Reading Group: Davidson: Actions, Reasons, and Causes

The Learning Zone Reading Groups
61

Been reading here and there in this post, so thought I’d enter.

I wonder if it’s a valid objection to Davidson that his collapse of reason with cause is single directional. That is, the Venn diagram would have causes and within it would be reasons (all reasons are causes, but not all causes are reasons). Why not flip that (all causes are reasons, but not all reasons are causes) or why not make all reasons causes and all causes reasons?

That is, he wants to make all reasons causes but not all causes reasons, but that just strikes me as an ad hoc manuever to construct a naturalistic model. Under naturalism, we assume all reasons have underlying physical (neuronal) explanation and then he goes on to argue that a reason must be a cause as well. Of course, confirmation bias demands that. Monism must exist in a naturalistic setting.

I go to the store because I need milk (the reason) and I go becuase my neurons and muscles fired a particular way (the cause). His move is then to say that reasons are causes.

The asymetry reveals bias. Why doesn’t he say all causes are reasons too? Why does the argument go only one way? As in, the earth spins because of gravity (the cause) but it also spins to sustain human life (the reason).

To say every cause has a reason is to say that there is a reason for all that happens. That view is pure teleology, but it seems as no greater leap to collapse the world into naturalism than into theism. It also lends itself to end up where we likely started, which is that reasons and causes are not ultimately reducible into each other, eliminating monism and back to dualism.

I tend toward Wittgenstein on the philosophical question here, which is that cause talk and reason talk are different games, neither of which can use metaphysical difference as the distinction. That’s as far as philosophy will get you at least. The rest are just assertions of worldview masquarading as necessity.

62

Yes, that is indeed what Davidson says here, and perhaps “flirts with” is unfair. But here’s what I mean: The complete equality of description (mental or physical) serves Davidson well when he wants to argue that a reason can be a cause. And it serves anyone well who is willing to treat “extension” in the logical sense of “the thing named or described,” full stop.

But we run into problems when we insist on looking more closely at “thing” (or “event”) and asking, What qualifies here? Especially: What sort of thing is going to count as “the very same” due to having the same extension as some other thing? The fact that we can say “‘Desire’ and ‘neural events a - j’ have the same extension” doesn’t help us understand what’s actually happening, since we have no idea what “desire” is, other than to note that it’s the word we use to denote a certain “mental event”. That’s the arm-waving aspect I alluded to earlier. This is really different from noting that the Evening Star and the Morning Star have the same extension, even if we grant that here too “extension” is not res extensa. We know what we mean by each star; we’re not wondering how “appearing in the morning” can be a thing at all.

It looks to me like these two statements are contradictory:

"It is not true that causation is extensional, extension is physical, so therefore the causally relevant description must be physical.”

“Extensionally, the metal disk and the coin are the very same.”

The second statement could only be true if the disk and the coin are “the very same” by virtue of physical extension. But the first statement denies this – rightly, I would say. We’d need the res extensa to be replaced by the “extensional context” in order to make the second statement fit the first.

It’s true that, as you say, “the physical description has no special place, not yet.” But that comes at a cost. Davidson should really give up “very same-ness,” which he’s loath to do. When he eventually discusses anomalous monism, I think we find him opting for an understanding of the physical that is indeed primary, though I’m open to the nomological construal you suggest; but that still makes it primary.

So: When you say, “That the metal disk has a certain value is a property of the disk under a certain intensional description, but that does not render it not a description of the disk,” I think I would disagree. By adding the “certain value,” we’re no longer describing the disk at all, but rather the coin. Unless, of course, we’re willing to go back to the “res extensa” idea of extension, in which case then yes, they are “the very same.” But then we give up the whole idea of being able to describe a mental event as a cause, since mental events and causes are not the very same, in this way.

This is all quite complicated, and I’m by no means confident that this must be right. I’m happy to consider other interpretations of Davidsonian causes.

63

I would also allow reasons to be causes under some circumstances – there are many discourses in which it would be incalculably tedious to keep reminding everyone that it’s the physical stuff that’s actually doing the causing. Moreover, such a reduction would also destroy the point being made, in many cases. It’s similar to the idea that a biologist, to speak “properly,” has to explain their subject matter in terms of particle physics. In our macro world, we need the “loose talk” of macro causation. To abandon it would be an affront to ordinary discourse.

But the case I had in mind wasn’t so much a denial that “reasons are causes and nothing else.” We both agree, no doubt, that reasons can also be explanations and justifications, so that’s not really the issue. What I think we should resist is saying that a reason, under that description (which, remember, is all we’ve got!) and without reference to an underlying physical theory, can be a cause. I mean, we can suggest that, but we’re so far from understanding how it can be the case that I think we ought to leave it alone.

And yes, I acknowledge that in a way this is nothing more than expressing a preference for “The reason I did X was Y” rather than “Y caused me to do X.” I understand the former, more or less, but not the latter.

Oh, definitely, and I wish I had more time to keep writing.

64

No interruption.

In the article, flicking the switch alerts the prowler; this is taken as the same event; but on a different account, flicking the switch caused the light to come on which caused the prowler to be alerted; the individuation of the events is different.

There is more than one way to individuate; to decide if it is one event or several, although casually linked… no fact of the matter but a difference of individuation.

However see footnote 2, “We would not call my unintentional alerting of the prowler an action, but it should not be inferred from this that alerting the prowler is therefore something different from flipping the switch, say just its consequence.”

Davidson has a criterion: individuation tracks what falls under a description of my doing something, including unintended upshots; what we are looking for is what he did. One might well offer an alternative, that flicking the switch caused the light to turn on which caused the prowler to be alerted; and notice here the move away from the agentive account: he flicked the switch, he turned on the light, he alerted the prowler. Both are valid, only one does what we are after, which is individuating the action.

This cuts to the structure of Davidson’s article. There is a question here that gets answered by causation, but not by other accounts: How on any other account do we explain an action in terms of a reason? If the movement of those arms and the poisoning were two different events, “I was only moving my arms up and down” would be a defence; because that movement was the act of poisoning the population, it isn’t. That’s the Wittgensteinian notion of meaning as use in Anscombe and Davidson; that there are differing descriptions of the same act because the one act does different things.

65

It is a mistake to read Davidson as being at odds with a Wittgensteinian account of action. That a reason rationalises an action only under a description is exactly an application of the meaning of an utterance as its use in a sentence. We have different descriptions of the same event, because we put that event to different uses.

I would offer the following, from p 691, as the centrepiece of the paper:

…a person can have a reason for an action, and perform the action, and yet this reason not be the reason why he did it. Central to the relation between a reason and an action it explains is the idea that the agent performed the action because he had the reason. Of course, we can include this idea too in justification; but then the notion of justification becomes as dark as the notion of reason until we can account for the force of that ‘because’.

If a reason explains an action, that action must happen because of that reason; anything less fails to provide the link between reason and action that is required in order that a reason be an explanation for an action.

This is why C2 is our only recourse.

66

There’s an ambiguity in “reason” that Davidson sets aside right at the beginning of the paper. Davidson is using reason as the explanation given by an agent in explaining their actions, not as a synonym for “cause”. A reason for the purposes of his paper is a pro- attitude together with some beliefs.

67

Wittgenstein wouldn’t disallow a reason to be a cause, but that’s just because contextually sometimes it is. Typically a cause references physical laws at play and disallows telological or intentions as proper forms of causative explanation. Importantly though, saying a cause must be X or Y would create a problem for Wittgenstein just because it nods towards essentialism and not usage. But just because we say X was a reason and Y was a cause draws no metaphysical distinction, but just one of usage.

But what is at stake for Davidson if he should acknowledge that reasons are causes but only to the extent usage overlaps and means to say no more?

68

I think I follow that correctly, but that would mean those explanatory reasons are causes for certain actions just as physical causes would be. In other words causes <> reasons, but some reasons are causes. That some reasons are not causes is because some reasons bring no action. Correct?

69

I think this is basically on the right track; ironing out the details is where this thread might best head.

Davidson presumably followed Quine in giving primacy to the language of physical science. In particular, he presumes too naive an approach to causation. Anscombe wrote quite a bit about causation. She might be described as following the spirit of Russell in downplaying the role of causation in scientific discourse. In particular, we might question the place of universal statements of one sort or another in setting out causal relations.

But this is a large topic, with much to consider. I’ve attached Anscombe’s paper. anscombe-causality-and-determination.pdf (1.5 MB)

On another topic: mixing Frege and Descartes will not do. “Extension” is used very differently in each; they are unrelated. In Descartes, extension is the stuff that something is; in Frege, it is a rule for substitution. It seems to me you continue to mix the two, to your confusion.

70

See the discussion of extension nearby. If two items are extensionally equivalent, one may be substituted for the other in a statement without change of truth value. Here, the reason and the cause are set up as extensionally equivalent, and hence the reason (the pro-attitude and belief) can substitute for the cause in explaining the act of some agent.

What you seem to want is to explain all causes in terms of reasons, but not all causes can be put in terms of pro-attitude and belief; the rock does not have a pro-attitude towards breaking the window.

So reasons are a sub-class of causes.

71

I hope, if you have been reading along, you can now see that a reason, as it is being considered here, is not a motivation, but an explanation. Of course it remains open that the reason might also be the motive. We might ask, what do we gain by adding this extra term? And perhaps it is not needed.

72

I hope you have been reading along, and if so, that it is apparent first that Davidson is offering an analysis of action, not a normative framework; and that there is a lot more going on here than a simple reduction to our actions being physical.

73

That’s a bit too quick. Take a look at the paper I’ve attached here and it might be clear that she does think actions are caused, but not that causes are *Nomological *in the way Davidson suggests.

74

What do you think, now that we have worked through the paper?

75

The idea appears to be, that reasons are reducible to biological and physical causes, and reasons are just the internal narratives we attach to those physical deterministic events.

I find this problematic in a number of ways. If you ask me to hand you something, you exert no relevant physical forces on me. My handing it to you cannot be reducible to an unbroken chain of physical deterministic causes that reduce to an invariable physical law. But more than that, I don’t accept determinism even about physical events. I think that this is an attempt to import logic to physics. Or it takes the necessity, invariance, and non-informational nature of logical operations and imports it to empirical relations, despite being unfounded, and counter to observation.

Ideas such as downward causation, and multiple realizability challenge determinism and reductionism as well. I favor pretty strong empiricism, myself, and prefer to keep abstract logical operations out of it as much as possible.

76

Surely not - it is that a description of an action in terms of a reason will be extensionally equivalent to a physical description. That at least must be seen as what Davidson suggests here.

Superman is not reducible to Clark Kent.

77

There are no distinction between superman and clark kent. Literally anything whatsoever you can say of one, you can say of the other. When superman raises his hand, so does clark kent. When superman eats a sandwich, so does clark kent. When superman stands on the marked spot, so does clark kent. There are not two things, but only one, and as such literally every single conceivable thing that can be said of one, can be said of the other. Does that apply here?

78

Well, no. Louis Lane believes Clark Kent needs glasses. She does not believe that superman wears glasses.

79

If I say “I learned that Clark Kent was Superman” does that mean the same thing as "I learned that Clark Kent was Clark Kent’?

Your error, I’d submit, is that meaning is referent and not use.

80

I don’t think it matters what you learned or thought about clark kent and superman. The same thing that is true of one is always true of the other. People may think that clark kent ran out for his sixth bagel that day, but in reality he is being deceptive so he can go supermaning. Or whatever else people learn, or think, or whatever.

1 Like