Physical reality is an explanatory posit: we posit it to explain various features of our experience, such as its prima facie non-randomness, and its usefulness as a basis for making predictions about further experience.
As with any explanatory posit, we’re naturally curious to know more about this one. We’re no more inclined to think that the nature of physical reality is exhausted by its explaining the regularity of experience than to think that the nature of genes is exhausted by their explaining biological heredity: we want to know what further nature the posit has, in virtue of which it explains what it does.
Different theories of the fundamental nature of physical reality (or “analyses of matter,” in an older parlance) are different accounts of this further nature. According to classic metaphysical realists, it’s a purely geometric and kinematic nature; according to modern day structural realists, it’s a purely mathematical or mathematico-causal nature; according to traditional idealists, it’s a sentient agentive nature; according to panpsychists, it’s a micro- or proto-phenomenal nature; according to Kantians, it’s a nature about which we can know nothing more than that it grounds various experience-causing powers. According to J.S. Mill and his followers (such as C.I. Lewis and A.J. Ayer), it’s an objective tendency for experience to occur certain ways, given that it occurs certain other ways.
The main thrust of my metaphysical writings is that Mill’s theory, which has come to be known as phenomenalism, is basically correct. Reality fundamentally comprises conscious experiences, and objective probabilities for certain experiences to occur conditional on the occurrence of other experiences. The probabilities constitute the physical part of reality.