The opportunity cost model proposed by Kurzban et al. is thought provoking, and we agree with much of it. It offers an ultimate explanation for why self-control seems limited, and it has the potential to move the field beyond simple and biologically improbable resource accounts of fatigue. However, we found the more proximal account of the limits of self-control to be lacking (see Scott-Phillips et al. Reference Scott-Phillips, Dickins and West2011). Specifically, the notion that opportunity costs drive self-control fatigue does not account for a number of relevant findings as they relate to the proximate processes underlying self-control and its failure. Most critically, the model's proximate account is based on a modern homo economicus that risks being just as inscrutable as the limited-resource model it is trying to replace. We discuss the strengths of the proposed model and its shortcomings, contrasting it with our own mechanistic revision of the limited-resource model of self-control (Inzlicht & Schmeichel Reference Inzlicht and Schmeichel2012).

We start by clarifying what we are and are not debating. We are not debating the consistent finding that engaging in self-control at Time 1 leads to declines in performance at Time 2. This basic effect has been replicated more than 100 times in independent laboratories across the world (Hagger et al. Reference Hagger, Wood, Stiff and Chatzisarantis2010a). It also maps onto the commonsense view that mental fatigue can lead to decrements in performance over time (Hockey Reference Hockey1983). We are also not debating the role of blood glucose as the physical resource underlying self-control and its depletion (Gailliot et al. Reference Gailliot, Baumeister, DeWall, Maner, Plant, Tice and Schmeichel2007). The mounting evidence points to the conclusion that blood glucose is not the proximate mechanism of depletion, even if the presence of glucose in the oral cavity can moderate the depletion effect (Hagger & Chatzisarantis Reference Hagger and Chatzisarantis2013; Kurzban Reference Kurzban2010a; Molden et al. Reference Molden, Hui, Scholer, Meier, Noreen, D'Agostino and Martin2012). What is debatable is the how of depletion. The dominant account of ego depletion (Muraven & Baumeister Reference Muraven and Baumeister2000) suggests that performance on self-control tasks decreases over time because it recruits and depletes a limited inner resource. Although results of many and varied experiments using the sequential-task paradigm are consistent with a limited-resource view, the resource in these studies is inferred, but never measured (Hagger et al. Reference Hagger, Wood, Stiff and Chatzisarantis2010a). So how does ego depletion work?

Kurzban and colleagues suggest that people engage in some complex, mostly unconscious calculation of the costs and benefits of continuing to pursue the current task versus the costs and benefits of pursuing some competing task. Some version of this view seems likely to be correct, but this account does not help us to understand or anticipate changes in the cost-benefit ratio. Nor does it explain why people sometimes engage in seemingly costly and effortful behavior following periods of high subjective effort; for example, going to lengths to aggress against others or to find and consume drugs (e.g., Muraven et al. Reference Muraven, Collins and Neinhaus2002; Stucke & Baumeister Reference Stucke and Baumeister2006). The proposed model also implies that people who monitor and who are generally aware of their phenomenological states should be especially likely to withdraw effort as subjective effort increases. But research has found the opposite: with people who are more self-aware being less influenced by previous acts of control (Alberts et al. Reference Alberts, Martijn and de Vries2011; Wan & Sternthal Reference Wan and Sternthal2008); with people more aware and accepting of their emotions particularly good at executive control (Teper & Inzlicht Reference Teper and Inzlicht2013); and with self-control fatigue being mediated by deficits in what can be construed as a form of self-awareness (Inzlicht & Gutsell Reference Inzlicht and Gutsell2007). These results are not easily explained by the opportunity cost model, but they can be explained by our own process model (Inzlicht & Schmeichel Reference Inzlicht and Schmeichel2012).

Like others (Botvinick et al. Reference Botvinick, Braver, Barch, Carter and Cohen2001; Strack & Deutsch Reference Strack and Deutsch2004), we construe self-control as being initiated by the competition between two opposing forces: the force that motivates the expression of an impulse versus the countervailing force that overrides the impulse. In this view, self-control fails after initial task exertions when impulses are relatively strong, when control is relatively weak, or through some combination of both of these factors. According to our process model (Fig. 1), self-control at Time 1 leads to shifts in motivation away from restraint and toward gratification, such that people become less motivated to control themselves and more motivated to self-gratify at Time 2. As part of this motivational shift, people pay less attention to self-control cues and more attention to reward cues. We also suspect as part of this motivational shift that people become less aroused by the prospect of goal failure or success and more aroused by the prospect of reward and immediate gratification.

Figure 1. The process model of self-control fatigue. Self-control failure tends to occur after initial self-control exertions because of shifts in motivation away from control and toward impulses and gratification. This shift in motivation consists of shifts in attention and shifts in emotional responding.

Our model is still preliminary, but it can accommodate data that give the resource model fits (e.g., Job et al. Reference Job, Dweck and Walton2010; Muraven & Slessareva Reference Muraven and Slessareva2003); it can also accommodate data that are left unexplained by the current opportunity cost model (e.g., Schmeichel et al. Reference Schmeichel, Harmon-Jones and Harmon-Jones2010). Whereas Kurzban et al.'s model is vague about how the calculation of utility changes over time, our model better specifies directions in the dynamics of “processing allocations,” by suggesting that it moves toward reward/gratification and away from conflict and further control. Most important, our model makes novel and testable predictions that run counter to the current model. Our model predicts, for example, that self-control at Time 2 can be maintained when people are given veridical negative feedback on their performance; it also predicts that increases in emotional acuity will increase, not decrease, control at Time 2.

More generally, we worry that the opportunity cost model makes a fundamental error: It assumes that people calculate costs and benefits in an objective, dispassionate manner. This hyper-rational view discounts seminal work in psychology on the follies of human decision-making (Kahneman & Tversky Reference Kahneman and Tversky1979) and modern economic takes on utility theory that allow for non-rational, hyperbolic discounting of the future (Ainslie Reference Ainslie1991).

We admire the authors' ultimate explanation for self-control fatigue, but we find that their proximate explanation falls short of accounting for observed patterns of data and is based on an outdated view of human rationality.