Skip to content
Check My Logic
Check My Logic

Valid form

Denying the consequent

Also known as modus tollens or modus tollendo tollens

Denying the consequent, better known by its Latin name modus tollens (modus tollendo tollens, “the way that denies by denying”), is the valid way to reason from an if-then statement when the result didn’t happen. If “if P, then Q” is true, and Q is false, then P must be false.

It’s a logical rule, valid by its form alone. You don’t need to know what P and Q stand for: any argument with this shape and true premises has a true conclusion.

Example

If the bakery is open, its lights are on. Its lights are off. So the bakery isn’t open.

In “if P, then Q”, P is called the antecedent (the “if” part) and Q the consequent (the “then” part). Modus tollens denies the consequent (Q is false) and concludes that the antecedent is false too.

Deny the consequent, not the antecedent

Here’s the same conditional with the other half denied:

If the bakery is open, its lights are on. The bakery isn’t open. So its lights are off.

This one is invalid. The bakers could be inside before opening time with every light on, and then both premises are true while the conclusion is false. That’s the fallacy of Denying the antecedent.

The two arguments both deny something, so they’re easy to mix up. What separates them is which half gets denied:

  • “If P, then Q” rules out exactly one situation: P true and Q false.
  • If you know Q is false, the only way P could be true is that ruled-out situation. So P is false. That’s modus tollens.
  • If you know P is false, the conditional has nothing to say. Q could go either way. That’s why denying the antecedent fails.

Another way to see it: “if P, then Q” means the same as its contrapositive, “if not Q, then not P”. Modus tollens is just modus ponens applied to the contrapositive. “If not P, then not Q” is a different statement that doesn’t follow from the original.

Harder to see than modus ponens

Although the two forms are equally valid, people accept modus tollens less readily than modus ponens. A meta-analysis of studies using abstract materials found that participants accepted modus ponens almost every time, but modus tollens noticeably less often. Researchers still debate why. Either way, a modus tollens that feels shaky may be perfectly valid, and one that feels obvious may really be denying the antecedent, so check which half was denied.

Valid isn’t the same as true

A valid argument only guarantees its conclusion if its premises are true.

If a restaurant’s food is excellent, it has a line out the door. This restaurant has no line. So its food isn’t excellent.

That’s a correct modus tollens, but the first premise is false: a restaurant can serve excellent food and still have no line because it’s new, hard to find or takes reservations. The form is fine; the conditional has exceptions. Modus tollens is only as reliable as the “if P, then Q” it starts from, so when an argument has this shape, check whether Q really always follows P. For more on the difference, see Valid vs. true.

Look-alikes

Four forms that look similar. Two are valid, two are not.

Form Valid?
If P, then Q. P. So Q. Yes: Affirming the antecedent (modus ponens)
If P, then Q. Not Q. So not P. Yes: denying the consequent (modus tollens)
If P, then Q. Q. So P. No: Affirming the consequent
If P, then Q. Not P. So not Q. No: Denying the antecedent

The valid forms affirm the antecedent or deny the consequent. The invalid ones do the reverse.

Sources

  1. Irving M. Copi, Carl Cohen and Kenneth McMahon (2011). Introduction to Logic (14th edition). Pearson.
  2. P. D. Magnus, Tim Button, Robert Trueman and Richard Zach, with J. Robert Loftis and Aaron Thomas-Bolduc (2025). forall x: Calgary. An Introduction to Formal Logic (Fall 2025 edition), section 19.2. Open Logic Project (open textbook, CC BY 4.0).
  3. Walter J. Schroyens, Walter Schaeken and Géry d'Ydewalle (2001). The processing of negations in conditional reasoning: A meta-analytic case study in mental model and/or mental logic theory. Thinking & Reasoning 7(2), 121–172.

Last reviewed 2026-09-13.