back

by throwoutway·2y ago·view on hn ↗
Which? Are they consistent? Are they useful?

From wikipedia:

> The incompleteness theorems apply to formal systems that are of sufficient complexity to express the basic arithmetic of the natural numbers and which are consistent and effectively axiomatized.

2 comments
Tarski's axioms for geometry are the standard example of a system which is both complete and consistent. It is even decidable - there is an algorithm which you can feed a statement and it outputs whether it is true or not.
Ok, how relevant is Tarski's though? Seems academically, they are weak

> We partially prove the theorem of [7] that Tarski’s (extremely weak!) plane geometry axioms imply Hilbert’s axioms. https://sciendo.com/article/10.2478/forma-2014-0017

Also, thats for geometric space but I dont see how that's particularly relevant to this. Elsewhere I find that it doesn't contradict Gödel

I don't understand your questions. They are relevant for geometry, they aren't really relevant to anything else but they're axioms for geometry so being relevant for geometry is what you'd expect.
IIRC, incompleteness theorems applies to arithmetic not geometry, the difference being subtle right?
The original incompleteness theorems apply to any logical system which can formulate and prove a certain chunk of arithmetic, yes. It turns out that the amount of arithmetic you actually need is incredibly small, so these results apply to many systems including some formulations of geometry.
Oh, by the way, it matters here that "weak" in the source you quote had a very specific technical meaning. If axioms are weak it means they apply to many things. Stronger axioms make more restrictions and apply to more specific things. So (arguably) axioms being weak is a good thing. Generally you want to work with axioms that are as weak as you can get away with.
Good to know thanks
Yes, they are consistent. Yes, there are useful systems.

See https://en.wikipedia.org/wiki/Agda_(programming_language) for a reasonably practical example:

> Agda is a total language, i.e., each program in it must terminate and all possible patterns must be matched.