The misconception: "I learned in algebra that multiplication distributes over addition: a \cdot (b + c) = a \cdot b + a \cdot c. Logical negation should behave the same way — just push the \lnot onto each term. So \lnot(p \land q) = \lnot p \land \lnot q."
That distribution is invalid. Negation and the AND/OR connectives do not obey a simple distributive rule. Pushing a \lnot across a \land only works if you also flip the connective from \land to \lor. The correct law is called De Morgan's law, and every missed flip silently rewrites your statement into a strictly stronger or weaker claim.
The correct rule, in bold
Two things happen when \lnot crosses a binary connective: the connective flips (\land \leftrightarrow \lor), and the \lnot moves onto each piece. Forget either step and you are off the correct formula.
Why the flip is necessary: p \land q is true only when both sides are true, so its negation is the broad claim "at least one side is false". "At least one" is \lor, not \land. The connective has to flip to capture "at least one failing" instead of "both failing."
The truth-table proof that distribution fails
Here is the side-by-side truth table for \lnot(p \land q), the wrong answer \lnot p \land \lnot q, and the right answer \lnot p \lor \lnot q:
| p | q | p \land q | \lnot(p \land q) | \lnot p | \lnot q | \lnot p \land \lnot q (wrong) | \lnot p \lor \lnot q (right) |
|---|---|---|---|---|---|---|---|
| T | T | T | F | F | F | F | F |
| T | F | F | T | F | T | F | T |
| F | T | F | T | T | F | F | T |
| F | F | F | T | T | T | T | T |
Compare the fourth column (\lnot(p \land q)) with the last two columns:
- The "wrong answer" \lnot p \land \lnot q matches only in rows 1 and 4. It misses in rows 2 and 3.
- The "right answer" \lnot p \lor \lnot q matches in all four rows. Identical truth values.
The wrong answer is strictly stronger than the correct negation — it demands that both p and q be false, whereas the correct negation only demands that at least one be false. Using the wrong one in a proof would be like proving "both patients are sick" when all you had evidence for was "at least one patient is sick."
A concrete example where the mistake costs you
Statement: "Raj has a ticket AND the train is running." Call it p \land q.
What does it mean for this statement to be false? The train service might be suspended. Or Raj forgot his ticket. Or both. Any one of those scenarios makes the joint claim false.
- Wrong negation \lnot p \land \lnot q: "Raj does not have a ticket AND the train is not running." This is the claim that both things fail simultaneously. That is far more specific than a plain denial.
- Correct negation \lnot(p \land q) = \lnot p \lor \lnot q: "Either Raj does not have a ticket OR the train is not running." This captures every scenario where the original fails.
The original claim fails the moment any one of the two parts fails. A correct negation must cover every failure mode, not just the one where both fail at once.
The intuition that rescues the flip
"p AND q" is true only when you hold both cards. "NOT (p AND q)" means "you are missing at least one of the cards." Missing at least one is a \lor situation — this card is missing, OR that card is missing, OR both. It is never a \land situation, because \land demands both failures at once.
By the same token: "p OR q" is true when you hold at least one card. "NOT (p OR q)" means "you are missing both cards." Both failures at once — that is a \land.
So the mental rule is: negation converts "at least one" to "all" and vice versa. "At least one success" negates to "all failures." "All successes" negates to "at least one failure."
Where the false analogy to distribution comes from
In algebra, multiplication distributes over addition: a(b + c) = ab + ac. The multiplier a gets copied to each term, and the + stays the same. You might expect logical negation to behave the same way — copy the \lnot onto each term, keep the \land the same.
But \lnot is a truth-table-flipping operator, not a multiplicative one. Watch carefully: when you flip truth values in p \land q, the role of the connective itself flips. A T outcome in p \land q becomes an F outcome in \lnot(p \land q), and T in \lor corresponds to "at least one is true" — which is the dual of \land's "both are true." The connective duality forces the connective flip.
If you want an algebra analogy that actually fits, think of \lnot like the operation "take the complement of the set." Taking the complement of A \cap B gives A' \cup B' — the intersection becomes a union, just as AND becomes OR. The set-theoretic De Morgan visualisation shows the same pattern on Venn circles.
Checklist for safely negating
When you negate any compound statement:
- Push \lnot inward one step.
- If it crosses \land, change it to \lor. If it crosses \lor, change it to \land. The connective must flip.
- Apply \lnot to each piece.
- Repeat on any remaining nested compounds.
Skip step 2 and your negation is wrong. The cleanest sanity check: build the truth table for your proposed negation and compare it to \lnot applied to the original. They must match row by row.
The set-theory twin
De Morgan's laws in set theory — (A \cap B)' = A' \cup B' and (A \cup B)' = A' \cap B' — are the exact same rule wearing different notation. If you've accepted that the complement of a union is an intersection (and vice versa), you have already accepted the logical version; they are just the same algebra applied to different objects. See De Morgan Intuition — Why Outside the Union Equals Intersection of Outsides for the pictorial version.
Related: Logic and Propositions · De Morgan for Logic — Swap Animation · De Morgan Intuition on Venn Circles · Recognise 'Neither A Nor B' Triggers De Morgan