Conversation Hackers

Olivier Morin and Sophie Claudel

Human argumentation is at the center of recent (and less recent) psychological work. We are learning a lot about our ability to argue. But the motivation behind human arguing is less well known. What makes us want to argue back at other people, even when we know they won't be convinced ? Internet Trolls know a few answers to that question. We are studying their culture from the inside.

"Just consider how terrible the day of your death will be. Others will go on speaking and you won't be able to argue back" – Ram Mohun Roy (HT: Hugo)

A few weeks ago, the web was all abuzz about with one of those stories people are so fond of discussing online. A Canadian woman, who couldn't work because of a depression, lost her sick-leave benefits over a few photographs that were displayed on Facebook. She was smiling on the photographs. The anecdote provoked widespread outrage and rekindled the endless debate over Internet privacy.

But the story in itself did not interest Steve that much. Where other people see a scandal, Steve sees an opportunity for fun. That night, he logged himself on a forum devoted to discussing the condition and problems of depressive people – one among a dozen medical forums where Steve, under a variety of aliases, is a regular. He quickly spotted the thread where the Facebook scandal was being discussed, licked his lips, and began typing something like this:

"It serves her right, if you ask me. You can't defraud insurance companies and think of yourself as a responsible person. It's not the victimless crime it appears to be. Depression is not a real disease anyways."

He clicked 'Send', and waited for the angry reactions to pour in.



He did not wait long: people rarely refrain from biting on Steve's baits. He relished every minute of the argument, every insult, every pathetic attempt at counter-arguing – and shared it all with friends over Skype. When the outrage abated, Steve poked the conversation back into existence with a few nasty comments about lazy depressed jobless people funding their parasitic lifestyle with taxpayers' money. It did not take long before the fun was back. When Steve grew bored with infuriating depressed websurfers, he might have gone nagging at a forum of semi-literate teenage girls, heaping trash on Twilight 2, with similar success. Or perhaps he just discussed the night's achievements with a couple of appreciative colleagues.




There are many young people out there who are looking for a fight. Some go to seedy bars. Some hang around on the wrong side of a stadium, wearing the wrong colours. Some rely on the confidentiality and relative security of a Fight Club. Steve and thousands like him look for trouble on the Internet. Every night, many hours a night, Steve haunts forums dwelling on human rights in China, blog threads considering flaws in the last version of Microsoft Vista, medical newsgroups debating flu vaccines – and he spoils discussion after discussion. His arguments span all the range of conversational perversion: from childish insult to intricate accusation, from in-your-face provocation to subtle insinuation, from blatant non sequitur to elaborate sophistry. For Steve is a conversation hacker, or, as they are better known, a Troll.

Trolls are shy creatures – some might say paranoid. Theirs is a barely legal hobby, and knowing it, they are careful to leave few clues as to their identity. Steve, for instance, did not disclose his real name (he never does) but neither did he allow us to use one of his usual pseudonyms (I coined a name for him). This post relies on the direct testimony of ordinary Trolls, on discussion threads and demonstrations of skill that Trolls provided us with, and on hundreds of hours of observed on-line trolling. That information was collected by Sophie Claudel. A regular on a variety of IRC newsgroups since the age of 13, she has daily interaction with Trolls, some of whom have become friends. Trolls, you see, have a life outside of trolling – a social life that looks just as rich and fulfilling as yours and mine, with conversations that are as pleasant and rewarding as anyone else's. Sophie, who does not troll herself, meets them in real life, on a regular basis, and almost every night on the Internet.

This special relation allowed us, we think, to explore the puzzling motivations of conversation hackers. There is a lot of Troll material on the web, but you find either lurid Troll tales meant to scare and fascinate the public (this NYT piece is typical) or boastful Trolls commenting on their strategy in a complacent way (we prefer not to attract their attention by linking to them). Both Trolls and anti-Trolls like to picture Trolls in a sensationalistic way. They are excited by the freakish, the predatorial, and the criminal. But the tall stories carried by Troll lore, though some of them are true (Myspace suicides, Bonzai Kittens, Craigslist traps, etc.) do not reflect the reality of ordinary trolling.

What is a Troll?

Steve would never let you call him a Troll. He sees himself as a person who likes to argue. In a way, that is not surprising: a Troll worth of the name cannot endorse the label in front of his victims, while he is 'trolling' them. Most regular forum or newsgroup users know about Trolls, and if they spot one, they will shun it, moderate it or refrain from 'feeding' it. Anti-Troll policies are on the rise, which has made the hobby more difficult of late, but also more exciting. Yet Steve's friends, off-trolling, will readily admit to being Trolls. And all of them will recognize a fellow Troll in Steve; some will even say he's the greatest they know. But Steve is so professional that he will never allow himself to let down his facade of sincere interest for argumentation. He won't come out as a Troll.

Many things might explain why a Troll hides. Some Trolls belong to Troll Leagues, organised groups that invade various websites, launch demonstrations of strength, and fight rival trolling leagues. Leagues, which can be very big, have protection imperatives and norms of confidentiality. But all the Trolls we know are free-lance: they hack conversations on their own. If Trolls like Steve won't come out, this might be because they are sincere, or because they never stop trolling, even their friends and relations. It is surprising to notice that shame seems to play no part at all in keeping Trolls into the closet.

How exactly to define trolling is a thorny matter. The fact that discussing it will inevitably attract Trolls does not help. We can see two possible ways of defining Trolls, one of them strict and the other less so.

The strict criterion for being a Troll is genuine cynicism: if a Troll cares at all for the topic he is discussing, that interest must come second. His top priority must lie in winning a rhetorical fight by using all available means, including spoiling the debate, nagging people, ranting endlessly, etc. This motivation must come first chronologically, too: a Troll enters a debate with the clear intention of making it go awry. For example, a person who simply got carried away by a discussion and, becoming pig-headed, started resorting to provocation and insults, is not a real Troll. This is the criterion given by our informants.

To illustrate this point, here's an example of a strategy a Troll once described : you take a sensitive topic (like the ban on minarets or the latest problem with Macintosh OS), and you build an argument around it. The conclusion of your argument is blatantly absurd, but every premise is correct, except one. The trick is to hide that wrong premise under an intricate discussion. You know that people will be so hasty to resist your conclusion that they will start by attacking the true premises. You have prepared a violent rebuttal for each objection, and you know that, since you are right on those points, some objective debaters might side with you, which will divide the discussion group (a crucial step). You hope that the discussion of your true premises will become so heated that, when someone finally notices the flaw in your argument, people will be too busy insulting you to care about that. This is the kind of cold-blooded, cunning, premeditated strategy that only genuine Trolls can devise.

But this criterion – being a cynical and lucid conversation hacker – seems a bit difficult to apply. A savvy Troll is careful not to appear cynical or manipulative in front of his audience, since that would exclude him from the discussion. Closeted Trolls like Steve will claim, perhaps sincerely, that they have no intention of spoiling conversations – that is something their contradictors do, with their stupidity and lack of good arguments. Also, every once in a while, a Troll who is not looking for trouble will discuss a topic he genuinely cares about – yet his old discussion habits will prevail, and his conversation style will strike everyone (except himself) as trollish.

Argumentation gone wild

That is why we would like to propose another way of defining Trolls, one that is less stringent, and takes into account the fact that Trollhood has blurry edges : it can be more or less severe, and even the meekest debaters might possess a tiny spark of it. In this definition, Trolls happen to possess to an extreme degree a motivation that is common to all humans : a motivation to argue. That motivation is specific to argumentation itself, and can be satisfied even when the usual goals of argumentation – convincing someone of thinking or doing something – have not been met at all. We readily argue with people we have no realistic hope of convincing. Trolls are special because 1) this motivation is very powerful in them and 2) they don't just seize occasions of satisfying it as they present themselves ; sometimes, they deliberately create these occasions, by setting up rigged conversations. We may note that the weird tastes of conversation hackers often brgin them to disrupt the usual rules of conversation, but we don't make that a criterion.

Let us explain why we think the motivation behind trolling is similar in nature (though different in degree) to the motivation behind human arguing in general.

Before they went to the dark side, most Trolls were just pig-headed debaters like many others – and if it were not for pig-headed debaters, Trolls would soon go out of business. Everyone who ever dealt with a Troll knows of the strong, nagging urge to argue back at him ; and they know, of course, that this urge must be repressed at all cost, for it is what Trolls feed on. Thus trolling is powered by the same basic motivation that it serves to satisfy : that crazy desire to get the last word in a conversation. Trolls exist because there is enough Trollhood in everyone of us for them to feed on. Our informants are keen to point out the existence of unconscious Trolls ; as one of them said, "those who do not know about trolling troll unconsciously". Others said they did not see the difference between a regular dead-end debate (citing a classroom discussion on Palestine that went awry) and successful trolling.

This is enough to show how similar a Troll and his victim can be. Indeed, they are sometimes undistinguishable, as we shall see.

Trolls who troll Trolls

You might be surprised to learn that Trolls readily engage in long debates with fellow Trolls – people, that is, whom they know to be perverse and cunning conversation hackers. Apparently, this does not detract them from wasting hours on fruitless debates that are blatantly rigged and full of sophistry. Few Trolls would be happy with debating only fellow Trolls (semi-literate teenagers and hard-boiled fundamentalists are so much tastier – even though they, too, might be trolling you). Yet most of them, every once in a while, enjoy having an absurd argument with another pig-head.

Things get weirder still when a Troll decides to hack a conversation that, unbeknownst to the Troll, is already full of Trolls in disguise. This happens more often than you might think. This forum, for example, is officially a discussion group of the Flat Earth Society. It claims a connection with the society that debated Alfred Russel Wallace over the Bedford Level Experiment – an experiment that allegedly proved that the Earth is flat. On the face of it, it is a well-meaning attempt at disclosing to the public the latest results and speculations of sincere crackpot scientists. The forum is open to discussions between 'Rounders' and 'Flatters', moderation being assured by both Rounders and Flatters. Flatters lay down their claims in the inimitable way of crackpot scientists, and Rounders react with the passion of self-righteous rationalists wasting their time on a benighted website.

But what really happens in the virtual lobbies of the Flat Earth Society is more twisted. There is probably not a single sincere proponent of Flat Earth Theory on the whole site. Rather (as far as we could guess), the forum seems to have been designed as a gigantic Troll bait. The presence of Trolls is openly acknowledged on the forum, as some important moderators of the site, Flatters and Rounders alike, have been unmasked. They have been spotted on hacked private forums, where they were boasting about their hoax. Apparently, the 'Flat-Earthers' who created the site were really Trolls who planned to attract Round-earthers, and confound them with silly arguments. Instead, other Trolls showed up and began arguing for both positions.

This is a fairly typical episode. Trolls are devout defenders of Science since, as one of them told Sophie, "I like to make fun of ignorance and stupidity. That's why attacking theories like creationism or the like is interesting. It's like hitting a big ant-hill. It tends to ridicule people". Outlandish claims about the Earth being flat, or 4000 years old, have great appeal for Trolls. But other Trolls know about this, and they often devise bogus parascientific claims just for the sake of courting controversy. This thread, triggered by a blog post thrashing a videogame for teaching Darwinian propaganda to children, is probably a case in point. It is useful to bear this in mind when one studies crackpot science on the web, as many crackpot scientists might actually be fakes trolling their audience.

Even when a debate is obviously designed by Trolls and for Trolls, trolling is rarely acknowledged as the true purpose of the conversation. At Flat Earth Society, with trolling being endemic and conspicuous everywhere on the forum, participants take great care not to come out as Trolls. Flat-Earthers (most likely to be suspected of trolling) insist on the sincerity of their beliefs. Trolls being unmasked are a cause for scandal. The reason why everyone feigns to take the question of trolling so seriously is, of course, because accusations of trolling offer endless opportunities for trolling about trolling. This thread for example, is typical: everyone claims to be the only sincere defender of Flat/Round Earth Theory, and accuses everyone else of being a Troll.

Hacked conversation can be surprisingly hard to distinguish from normal conversation. This is partly the result of trolling stategies – since Trolls these days are waxing furtive – but it also tells us something important about the nature of both conversation hacking and conversation in general. Both are fueled by a basic motivation for arguing, one that goes way beyond bringing someone to do or think what we want him to do or think. The need to argue for the sake of it varies from person to person (intellectuals on this blog being probably a bit on the dark side) and culminates in Trolls. But most people enjoy having a conversation even when all hopes of convincing anyone of anything are lost, and as a result, hacked conversations can be as enjoyable as conversations played by the rules. Entire communities of conversation hackers can find great argumentative pleasure in conversations that violate the most basic requirements for convincing and constructive discussion.

Philosophers as Trolls

A question remains. If we are right, and the possibility of trolling is so deeply ingrained in human nature, why did it take so long for Trolls to appear?

The ready answer is that anonymous conversations became much, much easier with the Web than ever before – as lack of anonymity makes trolling much more risky. True enough. Yet more or less impersonal discussions did exist before the creation of Usenet (1979) – in newspapers or gazettes, in the public places of big cities, etc. We should find Trolls there too.

Indeed, we can find them in some of the first public places where free conversation between strangers was allowed, on a variety of topics : the antique Forum, grandfather of the virtual forums of today, womb of all Trolls. There you may find the antique equivalent of Trolls : what people at the time called 'sophists' or 'philosophers' – two words that were used interchangeably by the man on the Forum. Many Sophists did not want to endorse the label – sophistry was frowned upon or downright illegal in many places – and insisted on being called Philosophers. But the average citizen did not distinguish much between all these varieties of arguers. It is clear from most outsiders' accounts that sophists/philosophers were perceived as disrupting the usual rules of conversation in a noxious way.

Two important men are having a careful conversation on military training. What do you call the guy who, having no particular competence or interest in the matter at hand, jumps in the conversation, systematically contradicts everyone with contrived arguments, ridicules the two competent discussants, orients the conversation on a completely different topic, then leaves the audience baffled and walks away, laughing? That Troll is Socrates in Plato's Laches. True, Plato's Socrates seldom hops in uninvited, and most of his interlocutors do not consider him noxious. Indeed one wonders why the whole city grew so irritated that they voted to condemn him to death. But Plato, like all philosophers and sophists, had a stake in defending his colleagues. In other views of Socrates (like Aristophanes' caricature), he is unmistakably trollish.

And Socrates was not your average philosopher or sophist. His colleagues' methods were much cruder. Take Diogenes, a hobo who combined unsollicited moral counselling with aggressive begging. Take travelling philosopher Stilpo, who, each time he entered a town, went on the forum, jumped on a soapbox, brandished an onion and claimed he could prove it was not a vegetable (Proof: a vegetable existed 100 years ago. This vegetable did not exist 100 years ago. Therefore, this is not a vegetable), then rebutted all contradictors and baffled the audience till the town went mad at him. There were hundreds of Stilpos at a time, in all parts of the world where annoying intellectuals were tolerated. The Chinese had their Trolls too, whose discussions would create Chinese Logic. And of course, just like Trolls, these early philosophers tended to make themselves quite unpopular in several places. There is a reason why Athens punished sophistry with banishment, or worse.

In History, the web has no rival as a Troll nursery. But micro populations of Trolls and semi-Trolls do appear, we think, wherever more or less impersonal discussions take place. How should we react?

The Glory of Trolls

People usually go for the easy and virtuous option: they moralize Trolls. Trolls are time-wasters, destroyers of the ethics of discussion, sociopaths of the Internet. We should look forward to banning them completely in the near future. That reaction is understandable, but, we think, counterproductive. Anti-Troll discourse is utopian. Does it make sense to forbid people to "try to impress others with their knowledge"? To "respond to incendiary materials"? To "send messages or post articles which are no more than gratuitous replies to replies"? (to quote from the authoritative Netiquette guidelines of Intel Corp) How do we define words like "incendiary" or "gratuitous"? If all things incendiary or gratuitous were removed from human conversation, it would be cleaner perhaps, but also a lot less fun.

Anti-Troll norms are hypocritical too: they are readily produced and used by Trolls themselves. Closeted Trolls are vocal Anti-Trollers. Even unrepentant Trolls are masters of anti-trolling, which they use, as we saw, for their own twisted purpose. Nowhere are Troll hysterias more prevalent than among Trolls. Usenet or IRC, discussion spaces that look like Anti-Troll fortresses if you read their presentation or their guidelines, are actually mighty Troll strongholds. Notorious conversation hackers can be found at the very top of the hierarchies of these forums and newsgroups. An important proportion of the Trolls we studied are also moderators on IRC newsgroups, and as such, they are proficient Troll-busters.

We suggest, instead, that peace could be made between Trolls and other humans. Conversation hackers are useful. Like other hackers, they test the boundaries of a system, and they force users to devise better systems. They strain human argumentation to its limits. Dealing with Trolls forces you to sharpen your arguments and keep a cool head. Sometimes you might even learn something from a Troll. Socrates was maddening, but he helped make some concepts clearer. And all these Greek and Chinese philosophers/sophists forced their interlocutors to revise the usual rules of argumentation and make them much more specific. Some modern logic was born from these efforts.

There is another reason to make peace with Trolls: they are much less alien than we'd like to think. Everyone has their inner Troll ; everyone has their urges to argue pointlessly with people they know they won't convince. Anti-Troll norms might keep our inner Trolls in check. But they might also foster a spirit of intolerance for other people's pig-headedness, and encourage us to deny our own trolling proclivities. Anti-Troll brigades are full of Trolls, Anti-Trolling being one of their best weapons. On the other hand, experienced Trolls gone to the bright side are better than most people at guiding arguments in interesting directions – which is why they often become newsgroup moderators. As usual, a system's hacker is often the best expert in the security of that system. Knowing that pig-headed discussions will never disappear from this world, there is sense in preferring to deal with proud and savvy Trolls, instead of clumsy, insecure and aggressive pig-heads. To quote our informant again: "those who do not know about trolling troll unconsciously". One might want to chose the conscious version.

That said, we know we barely scratched the surface of the topic. Many questions haven't been answered: are Trolls better at arguing or reasoning than the average geek? what kind of risks exactly does a Troll run? Is it true that, as rumors have it, corporations or states will pay Trolls, in addition to regular hackers, to bring down internet discussion spaces that go against their interests? We hope to address these issues some day.

Meanwhile, feel free to drop a comment below.

We are open to discussion.

Manual trackbacks:

Crooked Timber (don't miss the amazing comment thread)

Savage Minds

Quantum of Wantum

Michael Nielsen


Our post was also discussed on the ever fascinating Forum of the Flat Earth Society


(Olivier Morin's website is here.)


  • comment-avatar
    Pascal Boyer 16 December 2009 (17:13)

    Web forums and groups should be a great field for cognitive anthropologists – great and frustrating, as the taphonomy of internet exchanges is even more constraining than that of archaeological remains. The web gives us plenty of evidence for human communication, but it is highly skewed, produced by self-selected individuals, and by people about whom we cannot (legally or practically) obtain additional information. In my own, limited, experience and observation, “threads” (i.e. series of consecutive messages in response to some initial message, news item or opinion) are particularly interesting because [a] they rarely produce anything remotely informative, as far as the topic at hand is concerned, and [b] they almost invariably escalate into a shouting match and insults of varying degrees of subtlety. Why is that so? and why do people participate? Trolling may be one reason. I suspect the people who spend a lot of their time actively trolling probably overlap or at least share important features with other populations such as video game obsessives, various kinds of geeks and trainspotters (non-British readers, find a Brit to explain who these are). The interesting question perhaps is no so much what they do or why they do it, as why it has such an effect on internet conversations. Many threads on the net are very similar to clinical transcripts of conversations with schizophrenic patients. That is, the exchange seems in constant danger of swerving into a conversation ditch, so to speak, because one person misunderstands what someone said, then goes off on a tangent, which triggers equally off-topic comments from others, and so on. In such contexts repair (in the technical sense) seems very difficult – so that referential ambiguity cannot be clarified, leading to endless series of “that’s not what I said, what I said was…:” messages as well as utterly confusing embeddings of embedded responses to embedded messages. I had an example of that recently, following for a few days a thread about an article of mine (that’s sad perhaps, but one has to enjoy one’s fifteen minutes after all). The thread started in great fanfare with people saying that my argument was utterly misguided. So far, so good. Then someone pointed out that a special way I was misguided was in trying to argue for “p” (p is a hypothesis about group-selection that I do not agree with, and has been put forward by another evolution & religion person). Then someone said that it was typical I would argue for “p”, since all religious apologists agree with “p”. [For the record, I would rather have my leg chopped off than do religious apologetics]. Then someone suggested there should be a list of people like Boyer who insidiously use evolution and cognition to bolster their religious beliefs. Then someone said that that would be very much like communism, where they had lists of people with the wrong kinds of opinion. Then someone said that much-maligned communism was actually a great plan for society, only ruined by dictators and people who had not really read Marx. The following fifty or so increasingly vindictive messages were about communism and whether it has been a Good Thing after all. Then, out of the blue, someone wrote that they has bought a book of mine but it was unreadable garbage. [Yippee! I thought, we’re back on track, talking about the really interesting stuff] No-one took that bait. Someone pointed out that all the bad stuff about America was because of the minorities, at which point the moderator wisely chose to close the thread. Now a lot of that sounds like everyday conversations – but a lot of it does not, and it would be great to see what psychologists, conversation analysis folks and pragmatics people make of that. For instance, in regular conversations we assume a fairly stable set of interlocutors – so do we actually adjust to the fact that on the web the person you respond to is probably not there anymore, and someone else will reply to you? or do we just maintain that tacit assumption, which makes conversations so incoherent? and many other such questions. PB ================= NEXT MESSAGE If you ask me, that post by Olivier Morin and Sophie Claudel was terrible. Typical of French people – when will those Frogs learn to think? Anon-Renegade =================

  • comment-avatar
    Milo Price 21 December 2009 (04:04)

    The focus seemed to be on only one type of troll, being those who try to enrage people or disrupt conversation. What about those who assume the guise of someone stupid, yet benign rather than belligerent? Sure, there are many trolls who make false arguments or claims, and that’s probably the largest constituency, but what about people who, for example, ask to see a human-controlled helper in a virtual world and pretend to be confused for a longer time than it would take most other people to understand the answer to their question until the helper quits in disgust? The two aren’t as alike as one may think. Belligerent trolling is much easier to accomplish, because it’s easy to pretend to be stupid and angry and easy to get equally stupid and/or angry responses. When you pretend to be someone who’s just confused about the way the community, topic, or language works, however, the effect is much more subtle, and even if someone does suspect, it’s riskier to call the troll out on it.

  • comment-avatar
    Nick Argall 30 December 2009 (01:07)

    PB – I found your comparison between conversation with a schizophrenic and a troll-ridden thread fascinating. As someone who has been diagnosed with a dissociative condition, I’ve found that the practice I’ve had at finding constructive resolutions to arguments on the net has been useful practice at resolving arguments in my own head. While I don’t have much expertise with conversation analysis and pragmatics, I have a lot of experience with trolling… and, actually, I know some people who study linguistics. Hmmm….. MP – Yes, indeed, this good article is incomplete. Of course, I came here looking for a horse, and while finding a white horse satisfies me, I can understand that you might have wanted a horse that also had a saddle. A google search for ‘taxonomy of trolls’ yields some reasonable results. There’s a particularly cool troll taxonomy that I think I was linked to from alt.zen that had pretty pictures. I like pretty pictures. Speaking of alt.zen, that’s a place that has some of the highest-skilled trolls I’ve seen, putting most others to shame. They’ll even admit to being trolls, and point out that disapproving of trolls is contrary to Buddhist principles. (I remember a very cool conversation with Tang Huyen in which he criticised the other trolls for clumsy technique.) And yes, the ‘innocent question’ is one of the best trolling techniques when performing for an intellectual audience. Intellectuals pride themselves on their ability to answer questions, and if someone points out a fallacy inherent in the question, you can very effectively respond with “Oh, I see! Thanks for explaining that!” and move on to your next move (if so inclined). The highest art in trolling is not to call someone stupid, but to give them an opportunity to demonstrate their stupidity in front of a meaningful audience. The best trolls are able to acknowlege the skill it takes to avoid the traps that they lay, and celebrate that in others. =================== NEXT MESSAGE Anon-Renegade is at it again, I see. Does Steve (seriously, Steve, we all know, why persist in using that alias?) really think that anyone is impressed by his racist slurs? Just ignore him, he’ll go away when he realises he’s out of his league. Internet Sherriff ===================

  • comment-avatar
    Christophe Heintz 9 January 2010 (17:24)

    This ethnography of trolls is indeed rich of data relevant to the study of argumentative practices and its psychological bases. Olivier and Sophie actually make a psychological claim on the basis of their ethnography and some historical data: they assert that people are motivated to argue for the sake of it. We all like to argue, they say, and this type of motivation is stronger for trolls than for the average people (c.f. the first paragraph of their section “argumentation gone wild”). This is a strong and interesting claim and I believe it is worth more attention and further empirical studies. I am myself not convinced that a motivation to argue for the sake of it is indeed a human universal. I am not sure that we derive any satisfaction (or utility) from arguing per se. What motivates us to argue is rather the prospects of the results of a successful argument. The prospects are, mainly, convincing others, and, also, increasing one’s reputation as being a reliable source of information (e.g. knowledgeable, not easily fooled, smart, …) Here are some reasons to think that the motivation for arguing might not be pleasure derived from arguing. 1) According to Hugo Mercier and Dan Sperber’s evolutionary account of argumentative abilities, humans would have evolved argumentative abilities for, firstly, convincing others, and secondly, avoiding being deceived by others’ arguments. For these abilities to perform these two functions, do they need to be supplemented with a motivation to argue? Would this motivation increase our fitness as do our desires to have sexual intercourse? The answer is not obvious. I would think that we usually derive most satisfaction when arguing achieve its function and actually lead to convincing others. Arguing differs in that respect from copulating: we derive pleasure from copulating independently of whether we actually procreate. Unlike our drive for sex, the drive for arguing is probably the prospects of its functional effect: convincing. We understand the connection between arguing as a means and convincing as its goal (if we don’t, then we are unlikely to be good at arguing), and so we do not need a motivation for arguing other than the prospect of convincing. 2) Arguing requires cognitive efforts, and we tend to avoid such efforts unless their prospected results are worth the cost. I’d bet that the effort that is given to argumentation is, usually, directly dependent on the perceived stakes. Trolls might appear to be an exception there, but I suggest that this is not the case: trolls’ stake in argumentation is self-esteem, and this is why their motivation is to a large extent independent of the content of the discussion (more on this below). 3) Convincing means changing the mind of the audience in the desired way, which is the primary function of argumentation, but it also, as a important side effect, shows to the audience that we are smart and/or knowledgeable, and that we are not easily deceived. It shows to the community that we are a reliable source of information. I guess that we derive much satisfaction from this second effect. 4) It seems that most of us do not feel any joy or pleasure when we loose an argument. So we cannot be motivated for arguing independently of its expected success in convincing (or side effects such as increased prestige or increased self-esteem). What about Trolls? They usually do not convince and they do not obtain epistemic authority among the people with which they argue. Nonetheless, they do increase their self-esteem by showing to themselves that they can direct the discussion as they want. That increased self-esteem is the ultimate motivation of trolls is indicated by their avowal that they want to show that others are idiots — i.e. less smart than themselves. Also, some trolls cannot but boast about their achievements. Some evolutionary social psychologists have argued that motivation for increasing one’s self-esteem is a good proxy that lead people to eventually increase the esteem that others have of them (it would be the evolutionary function of self-esteem). But sometimes this goes astray, and this is what happens with trolls: they increase their self-esteem without increasing their social prestige. Yet, I’d bet that the more a community values argumentative skills, the more trolls they are in this community. Thus, trolls would not be so off-target.

  • comment-avatar
    Hugo Mercier 12 January 2010 (23:25)

    A few short comments on Christophe very interesting (and longer) comment. >humans would have evolved argumentative abilities for, firstly, convincing others, and secondly, avoiding being deceived by others’ arguments. For these abilities to perform these two functions, do they need to be supplemented with a motivation to argue? Well, we need to be motivated to argue in some cases at least, otherwise we would never argue, would we? >It seems that most of us do not feel any joy or pleasure when we loose an argument. Well, not in loosing per se, but if we are genuinely convinced by someone’s arguments, even if it is only after a long debate with us, then we may feel happy to have reached a better epistemic status. I think the problem is that arguing “for the sake of it” is not quite clear. If it means that people are not aware of the ultimate goal of arguing, then most of the things we do (including something with a clear evolutionary function like sex) is “for the sake of it”. So people arguing for the sake of it would not be a problem. However, even in this case, we could imagine that arguing would not be a “primary reinforcement”: it would only feel good because of its effects (e.g. acquiring better knowledge, getting other people to accept our point of view). The problem may emerge when arguing becomes a primary reinforcement: when we like to do it irrespective of its effects. I think this is what Christophe is alluding to, and in which case he is probably quite right that this is (or at least can be) problematic. The misguided self-esteem explanation would be an interesting hypothesis to test.

  • comment-avatar
    Olivier Morin 14 January 2010 (12:48)

    There is something about trolling that I find tremendously relevant to pragmatics on the web: trolls’ intricate use of irony. As Milo Price points out, trolling can be extremely diffuse and lacking in agressivity, sometimes to the point of being undetectable – just like socratic irony. Trolls have a distinctly ironic argumentative style. What does that tell us about pragmatics on the web? From a pragmatic standpoint, irony is essentially allusive and multilayered (you basically allude to something you or someone else might have in mind). That is why irony typically requires a rich shared background if it is to work at all. For example, a sentence like ‘your boyfriend has a nice new haircut’ can be ironic if told to someone whose boyfriend is completely bald, by someone who knows that. Common knowledge of what makes the sentence ironic is essential. But some forms of irony are more multilayered than others. Note that the more multilayered the irony, the more common knowledge you need to make sense of it. Troll jokes, like the French ‘mer il est fou’ (“bu he’s mad”) are typically ironic to the 3d or 4th degree. Now Troll Irony is relevant to Pascal’s comment because if Pascal was right, irony should be very difficult on the web. Web conversations, Pascal seems to assume, are pragmatically impoverished, and mutual understanding is constantly endangered. This can be expected from the lack of cues conveyed by gestures, pitch, volume, smiles, etc. in normal conversation. That point has been made time and again by critique of digital communication. But, as far as irony is concerned, one can easily see that web conversations are pragmatically very rich. Troll irony is fabulously intricate. Most famous Troll jokes and motos are ironic – like the slogan ‘Everything is serious on the Internet’. Some jokes, like the mythical ‘Pool’s closed due to AIDS’ are ironic to the nth degree: the funny thing about ‘Pool’s closed’ is that it is itself a reference to a nonsensical hoax, that was itself a tasteless way of joking about tasteless jokes about AIDS. Whether the joke was originally meant to be understood at the first, second or third degree is difficult to know. Such heavily multilayered jokes require a rich shared communicative background to be made sense of. True, Trolls use them once in a while to make fun of disoriented victims who just don’t grasp the irony (this is par of what makes ‘everything is serious on the internet’ funny). But they are also popular among non-trolls in communities like 4chan, where they work like shibboleths. Troll irony, as I see it, is richer and requires more pragmatic skills than most instances of irony. This belies the common picture of cyberspace as the realm of conversational pathology. Sick arguments do occur on the web, but they also occur around dinner tables, at any bar’s counter, at coffee breaks, etc. I remember reading, from a sociolinguist, the record of a live conversation between two people, A and B; A argued that the Chinese cannot smash at ping-pong. B disagreed and asked for reasons. Well, A replied, the Chinese don’t know how to hold a racket. Then what, B enquired, would the right way of holding a racket be, given that some of the world’s best ping-pong player go for the ‘wrong’ way? There is no right answer, A replied, because table tennis is not a real sport anyways. The discussion then proceeded to sink in the quagmire of ‘Let’s define sport’. The big difference between these conversations and those that take place on the web is that they will, most of the time, go unrecorded. We often take part in such silly conversations, we just didn’t realize it untill chats could be written down.

  • comment-avatar
    Ellen Mellon 13 March 2010 (08:49)

    This was an interesting read for the info about troll culture and interrelations. But the reason I was looking for more information about trolls was because I wanted to understand their motives, and I find myself just as confused as ever. This article tries to find the troll in all of us, so to speak, but what about sincerity? It’s true that non-troll people aren’t always sincere in their online arguments, but most of the time they at least care about what they’re arguing about. And what possible meaning can a discourse have if there is no sincerity? What is the point of arguing with someone when you don’t care what they think? And why devote so much time to it? Given all the reasons *not* to troll (foremost, a desire for authentic communication), what are trolls getting out of it that makes it rewarding for them? What do they feel when they troll?