Symbolic LogicNatural Deduction: Rules of Inference

Proof Strategy: Working Backward

Find the conclusion inside the premises, then plan which rule would produce it and what that rule needs.

Not started Jump to practice

Building Short Proofs gave the main idea: work backward from the conclusion. This lesson turns it into a short checklist, with the cases that trip people up most.

Key idea

Find the conclusion, or its letters, inside the premises. Plan the last step first: which rule would produce the conclusion, and what does it need? Then treat that need as a new goal. When the conclusion contains a letter that appears in no premise, Add is almost always how it gets in.

Why it works

Working forward, the number of possible lines grows quickly; working backward, each step narrows the search to what one rule needs. It is like planning a trip from the destination: to arrive by train, you must first reach the station. Each premise is a resource; the conclusion tells you which resources to spend.
Working backward.
The conclusion is …Likely last stepYou will need
the consequent of a conditional premiseMPthat conditional's antecedent
the negation of an antecedentMTthe negation of the consequent
a disjunct of a disjunction premiseDSthe negation of the other disjunct
a conjunct of a conjunctionSimpthe conjunction
a conjunctionConjeach conjunct
a disjunction with a new letterAddone of its disjuncts
a conditionalHStwo conditionals that chain
Planned backward, written forward.

Worked examples

Worked example: Plan backward

Premises: $t \to r$, $s \to t$, $s$. Conclusion $r$. Plan.
  1. $r$ is the consequent of $t \to r$: last step MP, needing $t$.
  2. $t$ is the consequent of $s \to t$: MP again, needing $s$.
  3. $s$ is a premise.

Answer: Line 4: $t$ (MP 2, 3). Line 5: $r$ (MP 1, 4).

Worked example: A new letter

Premises: $p \to q$, $p$. Conclusion $q \lor u$. How does $u$ get in?
  1. $u$ is in no premise.
  2. Get $q$ by MP, then Add.

Answer: Line 3: $q$ (MP 1, 2). Line 4: $q \lor u$ (Add 3).

Worked example: Two goals

Premises: $p \land q$, $p \to r$, $q \to s$. Conclusion $r \land s$.
  1. Last step Conj: need $r$ and $s$.
  2. $r$: MP from $p \to r$ and $p$ (Simp 1).
  3. $s$: MP from $q \to s$ and $q$ (Simp 1).

Answer: Six lines: Simp, MP, Simp, MP, Conj.

When you are stuck, try writing the goal on scrap paper and asking which premise mentions it. If a premise mentions it inside a conditional, the other side of that conditional becomes your next goal. If none mentions it at all, look for Add. If two premises mention it, try each; one of the routes usually comes out shorter.

Common mistake

Never using Add, because it seems to lose information. When the conclusion has a letter that no premise has, no other rule can produce it. Look for that clue first: it tells you the last step at once.

Practice

Pass with 3 right answers, the last 2 in a row. Each right answer earns 3 XP.

Loading practice…

Builds on

Leads to

Sources and further reading

  • Patrick J. Hurley and Lori Watson, A Concise Introduction to Logic, 13th ed., section 7.2, strategy
  • forall x: Calgary, an open introduction to formal logic (P. D. Magnus, Tim Button and others; Open Logic Project), chapter on proof strategies
@teddyrookbook GitHub