It is certainly easier to prove interesting theorems with formalism. You don't get caught up with such basic things like whether or not it is always possible to tell that one real number is bigger than another.
But formalism leads to having to accept conclusions that some of us don't like. I already referred to the existence of uncountably many things that cannot in any useful way ever be specified. If you include the axiom of choice, you get the Banach-Tarski paradox. Mathematicians debated that one for a while, but now generally accept it.
My favorite example of a weird conclusion comes from https://en.wikipedia.org/wiki/Robertson%E2%80%93Seymour_theo.... We can non-constructively prove the following facts. Any class of graphs that is closed under the graph minor operation (for example planar graphs), has a finite set of forbidden graph minors that completely characterize the graph (in the case of planar graphs, K5 and K3,3). In general, there is no way to find those forbidden graph minors. Even if you were given the complete list, you couldn't necessarily verify that the list was correct. You cannot necessarily even find an upper bound on how big this set is.
By "cannot necessarily" I mean, "it is sometimes unprovable".
In what sense can a finite set exist and be finite when it is unfindable, unverifiable, and has unboundable size?
To make this concrete, there are 17,523 known forbidden minors for the toroidal graphs. We don't know how to find more. We don't know if we have the full list. And we don't have an upper bound on how many more of them there are to be found.
You're free to accept this ephemeral claim to existence as actual existence. But this existence isn't very useful for us.
But formalism leads to having to accept conclusions that some of us don't like. I already referred to the existence of uncountably many things that cannot in any useful way ever be specified. If you include the axiom of choice, you get the Banach-Tarski paradox. Mathematicians debated that one for a while, but now generally accept it.
My favorite example of a weird conclusion comes from https://en.wikipedia.org/wiki/Robertson%E2%80%93Seymour_theo.... We can non-constructively prove the following facts. Any class of graphs that is closed under the graph minor operation (for example planar graphs), has a finite set of forbidden graph minors that completely characterize the graph (in the case of planar graphs, K5 and K3,3). In general, there is no way to find those forbidden graph minors. Even if you were given the complete list, you couldn't necessarily verify that the list was correct. You cannot necessarily even find an upper bound on how big this set is.
By "cannot necessarily" I mean, "it is sometimes unprovable".
In what sense can a finite set exist and be finite when it is unfindable, unverifiable, and has unboundable size?
To make this concrete, there are 17,523 known forbidden minors for the toroidal graphs. We don't know how to find more. We don't know if we have the full list. And we don't have an upper bound on how many more of them there are to be found.
You're free to accept this ephemeral claim to existence as actual existence. But this existence isn't very useful for us.