Drawing Lines: Trolling and Concept Formation

For weeks now on one of the most populated role-playing realms in World of Warcraft, a user has been hanging out and repeating explicit sexual scenarios in the trade chat channel, which is the conversation channel that has the largest number of participants.  What this looks like is a paragraph of text describing a scenario.  A role-playing realm is one where people play the game as their characters, rather than being themselves playing the game.

Initially, this user’s sexual scenarios were mostly bestiality scenarios.  They were icky, but ignorable.  A few people reacted, asking him to stop, reporting him to moderators, and trying to shame him into silence, but most people just ignored him.

The people ignoring him were following the age-old Internet wisdom, “Don’t feed the troll.”

Trolling is an area that researchers are increasingly interested in studying and it is a subject that overlaps with a range of online research.  All of the places I have studied contained trolls, but trolling has never been something that I considered in a systematic way.  Recently there has been an increasingly heated exchange on the Association of Internet Researchers (AoIR) listserv about trolls and trolling.  My thinking about trolling and the first iteration of the trolling discussion on the AoIR list led to some ideas I’d never fully formulated before about trolling.  It also led me to start reading some of the excellent work being done focusing on trolling specifically, some of which I’ll list at the end of the post.

For those unfamiliar with trolling, online, a troll is a user who tries to goad others into responding to him.  In the simplest sense, the troll is someone who is trying to get attention.  The most desired type of response is an emotional one.  Usually this means that the troll posts inflammatory, offensive, off-topic, or ludicrous material in the hope of generating an emotional response.

The troll is acting for his/her own amusement or, in some cases, the amusement of others.  Generally, this would mean that the troll does not actually believe what he has posted; the purpose would be only to elicit a response for entertainment’s sake.  The term is also popularly used to describe people who generate a strong, but pointless, response in others for the sake of attention.  For example, I’ve read people repeatedly refer to the Westboro Baptist Church members as the “ultimate real life trolls.”

The user posting disturbing scenarios in World of Warcraft was using just one of many trolling strategies—that of posting shocking content for a reaction.  This is a common, fairly simple, strategy.  Another common approach is using blatantly racist language.  But, there are also others.  There are people who engage in political arguments, people who post illogical (but usually infuriating) comments, bait and switch pranks, people who try to get people to do something, people who try to appear like allies but who are trying to undermine a group or user, and a lot of other things that fall into the category of goading or tricking people into responding.  For example, a famous “bait and switch” trolling tactic is “Rickrolling” someone—in which, User A tricks User B into clicking a link that User B believes is about one topic.  The link instead will lead User B to a video of Rick Astley singing “Never Gonna Give You Up” – which you can watch here.

Other researchers have created distinctions between types of trolls.  For example, there is work about “griefers” in video games.  A griefer is someone who harasses another player in an online game.  Possibly the most famous griefing incident when a group of people “crashed” and destroyed an in-game funeral in World of Warcraft.  The original video is pretty grainy, but someone cleaned it up and you can watch it here.

I have been reading a lot about methodology in the last few weeks as I attempt to sketch out a draft about the comparative method and online research.  While reading about concept formation and the role of concepts in comparative research, I have been mentally drawing on the example of trolls to help refine my own understandings of the technical intricacies of research methodology.  One of the loveliest articulations of concept formation I read was by Howard Becker (1998).  He points out, among other things, that concepts are relational.  In a discussion of social class Becker states (p.133):

Terms like ‘middle class’ or ‘working class’ only have meaning in relation to one another or to the ‘upper class,’ and the meaning is the character of the relationship…That seems obvious enough.  But it’s one of those obvious things that people acknowledge and then ignore.  How do they ignore it?  By imagining that a class, by having a characteristic culture or way of life, would be what it is no matter what system of relations it was embedded in.

He argues that concepts need to be put into the context of “the full set of terms they imply.”  (p.138)

I read this and then counted my new gray hairs.  How do I, as a self-described ethnographer, reconcile the close study of a given set of relations in a particular place with a comparative social science that demands a level of generalization?  Becker has advice: “forget the name entirely and concentrate on the kind of collective activity that is taking place.” (p.144)  Indeed, much like the “political elite” that is present in every human society and which is often an important aspect of a comparative study of a set of geographical-institutional spaces, trolls exist in many (most?) online communities.

In fact, much of the work done on trolls does treat them as a specific category of actor in all online spaces and/or as a type of phenomenon that emerges from online spaces. (Some researchers look at the psychology of the individual, but that is not my central interest here.)  But is a troll the antithesis of a normal or good user?  Or might a troll fit into an existing category of behavior(s) offline?

As I’ve been mentally framing literature on qualitative research using trolls, I have been reading some of the work that was brought up in the AoIR conversation, and I believe that the argument on the list, as well as the literature, serves as an excellent example of a conversation forming the parameters of a concept.  This is not to ignore that trolls and trolling have been discussed since the early days of the Internet—but rather to point out that one of the amazing aspects of studying the online world is the way in which so much is still open for conversation.  Although, this type of negotiation would be the case with any key concept.  Goldstone (2003) argues that in the case of revolutions the definition of “revolution” developed over time in response to our understanding of revolutionary events and that our understanding of revolutions has changed as more “empirical data” (revolutions) have become available.   And, an aspect  of the current debate on trolling is still about what behaviors “count” as trolling and what overall meaning can be ascribed to trolling is.

For example, one of the responders to the AoIR thread said that when a troll produces a positive outcome she is not usually then considered a troll.  Trolling can be nasty, but it can also be fairly benign, sometimes more like an entertaining practical joke or performance than a mean trick.  For example, I have a dear Internet friend who used to periodically post a thread on a forum we both used to frequent.  He would create a user name that was based on a universally hated pop cultural product (such as a movie) and then he would post a thread that was some variation on the following: “My sister and all her cheerleading friends are skinny dipping in our pool.”

He posted this knowing that immediately someone would ask him to post pictures.  When he was asked to post pictures he would tell people that he had to find the camera – and then he would tell them that he had the camera but he was afraid that the flash would alert his “sister” that he was taking photos.  There was a timing to this that he had developed—in that, he knew exactly how long to wait to post another question to his audience, he knew how to phrase the first post and thread title to capture attention, and he knew how to side-step the early accusations that he was a troll.

As the thread played out, he would tell them he’d managed to take photos but that he didn’t know how to upload them to his computer.  He would post a link that was obviously incorrect, something like, “C:/Program files/Photos/swimming.jpg.”  Because he would do this every few months, there were always users who knew it was a joke—and these people would often post in the thread saying that he was a troll.  But, as long as more people responded to him genuinely, the thread was a success.

Trolling can also lead to positive outcomes, such as annoying people who deserve to be annoyed or rallying people to stand up to a bully.  For example, a close friend of mine is fond of “trolling” abusive people in player-versus-player environments.  In most losing games in a player-versus-player environment, there is a point when someone begins insulting other players.  Sometimes this is targeted at a “weak link” and sometimes it’s more generalized abuse.  The insults are usually offensively phrased and most people do not respond to the insults—even when they are targeted at a single player.  However, my friend often responds by trolling the abuser.  An exchange might go something like this:

Random player: u all suck.
Friend: your game play is matched by your sparkling wit and intelligence.
Random player: shut up, f_ g.
Friend: you can’t win me over with sweet talk.

Very frequently, the original insulter will become incredibly angry that he is being challenged.  My friend is explicitly trying to generate an emotional response often from someone who is verbally bullying another player and he defines this as trolling, saying, “I can’t help but troll these people.”  My friend never uses offensive language or shocking terms; rather, he just intentionally “smartasses” the person into a rage. In many cases, my friend’s “trolling” seems to empower silent players to stand up to the original insulter.  The outcome is normatively positive, but he defines his actions as “trolling.”

Gerring (1999) points out that it is important for a well-formulated concept to be “intuitively clear” or “make sense.”  (p.368) He points out that familiarity is an essential component of a well-formulated concept; otherwise, the concept will not be understood or remembered.  And, I am clearly now working my way towards the argument that the “troll” is more than just negative behavior—and that the relational terms that we deal with trolls through may need to be interrogated.  Trolls have a bad reputation—and in many cases this is justified, as much of their behavior is pretty despicable.  Even in cases of more mild-mannered trolls, the subterfuge and the fact that the person being trolled is being tricked for the amusement of the troll and other observers, means that some level of meanness is involved.  However, the dichotomy created between troll and non-troll may obfuscate interesting insights—as shown by work that does not create strict dichotomies, such as Gabriella Coleman’s discussion of troll as trickster and Whitney Phillip’s argument that, “Rather than framing trolling behaviors as fundamentally aberrant, I argue that trolls are agents of cultural digestion; they scour the landscape, repurpose the most exploitable material, then shove the resulting monstrosities into the faces of an unsuspecting populace.”

Going back to the bestiality troller in WoW–the other day I logged in and discovered that the inappropriate sexual scenario troll had “escalated” his bestiality scenarios to an extremely graphic incest/pedophilia scenario.  This received a thunderous response from the general population who unified in its anger at him.

And, his escalation crystallized another question that had been nagging at the back of my mind about trolls and the relational nature of concepts in my research.  Why is he willing to post this incredibly disturbing content to receive a response, but other self-described trolls I know are not?

Around this time, I was on 4chan and someone posted a photograph that fell, without any gray area, into the category of child pornography.  The photo featured a child who looked to be under the age of 5.  While everyone (myself included) often focuses on the lack of rules on 4chan’s /b/ board, there is a very clear rule against child pornography (not to mention, a growing number of rules in general there over time), and the offending thread was removed within a minute by a moderator.

I wondered what makes someone willing to post something so extreme to receive a response while others draw invisible lines that they will not cross.

On the surface, it doesn’t seem odd that normal people are unwilling to say or post certain types of content.  Social norms and morality keep people from crossing certain lines in speech as well as in action.  But part of trolling is that the troll usually doesn’t believe what she is posting.  It isn’t just that audience members lose if they fall for a troll’s schtick, but that the purpose of trolling is entertainment, pure and simple.  And, buried in that is an idea that nothing really matters—words, images, and any other fuel for the trolling effort are just fodder for the moment of trolling (or the performance).

I wondered about trolls’ internal boundaries because I know quite a few people who used to troll.  But, in every case, there was a line that my friends wouldn’t cross.  My friend in World of Warcraft will only use humor and smartass comments to provoke a response from someone who is already being abusive.  I had a friend who would use extreme pornographic images to troll, but would never say anything racist.  Another e-buddy would be willing to say pretty much anything to make someone a respond, including highly offensive language, but he never posted any child pornography.  Part of Whitney Phillip’s research is on the trolling of Facebook suicide pages, and I know at least two people who trolled regularly for entertainment purposes who thought that this type of trolling was going too far.

The user posting in general chat in WoW made me focus on this question about the self-imposed boundaries that trolls give themselves—and suspect, although I have no evidence, that even the most extreme troll probably has some boundary that she doesn’t cross.  Gabriella Coleman refers to the idea of boundaries in her book chapter “Phreaks, Hackers, and Trolls: The Politics of Transgression and Spectacle” and says that even though it might seem that trolls have no moral restraint, in fact they do—specifically, while acknowledging that the dichotomy between online/offline is a false one, she articulates an idea that one’s pranks on the Internet should stay on the Internet (p.112).

Coleman’s “cultural line in the sand” (p.113) above, Whitney Phillips work on trolling and culture, and other scholars’ work has made me wonder about the social uses of trolling.  This is because one of the things I’ve tried to uncover in my research are the ways in which the individual beliefs that we import into online spaces then interact with the structures of online space and the norms of the communities in them.  Part of my argument about both groups in World of Warcraft and the subsections of the Anonymous communities rests on the idea that people already believe things when they arrive in those spaces.

It seems to me that trolling allows us to define the boundaries of acceptable behavior within a community, the transgressions that trolls engage in can also create important conversations, and a troll can rally a group to respond to behavior targeting someone as easily as he can rally a group to do something nasty.  Trolling does have a symbiotic relationship with the anonymity provided by many online spaces and certainly would have a relationship to the argument that anonymity allows a lack of empathy.  However, even though trolling has a symbiotic relationship with anonymity, people are also imposing boundaries and restrictions on themselves.  In other words, they are actually policing ourselves even as they transgress.  This is something that social scientists have observed and widely commented on in society at large – so it is no surprise that it would be happening online.   I would wager that to understand trolling, we need to understand these self-imposed limits as much as we understand the boundaries crossed.


As I was thinking about trolling and concept formation I read a bunch of amazing things–and I’ll list some of them here.

This is an incomplete (probably woefully so) list of work on trolling in no particular order.  Much of it I extracted from the AoIR discussion.  I also did quick searches of some journals such as First Monday and the Journal of Computer Mediated Communication.

Whitney Phillips has a website full of fascinating things among which, is an interview with a troll and discussion of her research.  You can read one of her articles, “LOLing at tragedy: Facebook trolls, memorial pages, and resistance to grief online,” in First Monday.  She’s also been profiled in various popular news sources such as Gawker, MSNBC, and Fast Company.

Gabriella Coleman has a blog post, “Hacker and Troller as Trickster,” and a book chapter “Phreaks, Hackers, and Trolls: The Politics of Transgression and Spectacle.” Both of these excellent pieces discuss trolls and trolling and engage in concept formation.

Judith S. Donath’s classic piece, “Identity and Deception in the Virtual Community,” discusses trolls as playing a game of identity deception and focuses on the costs of a troll to a community.

Susan Herring, Kirk Job-Sluder, Rebecca Scheckler, and Sasha Barab have an article published in The Information Society titled, “Searching for Safety Online: Managing ‘Trolling’ in a Feminist Forum,” that examines a troll interacting with a feminist discussion forum.

An io9 article about the psychology and behavior of trolls offline.

An awesome Oatmeal comic about playing online games as a grown-up.

NYTimes article, “The Trolls Among Us.”

Kelly Bergstrom’s piece in First Monday, “‘Don’t feed the troll’: Shutting down debate about community expectations on Reddit.com”

Erin Kissane wrote about the backlash against Anita Sarkeesian’s fundraising to support a documentary about women and video games.


Some of the work on concept formation that I have been reading is below.

Henry Brady and David Collier, eds. (2004), Rethinking Social Inquiry: Diverse Tools, Shared Standards. Rowman & Littlefiend Publishers.

Howard S. Becker (1998), “Concepts,” in his book, Tricks of the Trade: How to Think About Your Research While You’re Doing It.  Chicago: University of Chicago Press.

John Gerring (1999), “What Makes a Concept Good? A Criterial Framework for Understanding Concept Formation in the Social Sciences,” Polity, 31(3): 357-393.

Jack Goldstone (2003), “Comparative Historical Analysis and Knowledge Accumulation in the Study of Revolutions,” in Comparative Historical Analysis, Dietrich Reuschemeyer and James Mahoney, eds. Cambridge: Cambridge University Press, pp. 41-90.

Giovanni Sartori (1970), “Concept Misinformation in Comparative Politics,” The American Political Science Review, 64(4): 1033-1053.


I originally posted this on 9 July 2012 on my personal blog.

Comments are closed.