The Art of Trolling: Where the Word Comes From and What Research Shows
From the giant to the fishing line
The word troll has two older lives before the internet. As a noun for an ugly giant or dwarf it dates to 1610 and comes from the Old Norse word for a giant or demon. As a verb it names a fishing method in which a baited line is drawn through the water, and by 1972 US Navy pilots in Vietnam used the phrase “trolling for MiGs” for flying as bait.
The online sense appears in the early 1990s. The phrase “trolling for newbies” circulated in online discussion groups, and the Oxford English Dictionary’s earliest quotation for the internet meaning is from 1992. A post by Scott Cromar called “Trolling for Newbies”, dated 9 October 1992, is cited as an early example. Judith Donath’s 1999 paper on identity ambiguity in virtual communities is one of the academic treatments of the same early period.
The current meaning is harder. Trolling is the deliberate posting of offensive or provocative messages online, or similar behaviour offline, typically to provoke emotional responses or to manipulate how others perceive something. Wikipedia’s account notes that the label is subjective and is sometimes wrongly used for anyone with controversial or differing opinions.
Who trolls
Psychologists have looked for a profile. Studies cited in the Wikipedia article, among them work by Buckels and colleagues in 2014, link trolling to sadism and to the so-called dark tetrad of personality traits, and later work reports that men troll more than women. That picture of a small, antisocial minority was challenged in 2017 by Justin Cheng, Michael Bernstein, Cristian Danescu-Niculescu-Mizil and Jure Leskovec. Their paper “Anyone Can Become a Troll” won a best paper award at the CSCW conference.
The authors combined a controlled experiment with a long-running analysis of a large online news discussion community. They proposed two triggers: the individual’s mood and the context of the discussion, such as the exposure to other people’s trolling. In the experiment a negative mood and the sight of troll posts each raised the probability that a participant would troll, and together they roughly doubled it. A predictive model built on the observational data found that mood and discussion context together explained trolling better than a user’s own history of trolling did. The authors concluded that under the right circumstances ordinary people can engage in the behaviour.
What toxic comments do to readers
The second body of research looks at the audience. In 2014 Ashley Anderson, Dominique Brossard, Dietram Scheufele, Michael Xenos and Peter Ladwig published “The Nasty Effect” in the Journal of Computer-Mediated Communication. They used an online experiment with a nationally representative US sample. Participants read a neutral blog post about nanosilver covering risks and benefits, a largely unfamiliar topic for the public, and only the tone of the comments below it varied.
The result was more subtle than the headline. Uncivil comments alone had no significant direct effect on how risky readers thought the technology was. They did interact with prior views: among readers of rude comments, people with low support for nanotechnology perceived higher risk and people with high support perceived lower risk, so the audience was polarised. Religiosity also interacted with the comments: among those exposed to rude comments, highly religious participants reported higher risk perceptions than less religious ones. The model as a whole explained 17 percent of the variance in perceived risk, and the authors described their stimulus as fairly weak, since the exposure occurred only once.
A news summary from March 2013 described the experiment’s design, reporting that 1,183 subjects read a blog post on the pros and cons of nanosilver. Half saw civil comments and half saw rude ones, including one calling anyone who disagreed an idiot. The summary quotes Brossard as saying that rudeness may work as a mental shortcut for readers who need an anchor on a complicated question.
Responses
The oldest advice is to refuse the bait. The “don’t feed the trolls” approach holds that withholding attention may stop the behaviour, though Wikipedia notes that disruption can temporarily increase after contact stops, a pattern called an extinction burst. Other measures followed as the problem drew attention. The Wikipedia article records that in February 2010 users defaced Facebook tribute pages for two murdered children, that in 2012 the Daily Telegraph began a campaign against abusive “Twitter trolls”, that British law lists prison sentences of up to two years under the Malicious Communications Act 1988 or six months under Section 127 of the Communications Act 2003, and that Japan banned online insults in July 2022, punishable by up to one year of imprisonment.
The research suggests why a single remedy has not worked. If mood and exposure trigger trolling in ordinary users, then the design of a discussion matters as much as the character of its worst participants.
For other subjects on this site, the chess history hub, the computer chess hub and the page on correspondence chess are good places to continue.
Questions
When did the word troll acquire its internet meaning?
In the early 1990s. The earliest attestation in the Oxford English Dictionary is from 1992, and the phrase “trolling for newbies” was among the early uses.
Is trolling only done by a few antisocial people?
Studies link it to traits such as sadism, but a 2017 study found that negative mood and exposure to other trolling roughly double the chance that an ordinary user trolls.
Do rude comments change what readers think?
A 2014 experiment found no significant overall effect on risk perception, but rude comments polarised readers according to their prior support for the topic.