Hi, all. Welcome to my topic on the discussion of fallacies.
Please excuse the edits. I am a busy individual and my first versions were insufficient. Please see the AI summary to ascertain the direction of my text.
Introduction
Throughout this text I am working on the assumption that there are people who are somewhat too eager to call something a fallacy. However, calling something a “fallacy” is often not a sufficient critique, since such judgments rely on context dependence, normative complexity, the heterogeneity of the category, the risk of weaponization, and the pedagogical limits of presenting fallacies as a simple list. This topic will outline the basics of fallacy theory, its many complexities, and a few possible solutions. My central point is that fallacy theory fails when it treats arguments as isolated logical objects; it succeeds when it treats them as situated, dynamic, normative moves. The latter requires context and contemplation.
What Is a Fallacy?
According to the textbooks, the criteria for being a fallacy are that it: 1) must be an argument (or purport to be one); 2) falls short of some standard of correctness (deductive invalidity, inductive weakness, unjustified premise, ignored evidence, etc.); and 3) has a semblance of correctness, which helps to explain its persuasiveness. Fallaciousness is not a purely formal property. It is tied to standards of correctness (the logical aspect), dialogue goals (the dialectical aspect), and persuasiveness (the rhetorical aspect). These aspects are taken from Argumentation Theory for the purpose of orientation.
Fallacy vs Similar “Moves”
Philosophically, we must distinguish several notions. Errors in reasoning are cognitive mistakes (for example, miscalculation or misremembering a premise), not fallacies. Argumentative defects are structural or inferential flaws in the argument as presented, not fallacies. Biases are systematic tendencies in cognition (such as confirmation bias) that may cause errors but are not themselves arguments. Sophistry or deception is the intentional exploitation of the appearance of correctness in order to win, regardless of truth, and may not involve fallacies. A move can be biased without being a fallacious argument; it can be deceptive without being invalid; and it can be a genuine reasoning error without violating any dialectical rule. This already complicates the analysis of fallacies.
Fallacy Locations
Argumentation Theorists like to speak of the “structure” of arguments. Given this, fallaciousness can arise at different locations. These include unjustified premises; inferential failure (invalid deduction, weak induction, faulty abduction, etc.); and discussion rule violation (shifting the burden of proof, blocking critical questions, etc.). It can be useful to distinguish between local and global levels of fallaciousness. An argument can be locally defective (a weak inductive step) yet globally acceptable in a given discussion context (for instance, a provisional hypothesis in a scientific inquiry). Conversely, an inferentially valid argument can be globally fallacious if it derails the discussion (for example, by blocking or evading the issue). This should dissolve about 80% of internet “fallacy” accusations, which treat arguments as isolated logical objects rather than moves in a dialogue.
Norms and Normativity
Fallacies are normative notions that presuppose standards of good reasoning and good dialogue. These include epistemic norms (truth, justification, proper use of evidence—the logical aspect); pragmatic norms (effectiveness in achieving goals, coherence with conversational aims—the rhetorical aspect); and dialectical norms (rules for critical discussion, fairness, burden of proof, relevance to the issue). If these standards are violated, we say a fallacy has been committed. It is possible for different standards to be violated at different locations. Again, these aspects are taken from Argumentation Theory. My audience can be the judge of whether or not I have allocated them correctly.
Fallacy Locations × Norms: A Comparative Table
| Fallacy Location | Epistemic Norms (Logic) | Pragmatic Norms (Rhetoric) | Dialectical Norms |
|---|---|---|---|
| Unjustified Premises | Premises lack adequate evidence or justification. | Premises may be rhetorically effective (e.g. emotionally charged) but fail to advance truth‑seeking aims. | Premises violate rules of critical discussion by introducing claims without meeting burden of proof or by evading requests for justification. |
| Inferential Failure | Inference does not preserve truth or probability; conclusion is not adequately supported by premises. | Inference may be persuasive or goal‑effective (e.g. quick heuristic) but epistemically unreliable. | Inference blocks or bypasses critical questions that should be addressed in a critical discussion. |
| Discussion Rule Violation | Violations may indirectly undermine truth by preventing proper evaluation of evidence. | Violations can be rhetorically effective (e.g. dominating the conversation) but frustrate cooperative conversational aims. | Direct breach of dialectical rules: evading burden of proof, irrelevance, preventing the other party from challenging claims. |
This matrix shows why treating arguments as isolated logical objects misses much of what makes a move fallacious: the same location can violate different norms in different contexts, and the same norm can be breached at different locations.
Examples and Context‑Sensitivity
Context is crucial to ascertaining fallaciousness. Identical forms can be fallacious in one context and acceptable in another.
- Appeal to authority: fallacious when the authority is irrelevant or unqualified, legitimate when expert testimony is properly within the domain of expertise.
- Appeal to emotion: manipulative in scientific justification, sometimes appropriate in courtroom sentencing or moral deliberation where emotions are relevant to values.
- Ad hominem: fallacious when used to dismiss a claim solely due to personal traits, potentially relevant when assessing witness credibility or dialectical consistency.
- Ad populum: Popularity is usually irrelevant to truth, but in some contexts (e.g. conventions, coordination problems) majority practice can be significant.
- Straw man: Fallacious when representation is unfairly distorted to make refutation easier. But what counts as “faithful representation” is itself interpretive and context‑sensitive.
- Begging the question: Circular arguments can be deductively valid yet epistemically defective if they presuppose what is in dispute. Different varieties of circularity matter.
- Hasty generalization: Induction from insufficient evidence, but “sufficient” depends on sample size, variability, and stakes involved.
- False cause: Post hoc reasoning is often a heuristic for causal inference, distinguishing fallacious from legitimate causal claims requires careful consideration of confounders, mechanisms, and base rates.
- Slippery slope: Fallacious when causal chains are speculative, legitimate when there is strong evidence for each step and the slope is genuinely slippery.
- Equivocation: Fallacious when a term shifts sense mid‑argument, but polysemy and legitimate metaphor complicate the boundary.
- Composition/Division: Part‑whole inferences fail in some cases (emergence) but are valid in others (additive properties). Mereological principles apply.
These examples are a selection from hundreds. Ideally, to call something a fallacy, one must show not only that a pattern exists but also that in the context of discussion that it violates a certain norm in a certain location.
The Question of Irrationality
Argumentation theory holds that an argument is supposed to be acceptable to a rational audience, and norms of rationality are sometimes invoked when speaking of fallacies. But some “fallacies” may be ecologically rational heuristics that misfire in modern environments rather than being pure irrationality. From an evolutionary perspective, what looks like a fallacy may be a generally reliable strategy applied outside its proper domain. This is not irrationality so much as a mismatch of environments.
Fallacy Taxonomies
The standard treatment of fallacies sees them as a taxonomical list. Everyone should be familiar with Wikipedia’s list of fallacies. But such lists suffer from overlap and redundancy (for example, red herring, straw man, and irrelevant conclusion often blur together), proliferation (new “fallacies” are coined faster than they are theoretically grounded), and arbitrary individuation (fallacies are heterogeneous, so are they classified by form, content, function, or psychological mechanism?). Without principled criteria, taxonomies become a pedagogical convenience rather than theoretical insight. Memorizing them is fine so long as this fallibility is considered.
Controversial Cases
These cases are controversial because they are ubiquitous to standard reasoning, existing in an order prior to standard fallacy lists and not often discussed. Concerning induction, Hume’s problem shows that all inductive inference lacks deductive guarantee; the line between legitimate induction and hasty generalization is one of degree and justification. Concerning abduction (inference to the best explanation), this is warranted when hypotheses are compared on explanatory power, simplicity, and fit, but fallacious when used dogmatically or without considering alternatives. And concerning analogical reasoning, strong analogies can be highly informative, whereas false analogies are fallacious relative to the respects of similarity that matter for the conclusion. In each case, the “fallacy” label depends on how well the inference is supported, not on a simple pattern match.
The Mighty Fallacy Fallacy
The fallacy fallacy is the error of concluding that a claim is false simply because the argument given for it is fallacious. Bad arguments can support true conclusions, and good arguments may yet be unavailable. This alone shows that fallacy‑spotting cannot replace substantive evaluation. This means that calling a fallacy does not end the discussion—it begins the question of whether the conclusion might still be true. If we take this principle to the extreme, we end up with a world where every fallacy automatically makes its conclusion invalid, which collapses into the absurd result that all fallacies imply falsehood even when the conclusion is actually true. Thus the fallacy fallacy is a powerful meta‑check on fallacy‑spotting.
Weaponization of Fallacy Labels
In practice, “That’s a fallacy X!” sometimes functions as a rhetorical dominance move to silence opponents, as a thought‑terminating cliché that halts further analysis, and as a form of ad hominem via labeling (that is, “you’re committing fallacies, so you’re irrational”). Overdiagnosis—labeling any disagreeable argument as fallacious—undermines the diagnostic value of the concept and can constitute a meta‑fallacy: misapplying the very notion one claims to defend.
Infinite Regress
If every critique regarding an alleged fallacy can itself be accused of fallacy, we face a potential infinite regress. This motivates some philosophers to replace “fallacy” with finer‑grained evaluative tools (for example, specific norm violations). At minimum, it demands humility: fallacy attributions are themselves argumentative moves subject to their own kind of critique. This means discourse will sometimes feature somewhat tedious meta‑argumentation of fallacy viability in addition to regular argumentation.
Dialectical Obligations
Arguers owe each other honesty, clarity, fairness, and respect for the burden of proof and critical questions. The principle of charity is a major part of discourse and requires interpreting opponents’ arguments in their strongest plausible form before accusing “fallacy”. Over‑eager fallacy‑spotting often violates this duty, turning critique into a dominance game. All too often in more unscrupulous corners of the internet, argumentation is only a dominance game.
From Identification to Analysis
A mature philosophy of fallacies emphasizes reconstruction of arguments (premises, conclusions, assumptions), evaluation via argumentation schemes and critical questions, and cultivating intellectual virtues (open‑mindedness, humility) rather than mere fallacy avoidance. This shifts focus from “spot the fallacy” to “assess the reasoning in context”. It requires a certain noble disposition to think the best of poor argumentation.
Fallacy Backfiring
There may be a risk that teaching fallacies makes people better at committing them—that is, more skilled at disguising defects or accusing others selectively. This underscores the need for ethical enactment, not just pattern recognition. With this being said, I am not sure anyone actually intentionally uses fallacies in this manner. Rather, it is more likely that they issue a statement that they know might be accused of being a fallacy and hope for the best, but then again the way things are changing I suppose anything can be possible.
AI and LLMs (“Chatbots”)
Large language models often reproduce human fallacies present in training data and can generate superficially cogent but fallacious arguments. Using AI to analyse arguments may help scale detection, but cannot exactly replace human judgement about context and appropriate norms. On a related note, this extends to AI analysis of arguments beyond fallacies, where intention is falsely inferred from training data. It is very difficult to get through to someone who has been convinced by a chatbot that their interlocutor is arguing in bad faith due to its guessing their intention.
Criticisms of Fallacy Theory
Critics argue that “fallacy” is too vague and context‑sensitive for a unified theory, is ultimately biased towards adversarial, Western dialectics, and is overly focused on error rather than excellence in reasoning. Some propose eliminativism: abandon the category in favor of more precise normative descriptions (for example, specific rule violations, scheme misuses). This will not be necessary provided that fallacies are identified reflectively, since fallacies provide us with a useful way to prevent epistemic collapse.
Alternative Frameworks
Instead of spotting fallacies, alternative frameworks can be implemented. These include Bayesianism, where reasoning errors are modeled as violations of probabilistic norms rather than fallacy labels; virtue argumentation, which focuses on intellectual virtues and vices instead of fallacy avoidance; and (Walton’s) argument schemes, where fallacy lists are replaced with schemes plus critical questions, emphasizing constructive evaluation. These approaches retain the insights of fallacy theory while avoiding its oversimplifications.
Post‑Fallacy Discourse
New media environments generate new fallacious patterns (for example, sealioning, motte‑and‑bailey, Gish gallop—these tend not to be in standard lists). At the same time, “post‑truth” dynamics threaten the very norms that make fallacy talk meaningful. Once again, in this context, fallacy theory can serve as a bulwark against epistemic collapse, but only if used reflectively. And I suppose we need all the help we can get in fighting a post-truth catastrophe.
Conclusion
In summary, an effective philosophy of fallacies does not abolish the concept but disciplines its use: caution, nuance, and theoretical awareness are essential if “fallacy” is to remain a tool for understanding reasoning rather than a weapon in rhetorical combat. And finally, almost all of my audience will already understand the basics of what I am speaking of, hence the “intermediate” label. I am mostly writing this because I am terrifically bored and am waiting for something.
Thanks.
