You have just finished a proof that "if n is even, then n^2 is even." The next problem says: "show that 4 is even." You notice 4^2 = 16 is even, and you are tempted to write:
"n^2 is even (since 4^2 = 16), so by my theorem, n is even. Therefore 4 is even."
Stop. That argument is wrong in a way that is subtle enough to fool many students and prominent enough to have its own name: the fallacy of affirming the consequent, also called the fallacy of the converse. If you just proved P \Rightarrow Q and you observe Q is true, you cannot conclude P.
The example above happens to give a correct conclusion (4 is indeed even) by accident — because the statement "if n^2 is even, then n is even" is also true (its contrapositive form of a valid proof). But the reasoning is broken, and as soon as you apply the same pattern to a theorem whose converse is false, you produce actual errors.
The logical rule you used vs. the rule you needed
The valid move after proving P \Rightarrow Q is:
- Observe P is true for some specific object.
- Conclude Q is true for that object.
That is modus ponens, the workhorse rule of inference. It says: if P \Rightarrow Q and P, then Q.
What students do by mistake is the "reverse" move:
- Observe Q is true.
- Conclude P is true.
This is the fallacy. The truth table shows why immediately: P \Rightarrow Q is true in three cases (T,T; F,T; F,F), and in two of those three, Q is true. One of those two has P also true; the other has P false. Observing Q does not distinguish between them.
| P | Q | P \Rightarrow Q | Q is true here? | Can we conclude P? |
|---|---|---|---|---|
| T | T | T | yes | yes (accidentally) |
| F | T | T | yes | no — P is false! |
| T | F | F | — | — |
| F | F | T | no | — |
The second row is the counterexample to the fallacy. The implication is true, Q is true, but P is false. So "observing Q" cannot force P.
Why modus ponens is valid and affirming-the-consequent is not: modus ponens eliminates one of the two rows where Q is true (the one where P is false) by additionally requiring P. Affirming the consequent has no such elimination — it keeps both rows on the table, so it cannot pin down P.
A statement where the fallacy produces a real error
Consider the true theorem: "if it rained last night, then the road is wet." (P \Rightarrow Q where P is "it rained" and Q is "road is wet.")
Now suppose you wake up and observe the road is wet. Can you conclude it rained?
No. Someone could have washed the road. A pipe could have burst. A street cleaner could have driven by. Q is true, but P might be false. The theorem was valid. Your conclusion from the wet road is not.
In everyday life, affirming the consequent shows up constantly: "successful people work hard; I work hard; therefore I will be successful." The implication may be true (hard work helps), but "I work hard" is Q-like — it does not force P ("I will succeed"). Success also depends on luck, opportunity, and a dozen other factors.
A mathematical example where it matters
Consider the true theorem: "if n is divisible by 6, then n is divisible by 2." (P: 6 \mid n. Q: 2 \mid n.)
Now suppose you are told n = 8 (so Q is true — 8 is divisible by 2). Can you conclude P — that 8 is divisible by 6?
No. 8 is not divisible by 6. Yet Q is true. The theorem is true. But observing Q did not let you conclude P, because the converse "if n is divisible by 2, then n is divisible by 6" is false.
This is the cleanest illustration of the fallacy in action. Whenever the converse of your theorem is false, affirming-the-consequent produces wrong answers immediately.
The valid move when you want P from Q: prove the converse
If you actually want to use "Q is true therefore P is true," you need the converse, Q \Rightarrow P, as a separate theorem. The converse is a different statement from P \Rightarrow Q and must be proved separately.
When both P \Rightarrow Q and Q \Rightarrow P hold, the combined statement is the biconditional P \Leftrightarrow Q, written "if and only if." At that point, you can freely move between P and Q — but only then. See Converse vs Contrapositive — Why They Are Not the Same Thing for the full taxonomy.
An interactive — two theorems, two observations
The reflex that prevents the error
Before using a theorem P \Rightarrow Q, ask: "which side do I have?"
- If you have P, go forward. Conclude Q. Safe. This is modus ponens.
- If you have Q and want P, stop. You need a separate proof of the converse Q \Rightarrow P. Without that proof, concluding P is the fallacy.
- If you have \lnot Q and want \lnot P, go backward via the contrapositive. Safe. This is modus tollens.
- If you have \lnot P and want \lnot Q, stop. That is the fallacy of denying the antecedent, a cousin of the converse fallacy and equally invalid. P being false does not force Q to be false.
Two safe moves (modus ponens, modus tollens) and two fallacies (affirming the consequent, denying the antecedent). The pattern: safe moves go from hypothesis-side to conclusion-side or from not-conclusion to not-hypothesis. Fallacies go the other way.
The exam reflex
- Write the theorem in full: "if P, then Q."
- Identify what you have: is it P, Q, \lnot P, or \lnot Q?
- If you have P: conclude Q. If you have \lnot Q: conclude \lnot P (contrapositive).
- If you have Q or \lnot P: you cannot use this theorem to conclude what you want. You either need the converse (separate theorem) or a different approach entirely.
Mis-directing on this question — going backward through an implication you only proved forward — is how correct theorems get turned into wrong applications. Recognising the fallacy is the single most important sanity check before you apply any "if-then" theorem.
Related: Mathematical Proof — Direct Proof · Converse vs Contrapositive — Why They Are Not the Same Thing · Logic and Propositions · What Exactly Am I Allowed to Assume When I 'Assume P'?