
120 Computability Theory
That is, our witness (our true unprovable sentence) asserts, in a sense, that it is itself
not in the axiomatic theory that yields T. It is saying (under this rather free translation),
“I am unprovable in this axiomatic theory.”
Digression: In 1931, G
¨
odel did not have the development of computability theory
available to him. Instead, he proceeded directly to an ingenious construction of a sen-
tence that could be freely translated as saying “I am unprovable in this axiomatic
theory.” This sentence had to be true (if it were false, we would have a provable false-
hood), and hence unprovable in the axiomatic theory. Even better ...