# Provability of a theorem?

1. Apr 29, 2013

### cragar

If a theorem required an uncountable amount of information or symbols to prove it, would this mean it could be true but unprovable. Are we just limited because we can only write a countable number of symbols? Could the theorem be proved in some other sense?

2. May 3, 2013

### haruspex

Necessarily, a proof is of finite length. Of course, it may be unprovable in one formal system for this reason, but provable in a more powerful one.

3. May 3, 2013

### cragar

so in a certain model that theorem might require an infinite amount of information to prove it, but in another more powerful model it might only need a finite amount of information to prove it. When you say more powerful, do you mean different axioms of more axioms.

4. May 3, 2013

### haruspex

If different axioms it would be effectively a different theorem. It would have to be added axioms.

5. May 4, 2013

### cragar

ok, so If i had a theorem that required an infinite amount of information to prove it, but then if I just took that as an axiom I could could compress an infinite amount of information into a single finite statement.

6. May 4, 2013

### haruspex

Maybe, but I'm doubtful about the very concept of a theorem requiring an infinite proof. Proofs are not infinite by definition. All you can say is that the proposition is neither provable nor disprovable within the system, in which case, yes, you can add it (or its negation) as an axiom.

7. May 25, 2013

### lugita15

Well, at least normally. But look at this.