Experience as action

By Benj Hellie

Brains readers: sup. Lovely to be here — thanks to Gualtiero for signing me up. I’ll kick off my tenure here with the fifty cent tour of a paper I’m working on as part of work on my book MS, Conscious Life.

Title: ‘Experience as action’.
Thesis: the kinds of experiences are the kinds of actions.
Elaboration: (i) by “experiences” I’m talking about consciousness in general, not just perceptual “experience”; (ii) I mean that every action is an experience and every experience is an action (and “they” are of the same kind); (iii) I’m thinking of “actions” in the ordinary sense as including walks to the grocery store and writings of books; (iv) I think there is a “hard problem of consciousness”, in fact more than one, and my thesis is about consciousness in the sense that generates a hard problem — hence I think action generates a hard problem (because it = consciousness); (v) I’m not Alva, and I’m not saying that seeing is an action: in fact, I don’t think seeing is a kind of experience, though I do think looking is a kind of experience; (vi) the thesis is at the same level of “grain” as representationalism, the doctrine that the kinds of experiences are among the representational kinds: the difference is that I don’t think representation is especially important to consciousness, but instead think agency is all-important.
Argument for the thesis:
(1) K is a kind of action iff some possible K is attentive;
(2) Some possible K is attentive iff every possible K is conscious;
(3) Every possible K is conscious iff K is a kind of experience.
Conclude: K is a kind of action iff K is a kind of experience.
Central presupposition:
The “adverbial theory of attention”. I don’t think that there is any such thing (properly so-called, anyway) as “attending to a bottle you see”: instead, I think what you are doing is attentively looking at a bottle you see. So when I say that “some possible K is attentive”, what I mean is that in some possible cases, someone Ks attentively.
Have at it.


  1. Welcome to Brains, Benj.

    I’m having problems with (2)&(3).

    I’m disinclined to agree that there are non-trivial examples of Ks for which every possible K is conscious.

    (Trivial examples include conscious experiences. Arguably, “conscious experiences” may be read in a certain way wherein it’s trivially analytic that every possible conscious experience is conscious.)

    I’m having a hard time finding plausible that no possible experience is unconscious or that no possible action is unconscious.

  2. Ken Aizawa

    “every experience is an action … (iii) I’m thinking of “actions” in the ordinary sense as including walks to the grocery store and writings of books;”

    What do you say about reports of experiences while under complete muscular blockade, where actions in the ordinary sense of walking to the store are physiologically impossible?

  3. Yes, (3) is intended to be read as — not trivially analytic, but — following straightaway from some super-general metaphysics.

    So I’m thinking your complaint is with (2). You’re thinking a K can occur attentively despite some K being unconscious (where K is a “natural kind”). (If so it would seem that attentiveness isn’t a high grade form of consciousness, which is somewhat counterintuitive.)

    I wonder if you could give me an example of a token action that is not occurring consciously so that we can focus the discussion a bit — there are maybe three things that might be bumming you out, but instead of going for the full-court press I’ll let you speak your mind first.

  4. Ken — buncha possibilities.

    1. Ordinary actions are /included/ but perhaps not /exhaustive/ of the relevant class. Maybe there are some extraordinary kinds of action that we don’t have names for — maybe he’s performing some of those.

    2. Maybe he is performing some ordinary actions — but, sadly, the typical effects of those actions are masked by the muscular blocker.

    3. Inclined to regard reasoning and imagining as actions. Same for “looking” (which includes but isn’t exhausted by /looking around/ — pretty much what the psychologists would refer to as “focusing visual attention”) and “feeling” (which includes sort of “mentally probing the sensational condition of one’s body” — pretty much what the psychologists would refer to as “focusing proprioceptive attention”).

  5. Hi Bill, thinking dreaming is not an action, and also not an experience (similarly, think that *perceiving* is not an action). However, there are actions that one can perform *within a dream* that are experiences — looking at the bewildering scene, running from the threatening beast.

  6. Hi Arnold, the thesis is neutral on whether there are internal representations. It may be that whenever one acts, this involves an internal representation: if so, then I would think that whenever one has an experience, this involves an internal representation. But (1) whether this is so would seem to be an empirical question and (2) even if (in humans) it is biologically necessary that action is accompanied by representation, this might be metaphysically contingent.

  7. Benj, you say “… even if (in humans) it is biologically necessary that action is accompanied by representation, this might be metaphysically contingent.”

    1) As I understand your thesis, attention is an action/experience and is necessary for all other action/experience. If this is the case, isn’t the phenomenal experience of something to attend to (a representation) a *precondition* for attention?

    2) Could you elaborate on what you mean by the phrase “metaphysically contingent”?

  8. gualtiero


    (2) seems completely implausible, so I’m probably not understanding it.

    Is the following an instantiation of (2)?

    In some possible cases someone walks attentively iff every possible walking is conscious.

    I’m not sure what it means to say that walking is conscious. To the extent that I understand it, I take it to mean that the agent who is walking is conscious. But that is radically false. Even decerebrate cats can walk!

    Could you clarify things for me?

  9. Eric Thomson

    Benj: good to see you putting yourself out there at a serious blog rather than Pharyngula where the SNR gets sort of low.

    I have a few questions.

    1. Are you claiming that the three biconditionals follow from an analysis of the terms involved, or from stipulative definitions of the terms involved?

    2. Why make attention so central to your theory of agency? Walking to the store and driving a car are voluntary behaviors, involve some kind of agency. Can’t we engage in such actions while not attending? E.g., the action of driving I do all the time without paying attention to what I am doing, same with walking to the store. I am still doing both voluntarily, with agency, even if not attentively. When walking, the actual mechanics of walking is one of the furthest things from my (consicous) mind. Hence, this tight yoking of attention and agency is counterintuitive to me.

    3. On your claim (2), I worry that you are identifying consciousness too closely with attention, when in fact they can be disentangled. At least on the surface I have visual experiences of things in my visual periphery to which I am not attending. There is a lot of research suggesting that consciousness and attention aren’t the same, such as the seeming ability to attend to things that are not perceived (multiple lines of evidence are summarized in Kristof Koch’s review article Attention and Consciousness). What do you make of such research?

  10. 1) Not quite. The thesis is that some token actions/experiences are performed attentively, others inattentively; but attending is never the *kind* of any action/experience. So there is never any such thing as “attending to x”; the phenomena that go under this rubric are diverse, and include attentive performance of a range of acts (often perceptual acts, such as attentively looking at x).

    2) The thought is that even if it follows from the biology of humans that whenever we act, we represent, there might be some species that act without representing.

  11. Hi Gualtiero — thanks for the question, it’s a good one.

    that is indeed a good instance of (2); RHS gets a bit of clarification through its association with (3).

    On the poor decerebrated cats: one thought is that it is possible to “display A-ing behavior” without A-ing (somewhat inclined to endorse the converse as well). If so, I could say that the cats are merely “displaying walking behavior”.

    There’s some complexity and apparatus involved in making this seem like more than a desperate dodge. The first thing I would want to observe is that in the vast majority of ordinary cases we do not think it is possible (or at least have no clear cases of) display of A-ing behavior without A-ing. Imagine someone displaying the behavior of addressing a philosophy seminar without addressing a philosophy seminar, for example — really presses the boundaries of the conceivable. So at least a great number of instances of (2) seem to be true.

    Here’s another point: we seem to think of actions as associated with moral responsibility; but consider someone who, while sleepwalking, causes a death: it seems as if we would exculpate the person. So that seems to be a case where we would not qualify what happened as an action of walking, but as a mere display of walking behavior.

    If this is right then we might want to push through and grab (2) in full generality for sake of simplicity.

    I think there is an explanation of the fact, which you pick up on and which I grant, that there is some blurriness between A-ing and display of A-ing behavior. My thought is that concepts of action are based in the first instance in acquaintance with the first person: sort of how Chalmers thinks of concepts of qualia. But unlike Chalmers on qualia, I think these concepts have a priori entailments of (but not by) causal concepts — though there are nomologically necessary entailments back from the causal to the first-person concepts. So for most purposes, we can readily flip back and forth between the concepts of A-ing /properly so-called/ and the causal concepts of displaying A-ing behavior.

  12. Hi Eric, good to make contact in this haven of civility.

    (1) Not really thinking of any particular epistemic basis for the claims. More like, they seem initially plausible (or seem to be more or less equivalent to claims that are initially plausible) so let’s run with them.

    [Though hopefully not stipulative definitions! Then I would have established that we can so stipulate that ‘fnord’ and ‘yar’ are coextensive — I’m a bit too busy to worry about such gimmes.]

    (2) Indeed we can engage in walking to the store or driving inattentively. Absolutely the case that a lot of our actions are performed in the course of multitasking. Not claiming that /tokens/ are attentive iff agentive/conscious — rather, that /types/ are potentially attentive iff necessarily agentive/conscious.

    (3) Not exactly. Thought is that any kind of experience is potentially attentive. So if x is an F and attentive, x is an F and conscious; but sometimes x is an F and conscious but inattentive. Deny that seeing something is a kind of experience: when one sees a gorilla there is not thereby something it’s like to see that gorilla (one has to at least look at the gorilla).

    Concerning the research at issue, agree that o can occur consciously without occurring attentively, but deny that o can occur attentively without occurring consciously. Not entirely sure what is meant by claims about “attending to x”; since don’t think that perceiving x suffices for consciousness not certain what the relevance of such data would be; definitely inclined to think that one can attentively think about unperceived things; inclined to doubt that one can attentively look at unperceived things. Will check out the research, looks interesting; preliminary flip through suggests there may well be considerable fishiness in their use of ‘attention’.

  13. that’s the line, yeah.

    when I say that, I am of course making an appeal to a distinction that some readers of this blog might regard as objectionably fine-pointed. who cares about all this moral responsibility stuff, if the robot does XYZ?!?!?! — a not unreasonable question!

    of course consciousness and action are categories that we grasp before we come to the discussion. various research paradigms might find it of pragmatic value to manipulate our understanding of these categories in various ways to move forward with this or that project.

    that can be very useful, no slams against applied work or various forms of modelling intended. but the important thing is not to forget that we have done this. cos the risk is that we lose sight of the problem that got us interested in the first place: we wanted to understand the nature of consciousness but ended up with an understanding of “konsciousness”, a sort of theoretical notion, inside our system, that we need to bring back in contact with the notion of consciousness if we are to get a fully satisfying answer to our original question.

  14. Sorry for the delay in replying, Benj. I’ve been away in Vegas being very unphilosophical.

    So, the sorts of examples of unconscious experiences and unconscious actions I have in mind are the sorts that the Higher-Order Thought guys like to dine out on. Take, for example, the sorts of cases typically discussed in connection with the long-distance truck driver. There are some experiences (of lights changing, roads bending) and some actions (tapping on the brakes, down shifting) that sometimes occur attentively and sometimes not and sometimes occur consciously and sometimes not. This looks to be a rich source of falsifications of (2), (though I wouldn’t be at all surprised to find out that I’m reading (2) incorrectly). Anyways one sort of falsifying case might be as follows.

    On some occasion, TruckDriver attentively taps the brake. However, while some tappings are conscious, some are not. So it’s false that brake tapping is attentive if and only if every possible brake tapping is conscious, since there are brake tappings that are attentive even though there are some possible brake tappings that are unconscious.

  15. Hi Martin,

    It might, but I’m not sure.

    I guess the thing to focus on is the stuff about types v. tokens and kinds both natural and unnatural.

    Plausibly, my attempt to draw falsifying examples out of the truckdriver case can be done in terms of types instead of tokens. So, for example, my remarks can be restated in terms of the action-type, attentive brake-tapping.

    (Where there are tokens, there’s always some type or other to throw them under.)

    I anticipate that this particular type will be rejected as not forming a natural kind, which I’m happy to go along with. Since brakes aren’t natural, the attentive tappings thereof probably aren’t natural either. But the example arguably can be tweaked again, this time to something more plausibly hanging together as a natural kind. Try this: Instead of attentive brake tappings we’ve got attentive limb movements or something. . . .

    Anyway, I dunno. What do you think?

  16. Jason Leddington


    Interesting stuff. I’m sympathetic to the way you link attention and consciousness, but I’m puzzled about your take on “looking” vs. “seeing.”

    Consider Simons’ gorilla experiment on inattentional blindness. Suppose S fails to notice the gorilla on first viewing and then watches the video again and has the gorilla pointed out to her. It would be odd in the extreme for S to describe herself as having “seen” the gorilla but not “looked at it.” It would be much more natural for her to say: “I must have been looking right at the gorilla, but I didn’t see it!” In short, it seems that I can look at things that I don’t see, and that looking at X is necessary but not sufficient for seeing X. By contrast, you think that seeing X is necessary but not sufficient for looking at X. As this seems to abuse some ordinary and apparently unimpeachable ways of speaking, it seems to me that you have most likely gotten things the wrong way around.

    Consider the frequency with which people say things such as: “What do you mean you didn’t see him? You were looking right at him!” In this sense, looking at (or, say, watching) X seems to be merely a matter of having X “in view” in the right sort of way. For example, S counts as watching a movie just in case her gaze is directed at (just in case she’s looking at) the screen; and even if she later reports having *seen* hardly any of it (distracted, as she was, by thoughts of A), she presumably won’t deny having *watched* it. Similarly, suppose S is charged with watching the children: she’s *not* watching them if she’s in the other room, but she’s merely doing a *bad job* of watching them if she has them only distractedly in view (and so, doesn’t *see* what they’re doing).

    I’d be very interested to know how you handle such cases.

  17. Martin Roth

    Hi Pete,

    So let’s think about stroking one’s hair (I’m inclined to say that this is natural-kindy enough). There are attentive hair-strokings, but some possible hair-strokings are not done consciously. So (2) is false.

    The argument looks good to me, but I also suspect that I am misreading (2) (or am not following how some of the key terms are being used here).

  18. Hi guys, got away from the thread for a few days, sorry for the neglect.

    Inclined to think that a lot of talk of ‘consciousness’ is in fact talk about attentiveness. Dazed truckdriver and the like, absent-minded hair-strokings, etc — all cases of acting inattentively, rather than unconsciously.

    Main reason to say this is that it simplifies theory: not too incongruent with the pretheoretic sense of things, and (I’ve mapped it out, and can go into more detail if you’re interested) a view that grants 1 and the RL direction of 2 but abandons LR on grounds like the current looks /wacked out/.

  19. Hi, thanks for the question — interesting observation of the ordinary language; sounds pretty good to my ear. I’d be inclined to make a distinction and a re-interpretation.

    on the first: As Vendler noted, ‘see’ is used in a stative sense and an achievementive sense: the latter means ‘spot’ or ‘notice’, whereas the former concerns the kind of passive perceptual relation that forms the basis of spotting or noticing that I am thinking of.

    on the second: somewhat inclined to think that while they were looking in its direction, there’s a sense in which they weren’t looking “at” it — the latter seems to require a certain degree of tracking of the thing over time, perhaps, which is absent in this case.

Comments are closed.