Mathematical-Reasoning · Unit 3 · Video 4 · Interactive Practice

The Negation of an If-Then Is an And: Negating Conditionals and Proving Equivalences by Chains

IKey Formulas

EquivalenceNameWhat it gives
¬(P→Q)≡P∧¬Q\lnot(P \to Q) \equiv P \land \lnot QNegation of a conditionalThe negation of an if-then is an and
P→Q≡¬P∨QP \to Q \equiv \lnot P \lor QConditional as an orThe usual first link of a chain
P→(Q∨R)≡(P∧¬Q)→RP \to (Q \lor R) \equiv (P \land \lnot Q) \to RConditional with a disjunction in the conclusionOne extra assumption, ¬Q\lnot Q
(P∨Q)→R≡(P→R)∧(Q→R)(P \lor Q) \to R \equiv (P \to R) \land (Q \to R)Conditional with a disjunction in the hypothesisRR must follow from PP and from QQ

Key Insight: An equivalence can be proven without a truth table: a chain A≡B≡C≡⋯A \equiv B \equiv C \equiv \cdots in which every step is a known equivalence. The list of reasons beside the lines is the proof.

IIWhich Statement Is the Negation?

A negation of P→QP \to Q must have the opposite truth value in every one of the four rows.

IIIIf a<ba < b, Then a2<b2a^2 < b^2?

The negation is a<ba < b and a2≥b2a^2 \ge b^2, because ¬(a2<b2)\lnot(a^2 < b^2) means a2≥b2a^2 \ge b^2, not a2>b2a^2 > b^2.

💡 Challenge: find a<ba < b with a2=b2a^2 = b^2. There the conditional is false, and only the ≥\ge form of the negation is true.

IVProof by a Chain of Equivalences

Each line of a chain has the same truth value as the line before it in every row, so the first and last lines are equivalent.

¬(P→Q)\lnot(P \to Q)

VQuiz Questions

Problem 1 · Negate a Numerical Conditional

Given: the conditional “If −2<1-2 < 1, then (−2)2<12(-2)^2 < 1^2.” Which statement is its negation?

✅ Correct! ¬(P→Q)≡P∧¬Q\lnot(P \to Q) \equiv P \land \lnot Q, and ¬((−2)2<12)\lnot\big((-2)^2 < 1^2\big) is (−2)2≥12(-2)^2 \ge 1^2. Here 4≥14 \ge 1, so the negation is true and the original conditional is false.
❌ That negates only the conclusion. P→¬QP \to \lnot Q is not the negation of P→QP \to Q: when PP is false both conditionals are true. The negation of an if-then is an and.
❌ Not quite. The hypothesis is kept, not negated: ¬(P→Q)≡P∧¬Q\lnot(P \to Q) \equiv P \land \lnot Q. A conditional fails only when its hypothesis is true and its conclusion is false.
❌ Close, but the inequality is wrong. The negation of x<yx < y is x≥yx \ge y, not x>yx > y. For x=y=9x = y = 9, 9<99 < 9 is false and 9>99 > 9 is false too.
❌ Not quite. Use ¬(P→Q)≡P∧¬Q\lnot(P \to Q) \equiv P \land \lnot Q.
Show solution

Let PP be −2<1-2 < 1 and QQ be (−2)2<12(-2)^2 < 1^2. The negation of P→QP \to Q is P∧¬QP \land \lnot Q.

¬Q\lnot Q says it is not the case that (−2)2<12(-2)^2 < 1^2, which means (−2)2≥12(-2)^2 \ge 1^2. So the negation is

−2<1  and  (−2)2≥12.-2 < 1 \ \text{ and } \ (-2)^2 \ge 1^2.

Check: −2<1-2 < 1 is true, and (−2)2=4≥1=12(-2)^2 = 4 \ge 1 = 1^2 is true, so the negation is true. That matches the original: its hypothesis is true and its conclusion 4<14 < 1 is false, so the conditional is false.

Problem 2 · Negating When the Conclusion Is an Or

Given: statements PP, QQ and RR. Which statement is logically equivalent to ¬(P→(Q∨R))\lnot\big(P \to (Q \lor R)\big)?

✅ Correct! ¬(P→(Q∨R))≡P∧¬(Q∨R)≡P∧(¬Q∧¬R)\lnot\big(P \to (Q \lor R)\big) \equiv P \land \lnot(Q \lor R) \equiv P \land (\lnot Q \land \lnot R), by the negation of a conditional and then De Morgan.
❌ That is still a conditional. Negating only the conclusion fails when PP is false: then P→(¬Q∧¬R)P \to (\lnot Q \land \lnot R) is true, while ¬(P→(Q∨R))\lnot\big(P \to (Q \lor R)\big) is false.
❌ Check De Morgan. The negation of an or is the and of the negations: ¬(Q∨R)≡¬Q∧¬R\lnot(Q \lor R) \equiv \lnot Q \land \lnot R. With PP, QQ true and RR false your option is true, but P→(Q∨R)P \to (Q \lor R) is true there, so its negation is false.
❌ Not quite. The hypothesis PP is kept, not negated. With PP, QQ, RR all false your option is true, but P→(Q∨R)P \to (Q \lor R) is true there (false hypothesis), so its negation is false.
❌ Not quite. Apply ¬(A→B)≡A∧¬B\lnot(A \to B) \equiv A \land \lnot B with B=Q∨RB = Q \lor R, then De Morgan.
Show solution

A chain of known equivalences, with Q∨RQ \lor R playing the role of the conclusion:

¬(P→(Q∨R))≡P∧¬(Q∨R)[negation of a conditional]\lnot\big(P \to (Q \lor R)\big) \equiv P \land \lnot(Q \lor R) \qquad \text{[negation of a conditional]} ≡P∧(¬Q∧¬R)[De Morgan]\equiv P \land (\lnot Q \land \lnot R) \qquad \text{[De Morgan]}

Since an and of three statements is true exactly when all three are true, the grouping does not matter: P∧¬Q∧¬RP \land \lnot Q \land \lnot R. It is true in exactly one of the eight rows, PP true, QQ false, RR false, which is the only row where P→(Q∨R)P \to (Q \lor R) is false.

Problem 3 · Complete the Chain

Given: a chain proving (P∨Q)→R≡(P→R)∧(Q→R)(P \lor Q) \to R \equiv (P \to R) \land (Q \to R):

(P∨Q)→R≡¬(P∨Q)∨R[conditional as an or]≡  ?[De Morgan]≡(¬P∨R)∧(¬Q∨R)[ ? ]≡(P→R)∧(Q→R)[conditional as an or, twice]\begin{aligned} (P \lor Q) \to R &\equiv \lnot(P \lor Q) \lor R && \text{[conditional as an or]}\\ &\equiv \ \ \textbf{?} && \text{[De Morgan]}\\ &\equiv (\lnot P \lor R) \land (\lnot Q \lor R) && \text{[ ? ]}\\ &\equiv (P \to R) \land (Q \to R) && \text{[conditional as an or, twice]} \end{aligned}

What is the missing second line?

Which law justifies the third line?

✅ Correct! De Morgan turns ¬(P∨Q)\lnot(P \lor Q) into ¬P∧¬Q\lnot P \land \lnot Q, and or distributes over and: (¬P∧¬Q)∨R≡(¬P∨R)∧(¬Q∨R)(\lnot P \land \lnot Q) \lor R \equiv (\lnot P \lor R) \land (\lnot Q \lor R).
❌ Check De Morgan. The negation of an or is an and: ¬(P∨Q)≡¬P∧¬Q\lnot(P \lor Q) \equiv \lnot P \land \lnot Q. Writing ¬P∨¬Q\lnot P \lor \lnot Q is De Morgan for an and.
❌ Not quite. Only ¬(P∨Q)\lnot(P \lor Q) changes in this step; De Morgan replaces it by ¬P∧¬Q\lnot P \land \lnot Q, and the ∨R\lor R stays.
❌ Not quite. The step turns (something)∨R(\text{something}) \lor R, with an and inside, into an and of two ors, the pattern of P∨(Q∧R)≡(P∨Q)∧(P∨R)P \lor (Q \land R) \equiv (P \lor Q) \land (P \lor R).
Show solution

Line 2. De Morgan's law for an or: ¬(P∨Q)≡¬P∧¬Q\lnot(P \lor Q) \equiv \lnot P \land \lnot Q. Replacing it inside the line gives (¬P∧¬Q)∨R(\lnot P \land \lnot Q) \lor R.

Line 3. The distributive law P∨(Q∧R)≡(P∨Q)∧(P∨R)P \lor (Q \land R) \equiv (P \lor Q) \land (P \lor R), applied with RR as the outside statement (an or can be written in either order):

(¬P∧¬Q)∨R≡R∨(¬P∧¬Q)≡(R∨¬P)∧(R∨¬Q)≡(¬P∨R)∧(¬Q∨R).(\lnot P \land \lnot Q) \lor R \equiv R \lor (\lnot P \land \lnot Q) \equiv (R \lor \lnot P) \land (R \lor \lnot Q) \equiv (\lnot P \lor R) \land (\lnot Q \lor R).

Line 4. ¬P∨R≡P→R\lnot P \lor R \equiv P \to R and ¬Q∨R≡Q→R\lnot Q \lor R \equiv Q \to R, the first equivalence read right to left.

In words: RR follows from P∨QP \lor Q exactly when RR follows from PP and RR follows from QQ.

Problem 4 · Choosing What to Prove

Given: aa and bb are integers, and we want to prove: “If a⋅ba \cdot b is even, then aa is even or bb is even.” Which statement is logically equivalent to it, so that proving it proves the original?

✅ Correct! With PP: a⋅ba \cdot b is even, QQ: aa is even, RR: bb is even, the statement is P→(Q∨R)P \to (Q \lor R), equivalent to (P∧¬Q)→R(P \land \lnot Q) \to R. For an integer, “not even” is “odd”.
❌ The extra assumption is ¬Q\lnot Q, not QQ. (P∧Q)→R(P \land Q) \to R is not equivalent: for a=2a = 2, b=1b = 1 it is false (hypothesis true, bb odd), while the original is true there since aa is even.
❌ That is the converse. (Q∨R)→P(Q \lor R) \to P happens to be a true statement about integers, but it is not logically equivalent to P→(Q∨R)P \to (Q \lor R), so proving it proves nothing about the original.
❌ The conclusion is RR, not ¬R\lnot R. This is (P∧¬Q)→¬R(P \land \lnot Q) \to \lnot R. For a=1a = 1, b=2b = 2 its hypothesis is true (22 is even and 11 is odd) and its conclusion “bb is odd” is false, yet the original is true there since bb is even.
❌ Not quite. Name the parts PP, QQ, RR and use P→(Q∨R)≡(P∧¬Q)→RP \to (Q \lor R) \equiv (P \land \lnot Q) \to R.
Show solution

Let PP be “a⋅ba \cdot b is even”, QQ be “aa is even” and RR be “bb is even”. The statement has the form P→(Q∨R)P \to (Q \lor R), and

P→(Q∨R)≡(P∧¬Q)→R.P \to (Q \lor R) \equiv (P \land \lnot Q) \to R.

¬Q\lnot Q is “aa is not even”, which for an integer means “aa is odd”. So it is enough to prove: if a⋅ba \cdot b is even and aa is odd, then bb is even. The gain is one more assumption to work with, “aa is odd”, exactly as “33 is not a factor of aa” was in the video.

The other options have the forms (P∧Q)→R(P \land Q) \to R, (Q∨R)→P(Q \lor R) \to P and (P∧¬Q)→¬R(P \land \lnot Q) \to \lnot R; each differs from P→(Q∨R)P \to (Q \lor R) in some row of the truth table, so none is equivalent.

Solved: 0 / 4