The ‘‘self’’ of self-control

6 The ‘‘self’’ of self-control
What is the self? There is no single or simple answer to that question.
The term, which is a term of art (as a freestanding noun, the word
‘‘self’’ is hardly ever used in ordinary English) is used in a range
of sometimes conflicting ways; so many that some philosophers
despair at finding any unifying element underlying its disparate uses
and therefore recommend that we discontinue using it at all as the
name for a philosophical problem (Olson 1999). I suspect that these
philosophers are right to this extent: there is no single sense or even a
closely related set of senses, unifying all or even most of the central
uses of the term. There is therefore no problem of the self. Never-
theless, there are a number of problems of the self. In this chapter, we
shall consider just one.
The problem of the self upon which I want to focus is, briefly,
this: why is the self singular? This is not the first problem that
generally comes to mind when we consider the problem of the self.
However, it is a problem that is especially pressing for us. Consider
the evidence from neuroscience, reviewed in earlier chapters, that
the mind consists of a multitude of discrete modules and mechan-
isms. Consider, too, the extended mind hypothesis, according to
which mind can include or incorporate a set of tools external to the
self. Given that the mind consists of a motley collection of tools and
mechanisms, why is there a self at all? Why is this self experienced as
singular, and is this experience veridical?
Even this question, why is the self singular, can be broken
down into further, only tangentially connected, problems. We might
enquire into the so-called unity of consciousness; that is, into the
question of why our experience of our minds is of a single stream of
consciousness. That’s not the question I want to explore here,
interesting though it is. Instead, I want to focus on human behavior:
why, given the fact that our brains consist of many mechanisms, do

we act in such a well-coordinated manner? Briefly, the puzzle is this:
brains consist of many modules, each of which does its own thing;
many of these modules drive behavior, directly or indirectly (indir-
ectly: for instance, by producing desires or inclinations). Yet our
behavior looks remarkably consistent. How is the coordination pro-
blem solved? And (more importantly from the point of view of neu-
roethics), what light does the problem and its solution shed on how
we normally act, and on the range of pathologies of action to which
we are subject?
Is there a problem here at all? It might be thought that any
coordination problem would have been solved by evolution. After all,
all the modules which together constitute my mind/brain are in the
same boat: except in very rare circumstances, modules cannot behave
in ways that benefit themselves, from an evolutionary point of view,
without also benefiting the whole ensemble of modules. Defection
from a cooperative strategy will be punished, almost always, by a
lack of success at passing on the genes which promote such defec-
tion. Accordingly, we should expect the modules to work together.
They may be analytically separable, and they may dissociate either as
a result of brain injury or of clever laboratory manipulations, but in
the real world they will always work together.
This line of reasoning has a lot going for it. We ought to expect
that evolutionary pressures will ensure that the major coordination
problems have been solved. Indeed, the brain itself has a number of
mechanisms to ensure that discrete modules work together. There are
identifiable sites within the brain where information and representa-
tions from diverse sources are integrated. For instance, very basic bodily
information, from various sources, is integrated in the brain stem,
providing the basis for what Damasio (1999) calls ‘‘the proto self’’
(Churchland 2003). There is also evidence that emotions play a neu-

robiological integrating role by coordinating brain plasticity (Ledoux
2003). Neuroscientists have made significant progress in recent years
the ‘‘self’’ of self-control
198
in identifying mechanisms which contribute to solving the so-called
binding problem, the problem of how information from different
sources – about the shape, the color and the position of objects, for
instance – is integrated into a single complex perception. Temporal
synchrony seems to play a role in binding representations together
(Engel et al. 1999). Such representations enter, or are poised to enter,
consciousness, and therefore can play a role in guiding behavior that
is an intelligent response to information from many sources.
However, though some degree of unity is guaranteed by these
mechanisms, there is strong evidence that the binding mechanisms
do not provide us with the optimal degree of unity we need in order
to pursue the kinds of lives which we value. The unity they provide
us with is unity only over the briefest stretch of time, whereas the
kind of unity we need, to pursue fully human lives, is long-term
unity. I shall briefly sketch the reasons why we need such unity in
order to pursue the best kind of life, before turning to the evidence
that such unity can be lost, even in the absence of neurological
injury.
What sense of ‘‘self’’ is at issue in the question concerning the
unity of our behavior? The self in question is the self of self-control.
Now, self-control is a puzzling idea. When we say that someone has
lost self-control, we don’t mean that someone else is controlling
them. But if they are acting intentionally and voluntarily, and no one
else is controlling them, then they must be controlling themselves –
or so it seems. How should we make sense of this puzzle?
Consider an ancient paradigm of the loss of self-control, from

Plato’s Republic. Plato tells us the story of a certain Leontion, who
was outside the city walls when he noticed a pile of corpses, the
bodies of executed criminals. Leontion was fascinated by the corpses
and drawn to look at them, but at the same time he was repulsed and
disgusted, by the corpses and by his own desire:
For a time he struggled with himself and covered his eyes, but
at last his desire got the better of him and he ran up to the
the ‘‘self’’ of self-control
199
corpses, opening his eyes wide and saying to them, ‘There you
are, curse you – a lovely sight! Have a real good look!’
(Republic 440a).
Leontion lost his self-control. Yet his actions were not controlled by
another person; he acted, as Plato himself says, on his desire. It was
his desires that ‘‘got the better of him.’’ Leontion’s predicament
captures the essence of a loss of self-control. When someone loses
this kind of control, they act as they want to. However, their action
does not reflect their self-image.
That is not to say that conflict with one’s self-image is sufficient
for loss of self-control: after all, one’s self-image could be the product of
self-deception. Someone might regard themselves as trustworthy,
reliable and upright, and yet constantly behave in ways that are self-
interested, at the expense of the legitimate moral interests of others.
Such a person does not seem to suffer from a loss of self-control. In
addition to failing to cohere with one’s self-image, behavior which
reflects a loss of control also fails to cohere with a person’s values and
endorsed desires; values and desires that are actually in control of that
person’s behavior for significant stretches of time. Just what propor-
tion of my behavior must actually reflect a set of my values for those
values to count as mine I leave open; it seems likely, however, that at

least half of my behavior must be in accordance with my values or I am
wrong in thinking that they are my values.
I control myself, roughly, when my actions stem from my non-
self-deceptively endorsed values. Typical cases of loss of self-control
fit this mould: the woman who loses her temper and says something
she later regrets, the man who departs catastrophically from his diet
or from his vow to limit his drinking; the kleptomaniac who steals a
trinket they do not value or the heroin addict who feels powerless to
give up their drug; all these individuals control their actions, but
their behavior does not reflect their deepest values.
Now, why does it matter whether our behavior reflects our
deepest values? Why do we value self-control? As the examples of its
the ‘‘self’’ of self-control
200
loss we have just reviewed make clear, self-control is instrumentally
valuable to pursuing the kind of life we want. When we lack self-
control, we may find ourselves at the mercy of passing fancies. If I
cannot control myself, I cannot trust myself; to stop at one drink at
the bar, to pass on dessert, to hold my peace when someone says
something I find irritating. If my self-control is very badly damaged, I
will be stimulus-driven, and unable to follow a coherent life-plan.
Consider, first, some spectacular and pathological cases of loss of
self-control. People who suffer from utilization behavior, a disorder
caused by damage to the frontal lobes, respond compulsively to the
affordances – the suggestions for use – of objects in their immediate
environment. For instance, if a glass of water is placed in front of
them, they will typically raise it to their lips; if spectacles are placed
in front of them, they will put them on – and if a second pair is then
offered, they will put those on as well, despite the fact that they are
already wearing a pair. They will continue to respond in this way,

even after they are instructed not to (Estlinger et al. 1991; Lhermitte
et al. 1986). These patients seem literally unable to help themselves,
as do sufferers from imitation behavior, a social form of utilization
behavior. Patients with this condition will imitate an examiner’s
movements even when told not to and given negative reinforcement
(Lhermitte 1983).
Utilization behavior may, if it is severe, be incompatible with a
decent human life. Sufferers cannot count on their ability to carry
out plans, without being distracted from them by extraneous features
of their environment. But even when self-control is within the nor-
mal range, we may find its temporary loss a significant problem. In
the contemporary world, the ideal of authenticity is extremely
powerful. As a consequence, most of us believe it is very important to
pursue our own conception of the good life: an overarching image of
what it means, for us, to be human. We want to live a life that
expresses our central values, and we want that life to make narrative
sense: we want to be able to tell ourselves and others a story, which
explains where we come from, how we got to where we are, and
the ‘‘self’’ of self-control
201
where we are going. Indeed, as Marya Schechtman (1996) has agued,
human beings typically understand themselves and each other in
narrative terms; our very identity (in one sense of that term) is con-
stituted by the contents of a (largely true) narrative we tell about
ourselves. But imposing narrative unity on our lives requires that we
be able to shape our behavior, at least in its most significant respects,
in the light of the values we want our narratives to reflect and
express. Losing control over oneself threatens that ability; it may
result in our narrative taking turns we cannot endorse, or in constant
disruptions to its evolving arc.

A significant degree of self-control is also required for people to
live together in harmony and for the achievement of the kinds of
goods that human beings can only realize in society. I can only make
promises to others if I am able to ensure that I am (usually) capable of
behaving in the future as I now desire; in other words, if my present
self can exert a significant degree of control over the behavior of my
future self. If we are to coordinate our actions, and therefore if we are
to be able to realize the goods which come from divisions of labor,
then we must be able to trust one another to deliver on our word. If
we are to engage in intimate relationships, then we must coordinate
our activities, divide responsibilities and reliably be there for one
another. All of these activities take a relatively significant degree of
self-control.
Patients suffering from utilization behavior or similar disorders
may have to engage in complex calculations to prevent their behavior
disrupting their lives. They may have to carefully structure their
environments to enable them to carry out complex activities without
interruptions. Of course, the kind of inability to inhibit responses
characteristic of the frontal lobe patient is well outside the range of
ordinary experiences of loss of self-control; when you or I find our-
selves acting against our own all-things-considered judgment, we do
not experience ourselves as stimulus driven. However, it may be that
at least some ordinary losses of self-control can be illuminated by
these neurological disorders. Utilization behavior may be explained
the ‘‘self’’ of self-control
202
by a dysfunction of a specific inhibition mechanism: on this
hypothesis, the initial response of sufferers to objects and their
affordances is entirely normal. They, and we, respond to such affor-
dances automatically; the difference between us and them is that in

us the frontal lobe mechanisms which act to inhibit inappropriate
responses are intact, whereas theirs are not (Archibald et al. 2001). It
may be that losses of self-control within the normal range of human
experiences also reflect losses of, or weaknesses in the mechanisms
of, inhibitory control. Such losses could come about in a variety of
ways: developmentally, either as a result of an environment which
does not encourage its proper development or as a consequence of
innate factors, or as a response to environmental stressors which
temporarily overwhelm inhibitory resources. Some researchers have
suggested that common self-control problems, such as ADHD, might
be traced to dysfunctions in the same mechanisms that underlie
utilization behaviour (Archibald et al. 2001).
the development of self-control
Though the elements of a solution to the problem of coordinating the
various mechanisms that together constitute our minds is part of our
evolutionary heritage, nevertheless the extent to which self-control
is developed varies from person to person. We each need to learn to
control ourselves, as part of normal development, and some of us
learn the lesson better than others. Walter Mischel and his colleagues
have been gathering data on the development of self-control for
decades. Mischel’s team developed an experimental paradigm to test
children’s ability to delay gratification. Children were offered a
choice between two alternatives, one of which they valued more than
the other (for instance, one marshmallow versus two). The experi-
menter left the room, telling them that if they waited until he or she
returned, the child would get the more highly valued reward, but that
they could call the experimenter back at any time (by ringing a bell)
and receive the less highly valued reward at once. Children differed
greatly in their ability to wait for the second reward (Mischel 1981).
the development of self-control

203
Mischel’s studies have demonstrated a number of important
points. Perhaps most interesting is the discovery that ability to delay
gratification at age four is strongly predictive of a range of desirable
characteristics in adolescence and later: academic success, social
competence, attentiveness, concentration, and the ability to form
and execute plans (Mischel et al. 1989). This finding seems to con-
firm the claims made above, about the instrumental value of the
ability to control oneself for the achievement of goods we value. The
child, adolescent or adult who cannot exert a sufficient degree of
control over him or herself cannot successfully pursue academic
excellence (which is a project pursued over time and in which results
gradually accumulate) or prevent themselves being distracted by
immediate gratifications incompatible with success.
Mischel also found, more surprisingly perhaps, that the ability
to delay gratification seemed to depend in very important part on
the deployment of a set of skills. Rather than self-control simply
depending upon a mysterious faculty of ‘‘will-power,’’ the ability to
delay gratification depends crucially on self-distraction. This finding
was precisely the opposite of the experimenters’ hypothesis: rather
than focusing upon the reward for waiting, the good delayer thinks
either about something else entirely, or focuses upon aspects of the
reward that are not linked to the ways in which it is rewarding (for
instance, rather than think of the sweetness of the marshmallows,
the child thinks of them as fluffy white clouds). It seems that
focusing on those aspects of the rewards that make them desirable
ensures that self-control resources are overwhelmed relatively
quickly (Mischel 1989). Moreover, children who are good delayers
deploy these strategies spontaneously, and by the age of five under-
stand the kinds of strategies which are effective. Self-control is, or

depends upon, a set of skills and that suggests that it can be taught
and learned. In any case, Mischel’s work has shown clearly that the
coordination problem is not solved for us; though we are each
minimally unified simply by virtue of neurobiological mechanisms
that ensure we can function as organisms, we do not achieve an
the ‘‘self’’ of self-control
204
optimal, or even (for living a decent human life) a sufficient level of
unity of behavior without further effort and learning.
Further evidence for the all too grim reality of failures of self-
control is provided by a range of problems to which people are subject,
even in the absence of neurological damage. Most obviously, the
disorders on the obsessive-compulsive spectrum (obsessive-compul-
sive disorder (OCD), kleptomania, trichotilliamania, and so on) can be
understood as failures of self-control. Self-control in OCD, klepto-
mania or intermittent explosive disorder can be undermined in one or
more of several ways. One possibility is that such disorders make the
self-distraction strategies utilized by the children in Mischel’s studies
harder to deploy – perhaps by forcing the (subjectively perceived)
desirable qualities of the objects of the compulsions or obsessions to
the attention of the agent. If, as I shall shortly suggest, self-control is a
depletable resource, and it is depleted by focusing on these qualities,
then concentration on these qualities will quickly lead to its
exhaustion. Alternatively, OCD and other disorders might impact
directly on our reserves of self-control: on this hypothesis, agents’
abilities to deploy self-control strategies is unimpaired, but because
their self-control resources are already depleted, they cannot resist for
long. Of course, both explanations might work in concert to produce
failures of self-control. (Note, however, that neither explanation, nor
a combination of both, can be a full explanation of these disorders.

The hypotheses attempt to explain why sufferers give in to their
desires; in addition, we need an explanation for the sometimes bizarre
content of the desires: why do sufferers from trichotilliamania
experience the urge to pluck out their hair? Why do many sufferers
from OCD derive a brief respite from performing ritualistic actions?)
The other class of disorders of self-control to which human
beings are all too prone involve failures appropriately to regulate
consumption of rewarding substances, or engagement in rewarding
activities: food, alcohol, sex, drugs. We often attempt to explain these
disorders simply by invoking the notion of addiction, as though
that put an end to all further questions. But, while it is clear that
the development of self-control
205
addiction genuinely exists, and that it plays a role in explaining the
behavior of addicts, it is far from a complete explanation by itself. In
fact, I shall argue, the same basic mechanism is at work in all failures
of self-control, from disorders on the obsessive-compulsive spectrum
to addiction, from inability to delay gratification to failures of will-
power in dieting.
ego-depletion and self-control
Let’s begin with the reasons for the failure of the addiction hypoth-
esis as a complete explanation for the behavior of addicts. The
addiction hypothesis, in its most usual form, postulates that addic-
tive desires are literally irresistible. On this view, addiction destroys
agency itself: addicts do not choose to consume; they are impelled by
their addiction. The only difference between the addict and, say, the
man who, in Aristotle’s example of non-voluntary behavior, is car-
ried somewhere by the wind (Nichomachean Ethics, 1110a) is in the
location of the force that acts upon them: the wind impels from
without, while addiction impels from within. But in both cases, the

person is carried away, regardless of their wishes or beliefs.
This is the picture of addiction so eloquently expressed by
William James:
The craving for a drink in real dipsomaniacs, or for opium or
chloral in those subjugated, is of a strength of which normal
persons can form no conception. ‘Were a keg of rum in one
corner of a room and were a cannon constantly discharging balls
between me and it, I could not refrain from passing before that
cannon in order to get the rum;’ ‘If a bottle of brandy stood at one
hand and the pit of hell yawned at the other, and I were convinced
that I should be pushed in as sure as I took one glass, I could
not refrain:’ such statements abound in dipsomaniacs’ mouths.
(James 1890: 543)
James wrote more than a century ago, but the view he espoused is
still common. It dominates the common imagination, and it has
the ‘‘self’’ of self-control
206