort through. A moderately active email group, normally accounting for 3 to 6 pieces of mail a day, can become suddenly very active, causing increased pressure on participants to choose which notes to pay close attention to. This is particularly true when a flame war is going on. The happy, contented person will quickly delete the inflammatory, rude remarks he finds. The sad, dysphoric person, primed to pay attention to negative stuff, will react much stronger.

Research on the effects of mood offer one other factor, which is a good candidate to help account for the phenomena of flame wars. The occurrence of altruistic behavior is mediated by mood. Results from studies with children are the most consistent. Happy kids are helpful ones. Sad kids do not engage in prosocial behavior as much. These findings are not so clear as they relate to sad adults. Research supports all cases. Some find more prosocial behavior from sad adults, some find less. It has been suggested that this difference between kids and adults can be explained by looking at self focus as a moderator variable (Carlson, 1987). Adults who are self-focused (like children are all the time) are less likely to cope with a sad mood by engaging in prosocial behavior. The ANT allows for a node (and thus cognitions) to be activated in many ways. A self-focused sad person would have more internally co-active nodes currently engaged than an other-focused sad person. This cognitive as well as affective bias results in self focused sad people passing on the opportunities to help others. Normally, it would be a big leap of faith to connect a lack of prosocial behavior with the overtly negative rude behavior seen in flame wars. However, social relationships in cyberspace groups are based only on the attention, as explicitly expressed in text, that one member pays to another. Public expressions of agreement with another’s views is the element that creates social cohesion and the feeling of community. The findings of a decrease in prosocial behavior among sad, self-focused people can be interpreted, for cyberspace communities, as a propensity toward unkind behavior, because the option of simply ignoring others is not as valid in cyberspace. Many people do just “lurk” in cyberspace groups, never inputting at all to the discussion. For those people who are currently engaged in active cyberspace social relations, the effect of a sad, self-focused mood is likely to be expressions of disrespect basic nature of the relationship formed.

To summarize, findings in current research on mood and cognition as a bases of behavior suggest that evaluation of others, memory access and attention selection all show mood congruent effects in relation to judgments of self and others. The desire to help others, when self focused, becomes problematic for sad people. The ANT helps conceptualize these factors and the relationships between them. A sad mood and a self-focused frame of reference has been show to be a possible factor to help account for the relatively new phenomena of people being rude to each other, by way of computer-mediated communications to social groups.


Ellis, H. C., & Asbrook, P. W. (1991). The “state” of mood and memory research: A selective review. In D. Kuiken (Ed), Mood and memory; Theory research and applications (pp 1-21). Newbury Park, CA: Sage.

Niedenthal, P. M., & Setterlund, M. B. Emotion congruence in perception. Personality and Social Psychology Bulletin, 20, 401-411.

Mayer, J. D., Gaschke, Y. N., Braverman, D. L., & Evans, T. W. (1992). Mood-congruent judgment is a general effect. Journal of Personality and Social Psychology, 63, 119-132.

Clark, M. S. Moods and Memory: Some issues for social and personality psychologists to consider. In D. Kuiken (Ed), Mood and memory; Theory research and applications (pp 1-21). Newbury Park, CA: Sage.

Carlson, M., & Miller, N. (1987) Explanation of the relation between negative mood and helping. Psychological Bulletin, 102, 91-108.

Forgas, J. P., & Bower, G. H. (1987). Mood effects on person-perception judgments. Journal of Personality and Social Psychology, 53, 53-60.

Forgas, J. P., Bower, G. H. & Moylan, S. J. (1990). Praise or blame? Affective influences on attributes for achievement. Journal of Personality and Social Psychology, 53, 809-819.

Copyright © 1995 Storm A. King. All rights reserved. Published here by author.