#3 - A Problem for Free Will - Supervenience
§3. Supervenience and the Anomalousness of the Mental:
3.1 Incompatibility between supervenience and anomalousness of the mental:
We saw in §2.4 how Davidson argued for ‘nomological slack’ relating the mental and the physical. This allowed him to claim there are no strict laws on the basis of which we can predict and explain mental phenomena. None of the objections raised there were taken to be fatal to his claim for the anomalousness of the mental. However his supervenience thesis seems to represent a tightening of the relation between the mental and the physical suggesting a lawlike relation. To see why we need to consider an argument provided by Jaegwon Kim.
Kim writes; ‘If M supervenes on N [M and N are classes of properties], each property in M which is instantiated has a general sufficient condition in N … I don’t see how such generalizations could fail to be lawlike’. Kim’s claim is then that if someone bears a mental property M, and the mental supervenes on the physical, then, if that person’s physical state is P, supervenience will entail that anyone who is P has M. Kim does not have in mind here a mere psychophysical generalisation of the form (G) in §2.4 but something much stronger; a psychophysical law of the form:
(S) For any x, if x has the physical property P, then x has the mental property M.
A law like (S) does not amount to the reduction of mental properties to physical properties since the implication only runs from the possession of P to the possession of M. Reduction would entail a law like
(S') For any x, if x has the mental property M, then x has the physical property P.
Supervenience does not licence claims like (S'), so it could not be claimed supervenience led to property reduction. Non-reductive as (S) is, Davidson’s supervenience thesis does nonetheless seem to represent a psychophysical law that conflicts with the anomalousness of the mental. Supervenience might not licence the unrestricted doctrine of psychophysical correlations but it does not refute a limited correlation thesis that asserts the existence of a physical correlate for every mental event. But whether we should call generalisations like (S) laws or not it still seems to be the case that the very arguments Davidson uses against the possibility of psychophysical laws tell equally against his supervenience claims. This should be worrying for Davidson. If the attribution of mental states is to retain allegiance to its proper source of evidence; if it is to be governed by the normative principles of coherence, consistency and rationality, then which physical properties some object instantiates should be irrelevant to which mental properties it has. Consider the case of two people alike in all physical respects, one of whom has the mental property M. If we were to ask whether the other person also has M why should we let their physical similarity to the first person be a consideration at all? It might be the case that the principles of rationality which Davidson takes to govern the attribution of mental content, taken in conjunction with the other mental states already attributed to him, require us to deny the other person has M.
It seems, then, that quite generally that supervenience is incompatible with the thesis of the anomalousness of the mental. Davidson is of course right to say that laws connecting two sets of predicates would serve as a conduit through which the principles distinctive of one set are transmitted to the other set. He is also right to say that if such laws existed the possibility will then arise that the physical states of some person will pre-empt the attribution of mental states. However, given Kim’s argument there is some reason to think that Davidson’s supervenience claims would be psychophysical laws. But even if they are not laws they would still conflict with the autonomy of rationality constraints in governing the attribution of mental states.
There are further challenges confronting the doctrine of supervenience. Suppose the mental predicate ‘x fears that the world will be destroyed by an asteroid’ bears a relation to some set of physical predicates, S. If the mental predicate is true of one person and not of the other, then supervenience says that there will be some predicate in S that is true of one and not the other. The question is, which predicates might belong to S? A plausible response would be to say all the predicates that are true of the one who fears the end of the world. However, this means that supervenience will have to countenance the possibility of two people, one of whom feared the end of the world and one who did not, who are exact physical replicas except that one had a single eyelash longer than the other. These individuals, who differed in that one was fearful and one wasn’t, would satisfy the requirement that there was some predicate in the supervenience base that was true of one and not the other. Surely Davidson cannot countenance the possibility such a physical difference could account for their differing mental states.
By narrowing the supervenience base Davidson would avoid such absurdities. But what should and should not be included among the physical predicates on which the mental predicates supervene? It could perhaps be argued that mental predicates are supervenient on the physical predicates true of people’s brains. However, we could repeat the above example but limit the physical differences of otherwise identical people to some inert part of the brain, but in so doing we are in no better position than before. In fact what Davidson really wants from supervenience is the dependence of mental properties on just those physical properties of a person which are causally relevant to the manifestation of the mental property. The differences suggested by the above examples simply do not count because they play no causal role in the bringing about of other mental states and actions. But it is not at all clear that Davidson can legitimately have what he wants. Meeting this requirement would include those predicates that are potentially able to cause those actions and other mental states with which the given mental state might, given his normative and holistic considerations, be rationally connected. However, given he denies there are psychological laws under which mental states fall when described as mental states, we cannot say which other mental states, and which actions, they might cause. There is no way of saying which causal powers the underlying physical properties must have, and so no way of specifying which physical predicates should form the supervenience base for some mental predicate.
Davidson seems to have a dilemma. If he cedes that the appropriate physical predicates forming the supervenience base can be specified then he, contrary to the tentative findings in §2.4, repudiates the anomalousness of the mental. He can avoid this first horn only by allowing all-manner of causally irrelevant properties to be included in the supervenience base. In which case he would have to accept two otherwise identical people, differing only in, say, the length of a single eyelash, could have different mental content. This seems absurd.
However, Davidson’s position on supervenience has not remained static. Indeed it is no longer clear that he holds to the doctrine of supervenience thus far considered. When he first discussed supervenience in ‘Mental Events’ he said, ‘there cannot be two events alike in all physical respects but differing in some mental respect’ . However, in a later paper he seems to have changed his position. He writes, ‘we are … free to hold that people can be in all relevant physical respects identical while differing psychologically: this is in fact the position of “anomalous monism”’. We might concur that this is the position of anomalous monism, but it is not the position of anomalous monism as supplemented by supervenience. This does not change despite his modified doctrine of supervenience made in response to more recent discussions of propositional content, in particular twin-earth examples as offered by Putnam.
Take two worlds which are exact physical replicas of each other, differing only in the fact that on earth what we call H20 is on twin-earth some marcoscopically indiscernible substance XYZ. If I have some belief which I would express as ‘water is refreshing’, my twin-earth counterpart would have a belief which he would express in the same way. Yet although our brains are, by hypothesis, in type-identical states we have different beliefs because of the nature of the stuff our respective beliefs are about. My belief would be about water and his would be about XYZ. On an externalist view what my sentence means is not determined just by what is going on inside my head, but is affected by the stuff outside of me, the things that my beliefs are about. Given that Davidson also takes an externalist view of content he can no longer hold to supervenience in the way originally conceived.
However, even though he is forced to change his position on supervenience the problems it potentially raises are not serious, at least not as serious as that of the incompatibility between supervenience and the anomalousness of the mental. Indeed Davidson can accommodate his externalism. Even if he gives up the claim that mental states are supervenient on the physical states of the person whose mental states they are, they may be accommodated. He need simply extend the supervenience base to include not only predicates applying to the person, but also to predicates which qualify the physical environment. Such machinations do not however resolve the conflict between supervenience and the anomalousness of the mental. The problem persists even with a supervenience base so enlarged.
The extent of Davidson’s dilemma becomes apparent if we step back to §2.5 where we asked why Davidson held the supervenience thesis. His reason was because without it the mental is causally irrelevant. But it seems from the arguments presented that to reinstate the causal relevance of the mental while at the same time holding onto the anomalousness of the mental is impossible. As Honderich has pointed out; Davidson can insist that mental states are causally relevant and efficacious of physical events but this, when added to the principle of the nomological character of causally relevant properties, entails there are psychophysical laws and therefore the mental is not anomalous. Alternatively, he can insist on holding on to the anomalousness of the mental thesis but this, when taken with the principle of the nomological character of causally relevant properties, entails the mental is not a causally relevant property. In other words, the mental is epiphenomenal in that an intention to act becomes causally irrelevant to the action. Either option is fatal to establishing free will as I will now show.
Blog #4 - The Problems for Free Will