On Action

This post is about action, of both the mental and the physical sorts.

I mentioned earlier my hunch that the “standard story” (in sense of, e.g., Smith 2010) of action won’t work in the case of cognitive agency, e.g. explaining the sense in which judging is an action. But I will begin by giving some reasons to doubt the standard story of action even in its original domain of application: I do not think it does a good job of handling ordinary physical action either. This is the argument of my paper “The Instrumental Structure of Actions”.

What do I mean by the standard story of action? Consider Wittgenstein’s famous question: “What is left over if I subtract the fact that my arm goes up from the fact that I raise my arm?” (1958, sec. 621). Wittgenstein’s question suggests that a mere bodily movement is turned into an intentional action by the addition of something extra. According to the standard story of action, this extra ingredient is an appropriate set of causal relations to some independently specified mental states — perhaps, beliefs, desires, intentions and the like.

Like others, I find this approach to physical action unsatisfactory [1]. This is because it seems to me that it mischaracterizes the place of the mind in intentional action. On the picture that underlies the standard story, the bodily events that are the public manifestations of agency are, in themselves, mindless.[2] The mind is located strictly inside the organism. Bodily events are manifestations of mentality or intelligence only derivatively, in the sense that they are caused and directed by internal states.

This seems to me to be a mistake. We do better to locate intelligence or mentality in the ways in which an organism interacts with and makes its way through its environment. For sophisticated creatures like us, this occasionally involves activities of a theoretical and detached sort; but for most creatures, for most of the time, intelligence and mentality are manifested primarily in their engagement with the physical world.

What this means, I think, is that we should treat some physical movements (those that are actions) as also mental events. I intend this to be taken literally: certain bodily movements have intentional content, i.e., content of the sort that we typically ascribe to standard mental states like belief, desire and so on.

What is the argument for this? We generally do things bydoing other things. And some of the things that we do count as intentional actions in virtue of their role as instrumental parts of other things that we do. Think of walking to the kitchen to get a glass of water. Each step you take on your way to the kitchen is an intentional action on your part. Intuitively, this is because they are instrumental parts of the larger action of walking to the kitchen to get a glass of water. My contention is that the standard story of action does not have the resources to explain the status of such subsidiary acts.

As the argument for this claim goes through a number of different possible moves proponents of the standard story might make, it would take too long to reproduce here. I will only highlight a few salient points.

The relation two actions stand to each other in virtue of the one being an instrumental part of another cannot plausibly be reduced in non-intentional terms. Consider Mele's (1992) suggestion that the individual strokes of a swimmer count as intentional actions in virtue of "their particular relation" to the larger action of swimming a lap. This is fine as far as it goes, but Mele never attempts to specify what the "particular relation" is. The crucial point is no simple spatiotemporal or causal notion of parthood will do. The beating of the swimmer’s heart is causally necessary for her to swim a lap, and it takes place in the right place at the right time, and yet that does not make it into an intentional action.

Now, intuitively, the reason why mere causal relevance does not suffice for instrumental parthood is that it leaves out the agent’s own perspective on what she is doing. In some sense, the swimmer’s strokes manifest her grasp of how to swim a lap. The question is whether this idea can be cashed out in a way that is acceptable to the standard story of action.

I believe that it cannot be. This is because of the reductive ambitions of the standard story of action. This post is already getting rather long, so I am not going to reproduce the argument here (read the paper!), but the central thought is this. The standard story of action would need to analyze the mental state of grasping a way of, e.g., swimming a lap without reference to intentional action as such. As I see it, the problem for the standard story of action is that this cannot be done.

By way of illustration, notice that a simple dispositional account of grasping a way of swimming a lap is not going to work. A person who knows how to swim a lap may well have a disposition to move her arms and legs in the relevant ways in the relevant circumstances. However, she also has (or might have) a bunch of other dispositions to produce physiological responses in those circumstances (e.g., increased heart rate), and yet those responses would not thereby count as intentional actions. So which dispositions are relevant to taking the subject to know how to swim? Clearly, this question just is our earlier question about the instrumental parts of actions in a different guise. So the appeal to dispositions does not really advance the cause of reduction.

My suggestion is that we should simply give up on the reductive ambitions of the standard story. There is a way of grasping instrumental relations which simply consists in doing one thing by doing another. Our swimmer’s strokes are not the causal outcomes of an independently specifiable mental states that represents those strokes as means to the end of swimming a lap; they are her grasp of a way to swim a lap.

How should we understand such ways of thinking about instrumental relations that are embodied in actions? It is clear that a person who knows how to swim a lap retains that knowledge even when far from water — know-how is a standing state. Nevertheless, part of what it is know how to swim is, quite plausibly, to be able to recognize, online and in real time, what you need to do right now in order to swim. My claim is that such online instrumental judgments generally are embodied in your acting in the relevant ways.

At the beginning of this post I promised to say something about cognitive agency. My thought is this. As I already suggested in my previous post, it seems to me that the fundamental sense in which, e.g., judgingis an act cannot be explained in terms of how it was produced. Wittgensteinian arithmetic seems especially implausible in this case. But what has emerged from this discussion is that cognitive agency is actually notspecial in this regard: Wittgensteinian arithmetic fails across the board.

[1] I won't even try to give a representative list of names here. My own views have been heavily influenced by John McDowell, Michael Thompson and Kieran Setiya (who is more sympathetic to the standard story than the other two), while at Pittsburgh. 

[2] It’s important to emphasize that this is not meant to be an argument against all causal theories of action. The “standard story” of action has reductive ambitions that go beyond merely asserting that causation has a central role to play in a theory of action.

Comments

[Olle Blomberg]

Hi Markos!

As it happens, I was reading your paper just before I spotted your post. I found it really stimulating. I'm still thinking about it...but here is are some initial thoughts.

I think you are absolutely right that an adequate account of action must provide a treatment of the instrumental structure of action. I think you are (probably) also right that agents like us don't always represent this structure, and if we do, only represent parts of it. Neither will a merely causal story be able to capture the instrumental relations between component actions in an intentional activity like swimming for example.

So mere causation or mere mental representating won't do it. Your proposal, roughly, is then that the activities themselves embodies the instrumental structure and agent's grasp this structure by performing the actions for the sake of some purpose (where the purpose need not be represented by the agent, but may be "in" the movement itself).

You hint at an alternative though, which I think I would prefer (if it could be made to work). I'm not sure why you reject it and I'd be interested in knowing your reasons. The alternative is this: The instrumental structure is a functional structure, in which different actions are directed toward the same outcome (or 'culmination') in virture of being produced by mechanisms designed (by evolution or learning) to produce that type of outcome.

You hint at this alternative when you talk about what makes various occurrences all parts of a digestive process at the end of the paper, namely that all the occurrences are aimed at the digestion and extraction of nutrients. The idea here would be that the strokes of the swimmer are all parts of the simming activity in virtue of them all being produced by mechanisms that have been designed to produce the culmination of swimming or something like that.

In the paper, on p. 18, you say that the intentionality of actions cannot be reduced to the directionality that is involved in digestion, but this seems to me to be an interesting alternative to explore. That is, the intentionality of action might be reduced to the directionality of intentions *and* the digestion-type teleological directionality (here I'm having Steve Butterfill's paper "Two kinds of purposive action", European Journal of Philosophy, 2001 in the back of my mind).

In this layered picture, it wouldn't seem dead wrong to say that the bodily movements are intrinsically mindless, but considered in their context, they are still purposive (a dichotomy between minded and mindless seems unhelpful though). And it would be a reductive account, although it would proceed in two steps (intentional action-->intention+purposive behaviour, purposive behaviour-->teleological mechanisms+movement).

Hmm, anyway, do think such a reductive account is part of your target in the paper? Or would you say that telelogically purposive movements are precisely the kind of embodied cognitive events that you are talking about? Or something else?

Best
Olle

[Markos Valaris]

Thanks Olle, this is a great question. I did not have this view directly in mind in the paper.

In some ways I am very sympathetic to this picture: the instrumental structure of a swim is a matter of various motor skills being exercised for the purpose of swimming, where this is not to be explained by some kind of mental plan or blueprint standing behind the exercise of those skills but is. So there is a structural similarity with the way we individuate, e.g., a process of digestion: this, too, is a teleological process in which various capacities are recruited towards a common goal, where that is not to be explained in terms of a blueprint somehow standing behind the entire process.

But if we leave at that, proponents of the standard model will have a very reasonable complaint: what is the difference then between lower-level biological process and actions (which are simply higher-level biological processes, after all)? Intuitively, the difference is in the involvement of higher cognition and rationality. Normally, you are expected to know what you are doing and why, in a way that you are not expected to know anything about your digestive processes. Partly for this reason, you are also held responsible for your actions but not your digestive processes.

Now, the standard story of action tries to accommodate this point about the cognitive dimensions of action in terms of mental blueprints, which is exactly the picture that I reject. But, on my view, the answer to this is not to reject the requirement of cognitive involvement in action; it is rather to expand our view of what can count as cognitive, to encompass physical skills as well.

On the question of reduction, I would not assume that the picture I propose is inherently inimical to *all* sorts of reduction. It relies on a notion of teleology that does not reduce to being guided by a plan or blueprint, and a notion of content suitable to be “embodied”. Perhaps you can go at least some way towards reducing both of these in teleo-functional terms. (I suspect full-on reduction is not required for a sufficiently broad naturalism, but that’s more of a predilection than a developed position!)

[Olle Blomberg]

Thanks for your reply Markos. I think I wasn't very clear in presenting the alternative picture. My talk of it as being "layered" and the reference to Butterfill's paper were perhaps misleading (the reason I thought of Butterfill's paper was just that he combines ideas from teleological and causal theories, and the lesson I took from this is that there is no reason to see such ideas as being mutually exclusive).

On the alternative picture I had in the mind, as in the standard story of action, what sets intentional actions apart from bodily movements that aren't intentional actions is something that is *added to* those movements, namely an intention or a belief-desire pair. So, in this way, it is similar to the standard story. However, on the alternative picture, these bodily movements are not (always?) "mere" or "brute" bodily movements, but rather purposive bodily movements that have an instrumental structure (where the instrumental relations exist in virtue of "digestion-like" teleological directionality). Let's call this the standard+ story of action. So, the intentions are still crucial. They allow us, I think, to address the complaint you mention: What enables the proponent of the alternative picture to distingusih the bodily movements that are actions (e.g. Raji's step) from those that aren't (Raji's increase in heart rate) are precisely what is represented by intentions or other mental items. Raji's step is an action because it is aimed at the outcome that is represented by Raji's intention, but the heart rate increase (or digestive bowel movements) isn't because it isn't represented by Raji's intention.

So my idea is that while you are probably right that subsidiary actions are are real problem for the standard story of action, there is such a notion of instrumental structure available to the standard+ story of action. Put differently, there is a notion of parthood that Mele can use (provided that he accepts the standard+ story of action). Of course, this doesn't address the worries you mention about agent's knowledge or responsibility, but I think it at least allows a kind of reductive account to address the interesting problem of subsidiary actions that you present. (Btw, I don't think intentions or beliefs and desires are necessary for agency--I think spiders exercise agency for example--so you problem of distinguishing subsidiary actions from digestive processes doesn't go away, but I don't think it need to be a worry for the standard+ story of action.)

On a different note, and regarding the following:

"Consider a particular movement of Bob’s legs as he is running his marathon. My claim is that *that movement itself* should be thought of as a cognitive event: it embodies Bob’s grasp of a way for him to run a marathon." (p. 15)

"Taking a single step is itself a complex action, composed of many smaller movements. Moreover, I contend that such movements—placing a foot in front of the other, shifting her weight forward as she lifts her other foot and swings it forward, etc.—are also intentional actions on Raji’s part." (p. 7)

Concerning this: Are you saying that *any arbitrary* part of the movements of Bob's legs is an intentional (subsidiary, or sub-subsidary...etc) action that Bob performs? Or is it only the proper parts that fill slots (so to speak) in the instrumental structure of the activity? I think your view is the latter, but I'm not entirely sure…

(Btw, I realise that I'm not actually discussing your post, but rather your paper. If it's more appropriate to email you with comments and questions instead of posting them here, the let me know....)

[Olle Blomberg]

A clarification: I think that the problem of distinguishing actions from digestive processes is important, but it is not crucial for the standard+ story of *intentional* action, since this isn't a story of agency in general. (And sorry about all the typos in my last post...but I hope you get the gist of it!)

[Markos Valaris]

OK, I think I get the picture now.

There is a lot to like in such a picture, but here is the sort of concern I would have. In the case of instrumentally structured intentional actions there is reason to think that the instrumental relation of the parts to the whole is known to the agent. Perhaps the agent cannot articulate it in any great detail, but if you ask her "what are you doing with your legs and why?", she can say something like "I'm moving them like this because I'm walking to the kitchen". Moreover it seems to be no accident that she can say this; it seems essential to our taking the whole performance to be an action.[1]

I am not sure if this can be captured by the standard+ story. As I understand it, the standard+ story requires an overarching intention, and then a teleological mechanism that implements it; but the workings of the mechanism need not be available to the subject in a first-personal sort of way (at least this is what the analogy with digestion would suggest).

I suspect that the gap between the two approaches could be bridged by a suitable conception of know-how or skill, as distinct from e.g. your capacity to digest your food. I am happy with this more restricted version of the standard+ view; but then I doubt that you can give a fully reductive account of know-how! (This then would be the sort of non-reductive causal theory that I set aside at the beginning of the paper.)

[1] Incidentally, this is also how I would answer your other question, about which movements count as instrumental parts of an action.

[Olle Blomberg]

Thanks Markos. Yes, I can see that this might be a problem for standard+. Although I'm thinking that the overarching intention could give Raji non-observational knowledge that she is moving her body in the way required for walking. Perhaps that doesn't provide her with the fine-grained "grasp" that you are after though.

Anyway, here's a reason one might prefer the standard+ story. Instrumental relations between actions do not only exist within intentional singular activities, but also for intentional joint activities involving several agents. And one might be suspicious of the idea that jointly acting agents have the requisite non-observational grasp of these relations that you think is necessary for intentional agency. Although I should say that I'm open to the possibility that such a suspicion is actually misplaced or unfounded (as it happens, this is something I'm planning to do research on shortly).

[Markos Valaris]

Thanks, this is quite interesting. I have not thought about joint intentions or collective agency much. My initial instinct would be to try to push for a similar instrumental knowledge on the part of corporate agents as well, but I haven't thought any of this through. We do speak of corporates as possessing know-how; so perhaps there is also a sense in which when they put their know-how to use the should also count as knowing what they are doing. I find the area of group cognition and agency fascinating, but so far haven't found the time to really study it.