Here Davidson defied his contemporaries and the philosophical tradition. Wittgenstein , most notably, stopped at the explanatory role of reasons of actions and denied the possibility that a mental event could causally influence a physical event. Davidson notes that if we are to understand our everyday notion that we act because of some particular reason, we must acknowledge that a mental event causes a physical action—a concept anathema in philosophy and science, despite its implicit acceptance in everyone’s life (including philosophers and scientists!). Davidson:

In order to turn the first ‘and’ to ‘because’ in ‘He exercised and he wanted to reduce and thought exercise would do it’, we must, as the basic move [admit that] a primary reason for an action is its cause (Davidson 2001, 9–12).

We are now justified in putting a causal arrow between the mental states and the physical phenomenon (Fig. 3.1).

The cause for an action does not, of course, determine what the outcome will be; nor does each action have just one cause. For instance, I could start exercising for a completely different reason; I might start exercising in order to meet someone who also goes to this gym. In this case, it is not the desire to lose weight that causes me to exercise . In addition, by limiting his description of an action to what the agent actually does and the reasons for doing it, Davidson refuses to include the results of an action in its description. A driver who jams on the brakes to avoid hitting the car in front has indeed performed an action, whether or not he succeeds in preventing a collision. Intentions matter, as any attorney knows: A bullet fired in anger has far different implications than a bullet which is fired by accident, even if a person is killed in each case. Outcomes are not irrelevant in the law, but they are understood in the context of the intentions of the actor.²

Indeed, Davidson concludes that primitive actions , i.e., the ones that the agent actually performs, are the only actions that exist, and that as far as the effects of one’s actions are concerned, “the rest is up to nature” (Davidson 2001, 53). When I flip a light switch, illuminating the room and alerting the burglar, only flipping the light switch counts as my action .

Davidson does not deny that “an agent causes what his actions cause” (Davidson 2001, 59) rather, he maintains a sharp distinction between an agent’s actions and their causal effects. I joined and exercised at a gym; these were indeed my actions. And though we may often say, in addition, that I lost weight , losing weight was in fact the result of my action , not an action I in fact performed.

There is a real causal relationship between mental states (desires and beliefs ) and physical events (an action ). And yet we cannot say that there is a law that states that when one has such and such desire and such and such belief , a certain action will necessarily result (in reality, I still have not joined a gym). One might object here that this claim is inconsistent with Davidson ’s claim that desires and beliefs cause actions—for where there is a causal relation between two events, there must be a law that relates them. Davidson replies that the relation here is no different from many familiar causal relations between physical events:

I am certain the window broke because it was struck by a rock – I saw it happen; I am not (is anyone?) in command of laws on the basis of which I can predict what blows will break which windows (Davidson 2001, 16).

Suppose we accept that there are no laws which specify which action will result from a given reason: Why should this be? Why should reasons be causative but not determinative? One explanation of this is as follows: Beliefs and desires bring about action only when a wide range of other conditions are held constant. That is, actions always occur in an environment which powerfully shapes the outcome. For example, a desire to stay dry and a belief that it’s raining typically causes me to bring an umbrella on my walk to work; but not if I don’t believe that rain can get me wet, or if I believe umbrellas are ineffective for keeping me dry, or if I think the clouds will part as soon as I leave home, etc. These implicit beliefs shape whether my desire to stay dry will cause me to pick my umbrella up.

Not only do these implicit, background beliefs shape the outcome of an action , we also routinely alter our environment to make generally desired outcomes possible. For example, I may keep an umbrella near the front door so that when I desire to stay dry on a rainy day, I actually have an umbrella at hand when I’m leaving. The environmental regularities we establish are usually unseen—until they are disrupted by moving to a new home, a different city, or the like.

We have long been mistaken in thinking that there should be a law connecting reasons to actions owing to the ancient idea that actions are the result of practical syllogisms which mirror the form of the theoretical syllogism. From Aristotle on, the logic of action was understood to follow the form of a practical syllogism , a deduction from a major premise (all men are mortal) and a minor premise (Socrates is a man), to a conclusion (therefore Socrates is mortal). Similarly, an action could be seen as consisting of (1) of a major premise, a desire (loosing weight : All actions that make me lose weight are desirable), and (2) a minor premise, a belief (I believe that exercise is this type of action), from which follows (3) a conclusion. The action, then, follows immediately (so, immediately, I start exercising). Note that the action is supposed to follow from the conclusion immediately, just as the conclusion of a theoretical syllogism is supposed to do, without intermediaries. This was the syllogistic, deductive, conception of action: A reason explains an action by virtue of their logical relationship as premises to conclusion.