Supervenience is a way of describing the relationship between two sets of properties. Consider property set A and property set B. Saying that A is supervenient with respect to B is saying that for there to be a difference in A properties we must have a difference in B properties. A corollary to this is that if all B properties were found to be identical, we would find all A properties to be identical. An anti-corollary is that if all A properties were found to be identical, this does not imply that all B properties are necessarily identical.
Consider a concrete example: Numbers are properties of sets. We can have 10 pigeons or 10 marbles. The numerical properties (i.e. “how many there are in the set”) are supervenient upon the constitutive properties (i.e. “what the set is made of”). A properties (numerical properties) are supervenient upon B properties (constitutive properties). Thus, we cannot have a change in the numerical properties of either set without changing the number of pigeons or marbles (“A differences imply B differences”). If two sets (both composed of either pigeons or of marbles) have the same number of pigeons or marbles, then they will have the same numerical properties (“B identicality implies A identicality.”). However, for two other sets, one composed of 10 pigeons and one of 10 marbles, even though they have the same number of marbles as pigeons (“A identicality”), this does not mean that those numbers are the product of the same set of things (“B properties not identical even though A properties are identical”).
Another example: liquidity as a property of matter. Water at room temp is liquid, which means that if you shake your water bottle, the water sloshes around inside. Liquidity is supervenient with regard to the micro-molecular structure of H2O at room temperature. To change liquid behavior, we must change the micro molecular structure. If the micro molecular structure is sufficient to produce liquid behavior, then liquid behavior will be observed. However, liquid behavior is certainly not only seen in bottles of water at room temperature. It can also be seen in molten metal or glass in a kiln, for example. Thus, A properties are determined by B properties but are not absolutely and essentially reducible to B properties.
This model is useful for explaining the relationship of so called subjective experiences (“mind”) to so called objective experiences (“body”). It allows us to describe a relationship between mind and body that avoids both the problems of property dualism and the problems of epiphenomenalism. Property dualism is avoided because the mind is not defined as something separate from the body, any more than a number is something separate from a set or liquidity is something separate from water sloshing around in a bottle. Epiphenomenalism is avoided because the mind is not “reduced to” the body. Supervenience describes a real relationship between sets of properties that, though it describes a type of dependence of one upon the other, a real causal relationship, it does not make one set of properties infinitely determined by another.
Some may be tempted to accuse me of re-packaging property dualism and supporting a computer functional model of the mind, since this notion seems to imply that a mind in principle could be realized in another arrangement of micro molecular structure, just as liquidity can be realized in multiple forms. For example, we might conclude that a computer that is smart enough to beat a human in a game of jeopardy qualifies for just such “mind status” designation. The question of the multiple realizability of minds however, is a question not of what is theoretically possible but one of actual accomplishment. For example, I have never met a computer that has emotions or theory of mind, two essential properties of minds. And according to the theory of supervenience, if the A properties are different (in this case emotions and ToM as properties of “minds”) then the B properties must be different. Thus, if a computer does not exhibit properties that we consider essential to minds, it cannot be a mind. This does not, however, imply that someday we won’t be able to create a mind out of silicon chips, just that the arrangements of silicon chips we have created so far are not, in fact, minds, though they may behave at times in ways that make us think otherwise.
Leave a Reply