Jason Bridges

University of Chicago

Phil 31410, A03—lecture 4 notes

Reasons and Rationalizations

(Davidson, “Actions, Reasons and Causes”)


To begin, I want to say a few brief words that look back to what we just talking about and forward to what we’re about to talk about, before getting on to “Actions, Reasons and Causes”


I. Propositional attitudes

First on propositional attitudes, which is a crucial notion of “Actions, Reasons and Causes” and everything we’ll be talking about after.


Prop. attitudes are things like wishes, beliefs, desires, intentions, hopes.  They are called propositional attitudes because they are attitudes toward propositions.

Take my belief that that Chicago has good Indian food.  There are two aspects to this belief, an attitude and a proposition.


Firstly, there is that which I believe—in this case, that Chicago has good Indian food.  This is called a proposition.  Propositions are things that can be true or false.  Chicago is not itself a proposition; it makes no sense to say that Chicago is true or that Chicago is false.  But ‘Chicago has good Indian food’ expresses something that can be true or false, so we call what it expresses a proposition.  This particular proposition happens to be true, but other propositions, of course, are false.


Secondly, there is my attitude toward the proposition that Chicago has good food—in this case, my attitude is that I believe it.  But I might have disbelieved it, or doubted it, or even feared it.  There are countless attitudes one might take toward a given proposition.  For example, I might wish that Chicago had good Indian food.


Note that sometimes instead of talking about the proposition that the attitude is an attitude toward, philosophers talk about the propositional content of the attitude.  The propositional content of my belief that Chicago has good food is that Chicago has good food.  And sometimes this just gets shortened to talk of the content of my belief.  Content, then is what’s believed in a belief, or what’s desired in a desire, and so forth.


A quick point before moving on regarding a potentially distracting side issue.  Davidson, and other author’s we will discuss, treat states like wanting sleep or hoping for rain, etc., propositional attitudes.  But this might seem problematic, given the words I used to expresses these attitudes didn’t have the grammatical form that one would associate with a propositional attitude.

Let’s take a different case that raises the same point.  Compare:

I wish that Chicago had good Indian food.

I want to eat some Indian food.

I want Indian food.


In the first case we have a “that” followed by what on its own could serve as a declarative sentence.  So it expresses a proposition.  It can be true or false.

But in the second case what follows the attitude-verb is not something that could stand on its own as a sentence that is true or false, and so not something that expresses a proposition.  Still more so in the third case.  The third case is sometimes called an expression of an objectual attitude.  Typically, what follows “want” in particular, and also its cognate “desire”.

But can’t such attitudes as a desire to eat Indian food be an agent’s reason for doing something?  Indeed, Davidson’s own examples generally involve such attitudes.  He takes as an example to illustrate his view a case in which you bite your thumb at me because, as he puts it, you want to insult me.


What Davidson tacitly assumes, and what many philosophers assume is that these sorts of formulation are always short for a formulation that has the “that”-clause structure.  So the propositional attitude expressed by the last of our utterances, for example, would more fully be expressed as follows:

I want that I eat some Indian food.

This sounds awkward.  But it does seem true that if I want Indian food, what I want to obtain is that I eat some Indian food.

Whether this works in general, for all objectual and otherwise non-propositional attitudes is a question I won’t get into here.



II.  Intentionality

But first, another word on intentionality, a topic which came up last time but also bears on an understanding of propositional attitudes.


Chisholm introduced the term into the English language, prompted by some work of the late 19th century German philosopher Brentano.  Brentano identified this phenomenon of “intentionality” and claimed that it was distinctive of the mental.


“Intentionality” is a technical term; it is not especially connected to the idea of an intention (although intentions will be among the mental phenomena that exhibit intentionality).  Another phrase for it, Brentano says, is ‘directedness’ toward an object.  What is distinctive of mental phenomena, for Brentano, is that they’re directed toward objects.

Another way to put the idea is that mental phenomena, but not physical phenomena, are of or about  or toward something.  My fear of spiders is a fear of something, namely, spiders.  Spiders are not themselves of anything in this sense.  My belief that the sun is hot is about the sun, but the sun is not itself about anything.  My anger is directed toward you, but you are not in this sense directed toward anything.


Now, prepositions have loads of meanings/uses, as anyone attempting to learn a foreign language knows when they try to map uses of the prepositions of one language onto another.  So it’s not as if just using these three prepositions is supposed to fix the idea of intentionality.  It doesn’t, for the simple reason that these prepositions have unrelated uses.  (You are the daughter or son of someone; that doesn’t mean you have intentionality.)  So what matters hear is the particular examples we’re using.  They’re all supposed to exhibit rather clearly a common feature.  Call it directedness, aboutness, intentionality.


Now, we may not agree that all mental items have intentionality (e.g., pains).  But propositional attitudes do.


Chisholm refined the idea.  In particular, he characterized what he called the “intentional” use of language.  And he claimed that ascriptions of propositional attitudes must use language in this way.  Thus he initiated an application of that term to linguistic items, in particular the words we use to ascribe propositional attitudes and other things, rather than to propositional attitudes themselves.  This in turn led to the term, “intentional content”.


The major feature of the intentional use of language is nowadays called substitution failure.

One such failure is a failure of substitution of co-referential referring expressions.  Ordinarily, co-referential referring expressions are intersubstitutable salva veritate, i.e., saving truth.  If you have a sentence and replace a referring expression in it with an expression that refers to the same thing, you get a sentence with the same truth value.

For example:

X is a grad student.

That guy is a grad student. (said pointing at X)

If the former is true, and so must be the latter.


But now consider these two sentences:

Y believes that X is a grad student.

Y believes that that guy is a grad student.

Here we may well have a divergence in truth value.


Or consider a famous example, first discussed by our very own Leonard Linsky: Hesperus and Phosphorous…

This failure of substitutability is one of the things that is distinctive of intentional uses of language.  Note that sometimes when we are talking about the purely linguistic phenomenon, as opposed to the ‘metaphysical’ idea of intentionality this word is used instead: “intensional”.  The part after the phrase “Y believes that” is an intensional context.  The reasons for this terminology are beyond the scope of this course.


Analogously “that”-clauses of propositional attitude ascriptions exhibit a failure of substitution of descriptions that pick out the same things.

Y believes that the author of Prior Analytics was a genius.

Y believes that the founder of the Lyceum was a genius.

(Note that these phrase are called “definite descriptions” because they pick out exactly one item.  Sometimes people speak of what a definite description designates by which they mean the item that it picks out.  So we can speak of co-designating definite descriptions or co-extensional definite descriptions.


(An analogous point, by the way, also holds for co-extensional predicates.  The extension of a predicate is the group of objects that that predicate is true of.  The extension of the predicate “is blue” is the class of blue things.  The extension of “is a wolf” is the class of wolves.  Ordinarily co-extensional predicates are intersubstitutable salva veritate.  If you have a sentence and replace a predicate in it with a predicate with the same extension, you get a sentence with the same truth value.  But not in propositional attitude context.)


Now consider the following:

My taping of The Ring was intentional.

My taping over of Sex and the City was intentional.

These can differ in truth value as well.  If we accept Davidson’s view that, in the case as I described it, these descriptions pick out the same action, we must allow that “____ is (was) intentional” is an intensional context as well.  [or “quasi-intentional”, as Davidson puts it, given that the context doesn’t allow for non-existence of object.]


As we shall see, it may not be a coincidence that both propositional attitude ascriptions and talk of which actions are intentional both exhibit intensionality.  It may have to do with the fact that propositional attitudes are appealed to in the explanation of actions.



III. Reasons and rationalizations

The main aim of “Actions, Reasons and Causes” is to show that our ordinary explanations of people’s actions, in which we explain the performance of an action by giving the person’s reasons for doing what she did, are causal explanations.


Davidson calls explanation of actions in terms of the agent’s reasons “rationalization”.  So the thesis is that rationalizations are causal explanations.


A rationalization of an action A = An explanation of A in terms of the agent’s reasons for A


A rationalization of A is an explanation of why A occurred that consists of a specification of the agent’s reason (or reasons) for A.


Now, I don’t like the term “rationalization” myself in this context.  I prefer “reasons explanation”.  This has the advantage of showing on its face that what is at issue is a form of explanation (and it avoids the potential confusion with our ordinary usage of “rationalization”, in which a rationalization is an attempt to justify an action that has a whiff of dishonesty about it, an attempt, perhaps, to disguise the agent’s real reasons and posit other reasons that put her in a better light.

“reasons explanation” is kind of an ugly term. “rational explanation” is more appealing.  But at this stage it assumes too much for our purposes.


The first point to register in making sense of this definition is that when we speak of a person’s reason for doing something, as I mentioned in the first lecture, what is in question is not merely a reason in the undemanding sense in which any anything that explains something counts as a reason for what it explains.  If last night’s heavy wind explains why the tree in my yard fell over, then last night’s heavy wind is the reason the tree in my yard fell over.

But the heavy wind was not the tree’s reason for falling over.  The tree itself had no reasons.

Nor does the fact that what is being explained is something a person does ensure that the explanation will involve appeal to reasons in the sense we are concerned with here.  My clumsiness and distraction may be the reason I trip down the stairs, but they are not my reason for tripping down the stairs.


Contrast a case in which I trip down the stairs in a desperate bid to avoid an unpleasant conversation.  Here my wish to forestall the conversation is not merely the reason I trip.  It is my reason for tripping—it is the reason that I act upon, the reason upon which I base my action.


Just to have a label, I will call reasons in this sense agential reasons.  This is not terminology Davidson uses, but it’s less ungainly than constantly talking about an agent’s reasons or the reasons of an agent.

An agential reason is a reason upon which, for which, an agent acts.

The agential reason in the case just described, then, would seem to be my wish to avoid an unpleasant conversation.


Now, although Davidson is not especially explicit about it, he endorses the thesis that whenever an agent does something for a reason, that doing is an action.  Being done for a reason is sufficient to make the doing an action.  This suggests the following view of what it is for something to be an action:

An action is a doing of something for a reason.

This is in effect a second answer to the constitutive question about agency.  We exercise our agency in doing things for reasons.

Now recall the criterion of agent-hood discussed last time:

          An action is an intentional doing of something.

Since Davidson accepts both, this suggests he accepts this equation (thought, of course, the equation isn’t implied):


To do something intentionally is to do it for a reason.


And that is indeed a view that Davidson endorses.  We shall have occasion to talk about this equation later on.  We shall also see that there are various important distinctions and qualifications we need to make in characterizing reasons talk.  I’m going to suppress those issues for now.


Note that it is an implication of this answer to the constitutive question that all actions have reasons explanations (rationalizations).  All actions can be explained by citing a person’s reason for doing something.


Reflection on this equation also should make us realize that we need to refine the account of what a reasons explanation is.  For recall that an intentional doing of something can be an unintentional doing of something else.  Thus it should follow that a doing of something for a reason can be a doing of something else not for a reason.

And indeed that is the case.  I may break my leg by tripping down the stairs.  It doesn’t follow that I broke my leg for a reason.  My tripping down the stairs for a reason may be the same event as my breaking my leg not for a reason.

This means that one doesn’t have a reason for a given action period, but only for that action as described in a certain way, as Davidson would say, under a certain description.


A rationalization of an action A under the description d = An explanation of why A occurred that specifies the agent’s reason (or reasons) for A under a given description d.

Cash this out in terms of the example….


Now what is it to give a person’s reasons for an action?

This is not as simple a question as it might seem.  In fact, in an important respect, it will be the central topic of this course.  So we won’t be saying everything there is to be said about it today.

But we can ask, what on Davidson’s view is it to give a person’s reasons for her action?

It will help in what is to come to carefully distinguish the weaker and stronger claims that go into his view.


Recall our example.  My clumsiness does not seem a good candidate for being my reason for tripping (even if it is the reason I trip).  But a wish on my part, say to avoid talking with you, can be my reason for tripping.

Now, a wish is a propositional attitude.  In this case, then, it seems that my agential reason is a propositional attitude of mine.  And Davidson thinks this holds universally.


The most general, hence weakest claim that goes into Davidson’s view is this:

Davidson on reasons for action:

1. Psychologism about reasons: Agential reasons are propositional attitudes, or complexes thereof.


What seemed to be true in the case just described Davidson claims holds generally.  All reasons for which we act are propositional attitudes, wishes, beliefs, desires, what have you, or combinations of such attitudes.


Okay.  So let’s grant an agent’s reasons for doing what she does are her propositional attitudes.  Her reasons for doing that are that she wants this or believes that or what have you.

To this general thesis Davidson adds a more specific one.  To do so, he introduces the notion of a primary reason for an action.

The notion Davidson obviously takes to be very important to the case he makes in this article.  And as you know from reading the article, the idea is somewhat complicated.  But we don’t need to get into all of that complexity now.


What I want to do now is mention a somewhat more general thesis that the idea of a primary reason is intended to make more precise.

2. Means-ends-ism about reasons: Every action has an agential reason that consists of a belief and a pro attitude.  The pro attitude fixes the agent’s end.  The belief is that the action is a means to that end.


A pro attitude toward something is a favorable attitude toward that thing.  For example, a desire to eat ice cream is a pro attitude toward eating ice cream.  It involves seeing eating ice cream as favorable, positive, to be done—being pro ice cream eating.

Desires and wants aren’t the only pro attitudes.  If one feels that giving up one’s seat on a bus to an elderly person is morally obligatory, then even if one doesn’t want to get up, if one has a great desire to lazily remain seated, one nonetheless still has a pro attitude toward giving up one’s seat.

Moreover, says Davidson, we must not think that having a pro attitude toward something always involves seeing that thing as actually, objectively good in some respect.  His can of paint example (p. 686).

Thus in saying that the pro attitude fixes an end, we must be careful not to think of the idea of an end as essentially involving the idea of the good.  One can have the end of drinking a can of paint simply in virtue of having an inexplicable yen to do so, even if one sees nothing actually good in doing so.


So let’s take our example again.  My reason for tripping down the stairs is that I want to avoid talking with you—I have in this thin sense the end of avoiding talking with you—and I believe that tripping down the stairs is a handy means to achieving this end.


Davidson’s view is that this means-end structure is present in every case in which a person can be said to act for a reason.  As he points out, we don’t always, or even usually, need to specify both components in order to successfully explain an action.  But that is not because both components aren’t present; it’s because in specifying one the other is obvious and it would be pedantic to make it explicit.

If we say that my reason for tripping down the stairs was that I wanted to avoid talking to you, we don’t need to add that I believe tripping down the stairs is a means to doing so.  But that doesn’t mean that that belief isn’t an essential part of my reason.  Suppose I didn’t have that belief.  Then citing this desire is no longer a plausible explanation of my action.


I won’t now attempt to critically evaluate the idea behind every action is a reason having this means-end structure.  We shall return to it as well later in the course.