Short excerpt from ch. 5 of the book that I am currently writing, A Critical Introduction to Causal Powers & Dispositions
“By 2004, Mumford had revised his view of properties. In place of the idea that properties are such that they can be conceptualized both in terms of dispositional predicates, which are analytically connected to given causal roles, and in terms of categorical predicates, which are related to those same causal roles, but only a posteriori (tokens of which predicates are functionally identical across types) — in place of this, Mumford came to the position, advanced in Laws of Nature (and further refined in an article published that same year), that properties just are “clusters of powers,” as he put it. Xzq RG
I will say a bit more about the revised view in a moment, but a preliminary observation is in order. Already in Dispositions, Mumford was explicit that all properties are causally potent, as he put it. D zqw 164. Crucially, however, in 1998 he distinguished between neutral monism and any version of the claim that there are no real categorical properties. After 2004, by contrast, he took his continued commitment to the idea that no properties are causally impotent to preclude the existence of categorical properties.
The shift from neutral monism to pandispositionalism, as it has come to be called, is curious in that not only did Mumford contend in Dispositions that categorically characterized properties are just as efficicious as dispositionally characterized properties, he also made the point there that dispositionally characterized properties are themselves occurrent, in the sense that while powers can be had without being expressed, even when they are unexpressed they are had without qualification, in the present.
Given that he did not subsequently modify or retract either of these claims, it is tempting to suggest that what the ostensible repudiation of categorical properties signals is simply a definitional change: by 2004, any property that plays a causal role had come to be counted as dispositional (only), and only those that do not play such a role (if there are any) would be considered to be categorical. Since all properties are causally potent, from the perspective of neutral monism, it follows that, under the updated meanings of the terms, there can be so such thing as a categorical property, according to a neutral monist.
Confirming my diagnosis is Mumford’s comment, in 2004, that “if categorical properties are meant to be properties that are not powerful, then there is some doubt as to whether they exist.” In RG, 149-150 qx. Manifestly, that is not how he conceptualized them in Dispositions; indeed, at the end of the passage in which this sentence figures, Mumford refers back to his earlier work “for an attempt to make some sense of them.” Xzq RG 150.