The compactness theorem for first-order logic says that a set of first-order sentences has a model if and only if every finite subset of has a model. The forward implication is immediate. Conversely, if every finite subset is satisfiable, then is syntactically consistent: any proof of a contradiction would use only finitely many premises. The Godel completeness theorem then gives a model of .
The Upward Lowenheim-Skolem theorem says that a first-order theory with an infinite model has arbitrarily large models. Given an infinite model of and any cardinal , enlarge the language by constants for and add
Every finite subset can interpret its finitely many constants as distinct elements of the original infinite model. By the compactness theorem, the enlarged theory has a model containing at least distinct elements; its reduct is a model of .
Finally, suppose a set and a finite set axiomatize the same class. Let be the conjunction of the sentences in . Since , the set is unsatisfiable. Compactness gives a finite such that . Because , we also have , while is immediate. Therefore yes: the finite subset axiomatizes the same theory. This is finite subaxiomatization from an equivalent finite theory.
No. Let be the displayed infinite axiom set. Any finite subset mentions only finitely many of the two infinitude requirements. Choose at least as large as every index mentioned. A finite equivalence relation with exactly singleton classes and doubleton classes satisfies , but is not a first-order model of the complete theory.
If a finite axiomatization existed, let be the conjunction of its axioms. Then , so the compactness theorem gives a finite with . Every first-order model of would then satisfy the full first-order theory, contradicting the finite first-order model above. This is the finite subaxiomatization from an equivalent finite theory argument: finite axiomatizability would force one finite fragment of the explicit axioms to suffice.