Davidson: Actions, Reasons, And Causes

General Philosophy
1

Donald Davidson stands amongst the most influential philosophers of the late 20th Century. He was a student of Quine, continuing continuing the emphasis on logical structure, while bringing in developments from Tarski in order to contributed to epistemology and semantics. He also set out an approach to philosophy of action that was both at odds with and yet parallel to those of what was then loosely called Ordinary Language Philosophy.

There might be enough interest in the forums for us to take a look at a few of his more significant papers. So here is a start, from 1963: Actions, Reasons, And Causes.

The paper caused a stir at the time, eliciting direct responses from Elizabeth Anscombe and leading to a long secondary literature - 1776 citations on PhiPapers, placing it among those most cited. It begins a long discourse on action and mind, leading eventually to Davidson’s anomalous monism, an important contribution to the theory of mind and consciousness.

The article is discussed in some detail in Davidson’s SEP biography, by Jeff Malpas, with whom I was fortunate enough to have some association many years ago.

Again, my intent here is to crawl through the article in some detail, step by step. A PDF is attached for ease of access.

The question here is a simple one: when we give a reason for our actions, are we thereby setting out a cause for those actions?

My own response continues to be somewhat ambivalent. The arguments to be set out in what follows are cogent, but recognising the discomfort of Anscombe and Wittgenstein with the language of causation. The debate may partly dissolve once causation itself is reconceived in less deterministic terms. So nature of causation is also sitting here for our consideration.

To the introduction, then. The central question is asked off the top:

What is the relation between a reason and an action when the reason explains the action by giving the agent’s reason for doing what he did?

And the answer, that “rationalization is a species of ordinary causal explanation”. “Rationalization” here is understood as the reason given by the agent for theor actions.

actions-reasons-and-causes.pdf (464.1 KB)

2

How does this differ from motivation, which is as varied as the human species is.

3

By way of summary of part I:

So to part one, and almost immediately Davidson introduces a term of art: “pro-attitude”. The generality here is very broad, stretching to any action the agent is in favour of, and so covering everything form mere whim to considered decisions, from habits to changes of life. And it’s a logical, rather than a psychological, category. In that sense it is an almost empty notion, that one does something because one has a pro-attitude. So a pro-attitude is sometimes sufficient reason for why one does something; “I did it because I wanted to” at least sometimes rates as sufficient reason. And when it doesn’t, we add a suitable belief. The believe and the pro-attitude together form another term of art: a primary reason.

Such a pro-attitude, together with a belief, is taken as constituting a reason for an action. And together, the thesis goes, these are cause for that action.

4

Glad to see this thread. I’ll crack out my Davidson and chime in when I have a question or comment. So far, so good, and we can already see the problem Davidson is going to have with this: How to demonstrate that a pro-attitude aka a sufficient reason is the same thing as a cause? What stipulations must that involve? He can’t simply say, as you have him do, “one does something because one has a pro-attitude” (my emphasis). That is to beg the question, to already assert the equivalence of pro-attitude and cause.

1 Like
5

I will read it, and see what I think, but from the get go, I will say that we undoubtedly behave for reasons all the time. If the two of us spend the day together communicating, and I ask you to hand me this or that, to wait for me at such and such a place. To go get a drink or a meal because I’m hungry or thirsty, to ask someone for directions, and then at the end to drive me home. To name a few things, then it would be strange indeed to say that you doing all of those things had nothing to do with our communications. The reasons given throughout the day for them.

It is also the case that we can construct reasons for behaviors, moods, opinions, and lots of things that aren’t actually the reasons. We do that all the time too. I think we definitely do both things.

1 Like
6

I would defer to Mcdowell’s reading of Kant, and from that vantage object that Davidson is placing reasons within a casual framework, and that doesn’t do justice to the normative basis of rational explanation. A rational agent isn’t an organism whose internal states happen to cause behavior. A rational agent inhabits a world where things show up as reasons. The problem isn’t that reasons can’t cause actions, it’s that causal explanation can’t capture what makes a reason a reason.

Let’s say I see that there is a tree, and I believe there is a tree. According to Davidson, sensory input causes belief and belief causes action. But perception is already an encounter with the world as intelligible. The world itself enters the space of reasons, rather than acting as a cause of those reasons.

2 Likes
7

Hume certainly argued that actions are caused. Davidson follows Hume’s doctrine. Davidson’s account of primary reason as cause of human action can only make sense if you first accept that actions are caused, unlike Anscombe who does not believe that actions are caused (after all, actions are not like billiard balls that strike each other to cause a movement.)

Primary reason = desire + belief/idea

The issue is whether the primary reason can be called a cause or merely a practical response to the question “why did you do that?”

1 Like
8

A digression, before we move to II.

Anscombe raised the issue of intentionality under a given description, which Davidson make use of, in her Intentionality. She uses the example of someone seen moving their arms up and down.

“Why are you moving your arm up and down?”

“I’m pumping.”

“Why are you pumping?”

“I’m replenishing the water supply.”

“…To polish that lot off.”

The point is that one action can be described in many ways:

  • moving one’s arm up and down,
  • operating the pump,
  • pumping water,
  • replenishing the house’s water supply,
  • poisoning the inhabitants.

One action, various description, various levels of intent.

9

Yes, I think @Wosret, @Joshs, and @Elephant make good points. See also @Pierre-Normand’s post in the previous thread.

I would say that if we look at Davidson’s definition of a “primary reason” we will see the same ambiguity between reason-why-something-should-be-done (forward-looking) and reason-why-something-was-done (backward-looking). An efficient-causal understanding of reasons cannot achieve organic growth in this context. It must simply be assumed in the same moment that one points to the notion of “completed act” or “act that was in fact carried out.” Indeed, if “cause” is left general and open-ended, then it will not yield Davidson’s conclusion; but if it is defined in terms of mechanistic philosophy and efficient causality then the conclusion will be secure but it will lack plausibility for the reasons pointed out by others.

10

So to II.

I flip the switch, turn on the light, and illuminate the room. Unbeknownst to me I also alert a prowler to the fact that I am home. Here I do not do four things, but only one, of which four descriptions have been given.

Davidson is filling out details in the sort of account proffered by Anscombe. There is a lot going on. We have that he flipped the switch, one description of the event. We have that the flipped the switch in order to turn on the light. This is sufficient for an explanation of why he flipped the switch - we have a pro-attitude, wanting the light on, together with the belief that flipping the switch will turn on the light. Or we can equally have the pro attitude of illuminating the room, with the belief that flipping the switch will illuminate the room. What we do not have is a pro-attitude towards alerting the prowler. Hence, alerting the prowler cannot serve as a reason for flipping the switch.

All very simple.

Hence, C1:

Cl. R is a primary reason why an agent performed the action A under the description d only if R consists of a pro attitude of the agent toward actions with a certain property, and a belief of the agent that A, under the description d, has that property.

Now that’s the main body here, but there is much more going on. Hence the extensive footnote. The account is perhaps best read as stipulating the use of the terms involve. One action, under four descriptions, some of which set out intent, some of which don’t. The point here is that only some of these descriptions - flipping the switch, illuminating the room - is intended. Alerting the prowler is not intended.

The logic to which he is appealing is that of identity, and so substitution. Given the differing descriptions of the events, and yet their identity as the same event, we have an issue of The logic to which he is appealing is that of identity. Given the differing descriptions of the events, and yet their identity as the same event, do we have an issue for substitution? And how is this to play out? We have from the footnote that ‘A’ describes an action, ‘B’ describes something done voluntarily and ‘C’ a description of something done involuntary and unintentionally; and yet A=B=C… and the questions, “the same what?” and the expressed need for a generic term over these instances. it’s tempting to accept the adoption of “event” here, along with the caution; however the crux here is that no term will be suitable, and that what is needed instead is the activity of explanation elicited by “Why did you (he) do A?” It’s an appeal to a grounding explanations in acts.

The upshot of II so far is that rationalisation is description-sensitive.

11

I want to be pretty specific about the text here, so I wasn’t going to comment on not-so-specific posts. However I’d like to keep an eye on the parallels and differences with Anscombe, so this might deserve clearing up a bit:

Anscombe doesn’t deny that actions are caused, or that they can be. Her essay “Causality and Determination” argues against the Humean view that causation requires exceptionless laws; so for instance she thinks singular causal statements can be true without backing generalisations. That’s actually a point of convergence with Davidson’s own anti-Humean line in this paper. At issue also in all this is what sorts of strings of events count as causal, and that’s not a small issue. Anscombe’s own relation with cause is quite complex, not the least because for her causes are not necessitated - something with which I agree, but which is at odds withe the pop materialism and scientism found hereabouts.

12

So to the paragraph immediately after C1. The “certain property” here - in our example, the pro-attitude is towards turning-on-the-light-actions, not to the specific incident; it’s not a pro-attitude to me-turning-on-the-light-here-now, but towards the light being turned on; not towards the specific incident, but towards the general description. Hence

…the existence of the event is required by the truth of ‘I turned on the light’ but not by the truth of ‘I wanted to turn on the light’.

And the latter is the pro-attitude. So “I want that gold watch” has a physical object as its target, not a kind of action; but I want to buy that gold watch" has an action as its target.

Not just any old “want” will suffice.

13

To preview my own take on this, I agree with the above, McDowellian perspective; well put. With one emendation: It is an (additional) problem that “reasons can’t cause actions,” because without the “anomalous monism” apparatus, we don’t know how what is stipulated to be a mental or logical event (rationalization) can have a lawlike causal impact on the world. Heck, we can even remove “lawlike” and the problem remains.

Yes. And can a cause be description-sensitive? We’ll see how Davidson deals with that.

14

No spoilers. Thanks. We will get there. Folk tend to jump to the car chase.

1 Like
15

Right you are. I’ll keep in my seat. :wink:

16

I don’t follow this. If the person is afraid of burglars, they will flip the switch to scare away burglars, without knowing whether or not there actually is any burglars. Therefore a person might turn on the light to scare away the burglar when there isn’t any burglar.

This implies that under this descriptive formulation, if the end (reason) is the cause, then there could be a cause (turning on the light) without an effect (no burglars get scared away), when the reason for turning on the light is to scare away burglars. And a cause without an effect is incoherent.

Such is the case anytime the means fail to procure the end. And, it appears to be a flaw in the model because in this case, the reason for the act (to scare away burglars), doesn’t enter the description, when no burglars get scared.

17

Because ‘I wanted to turn on the light’ and ‘I turned on the light’ are logically independent, the first can be used to give a reason why the second is true.

That the act is within the scope of “I wanted to…” is what implies that it was intentional.

More tidying up. “Wanting” is suggested as a genus for pro-attitudes. That one wanted to do some act is deemed sufficient reason for doing that act. While a primary reason consists of a pro-attitude and a belief, we usually do not have to provide both - one can often be taken as granted in the given situation.

It is worth noting the brief argument concerning Ryle’s account at the bottom of p 689, since it involves a move from a single case to a general law. It is worth noting how Davidson points out that while the general law is dispensable, the primary reason is not. The “move from a single case to a general law” is Ryle’s move, which Davidson is then criticising.

And the next paragraph. Introducing an intention is changing the description of the situation. It is not introducing something new, in addition to, the situation. So “He is bobbing about” is redescribed as “She was knitting”, without introducing anything more into the situation. It’s to view the situation differently.

And finally, giving a justification is giving an explanation, but giving an explanation need not be giving a justification.

Most of this section has been terminological. A rationalisation includes a pro-attitude and a belief, although we usually do not need to state both; such a rationalisation gives the primary reason for the action. Talk in terms of reasons re-describes the action, it doesn’t add to it - and specifically it does not add an intention. And explanations function as justifications only if they are true. So justification depends on explanation, but not the reverse.

So much for part II.

18

Davidson’s account points to intentional act. In which case, he’s good. The criticism against this intentional act, however, is that it is sometimes just a one-off event, which other philosophers like Hume would deny its causality because a one off-event cannot or doesn’t establish synthetic causality, in which it becomes an empirical event that doesn’t have the character of true causality.

19

I disagree with your account of Anscombe based on Davidson’s argument. I think you may be mixing her account of external events with human actions originated from desires. As you can see from Davidson’s account of his disagreement with Anscombe, she doesn’t think that we can separate the desire to flip the switch from flipping the switch. There are no two events here. One is an external event, which is the flipping the switch, and the other one is a private knowledge of wanting to flip the switch. In causation, two events must be external.

The “string of events” that you find problematic is what Davidson calls the deviant causal chain which compromises the intention of the actor due to an array of thoughts, accidental act, and the resulting death of the climber at the end of the rope that the actor was holding on to. He let go of the rope, but he didn’t mean to, although it would be nice for him to let go of the rope because the other climber is heavy, and in the process of his thinking he got nervous, started shaking, and accidentally let go of the rope.

Now tell me the reason why the climber dropped to his death? The reason is the actor let go of the rope. Whether he intended it or not. Davidson disagrees.

If we are truly thinking of causality for the sake of causality — which by now I must say I’m with Anscombe that human actions cannot be in the same category as external events so therefore no causality — then we need to get rid of Davidson’s requirement of intention.

Actually I find this deviant causal chain as a non-issue. The problematic part is the intention that Davidson wants so much to shackle to the act, which to me is the flaw of his entire argument that human desires can be included as events—like external events.

20

Thank you for your post. There is a lot going on in there, some of which I agree with, and some of which I think problematic.

Rather than stop and address your post directly, I am going to proceed with the enterprise of working through the article. If you could point out were the problems you see occur as we go, we will have them in context and perhaps be better placed to see how they work.

Cheers.