The reason I referred to Descartes is that, in him, the idea of res cogitans and res extensa became clearly formulated. This division - between mind and body, spirit and matter - has become fundamental to the ‘grammar of culture’. It is part of the broader division introduced with modern science that divides the primary from secondary attributes of things, and the subjective and objective as domains of reference. It is referred to as ‘the Cartesian division’, subject of another post, Descartes’ Ghost.
But the dualism of Plato’s philosophy, modified but adapted by Aristotle, is the dualism of matter and form (hyle-morphē). And that is radically different. In Aristotle’s dualism, matter can’t exist without form (as there would be no answer to the question ‘what is that?’ for a thing without form), but forms are instantiated in matter.
Aristotle aside, what kinds of things are immaterial? The example I gave in this OP is the sign ‘=’. I argue that ‘equals’ is a distinct concept with a specific meaning, but that its nature is entirely intellectual (noetic or noumenal). And that idea can be extended to the entire domain of mathematical objects and relations. But I wouldn’t want to say that numbers are immaterial things. That’s where the Cartesian error creeps back in.
So, to respond to your challenge: yes, eyes, or at least, vision has a clearly physical or physiological basis - photosensitivity, transmission of signals via nerves, assimilation by the brain. This plainly operates in any organism with sight, and same too with h.sapiens.
But the non-material factor is interpretation - the identification of the object of perception, and its assimilation into the mind. This is the aspect that resists reduction to the physical because it is essential to the process by which the meaning of ‘physical’ is ascertained.
(That is the problem of recursion that was the subject of, for example, Douglas Hofstader’s I Am a Strange Loop, among other books.)
But where I’m going with that argument is that not only mathematics, but all formal symbolic systems, including the rules of logic, scientific conventions and laws - none of these can be reduced to or explained in terms of purely physical principles (which is what physicalism seeks to do, usually through a combination of neuroscience and evolutionary biology.)
This is at the basis of the interminable debates about platonism in philosophy of mathematics. The platonist (note small ‘p’) view is that numbers are real - to which they empiricist rejoinder is invariably, ‘well, where are they, then? Where in nature can numbers be found?’ I say, nowhere — but that they’re nonetheless real for anyone who can count.