Pythagorean said:
Since no one responded to this, I figure I'll reduce the workload for you:
What's the point you're trying to make? Apart from the fact Wigner represents the mysterian take on the question.
As I have argued, maths is "unreasonably effective" because it separates the constraints from the construction, the universals from the particulars.
For anything to exist, to develop and persist, it must be organised as "a system". It must have the self-organising form of global constraints and local constructions gone to no longer changing equilbrium.
Maths is a way of modelling this truth, without ever acknowledging this truth (it seems).
What maths does is freeze the global constraints (which in the systems view, are actually subject to development and change) and so greatly simplifies the task of modelling. The global constraints become axiomatic - things that just are. Things that themselves don't get explained. Then this sets up a world of all possible combinations of constructive action made permissible by a certain collection of fixed global constraints.
Take the famous case of Euclid's postulates of geometry. There was that axiom about parallel lines. It seemed very sound as an unchangeable constraint on reality. Then relax that constraint - make flatness something which has a developmental history rather than something that is frozen in - and a new more general view of geometry can be seen.
So what does that tell us about "unreasonable effectiveness"? It tells me that we start off assuming that the state of the world we appear to observe (populated by solid objects, limited to three flat dimensions, etc) is fixed that way. The global constraints just are.
Yet when we formulate that as a set of axioms, we then make it very clear to ourselves that they are assumptions. Which can be relaxed. And then in the history of maths, it became clear that relaxing the constraints - seeking the less constrained story that is more general - was a productive route for developing maths. The game became, let's throw overboard any fixed assumptions we can, because what we remove can always be added back in the form of a particular constraint on our imagined system.
So for example, you can shift from geometry (with its definite distances) to topology (with its indefinite distances). Distance becomes a constraint that can be added back into the more generalised description as need be. Just as curvature (or its lack) became an additive ingredient in the shift from Euclidean to non-Euclidean geometry.
So maths is an "unreasonably effective" approach to modelling because it does something unreasonable - freezes the global constraints of a system and pushes questions about their development, their reasonableness, right out of the frame.
You presume the global constraints as axioms. And if anyone queries this, you claim this is a free choice that commits you to no ontology. It is the mathematician's prerogative to state any axiom and explore its consequences just because it is interesting or beautiful (Wigner's argument). Maths claims this fundamental disconnect from reality, from experience. Even if, on closer examination, we discover the axioms being justified by their "natural logic" - and so derived in fact from experience.
Yet because maths has also taken a systematic approach to relaxing the constraints - becoming ever more general - it has paved the way for physics to do the same.
We live in a highly constrained reality (a universe with many very particular features). An unconstrained state is symmetric. A constrained state is symmetry broken. So to see our reality in terms of more general laws, we need to unwind the symmetry breakings. We must describe reality in less constrained terms.
Which is why maths is unreasonably effective. It is a method of successively relaxing constraints - while at the same time, keeping them always frozen, always something that can be added back in at will. This is a very tractable approach - it allows for calculation. All the dynamics gets reduced to the play of numbers - local, atomistic, additive, constructive action, or effective causality.
At the same time, maths is also very ineffective when it comes to the modelling of global constraints as self-organising, downward causal, developing and evolving, parts of the story.
There are new areas of maths that seem to be tackling this problem now. Hierarchy theory, infodynamics, the various other tools being used by systems scientists. And new more suitable brands of logic, like Peircean semiotics based on a logic of vagueness.