Reductio ad Trollum
Beware of the PAATroll

Alexandre Oliva

Reductio ad Trollum (dog Latin for "reduction (or argument) to Troll") is an elaborate form of "Reductio ad Hominem" that involves circular reasoning and "Reductio ad Hitlerum", the fallacy whose use triggers a popular consequence of Godwin's law that is often mistaken for the law itself. What makes this such an interesting and effective fallacious argument is that it is almost completely implicit, for it is usually proposed with a short sentence of strong emotional appeal such as "do not feed the trolls". The implied, disguised use of "Reductio ad Hitlerum" explains why it has so often had the opposite effect of an explicit use.

Reductio ad Trollum is often used to silence a controversial discussion, winning an argument against an interlocutor by labeling her as a troll, a scenario that makes the proponent of the fallacy a PAATroll (Positioned Against Alleged Troll). The popular wisdom that trolls only mean to cause harm, without bringing useful contributions to a community, fallacious leads to the conclusion that the claim proposed by the STroll (Supposed Troll) is incorrect. When used by community leaders, this fallacy often has the effect of evoking moral and verbal lynching, and ostracism of the STroll.


An Internet troll, or simply troll in Internet slang, is someone who posts controversial messages in an online community, such as an online discussion forum, with the intention of baiting other users into an emotional response.

It is well known that trolls are destructive to on-line communities, so various guidelines have been established to deal with them, the most common being not giving them the satisfaction of the emotional responses and thus the destructive disruption they seek. This is all captured by the advice "do not feed the trolls".

The fact that trolls are so well-known to be destructive makes them a perfect replacements for Hitler or Nazis in the "Reductio ad Hitlerum" fallacy of logic.

Begging the question

This is the circular reasoning fallacy of logic of assuming as true the statement one wishes to prove. It is quite effective when it is not obvious that the statement is contained in the assumptions, but in disguise.

Jumping to conclusions

This is a leap in logic, in which one concludes a statement holds even though the collected evidence is not enough to support it. Not all occurrences of "Reductio ad Trollum" involve jumping to conclusions, but argumentative uses thereof often do, involving both mind reading (basing conclusions on assumptions about what's in others' minds) and labeling (making inferences based on negative terms applied to someone or something).

Reductio ad Hominem

Disparaging a person making a claim without addressing its substance provides no evidence whatsoever as to its validity. When the disparaging is convincing as to the reputation or integrity of the proponent, it tends to reduce the faith others place on the unproven claim. Nevertheless, for not touching the substance of the claim, this argument is always a logical fallacy.

Reductio ad Hitlerum

This is a logic fallacy that appeals to the emotional rejection of Nazism, as a means of making a claim seem objectionable. Its most common form is "Hitler (or the Nazis) supported X, therefore X must be evil/undesirable/bad".

Reductio ad Hitlerum is rationally unsound for two different reasons: As a wrong direction fallacy (a type of questionable cause), it inverts the cause–effect relationship between why a villain and an idea might be criticized; conversely, as guilt by association (a form of association fallacy), it illogically attempts to shift culpability from a villain to an idea regardless of who is espousing it and why.

Godwin's law

In 1990, Mike Godwin proposed that the longer an online discussion extends, the more likely it is that someone will use a comparison involving Hitler or the Nazis. He did so as a means to avoid the overuse of hyperbolic comparisons.

There are many corollaries to Godwin's law [...] For example, there is a tradition in many newsgroups and other Internet discussion forums that once such a comparison is made, the thread is finished and whoever mentioned the Nazis has automatically "lost" whatever debate was in progress. This principle is itself frequently referred to as Godwin's law. It is considered poor form to raise such a comparison arbitrarily with the motive of ending the thread. There is a widely recognized codicil that any such ulterior-motive invocation of Godwin's law will be unsuccessful (this is sometimes referred to as "Quirk's Exception").'s_Law

Reductio ad Trollum

The form of the argument is to suggest that the proponent of an inconvenient claim is a troll, often without any evidence other than the controversy of the claim made by the STroll. The suggestion leads to the conclusion that the person is evil (i.e., like a Nazi), which, in turn, through "Reductio ad Hitlerum", makes any position held by the proponent, including the initial inconvenient claim, evil/undesirable/bad.

A more formal description of the reasoning involves the following predicates:

  • P1: The idea proposed by T is wrong
  • P2: T is a troll
  • P3: Trolls are evil
  • P4: T is evil
  • P5: Evil X proposes Y, therefore Y is wrong

The inference uses P1 as supporting evidence to conclude P2. This may or may not be fallacious reasoning, depending on what other evidence is available, but whether or not P2 is provable does not affect the fallacious nature of the entire argument.

The argument then uses P3 and P2 to infer P4, which is a logically sound inference step, and then concludes, through P5 (the "Reductio ad Hitlerum" fallacy) and P4, that P1 holds, closing the circularity.

Intent to troll and undecidability

Using the "do not feed the trolls" phrase in a debate, to imply an interlocutor is trolling, makes a judgment not only on the controversy of the position held by the STroll, but also on her evil intentions and behavior, because the destructive disruption caused by a troll is, by definition, deliberate.

Therefore, using this phrase as an argument in controversial discussions has become an effective way to disqualify an interlocutor, thus my suggestion of the mocking phrase "Beware of the PAATroll" to defend from abuses of the fallacy, a defense that should not to be abused itself.

The argument is effective in part because the element of intention, essential to characterize a troll, is subjective and essentially impossible to measure, let alone prove, which makes it very hard to defend from such an accusation, no matter how unsubstantiated it is. So, the accusation, in the presence of as circumstantial evidence as the persistent defense by the STroll of a claim assumed by the PAATroll to be wrong, "proves", through the disputable inference that starts the description of the circular logic process above, the evil intentions of the STroll, in that and in any other debates.

It is very easy to misinterpret and misunderstand utter motivations of an opponent in a debate, therefore when this accusation is brought up by someone with a vested interest in one of the positions (a PAATroll), most often by an active participant in the debate, it is more likely to be unfounded than when it is brought up by an unbiased person.

Furthermore, the urgent call to protect the community from an alleged aggressor, associated with trust on the caller built over time in the community, makes the accusation even more effective, for following leaders and reacting to aggressions have very strong emotional appeal. The emotional reaction to the call, even in spite of lack of sufficient supporting evidence, indicates it may be an instinctive process, which could thus explain the difficulty in involving the rational processes needed to detect such fallacies and halt inappropriate emotional reactions.

Nevertheless, whether or not the STroll is indeed a troll is irrelevant to the logical unsoundness of the entire "Reductio ad Trollum" fallacy.

Mirror corollary

Per the definition of Troll, quoted above, if the PAATroll "calls troll" in spite of the call being controversial, and does so with the intention of evoking an emotional response from the community, for example, as a distraction, as bait to change the subject to whether the STroll is indeed a troll, then the PAATroll is a Troll himself.


When a discussion is clearly off topic in a community, pointing that out and politely requesting the off topic discussion to move elsewhere is the best course of action. If the discussion persists, because a large number of active members of the community are interested in that topic, perhaps the discussion is on topic, after all. Failing that, one may conclude that there is intent to harm the community, and then "calling troll" may be appropriate, even though it says nothing about the validity of the claims in the controversy.

When a discussion has already been held exhaustively in a community, and the community agreed to no longer discuss it, politely pointing this out and providing pointers to the previous discussions for the education of new members may be the best course of action. It may be even more effective to start and maintain a community FAQ with such pointers for topics that have been exhausted, on which the community reached a consensus or agreed to disagree, and that the community decided to not discuss any further, which makes them effectively off topic. Then, when new members come by and bring them up, one can just point at the relevant FAQ entry. The pointers should also help the new member understand the reasons why the topic is banned, such that it's possible to tell whether, for example, discussing a new angle that hadn't been explored before would be welcome, or as undesirable as any repeated discussion. If the new member fails to respect the documented community decision, then "calling troll" may be appropriate, even though it says nothing about the validity of the claims in the controversy.

In many cases, it is tempting to use this fallacy as an elaborate (even if violent) form of censorship and proof by authority, to stop an argument that took a "wrong" turn. This use is discouraged, especially by PAATrolls, because, once the implicit "Reductio ad Hitlerum" fallacy is exposed, what would have been a trick to "win the debate" may turn into an "instant loss".

So, avoiding the use of "Reductio ad Hitlerum", in both its explicit and the implicit "Reductio ad Trollum" forms, is advisable. Warning a person about trollish behavior, and giving a chance for improvement before actually "calling troll", is not only polite, but it also reduces the likelihood of being counter-accused of using the "Reductio ad Trollum" fallacy to win the argument. This also provides for simplified accumulation and presentation of evidence, in case a later accusation needs to be backed up, and makes room for giving the benefit of the doubt, so important for undecidable judgments. This process also reduces the effectiveness of the "Beware of the PAATrolls" defense by actual trolls.

As unfair as being labeled troll may be, it happens, and Wikipedia's advice to Misidentified trolls is sound.

Unfortunately, trolls are a reality, even if telling them from honest but misguided community members is difficult. If it all comes to the worst, it is best if some member of the community that is neutral in a debate "calls troll", such that any automatic "win/lose" consequence does not apply. This person should ideally remind the community that the label and the treatment the community will then devote to the troll do not amount to a judgment on the substance under discussion, but rather on the methods used by the troll. The community can then carry on the debate, if there is interest, without prejudice to claims brought up by the troll.

Anecdotal evidence

I had web-searched for this fallacy name before writing the first draft of this article, but I had a misspelling, so I got no hits whatsoever. After writing the first draft and finding out I had a spelling error, I fixed it and searched again, and got only one hit:,1136,Scientists-Draw-Link-Between-Morality-And-Brains-Wiring,Robert-Lee-Hotz-WSJcom#43390

Interestingly, the PAATroll and the STroll had started snapping at each other after arguing some interesting points. The PAATroll appears to be a respected member in that community, according to another community member who jumped in his support. As predicted by Quirk's Exception to the popular misconception of Godwin's law, the intentional use of "Reductio ad Hitlerum", disguised as "Reductio ad Trollum", did not cause the discussion to end abruptly, and the consequences on the STroll appear to have been limited because he realized the nature of the fallacy and pointed it out.

This is all anecdotal evidence for the theory, of course, and I can't prove I hadn't seen that exchange before writing about the fallacy. (trust me!, what a useless phrase; if you do, it's not needed; if you don't, it makes no difference). Anyhow, it's still nice to see that the theory I derived from other experiences fit so perfectly the first (known to me) documented occurrence of the fallacy name.

The credit goes to some Henri Bergson, although certainly not the French philosopher who lived under that name from 1859 to 1941. Note that the evidence I have is not enough to determine whether Henri Bergson is a troll, but Luthien is clearly a PAATroll, since it is clear that he held an opposing position. What they are says nothing about the correctness of their claims, of course.

I found a later occurrence of the term "ad trollum", but, after a very superficial study of the situation, I don't see that that use of the term properly characterizes the claims of unhelpful participation. To me it looks more like a polite warning to the STroll that he was coming off as trollish, and that was not the first nor the last time. As of the writing of this article, the STroll is banned.

Several days later, I talked to someone about this article, and the person refers to it as 'ad trollium', with an 'i'. So I search the web for that and get a handful of additional hits, the earliest one from February 2006, which beats the first 'ad trollum' by 15 months. Oh well, I don't have a clue as to what the 'i' is doing here. I guess without it, it would be dog Latn.


Don't ever use this argument as a means to win a debate. False accusations could be very unfairly damaging to the interlocutor, and even to yourself, if you're found to abuse the argument. True but unsubstantiated accusations will just give the troll another controversial issue with which to disrupt the community.

Remember that, even if you have indeed identified a troll, "Reductio ad Trollum" is a fallacy, so it does not prove she is wrong and you are right with regard to the topic under discussion. Even if the discussion is about whether or not she is a troll, it would amount to circular logic just the same.

Actual trolls shouldn't find a safe harbor in this article; they will remain just as destructive and creative, and communities can deal with them just as effectively by making judicious and substantiated use of "do not feed the trolls" calls. But now misidentified trolls can use "Beware of the PAATroll", and a link to this article, against abuse.

This article is dedicated to all victims of "Reductio ad Trollum": both misidentified trolls and arguments discarded with prejudice.

Copyright 2008 Alexandre Oliva

All rights reserved for now, because this is still an early draft. Please don't publish the URL or quote this article before its final publication.

Permission is NOT YET granted to make and distribute verbatim copies of this entire document without royalty provided the copyright notice, the document's official URL, and this permission notice are preserved.