In his recent book “Ignorance and Imagination,” (Oxford 2006) Daniel Stoljar attempts to defuse anti-physicalist arguments (such as the knowledge argument and zombie-style conceivability arguments) by arguing that they are only plausible because we are ignorant of important physical truths. As Stoljar realizes, it is not enough simply to advance this “ignorance hypothesis.” We are obviously ignorant of many physical truths, but a case needs to be made that the gaps in our knowledge are of the basic sort that could explain away the metaphysical significance of our anti-physicalist intuitions. (Compare: simply advancing the hypothesis that we are ignorant about the location of boundaries for apparently vague concepts doesn’t resolve the sorites paradox unless it could be explained how any gain in knowledge could satisfy us that we have correctly located a precise boundary.) In an effort to flesh out the ignorance hypothesis, then, Stoljar flirts with the view—which I will call Categorical Phenomenalism—that our ignorance of the categorical bases of physical properties explains our ignorance about the way the physical features of our world determine the experiential features of our world. Although he does not explicitly embrace this view, first articulated by Russell, he hopes that its viability confers a good deal of plausibility on the more general ignorance hypothesis, providing just one way that it could be developed.

In this post, I want to question Stoljar’s strategy, not by attacking the basic plausibility of Categorical Phenomenalism, but by questioning whether Categorical Phenomenalism is in the end a physicalist position. Stoljar must not only provide a case where our ignorance of important facts explains our dualistic intuitions, he must provide a case where the relevant ignorance is ignorance of physical facts. I maintain that either Categorical Phenomenalism doesn’t provide a plausible “missing fact” which explains our anti-materialist intuitions, or the missing fact it does provide is not properly physical.

Categorical Phenomenalism basically holds that while physical science might tell us about all the physical objects in the world, it can at best give us all of their their dispositional properties. Science can tell us what objects will do under a certain set of conditions, but it cannot tell us what objects are intrinsically. Science simply does not tell us about “categorical” properties—non-dispostional properties that are metaphysically sufficient for the posession of certain disposittions. There must be categorical properties, however: dispositions must be grounded. There must be something in virtue of which an object has a certain disposition—a glass is not simply fragile, it is fragile in virtue of a property. Thus, there are properties that physical science does not tell us about, categorical properties, and the Categorical Phenomenalist says that these are also what are responsible for (if not identical to) phenomenal properties. Thus, the ignorance which explains our anti-materialist intuitions is an ignorance of the categorical properties of physical objects which science, by its very nature, does not tell us about.

Admittedly, this sketch is rather quick, and there are many places one might wish to interject. Must there be categorical properties? Is the dispositional/categorical distinction well defined? If so, is science really silent about the categorical properties? Why think that categorical properties must be relevant to experience? Let’s pass by these issues, for now. The question is, if categorical properties really are relevant to the generation of phenomenal experience (and not in virtue of the physically described dispositions they ground, but in virtue of their intrinsic natures), are they really physical?

Defining what it is to be a physical property (in the narrow sense) is a troubled issue. I am inclined to accept a negative definition along the following lines: A property is physical iff it is a contingent, fundamental property that is not irreducibly intentional or phenomenal. (Here I follow Crook and Gillett, Papineau, and others. For full citations and an argument very closely related to the one given in this post, see my “Emergentism and Supervenience Physicalism.”) In this context, such a definition might seem to beg the question, but the general idea behind it is this: to the extent that the mind/body problem is about the truth of physicalism, it is about whether or not mental properties must be posited as basic in some sense, or whether or not they are composed out of other, more basic things of the sort we believe physics is apt to study. If it turns out that paradigmatically physical things (rocks, electrons, etc.) actually have sentient souls, for instance, the world is not fully physical. (Of course you can use “physical” as you want, but if such things are physical, then the mind body problem wasn’t really about physicalism after all.) So too, it seems to me, if the categorical bases of paradigmatically physical dispositional properties are basic phenomenal properties, the world is not as we took it to be—it contains a different sort of fundamental stuff than physics describes, and that stuff is irreducibly mental. It is, it seems, a panpsychic universe, rather than a physicalist one.

The natural response is to back off and claim that perhaps the categorical properties are not themselves phenomenal, but instead combine in such a way to give rise to and explain phenomenal properties. I think this doesn’t help. To explain or give rise to phenomenal properties in the relevant sense, they must metaphysically necessitate them (perhaps only in certain conditions or when combined with other properties, etc.). If this is the case, then part of the nature of these categorical properties is irreducibly mental. Consider categorical property F. It is itself not phenomenal, but in a set of circumstances C it necessarily gives rise to P—a phenomenal property. Call this propensity to give rise to phenomenality in those conditions P(c). Perhaps F also grounds a set of dispositional properties of a more familiar sort, d1, d2…dn. To fully specify the nature of F, one must include (d1…dn, P(c)). Consider F*, which is just like F but it does not have P(c). It looks as if the only way to distinguish F from F* is to appeal to the fact that F necessitates phenomenal properties in C. That is, F cannot be exhaustively described without ineliminable, irreducible appeal to phenomenality. This “panprotopsychism” hardly seems more physical in the relevant sense thant the original panpsychism.

The result, I think, is that if Categorical Phenomenalism provides a case where ignorance of facts explains our anti-materialist intuitions, it is because those facts are themselves irreducibly mental, which is to say non-physical. If the only difference between an electron and a schmelectron is that electrons have a nature that gives rise to phenomenal consciousness, then to our surprise, electrons are not physical, but are only thought to be so because science cannot tell the difference between them and their schpectacular counterparts.

Posted by Robert Howell on Saturday, November 18th, 2006 at 6:01 pm.