Ben Martin 29/02/18
Benjamin.martin@ucl.ac.uk
PHIL2048: Intermediate Logic
Seminar Seven: Derivations - Solutions
So far, we have been proving that some formula f follows from a set of formulae G by using semantics. We have
been appealing to the meaning of the logical connectives, and so on.
We express these results with the double turnstile, ⊨. If f is a logical consequence of G, then G ⊨
f.
We can also demonstrate that some formula f follows from a set of formulae G by deriving f from G using a fixed
set of rules, known as derivation rules.
We express these results with the single turnstile, ⊢. If f can be derived from G, then G ⊢
f.
There are many different systems that can be used to prove derivations, but we will be using natural deduction,
which is one of the most popular kinds of proof calculus.
We are given certain rules of inference, and using these rules of inference (and only these), we can manipulate the
premises we are given to derive the intended conclusion.
As you know, each connective has its own introduction and elimination rules, and we give some to identity too.
RULES OF INFERENCE
We have our base rule:
(B) G ⊢
f whenever f Î G
Conjunction:
G ⊢
f, D ⊢
y (ÙI) G ⊢
f Ù y (ÙE)
G È D ⊢ f Ù y G ⊢
f G ⊢
y
Disjunction:
G ⊢
f G ⊢
y (ÚI) G ⊢
f Ú y, D È{f}⊢
g, S È{y}⊢
g (ÚE)
G ⊢
f Ú y G È D È S ⊢ g
Conditional:
G È {f}⊢
y (®I) G ⊢
f, D ⊢
f ® y (®E)
G ⊢
f ® y G È D ⊢ y
Negation:
G È {f}⊢
y, D È {f}⊢
¬y (¬I) G ⊢
¬¬f (¬E)
G È D ⊢
¬f G ⊢
f
1
Ben Martin 29/02/18
Benjamin.martin@ucl.ac.uk
Universal quantifier:
G ⊢
(f)[c/x] ("I) G ⊢
"xf ("E)
G ⊢
"xf G ⊢
(f)[t/x]
[On the assumption c doesn’t occur in f or G]
Existential quantifier:
G ⊢
(f)[t/x] ($I) G ⊢
$xf, D È {(f)[c/x]} ⊢
y ($E)
G ⊢
$xf G È D ⊢ y
[On the assumption c doesn’t occur in f, y or D]
Identity:
G ⊢
t » t (»I) G ⊢
(f)[t/x], D ⊢ t » u (»E)
G È D ⊢
(f)[u/x]
PROOFS:
LEMMA 2.16: Show that for all L-formulae f, y and g,
[4] That we can establish {(f Ú y), (f ® g), (y ® g)} ⊢
g
{(f ® g)}⊢
(f ® g) (B) {f} ⊢
f (B) {(y ® g)}⊢
(y ® g) (B) {y} ⊢
y (B)
{(f ® g), f}⊢ g (®E) {(y ® g), y}⊢ g (®E) {(f Ú y)} ⊢(f Ú y) (B)
{(f Ú y), (f ® g), (y ® g)} ⊢
g (ÚE)
LEMMA 2.17: Show that for all L-formulae f, y, g and d,
[6] That we can establish {(f ® g), (y ® d)} ⊢ ((f Ù y) ® (g Ù d))
{(f Ù y)} ⊢
(f Ù y) (B)
{(f ® g)} ⊢
(f ® g) (B) {(f Ù y)} ⊢
f (ÙE) {(f Ù y)} ⊢ y (ÙE) {(y ® d)} ⊢
(y ® d) (B)
{(f ® g), (f Ù y)} ⊢
g (®E) {(y ® d), (f Ù y)} ⊢
d (®E)
{(f ® g), (y ® d), (f Ù y)} ⊢
(g Ù d) (ÙI)
{(f ® g), (y ® d)} ⊢
((f Ù y) ®
(g Ù d) (®I)
2
Ben Martin 29/02/18
Benjamin.martin@ucl.ac.uk
EXERCISES:
EXERCISE 2.17: Show that for all L-formulae f, y and g,
[2] That we can establish {(f ®y), (f Ù g)} ⊢ (y Ù g)
{(f Ù g)} ⊢
(f Ù g) (B)
{(f ®y)}⊢ (f ®y) (B) {(f Ù g)} ⊢
f (ÙE) {(f Ù g)} ⊢
(f Ù g) (B)
{(f ®y), (f Ù g)} ⊢
y (®E) {(f Ù g)} ⊢
g (ÙE)
{(f ®y), (f Ù g)} ⊢
(y Ù g) (ÙI)
[5] That we can establish {(f ® ¬y), (g ® y)} ⊢ ¬(f Ù g)
{(f Ù g)} ⊢
(f Ù g) (B) {(f Ù g)} ⊢
(f Ù g) (B)
{(f ® ¬y)} ⊢ (f ® ¬y) (B) {(f Ù g)} ⊢
f (ÙE) {(f Ù g)} ⊢
g (ÙE) {(g ® y)} ⊢ (g ® y) (B)
{(f ® ¬y), (f Ù g)} ⊢ ¬y (®E) {(g ® y), (f Ù g)} ⊢ y (®E)
{(f ® ¬y), (g ® y)} ⊢ ¬(f Ù g) (¬I)