16.05.2014 Views

C:\Documents and Settings\dave\Desktop\Godel, Escher, Bach ...

C:\Documents and Settings\dave\Desktop\Godel, Escher, Bach ...

C:\Documents and Settings\dave\Desktop\Godel, Escher, Bach ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

the day: how to prove a system consistent. Gödel found a simple way to express the<br />

statement "TNT is consistent" in a TNT formula; <strong>and</strong> then he showed that this formula<br />

(<strong>and</strong> all others which express the same idea) are only theorems of TNT under one<br />

condition: that TNT is inconsistent. This perverse result was a severe blow to optimists<br />

who expected that one could find a rigorous proof that mathematics is contradiction-free.<br />

How do you express the statement "TNT is consistent" inside TNT It hinges on<br />

this simple fact: that inconsistency means that two formulas, x <strong>and</strong> x, one the negation of<br />

the other, are both theorems. But if both x <strong>and</strong> -- x are theorems, then according to the<br />

Propositional Calculus, all well-formed formulas are theorems. Thus, to show TNT's<br />

consistency, it would suffice to exhibit one single sentence of TNT which can be proven<br />

to be a nontheorem. Therefore, one way to express "TNT is consistent" is to say "The<br />

formula -0=0 is not a theorem of TNT". This was already proposed as an exercise a few<br />

pages back. The translation is:<br />

---3a:TNT-PROOF- PAIR{a,SSSSS<br />

SSSSSOIa'}<br />

223,666,111,666 S's<br />

It can be shown, by lengthy but fairly straightforward reasoning, that-as long as TNT is<br />

consistent-this oath-of-consistency by TNT is not a theorem of TNT. So TNT's powers<br />

of introspection are great when it comes to expressing things, but fairly weak when it<br />

comes to proving them. This is quite a provocative result, if one applies it metaphorically<br />

to the human problem of self-knowledge.<br />

TNT Is ω-Incomplete<br />

Now what variety of incompleteness does TNT "enjoy? We shall see that TNT's<br />

incompleteness is of the "omega" variety-defined in Chapter VIII. This means that there<br />

is some infinite pyramidal family of strings all of which are theorems, but whose<br />

associated "summarizing string" is a nontheorem. It is easy to exhibit the summarizing<br />

string which is a nontheorem:<br />

u S's<br />

Va: 3a':<br />

To underst<strong>and</strong> why this string is a nontheorem, notice that it is extremely similar to G<br />

itself-in fact, G can be made from it in one step (viz., according to TNT's Rule of<br />

Interchange). Therefore, if it were a theorem, so would G be. But since G isn't a theorem,<br />

neither can this be.<br />

Now we want to show that all of the strings in the related pyramidal family are<br />

theorems. We can write them own easily enough:<br />

On Formally Undecidable Propositions 444

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!