Reason is utterly insufficient to determine if there are infinite sets.
The only thing that guarantees the existence of an infinite set is the axiom of infinity. — fishfry
I should not have said 'formally'. I meant informal deduction. My point was to contrast informal deduction with "mind's eye" visualization. I reason from the notion that there is the property of being a natural number. And I don't claim that reasoning is as rigorous in avoiding paradox as axioms. My point was at a very broad level of generality, not necessarily responsible to formal contradictions. I should have counted on someone not recognizing that my remarks are merely in the spirit of personal reflection. Moreover, I am not advocating such reflections as philosophical conclusions that others necessarily should adopt.
If we adopt the axiom of infinity, there are infinite sets. If we reject the axiom of infinity, there are no infinite sets. It's as simple as that. — fishfry
It's not as a simple as that, and it is not correct:
(1) To reject the axiom of infinity is to not include it in the axioms. Not including the axiom of infinity does not entail that there are no infinite sets. Rather, to entail that there are no infinite sets requires both not including the axiom of infinity and also adopting the negation of the axiom of infinity.
(2) The axiom of infinity is used for many axiomatizations. It is not precluded that one may axiomatize differently and still derive the theorem that there exist infinite sets. It is fine to say that with ZFC, for example that the axiom of infinity is required to prove the existence of an infinite set; but my remarks were not specific to axiom systems.
The existence of an infinite set is purely a matter of accepting or rejecting the assumption that says there is an infinite set. — fishfry
Again, I should not have said 'formal'. I meant informal deduction.
It's easy: I understand the property of being a natural number; and I understand the idea that given a property, there is the set of things that have that property (with some restrictions on that principle).
— TonesInDeepFreeze
Your principle leads directly to a contradiction. The restriction needed to patch the problem is restricted comprehension, which would already require there to be an infinite set that's a superset of the infinite set you wish to conjure. — fishfry
(1) The axiom of infinity provides that there is a successor-inductive set. The axiom of separation provides deriving the set of natural numbers from the fact that there exists a successor-inductive set. w is the unique set that is a subset of all successor-inductive sets. There is no particular proper superset mentioned. Only that there is a successor-inductive set (whether or not a proper superset of any other successor-inductive sets). Then, after we have proven that there is a unique set that is a subset (notice: not even presuming that it is a proper subset) of all successor-inductive sets, we prove that there are proper supersets of w that are successor-inductive (e.g. w+w).
(2) I only described (subjective) reasoning, not even associated with any particular axioms. I mentioned that the principle "for every property there is a set" would not be taken without restriction (if I were to expand on that remark, it would be that such properties as "is not a member of itself" or more generally, unrestricted comprehension don't fly). Taking the property "is a natural number" to provide the set of natural numbers is not contradictory.
One problem though does haunt me: Every arithmetical statement is either true or false.
— TonesInDeepFreeze
No. Just no. Given an arithmetical statement (a syntactic entity) AND AN INTERPRETATION of the symbols of that statement (semantics) then the statement becomes true or false. — fishfry
Of course truth is relative to models. But mathematicians and philosophers most often take the liberty of saying "true" or "true in arithmetic" or "true statement of arithmetic" implicitly to mean "true in the standard model for PA". When a mathematician or philosopher says "it's a true statement of arithmetic", it is implicit that more formally that would be "is true in the standard model of PA".
Indeed, at least a few times in the past, I've mentioned to you that in greater formality, of course, sentences are only true or false per a given model or per a class of models (by the way, 'class' also is used informally (but such that with other locutions we can avoid 'class') there since, at least in Z set theories, there does not exist the proper class of all the models of a theory.
So when I say "true statement of arithmetic" you may always regard it at the usual shorthand for "true in the standard model of PA".