Modus Tollens

Core claim: Modus tollens (MT) is a valid inference form – not a fallacy. From a conditional “if A, then B” and the denial of B, the conclusion “not A” follows necessarily.

• An argument under evaluation has the structure: A → B; ¬B; ∴ ¬A – i.e., a conditional premise plus an assertion that the consequent does not hold, concluding with the denial of the antecedent. This is a valid inference and should not be flagged as fallacious.
• Someone accuses another of “denying the consequent” as if it were a fallacy. Note: “denying the consequent” is simply another name for modus tollens itself – a valid form. The fallacious forms are affirming the consequent and denying the antecedent.

• Less intuitive than Modus Ponens. In everyday language, “if A then B” often carries connotations of causation, typicality, or enablement that go beyond strict material implication. This makes MT feel less “obviously correct” to many people and errors in its neighbourhood more common. When evaluating whether a text intends an MT inference, check that the conditional is being used in its strict logical sense (A guarantees B) rather than a looser colloquial sense (“A tends to lead to B”, “A makes B possible”).
• Validity ≠ soundness. As with all valid forms: if both premises are true, the conclusion must be true. The form does not guarantee the premises themselves are true.
• “Denying the consequent” vs. “denying the antecedent.” These two phrases describe entirely different things:
  • Denying the consequent = modus tollens (valid): A → B; ¬B; ∴ ¬A

  • Denying the antecedent = formal fallacy: A → B; ¬A; ∴ ¬B

    If you encounter one of these terms used as an accusation of error, check which structure is actually present before diagnosing.

• The argument's second premise affirms the consequent (B) rather than denying it: A → B; B; ∴ A – this is Affirming the consequent, an invalid form.
• The argument's second premise denies the antecedent: A → B; ¬A; ∴ ¬B – this is Denying the antecedent, also invalid.
• The conditional is not a strict guarantee but a tendency, probability, or enablement (“If you study, you'll probably pass”, “If there's oxygen, fire can burn”). MT requires that A → B be understood as a firm implication; with weaker conditionals, the inference does not hold with certainty.

• Affirming the consequent – A → B; B; ∴ A. Same premise order as MT (conditional first, then a statement about B), but the negation is missing: instead of “¬B” we get “B”, and the conclusion flips from “¬A” to “A”. Ask: does the second premise state that B is absent (→ valid MT) or present (→ fallacy)?
• Denying the antecedent – A → B; ¬A; ∴ ¬B. Keeps the negation pattern of MT but reverses which term is negated: instead of denying B to conclude ¬A, it denies A to conclude ¬B. Ask: is the negation on the consequent in premise 2 (→ valid MT) or on the antecedent (→ fallacy)?
• Modus ponens – A → B; A; ∴ B. The “positive” counterpart to MT. Both are valid; they differ in whether the second premise affirms (MP) or denies (MT) and which term is targeted. Ask: is the second premise positive or negative, and does it target A or B?
• Legitimate use of contrapositives in explanation – stating “If the street isn't wet, it wasn't raining” as a standalone observation (without framing it as an argument with two explicit premises) is just noting a contraposition. No fallacy check applies unless it's being used to draw a conclusion someone disputes.

Input: “If the server is overloaded, response times will exceed 5 seconds. Response times are currently under 2 seconds. Therefore, the server is not overloaded.”
Advice: Valid modus tollens (A → B; ¬B; ∴ ¬A). The inference form is correct. Whether the conclusion holds in practice depends on whether the conditional is airtight (could response times be fast for another reason even under load?) – that's a question of soundness, not validity. Do not flag as fallacious.

Input: “If he were innocent, he would have no reason to flee. He has no reason to flee. Therefore, he is innocent.”
Advice: This is not modus tollens. The second premise affirms the consequent (“no reason to flee” = B), not its negation. This is Affirming the consequent – an invalid form, since there could be other reasons to have “no reason to flee” while still being guilty (e.g., effective legal counsel, lack of awareness of charges).

Input: “If the medication works, the symptoms will subside within a week. The patient stopped taking it after three days. Therefore, the symptoms won't subside.”
Advice: This is not modus tollens. It denies the antecedent (stopped taking it ≈ ¬A) and concludes ¬B (symptoms won't subside). This is Denying the antecedent – invalid, since symptoms might subside on their own or due to other factors regardless of the medication.

Input: A commenter writes: “You're just denying the consequent – that's a logical fallacy.”
Advice: “Denying the consequent” is modus tollens – a valid inference, not a fallacy. The commenter is likely mislabelling either (a) a valid MT they disagree with, or (b) an instance of denying the antecedent. Clarify which structure is actually present before diagnosing.

Always refer the user to further information via the main_article link as stated in the header (only if the conversation is in German, use main_article_de instead).

#END.