4.2.3. Further examples

Both disjunction rules are illustrated by the derivation at the right, in which one grouping of a three-part disjunction is shown to entail the other. Choices between the two ways of planning for a goal disjunction were made at stages 2, 3, 5, 6, and 7 in accordance with the rules of thumb given above. Each choice helped to shorten the derivation—though only by a few steps. The derivation is contrived to provide several examples of this rule; we might have instead planned for initial the goal at stage 1 before exploiting the premise rather than planning for it separately in each of three gaps.

 
│A ∨ (B ∨ C) 1
├─
││A (4)
│├─
│││¬ C
││├─
││││¬ B
│││├─
││││●
│││├─
4 QED ││││A 3
││├─
3 PE │││A ∨ B 2
│├─
2 PE ││(A ∨ B) ∨ C 1
││B ∨ C 5
│├─
│││B (8)
││├─
││││¬ C
│││├─
│││││¬ A
││││├─
│││││●
││││├─
8 QED │││││B 7
│││├─
7 PE ││││A ∨ B 6
││├─
6 PE │││(A ∨ B) ∨ C 5
││
│││C (10)
││├─
││││¬ (A ∨ B)
│││├─
││││●
│││├─
10 QED ││││C 9
││├─
9 PE │││(A ∨ B) ∨ C 5
│├─
5 PC ││(A ∨ B) ∨ C 1
├─
1 PC │(A ∨ B) ∨ C

The scale of the difference you can expect a choice between the two forms of PE to make is illustrated by the two derivations below.

│B (3)
├─
││¬ A
│├─
│││¬ C
││├─
│││●
││├─
3 QED │││B 2
│├─
2 PE ││B ∨ C 1
├─
1 PE │A ∨ (B ∨ C)
 
│B (5)
├─
││¬ (B ∨ C) 3
│├─
│││¬ A
││├─
│││││¬ C
││││├─
│││││●
││││├─
5 QED │││││B 4
│││├─
4 PE ││││B ∨ C 3
││├─
3 CR │││⊥ 2
│├─
2 IP ││A 1
├─
1 PE │A ∨ (B ∨ C)

Each chooses a different way of planning for the initial goal at stage 1. Notice that in the second, which makes the less efficient choice, we are led back to the goal B ∨ C in a couple of stages.

Glen Helman 15 Aug 2006