An examined document · read the margins
Argument from Fallacy
Inferring that a conclusion must be false merely because one argument offered for it commits a fallacy — confusing a defect in reasoning with a defect in the truth of the claim.
Argument from fallacy (the "fallacy fallacy," argumentum ad logicam, fallacist's fallacy) is a normative error in informal logic, not an experimentally measured cognitive bias: it treats "this argument for C is fallacious" as if it established "C is false." Because a fallacious argument can still have a true conclusion — fallacies are errors in the inference, not verdicts on the truth-value of the claim — the move is itself invalid, structurally a special case of denying the antecedent. The concept is a philosophers' and logicians' construct, established by analysis and example rather than by psychology experiments, so it has no effect sizes, samples, or replication record.
The error rests on conflating two independent properties of an argument: the validity/soundness of the inference and the truth-value of the conclusion. To call an argument fallacious is to deny a sufficiently strong logical connection between premises and conclusion; it says nothing about whether the conclusion is in fact true. Formally, argument from fallacy has the shape 'If P then Q; the argument for Q is fallacious; therefore not-Q' — a special case of denying the antecedent, where the negated antecedent is an entire argument rather than a single proposition. As with any denial of the antecedent, the consequent may still be true. Aberdein (2023) reframes it as an instance of the 'Owl of Minerva problem' (Aikin 2020): the very tools devised to detect bad reasoning (fallacy labels) create a new way to reason badly — flagging a 'fallacy' to dismiss a conclusion. Lycan (1996) isolates a related but distinct sense: imputing fallaciousness to a disliked view without showing any actual error, which shades into begging the question / argumentum ad lapidem.
There is genuine scholarly disagreement over what 'the fallacy fallacy' even is. Aberdein (2023) distinguishes (a) misdiagnosis of fallacy — calling something a fallacy when it isn't (Lycan 1996; Hedden 2019) — from (b) the supposition that a fallacy's conclusion must be false (Cotton 2019; Aikin 2020). Fischer's (1970) original 'fallacist's fallacy' is broader still, bundling several propositions including a moralizing one about the arguer's character. A separate live debate (Walton; informal-logic literature) questions whether 'denying the antecedent' is always fallacious at all, which bears on whether argument from fallacy is strictly a formal fallacy or a defeasible/contextual one.
Further annotations
case · Wiles's flawed 1993 proof of Fermat's Last Theorem (1993)
Andrew Wiles announced a proof of Fermat's Last Theorem in June 1993; in September 1993 a critical gap was found (the Euler-system bound via the Kolyvagin–Flach method was incomplete). The argument as given was defective — yet the conclusion (the theorem) was true, and Wiles and Richard Taylor published a corrected proof in 1995. The Fallacy Files cites this as the canonical illustration: an argument's fallaciousness does not make its conclusion false. It is an illustrative historical episode, not a documented instance of someone actually committing the fallacy.
case · Historians' inference of falsity from fallacious method (1970)
Fischer's 1970 book documents the fallacist's fallacy as a recurring error in historical scholarship — historians rejecting a colleague's conclusion (or imputing the author's bad faith) on the ground that the supporting argument was structurally fallacious. This is the originating documented context for the named fallacy, drawn from Fischer's analysis of real historiographical reasoning, though Fischer presents it as a pattern rather than as a single dated dispute.
Catch it in the act
Watch for the move "that's a logical fallacy, so you're wrong" — where pointing out a flaw in one argument is treated as having refuted the conclusion itself. The tell is a leap from "this reasoning is bad" to "this claim is false," with no independent case made against the claim. Especially common in online debate ("ad hominem, therefore false") and after spotting any named fallacy in an opponent's case.
Cross-references in the margin
- Denying the antecedentparentArgument from fallacy is standardly analyzed as a special case of denying the antecedent, where the negated antecedent is an entire fallacious argument rather than a single false proposition.
- Argumentum ad lapidem (appeal to the stone)mechanistically-linkedDismissing a claim as absurd without argument; the Lycan (1996) 'misdiagnosis' sense of the fallacy fallacy shades into ad lapidem when fallaciousness is asserted but never shown.
- Genetic fallacysiblingBoth judge a conclusion by a feature of how it was reached (its origin / the quality of one argument) rather than by the conclusion's own merits.
- Ad hominemeasily-confusedOften co-occurs in examples: an interlocutor commits ad hominem, and the target wrongly infers from that fallacy that the interlocutor's conclusion is false — itself the argument from fallacy.
- Bad reasons fallacy (argumentum ad logicam, broad sense)parentThe Fallacy Files treats argument from fallacy as a subfallacy of the broader 'bad reasons fallacy' — inferring a conclusion's falsity from the badness of an argument for it, of which committing a fallacy is one species.
- Begging the questionmechanistically-linkedAberdein/Lycan note that asserting an opponent's argument is fallacious without demonstrating it can itself beg the question.