This edited volume discusses the legitimacy of satisficing as a rational
model for practical reason and for ethical deliberation. The idea of
satisficing was first conceived by Herbert Simon, who argued that the
maximization of expected utility in our choices was an ideal of rationality not
generally possible for us, because we lack the cognitive abilities and information
relevant for an accurate identification of the best alternative. Instead of
maximization, Simon suggests that rationality requires of us only to choose
among the alternatives that guarantee a satisfactory
outcome.

To illustrate this, consider the following example. Kate is offered
chocolates by a friend, after having had a nice meal and desert. She declines
the chocolate offer, because she ‘has had enough’. One might think that she is
satisficing, to the extent that she refuses an opportunity to maximize her
enjoyment. But this conclusion would be too quick. Kate might actually be optimizing from a larger perspective.
This would be the case if refusing the chocolates, ‘locally satisficing’, would
be a way of maximizing other goods, that is, for the sake of a ‘global optimum’
(this expression comes from David Schmidtz’s chapter on ‘Satisficing as a
Humanly Rational Strategy’). For instance, she may not want to take time off
her work or she may be watching her weight. This is the view that Michael Slote,
in his chapter ‘Two Views of Satisficing’, labels ‘instrumentally rational
satisficing’. On the instrumentally rational version of satisficing,
satisficing locally is done so as to optimize on a broader scale. Kate is thus
here using an instrumentally rational satisficing strategy.

For Kate to be using an intrinsically rational satisficing strategy, two
necessary conditions have to obtain. First, there should be no other expected
loss of utility with respect to another aspect of her life that is caused by
her eating the chocolates–i.e. Kate does not believe that eating chocolates is
unhealthy, she is not worried about her weight, her teeth are in a good state,
her work would not suffer from this break, etc. Second, Kate believes that she
would enjoy eating the chocolates, that is, there would be a gain in her
utility by eating the chocolates despite having eaten already enough. If these
two conditions are met, Kate is employing what Slote has coined an
‘intrinsically rational satisficing’ strategy. Proponents of the satisficing
model of practical reason would claim that she is perfectly rational in
employing an intrinsically rational satisficing strategy.

A slightly different example is Kate’s search for toothpaste. Here Kate
has to consider each toothpaste successively. There are thus here two
dimensions along which Kate can satisfice: she may satisfice by choosing the
second-best toothpaste or she can satisfice by deciding to stop considering
further alternative toothpastes. It is the latter dimension on which the
authors of the volume focus on. If Kate takes one of the first satisfactory
toothpaste she finds, is she acting rationally? After all, she fails to
maximize to the extent that she is not aiming at finding the best toothpaste.
But again, Kate’s behavior can be interpreted as an optimizing strategy, since by choosing the first satisfactory
alternative she is saving time and energy that she wants to use for more
worthwhile activities. For her to really satisfice, Kate must aim at finding a
satisfactory rather than the best toothpaste given her time constraints. Kate may even take more time to
satisfice, since there is no guarantee that a satisfactory toothpaste would be
easily found. Kate may have very demanding criteria of what a satisfactory toothpaste
is and may need to go to another shop in order to find a toothpaste that is
satisfactory for her. This highlights the fact that satisficing and maximizing
in cases in which the alternatives are presented successively have different
stopping rules. This is a point that Schmidtz make in his contribution
‘Satisficing as a Humanly Rational Strategy’. Schmidtz opposes optimizing and
satisficing on the basis of the kind of stopping rules they use. Optimizing
entails ending one’s search for alternatives when the best alternative has been
found. Satisficing entails terminating one’s search when a satisfactory
alternative has been found. Schmidtz argues that if one stops searching before
having surveyed all the alternatives one may still be optimizing by taking into
account time constraints. A real satisficer does not stop because looking for a
better alternative would take off some time, but because the alternative
currently considered is ‘good enough’.

In this collection, Michael Byron, James Dreyer, David Schmidtz and Jan
Naveson deny the possibility of ‘intrinsic rational satisficing’. These authors
would claim that, for Kate’s choice to be rational, it must be done for the
sake of a global optimum. There are different ways in which some of the authors
of this contribution have nevertheless tried to make sense of Kate’s
satisficing strategy. In order to justify why satisficing is rational, one can
start by mentioning that it is an appropriate way of taking quick rational
decisions, given that to take the best decision would often be out of our
reach. But the question remains why satisficing would be superior to an
optimizing strategy that takes into account time constraints. The three main
ways to make sense of such satisficing strategies are to interpret these cases
as cases of incommensurability, cases in which satisficing is virtuous, and
cases that apply the notion of supererogation to rationality. 

First, there are cases of incommensurability. Once it has been granted
that two alternatives are really incommensurable, that is, that we cannot
compare them using a single currency, there is no way we can maximize, since
there is no best alternative. Satisficing can make sense of this situation, to
the extent that it presents the two alternatives as satisfactory and allow to
randomly pick one of the two. Schmidtz describes this situation as lacking any
global optimum. Some alternatives may be so different in kind that, although we
could compare them, we could not say of any that it is superior or inferior to
the other. Moreover, incommensurability is also a feature of cases in which
different alternatives are roughly on a par, that is, neither of the options is
significantly better than the other. This seems to be the case for most of us
when we are looking for toothpastes.

Henry S. Richardson, in his chapter ‘Satisficing: Not Good Enough’
claims that satisficing is plagued with the same problem as maximization, since
it also relies too confidently on the use of a metric in order to understand
practical reason. Whereas maximizing converts everything into utility,
satisficing converts everything into preference satisfaction. Richardson
suggests that a better way of thinking about practical reasoning is to look at
how practical commitments function in deliberation. Richardson believes
satisficing is wrong-headed because it converts alternatives into a metric,
although a simple one (satisfactory/ unsatisfactory).

In ‘A New Defense of Satisficing’, Michael Weber claims that
incommensurability enters the picture at another level. He introduces a
distinction between the broad perspective of the whole life and the narrower
momentary perspective. From the momentary perspective, one can act in a way
that would produce satisfactory but not optimal outcomes in the broader perspective
of one’s life as a whole. One is rationally permitted to have a satisfactory
life instead of the best life if it is done for the sake of the momentary
perspective. However, this line of argument does not persuade me. I do not take
the temporal perspective to be incommensurable with life as a whole, since
different temporal perspectives constitute life as a whole. It would not be
rational for me to act so as to favor the temporal perspective I am in, given
that I will be in another temporal perspective later on. Some individuals are
motivated to actually favor the temporal perspective they are in, but they are
not rational in doing so, given that life as a whole is composed of all the
different temporal perspectives. I take these forms of satisficing to be cases
in which one aims at optimizing from the momentary temporal perspective and one
fails to be fully rational in doing so (this view has been mainly defended by
Parfit and Nagel among others). 

Second, satisficing has as well been interpreted by Slote as a plausible
strategy to the extent that it is an expression of the virtue of moderation. In
his paper ‘Could Aristotle Satisfice?’, however, Byron claims legitimately that
it would be begging the question to require that theorizing about practical
reason be responsive to moral virtues like moderation. Moreover, he argues that
to satisfice is not necessarily to exemplify the virtue of moderation. As we
saw in the toothpaste example, satisficing may be immoderate, as Kate may have
very high standards for a satisfactory toothpaste. Byron compares the notion of
satisficing with the notion of temperateness in Aristotle and claims that
temperateness in Aristotle is not an instance of satisficing, because it is an optimal behavior–just the right amount
of emotion in the appropriate circumstances.

Interpreting satisficing as a virtue may be problematic for another
reason that has not been raised in this collection. We can test our intuitions
by introducing some science fiction. Let us imagine that there are knowledge
pills that can be swallowed by individuals to effortlessly acquire knowledge.
Instead of turning down some additional gustatory pleasure, imagine that Kate
is turning down a pill containing 12 units of geographical knowledge and
ingests instead one with only 6 units of geographical knowledge in it. Assuming
that Kate highly prizes geographical knowledge, it seems to me that we could
not plausibly say that it is virtuous of her to choose to ingest the pill that
has only 6 units of knowledge. Hurka’s chapter on ‘Satisficing and Substantive
Values’ points to a related problem. Hurka argues that, although a satisficing
view can be adopted independently from any substantive view about the good, it
would not be plausible for someone holding an objectivist conception of the
good to endorse satisficing. 

Third, another strategy is to compare the role satisficing has in
practical deliberation with the role supererogation has in ethical thought. In
ethics, supererogation is taken to imply that individuals are not morally
required to act perfectly, but only to act in the right way. There is a
distinction between what you are morally required to do and what it would be
praiseworthy for you to do without being required of you. This is not usually
taken to be the case in rational decision-making processes, but Slote argues
that the same applies to rationality.

In his chapter ‘Why Ethical Satisficing Makes Sense and Rational
Satisficing Doesn’t’, James Dreier claims that supererogation is possible in
the ethical domain only because it involves two competing perspectives: that of
the saint and that of the dutiful individual. However, Dreier denies that
rationality can be similarly divided into two perspectives, since rationality
is about appreciating whatever reasons one actually
has. Moreover, in decision theory, if an agent prefers one alternative, this
choice attributes higher utility to the alternative in question, so an
individual can’t fail to maximize his expected utility. No case of rational
satisficing can make sense on Dreier’s view. Christine Swanton, in her paper on
‘Satisficing and Perfectionism in Virtue Ethics’, seems to deny the importance
of ethical satisficing. She interprets satisficing in moral terms as referring
to the satisfactorily right as opposed to the maximally right action. She aims
to demonstrate the possibility of an undemanding virtue ethics that doesn’t
require a commitment to a satisficing criterion of the right.

This edited collection is intended for an academic audience but is accessible
without background knowledge in the topic. A distinctive quality of this
collection is the care shown by the contributors in responding to each other.
This, however, also tends to create a sense of repetition and overlap, although
one does end up having a more definite sense of the problems raised by the
satisficing strategy. Unfortunately, the last chapter, by Tyler Cowen, stands
out as having little to do with the debate on satisficing, but it does give a
useful overview of various economic conceptions of rationality.

Moreover, some terminological vagueness about satisficing makes the
debate unnecessarily complicated. As mentioned above, most authors hold that
cases of what Slote labels instrumental satisficing are actually instances of
optimization. I also think that it is not helpful to describe these cases as
instances of satisficing, given that these are cases in which ‘satisficing’
locally is best for the person overall. Similarly, James Dreyer and Christine
Swanton start from a different understanding of ethical satisficing. It is not
even clear to what extent the notion of satisficing adds anything to the
ethical distinction between the right and the supererogatory.

In the end, the case for the rationality of satisficing as a model of
practical reason isn’t altogether convincing. Cases of incommensurability in
which two alternatives are on a par or roughly equal seem to be cases that are
most likely to justify a rational satisficing strategy. However, cases of
incommensurability are cases in which optimization is impossible and both
alternatives considered are satisfactory. So a satisficing principle of
rationality doesn’t really have any work to do.

In cases in which one alternative is superior to another, it seems
rational to go for it and irrational not
to.  To take the example of the
person who believes that her career is good enough as it is. It seems
reasonable to think that she could indeed be satisfied with it, but that this
may be due to the implicit assumption that she actually does not want to bear
the negative consequences associated with having a better career, or that she
is suffering from a weak will and is not able to resist it. Let us thus imagine
that this person is able to choose between having either life A with a good
career, or life B with an excellent career. If all things remain equal, there
is a strong intuition behind the claim that it would not be rational for her to
prefer life A. But maybe the whole collection did not put enough emphasis on
the original reason for introducing the notion of satisficing: the discrepancy
between real life decision-making and abstract models of practical reason. No
one faces neatly ranked alternative lives and that is what makes the idea of
satisficing plausible. Satisficing is a rational strategy only to the extent
that there is uncertainty about what would be the best alternative in real-life
situations. If this is correct, satisficing is best seen as a strategy for
increasing one’s chances of ending up with the best alternative and would thus
not be an alternative to maximization but an attempt to maximize in conditions
of uncertainty.

 

© 2006 Alexandra Couto

Alexandra Couto is a PhD candidate at Oxford University.

Categories: Philosophical, Ethics