Functionalism, Qualia and Causation

Abstract

from the pragmatics of the situation for a moment and proceed thus. Let c be the cause of e. Then a causally efficacious feature, F, of c is one such that if c had not been F then c would not have caused e. (This is not circular-we are only trying to define causal efficacity, and it is not surprising that the notion of causation itself is useful in this task. This definition means we can neglect worrying about situations where, by accident, an event like e occurs even though c is not F.) Pragmatically speaking, the causal efficacy of c is that causally efficacious feature which makes the cause-effect relationship, between c and e clear to my questioner. Naturally, causally efficacious features are ones that figure in causal laws. Thus to give a true causal explanation is to locate the causal efficacy of a certain event. We may also and no doubt often do give false causal explanations which nonetheless satisfy our questioner. This will occur when the feature accepted as causally efficacious is in fact not a feature which figures in the correct causal law relating the events at issue. To make, however, a location of causal efficacy is to claim that a certain feature will or does figure in This way of looking at things was suggested to me by a paper of Peter Hess, 'Actions, Reasons and Humean Causes' in Analysis, June, I98I. This content downloaded from 157.55.39.35 on Wed, 31 Aug 2016 04:34:36 UTC All use subject to http://about.jstor.org/terms FUNCTIONALISM, QUALIA AND CAUSATION I79 such a law. Note that in our home run example, 'swinging the bat' would certainly seem to be a causally efficacious feature of the event which caused the home run but it is hard to imagine a situation where one could locate the causal efficacy of the event in this feature of it. That the swinging of the bat occurred at 9: i9 p.m. does not seem to be a causally efficacious feature. I have gone to these lengths in the discussion of causal explanation in order to contrast it with another way of explaining the occurrence of an event. Suppose you ask why the numeral 'i' was printed and I reply, 'Because you struck the button marked i '. Is this a causal explanation? I think that normally it is not. There is no causal law which relates the pushing of i-buttons with the subsequent printing of a 'I' (as the existence of elevators attests). One might object here that I am ignoring my own remarks made just above, and that further conditions must be imposed here, or further specifications of the questioner's background knowledge be given, in order to develop a causal explanation. There is indeed something to this objection and we will see just what below. For the moment, consider another direction of clarification we might take. I say instead in reply, 'Because you pushed the i -button and this is a typewriter'. This is still no causal explanation-there are broken typewriters. What is more important, a working or functioning typewriter must print a 'I' if the i -button is depressed. And when I tell you it is a typewriter it is clear that I mean you to understand that it is a properly functioning typewriter. Let us call this form of explanation functional explanation. One functionally explains a certain event by claiming that the event is the output of a system which is a certain functional system, that is, which is a realization of a certain abstract functional systemdescription. Functional explanations are helpful since one may not know that a given system meets any particular functional description. Functional explanations are thus characterized by the noncontingency of the link between explanans and explanandum. We can illustrate these points by considering the widget. The widget, I say, is any device which meets the following functional description: