Formal Arguments and Common Fallacies: A Practical Guide

What It Is

An argument is a structured piece of reasoning in which one statement (the conclusion) is claimed to follow from one or more other statements (the premises). Logic is the study of which arguments do and which do not deliver this claimed support. Practical argument-reading is the skill of identifying the structure in a piece of natural-language text and then assessing whether the reasoning is good.

This page has two halves.

Half 1: How to read an argument. Identify premises and conclusion. Identify the kind of inferential connection claimed (deductive, inductive, abductive). Assess validity (does the conclusion follow if the premises are true?) and soundness (are the premises actually true?).

Half 2: The standard fallacy catalog. Patterns of bad reasoning common enough to have names. Knowing the catalog is a defensive skill: many bad arguments belong to a recognized family. Diagnosing the family is most of the work of refutation.

This page assumes What Is Logic? and Validity vs Soundness for the basic concepts; it extends them toward practical argument-reading. For the formal language in which validity is precisely defined, see Propositional Logic and Predicate Logic. For deductive proof technique on the math side, see Basic Logic and Proof Techniques.

Half 1: Reading an Argument

Step 1: Locate the conclusion

Find the statement the author is trying to establish. Conclusion-indicator words: therefore, hence, thus, so, it follows that, consequently, accordingly. Sometimes the conclusion is stated first ("Capital punishment should be abolished. Here is why: ...") and sometimes last ("Studies show X. The legal regime supports Y. Therefore, the system is unjust"). The conclusion is the load-bearing claim the rest of the argument supports.

Step 2: Locate the premises

Find the statements offered as support. Premise-indicator words: because, since, for, given that, assuming that, on the grounds that, in light of. Premises can be themselves complex. They can be stated explicitly or assumed implicitly. Implicit premises (called suppressed or enthymematic premises) are the most common source of disagreement: people argue about the implicit premise, not the stated ones.

Step 3: Diagram the inferential structure

The simplest argument has the form: $P_1, P_2, \ldots, P_n \therefore C$ . More complex arguments have sub-arguments: intermediate conclusions that serve as premises for further inferences. Diagram these as a tree:

\frac{P_1 \quad P_2}{C_1} \quad \frac{P_3 \quad P_4}{C_2} \quad \therefore C

The argument has two sub-arguments converging on the main conclusion. Identifying the structure is half the work.

Step 4: Classify the inferential kind

Three families.

Deductive. The conclusion is claimed to follow with necessity: if the premises are true, the conclusion must be true. Standard mathematical proofs are deductive.
Inductive. The conclusion is claimed to follow with high probability: if the premises are true, the conclusion is likely to be true. Generalizing from a sample, predicting tomorrow from yesterday, inferring a pattern from cases.
Abductive. The conclusion is claimed to be the best explanation of the premises (the evidence). The premises are observations; the conclusion is the hypothesis that, if true, would best explain them. Most everyday and scientific reasoning is abductive.

Each kind has its own standards. Deductive arguments are assessed by validity. Inductive and abductive arguments are assessed by strength (how well the premises support the conclusion) and other criteria (representativeness of evidence, explanatory power of the hypothesis).

Step 5: Assess validity (for deductive arguments)

A deductive argument is valid iff there is no possible situation in which all premises are true and the conclusion is false. Validity is a property of form; it does not depend on whether the premises are actually true. The argument

\text{All philosophers are immortal.} \quad \text{Socrates is a philosopher.} \quad \therefore \text{Socrates is immortal.}

is valid (the form is: all $F$ are $G$ ; $x$ is $F$ ; therefore $x$ is $G$ ). The first premise is false; the argument is unsound but still valid.

See Validity vs Soundness for the four-case breakdown (valid + true premises, valid + false premises, invalid + true premises, invalid + false premises) and worked exercises.

Step 6: Assess soundness (for deductive arguments)

A sound deductive argument is valid AND has actually-true premises. Sound arguments deliver their conclusions; valid-but-unsound arguments do not. Most real-world disputes are about premise-truth, not validity: people accept the inferential connection but contest one of the premises.

Step 7: For inductive / abductive arguments, assess strength

For inductive arguments: is the sample representative? Are the cases relevantly similar? Is the inference base large enough? For abductive arguments: is the hypothesis genuinely the best explanation, or are there alternatives that explain the evidence at least as well? Inference to the best available explanation is not inference to the best possible explanation; if the search over candidate hypotheses was shallow, the abductive inference is weak.

Half 2: The Standard Fallacy Catalog

A fallacy is a systematic pattern of bad reasoning. Some fallacies are formal (the argument form is invalid; the fallacy is in the inferential structure). Others are informal (the form may be valid but a substantive error in content, context, or relevance makes the argument bad). The distinction is useful but not sharp; some classifications place ambiguous cases in either bucket.

Formal fallacies

Affirming the consequent.

P \to Q. \quad Q. \quad \therefore P.

The argument form is invalid: the conditional $P \to Q$ does not say that $Q$ requires $P$ as cause; $Q$ might hold for many other reasons. Example: "If it rained, the street is wet. The street is wet. Therefore it rained." The street might be wet because someone watered the lawn, the fire hydrant burst, a car spilled water, and so on.

Denying the antecedent.

P \to Q. \quad \neg P. \quad \therefore \neg Q.

Same kind of mistake from the other direction. The conditional says $P$ suffices for $Q$ , not that $P$ is necessary. Example: "If she studied, she passed. She didn't study. Therefore she didn't pass." She might have passed because the test was easy or because she already knew the material.

Both formal fallacies are recognizable by writing out the structure in symbolic form. The valid versions are modus ponens ( $P \to Q, P \therefore Q$ ) and modus tollens ( $P \to Q, \neg Q \therefore \neg P$ ). The fallacies confuse the position of the negation or the direction of inference.

Informal fallacies

Ad hominem (against the person). Rejecting an argument by attacking the person making it, rather than by addressing the argument's premises or inferential structure. "You can't trust her on climate; she's funded by environmental groups." Why it looks compelling: the speaker's interests are sometimes relevant to credibility. Why it fails: even biased speakers can give valid arguments with true premises; the argument should be evaluated on its merits.

A subtype, tu quoque ("you too"), rejects an argument by pointing out the arguer's hypocrisy. "You smoke yourself; you can't tell me smoking is bad." The arguer's hypocrisy does not bear on the validity or truth-of-premises of the argument.

Equivocation. Using a term in two different senses across the argument, exploiting the ambiguity to slide between meanings. "The end of life is its goal; death is the end of life; therefore death is life's goal." The first "end" means purpose (telos); the second means termination. The argument trades on the shift.

Detection: identify the key terms, check whether they are used consistently throughout the argument. Most equivocation fallacies involve a term with both a thin and a thick reading.

Straw man. Misrepresenting an opponent's position in a weaker or more extreme form, then refuting the misrepresentation. "Critics of the immigration bill say we should open the borders entirely. That's absurd. So the bill should pass." If the critics' actual position was "the bill is too restrictive on a particular technical clause," the response refutes nothing. The remedy is to argue against the strongest version of the opposing position (the steelman), not the weakest.

Slippery slope. Arguing that one action will lead, by a chain of causal steps, to an unacceptable outcome, without establishing the strength of any link in the chain. "If we allow this exception, the next thing we know we'll have abandoned the principle entirely." Why it looks compelling: chains of small steps do sometimes lead to large changes. Why it fails: the argument requires that each step in the chain be highly probable, and the rhetorical use of slippery-slope arguments typically does not establish this for any link, much less all of them.

A causal slippery slope (each step actually causes the next) is sometimes legitimate when each step's high probability is genuinely supported. A logical slippery slope (each step is logically forced by the previous) is almost never legitimate.

False dichotomy (false dilemma). Presenting two options as exhaustive when they are not. "You're either with us or against us." "Either we accept full censorship or we accept total chaos." Real situations almost always have intermediate options or different framings entirely. The remedy is to ask: what are the actual options here?

Appeal to authority. Citing an authority figure's pronouncement as evidence for a claim outside her domain of expertise, or in a domain where her authority is contested. "Linus Pauling won two Nobel prizes, and he says vitamin C cures the common cold; therefore it does." Nobel-winning chemists are not authorities on clinical medicine. Why it looks compelling: in domains where the authority does have expertise, citing her is legitimate. Why it fails: the citation only works when the authority's expertise actually covers the claim.

Post hoc ergo propter hoc (after this, therefore because of this). Concluding from temporal sequence that the earlier event caused the later. "I started taking the supplement, and my cold cleared up the next week. The supplement cures colds." Colds clear up on their own; the temporal correlation does not establish causation. The remedy is to ask what would have happened without the supposed cause (the counterfactual baseline).

Begging the question (petitio principii). Using the conclusion (sometimes restated) as one of the premises. "We know God exists because the Bible says so, and the Bible is true because it is the word of God." The conclusion (God exists) is presupposed in the second premise. The argument is technically valid (if the premises are true, the conclusion must be); it just gives no reason to accept the premises.

A subtle version: circular arguments where the supporting premises require the conclusion to be established by some independent route. Hard to spot because the circle can be wide.

Hasty generalization. Drawing a general conclusion from a sample too small or too unrepresentative to support it. "I met three rude waiters in this city; the people in this city are rude." The remedy is to ask whether the sample size and selection adequately support the generality.

Composition and division.

Composition: inferring from a property of parts to a property of the whole. "Every brick in this wall is light; therefore the wall is light." Light bricks make a heavy wall when there are enough of them.
Division: inferring from a property of the whole to a property of the parts. "The orchestra plays brilliantly; therefore each musician plays brilliantly." A merely competent ensemble can produce a brilliant whole.

Genetic fallacy. Rejecting a claim because of its origin rather than its content. "Marx's theory of capital was developed by a man who lived on his friends' charity; therefore the theory is unsound." Origin can be evidence about credibility in some cases, but the content of the theory must be assessed on its own. A useful idea can come from an unreliable source; a bad idea can come from a reliable one.

Appeal to emotion / appeal to consequences. Substituting an emotional response (or an undesirable practical consequence of believing $P$ ) for evidence about whether $P$ is true. "If determinism is true, nobody is morally responsible, and that would undermine the criminal justice system; therefore determinism is false." Whether determinism is true depends on the metaphysics, not on what conclusion we would like to be able to draw about responsibility.

Worked Practice Exercise

For each argument: state the conclusion, identify the premises, classify the kind of inference, assess validity (for deductive) or strength (for inductive / abductive), and identify any fallacy.

Argument 1.

If interest rates rise, housing prices fall. Housing prices are falling. Therefore interest rates are rising.

Conclusion: interest rates are rising. Premises: (i) if interest rates rise, housing prices fall; (ii) housing prices are falling. Kind: deductive. Validity: invalid. Form: affirming the consequent. The conditional says rising rates imply falling prices, not the converse. Prices could fall for many other reasons (rising unemployment, oversupply, regulatory changes).

Argument 2.

Every philosopher I have met in this department has been a Kantian. Therefore every philosopher in this department is a Kantian.

Conclusion: every philosopher in this department is a Kantian. Premises: every philosopher the speaker has met in the department is a Kantian. Kind: inductive (generalizing from a sample). Strength: depends on the sample. If the speaker has met three of fifty department philosophers, the inference is hasty generalization. If the speaker has met forty-eight of fifty, it is reasonable but the remaining two might falsify the generalization.

Argument 3.

Either we adopt the proposed reform or the system collapses. We cannot let the system collapse. Therefore we must adopt the proposed reform.

Conclusion: we must adopt the reform. Premises: (i) either reform or collapse; (ii) cannot let collapse. Kind: deductive. Validity: valid (form is disjunctive syllogism: $A \lor B, \neg B \therefore A$ ). Soundness: depends on premise (i). The premise looks like a false dichotomy: incremental reform, modified versions of the reform, and various other interventions are all options. The argument is valid but unsound because of a contested premise.

Argument 4.

Smith argues that capital punishment is unjust. But Smith was convicted of a felony in his twenties. So we should reject Smith's argument.

Conclusion: reject Smith's argument. Premise: Smith has a felony conviction. Kind: ostensibly inductive (credibility-based). Strength: zero. The argument is ad hominem. Smith's conviction is irrelevant to whether his argument about capital punishment is valid or has true premises. The argument should be evaluated on its own structure and content.

Argument 5.

Every previous time the economy has entered a recession during a presidential election year, the incumbent has lost. The economy is in recession. Therefore the incumbent will lose.

Conclusion: incumbent will lose. Premise: every prior election-year recession has led to incumbent loss. Kind: inductive (generalization from base rate plus current evidence). Strength: depends on the base rate (how many prior election-year recessions?) and on whether the current situation is relevantly similar to the prior ones. If the base rate is large and the current situation is similar, the argument is reasonably strong; if the base rate is small (e.g., three prior cases) the inference is weaker. Not a fallacy as stated; the argument is honest about its inductive character.

Common Misconceptions

"A valid argument is a true argument." No. Validity is about form: the inferential connection between premises and conclusion. A valid argument with false premises has a false conclusion (or might happen to have a true one by coincidence) and is unsound. Validity is a necessary but not sufficient condition for a good deductive argument; truth-of-premises is the other necessary condition.
"Naming a fallacy refutes the argument." Not on its own. Identifying a fallacy is a diagnostic; the work of refutation is showing that the diagnosis fits the particular argument under discussion. Many real arguments contain superficial features that look like a fallacy but on closer reading do not actually commit one. A bare label is not yet an argument.
"All slippery slope arguments are fallacies." No. Some slippery slopes are real: small steps do sometimes cumulate into large outcomes, and the historical record sometimes supports the projection. The fallacious slippery slope is one that asserts a causal chain without supporting the strength of the links.
"Ad hominem means any negative comment about a person." No. Ad hominem is the specific move of rejecting an argument because of features of the arguer. Negative observations about a person that do not function as a substitute for engaging with their argument are not ad hominem fallacies.

Comparisons to Related Views

Term	Definition	Quick example
Argument	Set of statements: premises and a conclusion claimed to follow	"All $F$ are $G$ . $x$ is $F$ . Therefore $x$ is $G$ ."
Premise	A statement offered as support	"All philosophers are mortal."
Conclusion	The statement the argument is trying to establish	"Socrates is mortal."
Validity	No possible situation has all premises true and conclusion false	Modus ponens: $P \to Q, P \therefore Q$
Soundness	Validity AND actually-true premises	A valid argument with true premises
Fallacy	Systematic pattern of bad reasoning	Affirming the consequent, ad hominem, equivocation
Modus ponens	$P \to Q, P \therefore Q$ (valid)	"If it rains, the street is wet. It rained. So the street is wet."
Modus tollens	$P \to Q, \neg Q \therefore \neg P$ (valid)	"If it rains, the street is wet. The street is not wet. So it did not rain."
Affirming consequent	$P \to Q, Q \therefore P$ (invalid)	"If it rains, the street is wet. The street is wet. So it rained."
Denying antecedent	$P \to Q, \neg P \therefore \neg Q$ (invalid)	"If she studied, she passed. She didn't study. So she didn't pass."

Go Further

Stanford Encyclopedia of Philosophy, "Fallacies" by Hans Hansen. The canonical survey, with historical depth.
Stanford Encyclopedia of Philosophy, "Informal Logic" by Leo Groarke.
Stanford Encyclopedia of Philosophy, "Argument" by John Hawthorne, and "Argumentation Theory" by Frans van Eemeren.
Walton, Douglas. Informal Logic: A Pragmatic Approach. 2nd ed. Cambridge University Press, 2008. The clearest single textbook on informal-logic fallacy analysis.
Tindale, Christopher W. Fallacies and Argument Appraisal. Cambridge University Press, 2007. Useful pairing.
Govier, Trudy. A Practical Study of Argument. 7th ed. Wadsworth, 2010. Widely used for practical argument analysis.

For the symbolic-logic backbone of validity, see Propositional Logic and Predicate Logic. For deductive proof technique applied to mathematics, see Basic Logic and Proof Techniques and Proof by Contradiction.