Printer Friendly

Dretske on metaphysics and freedom.

According to Fred Dretske (1988, 1992) an action is a complex event which consists of an internal and an external part. To put it roughly, the internal part is the reason which explains the action and the external part is a bodily movement caused by the reason. When a person acts her bodily movements are caused by the reasons which explain her action. Dretske calls this kind of analysis the component theory of action. Actions are causal processes or causings. Thus, Dretske would say that
   My raising my arm = My reason for raising my arm causing
   my arm's going up.

In Dretske's view the action of my raising my arm is not identical with my arm's going up but has it as a component. If I raise my arm in order to vote, then my action of raising my arm is my desire to vote and my belief that by raising my arm I vote together causing my arm's going up.

Dretske's view on the ontology of action differs significantly from what he (1992:2) calls the orthodox view. This view whose best-known proponent is Donald Davidson (1963) says that actions are those bodily movements that are caused, in the right way, by the reasons for acting. Thus, if my arm goes up because of my desire to vote and my belief that by raising my arm I vote, then my arm's going up is an action of mine. In the orthodox view, then, the following holds:
   My raising my arm = My arm's going up when my arm's going up is
   caused in the right way by the reasons which explain my arms's
   going up.

In this orthodox view, actions are identical with bodily events in the sense that each action is identical with some bodily movement. But, of course, the converse does not hold. It is not the case that every bodily movement is identical with some action.

In the orthodox view it is a conceptual truth that reasons cause actions--nothing counts as an action if it is not caused by the reasons that explain it. According to Dretske, this view threatens our status as autonomous agents. He (1992:3) writes:
   If an intentional act--say, wiggling one's ears in order to earn a
   dollar--is caused by one's beliefs and desires (the reasons one has
   for wiggling one's ears), then, by the transitivity of the causal
   relation, it appears to follow that it is (also) caused by whatever
   causes one to have those beliefs and desires. But the causes of
   belief and desire are often (in fact, if we trace the causal chain
   far enough backward, always) factors over which one has no control.
   So, intentional behavior is often (or always) something one is made
   (caused) to do by factors over which one has no control. This,
   however, robs intentional behavior--and, presumably, also voluntary
   action--of its autonomy. Deliberate acts--Jimmy wiggling his ears
   to earn a dollar--have the same causal structure as does a bell
   that rings because a button is pushed. The only difference is the

Dretske's worry is that if somebody, she, gives a reason for somebody, him, to do an action, then she makes him to perform that action. She causes the reason and the reason causes the action. But by transitivity of the causal relation she causes his action. It would indeed be problematic if by providing somebody with reasons for acting, no matter how good, we would make the actions performed from those reasons unautonomous. What Dretske is concerned to show is that his theory of action which is in accordance with universal causation does not entail such absurdity. Dretske's aim is, then, not to give a full defence of compatibilism but to make a more modest point, i.e., that the fact that a reason for which the agent acts is ultimately caused by things under the control of the agent does not make the act unfree or unautonomous. So what Dretske wants to show is that in his component theory of action reasons do not cause the actions they explain even though they cause the events associated with those actions.


In Dretske's (1988:38, 1992:8-9) component theory of action, reasons are parts of actions and an action is the reason causing the event associated with the action. But now, Dretske claims that if the reason for which an agent performs an action were also its cause, then a part of the action should be a cause of the whole action. But Dretske (1988:38) thinks that it is false that a part of a whole could causally contribute to the existence of the whole. Thus, the following principle should be accepted:
   (P) If c and e are events and c causes e, then it is not true that
   c causes c's causing e.

No doubt, (P) has some intuitive plausibility. Causings are complex events which have reasons as constituents and it seems natural to say that a constituent of a whole cannot cause the whole of which it is a constituent. In defending his view, Dretske (1988:38) cites Thalberg (1977:74):
   Not only does the whole (the process) not cause its parts (the
   product); the parts don't cause the whole. As Thalberg puts it, a
   whole action cannot result from one of its ingredient events.

Also Terence Horgan accepts the Thalberg-Dretske view. In commenting Dretske's theory of action, Horgan (1991:85) writes:
   Since a causal process cannot be caused by a part of itself,
   [Dretske] must maintain that reasons do not in fact cause the
   actions they rationalize.

Dretske, Horgan and Thalberg all seem to accept (P) almost as a self-evident truth. (1) In the next section, my aim is to show that there are good reasons for rejecting (P).


Roderick Chisholm does not join the chorus of accepting (P). Chisholm (1979:371) who also adopts a component theory of action believes that the following principle holds:
   (P1) If c contributes causally to e, then c contributes causally to
   c's contributing causally to e.

For Chisholm, this principle is important because according to him the parts of an action include an undertaking, which is much like an act of the will, and agency is explained in terms of this undertaking. In Chisholm's (1976: chapter II) theory a person is the agent of those things that her undertakings cause and if her undertakings did not cause her actions, then she would not be the agent of her actions and that would be absurd. According to Chisholm, then, a part of an action may be a cause of the whole of which it is a part.

This is a rather interesting case of a conflict of intuitions. I believe that this conflict shows that Chisholm on the one hand and Dretske (and Thalberg and Horgan) on the other seem to have different concepts of causation in mind. In what follows I try to explain these sorts of causation.

Thalberg presents his denial of part-whole causation (i.e., part as the cause and whole as the effect) by saying that a whole action cannot result from one of its ingredient events. This, I believe is true but to say that there is part-whole causation is not necessarily to say that the whole results from one of its parts. To infer 'x results from y' from 'y causes x' is to accept a concept of causation that was widely held before Hume. For example, Descartes, Spinoza and Leibniz thought that a cause must come from its effect. The effect, these philosophers think, must get all its reality from the cause and in a sense is involved in the cause. "For where, I ask, could the effect get its reality from, if not from the cause" writes Descartes in The Meditations (1984:28). I call this concept of causation as the generation view of causation (g-causation, for short) because in this view effects are, as it were, made from the material involved in the causes. (2)

In g-causation the denial seems plausible. If a part g-caused its whole, then this whole should somehow come from its part and this would mean that this part should give birth to itself and this sounds impossible. We cannot, for example, say that a leg of the table g-causes the existence of the table because this would mean that the whole table, the leg included, is generated by the leg.

But is the denial of part whole causation plausible in more recent analyses of causation? Let us consider first a simplified version of the regularity analysis of causation. In the spirit of regularity theory we could make the following definitions:

(i) c causes e as follows iff c is an indispensable part of a sufficient causal condition d of e;


(ii) d is a sufficient causal condition for e iff it is a law of nature that d is followed by e.

It seems evident that also in the regularity theory a leg of the table cannot be viewed as a cause of the whole table because there are no laws connecting legs of the tables with tables. Legs may be considered as necessary for the existence of certain kind of tables but this kind of relation is not a causal relation. Moreover, tables and legs as continuants are not suitable relata for causal relations because it is widely believed that causality holds between events which have temporal parts.

However, actions are events. In Dretske's analysis they are complex events which have other events as their parts, and it turns out that it is not at all clear whether regularity theory denies that parts of events cannot causally contribute to the existence of the events they are parts of. In fact, I believe this to be true when causings are at issue, i.e. events whose temporal parts are related to each other through causality so that a prior part of that event is the cause of the next part.

Let us consider causes of causings in regularity theory. Suppose that c causes e. Now, c is an indispensable part of a sufficient causal condition for e which sufficient condition is connected to e via a law of nature. But this sufficient causal condition of which c is a part must also be a sufficient causal condition for c's causing e, and c has to cause c's causing e. This can be seen as follows. Suppose d is a sufficient causal condition of e and that c is a part of d. Now, when d occurs, c causes e. So, when d occurs c's causing e occurs which means that the occurrence of d is sufficient for the occurrence of c's causing e. The same law or laws that connect d and e connect also d and c's causing e. But because the parts of a sufficient causal condition for an event are causes of that event, all parts of d are causes of c's causing e. However, c is a part of d, and so c causes c's causing e. In causings, part-whole causation is, then, in accordance with the regularity analysis of causation.

If David Lewis's (1972) counterfactual analysis of causation is accepted the result seems to be similar. In this analysis 'c causes e', if the counterfactual 'if c had not occurred then e had not occurred' is true. But in Dretske's theory, the reason for action is necessary for the occurrence of the action, i.e. without the reason which explains the action, the action would not have occurred. Thus, also in Lewis's analysis of causation the action which is a causing is caused by the reason which is a part of that causing.

Maybe some feel that what has been said faces the following objection. The acceptance of part-whole causation in causings allows the possibility of a conceptual connection between the effect and its cause. One might even say that in this special case one cannot think of the effect without thinking of its cause. But it has been claimed (for example, von Wright 1971) that causes and effects cannot be conceptually connected to each other. It is not quite clear to me why causal relations should meet this constraint. Of course, to exclude cases of self-causation, it is natural to hold that cause and effect have to be distinct entities. Moreover, it is natural to think that a continuant, such as a table, cannot be the cause of its component parts. However, it might be claimed that the existence of the components of a continuant do follow from the existence of the continuant. For example, according to mereological essentialism, if x is a part of y, then x is necessarily a part of y. Thus, in a certain sense the parts of a continuant can be seen to follow from the whole. However, it seems clear that parts of a continuant do not causally follow from the continuant. But, of course, the part-whole causation in causings is not an instance of self-causation or an instance of a continuant having necessarily its parts.

In fact, it seems that it is not difficult to give examples of causes and effects being conceptually connected to each other. Suppose I tear a sheet of paper in two pieces. It is arguably the case that it is my tearing the paper in two pieces which causes the paper to be torn. However, my tearing the paper is conceptually connected to the paper being torn. Dretske would say that in his component theory of action, the tearing does not cause the paper to be torn. But this sounds intuitively implausible. If somebody asks why the paper is in two pieces, the answer "Somebody tore it" seems to identify a cause of that state of affairs.

I believe, then, that Dretske is not right in his denial of part-whole causation. So also his claim that reasons do not cause the actions they explain is false even in his component theory of action.


I would like to thank a referee of this journal for helpful comments.


Chisholm, Roderick M. (1979) "Objects and persons: revisions and replies". In Essays on the philosophy of Roderick Chisholm, 317-388. Ernest Sosa, ed. (Grazer Philosophische Studien, 7-8.) Amsterdam: Rodopi

Chisholm, Roderick M. (1976) Person and object. London: George Allen and Unwin.

Davidson, Donald (1963) "Actions, reasons and causes". Journal of Philosophy 60, 685-700.

Davison, Scott A. (1994) "Dretske on the Metaphysics of Freedom". Analysis 54, 115-123.

Descartes, Rene (1984) The philosophical writings of Descartes. Vol. 2. Cambridge: Cambridge University Press.

Dretske, Fred (1988) Explaining behaviour. Cambridge, Mass.: The MIT Press.

Dretske, Fred (1992) "Metaphysics of freedom". Canadian Journal of Philosophy 22, 1-14.

Horgan, Terence (1991) "Actions, reasons, and the explanatory role of content". In Dretske and his critics, 73-101. Brian P. McLaughlin, ed. Oxford: Basil Blackwell.

Koistinen, Olli (2002). "Causation in Spinoza." In Spinoza: Metaphysical Themes, 60-72. Olli Koistinen and John Biro, ed. New York: Oxford University Press.

Lewis, David (1973) "Causation". Journal of Philosophy 70, 556-567.

McCann, Hugh, A (1993) "Dretske on the Metaphysics of Freedom". Canadian Journal of Philosophy 23, 619-630.

Thalberg, Irving (1977) Perception, emotion and action. Oxford: Basil Blackwell.

Von Wright, Georg Henrik (1971) Explanation and understanding. London, Routledge and Kegan Paul.

Olli Koistinen

University of Turku


Olli Koistinen

Department of Philosophy

20014 University of Turku



Tel.: +3582 3335 416

(1) Also Scott A. Davison (1993) and Hugh J. McCann (1994) are in favour of (P).

(2) I have considered g-causation more fully in Koistinen (2002).
COPYRIGHT 2009 Estonian Academy Publishers
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2009 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Fred Dretske
Author:Koistinen, Olli
Article Type:Report
Geographic Code:4EXES
Date:Jun 1, 2009
Previous Article:Research and teaching of history at the University of Tartu at the end of the 19th and the beginning of the 20th century.
Next Article:Wittgenstein's Tractatus 3.333 and Russell's paradox.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters