Jason Bridges

University of Chicago

Phil 31410, A03—lecture 5 notes


(Davidson, “Actions, Reasons and Causes”)


I. Causalism

The third main claim Davidson is concerned to defend in “Actions, Reasons and Causes” is this:


Causalism: Reasons explanations (a.k.a. rationalizations) are causal explanations.


The idea of a causal explanation is very fundamental.  There have been many attempts to analyze the notion of causation, and consequently of causal explanation, but that is not a topic we can get into here.  (Although we shall shortly talk about the relationship between assertions of singular causal relations and causal explanations.)

A causal explanation is an explanation that explains by saying what produced, what brought about, what led to, the thing in question.  If I explain the abrupt ending of our lawn party by saying that an elephant stampede led to everyone’s running away, that I am offering a causal explanation of the abrupt ending of our lawn party.

Obviously, causal explanation is an extremely ubiquitous and familiar strategy of explanation.  Davidson’s view is that rationalization, the citing of a person’s reasons to explain her actions, is a species of this familiar strategy.


He spends a bit of time giving positive reasons for this thesis, but he spends most of the time trying to defuse objections to it.  That is probably because he thinks causalism is the default position.  It is the commonsense view, as he sees it, and the onus lies on those who would wish to deny it.


But before we consider his counter-objections to the objections to causalism, let’s look very briefly at his positive argument for it.

In fact, what he says on this score is a positive argument only in a weak sense.  For the thrust of the argument is that there is no alternative available, no other way of understanding the connection between agential reasons and actions such that citing the former can be seen to explain the latter.


One suggestion that was prevalent in the anti-causalist literature of the time is that citing the agent’s reasons for her action explains it by placing it in a pattern.  Davidson’s response is: sure.  But we have not arrived at an adequate understanding of the force of these explanations until we say what kind of pattern they serve to place actions in.  One plausible possibility, he suggests, is the familiar pattern of cause and effect.


Another suggestion is that citing a person’s reasons for her action explains the action by saying what justifies it rather than what causes it.


Now, Davidson accepts that rationalizations justify the actions they explain.  They do so by showing that the agent had a pro attitude toward something he believed the action was a means to.

So Davidson accepts what we might call:


The weak normativity thesis:  A successful reasons explanation of an action justifies the action.


It is crucial for evaluating this thesis to remember that the kind of justification at issue here is “anemic” (p. 9), and this in two respects:

1.     Justifying an action, in the relevant sense, is just a matter of showing how the action seemed justified to the agent at the time that she performed the action.  A reasons explanation can meet this condition without implying that the agent was correct to regard the action as justified—and so without implying that the action was justified objectively speaking.  For example, I might explain why your voting in favor of a particular referendum seemed justified to you, and it would be perfectly consistent with that explanation for me to add that your reasons are bad ones and thus that your vote was totally unjustified.

2.     Davidson says that a rationalization justifies by showing that “from the agent’s point of view…there was something to be said for the action” (p. 9).  This in turn is a matter of the agent’s have a “pro attitude” toward a type of action (plus her belief that the action in question belongs to that type).  In the second part of the course we will look closely at the question of what pro attitudes might be.  Most philosophers take the central case to be desire; we will consider at length whether this is so.  For now, we should note that Davidson understands the category in a very weak, inclusive way.  To take his example, even an inexplicable yen to drink a can of paint counts as a pro attitude toward paint drinking.  The sense in which “there is something to be said for the action”, apparently, is that it would satisfy this yen.  One might wonder whether that really captures a meaningful form of justification, even granting that what is at stake is just justification from the agent’s point of view.  We’ll return to this.


But the suggestion now on the table is stronger.  It might be put this way:


The strong normativity thesis: A reasons explanation succeeds just so long as it justifies the action.


In other words, the thesis is that to give a reasons explanation of an action simply is to justify the action (in the subjective sense of justification identified above).  Once we have done the latter, we have done the former.


The strong normativity thesis is inconsistent with causalism.  Causalism holds that a reasons explanation explains an action by saying what caused it.  But the strong normativity thesis denies that a reasons explanation need make any causal claim, for it holds that the explanation need do nothing other than show why the action seemed justified to the agent.

The weak normativity thesis, on the other hand, is compatible with causalism.  There is no logical inconsistency in imposing these two requirements on a reasons explanation: that it show that the action was justified to the agent, and that it show what caused the action.


A reasons explanation of an action, recall, gives the agent’s reason or reasons for it. We can put the difference between the two theses this way. 

Let P be an agent and A an action he performed.  The weak normativity thesis holds:


1. If we give P’s reasons for A, we have justified A.


But the strong normativity thesis also has the converse implication:


2. If we justify A, we have given P’s reasons for A.


But 2 is obviously false.  Suppose I wish to avoid an unpleasant conversation with you, and believe that tripping down the stairs is an excellent and cost-free way of doing so, and suppose I then trip down the stairs.  Then, in the relevant sense of justification, this belief and wish justify my tripping down the stairs.  Therefore, on the view we’re considering, they must have been my reason for jumping down the stairs.

But that does not follow.  For all we’ve said, I might have tripped down the stairs for another reason entirely.  Perhaps I was ordered to do so by my commanding officer.  And this is the reason I acted upon.


Davidson’s question is: what is the difference between the case in which the wish/belief pair was my action and the case in which it wasn’t? There is no difference in respect of justification.  So what could the difference be except in what caused my tripping?  In the one case, what produced, what brought about, my tripping was my wish and belief.  In the other case, it wasn’t.


II. Objections to causalism

Davidson considers a number of possible objections to causalism.  One, that causal relations must be subsumed by general laws and that rationalizations don’t depend upon or point to such laws, we shall discuss at length next time.


The responses to most of the other objections are pretty straightforward, so I leave them to your perusal.


But let me say a quick word about one, concerning the infamous “logical relations” argument.  Davidson’s treatment of it dovetails well with his treatment of the issues we talked about last time.


The objection is this, to quote Melden: “A cause must be logically distinct from its effect.”  But an agent’s reason for her action is not logically distinct from the action.


What does this mean?  Here’s one possibility, which is not discussed in precisely this form in the article.  Suppose I want to turn on the light, and that this want rationalizes my flicking the switch.


Statement of reason: I want to turn on the light.

Statement of action: I flick the switch.


Now the want only serves as part of my reason for flicking the switch if I believe that, in flicking the switch I will turn on the light.  This means that we must be able to redescribe the action thusly:


I turn on the light.

In fact, this is not quite the case: I could be wrong that in flicking the switch I will do this.  But then at least my actions must be describable as:

          I try to turn on the light.


So in order for my wanting to x to be my reason for doing y, it must be the case that my doing y can be redescribed as my doing x or at least trying to x.  This general proposition seems necessarily true.  (Actually, it isn’t.  But we’ll ignore that.)  And the thought is that this shows that there’s too close a “logical” link between the want and the action for the one to be part of the cause of the other.


Davidson’s response is to point out that effects can always be redescribed in terms of their causes, and vice versa.  So suppose A causes B.


Cause: A

Effect: B


But then it must be the case that A can be redescribed as:

Cause: The cause of B


It is a completely generally and necessary proposition that if A is the cause of B, then the cause of B is the cause of B.  Does this mean that there’s too close a logical link between these events for one to be the cause of the other?  Then nothing ever causes anything.