Proof Techniques
Context: FIT1058_MOC, FIT2014_MOC · the named patterns a proof follows · matched to the claim’s quantifier · Mathematical Induction is the fifth Judging quality: Proof Critique (Good, Bad and Ugly Proofs) — valid vs invalid vs merely graceless.
Quick Revision
- 🎯 Objective: pick the argument pattern matching the claim’s logical shape ➔ construction / cases / contradiction / symbolic manipulation (+ induction).
- 📦 Core Components: construction | cases | hard-direct contradiction | identities symbolic manipulation.
- ⚡ Key Constraint: a single example proves but never ; exhaustion needs a finite (or finitely-partitioned) domain.
📝 Core
1. Symbolic Manipulation
- Method ➔ a chain of law-justified equalities transforming one expression into another.
- Freedom ➔ start from either side (often the more complex, simplifying toward the other).
- Scope ➔ algebra, set laws (De Morgan, distributive, idempotent).
2. Construction (Witness)
- Proves ➔ an existential by exhibiting one concrete object.
- Common mistake ➔ a single example never proves a universal — it only illustrates.
3. Cases (Exhaustion)
- Method ➔ split the domain into finitely many cases that collectively cover everything; prove each.
- Rules ➔ cases may overlap (redundant but valid); they must be exhaustive; infinite domains partition (e.g. vs ).
4. Contradiction (Reductio)
- Method ➔ assume the negation, derive an absurdity (), conclude the assumption false.
- Canonical ➔ Euclid’s infinitude of primes; interesting-number (Well-Ordering); is irrational (assume in lowest terms both even contradiction); Cantor’s diagonalisation (Countability and Cantor Diagonalisation).
- ⚠ Don’t over-apply ➔ if a direct proof exists, wrapping it in contradiction makes it ugly, not stronger (Proof Critique (Good, Bad and Ugly Proofs)).
5. Proving Equalities (the two-sided strategies)
- Set equality ➔ prove and separately.
- Subset relation ➔ take a general member of , name it, use ‘s definition, follow the consequences, aim at ‘s definition. (the blueprint used for in Formal Languages (Alphabets, Words, Languages))
- Numerical equality ➔ transform into algebraically if you can; if not, prove and .
⚖️ Core Decision Matrix
| Claim shape | Technique | Key first move |
|---|---|---|
| construction | exhibit one witness | |
| (finite domain) | cases / exhaustion | partition, prove each |
| equality / identity | symbolic manipulation | transform one side |
| hard to attack directly | contradiction | assume the negation |
| Mathematical Induction | base + inductive step |
When It Flips: method follows the goal — the quantifier picks the pattern. Exhaustion needs a finite (or finitely-partitioned) domain; contradiction needs a true axiom framework, so a derived absurdity convicts only the assumed negation. Real proofs often combine several.
📊 Exam Execution Trace
Manual Execution Trace
Matching claims to techniques:
| Step / State | Claim | Logical shape | Technique |
|---|---|---|---|
| 0 (Init) | — | — | — |
| 1 | ” even prime” | existential | construction (witness 2) |
| 2 | ” even “ | universal, finite cases | cases (, ) |
| 3 | ” irrational” | hard-direct | contradiction |
⚠️ Common Mistakes
- 💡 Construction ≠ universal proof ➔ exhibiting one witness settles only; a claim needs cases/arbitrary-element, and contradiction relies on a true framework so the lone false assumption is convicted.
🧠 Active Recall
Match each proof method to the claim it suits, and give construction's common mistake.
- Hint: Quantifier ↔ method.
Answer
- Short answer: Symbolic → identities; construction → ; cases → finite ; contradiction → hard-direct.
- Why: One example ≠ ➔ a witness proves only existence; a universal needs every case or an arbitrary element.
Outline Euclid's proof of infinitely many primes and why it is by contradiction.
- Hint: Construct , force a missing prime.
Answer
- Short answer: Assume finitely many ; has remainder 1 under each , so a prime outside the list exists.
- Why: Contradiction ➔ this refutes the completeness assumption (Fundamental Theorem of Arithmetic guarantees has a prime factor).
In the "every natural is interesting" proof, which principle is used and why does it fail over ?
- Hint: Well-Ordering.
Answer
- Short answer: Well-Ordering: every non-empty has a smallest element; the smallest “uninteresting” number is thereby interesting — contradiction.
- Why: not well-ordered ➔ it extends to , so a non-empty integer set need not have a least element; the construction breaks.