David Papineau 1 Introduction I take myself to be a physicalist. I hold that all facts, including such prima facie non-physical facts as mental and biological facts, metaphysically supervene on the physical facts.
However, I do not have any views about the relationship between macroscopic and microscopic facts. I am neutral on such questions as whether big things are always made of small things.
Recently I have become worried about this combination of views. This is because many other philosophers seem to think of physicalism as some kind of commitment to the primacy of the microscopic. In their view, physicalism doesn’t just say that everything is physical. It also says that everything is microscopically determined.
Here are some representative quotations:
‘[Physicalism is] the doctrine that actually (but not necessarily) everything non-microphysical is composed out of microphysical entities and is governed by microphysical laws’ (Pettit 1994, 253. See also Petit 1993).
‘Any thing of any level except the lowest must possess a decomposition into things belonging to the next lower level. In this sense each level, will be as it were a “common denominator” for the level immediately above it’ (Oppenheim and Putnam 1958).
‘The bottom level is usually thought to consist of elementary particles, or whatever our best physics is going to tell us are the basic bits of matter out of which all material things are composed. As we go up the ladder, we successively encounter atoms, molecules, cells, larger living organisms, and so on. The ordering relation that generates the hierarchical structure is the mereological (part-whole) relation: entities belonging to a given level, except those at the very bottom, have an exhaustive decomposition, without remainder, into entities belonging to the lower levels’ (Kim 1998).
‘[Physicalism requires] a mereological structure, ordered by the part-whole relation . . .’ (Schaffer 2003).
Perhaps I have been missing something. Despite my lack of interest in the issue, maybe physicalism does entail that everything is microphysically determined.
But there is another possibility. Perhaps there are two separable theses associated with physicalism, and the philosophers just quoted are unjustifiably running them together. This is what I shall argue in this paper. I shall distinguish physicalism per se from a further thesis about microphysical determination, and I shall argue that these two theses are independent. Physicalists don’t have to be Microphysicalists.1 2 Two Theses Let me start with what I take to be the basic content of physicalism.
(P) All facts metaphysically supervene on the physical facts.
For clarity, I shall capitalize this thesis henceforth as ‘Physicalism’.
Now consider this further claim about the nature of the physical facts themselves.
(M) All physical facts metaphysically supervene on the microphysical facts.
I shall call this thesis ‘Physical Microscopism’.
On the surface, it certainly looks as if these two theses could be independent. Physicalism is a doctrine about the relationship between prima facie non-physical things and physical things. It says that the mental, biological, meteorological and other prima facie non-physical things—that is, those things that can be directly identified using mental, biological, meteorological and other non-physical vocabulary—are not in fact ontologically distinct from physical things. Physicalism thus tells us how prima facie non-physical realms relate to the physical realm.
Physical Microscopism, by contrast, doesn’t say anything about the relationship between the physical and other realms. Rather it is a doctrine about how things go within physics itself. It says that all physical facts are fixed by microphysical facts. It doesn’t say anything about prima facie non-physical things.2 The theses expressed in the quotations above can be viewed as the conjunction of Physicalism and Physical Microscopism. Let us define ‘Microphysicalism’ as the thesis that
(A) All the facts supervene on the microphysical facts.
Microphysicalism so defined is equivalent to the conjunction of (P) and (M).
To verify this equivalence, note first that, if (P) everything supervenes on physical facts and (M) all physical facts supervene on microphysical facts, then (A) everything supervenes on microphysical facts. Conversely, if (A) everything supervenes on microphysical facts, then immediately (M) everything physical supervenes on microphysical facts, and also (P) anything prime facie non-physical supervenes on microphysical facts and so a fortiori on physical facts.
The Microphyicalist doctrines quoted above are thus be committed to the conjunction of Physicalism and Physical Microscopism.3 By the same coin, there are philosophers who deny both Physicalism and Physical Microscopism. Not only do they defend the traditional dualist view that non-physical realms like the mental are ontologically separate from the physical realm, but they also maintain holist doctrines about the physical realm itself, insisting that certain kinds of physical wholes are metaphysically more than the sum of their microphysical parts. (Thus Crane and Mellor’s influential ‘There is no Question of Physicalism’ (1990) defends a version of this extreme anti-Microphysicalism.)
However, I shall be arguing that it is not mandatory to tie Physicalism to Physical Microscopism in this way. By way of preliminary support for this claim, note that the other two combinations of assertion and denial of Physicalism and Physical Microscopism also make perfectly good initial sense.
Thus there is the possibility of defending Physical Microscopism while rejecting Physicalism. I would have thought that this was Descartes’s view, for example. Even though Descartes is a paradigm dualist about the relation between the mental and physical realms, within physics itself he certainly looks like someone who thinks that the microphysical facts at least fix all the physical facts. We might also expect some contemporary dualists, such as David Chalmers, similarly to uphold this combination. There is no obvious reason why their commitment to an ontologically separate mental realm should force them to any kind of holism within physics itself.
The other possibility is Physicalism without Physical Microscopism. This is the option that interests me. The Microphysicalist quotations above suggest that once you are a Physicalist, then this will carry Physical Microscopism in its train. But why should this follow? Suppose I am a Physicalist about the mind. I think that the mental level is determined by the physical level. There is nothing more to the mind than the brain. Why should this commit me to any view in particular about the way things go within physics? Why shouldn’t I hold that physical wholes transcend what is determined by their microphysical parts? Such a within-physics holism would seem perfectly consistent with my rejection of Cartesian dualism. Can’t I still identify mental facts with macrophysical facts, even if I that think those macrophyical facts transcend what is determined by microphysical parts?
This anyway is the possibility that will concern me in the rest of this paper. Can one be a Physicalist without embracing Physical Microscopism? Equivalently, must a Physicalist be a Microphysicalist?
3 Motivations for Microphysicalism Why might anybody think that Physicalism requires Physical Microscopism? Are the Microphysicalist views expressed in the earlier quotations just an oversight, betraying insufficient thought about the nature of Physicalism? Or is there some more principled reason for linking Physicalism to Physical Microscopism?
I can think of two possible reasons for forging this link. The first is to do with the meaning of ‘physical’. The second relates to the availability of arguments for Physicalism. Let me consider these possible reasons now, as they will allow me to introduce some points that will be useful later. I shall take them in turn.
The difficulties involved in defining ‘physical’ are well known. As Carl Hempel (1969) pointed out many years ago, Physicalists cannot simply define this term in terms of the categories recognized in contemporary Physics Departments. This is because current physics is a work in progress, so to speak—future discoveries will no doubt add to and subtract from the categories recognized by current physical theory. So a ‘Physicalism’ that asserts that everything supervenes on currently recognized physical categories will almost certainly prove false. Nor is it much of a solution, Hempel added, to define ‘physical’ by reference to the categories that will be recognized by future Physics Departments—at the ideal end of enquiry, perhaps. To the extent that we currently lack any clear idea of what those categories will be, this would remove any substantial content from Physicalism.
In the face of this dilemma, one possible solution is to define ‘physical’ in terms of ‘microphysical’. That is, we might read ‘physical’ as encompassing only what is microphysically determined. Philip Pettit understands ‘physical’ in this way. The passage quoted earlier is part of an argument designed to show that ‘physical’ can be defined as ‘composed out of microphysical entities and governed by microphysical laws’. By this proposal, Pettit hopes to counter the view that there is no good way of understanding ‘physical’ and that ‘Physicalism’ is therefore an empty doctrine.
Now, if we do define ‘physical’ as Pettit does, then Physical Microscopism will become a definitional truism. All physical facts will inevitably supervene on microphysical ones, for if they didn’t they wouldn’t be ‘physical’. And therewith the Physicalist claim that everything is physical will automatically collapse into the Microphysicalist thesis that everything is microphysically determined.
However, there are alternatives to Pettit’s definition of ‘physical’ as microphysically determined. These will leave it open whether or not everything physical is microphysically determined, and therewith allow for versions of Physicalism that are not committed to Physical Microscopism.
For a start, there is the option of defining ‘physical’ negatively, as covering anything that can be directly identified without using some distinguished terminology. For example, we might count as ‘physical’ anything that can be directly identified using non-mental terminology. Or we might define it somewhat more restrictively, as anything that can be directly identified without using mental or biological terminology. This is the way of understanding ‘physical’ that I myself favour. In my book Thinking about Consciousness (2002) I argue for an understanding of ‘physical’ as inorganicallyidentifiable. The idea here is that we start with a distinguished inventory of mental and biological terms, and then pick out the physical realm as anything that can be directly identified without using those terms. (Note that the physical realm is here anything that can be so identified, not things that can only be so identified. Physicalists will of course hold that some parts of that physical realm can also be identified using mental or biological terms.)
Some philosophers favour a yet further option, one that takes off from Hempel’s dilemma. The idea here is to appeal to the categories represented by current Physics Departments, but to allow some wiggle room for future developments. So we might think of ‘physical’ as referring to all those categories that bear some resemblance to the categories recognized in contemporary Physics Departments. For example, ‘physical’ might be understood as equivalent to something like ‘displaying mathematically simple and precise behaviour’. I shall call this the ‘resemblance’ conception of ‘physical’ in what follows.
I shall not choose between these different understandings of ‘physical’ in this paper. It will be enough for my purposes to show that they allow various senses in which Physicalism might hold without Microphysicalism. But it will be useful to make one further point about the meaning of ‘physical’. Suppose we have fixed on one of the above definitions of ‘physical’. It will be convenient for the purposes of this paper to understand ‘physical’ recursively, in the sense of including any categories that supervene on the so-defined physical realm, even if they do not themselves fit the base definition. For example, suppose we equate ‘physical’ with ‘inorganically identifiable’. Then it may be that facts about insects supervene on the physical realm so-defined, but that there is no way of stating insect facts using inorganic terminology. (Suppose that insect facts are ‘multiply realized’ at the inorganic level, in a way that precludes any uniform inorganic specifications of such facts.) Even so, I will take the supervenience of the insect facts on the physical facts to qualify them as ‘physical’.
This recursive way of understanding of ‘physical’ would not necessarily be appropriate for all philosophical purposes. For instance, if our focus were on physical explanation, it would be confusing to hold that certain facts were physically explainable just because they could be explained in terms of entomological facts that supervene on physical facts, even though there was no question of specifying those entomological facts in physical terms. But our interest here is with ontology, not explanation, and in particular with which categories supervene on the physical facts and which do not. Given this, it will suit my expository needs to count anything in the former category as ‘physical’.
I turn now to the other possible reason for equating Physicalism with Microphysicalism, namely, the demands of providing an argument for Physicalism. Even if there are ways of understanding ‘physicalism’ that do not automatically collapse Physicalism into Microphysicalism, it could nevertheless be that the only way of arguing for Physicalism argues for Physical Microscopism too.
Thus consider this inductive argument: all facts so far subject to scientific scrutiny have turned out to supervene on the microphysical facts; so all the facts supervene on the microphysical facts. Some philosophers take this to be the primary rationale for embracing Physicalism. (Cf Rey 2002.) Now, if this kind of inductive argument were the only available argument for Physicalism, then clearly any justification of Physicalism would justify Microphysicalism too. Our rationale for thinking that all facts supervene on the physical facts would essentially depend on the lemma that they all supervene on the microphysical facts. So our rationale for Physicalism would endorse Physical Microscopism along the way.
However, the above inductive argument is not the only possible argument for Physicalism.4 There are alternatives that are quite free of any assumptions about microphysical goings-on. Thus consider the ‘causal argument’ that goes: prima facie non-physical facts like mental and biological facts have physical effects; all physical effects have physical causes (‘the causal completeness of the physical’); so those prima facie non-physical facts must supervene on physical facts (or we would have unacceptable overdetermination). This is the argument for Physicalism that I myself favour. As we shall see below, this argument need not commit us to any claim that all the physical facts supervene on the microphysical facts. The crucial premise—the causal completeness of the physical—need only claim that all physical effects have physical causes, not that they have microphysical causes. And then this argument will only commit us to the conclusion that prima facie non-physical facts must supervene on physical facts, not that they must supervene on microphysical facts. The causal argument will thus remain available even to those Physicalists, like myself, who wish to remain neutral on the issue of Physical Microscopism.5 4 Species of Emergence My aim is to show that we can deny Microphysicalism without denying Physicalism. That is, I want to show that Microphysicalism might fail, not because there are non-physical facts, but rather because some physical facts fail to supervene on the microphysical facts. In such a case, we would have a violation of Physical Microscopism, but not of Physicalism.
I won’t be concerned here to make a positive case for any such violations of Physical Microscopism. As I said at the beginning, my first commitment is to Physicalism, not to any views about microphysical determination. So my aim is only to establish conditional claims of the form: even if certain facts are emergent vis-à-vis the microphysical realm, Physicalism can still be true. I shan’t defend the antecedents of these conditionals. My interest in not in microphysical emergence as such, but rather in the fact that Physicalists don’t always need to reject microphysical emergence.
Of course, not all kinds of microphysical emergence are compatible with Physicalism. Cartesian dualism, for example, posits microphysically transcendent facts that would clearly violate Physicalism. This is because Cartesian minds would not only transcend the microphysical realm, but the physical realm too. To support my thesis, I need microphysically emergent facts that would remain genuinely physical.
Some of the microphysically emergent facts I consider below will fail to support my thesis. This is because it will prove difficult to avoid the conclusion that they would not count as physical. In the face of these particular species of microphysical emergence, Physicalists cannot of course stand neutral. They must reject any emergent facts that would transcend the physical realm, just as they must reject Cartesian minds. Fortunately, as we shall see, there are good arguments for denying those variants of microphysical emergence that would also transcend the physical realm.
5 Humean Supervenience Microphysicalists claim that all the facts, including the macrophysical facts, supervene on the microphysical facts. The strength of this claim depends on what gets included in the ‘the microphysical facts’. Austere understandings of the microphysical facts make for strong versions of Microphysicalism. Such strong versions will be comparatively easy to deny. By contrast, the more that gets included in ‘the microphysical facts’, the less easy it will be to show that there are facts that transcend the microphysical facts.
A particularly strong version of Microphysicalism would correspond to David Lewis’s doctrine of ‘Humean Supervenience’:
(HS) All the facts are metaphysically determined by the intrinsic properties of spacetime points plus the spatiotemporal relationships between those points.
This asserts that any world which agrees with the actual world on the ‘Humean mosiac’ of spacetime points and their intrinsic properties will contain all the facts that are present in the actual world. This is an extremely strong doctrine. It countenances no ‘external relations’ between spacetime points except their spatiotemporal relationships. Every other relational fact is fixed by the intrinsic properties of the points and the way these points are arranged in space and time.
Suppose we agree that the intrinsic properties of spacetime points are all physical properties. Humean Supervenience will then amount to a very strong form of Microphysicalism. Because it is so strong, it is easy for it to be false. In particular, it will be false if a non-Humean view of laws is true. The Humean view is that laws depend on nothing more than the ‘constant conjunctions’ of particular facts displayed by the actual world. So any view on which laws transcend such facts of constant conjunction will contradict Humean Supervenience. Any such view implies that a world can agree with this world on the Humean mosaic yet differ on the laws.
I take this to illustrate a minimal sense in which one can be a Physicalist while rejecting Microphysicalism. If we equate Microphysicalism with Humean Supervenience, then anybody who rejects a Humean views of laws will be rejecting Microphysicalism. But nobody, I take it, would want to argue that a non-Humean view of laws amounts to a violation of Physicalism. This would only follow if non-Humean laws must in some sense themselves be non-physical, and there seems no reason to hold this. Certainly many actual Physicalists embrace this kind of non-Humeanism about laws without feeling that it somehow undermines their Physicalism.
Still, I don’t suppose that this point will worry any of the philosophers who think that Physicalism requires Microphysicalism. This is because they are unlikely to understand Microphysicalism as making the extreme claims of Humean Supervenience, and in particular as requiring a Humean view of laws. Just as Physicalists in general will say there is nothing non-physical about non-Humean laws, so those who equate Physicalism with Microphysicalism are likely to say that there is nothing non-Microphysical about non-Humean laws either. They will thus be happy to add
non-Humean laws to Lewis’s Microphysicalist supervenience base, and thereby weaken the relevant supervenience doctrine: to fix all the facts, it is not enough just to fix the intrinsic properties and spatiotemporal arrangements of spacetime points—we must also fix the laws that govern the causal interactions between those points. These laws themselves need not supervene on the properties and arrangements of spacetime points.
This doesn’t mean that those who want to equate Physicalism with Microphysicalism will place no restrictions at all on the laws present in a given world. They will typically insist that the only basic laws are microphysical laws. There may be genuine macroscopic laws, but if so they will be derived from the microscopic laws. As Pettit puts it, ‘. . . once the microphysical conditions and the microphysical laws have been fixed, then all the crucial features of a world like ours will have been fixed; viz., all the other laws that obtain at the world . . .’ (1993 p 219). From this point of view, while we might have to add non-Humean laws to get an adequate Microphysicalist supervenience base for all facts, it will be enough to add microphysical non-Humean laws. There are no further laws that are not determined by microphysical laws plus arrangements of microphysical initial conditions. So now we have another Microphysicalist supervenience thesis, one that places restrictions specifically on laws.
(L) All the laws are metaphysically determined by microphysical laws and microphysical initial conditions.
6 Broad-Style Emergent Laws I now want to consider whether a Physicalist can deny (L) and yet remain a Physicalist. That is, would the existence of macroscopic laws that are not dependent on microphysical laws and initial conditions somehow contradict Physicalism?
This will prove a less than straightforward matter. In this section I shall argue that there is no immediate reason why Physicalists should not countenance macroscopic laws that do not depend on microscopic ones. However, the situation is complicated by considerations to do with force fields. I shall consider these complications in the next section.
A first question to address is what exactly qualifies a law as microphysical. We can take a microphysical law to be one that applies inter alia to small physical systems. (We needn’t worry about what precisely qualifies a physical system as ‘small’—the issues will come out the same wherever we draw this line.)
Note that there is nothing in this definition of a microphysical law to require that it applies only to small physical systems. It may be that microphysical laws are formulated in such a way that they apply uniformly to both small and large physical systems.