Skip to content

Neil Levy’s 2nd Leverhulme Lecture: “The Science of Self-Control”

Yesterday Neil Levy delivered the second of three Leverhulme lectures. The topic this time: “The Science of Self-Control.” In these lectures, Levy is setting two views against each other. The first is a view that emphasizes willpower – when tempted, one must grit it out. The second is a view that emphasizes self-management – the way to avoid temptation is to objectify ourselves, understand what triggers failures of self-control, and put ourselves in environments without temptation. Like Ulysses aware of the nearness of Sirens, we ought to find ways to tie ourselves to the mast.

What does science have to do with this? Levy argues that science is indicating the preferability of a self-management view.

Levy began by reviewing a leading view in the psychology of self-control – the so-called ego depletion view (associated most frequently with the work of Roy Baumeister and colleagues (e.g., Baumeister et al. 1998)). On this view, self-control (or willpower) is seen as a depleteable resource. Our capacity to exercise self-control is akin to a mental muscle – while practicing the exercise of self-control can enhance this capacity, there is only so much this capacity can do over short stretches of time. So, if you have had to use this capacity throughout your day, you will be more likely to give in to the next temptation that comes along.

Levy does not deny that we are able to exercise something like willpower in some circumstances. But the effectiveness of the use of willpower, he argues, depends heavily on the environment an agent is in. Even if we can sometimes use willpower to resist temptation, we ought to manage ourselves to ensure we are not relying entirely on willpower.

Levy’s rationale for thinking this is based partly in a different model of self-control – one Levy is more inclined to endorse (although Levy wants to tweak elements of this model in certain ways). This is the opportunity cost model of Kurzban et al. (2013). According to this model – very roughly – various psychological mechanisms are often engaged in scanning the environment and performing calculations about the value of engaging in one mental task or another. If the task or activity in which an agent is currently engaged is deemed less valuable than the next best alternatives, then these mechanisms generate a signal that agents will experience as the effortfulness of the current task. The aversiveness of this experience will tend to prompt a shift to an alternative task or activity.

On its own, this is not a model of self-control. Levy noted, however, that this model might be able to explain self-control failures as instances when the mechanisms gauging value in the environment disagree with one’s personal-level values. Suppose one is on a diet, and thus one highly values healthy food. The mechanisms don’t care. So if the mechanisms put a heavy emphasis on some task or activity (focusing on and eating a brownie) as opposed to some other task or activity (writing a blog post) one will be likely to experience writing the blog post as effortful, and to experience comparatively higher motivation to think about and eat the brownie.

In his lecture, Levy spent a good deal of time arguing that some version of the opportunity cost model can better explain self-control failures than can the ego depletion model. We know that people get worse at self-control tasks over certain stretches of time. The ego depletion model explains this in terms of the depletion of a limited resource. But the model has a difficult time specifying the nature and contours of the resource. According to Levy, one leading suggestion – that the resource is closely related to glucose levels – is not borne out by the evidence. While it is true that consuming sugar after a self-control task leads to better performance on a subsequent task, it is also true that simply swishing a sugary substance around in your mouth and spitting it out has the same effect. It seems that sensing the presence of glucose is enough to reinvigorate the ‘depleted resource.’

Levy thinks we should thus conceptualize the way we deploy self-control resources as a function of various computations psychological mechanisms make, rather than a function of an actual store of willpower. Following on the work of Kurzban and colleagues, a number of psychologists have made suggestions regarding the nature of these computations. To these, Levy adds his own.

Levy introduces the thought that self-control mechanisms display discount functions. When a reward is far away, low value is placed on it. But as a reward nears, increasingly high value is placed upon it. To this, Levy adds the thought that we should expect these mechanisms to be sensitive to environmental features. For example, self-control capacities are predicted by socioeconomic status (SES), which Levy hypothesizes marks a difference in environment (rather than genetics). Low SES children develop in non-stable, non-secure environments, in which there is little point to delaying gratification. The opposite is true for high SES children. So the discount functions of these groups operate differently. Even so, Levy argues, we should expect continued sensitivity to the environment.

What does this have to do with the effectiveness of sensing glucose in one’s environment? Levy speculates that sensing glucose operates as a signal of environmental richness, which signals little need to manage self-control resources (as Levy put it, after such a signal I am more likely to explore my environment, rather than exploit the resources at hand).

But why, then, does sensing glucose only help those who have just performed a self-control task? (If you have not recently performed a self-control task, and you sense glucose, your performance on an upcoming self-control task will not improve.) Levy suggests that perhaps the mechanisms responsible for calculating opportunity costs are normally down-regulated, and that engaging in a self-control task up-regulates them (makes them more active). What sensing glucose does is to down-regulate them again – to make agents worry less about utilizing self-control resources.

Levy concluded his lecture by emphasizing how these ruminations on the workings of various self-control mechanisms fit within his bigger picture. For Levy, the nature of the environment plays a major role in self-control success or failure. While we do have some willpower, our success at self-control is not explained or predicted by amounts of willpower, but rather by the environments in which we act. And we can manipulate these, namely, by self-management – by putting ourselves in stable and rich environments suited to our goals. But this, it seems, requires a fairly high degree of self-knowledge – understanding how our self-control mechanisms color the world with value, and thus how our motivational systems incline us to act (whether or not we want them to so incline us).


Baumeister, R. F., Bratslavsky, E., Muraven, M., & Tice, D. M. (1998). Ego depletion: is the active self a limited resource?. Journal of personality and social psychology74(5), 1252.

Kurzban, R., Duckworth, A., Kable, J. W., & Myers, J. (2013). An opportunity cost model of subjective effort and task performance. Behavioral and Brain Sciences36(06), 661-679.


Please follow this link to listen to the podcast of Neil Levy’s 2nd Leverhulme lecture:

Share on