Understanding the nature of human selfhood and personal identity has been a preoccupation of multiple fields of inquiry and many thousands of scholars. My own ongoing interest in that topic led me in the 1980s to the study of self-regulation, which I thought might provide important keys. Regulating itself is not just one more thing that the self does: It is often a basic part of almost everything the self does.
The central importance of self-regulation to the theory of self is, to be sure, hardly the only reason to study self-regulation. Many other researchers have been attracted by its practical and applied benefits. Inadequate, misguided, or otherwise failing self-regulation has been linked to a breathtakingly broad assortment of personal and societal problems, including overeating and obesity, cigarette addiction, alcohol and drug addiction, violence and crime, prejudice, underachievement in school and work, lack of compliance with medicine, inadequate physical exercise, debt and bankruptcy, spousal abuse, failure to save money, and premature death.
This chapter is intended to provide an overview of the evolution of my inquiries and thoughts about self-regulation. It will feature the strength model, which has gradually grown into a highly influential but also controversial account of how self-regulation functions. For a thorough treatment of the findings, theoretical developments, and rival accounts, see Baumeister and Vohs (2016).
Basic Idea and Early Thoughts
I started to study self-regulation as a potentially useful platform to help understand the self. Not having any clear understanding of self-regulation beyond the Carver and Scheier (1981, 1982) feedback loop, I began by reading lots of work. My practice in these things is to read more data than theory, so I mostly skipped other accounts of how self-regulation might work in process and instead read studies on people trying to quit smoking, overeating, spending too much, and the like. Along with two colleagues, we pulled the material together into a book, Losing Control: How and Why People Fail at Self-regulation (Baumeister, Heatherton, & Tice, 1994). In it, we concluded that many findings seemed to fit the idea that self-control depended on a limited energy supply akin to the folk notion of willpower, but certainly there was nothing conclusive, and other theories were still viable.
I completed the book while on sabbatical and sent it back to my graduate students who were working away, temporarily orphaned as it were by my (and Ticeās) absence. Mark Muraven was then a graduate student in our laboratory, and he decided to try some experimental tests of the idea. I do not remember how much we communicated on this, and so the first studies may have been mainly his work or may have been developed in discussions with Dianne Tice. But his experiments worked well, indeed better than his previous studies on other topics, and so he latched onto this line of work.
Muraven was indeed the first author on the original paper on regulatory depletion (Muraven, Tice, & Baumeister, 1998). It established the methodological basis for much subsequent work, which has come to be described as the ādual task paradigm.ā It involves two different tasks that both require self-regulation. The first one is used to deplete the participantās willpower, so to speak. The second one measures the consequences.
In our early writings (Muraven et al., 1988; also Baumeister, Bratslavsky, Muraven, & Tice, 1998), we presented the studies as testing three competing models of self-regulation. One was that it depends on a limited energy supply. Another was that it was essentially an information-processing system. The third was that it was a skill. All these views had been advocated in the literature (as reviewed by Baumeister et al., 1994), and all were inherently plausible. They made competing predictions about how people would fare on a second self-regulation task as a result of having already done a different one. As we reasoned, if self-control depended on energy or strength, then the first one would deplete it to some degree, so performance would suffer on the second task. If self-regulation were mainly a cognitive process, then the first task should activate the self-regulating modules in mind and brain, and so performance should be facilitated and improved on the second task. And skill does not change from one trial to the next but should only show gradual improvements over considerable practice. Our results have consistently favored the strength model, but the early studies were done with an acute sense that things could go differently.
Building the Strength Model
The idea that the self consisted partly of some kind of energy was quite radical in the early 1990s (and still is, in many quarters). Information-processing models reigned everywhere. To find a theorist who had taken seriously the idea that energy dynamics were central to understanding the self, we had to go back to Freud. Hence, we used the term ego depletion (using Freudās term āegoā to refer to the self) to refer to the state following exertion of self-control, marked by diminished capacity for further self-regulation. The idea was that some kind of energy resource gets depleted when one exerts self-control, leaving less available for further exertions.
The analogy of a muscle becoming fatigued was helpful from the start and over the years has been a useful heuristic. One exerts a muscle, and afterward it is tired and weak. In apparently similar fashion, one exerts self-control, and afterward the capacity for self-control likewise seems tired and weak. (Reports of fatigue and tiredness have been one subjective sign of ego depletion; see Hagger, Wood, Stiff, & Chatzisarantis, 2010, for a metaanalysis; Chapter 10.)
The muscle analogy was extended in two important ways in subsequent work. One is that the effects of fatigue show up long before the muscle is so exhausted that it is incapable of further responses. Athletes follow initial exertion by conserving their strength. In a similar fashion, self-controllers may follow initial exertion of self-control by conserving their remaining resources. The alternative (car engine) metaphor, that ego depletion indicates that the brain has run out of gas, does not correspond to observations. Muraven and Slessareva (2003) showed that people could overcome ego depletion if they were sufficiently motivated. Muraven, Shmueli, and Burkley (2006) provided abundant evidence that ego depletion effects are conservation effects: People skimp more on current self-control if they anticipate further demands for self-control, and so forth.
The other key aspect of the muscle analogy is that muscles can become strengthened over time by frequent exercise. The short-term effect of exertion is to tire the muscle, but after it recovers from fatigue, it is a little bit stronger. Muraven, Baumeister, and Tice (1999) contributed some preliminary findings indicating that people who exercised self-control regularly for 2 weeks (mainly trying to improve their posture as they went about their daily activities) showed improvements on laboratory tests of self-control (unrelated to posture). Their baseline performance was unchanged, but they showed slower decrements in response to ego-depleting tasks. In physical terms, the improvement was not so much in the muscleās raw power as in its stamina. Various other studies have shown improvements in self-control performance as a result of regular exercise (for a review, see Baumester, Gailliot, DeWall, & Oaten, 2006). Perhaps most impressively, Muraven (2010) showed that smokers were more successful at quitting smoking after having exercised their self-control āmuscleā for 2 weeks.
Beyond Metaphor? Glucose Dynamics
A failed study by Matt Gailliot opened a new perspective on ego depletion and suggested that it might be possible to move beyond metaphors of strength and energy to invoke genuine, observable physiological processes. Like presumably everyone else working on self-regulation, I assume that it has physiological bases, even though identifying the precise brain circuits has been remarkably difficult to do (see Chapters 13 and 14). The original hypothesis was as follows: Given that resisting temptation weakens self-control, then perhaps indulging in temptation would improve or strengthen it. Participants were depleted first, and then some had an ice cream treat before the second task. Consistent with that hypothesis, the ice cream did lead to improved self-controlābut so did an unappetizing milk shake, made with half-and-half (milk and cream), no real ice cream, and no sugar, and thus consisting of a large dose of unsweetened dairy glop. Thus the pleasure of eating the ice cream was not necessary to counteract ego depletion. Reflecting on this, Gailliot and I began to think; perhaps it was the calories? Calories from food are, after all, energy.
Gailliot began to read up on glucose. Psychologists had not found glucose interesting, but nutritionists had documented various changes in behavior associated with glucose problems, and many of these suggested low self-control. A review of this literature concluded that low glucose impaired self-control (Gailliot & Baumeister, 2007a). We also undertook a program of laboratory studies to illuminate how glucose might work in the laboratory.
The experiments yielded three tentative conclusions about glucose and self-control (Gailliot et al., 2007). First, blood glucose levels drop from before to after ego-depleting tasks. (This one has not consistently replicated, however, and we no longer believe this to be generally true; at least, the picture is much more complicated.) Second, low levels of blood glucose predict poor performance on self-control tests. (This confirms what the abundant data from studies of nutrition, diabetes, and related areas had already suggested.) Third, getting a dose of glucose counteracts ego depletion. This was a novel prediction (although in a sense the milk shake study had shown it), but it has held up well in subsequent studies in our and othersā labs.
The idea that glucose was the resource behind self-control, the active ingredient in willpower, was seductive, but almost certainly the picture is far more complicated than that. Molden et al. (2012)...