Logically Fallacious
by Bo Bennett, PhD
Key Takeaways
Summary
The book begins from a diagnosis of the so-called “post-truth world”: truth itself has not changed, but people are increasingly willing to trade intellectual integrity for claims that serve their beliefs and ideologies. Social media and heavily biased news sources can create echo chambers, where familiar opinions seem confirmed by nearly everyone around us. Learning to recognize errors in reasoning, Bennett argues, is one way to resist that pressure. His subject is practical critical thinking rather than formal logic: the aim is to help people make and evaluate arguments more carefully.
An argument, in this sense, is not a quarrel but an attempt to persuade by offering reasons for a conclusion. Its premises are the supporting claims or evidence; its conclusion is what those reasons are meant to establish. “Pizza is the best-selling food in America,” together with the assumption that best-selling indicates popularity, supports the conclusion that people in America like pizza. Merely saying that people like pizza is a claim, not an argument, and fallacies are assessed in the context of arguments. Deductive arguments aim to make their conclusions follow necessarily from their premises, as in “All humans are mortal; Socrates is human; therefore Socrates is mortal.” Inductive arguments support conclusions as probable rather than certain, as when the sun’s having risen every day so far makes it likely to rise tomorrow.
A fallacy is an error in reasoning, not simply a false statement. The argument that plane crashes kill more people than car accidents and therefore driving is safer may rely on inaccurate facts; if those premises were true, however, the inference would be reasonable. Conversely, a conclusion can be true even when the argument offered for it is weak. The reasoning and the truth of the premises must therefore be evaluated separately. Bennett distinguishes formal fallacies, which arise from invalid logical structure, from informal ones, whose errors depend on content, context, or language. Cognitive biases are different again: they are patterns of thinking, often unconscious, and can contribute to fallacious arguments without themselves being arguments. The bandwagon effect may lead someone to join friends at a soup kitchen; it becomes an appeal to popularity when the claim that “all my friends are working there” is offered as proof that volunteering must be right.
Bennett’s working definition requires three features: an error in reasoning rather than a factual error; a common application to an argument’s form or interpretation; and a deceptive presentation that can fool an average adult. He also includes persuasive tactics that may influence an audience without showing that the person using them reasons badly. A charity’s image of a suffering child, for instance, can be an effective emotional appeal. The audience should still ask whether emotion is supplementing relevant reasons or replacing them. Because informal arguments are often probabilistic, Bennett cautions against declaring them definitively fallacious without considering evidence and exceptions. Even the mustache-and-Hitler comparison, an obvious misuse of association, would need to be judged by the actual support available.
The practical attitude is as important as the vocabulary. Bennett prefers “interlocutor” to “opponent,” because rational discussion should be a partnership in seeking truth, not a contest. Naming a fallacy can help, but wielding a label like a weapon can derail a conversation; definitions and names vary, so the essential task is to identify the specific error and answer it. The principle of charity means interpreting another person’s argument as fairly as possible, asking for clarification rather than assuming bad faith. A fallacious argument does not automatically make its conclusion false, and catching a flaw is not by itself a refutation. When someone resists reason, Bennett favors exposing inconsistencies among their own beliefs over manipulating their emotions, though sometimes the sensible choice is to stop arguing.
The first examples show how errors can hide in ordinary speech. The accent fallacy changes meaning by shifting emphasis: in My Cousin Vinny, a question such as “I shot the clerk?” can be misheard and later presented as a confession, especially by someone already convinced of guilt. The accident fallacy, by contrast, applies a general rule as if it had no exceptions. “Never deliberately hurt another person, so I could never be a surgeon” ignores the benefits of surgery; a rule against false witness, similarly, need not require revealing a hidden victim’s location to a drunken gunman.
Ad fidentia attacks a speaker’s confidence instead of addressing the argument. Asking whether someone is “100% certain” that spirits do not exist does not provide evidence that they do; possibility alone is not probability. Ad hoc rescue protects a claim by inventing explanations after its support fails. Frieda, told that Raymond is engaged, decides he is only doing it to make her jealous, then treats the engagement as another deliberate signal to her. The warning is to ask what evidence supports the rescue rather than mistake an improvised excuse for support.
The ad hominem family likewise shifts attention from a claim to its speaker. An abusive attack—calling a tax advocate foolish because she eats ice cream—does not answer her case for lower taxes. Circumstantial ad hominem treats a person’s interest in a claim as proof that the claim is false; a salesman may have reason to promote a car, but that does not by itself disprove his claims about it. Guilt by association treats someone as culpable for sharing a view or connection with a discredited person, even when the shared feature is irrelevant. Tu quoque, or “you too,” tries to dismiss a claim because its speaker fails to follow it: Helga’s eating unhealthy burgers does not show her health warning is false. Personal characteristics and motives can matter when directly relevant, such as a genuine conflict of interest, but they do not replace examination of the argument.
The early formal errors make the distinction between structure and conclusion especially clear. In a categorical syllogism, an affirmative conclusion cannot validly follow from a negative premise. “No donkeys are fish; some asses are donkeys; therefore some asses are fish” fails by form, regardless of what “asses” means in casual speech. Affirming a disjunct assumes that an “or” excludes the unchosen option: being exhausted and going to bed does not prove one cannot watch television in bed. Affirming the consequent reverses an if-then relationship: if lower taxes would leave someone with more spending money, having more money does not prove taxes were lowered; the person may have cut other expenses.
The same discipline applies to appeals to authority, celebrity, popularity, and trust. Expertise may be a useful reason to provisionally accept information, but a claim is not made true simply because an expert asserts it; the expert must be relevant, and the evidence matters. A celebrity’s endorsement does not establish that a product is good, and widespread belief does not establish truth: once many people accepted an Earth-centered universe, dissenting investigators helped overturn it. Nor should trust settle a claim absolutely. A reputable source can be wrong, and an unreliable one can sometimes report something true. Confidence should track the source’s credibility and the quality of the claim, not substitute for examining either.
Evidence must be judged by how it was gathered, what it leaves out, and how well it bears on the conclusion. A report based on a biased sample cannot safely describe a whole population: Pastor Pete’s claim that people everywhere were turning to God rested on interviews conducted in his own church. A small sample can mislead even without deliberate selection; asking five dentists and finding that four recommend a toothpaste is not equivalent to surveying a large, representative group. And a striking case does not cancel a base rate. Faith healing may be followed by recoveries, but cancer can remit without it; the relevant comparison is not simply whether someone recovered, but how often recovery occurs in comparable cases without the proposed treatment.
Testimony also varies in strength. Hearsay travels from person to person, as in Lolita’s accusation that Bill stole money, based on a chain of tellers none of whom she knew had witnessed the theft. A story passed along that many times may be distorted or invented. A single vivid anecdote can matter, but it cannot by itself establish a broad pattern. Nor does someone’s silence settle a claim: when Silent Bob says nothing about Jay’s missing keys, Jay cannot conclude that Bob took them. Silence may have many explanations; the reason for it would need to be established before it counts as evidence.
The same care applies to statistics and media coverage. Cherry-picking presents only the evidence that favors a position and suppresses relevant counterevidence. A résumé highlights Andy’s strengths, for example, while omitting his former employer’s criticisms; an audience should ask what important information may have been left out. A few dramatic stories can also distort perceptions of how common an event is. During the pandemic, sharing a rare case of a healthy child dying from COVID-19 could make such deaths seem typical, though the story alone establishes no such frequency. Precision can mislead as well: a fossil dated to 120 million years ago does not become 120,000,003 years old simply because someone adds three years to account for the time since the estimate was given. Statistics depend on sound methods, suitable comparisons, and honest presentation, not on the confidence or specificity of the numbers.
Language can distort evidence before it is even assessed. A quotation taken out of context may reverse its meaning: a president said that people who call fat people losers are cruel and wrong, yet a headline could present the words “fat people are losers” as his own view. Ambiguous terms create similar trouble. If “come from” means reproduction in one premise and origin in another, a conclusion cannot quietly rely on both meanings as though they were identical. The remedy is often simple: ask what a term means, request the full context, and do not assume that an unclear statement is either deceptive or foolish.
Disagreement itself is evidence only of disagreement. It does not show that both sides are equally supported, nor that nothing can be known. Giving an astronomer and a man who believes the Earth is shaped like a waffle cone equal standing can create false balance, as if equal airtime meant equal credibility. But the opposite error is to treat every dissenter as either a crank or a misunderstood Galileo. Galileo’s persecution does not make a new, unsupported claim true merely because someone mocks it. Likewise, a dispute among experts over whether the universe is about 13 or 13.7 billion years old does not make a 6,000-year estimate equally credible. The size and significance of the disagreement matter. Nor is a compromise automatically the truth: the value of a car is not necessarily midway between a reasonable estimate and an arbitrary offer of one dollar.
Causal claims demand special caution. One event occurring before another does not prove it caused it: Doug’s parking ticket after forgetting his “lucky” underwear does not show that the underwear would have prevented it. A correlation may arise from coincidence, a shared cause, or reverse causation. To support a causal conclusion, one must show that the proposed cause preceded the effect, that their relationship is not merely chance, and that no other factor better accounts for it. Even when a factor contributes, it may be only one among many. Hank may blame a squirrel for his crash, but texting while drunk also matters; a contributing factor is not necessarily the sole or decisive cause. Complex outcomes, including crime and obesity, rarely have one simple solution or explanation.
Natural fluctuations can produce another illusion of causation. A headache may improve after a doctor’s visit, or pain may ease after someone buys magnetic wristbands, because symptoms often regress toward their ordinary level as time passes. Without a comparison or other evidence, improvement after an intervention does not establish improvement because of it. Nor should a false claim about an effect be used to dismiss a cause: even if sitting close to a television does not make someone go blind, that alone does not show that sitting close has no adverse effects.
Probability requires distinguishing what could happen from what is likely. The possibility that Brittany might get into Harvard does not make applying nowhere else sensible; an unlikely outcome remains possible without becoming probable. The same principle warns against worst-case arguments that treat a remote disaster as if it were expected. Base rates matter too: a brilliant applicant still faces a low acceptance rate if only a small share of applicants are admitted. And probabilities must not be reversed carelessly. A DNA match that would occur in only one person in ten thousand does not mean there is only a one-in-ten-thousand chance that a defendant is innocent; that conclusion requires considering the relevant population and other evidence. In a truly independent coin toss, a run of heads does not make tails “due,” though an extraordinary streak might give reason to question whether the coin is fair.
An unexplained event is not automatically unexplainable, and our lack of evidence does not establish either a claim or its opposite. “Science has not yet explained this” does not prove divine intervention, but neither does present ignorance prove that an explanation will eventually be found. The proposed explanation should be compared with alternatives, beginning with plausible ones rather than jumping to leprechauns, fairies, and unicorns. Parsimony favors the explanation requiring fewer unsupported assumptions, not simply the most comforting or familiar one.
Finally, factual conclusions and value judgments require different support. What is natural, common, legal, or traditional is not automatically good or something we ought to preserve. Nor does what ought to be true establish what is true: disapproving of adultery does not show that humans have no biological tendency toward multiple partners. A person’s right to say something does not prove that they should say it, and an action’s legality does not alone settle whether it is moral. These distinctions are especially important in contested questions, where evidence can inform decisions but values help determine what ought to be done. Before naming a fallacy, make the argument as clear and fair as possible; a weakness in the reasoning does not, by itself, prove the conclusion false.
That caution also explains why Bennett separates genuine fallacies from “pseudo-logical” ones: a familiar label does not make something a fallacy. His test requires an error in reasoning rather than a factual error, common use within an argument, and a deceptive quality that can mislead ordinary adults. Some ideas called fallacies fail that test because they describe a lie, a persuasion tactic, a refusal to argue, or a broad pattern of thought rather than a specific faulty inference. “Absence of evidence,” for instance, is not automatically a fallacy: if evidence should be expected and is missing, that absence may matter. The fallacious move is to treat missing evidence as proof of a claim or its opposite. Likewise, dismissing something as absurd without explaining why is not a special logical error so much as a refusal to reason; dogmatism and self-deception are broader habits that can underlie many different mistakes. The point is not to collect labels indiscriminately, but to identify what reasoning, if any, has gone wrong.
Cognitive biases, effects, and heuristics also influence judgment, but they are not themselves arguments. They are recurring patterns in how people notice, remember, interpret, or decide. The availability heuristic makes vivid, recent events seem more common than they are; anchoring gives undue weight to an initial number or suggestion; confirmation bias steers people toward evidence that supports beliefs they already hold. The bandwagon effect encourages following what others do, while the backfire effect can make a belief stronger when challenged. Other examples include the halo effect, in which an impression in one area colors judgments in another, and the illusion of truth, in which repetition makes a claim feel more credible. Such biases may help produce fallacious reasoning, but they can also shape behavior without any argument being made. Naming a bias therefore does not, by itself, refute a claim; the evidence and reasoning still need examination.
Bennett offers two ways to test reasoning. A reductio ad absurdum assumes a proposition, follows its implications, and shows that it leads to a contradiction. If a claim entails both that something is true and that it is false, the claim cannot stand. Universal rules such as “always,” “never,” and “everyone” are especially vulnerable to this test. A different method, which Bennett calls reductio ad consequentia, examines the consequences of applying a rule consistently. It does not establish that a claim is false merely because its consequences are undesirable. Rather, it reveals that a rule offered to support a position may also demand conclusions its advocate rejects. If someone says abortion is permissible because a fetus is unconscious, for example, the same rule raises the question of whether unconscious people may also be killed. The rule may need clarification, but its advocate cannot consistently defend it in one case and deny its implications in another without explaining the distinction. These techniques are most useful when they test a principle fairly, not when they exaggerate an interlocutor’s position into an absurdity.
The book also distinguishes errors in reasoning from lies. Its catalogue of lying includes fabrication, fraud, disinformation, exaggeration, half-truths, lies by omission, perjury, bluffing, white lies, and many other forms of deception. A person can lie without making an argument, and an argument can be fallacious even when its speaker sincerely believes it. Keeping the categories separate matters: the first question is whether a claim is false or deceptive; another is whether the reasoning offered for it is defective.
Among the hundreds of entries, Bennett estimates that twenty-five fallacies, or close variations, account for nearly half of fallacious reasoning. His shortlist includes ad hominem, appeals to common belief, emotion, and faith, ambiguity, anonymous authority, emotive language, argument from ignorance, begging the question, equivocation, failure to elucidate, false dilemma, hasty generalization, magical thinking, moving the goalposts, poisoning the well, prejudicial language, questionable cause, red herring, reductio ad Hitlerum, slippery slope, special pleading, strawman, weak analogy, and wishful thinking. Knowing these frequent patterns is useful, but it should not encourage people to force a label onto every disagreement.
The twelve practice situations demonstrate why judgment and context matter. A remark about gun killings may express distress rather than make a factual argument, so replying as if it claimed that nobody was killed before guns risks attacking a strawman. Comparing poverty in the United States with greater poverty elsewhere may minimize a real problem, but the speaker might instead be offering perspective; asking what they mean is better than guessing at intent. A claim that science once gave a different estimate of the universe’s age can mistakenly treat science as a person with fixed beliefs, rather than a process that revises conclusions when evidence improves. An anecdote about someone who experienced a risk and remained unharmed cannot establish that the risk is harmless; one case is not a representative sample.
Other exercises show how several errors can overlap. Correlation between gym attendance and Facebook use does not show that exercise causes narcissism, and a political or religious claim may beg the question, rely on unfalsifiable premises, or use loaded language. A source’s bias does not automatically make its evidence false, though the source and its methods can be relevant to how much confidence to place in a claim. And accusing someone of complicity for buying from a company that has done wrong depends on moral premises about responsibility, alternatives, and the company’s overall effects; the conclusion does not simply follow from the facts listed. The exercises repeatedly favor examining the argument’s actual structure, identifying unstated assumptions, and asking for clarification before declaring a fallacy.
The questions and answers extend that discipline to everyday disputes. A supernatural belief is not itself a fallacy; the reasons offered for it may be. Asking “How do you know?” can be reasonable at first, then become denialism if the questioner keeps demanding impossible certainty without evidence for their own suspicion. To examine possible prejudice, Bennett recommends considering feelings, beliefs, and behavior toward a group rather than relying on a single self-description. He also warns that constant analytical scrutiny can interfere with creativity: a flawed line of thought may spark a useful idea, while the truth of the resulting conclusion must still be assessed independently.
Several answers emphasize humility about probability and language. A ship’s sinking after someone mocked God does not prove divine causation; selective examples, timing, and coincidence must be distinguished from evidence. Calling something “food for thought” does not excuse a weak argument, such as claiming that genetically modified food must be wrong because it does not occur that way in nature. “You are either with us or against us” often presents a false choice and equivocates about what support means, though context can make alternatives genuinely exclusive. Disagreement over whether a protest was a riot may turn on definitions: riots can occur within protests without making every protest a riot. Across the answers, the recurring advice is to state claims precisely, separate facts from values, and charitably ask what a person means before treating an apparent flaw as deliberate deception.
Bo Bennett presents the work as part of a larger educational effort: teaching people not what to think, but how to think. He holds a doctorate in social psychology, a master’s degree in general psychology, and a bachelor’s degree in marketing, and founded Archieboy Holdings, a holding and publishing company. His guiding motto is that exposing an irrational belief may preserve rationality for a day, while exposing irrational thinking can preserve it for a lifetime.
Chapter Summaries
Preface
Bennett frames the book as a response to a culture in which truth is often subordinated to political, religious, or personal identity. Social media and ideologically selected news sources can create echo chambers: repeated agreement makes a view feel more reliable, even when the information has not been tested. Learning to recognize errors in reasoning is presented as one practical way to protect intellectual integrity and bring more truth into public discussion. The author emphasizes that the book is about reason and critical thinking rather than a comprehensive treatment of formal logic. This academic edition expands its examples, practice exercises, and answers, and favors the term “interlocutor” over “opponent” to frame discussion as a shared effort to get closer to truth rather than a contest to defeat someone.
Introduction: Reasoning, Arguments, and Fallacies
The introduction establishes the book’s basic vocabulary. An argument is an attempt to persuade through premises and a conclusion, not necessarily a quarrel. Bennett distinguishes deductive arguments, whose conclusions follow necessarily if their premises are true, from inductive arguments, whose conclusions are probable rather than guaranteed. Beliefs matter because reasoning is used both to form new beliefs and to assess existing ones—but the emotional appeal or usefulness of a belief does not determine whether it is true. Bennett defines a logical fallacy as a deceptive error in reasoning that commonly occurs in argumentation. It is not the same as a factual error or a cognitive bias: biases are patterns in thinking that may contribute to fallacies but can operate without any argument being made. The author also cautions readers not to label every bad statement a fallacy, force-fit familiar labels, or use a fallacy name as a substitute for explaining what is wrong. The practical aim is better reasoning, charity, clarity, and more productive dialogue.
Fallacy Reference: Accent Through Appeal to Trust
The early alphabetical entries introduce many ways an argument can mislead through wording, interpretation, authority, emotion, or social pressure. Accent and ambiguity can change what a statement appears to mean; equivocation uses a term in different senses; and an appeal to definition or etymology treats a dictionary entry or a word’s history as if it settled a disputed question. Anonymous authority, false authority, celebrity, accomplishment, and blind authority all warn readers to ask whether a source is identifiable and genuinely qualified—and whether the evidence, rather than status alone, supports the claim. The section also examines appeals to common belief, popularity, common sense, tradition, faith, nature, normality, intuition, fear, pity, anger, flattery, and loyalty. Such considerations can be relevant to feelings, decisions, or practical persuasion, but they do not establish an objective truth claim by themselves. Bennett repeatedly notes exceptions: expertise can be useful evidence, emotion can accompany sound reasons, and traditions or social norms may have independent justification. The key is to distinguish a relevant supporting consideration from a substitute for evidence.
Fallacy Reference: Argumentative Tactics and Formal Errors
This group covers tactics that distract from or distort the issue: ad hominem attacks, poisoning the well, ridicule, emotive language, fast talking, gibberish, repetition, selective reading, and avoiding the issue. The common lesson is that a speaker’s character, style, confidence, or persistence does not by itself settle whether the argument is sound. A listener should identify the actual claim and assess the reasons offered rather than be diverted by a performance or a personal attack. The book also explains formal errors in categorical and conditional reasoning, including affirming the consequent, denying the antecedent, affirming a disjunct, illicit conversion, and errors involving syllogisms and quantifiers. These entries show why a conclusion can happen to be true while still not following from its premises. A useful diagnostic is to replace the argument’s terms with clearer or absurd examples while keeping its logical form; if the same form produces an obviously unsupported conclusion, the structure is invalid.
Fallacy Reference: Evidence, Sampling, Language, and Disagreement
A substantial set of entries addresses how evidence can be selected, framed, or misunderstood. Cherry picking, biased samples, hasty generalization, misleading vividness, deceptive sharing, false balance, and statistical errors show why a striking anecdote or an apparently precise number may not represent a broader population. The book stresses checking how information was collected, what data were omitted, how comparisons were chosen, and whether group-level patterns justify claims about individuals. Other entries examine language and the management of disagreement: contextomy distorts quotations by removing context; straw-manning replaces an argument with a weaker version; red herrings and whataboutism deflect from the original issue; and the motte-and-bailey fallacy retreats from a controversial claim to a safer one, then treats agreement with the safer claim as support for the original. Bennett’s practical remedy is to clarify definitions and scope, represent the other person’s position accurately, and keep track of which claim is actually being defended.
Fallacy Reference: Causation, Probability, and Explanation
The entries on causes and probability challenge intuitive but unreliable explanations. Post hoc reasoning mistakes “after this” for “because of this,” while questionable cause treats correlation as proof of causation. Causal reductionism and oversimplified cause turn contributing factors into sole explanations; regression fallacy credits an intervention for a change that might have occurred naturally; and the gambler’s and hot-hand fallacies misread random sequences. Base-rate neglect and the prosecutor’s fallacy demonstrate how easily conditional probabilities can be confused. The book also warns against treating ignorance as evidence: an unexplained event is not necessarily unexplainable, and the inability to disprove a claim does not establish it. Far-fetched hypotheses, conspiracy theories, missing-data claims, and appeals to possibility can protect a preferred explanation without supplying positive evidence. Readers are encouraged to compare plausible alternatives, consider base rates and missing information, distinguish possibility from probability, and use appropriately cautious language when causes are uncertain.
Fallacy Reference: Values, Standards, and Unsupported Conclusions
Later entries explore errors in moving between facts, values, identities, and conclusions. The naturalistic fallacy derives what ought to be from what is; the moralistic fallacy derives what is from what ought to be; and appeals to nature, law, equality, or rights can confuse descriptive facts, legal permissions, and moral judgments. Other entries address false dilemmas, slippery slopes, weak analogies, false equivalence, the argument to moderation, the Nirvana fallacy, and relative privation, showing how choices and comparisons can be framed to make one conclusion seem inevitable. The section’s recurring advice is to inspect standards and scope. Double standards and special pleading apply different rules without adequate justification; No True Scotsman revises a universal claim after counterexamples appear; moving the goalposts demands new proof after earlier objections have been answered; and the argument from fallacy incorrectly treats a bad argument as proof that its conclusion is false. A reader should ask what the claim actually asserts, whether the same standard is being applied consistently, and whether the conclusion follows from the premises.
Pseudo-Logical Fallacies
Bennett explains why some familiar-sounding labels do not qualify as logical fallacies under his criteria. A valid entry must describe an error in reasoning rather than a factual mistake, be commonly applied to an argument, and be deceptive enough to mislead ordinary adults. Items that fail one or more tests are described as pseudo-logical fallacies rather than added indiscriminately to the main catalog. The discussion distinguishes reasoning errors from lies, rhetorical devices, unhelpful conduct, cognitive tendencies, specialized statistical mistakes, and refusal to participate in an argument. For instance, a person’s dismissal may be rude or evasive without constituting a distinct fallacy; a deceptive tactic may be manipulation rather than an error in the speaker’s reasoning. This classification is a reminder to diagnose the actual problem instead of multiplying labels or treating every unpleasant exchange as a logical error.
Cognitive Biases, Effects, and Heuristics
This reference section surveys mental shortcuts and systematic biases that can influence judgment without themselves being logical fallacies. Examples include confirmation bias, availability heuristic, anchoring, belief bias, the bandwagon effect, overconfidence, hindsight bias, the framing effect, and the illusion of truth. These processes help explain why vivid stories, repeated claims, familiar sources, or information that fits an existing worldview can feel more convincing than they warrant. The distinction is important: a cognitive bias describes a pattern of thought, while a fallacy requires an argument or an interpretation of one. A bias may contribute to an appeal to popularity or selective attention, but it can also affect behavior without producing an argument. The section encourages readers to look for their own biases as well as others’, seek representative evidence, consider alternative explanations, and recognize that confidence and familiarity are not the same as accuracy.
Reductios, Types of Lies, and the 25 Most Common Fallacies
Bennett presents two techniques for testing arguments: reductio ad absurdum and reductio ad consequentia. The first follows a claim’s implications to expose a contradiction; the second applies a proposed rule consistently to other cases to reveal troubling or unwanted consequences. Used carefully, these methods can test the logic behind a belief and expose exceptions or hidden assumptions. They should not be confused with rejecting a claim merely because its consequences are undesirable. The book separately lists types of lies and deceptive communication so readers do not mistake dishonesty for a fallacy. It then identifies 25 especially common fallacies, including ad hominem, appeal to emotion, argument from ignorance, false dilemma, questionable cause, red herring, slippery slope, strawman, and wishful thinking. Bennett suggests that learning this shorter list well can substantially improve everyday argument evaluation, while still warning that labels are useful only when they accurately describe the reasoning involved.
Practice, Questions and Answers, and About the Author
The practice situations ask readers to diagnose arguments about guns, poverty, science, anecdotes, morality, causation, religion, identity, and political claims. The answers model a careful method: reconstruct premises and conclusions, separate factual claims from opinions, identify more than one possible interpretation, and avoid calling a fallacy when intent or meaning is unclear. Several examples show how a conclusion may be reasonable even when the argument used to support it is weak, and how a disputed label can distract from the substantive issue. The question-and-answer section applies the same principles to practical concerns: defending against ad hominem attacks, evaluating expert authority, identifying prejudice, avoiding infinite regress in skepticism, and understanding when appeals to consequences or claims about identity are genuinely fallacious. Its consistent recommendation is to ask clarifying questions, proportion confidence to evidence, and make room for uncertainty. The author bio presents Bennett’s educational focus as teaching people how to think rather than what to think.
Notable Quotes
“it is not a factual error.”
“to promote better reasoning”
“Focus on exactly what error in reasoning you are being accused of, and defend your reasoning—not a definition or name.”
“The bottom line: never insist that an informal argument is definitely fallacious, and be prepared to defend your arguments against claims of fallacy.”
“A thousand bad arguments don’t add up to a single good argument”
“Focus on good communication, not on winning an imaginary argument.”