Properties supervenient

However, the split between monistic and dualistic systems of belief often borders along the concept of supervenience (characterization of a relation that emergent properties bear to their base properties), see also Ref. [6] for a modern appraisal. To use philosophical relationship terminology if one states that biology (naturally) supervenes on physics, i.e., when physical (spatiotemporal) facts about the world determine biological facts, it is nevertheless called into question whether biology logically supervenes... 

Third, many more than two discourses (apart from chemical and physical) need to be distinguished there is manifest water, the physical, physico-chemical, chemical, and biochemical macroproperties of water, the (chemical) molecular microdiscourse, and the (physical) quantum mechanical discourse, and that s only the beginning. Fourth, the notion of reduction has to be spelled out (a) What is it that is being reduced (theories, concepts, properties, natural kinds, laws, explanations) (b) How does it differ from related notions such as replacement, elimination, integration, supervenience, and emergence (c) What sort of relation is it (dependence, identity,. . . ) ... 

It has been argued, though, that supervenience is not itself a dependence relation, but only a modalized covariance of properties. On this view, it is at best the sign of ontological dependence, related to it as correlation is to causation, see Horgan 1993. 

Although there is notoriously no unanimous agreement on what the supervenience relationship is, the most popular view is that supervenience is a relationship of asymmetric dependence. Two macroscopic systems which have been constructed from identical microscopic components are assumed to show identical macroscopic properties, whereas the observation of identical macroscopic properties in any two systems need not necessarily imply identity at the microscopic level. In simpler terms, the phenomena we study in some secondary science are thought to be ontologically dependent upon relationships at the primary level whereas macroscopic identity need not imply microscopic identity. 

As an example to contextualise the supervenience argument, Scerri McIntyre (1997) consider the property of smell. If two chemical compounds were synthesised out of elementary particles in an identical manner, they would share the same smell. Similarly the supervenience argument would entail that if two compounds share the same macroscopic property of smell, we could not necessarily infer that the microscopic components from which the compounds are formed would be identical. Such scenarios can be explored by biochemists and neurophysiologists but whatever the outcome, the question of supervenience of chemistry on physics will depend on empirical facts and not on philosophical considerations. 

The case of supervenience highlights the role of empirical chemical research in establishing at least some aspects of the relation between microscopic and macroscopic systems. One educational implication would then involve an emphasis on the significant role of empirical research in chemical inquiry. As an example educational scenario, the question of supervenience can be raised at secondary education through case studies investigating the relationships between the colour, smell and texture, and microscopic properties such as molecular structure and bonding. 

What I want to focus on here is Davidson s argument that his view does not lead to the epiphenomenalism of mental properties. Davidson seems to argue here that the supervenience of mental properties on the physical will ensure their causal efficacy. For example, Davidson says ... 

If supervenience holds, psychological properties make a difference to the causal relations of an event, for they matter to the physical properties, and the physical properties matter to causal relations. It does nothing to undermine this argument to say But the mental properties make a difference not as mental but only because they make a difference to the physical properties . Either they make a difference or they don t if supervenience is true, they do (1993 14). 

Davidson, I think, is led to this particular line of defense because of his views about causation. Davidson holds an extensionalist view of causation - i.e., that causation is a relation between non-abstract particular events. Thus he says, But it is also irrelevant to the causal efficacy of physical events that they can be described in the physical vocabulary. It is events that have the power to change things, not our various ways of describing them (Davidson 1993 12). Since it is events that are causes, then any properties that are relevant to what a particular event causes are causal properties. Davidson says, But properties are causally efficacious if they make a difference to what individual events cause, and supervenience ensures that mental properties do make a difference to what mental events cause (1993 15). So since mental properties supervene on physical properties, and physical properties make a difference to what an event causes, mental properties also make a causal difference. 

Now there are other physical properties this ball has, which are epiphenomenal with respect to the ball landing on the bottom. The fact that the ball is blue (as opposed to red) or plastic (as opposed to wood) has no causal bearing on the fact that it sorts to the bottom. Yet these properties are physical properties, and they supervene on the local microstructure of the b lll. These are both properties we could cite as properties of event A (my shaking the sorter with the blue plastic ball) and of event B (the blue plastic ball sorting to the bottom). The fact that a property is physical, or supervenient, or is a property had by the event that is a cause does not mean that this property is a causal property. 

Thus, Davidson s claim that supervenience saves mental causation falls short. Davidson is wrong to think that the question, hi virtue of which properties did event A cause event B makes no sense. To save AM fi om the epiphenomenalism of mental properties, we need a way to show how it is possible for causation to be in virtue of mental properties. The question is whether the anomalism of the mental will make it impossible for mental properties to be the one s in virtue of which events are causes. 

Let us now take a preliminary look at how to make sense of causation in virtue of irreducible mental properties. If we are going to assume, for the purposes of this discussion, that Davidson has not quite done enough to get property causation, then we need to find a different way to make sense of property causation. Before we can do this, we need to make sure that this project is even possible. If strict laws will rule out causation in virtue of irreducible mental properties, then it is hard to see how any approach other than Davidson s appeal to supervenience and non-strict laws could possibly save property causation. If, however, strict laws do not rule out causation in virtue of irreducible mental properties, then we will be in a position to try to find out how irreducible mental properties can be causal. 

This type of objection to nonreductive materialism is central to the work of Jaegwon Kim. Kim discusses this objection extensively in the various versions of his exclusion argument or supervenience argument. He holds that the only way to give mental properties a causal role is to functionally reduce them to physical properties, since physical causation will always rule out causation in virtue of irreducible mental... 

If, however, we reduce mental properties to physical properties, then mental properties will simply inherit the causal powers of their supervenience bases and will thus be causal in this derivative sense. 

This principle could be taken to mean that every good causal story we can tell has to be couched in physical terms or that all causation is always in virtue of physically definable or reducible properties. Taken this way, however, this principle would just beg the question against nonreductive materialism. Under physical causal closure, it is not that we can never invoke physically irreducible higher-level properties as causes for a physical event. We may very well cite some irreducible mental property as the cause of some physical event. The point is that this mental property, if it is to be causal, must somehow be anchored in the physical world. It must be tied to the physical by some sort of supervenience or realization relation. This close tie to the physical will always make it possible, in principle, for us to tell a causal story in terms of physical properties, even though the best causal story might not always refer only to physical properties. Thus, we will never be forced to go outside the physical domain and refer to physically irreducible properties in order to find a sufficient cause for a physical event. So the principle of physical causal closure requires that if we pick any physical event, it will always be possible, in principle, for us to find a purely physical causal chain for that event. So if we take any physical event P, we should be able to find some property or group of properties P, such that P is physical and P is a sufficient cause for P. ... 

The second assumption that leads to the exclusion problem is the idea that any physicalist view is committed to some sort of notion of mind-body supervenience. There are many different notions of supervenience, but at minimum, a physicalist must hold that every mental property has some sort of physical base which instantiates it. The mental depends, for its existence, on some sort of underlying physical state and does not constitute its own ontologically independent domain (Kim 1998 41). Thus, if we take any mental property M, it will have some underlying set of physical properties P on which it supervenes. ... 

Kim formulates what he takes mind-body supervenience to entail as follows Mind-body supervenience Mental properties supervene on physical properties in the sense that if something instantiates any mental property M at t, there is a physical base property P such that the thing has P at t, and necessarily anything with P at a time has M at that time (1998 39). ... 

Notice that if we accept this notion of mind-body supervenience, then the existence of the base property P will always be sufficient to bring about the supervenient property M. 

This notion of mind-body supervenience, however, seems a little stronger than what is necessary for physicalism, if we take P to be only local, non-relational properties of tiie... 

At the very least, M and P must be related by supervenience. However, it may be that mere supervenience is not enough to give us an adequate accoimt of mental causation within a physicalist framework. We might want to hold a view like Davidson s, where not only does M supervene on P, but P realizes M, and M and P are properties of a single event. So as properties, M and P are distinct, but as event, M and P are token identical. Whether physicalism requires mere supervenience, a picture like Davidson s with realization and event identity, a functional reduction of M to P as Kim advocates, or a fullblown type identity of M and P will be discussed shortly. 

Notice that this only holds true on Kim s notion of supervenience, where P is limited to local physical properties. On a weaker notion of supervenience, where P can include context-dependent properties, M s physical supervenience base would be sufficient for the occurrence of M. In this case, P would be sufficient for the occurrence of M, but P would be understood as M s local supervenience base in the proper context. 

This is actually not quite right. Because M is multiply realizable, M could have occurred even if P had not occurred in the case where some physical property other than P instantiated M. So M could have had a different supervenience base other than P, say P2. The point is that M needs some sort of physical supervenience base to instantiate it and cannot occur independently of its supervenience base. Whatever supervenience base does instantiate M, it will be related to M by supervenience, and it and M will not be independent of one another. 

But recall that the exclusion principle is not restricted to simply mental properties. The principle holds that no event can have two distinct properties unrelated by reduction as causes, if one of the properties is a sufficient cause. But the ball sorter example is a clear case where we have two distinct causes, unrelated by reduction. The lower-level property is a sufficient cause, yet the higher-level property is clearly a causal property as well. If the revised exclusion principle were true, it would rule out not just the causal efficacy of mental properties, it would also force us to hold that the property of size cannot be causally efficacious with respect to the balls sorting. In fact, we would have to say that no distinct and irreducible higher-level supervenient property is ever causal, and that all real causation always happens at the lower level. This is an unpleasant result and not one that we should be willing to accept. Thus, we ought to reject the exclusion principle on these grounds. 

I will soon discuss in greater detail what it means for a level to be causally closed. For now, I will just say that if a level is causally closed, then any event that occurs at that level will have a sufficient cause that refers only to properties at that level. Now let us assume that L - 1 is not a causally closed level. That means that for any event that occurs at level L - 1, we may sometimes have to look to properties at levels above or below L -1 to find a sufficient cause for that event. Now say that L is supervenient on and irreducible to L - 1, and L -1 is closed with respect to level L. This means that for any event with a cause that occurs at L, we will always be able to find a sufficient cause for that event at level L - 1. If we were to look at an event and its cause, both of which had L and L - 1 properties, we would have to apply the exclusion principle. Since properties at L are not reducible to properties at L - 1, the exclusion principle requires that we rule out either the property at L or the property at L - 1 as causal. If L -1 is closed with respect to L, then this event with L properties will have a sufficient cause at level L - 1. That means we will be forced to choose the L - 1 property as causal. This property will preempt the causation of its supervenient property at level L. 

So it seems that in this scenario, where we have closure at L - 1 with respect to L, the exclusion principle and seepage will both apply. The causal powers of properties at L will always be preemped by properties at L - 1. Causation need not seep down to a lowest closed level for the exclusion argument to work. Instead, it seems that if the exclusion principle applies to all irreducible supervenient levels, then whenever there is closure of a subvenient level with respect to its supervenient level, causation will seep... 

---