Q
Question
Understand why fallacies matter — and why naming them is a philosophical act, not a rhetorical trick

Most bad arguments do not announce themselves as bad. They arrive dressed in the clothes of good reasoning — confident, structured, emotionally compelling. They look like they are giving you a reason to believe their conclusion. The problem, on inspection, is that they are not. The premises are irrelevant, or the evidence is too thin, or the terms shift halfway through, or a false choice has been smuggled in where many options actually exist.

A fallacy is a recurring pattern of defective reasoning. The word matters. Not every mistake in thinking is a fallacy — a fallacy is a pattern, one that recurs often enough and is recognisable enough to have earned a name. The catalogue of fallacies that follows is the accumulated critical intelligence of two and a half thousand years of philosophy, rhetoric, and logic: a list of the moves that look like argument but aren't.

But before we catalogue them, a crucial distinction.

Formal versus informal fallacies

Not all fallacies are the same kind of thing. There are two fundamentally different ways an argument can go wrong, and they have different names.

Formal fallacy
Invalid logical form
The argument's conclusion does not follow from its premises because the logical structure is broken — regardless of what the premises are about. The error is in the form, detectable by inspection of the argument's shape alone.
"All dogs are mammals. All cats are mammals. Therefore all dogs are cats." — Invalid regardless of what terms you substitute for dogs, cats, and mammals.
Detected by: testing whether the conclusion must follow from the premises (Article A3 — validity)
Informal fallacy
Defective content or context
The argument's logical form may be technically valid, but it fails because the premises are irrelevant, the evidence is inadequate, an assumption is unwarranted, or language is used ambiguously. The error depends on the argument's content, not just its shape.
"You shouldn't listen to her argument about climate policy — she drives a large car." — The form could be valid (if the premise were relevant), but the premise is irrelevant to the argument's quality.
Detected by: asking whether premises are relevant, sufficient, well-grounded, and clearly used

This article is entirely about informal fallacies. They are organised into four categories based on the type of error they make — a classification that connects directly to the evaluative vocabulary of A1–A3:

Relevance fallacies offer premises that are simply irrelevant to the conclusion — they fail the soundness test because the premises, even if true, give no support to the conclusion at all.

Weak induction fallacies offer premises that are relevant but insufficient — the evidence is real but too thin, too biased, or too distorted to make the conclusion probable.

Presumption fallacies sneak unwarranted assumptions into the argument's structure — the conclusion depends on a hidden premise that has not been established and may be false.

Clarity fallacies exploit ambiguity in language — either misrepresenting an opponent's argument or using the same word with two different meanings to make an inference appear to follow when it does not.

With that map in hand, let us examine the twelve most important informal fallacies.

U
Unpack
Map the twelve most important informal fallacies across four categories — with examples and precise diagnoses
Category 1 — Relevance Fallacies
The premise offered is simply beside the point
Ad Hominem
argumentum ad hominem — "argument against the person"
Relevance
Why it fails: The premise attacks the person making the argument rather than the argument itself. A person's character, motives, or circumstances are almost always irrelevant to whether their argument is sound. Even a dishonest person can make a valid argument; even a saint can make an unsound one. The fallacy trades on the psychological tendency to distrust sources rather than evaluating the logic on its own terms.
"We can dismiss her critique of the government's economic policy — she's a member of the opposition party and has a personal interest in making the government look bad."
The critic's political affiliation tells us to check her argument carefully for bias — but it does not refute the argument. To refute it, you must show that a premise is false or the inference is invalid. This conflates the reliability of a source with the quality of what they are arguing.
Appeal to Authority
argumentum ad verecundiam — "argument to reverence"
Relevance
Why it fails: Expert testimony can be legitimate evidence — but only when the authority is genuinely expert in the relevant field, the field admits of expert knowledge, and there is no strong contrary expert consensus. When authority is invoked outside these conditions — a physicist's view on economics, a celebrity's view on medicine — it provides no genuine support for the conclusion. The premise (expert says X) is irrelevant unless the expertise applies.
"This alternative therapy must be effective — a Nobel Prize-winning physicist has endorsed it."
Physics expertise does not transfer to medicine. Even within medicine, one expert's endorsement against a strong contrary consensus carries little weight. The premise is present but not relevant to the domain of the conclusion. Note: citing established expert consensus on matters within a field is not this fallacy — it is legitimate inductive reasoning.
Appeal to Popularity
argumentum ad populum — "argument to the people"
Relevance
Why it fails: The number of people who hold a belief is irrelevant to whether the belief is true. For most of human history, the majority believed the Earth was flat and the Sun orbited it. Popular belief is a social fact; truth is not determined by voting. The premise (many people believe X) simply does not support the conclusion (X is true).
"Most people intuitively feel that we have a duty to give to our own family before strangers. Therefore Singer's argument that we should give as much as possible to distant strangers is simply wrong."
The widespread intuition is data about how people feel — it is not an argument for why the intuition is correct. To refute Singer, you need to show that one of his premises is false or that his inference is invalid. Pointing to a contrary intuition raises a challenge, but it does not by itself resolve the philosophical question.
Tu Quoque
tu quoque — "you too" / whataboutism
Relevance
Why it fails: Pointing out that the person making an argument is themselves guilty of inconsistency does not address the argument. "You also do X" is a premise about the arguer's behaviour, not about the soundness of the argument. Hypocrisy is real and worth noting — but it cannot refute a philosophical claim. A doctor who smokes can still be right that smoking causes cancer.
"You argue that we should reduce carbon emissions — but you flew to an overseas conference last month. Your argument is hypocritical and therefore invalid."
The arguer's travel habits are irrelevant to whether the scientific and ethical case for emissions reduction is sound. Hypocrisy is a separate moral charge; it does not touch the argument. To refute the emissions argument, show that a premise is false or the inference fails. The ad hominem and tu quoque are closely related — both attack the person, not the argument.
Category 2 — Weak Induction Fallacies
Evidence is offered, but it is too thin, biased, or distorted to support the conclusion
Hasty Generalisation
secundum quid — "according to something"
Weak induction
Why it fails: A general conclusion drawn from too small, too limited, or too unrepresentative a sample makes the inductive argument weak. The premises may be entirely true — the observed instances may really be as described — but there are not enough of them, or they are too similar, to support a generalisation. As Article A2 showed, inductive strength depends critically on sample size and diversity.
"Every philosopher I've studied has been European or North American. Therefore the great philosophical tradition is essentially a Western one."
The premise reflects a genuine but parochially limited sample — shaped by which philosophical traditions have been historically translated, published, and included in Western curricula. It ignores Confucian philosophy, Buddhist philosophy, Nyāya logic from India, African philosophy, and Islamic philosophy. The sample is not just small but systematically biased toward one tradition.
Post Hoc Ergo Propter Hoc
post hoc ergo propter hoc — "after this, therefore because of this"
Weak induction
Why it fails: Sequence is not causation. That B followed A does not mean A caused B — the correlation may be coincidental, or both A and B may be caused by a third factor, or the causal direction may be reversed. This is one of the most common errors in everyday reasoning and one of the most damaging in science, medicine, and policy. Inductive strength requires more than temporal sequence: it requires controlled comparison, ruling out confounders, and ideally an identified causal mechanism.
"After the introduction of universal education, crime rates rose. Therefore universal education causes crime."
Both trends could share a common cause (urbanisation and industrial change), the causal direction could be inverted (rising crime preceded education expansion in some areas), or the correlation could be spurious. No controlled comparison has been made; no mechanism has been identified. The inductive leap from correlation to causation is entirely unsupported.
Slippery Slope
(informal name — no standard Latin)
Weak induction
Why it fails: The argument claims that accepting position A will inevitably lead, step by step, to an extreme and unacceptable position Z — without establishing that the intermediate steps are likely or even plausible. When the steps are genuinely probable and each follows from the last, a slippery slope argument may be legitimate. When the chain of consequence is merely asserted rather than argued, each link needs to be defended. The fallacy lies in treating an imagined cascade as though it were established fact.
"If we allow voluntary euthanasia for the terminally ill, we will soon be pressuring elderly people who feel they are a burden to end their lives, and before long the state will be deciding who lives and who dies."
Each step in this chain — from voluntary to coerced to state-directed — is a substantive empirical and political claim that requires independent support. Evidence from jurisdictions with legalised voluntary euthanasia (the Netherlands, Belgium, Oregon) would be relevant, and that evidence is contested. The argument as stated simply asserts the cascade. Whether this is fallacious depends on whether the causal links are established — not on whether the conclusion is unpleasant.
Category 3 — Presumption Fallacies
An unwarranted assumption is hidden in the argument's structure
Begging the Question
petitio principii — "assuming the beginning"
Presumption
Why it fails: The argument's conclusion is hidden inside one of its premises — so the argument is circular. The conclusion follows from the premises only because the premises already assume the conclusion. Such an argument is technically valid (the conclusion does follow!) but cannot convince anyone who does not already accept the conclusion, because accepting the premise requires having already accepted what the argument is supposed to prove. Note: in everyday usage "begging the question" has come to mean "raising the question" — this is a distinct, philosophical meaning.
"The Bible is the word of God. The Bible says God exists. Therefore God exists. We know the Bible is the word of God because God would not deceive us, and God exists."
The claim that the Bible is the word of God assumes God exists — which is what the argument is trying to prove. The conclusion (God exists) is embedded in the premise (Bible = word of God). The argument goes in a circle. Someone who already accepts both the premise and the conclusion will not notice the circularity; someone who does not accept the conclusion will immediately see that the premise is no less in need of justification than the conclusion.
False Dilemma
also: false dichotomy, bifurcation fallacy
Presumption
Why it fails: The argument presents only two options when more exist, then eliminates one to force the conclusion that the other must be accepted. The hidden premise — that only these two options are available — is false. This fallacy is particularly powerful rhetorically because eliminating one unacceptable option makes the remaining one seem inevitable. The antidote is always to ask: are these really the only possibilities?
"Either you support this war, or you support the terrorists. You don't support the terrorists — so you must support this war."
This presents a false dichotomy. One can oppose both the specific military action and terrorism; one can support diplomacy, sanctions, or other alternatives; one can oppose the war on strategic grounds while condemning terrorism. The hidden premise — that only two positions exist — is simply false. Once exposed, the argument collapses: accepting the second option (rejecting terrorism) does not force acceptance of the first (supporting this specific war).
Loaded Question
also: complex question, plurium interrogationum
Presumption
Why it fails: The question presupposes something that has not been established — so any direct answer will implicitly concede that presupposition. The classic form forces a "yes or no" on a question that contains a hidden unproven assumption, so that answering the question at all is itself a trap. The correct response is always to identify and reject the presupposition before answering.
"Have you stopped cheating on your exams?" (asked of a student who has never cheated)
Answering "yes" concedes that cheating happened. Answering "no" concedes it is still ongoing. Either answer accepts the false presupposition. The only honest response is to reject the question's premise: "That question presupposes I was cheating, which I was not." In philosophical contexts, this appears as: "Do you still believe the soul is immortal?" — directed at someone who never held that view.
Category 4 — Clarity Fallacies
The argument exploits ambiguity in language or misrepresents what is being argued
Equivocation
aequivocatio — "equal naming"
Clarity
Why it fails: A key term is used in two different senses within the same argument, so the conclusion appears to follow when it actually does not. The argument is valid only if the term has the same meaning throughout; the shift in meaning creates the appearance of a logical connection that does not exist. Equivocation is one of the subtlest fallacies because it can be hard to spot when the two meanings are close.
"The law of gravity says that all bodies obey the law. Criminals are bodies. Therefore criminals must obey the law."
"Law" is used in two senses: a descriptive natural law (which all physical objects necessarily follow) and a prescriptive legal law (which humans can choose to violate). "Bodies" shifts from physical objects to persons. The argument appears valid because the same words are used, but the concepts are entirely different. Replacing "law" with its two distinct meanings immediately exposes the non-sequitur.
Straw Man
homo stramineus — "straw man"
Clarity
Why it fails: The arguer misrepresents an opponent's position — making it weaker, more extreme, or more easily refuted — and then attacks that misrepresented version. Even a perfectly sound refutation of the distorted version leaves the original argument entirely untouched. The straw man violates the principle of charity (A1): you must engage with the strongest version of an argument, not a fabricated weak version. It is a clarity fallacy because the argument being addressed is not the argument that was actually made.
"Philosophers who argue for free will are claiming that human choices are completely uncaused and random. But random events cannot constitute genuine agency — so free will is incoherent."
This misrepresents libertarian free will, which holds that choices are neither causally determined by prior events nor random, but involve agent causation — a distinct third option the straw man ignores. Compatibilists don't claim choices are uncaused at all. Even if the refutation of "random, uncaused choices" is sound, it refutes a position that virtually no serious philosopher holds. The original argument is not addressed. Identifying the distortion is the first step in any philosophical response to a straw man.
E
Examine
Catch fallacies in the wild — three worked analyses showing the full diagnostic process

Naming fallacies is easy. The skill philosophy actually demands is different: encountering an argument that feels compelling, pausing, and asking systematically which of the criteria from A1–A3 it fails to meet. The following three cases are drawn from political philosophy, applied ethics, and a real pattern of historical reasoning. Each is analysed using the full spotter's method.

Case 1 — Philosophical debate on capital punishment Relevance — Ad Hominem
The argument as encountered
"Professor X argues that capital punishment is never morally justified. But Professor X is a committed pacifist who opposes all forms of state violence — she can hardly be expected to assess this question objectively. Her argument should therefore not be taken seriously."
1
What is the conclusion? What are the premises?
Conclusion: Professor X's argument should not be taken seriously. Premise: She is a committed pacifist with a prior ideological commitment against state violence. This is a single-premise argument — and everything depends on whether that premise is relevant.
2
Is the premise relevant to the conclusion?
No. The professor's ideological commitments are relevant to her potential biases — but biases do not make arguments unsound. The argument about capital punishment either has valid form and true premises or it does not. Her personal commitments do not change the logical relationship between her premises and her conclusion. We need to engage with the argument, not the arguer.
3
What should the response be?
Identify the fallacy explicitly: "This is an ad hominem — it attacks the source of the argument rather than the argument itself. Even if Professor X's pacifism creates a bias, that tells us to scrutinise her premises carefully, not to dismiss the argument without examination. To refute her argument, you need to show that a specific premise is false or that the conclusion does not follow."
Diagnosis Ad Hominem — Relevance fallacy The premise (she has an ideological commitment) is real but logically irrelevant to whether her argument is sound. The criterion violated: the premise provides no support for the conclusion (A3 — soundness).
Case 2 — An argument about drug policy Presumption — False Dilemma
The argument as encountered
"Either we maintain strict criminal penalties for drug possession, or we legalise all drugs and accept the social chaos that follows. Since we cannot accept social chaos, we must maintain strict criminal penalties."
1
What is the conclusion? What are the premises?
Conclusion: Strict criminal penalties must be maintained. P1: Either strict criminalisation or legalise-all-drugs — no middle ground. P2: Legalising all drugs leads to social chaos. P3: Social chaos is unacceptable. The argument is deductively valid — if P1–P3 are true, the conclusion follows. The question is whether the premises are true.
2
Is P1 true?
No — and this is the false dilemma. The binary presented ignores a range of existing and proposed middle positions: decriminalisation without legalisation (Portugal's policy approach); legalisation of specific substances with strict regulation (cannabis in multiple jurisdictions); harm reduction models that neither fully criminalise nor fully legalise; differential treatment of different substances. The "either/or" of P1 is a false dichotomy.
3
What is the right response?
"This argument presents a false dilemma: the choice between strict criminalisation and full legalisation without regulation is not exhaustive. A third option — decriminalisation with regulated access to specific substances, as adopted in several jurisdictions — undermines the dilemma's premise entirely. Once P1 is rejected, the argument fails even if P2 and P3 are true."
Diagnosis False Dilemma — Presumption fallacy The argument is valid in form but unsound: P1 presents only two options when many more exist. The criterion violated: a key premise is false (A3 — soundness). Once P1 is shown false, the deductively valid argument still fails.
Case 3 — A historical pattern of reasoning Weak induction — Post Hoc & Hasty Generalisation
The reasoning pattern
"Since women gained access to higher education in the early twentieth century, the birth rate has steadily declined and family breakdown has increased. The evidence clearly shows that women's education damages society's fundamental social structures."
1
What inductive claims are being made?
Two: (a) women's education caused the decline in birth rates; (b) women's education caused increased family breakdown. Both claims infer causation from the temporal sequence of events — two hallmarks of post hoc ergo propter hoc. Additionally, "increased family breakdown" may itself be a contested or undefined measure.
2
What alternative explanations exist? What would a strong inductive argument need?
The same time period saw: industrialisation and urbanisation, contraception becoming available, women's economic independence, changing social norms, housing costs, and medical advances in family planning. Any or all of these could explain the birth rate change — and all occurred simultaneously. A strong causal claim requires controlled comparison, ruling out alternative explanations, and ideally an identified mechanism. None of that appears here. The correlation also rests on undefined and potentially contested terms like "family breakdown."
3
What does the evidence from better-designed studies show?
Cross-national data consistently shows that higher women's education correlates with improved child health outcomes, lower infant mortality, and — in developing economies — better long-term social stability. These findings involve controlled comparisons that this argument entirely lacks. The claim as stated is not merely weak — it is contradicted by the stronger evidence available.
Diagnosis Post Hoc + Hasty Generalisation — Weak induction The inductive argument is weak on both counts: it confuses temporal sequence with causation, and draws a sweeping causal conclusion without ruling out confounders. The criterion violated: the premises, even if true, do not make the conclusion probable (A3 — strength/cogency).
S
Synthesise
Build a systematic method for detecting fallacies in any argument — applying the full A1–A4 toolkit

The fallacies in Unpack are not a list to memorise and match. Used that way, fallacy-spotting becomes a parlour game — slapping labels on arguments without doing the philosophical work of showing precisely why each label applies. The point of having names for fallacies is to give you a vocabulary for a structured diagnostic process.

The method below extends the six-step evaluation framework from Article A3 with four targeted fallacy-detection questions. These questions correspond directly to the four categories introduced in the Question tab. Together, they form a complete analytical procedure — the "fallacy spotter's method."

The Fallacy Spotter's Method — Applied Across A1–A4
1
Reconstruct the argument in standard form and interpret charitably (A1)
Before looking for fallacies, ensure you have the best version of the argument. Attacking a weak version when a stronger one is available is itself the straw man fallacy. Write out P1, P2 … ∴ C explicitly before proceeding.
2
Are the premises relevant to the conclusion? (Relevance fallacies)
For each premise, ask: if this premise were true, does it give any genuine support to the conclusion? Premises about the arguer's character, motives, or circumstances are almost always irrelevant. Premises about what people believe do not establish what is true. Premises drawn from authorities outside their domain of expertise carry no weight.
Ad hominem Appeal to authority Ad populum Tu quoque
3
Is the evidence sufficient to support the conclusion? (Weak induction)
For inductive arguments: is the sample large enough? Is it representative, or biased toward one type of case? Has correlation been confused with causation — and have alternative explanations been ruled out? Has each step in a causal chain been independently established? The conclusion of an inductive argument should not go further than the evidence genuinely supports.
Hasty generalisation Post hoc Slippery slope
4
Are there unwarranted hidden assumptions? (Presumption fallacies)
Identify all the unstated premises the argument requires (enthymemes — from A1). Ask whether each is established or merely assumed. Does the argument assume only two options when more exist? Does a premise quietly assume the conclusion? Does a question assume something that has not been proved?
Begging the question False dilemma Loaded question
5
Are key terms used consistently? Is the argument I'm addressing actually the argument that was made? (Clarity fallacies)
For each key term: does it carry the same meaning in every premise and in the conclusion? If the meaning shifts, replace each occurrence with the specific meaning being used and see if the argument still holds. Also: is the argument I'm about to refute actually the one that was offered, or a weaker distorted version of it?
Equivocation Straw man
6
Name the fallacy — and explain exactly why it fails in A1–A3 terms
In examined responses, it is never enough to say "this is an ad hominem" — you must explain precisely which criterion from A1–A3 is violated and why. "This is an ad hominem fallacy. The premise about the arguer's character is irrelevant to whether the argument is sound — it fails to give any support to the conclusion, meaning the argument is unsound even if valid in form."

One further caution deserves emphasis: the goal is not to "win" by attaching a label. It is to understand, precisely, why an argument fails — and to be honest about when an argument does not fail. Some slippery slope arguments are legitimate (when the causal chain is established). Some appeals to authority are legitimate (when genuine expert consensus exists within a relevant field). Some generalisations are strong (when the sample is large, diverse, and controlled). The method should make you more precise, not more dismissive.

Philosophy demands that you take arguments seriously — including arguments you disagree with. The principle of charity requires that you identify the best version of an argument before you evaluate it. Fallacy-naming should be the last step in a careful analysis, not a substitute for one.

T
Transfer
See how fallacy recognition applies far beyond philosophy — in politics, science, and everyday critical thinking

Fallacies in political discourse

Political debate is arguably the domain in which informal fallacies are most densely and most consequentially deployed. Several structural features of democratic politics make this almost inevitable: politicians must persuade large audiences quickly, emotional engagement drives electoral outcomes more reliably than logical precision, and the media environment rewards simplification over nuance.

The false dilemma is the political fallacy par excellence. "You're either with us or against us," "you either support border security or you support open borders," "you either back economic growth or you care about the environment" — these framings do political work precisely by closing down the space for nuanced positions. Once a binary is accepted, the elimination of one unacceptable option forces acceptance of the other. Recognising the false dilemma means always asking: where are the third, fourth, and fifth options?

The ad hominem is the second great weapon of political discourse. Rather than engaging with a policy argument, political opponents routinely attack the motives, character, or associations of the person making it. This is psychologically effective — it is easier to distrust a source than to refute an argument — but it is always logically empty. A corrupt politician who proposes a good policy has still proposed a good policy. The corruption is a separate matter to be addressed separately.

Fallacies in scientific communication

Scientific reasoning itself is disciplined to avoid these errors — the controlled experiment is specifically designed to rule out post hoc reasoning, and peer review is designed to catch hasty generalisations. But the communication of science to the public is deeply vulnerable to fallacious interpretation.

Headlines regularly commit the post hoc fallacy: "Coffee drinkers live longer — here's why." The observed correlation (coffee consumption associated with longer life) is reported as though causation has been established, eliding the enormous methodological work required to move from correlation to causal conclusion. Epidemiology routinely produces correlations; establishing causation requires controlled trials, dose-response relationships, plausible mechanisms, and replication across diverse populations.

The appeal to authority becomes particularly consequential when it is invoked to dismiss scientific consensus rather than support a minority view. "Many scientists disagree with the consensus on climate change" — when the number of such scientists is vanishingly small, and their expertise is not in climate science — is an appeal to manufactured authority. Understanding how legitimate expert testimony differs from the appeal to authority fallacy is one of the most important skills for scientific literacy.

A note on charity and critical generosity

The tradition of identifying fallacies stretches back to Aristotle's Sophistical Refutations — a catalogue of the tricks used by sophists to win arguments they should have lost. The impulse to name and expose bad reasoning is as old as philosophy itself, and it is a genuinely important intellectual practice.

But there is a misuse of fallacy-naming that philosophy should resist: using labels as a substitute for engagement. When a fallacy label is applied too quickly — before the argument has been charitably reconstructed, before the specific premise or inference that fails has been identified — it becomes a rhetorical move, not a philosophical one. The slippery slope becomes a thought-terminating cliché rather than a substantive observation that the causal links have not been established. "That's an ad hominem!" becomes a way to refuse to engage with a legitimate point about the arguer's track record.

The principle of charity, introduced in A1, is the corrective. Always reconstruct the argument in its strongest form. Identify exactly which step fails. Name the fallacy after showing why it applies. This sequence — reconstruction, analysis, diagnosis — is philosophy, not rhetoric.

Connecting to Article A5

You have now worked through the informal fallacies — the ways arguments fail because of problems with their content: irrelevant premises, insufficient evidence, unwarranted assumptions, or ambiguous terms. These errors are detectable by careful attention to meaning and logical relationship, without any formal symbolic machinery.

Article A5 asks a different question: can we represent the logical structure of arguments in a purely formal, symbolic language — one where validity can be checked mechanically, without depending on your ability to "see" whether a conclusion follows? This is the project of propositional logic, and it takes you into a different mode of philosophical reasoning altogether: precise, symbolic, and provably complete in its own domain.

The question to carry with you into Article A5
Throughout A1–A4, we have been testing arguments by asking whether a conclusion "follows" from premises — and relying on our intuition to tell us when it does. Is there a more rigorous way? Can we represent logical relationships in a symbolic language precise enough to check validity by mechanical rules alone, without relying on intuition at all?
Article A5 introduces propositional logic — the formal language that Aristotle began and Frege, Russell, and Whitehead completed. It is the foundation of all formal reasoning, and it will change how you understand the arguments you have already encountered in this package.