Two ways are considered of formulating a version of retentive physicalism, the view that in some important sense everything is physical, even though there do exist properties, e.g. higher-level scientific ones, which cannot be type-identified with physical properties. The first way makes use of disjunction, but is rejected on the grounds that the results yield claims that are either false or insufficiently materialist. The second way, realisation physicalism, appeals to the correlative notions of a functional property and its realisation, and (...) states, roughly, that any actual property whatsoever is either itself a physical property or else is, ultimately, realised by instances of physical properties. Realisation physicalism is distinctive since it makes no claims of identity whatsoever, and involves no appeal to the dubious concept of supervenience. After an attempt to formulate realisation physicalism more precisely, I explore a way in which, in principle, we could obtain evidence of its truth. (shrink)