A convergence of events and reading has me thinking about trolls. First, I recently began reading Bill Eddy's High Conflict People in Legal Disputes, which I think will be relevant to conflict management (online or otherwise) and trolls. Second, in reviewing the literature on gender dynamics in online community I've been reading a lot of excellent work from Susan Herring, who with her co-authors wrote Searching for Safety Online: Managing "Trolling" in a Feminist Forum (HerringEtal2002sso). Therein they follow Andrew (Andrew2010tf) in specifying three criteria for trolls: sending messages (1) that appear outwardly sincere, (2) are designed to attract predictable responses or flames, and (3) waste a group's time by provoking futile argument. Third, a natural question arising from my focus on good faith collaboration is the flipside, which includes trolls. Finally, I recently read Penina Shachaf's and Noriko Hara's Beyond Vandalism: Wikipedia Trolls (ShachafHara2010bvw).
In the latter work, the authors interview 15 admins from the Hebrew Wikipedia and conclude their trolls can be likened to hackers (i.e., script kiddies). Trolls are characterized as "engaged in intentional, repetitive, and harmful actions;... that violates Wikipedia policies;... have high interest and destructive involvement within the Wikipedia community;... work in isolation under hidden virtual identities." Specific problematic behavior includes argumentative writing, cursing, aggressive threats, offensive content, and sock puppetry. A worthwhile contribution, no doubt, however I felt the section on troll motivations needs more qualifications since it is dependent only upon administrator interviews. I found the placement of trolls in the context of hackers to be odd given the many meanings of the word. And I think the interesting problem is not how to distinguish a troll from a "hacker" but from a well intended but exceedingly difficult person -- a "high conflict person" in Bill Eddy's book. Finally, my attention was attracted to the findings that trolls are not typically ideologically motivated and work in isolation.
The authors are careful to note that their findings are not necessarily transferable to other situations. Accordingly, in my experience I think I have seen fiscally and ideologically motivated trolls. Also, a collective form of trolling seems to me quite common in the Web 2.0 (crowd-source) context, which I tend to refer to as a "troll-plex." This means that the behaviors typically attributed to the individual troll may now be decentralized among a group of participants who may individually exhibit a spectrum of behavior, but together share a target, a culture (perhaps mean-spirited "lulz") and/or venue where they can discuss and/or coordinate activity which in sum satisfies all the criteria of trolling by a single individual -- but with perhaps an even greater impact. For instance, the attacks upon Kathy Sierra ranged from critiques, to mean-spirited attacks, to scary and illegal threats, but one could find all these participants discussing their behavior among themselves, such as at the "mean kids" site. I found Herring et al.'s discussion of the feminist troller "Kent" particularly relevant to Web 2.0 trolling if one imagines that varied people enacted different aspect of Kent but also discussed and reinforced their collective behavior:
Kent had the outward manifestations of sincerity, implying that he will modify his behaviors if people will engage and explain their concerns to him. "In these messages, Kent attempts to present himself as rigorous and principled as regards the rules of debate, and potentially cooperative if others meet his conditions (providing proofs, answering his questions, etc.) rather than provoking response through conscientiousness, he appeals to others to respond in good faith. He attempts to appear outwardly sincere (criterion 1 of our definition of trolling) as a means of drawing others into responding." And he further attempts to provoke futile argument; his rhetorical strategies show a fundamental uncooperativeness and perverseness in rhetorical strategies including: refusing to acknowledge others' points, willfully misinterpreting others' motives and views, and attacking others for ignoring him. The paradox is that he is expending considerable effort to be read and responded to, but taunts women who do not respond as weak. Finally, he enacts ideological manipulation: if they permit him to stay, their community is polluted with antifeminist harassment, if they ban him they risk being labeled censorious (HerringEtal2002sso).
Barry Kort on 2010-11-05
Since you released my last comment without responding to it, I'm tempted to imagine that you're still thinking about it. If so, I sincerely hope that you will drill down your thinking to ever deeper levels, Joseph, and then share with us your newest thoughts on recognizing good-faith critics who, on occasion, are unfairly dismissed as trolls of the nefarious variety
@Joseph: "Of the three conditions above (appear sincere, designed to evoke inflamed response, and disrupt the group) the second is the most difficult as it speaks to motives/intentions, which is difficult to claim."
Oh it's easy enough to claim, but then it's common to make claims which are ungrounded flights of fancy, especially when it's a claim about the other guy's state of mind, beliefs, desires, or intentions. There is a fine line between a provocative and thought-provoking question and an inflammatory one. Those of us with years of experience in academia still struggle to craft provocative questions which promote insightful thought without inflaming those whom we are sincerely asking to stop and think ever more deeply.
Theorizing about an adversary's intentions is a popular way to aggravate and antagonize a good-faith critic whose original intention is to work collaboratively toward the development of better organizational practices. I wonder how many trolls of an antagonistic nature evolved into that unbecoming nature when a constructive and collaborative overture went sour.
Barry Kort on 2010-11-03
Socrates was considered to be a gadfly. Would someone today, who behaved like Socrates, be labeled as a troll?
King Henry considered Beckett to be a "meddlesome priest." Would someone today, who behaved like Beckett, be labeled as a troll?
Pope Urban was not too thrilled with Galileo, either. Would someone today, who behaved like Galileo, be labeled as a troll?
I have an idiosyncratic definition of troll. I define a troll as someone who asks good questions that one doesn't want to have to answer.
Most unwanted questions are not good questions of the sort that need to be asked. A good question makes one stop and think. Not everyone appreciates having their thinking arrested by a good question of that sort.
My observation is that WikiCulture tends to mislabel people as trolls, even when they are sincerely asking good (if uncomfortable) questions that need to be asked if an organization is to evolve in a positive direction.
Peter Damian on 2010-11-05
Criterion (1) is a necesssary but not sufficient condition. It's not enough to appear outwardly sincere, it's crucial that the sane, clever and reasonable people on the forum (i.e. the other trolls) instantly see the point of the troll. It's the clueless or humourless or too-earnest naifs that the troll is aimed at.
Barry Kort on 2010-11-04
@Joseph, I don't know if you are old enough to have been a fan of the TV series, Columbo, starring Peter Falk, but he played a character who might be likened to a "troll" in the sense suggested by Peter Damian, above.
At all times in human history, there are widely held beliefs which are simply incorrect, and which become ripe for overthrow. Occasionally, these are highly controversial, as in the well-known cases of Galileo and Darwin. In WikiCulture, there are comparable examples of considerably lesser import, but nonetheless controversial enough to become the subject of your research.
It occurs to me that it would be instructive to look at those upstarts who had the temerity to overthrow popular consensus, when that consensus was sadly misguided. What techniques were employed, and what defenses did the misguided majority employ to ward off the upstarts who would overthrow their sacred (if mistaken) beliefs?
Did Puck and Columbo know what they were doing?
Would their approach have worked in WikiCulture?
Seth Finkelstein on 2010-11-04
@Joseph - One has to be very careful about constructions such as "... that one can participate in a larger group phenomenon". There's a danger of broad-brush smears or guilt-by-association. In fact, that's exactly what was so dangerous about the Sierra incident. One of the accused wrote - and remember, these are his words, not mine - "I think her response, as it pertains to anything I personally wrote, was unjustified -- but highly effective -- character assassination. As a result, I'm sure I'll be explaining for years to come that I'm not really an ax murderer and child molester. Nice work."
Or, from another direction, I'm sure you would not accept a characterization of been part of a con-job huckster "larger group phenomenon" dishonestly selling Wikipedia, even if you're not dishonest yourself. I'm absolutely certain you'd say your own work should not be tainted by anyone else's misrepresentations or rewrites of history for commercial motivations. Right?
Which is not to deny institutional incentives. But you should be fair in application. If there's a troll-plex, there's also a hype-house.
Kelly Martin on 2010-11-04
"engaged in intentional, repetitive, and harmful actions;... that violates Wikipedia policies;... have high interest and destructive involvement within the Wikipedia community;... work in isolation under hidden virtual identities" describes not only classic trolls but also both the highly aggressive point of view pushers that dominate some areas of Wikipedia discourse, and many Wikipedia administrators. In the latter two cases, however, it is often not broadly recognized that the behaviors in question are harmful. A more accurate statement is that the identification of such behaviors is actively suppressed because the actors are socially powerful enough to avoid sanction.
Wikipedia's internal use of "troll" is idiosyncratic; at Wikipedia a "troll" is anyone who refuses to quietly accept the "consensus" decreed by Wikipedia's elites. This technically could be described as failing to follow Wikipedia policy, as Wikipedia policy is whatever its elite says it is (notwithstanding what the published policy documents might say), but it's not what most people think of by a troll. This especially applies if the individual speaks in opposition to the will of the elite more than once, persists in forwarding an argument after an elite has made it clear that it will not be accepted, or does anything to try to increase public awareness of the dispute.
The key point here is that one can be described as a "troll" for violating "consensus" long before a reasonable person would believe that a consensus had been formed, and the epithet "troll" is routinely used in Wikipedia dialogue as a epithet to marginalize and exclude those whose opinions differ from those preferred by the elite.
Joseph Reagle on 2010-11-04
btw: I've also been speaking to Luke Simcoe on this topic and agree with his sentiments  in terms of the difficulty of properly situating trolling in the realm of human behavior, especially in not seeing hacking and trolling in the same genus of person.
Joseph Reagle on 2010-11-04
@Barry: I appreciate your point that the term troll is sometimes misapplied, but I wouldn't advocate a move towards idiosyncratic definitions as it just makes it all the more confusing. As I note in the book: "Obviously, those who cannot appreciate the relative weight of well-supported claims (i.e., the consensus of peer-reviewed research supporting evolution) will have a difficult time at Wikipedia. However, I would not actually consider such contributors as "trolls." (While the term has taken on a general pejorative function, trolls properly signify those who post controversial or irrelevant messages with the intention of disrupting an online community.)" This means we need to consider our definitions more clearly. Of the three conditions above (appear sincere, designed to evoke inflamed response, and disrupt the group) the second is the most difficult as it speaks to motives/intentions, which is difficult to claim. I'm of the opinion that there are people that are sincere, have frequent intrapersonal/intergroup difficulties, and whose contributions are not constructive but disruptive to a group. I'm beginning to think one should use the term "high conflict" person/behavior over "troll" in such cases, but it is a fuzzy boundary.
@Seth: the point of the notion of "troll-plex" is to indicate that one can participate in a larger group phenomenon that is troll-like (or harassing) even if not every constituent member is trolling or harassing.
An interesting question to ask is: had I participated in the initial responses to Gorman -- who noted harassing responses and was trolled as well -- would I have been part of the troll-plex? While he had experienced a troll-plex, I did not participate since I didn't meet criteria 3.
1. a varied/decentralized group of participants who may individually exhibit a spectrum of oppositional behavior (appreciated critique to threats);
2. share a target;
3. share a culture (perhaps mean-spirited "lulz") and/or venue where they can discuss and/or coordinate activity with those elsewhere in the spectrum;
4. which in sum satisfies all the criteria of trolling by a single individual.
Of course, this is just some thinking-out-loud, initial musing on a blog.
Peter Damian on 2010-11-04
In defence of the Troll. The character of the troll is fascinating. I had many usenet encounters with the late Torkel Franzen http://en.wikipedia.org/wiki/Torkel_Franz%C3%A9n. He was both a gifted mathematician (he wrote a very accessible book on Godel's theorem) and a self-described troll. He wrote an excellent guide to trolling which unfortunately has been taken down from his university site, but I have a copy somewhere. His technique was to take on usenet cranks of the physics and mathematics variety ("cantor was wrong") and gently, sometimes viciously and cruelly tease them by appearing outwardly sincere and encouraging then gradually leading them towards a horrible black hole of self-contradiction. The purpose was partly to amuse those who were in on the joke. But to those who knew him well, he was a kind and well-intentioned person whose purpose was to teach and to educate above all else.
Another site explains the same technique as follows. "The well-constructed troll is a post that induces lots of newbies and flamers to make themselves look even more clueless than they already do, while subtly conveying to the more savvy and experienced that it is in fact a deliberate troll. http://www.urban75.com/Mag/troll.html "
Did trolls exist before the internet? Are they one of the basic human character types, who have existed since humans existed, or are they a Web 2.0 phenomenon? I suggest the former. Trolls (under different names) are a stock character in literature. The character of Puck is a shrewd and knavish sprite who causes minor catastrophes and embarrassing situations. There is also the stock literary relationship between the well-intentioned but naïve and self-important or buffoonish master versus the shrewd and astute servant. Wotan/Loki, Don Quixote/Sancho Panza, Mr Pickwick/Sam Weller, Wooster/Jeeves. Not forgetting Sergeant Wilson and Captain Mainwairing, for those who can remember!
A famous pre-internet troll is still in print http://www.amazon.co.uk/Henry-Root-Letters/dp/0297777629 .
From the point of view of the naïve, the troll can often appear disruptive, vicious and dangerous. "you must understand that in the last days scoffers [empaiktes] will come, scoffing and following their own evil desires (2 Peter). But the troll has an important and vital social function : to puncture self-importance and self-delusion, using his (or her) standard weapons of humour (sometimes cruel), irony.
So don't knock trolls! They have an important social purpose. The world would be a worse and more dangerous place without them. And a lot less funny.
Seth Finkelstein on 2010-11-03
Regarding "... ranged from critiques, to mean-spirited attacks, to scary and illegal threats, but one could find all these participants discussing their behavior among themselves, such as at the "mean kids" site."
This is very mistaken, and though formally not libelous, it has that flavor. NONE of the "mean kids" engaged in "scary and illegal threats", and some were furious at being slimed that way. See my column on the incident:
"Accusations of sex and violence were bound to grab the headlines"
I assume you're merely echoing the common inaccurate account. However, this should give you pause in terms of doing scholarship.
For Wikipedia, it would be a strawman were I to contend that the words "critic" and "troll" are synonyms. But too many Wikipedians do have a pronounced tendency to dismiss criticism - or, bluntly, expressions that one dislikes their institution or behaviors - as trolling.