A theory of Justice: Revised Edition

The Methods of Ethics,
7th ed. (London, Macmillan, 1907), pp. 111f.
366
Goodness as Rationality

---

person is not to be faulted for any discrepancy between his apparent and
his real good.
The notion of deliberative rationality is obviously highly complex,
combining many elements. I shall not attempt to enumerate here all the
ways in which the process of reflection may go wrong. One could if
necessary classify the kinds of mistakes that can be made, the sorts of
tests that the agent might apply to see if he has adequate knowledge, and
so on. It should be noted, however, that a rational person will not usually
continue to deliberate until he has found the best plan open to him. Often
he will be content if he forms a satisfactory plan (or subplan), that is, one
that meets various minimum conditions.
15
Rational deliberation is itself
an activity like any other, and the extent to which one should engage in it
is subject to rational decision. The formal rule is that we should deliberate
up to the point where the likely benefits from improving our plan are just
worth the time and effort of reflection. Once we take the costs of delibera-
tion into account, it is unreasonable to worry about finding the best plan,
the one that we would choose had we complete information. It is per-
fectly rational to follow a satisfactory plan when the prospective returns
from further calculation and additional knowledge do not outweigh the
trouble. There is even nothing irrational in an aversion to deliberation
itself provided that one is prepared to accept the consequences. Goodness
as rationality does not attribute any special value to the process of decid-
ing. The importance to the agent of careful reflection will presumably
vary from one individual to another. Nevertheless, a person is being
irrational if his unwillingness to think about what is the best (or a satis-
factory) thing to do leads him into misadventures that on consideration he
would concede that he should have taken thought to avoid.
In this account of deliberative rationality I have assumed a certain
competence on the part of the person deciding: he knows the general
features of his wants and ends both present and future, and he is able to
estimate the relative intensity of his desires, and to decide if necessary
what he really wants. Moreover, he can envisage the alternatives open to
him and establish a coherent ordering of them: given any two plans he
can work out which one he prefers or whether he is indifferent between
them, and these preferences are transitive. Once a plan is settled upon, he
is able to adhere to it and he can resist present temptations and distrac-
tions that interfere with its execution. These assumptions accord with the
15. On this point, see H. A. Simon, “A Behavioral Model of Rational Choice,” 
Quarterly Journal
of Economics,
vol. 69 (1955).
367
64. Deliberative Rationality

---

familiar notion of rationality that I have used all along (§25). I shall not
examine here these aspects of being rational. It seems more useful to
mention briefly some ways of criticizing our ends which may often help
us to estimate the relative intensity of our desires. Keeping in mind that
our overall aim is to carry out a rational plan (or subplan), it is clear that
some features of desires make doing this impossible. For example, we
cannot realize ends the descriptions of which are meaningless, or contra-
dict well-established truths. Since 
is a transcendental number, it would
be pointless to try to prove that it is an algebraic number. To be sure, a
mathematician in attempting to prove this proposition might discover by
the way many important facts, and this achievement might redeem his
efforts. But insofar as his end was to prove a falsehood, his plan would be
open to criticism; and once he became aware of this, he would no longer
have this aim. The same thing holds for desires that depend upon our
having incorrect beliefs. It is not excluded that mistaken opinions may
have a beneficial effect by enabling us to proceed with our plans, being so
to speak useful illusions. Nevertheless, the desires that these beliefs sup-
port are irrational to the degree that the falsehood of these beliefs makes
it impossible to execute the plan, or prevents superior plans from being
adopted. (I should observe here that in the thin theory the value of know-
ing the facts is derived from their relation to the successful execution of
rational plans. So far at least there are no grounds for attributing intrinsic
value to having true beliefs.)
We may also investigate the circumstances under which we have ac-
quired our desires and conclude that some of our aims are in various
respects out of line.
16
Thus a desire may spring from excessive generali-
zation, or arise from more or less accidental associations. This is espe-
cially likely to be so in the case of aversions developed when we are
younger and do not possess enough experience and maturity to make the
necessary corrections. Other wants may be inordinate, having acquired
their peculiar urgency as an overreaction to a prior period of severe
deprivation or anxiety. The study of these processes and their disturbing
influence on the normal development of our system of desires is not our
concern here. They do however suggest certain critical reflections that are
important devices of deliberation. Awareness of the genesis of our wants
can often make it perfectly clear to us that we really do desire certain
things more than others. As some aims seem less important in the face of
critical scrutiny, or even lose their appeal entirely, others may assume
16. For the remarks in this paragraph, I am indebted to R. B. Brandt.
368
Goodness as Rationality

---

an assured prominence that provides sufficient grounds for choice. Of
course, it is conceivable that despite the unfortunate conditions under
which some of our desires and aversions have developed, they may still fit
into and even greatly enhance the fulfillment of rational plans. If so, they
turn out to be perfectly rational after all.
Finally, there are certain time-related principles that also can be used
to select among plans. The principle of postponement I have already
mentioned. It holds that, other things equal, rational plans try to keep our
hands free until we have a clear view of the relevant facts. And the
grounds for rejecting pure time preference we have also considered (§45).
We are to see our life as one whole, the activities of one rational subject
spread out in time. Mere temporal position, or distance from the present,
is not a reason for favoring one moment over another. Future aims may
not be discounted solely in virtue of being future, although we may, of
course, ascribe less weight to them if there are reasons for thinking that,
given their relation to other things, their fulfillment is less probable. The
intrinsic importance that we assign to different parts of our life should be
the same at every moment of time. These values should depend upon the
whole plan itself as far as we can determine it and should not be affected
by the contingencies of our present perspective.
Two other principles apply to the overall shape of plans through time.
One of these is that of continuity.
17
It reminds us that since a plan is a
scheduled sequence of activities, earlier and later activities are bound to
affect one another. The whole plan has a certain unity, a dominant theme.
There is not, so to speak, a separate utility function for each period. Not
only must effects between periods be taken into account, but substantial
swings up and down are presumably to be avoided. A second closely
related principle holds that we are to consider the advantages of rising, or
at least of not significantly declining, expectations. There are various
stages of life, each ideally with its own characteristic tasks and enjoy-
ments. Other things equal, we should arrange things at the earlier stages
so as to permit a happy life at the later ones. It would seem that for the
most part rising expectations over time are to be preferred. If the value of
an activity is assessed relative to its own period, assuming that this is
possible, we might try to explain this preference by the relatively greater
intensity of the pleasures of anticipation over those of memory. Even
though the total sum of enjoyment is the same when enjoyments are
17. This name is taken from Jan Tinbergen, “Optimum Savings and Utility Maximization over
Time,”