Symposium on Joshua Shepherd’s “Halfhearted Action and Control”

Welcome to our fifth Ergo symposium. This week we are showcasing Joshua Shepherd’s paper “Halfhearted Action and Control”, with commentaries by Andreas Elpidorou (Louisville), Nora Heinzelmann (Munich), and Zachary Irving (Virginia). Let me begin by thanking all the participants for their great work!

Shepherd introduces his topic with a possibly familiar scenario: “It is 3 p.m. You are grading papers, as you must. But you don’t really want to. It is sunny, and you want to be in the park. Your legs hurt, and you want to take a walk. You await an important e-mail, and you want to check for it. You both want to do something else, and you do not want to do the thing you are doing. You grade the papers intentionally. But you also grade them halfheartedly” (259). What is it, though, for you to act halfheartedly? Shepherd claims that it is for you to act with weak overall motivation.

Shepherd begins in §1 by explaining what it is for an agent to possess some overall level of motivation to act in a certain way and what it is for an agent’s overall level of motivation to be weak or strong. According to Shepherd, an agent’s overall level of motivation with respect to some action, A, consists in a complex set of dispositions, including especially dispositions to initiate, sustain, and guide A-ing (262, 264-5). And, on his view, to say an agent’s overall motivational level with respect to A-ing is weak (or strong) is to apply a gradable predicate mapping the agent’s overall level of motivation to A onto a scale, where location on the scale is determined relative to a specific class and standard of comparison (263). As Shepherd explains, differences in the class and standard of comparison can yield different theoretically illuminating measures of strength or weakness of an agent’s overall motivation with respect to an action (264). For example, one’s comparison class could be the agent’s overall motivational level for a certain type of action across a wide range of different physical conditions. This would provide a measure of how strongly motivated an agent is during a particular performance as compared to her other performances of the same act-type. Alternatively, one’s comparison class might be an agent’s typical overall motivational level for all of the act-types of which she is capable, yielding a measure of how strongly motivated the agent typically is to perform a certain type of action (e.g., grading) compared to other act-types (e.g., playing with one’s dog). Similar flexibility is permitted for one’s standard of comparison—e.g., one might focus on how overall motivation modulates an action’s initiation, sustenance, or guidance (265). For the remainder of the paper, Shepherd focuses on the relation between the agent’s overall motivational level to A and A’s guidance as his standard of comparison.

Next, in §2 Shepherd argues that a central manifestation of being weakly motivated to A is that an agent possesses less control when intentionally A-ing than she would possess were she more motivated to A. Appealing to his earlier work on agential control, Shepherd suggests that an agent possesses control over an action, A, to the degree to which the agent can repeatedly and flexibly bring about a match between her intention to A and her A-ing, where repeatability makes one’s success reliable rather than merely lucky or one-off, and flexibility makes one’s success not merely situation-dependent and rote, but robust in the face of environmental variance. With this as his model of control, Shepherd marshals empirical evidence suggesting that that being weakly motivated to perform a task reliably impairs agents’ control over task execution—specifically, by increasing proneness to distraction (268).

In the third and last part, Shepherd identifies a puzzle about intentional agency during knowingly halfhearted action. To illustrate, consider the dancer, Arezoo. At the end of a long day of practise, Arezoo attempts a routine she knows to include a concentration-demanding maneuver. She also knows that she is weakly motivated and won’t execute the maneuver successfully unless she effortfully increases her concentration. She performs the routine but, because she is weakly motivated, she poorly executes the concentration-demanding maneuver. Shepherd proposes that, at least on the face of it, a knowingly halfhearted action like Arezoo’s implies rational incoherence on the part of the agent. Specifically, it seems to require that the agent simultaneously commit to a certain action-plan (e.g., embodied in Arezoo’s intention to perform the dance routine) and knowingly accept deviating from that action-plan (e.g., embodied in Arezoo’s awareness of her own halfheartedness prior to executing the routine). While Shepherd acknowledges that there are various strategies one might pursue in order to explain away the appearance of rational incoherence, he points out that such strategies have yet to be explored systematically.

You can find the target article, commentaries, and Shepherd’s response below.

Target Article:

Joshua Shepherd: Halfhearted action and control


Andreas Elpidorou: The costs and benefits of doing something halfheartedly

Nora Heinzelmann: Motivation and control

Zachary Irving: Halfhearted creative action: A puzzle about control

Author’s Response:

Joshua Shepherd: Replies for Brains Symposium on ‘Halfhearted action and control’


  1. Wesley

    This seems like a reasonable enough account to me, but I was left wondering what it is an account of, exactly, and what it’s ultimately answerable to. For example, is this an analysis of the concept answerable to how the word “halfhearted” strikes us, a prescriptive account of what such a concept should be like, a theoretical construct posited to elucidate a set of philosophical phenomenon, something else?

  2. Josh

    Hi Wesley, good question.
    It’s more the theoretical construct idea that you suggest. In writing the paper, I was thinking there is an interesting phenomenon of acting with weak motivation. And it turns out to be constructed in all sorts of ways, since there are many ways that motivation might vary. And in thinking about motivation in action, there seemed to me to be connections to interesting issues in philosophy of action, ethics, epistemology, cognitive science. Many others have noticed such connections, of course – the paper was an attempt to say something about the phenomenon that made sense to me, and focused issues in a way I found fruitful, and that could highlight the phenomenon as of philosophical interest. One thing I’ve really appreciated about the comments Andreas, Nora, and Zac offer is that they suggest what I’ve (somewhat artificially, of course) called halfhearted action is interesting for reasons I suspected, and for reasons I hadn’t envisioned.

  3. Aaron Henry

    Hi Josh,

    I was wondering if you might clarify or expand on your response to Nora’s proposal that halfheartedness in action does not consist in possessing less control over an action, but in allocating less control to it or exercising less control over it. Intuitively, Nora’s distinction seems important. For example, one might expect it to have normative implications. (E.g., it intuitively makes a difference to responsibility whether an agent acting with weak motivation lacks the ability to perform any better than they are performing or instead possesses the ability to do better but doesn’t exercise the ability). I’m curious to hear your thoughts, because your response to Nora seemed to suggest you think the distinction between not possessing control and possessing control but not allocating it might simply depend on how we as theorists classify things—e.g., whether we regard processes of resource allocation as internal or external to control possession (where you are talking about ‘zooming in’ and ‘zooming out’). That made it seem (to my mind) like you regard the distinction between failing to exercise control and failing to possess control as relative to some choice of the theorist, and so not as substantive as one might initially have supposed. Is that right?

  4. Josh

    Hi Aaron, thanks for the question.
    In my response to Nora, I was trying to think through the process of allocating control (or, as I would prefer to think of it, control resources). I was quite tentative in considering various possibilities, in part because a lot of this looks empirical, and I’m not confident that one or another possibility is the best. I think a number of different process-types might be thought of as (contributions to) allocation of control resources. Perhaps the one Nora had in mind was explicit – something like a decision of which an agent is aware. That’s the one that most obviously has normative implications, I think.

    I don’t want to turn Nora’s suggestion regarding allocation into something trivial, or into a matter of theoretical choice. I would agree that in some cases, such a decision occurs, such a decision has normative (potentially both moral and epistemic) implications, and such a decision is important for explaining the agent’s performance in that case. I think we can accept that and also accept that the weak motivation plays an additional role in explaining lowered control in the same case. (It’s not like every aspect of the control an agent exercises is up to the agent in any robust sense.) Further, the weak motivation may play a normatively relevant role in explaining the allocation decision – e.g., if weak motivation makes it very unlikely that the agent will do anything other than decide to allocation less control. (And accepting that there are cases like this does not commit one to denying the existence of other cases in which control resources are not devoted to the task at hand because of processes of which the agent is unaware.)

    One thing perhaps worth highlighting about cases in which the agent explicitly allocates less control resources is this. Nora said that we could think of cases like the tired dancer as special cases of the puzzle of weakness of will. I do think akrasia is going on in those kinds of cases. But I think the puzzle I was trying to highlight is somewhat distinct, in that it depends upon the agent knowingly undermining her own commitment to her action plan. The difference is between acting against a belief about what is best to do, and acting contrary to an intention to A (or, if you are a cognitivist about intention, contrary to a special kind of belief . . . actually I think the puzzle might be even more interesting if you went full Marušić and Schwenkler about intentions).

  5. Aaron Henry

    Thanks, Josh. That’s all interesting and helpful for appreciating the spirit of your reply to Nora. If I understand you, you’re not downplaying Nora’s distinction, but noting that what we make of her proposed alternative—i.e., that weak levels of motivation lead to decreased control allocation *as opposed to* decreased control possession—will partly depend on empirical details surrounding the mechanisms of resource allocation (e.g., to what extent these processes are explicit or more automatic, etc.). If that’s right, then I think my original comment was indeed off-base.

    Regarding your last point, I’ll need to think more about that, as I’m still working on distinguishing your puzzle from the others that you mention in the text. I’m inclined to see your puzzle as a version of what you call a ‘failure of synchronic self-control’, and (I’m not sure if this is different) it reminds me of discussions of temptation in literature on diachronic rationality (e.g., Bratman, Holton). But I think that’s because I find it natural to imagine the tired dancer changing her mind before or during the difficult maneuver, thereby succumbing ‘in the heat of the moment’ to the temptation to do it halfheartedly. I take it, though, your idea is that, at least on the face of it, she never abandons her original intention to perform the routine well.

  6. Josh

    Hi Aaron,
    That’s right. A way out of the puzzle would be to say the agent changes their intention at the last minute (to an intention to A halfheartedly?). That doesn’t seem like the best way to understand the case (to me). Or, even if that can happen in some cases, a set of the problem cases seems to me to remain. (I’m not thinking of the agent having an intention to A well, only that intending to A embeds a commitment to execute the plan, and knowing one will deviate from that and acting anyway seems problematic.)

  7. Nora

    Dear Aaron and Josh,
    These are interesting thoughts and I agree — a lot more can and probably should be said about how exactly the differences between allocations of control and possession of control are to be spelled out, and a lot will depend on empirical matters. This is perhaps one useful lesson we can take home from this symposium (at least for me Josh’s response and the discussion here has given me a lot of food for thought about these matters). And, as you, Aaron, say, the issue is probably be quite important because of the normative strings attached.
    Regarding the puzzle, just one small additional comment. Josh, you say that, in contrast to the puzzle about akrasia, “the puzzle I was trying to highlight is somewhat distinct, in that it depends upon the agent knowingly undermining her own commitment to her action plan. The difference is between acting against a belief about what is best to do, and acting contrary to an intention to A (or, if you are a cognitivist about intention, contrary to a special kind of belief . . . actually I think the puzzle might be even more interesting if you went full Marušić and Schwenkler about intentions).”
    Well, if you take a somewhat non-standard view about what weakness of the will is, like Richard Holton, then you can say that weakness of the will just IS failure to persist in, and a fortiori act upon, a (previously formed) intention. That seems to be very close to your puzzle on half-hearted action. Of course there is a big debate whether this account of weakness of the will is plausible… (Your point in brackets is actually very interesting, that had not occurred to me.)
    Sunny greetings,

  8. Josh

    Hi Nora,
    Point taken about Holton and weakness of will.
    About the parenthetical point – yeah, I’ve been thinking about cognitivism lately. I do wonder how the cognitivist might construe this issue. I’m not a cognitivist, but I think it’s an interesting thing to try and work out.
    Let me know if your thoughts develop on these issues. I’m certain I haven’t nailed everything down in the best way regarding halfhearted action. So I’m indebted to you and the other commenters and discussants for pressing such interesting questions and lines of thought.

Comments are closed.