Jason Bridges

University of Chicago


Phil 31410, A03—lecture 3 notes

More on the ontology of action

(Davidson, “Agency”)


I. Summary of main elements of Davidson’s account of ontology of actions

So let’s summarize what we were talking about last time.  Then I’m going to say just a bit more on some of further related issues.  I hadn’t planned on going into them, but interesting questions were raised last time and it can’t hurt.


Davidson on actions:

1.     An action is a person’s intentional doing of something.

2.     An action is an event, and like other events, can be picked out various descriptions.

3.     To be an action, an event need be intentional only “under” one of its descriptions


This is Davidson’s way of saying that an action can be an unintentional doing of something, so long as it’s an intentional doing of something else.


4.     One way an action can be described is in terms of its effects.


If I describe an action of yours as a killing of Bob, then I have described it in terms of one of its effects, namely Bob’s death.  An action simply can’t be a killing of Bob unless it results in Bob’s death.

That is to say:

“Your action was a killing of Bob” implies “Your action caused Bob’s death.”

Note that I put it in this awkward way because what might at first seem a more straightforward way of putting the point seems to me to be not quite true:

“You killed Bob” implies “You performed an action that caused Bob’s death.”

Suppose Bob is allergic to your sweat.  He touches you and dies as a result.  Here it would be true to say that you killed Bob, but you didn’t perform any action that caused his death.

What’s right is to say that, if it’s in performing an action that you kill Bob, then that action causes Bob’s death.

Many of our ordinary ways of describing actions carry the implication that the actions result in certain kinds of effects.  For example, speaking of an actions as a rescuing (of someone), a boiling (of, say, water), a destroying, etc.


5.     If a person does x by doing y, then her doing x is the very same event as her doing y.


Thus if the queen destroys the kingdom by moving her hand, her destroying of the kingdom is the very same event as her moving of her hand.  Remember Leibniz’s law: if A is identical to B, then A has all the same properties as B.  Thus since these two events are identical, they have all the same properties.  In particular, they begin at the same time and end at the same time.  Thus the queen’s act of destroying the kingdom, since it’s identical to her moving of her hand, ends as soon as her moving her hand is over.

This can seem counterintuitive.  And there may well be reasons to worry about this.  But one thing Davidson wants to point out is that the claim doesn’t imply that the destruction of the kingdom happened at the same time of the moving of the Queen’s hand.  It didn’t.  In fact the moving of the Queen’s hand, via various intermediate causes, ended up causing the destruction of the kingdom.  Since one is the cause of the other, they can’t be the same event.


So where as one might be inclined to say that the “by”-relation holds between two actions when one causes the other, Davidson claims the truth is somewhat subtler:


6.     Typically, “by” connects descriptions of a single action in accordance the causal order of the relevant effects.


Thus, in our example from last time, because:

The Queen’s action à the poison’s dripping into the King’s ear àthe King’s death.

We can describe the Queen’s action both as her pouring poison into the King’s ear and her killing the King

Since these three events on the line are distinct and causally related, they must happen in temporal order.  But the Queen’s action, of course, takes place at the same time as it takes place, however described.  (There might be significant overlap between the queen’s pouring the poison in to the King’s ear, and the poison’s dripping into the King’s ear, but the one precedes the other at least by a bit.  It’s not atypical for a cause to overlap in time with its effect, in cases where they both take time.  The wind’s blowing causes the tree to bend.


A different objection to 5 might be developed as follows.  Suppose the Queen hypnotizes herself on Monday to kill the king on Monday.  This action leads to a further action on her part, killing the King, which happens on Tuesday.

You might say: then the Queen killed the King by hypnotizing herself.  And these are distinct events related as cause and effect, which would contradict Davidson’s analysis.  But in fact it doesn’t quite seem right to say the Queen killing the king by hypnotizing herself.  She killed the king by poisoning him.  What we might say is that she got herself to kill the king by hypnotizing herself.  That is to say, in hypnotizing herself, she did something that caused herself to kill the king later on.  And that’s consistent with Davidson’s analysis.  The queen’s hypnotizing herself is identical to her doing something which caused her to kill the king later on.  And her pouring poison in the king’s ear is identical to her killing him.

[Hypnotizing case is bad example.]

Map this out.


Here’s another example that may help clarify the point.  Suppose the queen hypnotizes herself to boil water at a later time.  And she does it by boiling water (the water contains some hypnotic drug).  Would we then want to say:

The queen boiled water (today) by boiling water (yesterday).

This would be a counterexample to Davidson’s view.  The queen performs one action of boiling the water by performing another action of boiling the water.  Two distinct events.

But in fact, this seems false.  What seems fair, though, is to say this

The queen got herself to boil water today by boiling water yesterday.

But both of these are descriptions of yesterday’s action.  Yesterday the queen performed an action that was both a boiling of water and a getting herself to boil water.

There are further complicated objections of this sort, be we should move on.


Note finally, that I’ve hedged 6 with a “typically”.  Davidson doesn’t speak to this, but there do seem to be causes in which we do one thing by doing another but there’s no corresponding effects related causally.

I greeted you by waving my hand.

Doesn’t seem that there is an event, a greeting of you, that is caused by my action of waving my hand.

This suggests the “by” talk captures a more general notion of dependency than causal dependency.

It’s consistent with this, though, to suppose that in the typical case what generates the relations of dependency among the different things done in an action is the causal order of the ensuing effects.


A related issue that must be kept distinct is this.  Sometimes an action has other actions has parts.  This is an instance of a more general point about events.  Some events are composed of a series of smaller events.  Thus the Civil War has as a part the battle at Gettysburg, which is itself an event, as well as the battle at Antietam and so on.

Similarly, suppose I set the table.  My setting of the table is certainly an action.  But it consists of a number of identifiable smaller actions: my putting one glass here, and a fork there, and so on.

Objects, of course, are also composed of parts, which are other objects.  Your body is composed among other things of your leg, which is itself an object.

Nothing untoward here.


II. Primitive actions

Finally, a word on Davidson’s talk of “primitive actions”.


It might seem on reflection that the list of descriptions of actions linked by “by” must come to an end.  At some point there must be something one can just do, not by doing something.

One does x primitively = One does not do x by doing something else.


What Davidson thinks we can just do, do primitively, is move our bodies.

Davidson agrees: he thinks what we can just do, what we can do by not be doing anything else, is move our bodies.

Thus in our example, there is nothing by doing which the queen moves her hand.  She just moves her hand.


Why might one think this?  One reason is that that seems to be where the buck stops in the deliberation that leads to actions.  The Queen might set out to destroy the kingdom and ask the best way to go about doing that.  She might decide killing the king is the way to go.  But how to kill the king?  By pouring poison in his ear.  But how to do that?  By moving her hand just so.  But how to go about moving her hand?  At this point the questions give out.  One does not have to reflect on how to go about moving one’s arm, say.  Certainly one doesn’t need to think about what muscles to contract or what have you.  One just moves one’s arm.


This train of thought is not decisive.  One might point out that in general we don’t need to take this final step.  We don’t think about how to move our bodies in order to pour something.  We just pour it.


And note that some philosophers think that we need to take one further step back: what we can just primitively do is try to move our bodies.  If we do move our bodies, we do so by trying to move our bodies.  Davidson doesn’t agree, and I won’t pursue that matter here.


So suppose we accept that we can just do is move our bodies.  Now, we might think it would be helpful to classify actions in terms of whether they are doings of things that one can just do, which we’ve deciding are movings of our bodies, or doings of things by doing other things.  Let’s call those actions that are doings of things not by doing anything else “primitive actions”.


Go ahead, says Davidson.  But then note that all actions are primitive actions.  Consider our example again.  Suppose that what the queen does not by doing something else is move her arm.  So her moving of her arm is a primitive action.  But note that on his picture her moving of her arm is her killing of the king.  So her killing of the king is a primitive action.  And so on.


7.     What a person can do primitively (not by doing something else) is move her body.


Note the conclusion:


8.     Thus every action is a person’s moving of her body


One last issue on this.  Consider these sentences:

I boiled the water.

The water boiled.

This shows that “boil” can be both a transitive and an intransitive verb.  A transitive verb is one that takes a direct object, where the direct object is a noun phrase that identifies the thing that is the recipient or otherwise the object of the activity described by the verb.

Many verbs can be used both transitively and intransitively.


Now, it’s natural to think, and a view that Davidson famously defends elsewhere, that each of these sentences is made true by the occurrence of an event of a certain kind.  That is to say, each implies a statement asserting the existence of an instance of a certain type of event.  Thus, respectively:

An event of my boiling the water occurred.

An event of the water’s boiling occurred.

An event of my boiling the water is very likely an action.  It’s very likely that in boiling the water, I was doing something intentionally.  An event of the water’s boiling is definitely not an action.

Now what makes an action by me a case of my boiling of the water?  On Davidson’s view, to describe an action as such is to imply it had a certain kid of effect, namely, an event of the water’s boiling.  This latter is not the same event as the action, it is a distinct event which is caused by the action.

Thus we’d be in a position to say something like the following:

My boiling the water caused the water’s boiling.

When you take a verb and used it in a noun phrase, you’re said by grammarians to have nominalized the verb.  And the resultant noun phrase is said to be a nominalization of the original verb phrase.  Notice that the respective uses of “boil” retain their respective transitive and intransitive forms….


But now take a phrase like “the boiling of the water”.  This phrase is ambiguous.  It can either refer to an event of someone’s boiling the water or of an event of the water’s boiling.  So it can either involve a disguised version of the transitive use or it can involve the transitive use.

Thus we might been in a position to say something like:

The boiling of the water caused the boiling of the water.

In the first case, “the boiling of the water” picks out an action, an event of someone’s boiling the water, and in the second case picks out the effect of that action, namely an event of the water’s boiling.


One reason I mention this is because it may be that this ambiguity lies behind some of the resistance to Davidson’ view of the “by” relation that we’ve been considering.  But I won’t elaborate.


And the other reason I mention this is this.  Consider:

I moved my arm.

My arm moved.


One might hold on the face of it that we ought to be able to say everything about this pair that we just said about the last pair.

Thus that they respectively imply:

An event of my moving my arm occurred.

An event of my arm’s moving occurred.


And moreover, that what makes an action an event of my moving my arm is that it causes an event of my arm’s moving.  Thus:

My moving my arm caused my arm’s moving.

But notice what this implies.  That the action of my moving my arm actually occurs, or at least begins, before my arm moves.  Thus even if on Davidson’s view all actions involve a person’s moving her body, all actions occur prior to, because cause, the agent’s body’s moving.  Elsewhere, Davidson indicates that he rejects this view.  But it seems to be a natural implication of what he says.


Does this view imply that actions aren’t bodily movements and that they actually occur prior to them?  It depends what you mean by “bodily movements”.  It doesn’t suffice to resolve this to say that a bodily movement is an event of a moving of a body, in light of the ambiguity just mentioned.  For it seems on analogy to what we said about boiling we can intelligibly say something like this:

The moving of my arm caused the moving of my arm.

In the first case, the phrase picks out an action, my moving my arm, in the second case the effect of that action, my arm’s moving.  The question whether actions are bodily movements would then turn on the question in which of these senses do we want to say that the bodily movement is the moving of my arm.


Further reading:

Judith Jarvis Thomson, “The Time of a Killing”, “Individuating Actions”

Anscombe, ?, “Under a Description”

Davidson, “Problems in the Explanation of Action”

Hornsby, Actions