A BROAD UTILITARIAN THEORY OF VALUE AND MORAL VALUE .lames A. Yunker
INTRODUCTION An important thread running throughout the history of moral philosophy has been
the perennial controversy between
interpretation
of
interpretation.
The teleological interpretation asserts that the moral
goodness
of
moral
an action
value
can only
and
Plato's deontologicaI
Aristotle's
teleological
be assessed on the basis
of
its
consequences in terms of human welfare, and that moral judgments are therefore based on rational, cognitive consideration of consequences. The deontological interpretation asserts that the moral goodness of an act is inherent or immanent within the act, and that moral judgments are therefore based on intuitive, non-cognitive appreciation of the act itself.
Utilitarianism is an archetype of the teleological interpretation,
while Kant's Categorical Imperative is an archetyr~e of the deontoloqical interpretation. The intention of this essay is to propose a novel interpretation of utilitarianism
based upon an appreciation of tile modes of thought
applied in confronting various socio-economic policy problems. employ
both
noncognitive
modes of
thought
cognition.
Non-cognitive intuition is used to specify the assumptions
concerning the socio-economic environment.
intuition
These
and rational
Rational cognition is used
to extract implications from the assumptions.
When a declaration is
made, on the basis of these implications, that policy A is superior to policy B, this declaration is a j.dgment based upon both reason (the extraction
of
implications)
and
intuition (the
specification
of
assumptions). Judgments concerning the moral value of a particular action of a particular concerning alternative.
individual the
involve
desirability
the of
same
elements
a particular
as do judgments
socio-economic
policy
We make intuitive a~umptions about the circumstances of
the act, and the basis of these assumptions, we proceed rationally to the consequences.
These consequences are presumed (as opposed to actual
30 or certain) because they are based on intuitive assumptions which may be invalid.
The presumed consequences dictate the moral judgment of
the action.
[s the judgment therefore based on non-cognitive intuition,
or on rational assessment of consequences?
Clearly, it is based upon
both. Moral judgment is best described as teleological in nature rather than deontological.
First, intuition has a role in moral judgment not
because it is satisfactory in itself, but because of the limitations in our knowledge which force us to rely upon it.
The essence of our humanity
is a drive to expand our knowledge that we may be less reliant on intuition
both in social policy
judgment.
formulation
and in individual
moral
Second, it is misleading to polarize our conceptions of reason
and intuition.
Intuition
is b e t t e r regarded as a form of imperfect
reasoning; "imperfect" since it is unconscious, and subject to a high degree both of random error and systematic bias error, but "reasoning" because it is based fundamentally on our perceptions and understanding of the world.
Third, both reason and intuition are enlisted, in moral
judgment, to assess probable consequences of actions.
A B R O A D F O R M U L A T I O N OF UTILITARIANISM The broad formulation of utilitariansim represents an effort to enunciate a concept of utilitarianism that is consistent with its standard definition, and reflects the mode of thought motivating policy
analysis
formulation
is
by
economists
significantly
and
other
different
modern socio-economic social
from
scientists.
many
of
the
This more
particularistic versions of utilitarianism to be found in the literature. Some of these particularistic versions seem to have been set up as straw men by philosophical opponents of the utilitarian concept of goodness and
morality.
[t
should be noted that
the broad formulation
of
utilitarianism in itself constitutes a response to some of the criticisms which are not really directed at the concept itself but rather against specific formulations of it regarded as inadequate by the c r i t i c .
As a
general concept, utilitarianism only insists that evaluations of both good and bad,
and
moral
good
and bad~ must
be based ultimately
on
31 assessments of social and individual consequences of actions.
As a result
of the modern emphasis oil time and u n c e r t a i n t y in social policy analysi% however, the consequences of interest are the l o n g - t e r m consequences rather than the present or current consequences, and these consequences must be viewed as p r o b a b i l i s t i c rather than certain. problem
is one of dynamic
decision-making
In other words, the
under uncertainty
rather
than one of static decision-making under c e r t a i n t y . We conceive a society consisting of I human individuals (i = 1 , _ , I ) over an interval behavior economic
of T periods of time (t = 1 , . , T ) .
of individual theory,
We indicate the
i at time t with the symbol Bit.
Bit
would
represent
a
vector
of
In standard continuously
measurable consumption levels of various goods and production levels of various p r i m a r y
factors.
activities
primarily
not
consumption.
In more general terms, Bit includes many defined
In the c o n t e x t
in
terms
of
production
and
of moral judgment, Bit might often be
taken as dichotomous variables:
the individual either does, or does not,
perform
set of circumstances
some act
in a given
(e.g.,
save the
drowning child, lie to the policeman~ run away from the battle). The u t i l i t y f u n c t i o n for individual i at time t is w r i t t e n as follows:
Uit = IJit(Sit,...,iSit,uit)
(i)
This f u n c t i o n indicates that i's u t i l i t y at time t depends on the behavior patterns
of all other individuals
as well as his or her own behavior
p a t t e r n , and on a random disturbance t e r m uit which represents unknown and uncertain factors influencing his or her u t i l i t y .
A simple a p p l i c a t i o n
of u t i l i t y interdependencies in economic theory is the s p e c i f i c a t i o n of envy through the a ~ u m p t i o n that u t i l i t y of individual i is a decreasing function of consumption of individual j if consumption of individual j exceeds consumption of individual i. We propose that there is a natural tendency in each individual i to choose behavior
at any time
t in order to maximize
the l o n g - t e r m
expected u t i l i t y tllat he'or she will receive over the planning period t = 1,...,T. This l o n g - t e r m u t i l i t y is denoted: T
ui = t _Z0 E(Uit)
(2)
32 where E is the expected value operator. Turning next to society, we propose a set of social oolicy decision variables; Xjt (j = 1,...,J).
These are analogous to the behavioral control
variables of the individuals (Bit).
The Xjt variables include formal
enforcement variables (laws governing individual behavior, together with police, judicial and punitive mechanisms, and institutions), and informal enforcement
variables (general moral principles governing individual
behavior, together with mechanisms and institutions intended to oromote adherence to them, such as indoctrination of the young, unfriendly treatment
or ostracism of offenders,
and so on).
The social policy
decision variables include both continuously measured variables such as tax rates, and dichotomous qualitative existence
or
non-existence
of
legal
variables manifested by the or
customary
behavioral
prescriptions o!~erative under certain sets of conditions. The social welfare function at time t is written as follows: Wt = W t ( U l t , . . . , U i t , vt)
(3a)
This function indicates that social welfare depends on the individual utilities of all individuals in the society, and on a random disturbance term v t which represents unknown and uncertain factors influencing social welfare.
Equation (3a) represents the evaluative social welfare
function~ the instrumental social welfare function shows the dependence of social welfare on social policy decision variables, and is w r i t t e n as follows:
w t = wt(xlt,...,xjt , wt) In (3b), w t is another random
(3b) disturbance term which will have some
degree of overlap with v t, but which contains additional uncertainty elements because (3b) is more fundamental than (3a). There is a natural tendency in society to choose social policy at any time t in order to maximize the long-term expected social welfare over the planning period t = 1,...,T.
This long-term expected social
welfare is denoted: T u = t ~= O E(Wt) where (as in equation (2)), E is the expected value operator.
(b,)
33
The maximization problems of the individual and society are both constrained
optimization
problems.
The
individual
i
confronts
a
constraint that may be w r i t t e n in i m p l i c i t form as fgllows:
Fi(BIt,...,BIt, • ZIt,...,ZKt, rit) = 0 (5) where tit is the usual rando,n d,~drbance variable. The Z,~t variables (k = l,...,K)in equation (5) represent exogenously determined parameters embodied both in the environment (e.g., resource constraints) and in human individuals (e.g., survival needs in food, water, air). A standard example of an individual constraint in economic theory is the budget constraint:
the individual's after-tax income from personal labor,
property ownership and financial gifts must 9qual expenditures on priced consumption goods.
In such a
constraint,
labor provision and
consumption of goods represent the Bit behavior vector of individual i, gifts represent behavior vectors of others, the tax rate is a socially determined
Xjt variable,
and the endowment of property
is an
exogenously determined Zkt variable. The basic notion of a constraint utilized in economics ,nay be generalized to almost any i n t e r a c t i o n a l
situation.
F'or example, if individual A hits individual B, B may very w e l l
hit A back. This is a constraint as far as A is concerned. The tendency to maximization by individual i of u i as given in equation (2) subject to the constraint given in equation (5) determines Bit as a function of the X j t and Zkt variables: Bit = Bit(Xlt,...,Xjt, Zlt,...,ZKt,Sit) where sit is the usual random disturbance variable. interpreted
as
a set
maximization problem.
of
constraints
entering
(G) Equation (S) can be the
social
welfare
This is the critical link between the evaluative
social welfare function (3a) and the instrumental social welfare function (3b).
Substitution
of the behavioral functions (S) into the individual
u t i l i t y functions (1) manifests the dependence of individual utilities (Uit , i = 1,...,I) on social policy ( • form individual
utility
J = 1,...,J). Substitution of these reduced
functions into the social welfare function (3a)
manifests the dependence of social welfare on social policy, which is indicated explicitly in the instrumental social welfare function (3b). The full social problem is the maximization
of long-term social
34
welfare ~ rather than the maximization of instantaneous random social welfare Wt. It sllould be noted that it would be meaninqless to soeak of maximizinq W t.
Wt is a random variable since it depends on all of the
random variables soecified above: 'Jit, vt, wt, t i t , and sit. It is sensible to speak of maximizing the exoected valt]e of a rar~L;,J,l~ variable, in this case
E(Wt).
But
the
crit3rion
r)rooosed
is
the
summation
of
instantaneous F(W t) over a ~eriod from t = ,3 to t = T. The maximization of w is accomolished subiect to two fundamental constraints.
The first is the set of Zkt variables:
determined
parameters
involved
in
the
the exogenously
fundamental
physical
characteristics of the natural universe and the human organism.
The
second is the tendency toward individual utility maximization displayed by each individual which results in eq,Jation (6). The asterisk superscriot denotes the socially ootimal solution in the social policy variables. denoted by
~o*.
Thus •
produces the maximum level of u,
Throuqh the bellavioral
functions (6), the socially
o0timal Xjt produce the socially ootimal 13it behavioral r~atterns of the individuals in society. It should he unobiectionable to refer to the o,otimaI ' ~ t as good, desirable or beneficial social policy, and to the ootimal F~it as good, desirable and beneficial individual behavior. optimal X~t and ~ t
I~ut should we refer to tile
as nnorally qood? On the contrary, moral goodness,
as opposed to simole ~oodness, is a considerably more subtle concel~t than that which has been delineated thus far.
In the first olace, moral
goodness is a concept pertinent to human behavior, not to social oolicy. We often speak of immoral laws (e.g., 3ira Crow laws in the DostPxeconstruction South, anti-Semitic laws in Nazi Germany). But the laws themselves are inanimate, non-volitional objects.
The most we can say
with any s~mantic orecision is that the making of some laws and their subsequent enforcement (may have) represented immoral actions by immoral oeople. 3udgments on the moral quality of various actions are based on two seoarate concepts of behavior.
C)nly where the conceots are in fact
35
differentiated
qualitatively
or quantitatively
does the possibility
moral judgment, whether positive or neqative) arise. proposed
Bit
as the
behavior
expected social welfare0J.
pattern
which
of
We have already
maximizes
long-term
Now define Bi~t as the behavior l~attern that
maximizes loncl-term expected individual utility U i. The oossibUity of f~oral judgment only arises in eases of a oerceived difference between behavior (B~t)_ that maximizes the individual's
own
welfare
and
behavior (B~t) that maximizes social welfare (~). One can object that in arriving at ootimal social r)olicy X j t , society incorporates
as
a
constraint
the
utility
maxi,nizing
behavior
of
individuals embodied in the behavioral relation (6). Thus there should be * no difference between Bit
and Bit 0
. Socially desirable be)lavior should
be that which most benefits the individual.
However, that there was not
always and everywhere a perfect corresl~ondence between the interests of one individual and those of the rest of society is amply demonstrated simply by the e•
of a body of formal
laws and regulations
enforced by police forces and rJrisons, and a body of generalized moral principles enforced informally by spoken and non-spoken aparobation and disapprobation.
The system of rules and enforcement mechanisms is
intended to bring self-interested behavior into better alignment with
socially optimal behavior, i.e., to make B~t approximate ~it" But it is a.oparentlv i,noossible to eliminate all divergences between the two behavior patterns, and would be excessively costly to attempt to do so. The development and imr)lementation of the concept of moral qoodness is analogous to the a t t e m p t by Society to have its members internalize ~tand'~rds
and
attitudes
which
promote
law-abiding
behavior consistent wit)l generalized moral principles.
behavior
and
This reduces the
toll taken on society's resources by the enforcement system. The individual I~as a tendency to choose behavior that maximizes u i,
Therefore, the behavioral relationships of equation (6) are based
on a tendency and not a certainty.
This is one reason why a random
disturbance must a orJear in them. Moral behavior, whether positive or negative, can only occur if there is a distinction between self-serving behavior (B~t) and socially
36 !
oriented behavior (B~t).
In such cases, if actual behavior (denoted Bit)
approaches or equals ~ t ' we condemn it as morally bad~ if it approaches or equals Bit ~ we laud it as morally good. Moral goodness can only come about if the individual sacrifices his or her own interests in favor of social interests~ and moral badness can only come about if the individual puts his or her own interests ahead of those of the rest of society. It is one of the abiding paradoxes of moral value that the existence and application of the concept tends toward the extinction and negation of the reality.
The obvious purpose of laws and enforcement is to detery
to prevent~ and hopefully to exterminate illegal behavior which at one time was usually rather common and considered merely immoral.
But of
course there is no moral value now in obeying laws simply out of fear of punishment.
Informal enforcement mechanisms such as ostracism and
verbal abuse are different in degree but not in quality from formal enforcement mechanisms such as fines and imprisonment.
Thus, for
example, at the moment it might be considered immoral not to give to a worthy charity such as the United Way. [mmoralbehavior (not giving to the United Way) might elicit verbal abuse.
If someone gives to the
United Way under these conditions, is he or she acting morally?
The
more personal adversity to verbal abuser the less moral value will be attributed to giving to the United Way.
When one thinks closely about
it, remarkably few opportunities are presented to anyone in society to act in a truly or fully or definitely moral manner. generally
speaking,
correspondence
society
between
is
quite
self-interested
successful behavior
This means that at creating a 0 Bit and socially
oriented behavior B ~ it" An essential characteristic uncertainty.
of moral value is the element
of
We speak far more naturally about moral judgment than we
do about moral knowledge.
The essence of judgment is arriving at a
conclusion on the basis of imperfect or incomplete knowledge.
It is
important to distinguish, therefore, among (a) B{~, the behavior that actually maximizes ~0~ (b) Bi"t, the behavior that an outside judge believes that i believes will maximize ~ estimate of Bit).
(the outside judge's estimate of i's
We also define that actual behavior that maximizesu i
37 ^o
n
(B{t) , i's estimate of that behavior (B{t) , and the outside judge's estimate of
i% estimate of that behavior (B~t). Similarly we define actual
consequences,
estimated
consequences,
and
estimated
estimated
consequences with reference to social welfare for socially optimal and ^
^ ,., ~, , ,,,0, ~0,~, and with reference to self-interested behavior (~*, u*., individual utility for socially optimal and self-interested behavior (u~, ['~I' ~I' uO' u:~'u~" The same definitions may be applied to actual behavior ( Blt, ~ i t ' ~ i t ) , its social consequences(~', &', ~}, and its This notation is summarized in Table individual consequences (uI ,~,i,O~. ~' l.
Table i: A Schema of Behavior (Bit) related t o Social Consequences (W) and Individual Consequences (Ui) self-serving behavior (maximizes Ui)
true
B~
Wo
actual behavior
B' ~
U!
it
U*
1
3.
1
f
e s t i m a t e d by individual i i
1
i's estimate itself estimated by outside judge
W*
W'
i t ~
it'-~ 00
s o c i a l l y - o r i e n t e d behavior (maximizes W)
[,7
it-..~
9 W*
1
1
l
1
A theory of moral judqment based on these concepts shows that the moral judgment
made on a oarticular
action (M) is a m u l t i p l i c a t i v e
function of an estimated index of morality (r~) which lies somewhere in the range from 0 (higll|y immoral) to 1 (highly moral) inclusive, and an estimated
level of consequences (~), which is defined as the gain to
38
society plus the Loss to the individual as a function of the actual behavior of the individual.
The estimated index of morality is itself the
product of two numbers, both between 0 and 1 inclusive, the first being the proportion achieved of the potential gain to society from highly moral behavior by the individual, and the second being the proportion achieved of the potential Loss to the individual from highly moral behavior. These concepts may be expressed formally as follows; M = m C,
(7)
where:
_
_
(8)
~= (~=,_~0)
= _ 3~ (u*. 1
c = (=' - =0) + (
].
_ ui )
(9)
As the index of morality (r~) has an inclusive range from 0 to l, it seems natural to suggest that the value of .5 should be the dividing line between behavior deemed immoral and behavior deemed moral.
The
level of consequences (~) has a range from 0 up to the oositive number (~, _ ~0 ) + (~0 _ ~ ) .
(It should be noted that both terms in the last
expression are positive because by definition ~ >
~0 at~d ut0> ~.*. These
two inequalities are fundamental to the entire discussion.) If M is below the midpoint
of this range, the corresponding behavior is deemed
immoral, and if M is above the midpoint of the range, the corresponding behavior is deemed moral. What about generalized moral precepts, such as the universal injunctions
to
truthfulness,
non-violence,
and reliability?
cooperation,
helofulness,
We may specify pattern
embodying
friendlinessp
B + as a generally
recommended
behavior
these
desired
characteristics.
Over the great majority of the multitude of events in
the course of human existence, behavior approximating B + will at the same time closely approximate the social welfare maximizing individual
39
behavior pattern B~t.
It is reasonable to suppose that the rules were
proposed in the first place because of this presumption. We have just proposed, in equations (7), (8), and (9), a seemingly quite complicated methodology for arriving at a moral judgment. A nonutilitarian may very well balk at this formulation, and propose instead that in practice moral judgments are made on the basis of how closely a given behavioral pattern conforms to the traditional moral standard embodied in B +, The really difficult problems of moral judgment are not in
fact
rendered
principles.
any easier
simply
by invoking
traditional
moral
The reason is that moral problems are moral problems
because of obvious conflicts between traditional moral principles.
For
example, A is walking down the street and witnesses B attacking C. A attacks B in defense of C.
Is it moral?
moral precepts is of no help.
Reference to the traditional
The precept of non-violence is in direct
conflict with the precept of helpfulness.
We should have to know much
more about the circumstances and consequences of the incident in order to render a moral judgment.
For example, if B is a strong young
hooligan preparing to stab a defenseless old lady C in order to steal her severely limited funds, and the rescuer A is a frail, elderly ,-nan with a heart condition, we would probably deem the intervention moral. On the other hand, if B and C are two small children squabbling over a piece of candy, and the rescuer A, a robust adult, beats B senseless, the act would probably be deemed immoral, particularly if A's over-reaction was motivated by a desire to indulge a sadistic streak. The theory of moral judgment proposed here is fully consistent with the general notion of moral judgment to be derived from an overall appreciation
of
both
the
literatures on the subject.
philosophical
and
social
psychological
The major determinants of moral judgments
of acts are often enumerated as follows:
the judge's perception of the
circumstances and consequences of the act; the judge's perception of the actor's responsibility, motivation, intention, awareness, and sacrifice; and finally, the personal characteristics of the judge.
Each of these
dimensions of moral judgment may be related briefly to the proposed theory.
Circumstances
(the Xjt
and Zkt variables,
as well as the
40 behavior patterns B of all individuals other than i), together with i's behavior (Bit), determine consequences (wand t~) through the functions (1), (2), (3a), and (4).
Individual i's responsibility for the consequences
varies depending on the behavior pattern (act) and the circumstances. Consequences are inherently uncertain to an individual because of imperfections in his or her knowledge of the relevant functions, and because of the random variation introduced into the outcomes via the random disturbances uiu vt, w t, tit, and sit. A useful way of envisioning the consequences of actions is that a given action changes the perceived probabilities of occurrence of.various outcomes. If the probability shift is dramatic (e.q., the act causes the probability of a particular outcome to rise from .001 to .999), we agree that the subject's responsibility for the outcome is virtually complete.
But in many cases a given action
does not affect radically the probability of a given outcome.
It is a
matter of subjective judgment as to how high the change in probability must be for the subject to be deemed responsible as opposed to not responsible.
Subjective judgment is invoked to establish an arbitrary
demarcation point along a continuous spectrum of responsibility. A shift in the probability structure of outcomes as a result of an action shifts the expected value of the outcome.
P~esponsibility may thus be gauged
in terms of the shift in expected value of consequences; a small shift means small responsibility and a large shift means large responsibility. The
very
same
responsibility
act
could
or a low
therefore
result
in
a high
degree of responsibility
degree of
depending on the
circumstances. The social and individual consequences defined above as criteria for moral judgment (wand ui) are expected value concepts and hence incorporate the notion of resl)onsibility. Motivation, intention,
and awareness are incorporated through
making moral judgment dependent no on actual consequences ~and u~, but on the perceived consequences by the subject:
~ and ~s
The
perceived consequences may diverge quite substantially from the actual consequences.
Motivation or intention is frequently broken down into
desired ends versus undesired consequences.
For example, a robber,
endeavoring to obtain funds to feed his hungry family, encounters
41 opposition while robbing a liquor store and kills the proprietor.
The
robber's primary intention was not to kill but to feed his hungry family. We judge the morality of the killing less harshly than we would have had the robber entered the store without a hungry family but with the intention of killing the proprietor for the purpose of bolstering his (the robber's) sensations of dominance.
This is incorporated into our theory
as follows:
a robber committing a robbery as a pretext for killing enters
the
with
store
a much lower estimated
expected value of
social
welfare ~ than a robber who wants to feed his hungry family (because in the former case there is not a hungry family to be benefited by the act but t h e r e is an innocent individual, the proprietor, to he drastically harmed by it). latter.
Thus M for the former is much lower than M for the
Another subdivision of motivation or intention is into primary
ends versus secondary ends, as, for example, in the statement:
"It was
morally good for Martin Luther King to pursue equal rights because his primary motivation was to help mankind and not to enhance his personal recognition
and
prestige."
The
statement
indicates
that
King's
perception of the consequences of his behavior was that it set social welfare ~ near its upper limit of ~ a n d expected personal utility u i near ^
its lower limit of vi, and that moreover, substantial distances existed between ~ and ~o and 6 i and divergence between
60i" to the extent that there is less of a
u i and u~, there is less sacrifice attached to
an
action, and hence less moral value in it. The concept of sacrifice is fundamental to any iml3utation of moral quality.
The individual must sacrifice his or her own interests in favor
of those of society for the act to be moral; or he or she must sacrifice the interests of society in favor of his or her own interests for the act to be immoral.
We capture this element through distinguishing behavior
which maximizes the individual's own welfare (E~it) and its attendant individual
and social
consequences
(
uP, u0), from behavior
which
maximizes society's welfare (F~it) and its attendant individual and social * consequences ( ui,
~*).
If there is no distinction between the two
behaviors in a given set of circumstances (i.e., if i~it = Bit,ui,0 -u.*_ l and ~01 = ~o*),
then there is no possibility of moral imnutation.
This 13pint is
42
expressed technically in the model through the dependence of moral judgment M on the index of morality m (equation (7)), which itself depends on two ratio terms (equation (8)) the denominators of which are both zero when B~t = l'~it.
In mathematical logic, division by zero is
indeterminate and undefined. We come finally to the characteristics
of the judges.
Moral
imputations are unavoidably judgmental, and involve not one but two estimations, both subject to sampling error. consequences
of
his
actions,
estimates of the consequences.
The actor estimates the
and the judge
estimates
the
actor's
It is quite obvious from the voluminous
experimental literature on both legal and moral judgment that human judgments are influenced at least as much, and possibly more, by the personal characteristics of the judges as they are by the objective facts of the case under consideration.
In particular, there are observed
variations in judgment by different judges considering exactly the same evidence on a case, and a considerable proportion of this variation is statistically
related to personal characteristics of the judges such as
age, religion, personality variables, political and social attitudes, status, occupation, and so on (Staub, 1978, 1979). Judgment is only required in problem cases, and problem cases are problem
cases
because
different
people
disagree
regarding
inferences and implications to be drawn from the available data.
the Thus
the e~ence of moral judgment as applied to a moral problem is that the final judgment will differ among different judges. [t may well be that as the available data on a given problem accumulates and deepens, the dispersion in judgment will shrink (i.e., the standard deviation of the sampling distribution of judgments will decrease), until finally a given action
would
be
virtually
universally
universally acclaimed as moral. the different
condemned
as immoral,
or
But until this uniformity is achieved,
personal characteristics of the judges affect how they
perceive and evaluate the evidence bearing on a given problem. Judge ,4 will tend to arrive at a different ~ and ~0than will judge B. One of the more absurd criticisms of utilitarianism is that it is too pat in that it proposes simplistic solutions to complex problems.
The
43 fact is that the two judges both could be enthusiastic utilitarians who would agree totally with everything that has been put forward in this paper--and yet there remains literally an infinite number of grounds for disagreement between them regarding a specific judgment to be made in a specific
problem situation.
economists
are
nearly
all
For example, consider the fact that united
in
allegedly
simple-minded
utilitariansim, yet they remain renowned for their inability to arrive at a reasonable consensus on, for example, the best way to fight stagflation. It is a far more serious criticism of utilitarianism that it is devoid of any practical usefulness simply because it is such a general formulation. non-utilitarian
clearly has nothing better to offer
The
since his or her
wisdom, stripped to its essentials, amounts merely to tautology cure tradition:
"It is morally right because it is morally right, and we know it
is morally right because we have always known it is morally right." Utilitarianism overall
offers an appealing formulation of the problem in an
sense,
and
correct
overall
problem
formulation
is almost
certainly a necessary pre-condition for arriving at a valid and genuine solution to any specific sub-problem in the category. PROBLEMS IN THE APPLICATION OF BROAD UTILITARIANISM All the functions used to specify socially desirable social policy (X~t) and individual
behavior
(B~)
were
completely
general
mathematical
functions.
In order to apply utilitarianism strictly or formally to any
specific problem, we must proceed to the opposite end of the spectrum-to the complete specification of the mathematical functions, i.e., to the specificat~',n
nf
both
the
parameters of the function.
m~thgm~tica!
form
and the
numerical
Between complete generality and complete
specification, there are two intermediate stages, the specification of general properties, and the specification of mathematical form. The full spectrum is indicated and illustrated in Table 2.
44
Table Z: The Spectrum between Complete Generality and Complete Specification of Relationships Term
Concept
Example
complete generality
specification only that a relationship exists
intermediate generality
specification only of the qualitative properties of a relationship
dy/dx > 0
intermediate specification
specification of an explicit mathematical form
y=a+bx
complete specification
specification of the numerical values of the parameters of an explicit mathematical form
y=2+3x
Over
the
eenturies,
empirically
natural
scientists
have
y = f(x)
uncovered
a host
validated, completely s~)ecified relationships involving all
kinds of variables, from Boyle's PV = T to Einstein's E = me 2. though
economics
scientific
of
is generally
Even
deemed to be the oldest and most
of the disciplines concerned with social inquiry, and even
though the pages of economies journals are filled with impressive looking mathematical
formulae,
tables,
economies that even remotely laws.
and
figures,
there
is
nothing
in
eompares with Boyle's and Einstein's
In fact, it is an indication of the immaturity of economics as a
scientific discipline that a great deal of theoretical work is conducted exclusively
at
the
level
properties are postulated
of
intermediate
generality.
Qualitative
for certain fundamental relationships, and
mathematical reasoning is employed to derive qualitative oroperties of derived relationships.
For examl)le, in examining the microeconomic
roots of inflation, qualitative properties are specified for the firm's production function,
the firm's objective
is soecified (usually profit
maximization), and qualitative implications are obtained concerning the effect of input prices and output prices.
Economists are reluctant to
move forward to the stage of intermediate specification because little or no consensus exists as to the most likely forms of the fundamental relationships.
For example, three specific forms are widely utilized to
45 represent the production function of the firm:
linear (L)~ Cobb-Douglas
(CO), and Constant Elasticity of Substitution (CES). choice
among the three is dictated
mathematical
far
In practice~ the
more by considerations
tractability than by considerations of plausibility.
of The
CES production flfnetion conventionally is deemed the most realistic of the three because it is the most general and flexible7 but it is also the most complicated in an algebraic sense. If, as often happens~ the other components of a given mode] are sufficiently complex that incorporation of a CES production function would abrogate the possibility of algebraic analysis, it is conventionally
acceptable to use the simpler CO or L
forms. The
route
from
intermediate
specification
to
complete
specification in economics (and other social sciences) is chiefly along the rough road of statistical analysis of non-experimental data generated by the vast complexities of real-world social interaction.
The statistical
tool which is most favored in economics is regression analysis.
The
classical
the
linear
regression
model
originally
was developed
in
nineteenth century for the purpose of analyzing experimental data. [t is very possible that many of the a~umptions
of the classical Linear
regression model are not satisfied by the data sets obtained from the real world. An entire branch of statistics~ called econometrics, has been developed inclined
by
economically
economists~
which
inclined
statisticians
exclusively
studies
and
statistically
breakdowns
assumptions of the classical linear regression model.
in
tile
As a result of the
development of this field~ economists are now frequently to be found estimating their postulated relationships with techniques such as three stage least squares regression in preference to ordinary least squares regression. Economists who employ recently developed9 highly sophisticated estimation techniques are unlikely to be particularly confident of the validity and accuracy of their estimated relationships.
The field of
econometric analysis has brought to light a formidable and intimidating array of identification and estimation problems to which economists had been oblivious as recently as the 19213s and 1930s.
The solutions which
46 have been proposed to these problems are based mostly on assumptions which seem only marginally more realistic than the as.~umptions of the old classical linear regression mode[.
Econometric thinking over the
past few decades has succeeded in throwing most estimations based on ordinary least squares into severe doubt; it seems natural to expect a similar fate in the future for estimations based on currently respectable methods suctl as three stage least squares. All this implies that a great deal of uncertainty necessarily is attached
to
social
scientific
analysis
designed
to
deepen
our
understanding of the socio-economic environment that we may improve socio-economic policy.
In the voluminous professional literature
of
economics, one rarely encounters strongly worded policy prescriptions; in fact, one rarely encounters any policy prescriptions at all. Even when presenting results which potentially could strongly affect our evaluation of a given institution or policy, it is customary to assert relevance to judgment in a very
diffident
manner:
e.g., "these results may be
relevant to our consideration of such and such a policy."
This custom is
the result of the almost universal recognition and acceptance by welleducated
economists
empirical
economic analysis are not nearly as solid as those in the
natural
sciences
that
which
the
are
foundations
being
used
of both
as
theoretical
informal
and
standards
of
comparison and emulation. Economic analysis .is applied to economic oolicy problems in such areas as:
What rates of taxation should be applied to various individual
income levels?
What benefits ought to be paid to the elderly under the
Social Security system? grow at a constant rate? education,
Should the money supply only be allowed to How much subsidation should there be of
urban mass transit,
cultural
endeavor?
There are many
additiSnal social policy problems generally deemed non-economic, which do, however, normally have economic aspects or implications, such as: What should be the national defense budget? procured?
What weapons should be
How much foreign aid should the United States donate to the
less developed nations?
Should there be capital punishment?
should be the law on abortion?
What
Turning to individuals, the idea of
47
socially beneficial individual behavior can be formulated in questions such as these: How fast should people drive on the highway? How much should they give to charity?
How much should they save?
they approach the socialization of their children?
How should they
allocate their expenditure between liquor and books? consider questions concerning moral judgment. rain-slicked motorist
highway
crowded with
traffic~
How should
Finally, we may
For example, given a
how fast
does a given
have to be driving for his or her behavior to be deemed
immoral? Answering any of these questions precisely requires specification of all relevant functions down to the level of coml~lete specification.
It
should be apparent from the brief review above of problems confronting economic policy analysis that this specification process is not easy or straightforward.
Just as it is not easy to answer questions pertaining to
the socially optimal income tax scheduled neither is it easy to answer questions about desirable national defense policy9 proper socialization of children~ or the relative immorality of fast driving on a crowded road in rainy weather. Important
differences
do
exist
between
discussions
of~
for
instance~ optimal taxation in the professional economic [iterature~ and discussions of the morality of fast driving.
There is currently a rather
substantial literature on optimal taxation in which all the functions necessary for an analytical solution to the overall social welfare problem are specified, t potential judgment.
Analyses of the morality of fast driving typically review
factors bearing on the situation that
affect
the morality
Factors which increase the potential costs of fast driving to
the rest of society (such as wet pavementy heavy traffic~ etc.) tend to
1[t should be noted that in the great bulk of this Ziterature~ the relevant functions are specified only to the level of intermediate generality, This permits the specification of the general properties of the optima] tax schedule~ but not the appropriate numerical values of tax rates at various income levels.
48 produce the judgment of immorality.
Factors which increase the
potential gains from fast driving to the rest of society (e.g., a doctor rushing to a hospital to save a critically injured person) tend to produce the
judgment of morality.
But these factors are not integrated
consciously and explicitly into mathematically and ~lu,-ner~cally specific functions in order to arrive at a conclusion, as is the case in the or~timal taxation literature.
Nevertheless, the underlying functions do exist and
are being used by the iudge in order to evaluate the morality of fast driving under a specific set of circumstances.
The fact that they are
used unconsciously rather than consciously does not negate nor diminish their reality.
One does not refute biological phenomena such as
respiration and blood circulation on the basis that human beings are unconscious of respiration and blood circulation.
The same problems
which beset formal analyses carried on by means of conscious reasoning grounded on conscious assumptions (as in optimal taxation) also beset informal analyses carried on by means of unconscious reasoning grounded on unconscious assumptions (i.e, by means of intuition). assumed
r~athematical
forms
or the numerical
Namely, the
parameters of the
relevant functions might ')e invalid. The various problems confronting the specifications necessary in order to apply utilitarianism to practical problem situations may be subdivided into;
(17 those pertaining to the nature of the evaluative
social welfare function Wt = Wt ( U l t , . . , t J i t , vt) , i . e , to the evaluation of individual utility consequences (Ui, i = 1,..~I); and (2) those pertaining to
the
causative
relationships
between
the
independent
variable
circumstances (Zkt), individual behavior (Bit) and social policy (Xjt) on the one hand, and the dependent individual utility consequence variables (Ui,
i = 1,...,I) on the
other.
These causative relationships
are
summarized in the behavioral constraint set shown as equation (6). Problems of the first type pertain to the form of the evaluative social welfare function (equation (3a), and that problems of the second type pertain to the form of the behavioral constraint set (equation (G)).For example, the optimal taxation literature has developed a relatively plausible technical specification of the ancient equity versus efficiency
49 problem, creates
The assumption of diminishing marginal u t i l i t y an incentive
to
redistribution"
a given
dollar
of income of
income
transferred from a rich household to a poor household decreases the utility of the rich household less than it increases the utility of the poor household,
On the other hand~ taking a dollar away from the earned
income of the rich household may reduce the labor provided by the rich household~ and paying that same dollar as a subsidy to the poor household may also reduce the tabor provided by the poor household, Thu% as the pie is divided more evenly~ it is probable that the size of the pie will shrink,
The optimal position will normally lie somewhere between the
extremes of no redistribution and complete redistribution~ which would make the incomes of tile two households equal~ after-tax and a f t e r subsidy. But the exact position will depend on the degree of diminishing marginal
utility
(i.e.,
income)~
and on the
how quickly sensitivity
marginal
utility
(or elasticity)
of
diminishes
with
the labor supply
responses of the rich and poor households to redistribution.
These are
problems of specification of the behavioral constraint~ of attempting to perceive the true nature of the world~ or the true effect of a given set of circumstances~ individual behavior~ and social policy on individual utility consequences. These are not easy problems. [n economics at the present time~ statistical estimates of household suopIy elasticities of labor are considered highly speculativey and numerical estimates of the marginal utility/income relationship almost completely speculative. Another problem is in specifying the sociaI welfare function.
For
example~ there are two households in the society~ one rich and one poor~ and a proposed plan of redistribution would reduce the utility of tile rich household from 15 to tO and raise the utility of the poor household from 5 to 8.
Whether this shift is desirable or undesirable depends on the
social welfare function.
Three analytical social welfare functions are
commonly used in contemporary economic policy analysis: the Bentharn or
sum-of-utilities
formulation~
the
Nash
or
product-of-utilities
for muIation~ and the F{awls or maximum formulation.
According to the
Bentham criterion~ the shift is undesirable because it reduces the sum of utilities from 20 to 18. But according to the Nash and Rawls criteria~ it
50 is desirable, in the first case because the product of utilities is raised from 75 to 80, and in the second because the utility of the lowest-utility household is raised from 5 to 8. An abstruse debate, partly philosophical and partly
technical,
is continuing in the economics journals on the
relative merits of these formulations with no consensus in sight.
For
practical purposes of policy analysis, it is more or less permissable to use any one of the three functions which is mathematically convenient to a particular application. produce
different
results,
But the different as
was
functions ,nay easily
demonstrated
in
the
simple
redistribution example above. One
issue in connection with the social function is worthy of
special mention.
Economists who use explicit, analytical social welfare
functions in policy analyses are invariably pure egalitarians. That is, the proposed social welfare function weights the utility of each individual or household in exactly the same way (as in the Bentham, Nash, and F~awis functions). This may be one of the reasons why real-world policy-makers normally go about their business with nearly total disregard for the prescriptions of the
economists.
It seems
clear first, that each
individual's personal version of an appropriate social welfare function incorporates unequal weighting of individuals; and second, that the operative social welfare function in real-world social policy formulation incorporates unequal weighting. closer
to
himself,
whether
Each person tends to weight those in
familial,
geographical,
national,
occupational, or social status terms, higher than those more remote from himself or herself (Eisenberg-Berg and Neal; and Hamilton). At the social level, society weights the utility of each individual in prooortion to the political power wielded by that individual. As political power is unequally distributed, so also is the weighting of individual utilities in social policy formulation. This situation appears to be basically unsatisfactory.
Our mental
ideal of the best judge of social policy and morality, is that represented by a completely disinterested, impartial third party to w h o m human
individual would mean
anything more
no one
or less than any other
human individual. However, every person is on the inside looking out at
51
the rest of society, and there is a natural tendency for each individual to interpret human welfare largely in terms of his or her own personal welfare.
Opinions differ on the natural mix of egocentrism and altruism
in the raw human personality. The purpose of socialization of the young is to discourage natural egocentric impulses and behavior, and encourage natural socially cooperative impulses and behavior. The simple fact that socialization is necessary, that positive actions hays to be encouraged and negative actions discouraged, strongly suggests that while human individuals may not be born into the world lusting after the blood of their fellows, they are born into the world with unsatisfactory, suboptimal personaliti es. In
terms
of
the
mode[
developed
herein,
socialization can be interpreted in two ways.
the
purpose
of
The first purpose is to
educate the young individual in the constraints he or she faces in his or her efforts to maximize his or her own utility:
to deepen his or her
awareness and understanding of equation (5).
This encourages the
individual
to
embrace
pro-social
behavior
and
eschew
anti-social
behavior because it is in his or her own interest to do so. This purpose has been characterized by Freud as the development of the ego (which controls the id), aID.: hy Kohlberg as the devglopment reasoning.
of premoral
The second purpose is to develop the concept of a social
welfare function in the individual.
An individual is born with a social
welfare function of a sort--but it weights the utility of the self far too much and the utilities of others far too little.
The socialization process
attempts to spread about more widely the weight attached by a given individual to the utility of various people in society. This encourages the individual
to
embrace
pro-social
behavior
and eschew
anti-social
behavior simply because i t benefits others, without reference to the individual's own welfare.
This purpose has been characterized by Freud
as the development of moral reasoning. 2
2Lawrence Kohlberg actually hypothesizes three levels in moral reasoninq: premoral (do the right thing to avoid nunishments and gain rewards); conventional (do the right thing because it is good to help other people and keep society functioning); postconventional (do the right thing because it is the right thing). Two sub-categories are perceived within each of the three levels.
52 While opinions differ as to the relative importance of external (purpose 1) and internal (purpose 2) enforcement of pro-social behavior, it seems fairly clear that both purposes are achieved to some extent, but both efforts fall short of the ideal conception.
Thus while most people
behave in a tolerable manner, it is easy for us to imagine improvements in the social behavior of others~ and, with somewhat more effort, in our own. This discussion points up the variability to be expected in the social welfare function as perceived by different people. Each individual tends to weight himself, and those close about himself, more highly than he weights others in his perceived social welfare function. socialization
can
reduce
this bias, but (~robably)
Intensive
not eliminate
it
altogether. Thus when w e ask what is the social function, or even what should be the social welfare function, the answer is not clear and unambiguous.
An
inevitable
social welfare function.
uncertainty attaches to the evaluative
Similarly an inevitable degree of uncertainty
attaches to the various properties of the natural and h u m a n environment which determine the constraints subject to which social welfare is to be rnaximized.
We
have used the various random
variab|es incorporated
into the formal model of the previous section to indicate these sources of uncertainty. Oecision-making under uncertainty forces us to rely on intuition. Intuition is invoked to specify the exact locations of the various random variables in the model, and to soecify their probability distributions. While the
non-utilitarian believes that social i~oliey evaluation and
evaluation of morality should be described as non-rational, they are m u c h better described in terms of unconscious rationality or imoerfect rationality.
First, in
specifying
the
locations
and
probability
distributions of the random variables, the individua! is relyinq, at least fuzzily, on his or her other perceptions and knowledge of the world. Second9 in
proceeding
from
these
probability
distributions
to
the
consequences in terms of long-term expected value of individua! utility and social welfare, a fuzzy reasoning process is utilizedp involving fuzzy estimates of the various functions specified.
53 CONCLUSION A s u f f i c i e n t l y broad f o r m u l a t i o n
of u t i l i t a r i a n i s m avoids many if not
most of the d i f f i c u l t i e s t h a t its c r i t i c s have a t t r i b u t e d to the u t i l i t a r i a n concept.
One might be tempted to go so far as to suggest t h a t in r e a l i t y
~lo viable alternative exists to broad u t i l i t a r i a n i s m in p r a c t i c a l social policy
evaluation
and
moral
judgment.
At
u t i l i t a r i a n i s m does not solve all our problems.
the
same
time,
broad
All that it can accomplish
is a roughly accurate f o r m u l a t i o n of any given problem in social policy or moral
judgment.
That
this is likely
to be superior
to the non-
u t i l i t a r i a n ' s proposal t h a t we dispense altogether w i t h such bothersome notions as problem formulation has become more evident as a result of this discussion.
However the degree of s u p e r i o r i t y is not exaggerated.
Most
issues
problem
information
are
sufficiently
sufficiently
imperfect,
difficult,
and
we are
forced
that
our to
relevant rely
on
i n t u i t i o n to a very large e x t e n t regardless of how much we wish to rely instead on conscious reasoning about social welfare consequences. C e r t a i n key aspects of broad u t i l i t a r i a n i s m serve to distinguish it from the many specific formulations which have been set up as straw men by the c r i t i c s of u t i l i t a r i a n i s m . social welfare
F i r s t , the individual u t i l i t y and
consequences of any given socia! policy or individua!
behavior pattern rarely can be predicted w i t h any degree of c e r t i t u d e . These consequences have to be regarded as random variables, and the most t h a t a given policy or act can a f f e c t is the probability d i s t r i b u t i o n over the range of conceivab!e outcomes.
Desirable oo!icy or behavior is
defined as that which has a positive e f f e c t on the expected value of social welfare. referred
Second, it is to be stressed t h a t the social welfare
to is the sum of instantaneous socia! welfares over a fair!y
substantial period of time
extending into
the future.
The relevant
c r i t i c i s m is not expected social welfare at the present (or instantaneous) moment,
but t h e
extended
period
utilitarianism forms
they
function:
sum of instantaneous social welfares over a fair!y into
the
future.
Third,
the broad
formulation
of
ass~rts only t h a t relationships exist, not what specific take
on.
One
the re!ationship
b..y
relationship
that trans!ates
is the
social
individual u t i l i t i e s
welfare into
a
54 measure of social welfare.
Broad utilitarianism does not dictate~ for
example~ that this function follows the Bentham sum-of-utilities form. Neither
does it
dictate
that
in
the
individual
utility
functions,
consumption of material goods and services is the only determinant~ the most
impo;'tant
determinant~
d e t e r m i n a n t of individual utility.
nor
even
a
particularly
important
Our knowledge is far too limited at
this point for confident specification of the e x a c t forms of the r e l e v a n t functions.
For this reason~ utilitarian evaluation of social policies and
individual moralities must be regarded as problems in decision-making under uncertainty--a high degree of uncertainty. The formulation of broad utilitarianism proposed herein includes a sharply
defined
distinction
between
simple
goodness
and
moral
goodness. Judgments on morality9 as oDoosed to judgments on the more general concept of desirability (all morally good actions are desirabl% but not all desirable actions can be described as morally good)~ are confined to individual behavior as opposed to social policy. For a given action by a given individual to be judged morally good or morally bad~ the judge must perceive that the individual in question perceives a conflict between the course of action that would most benefit his or her own long-term expected individual utility and the course of action that would most benefit long-term expected social welfare.
The morally
good course of action is the latter course of action (which benefits longterm expected social welfare); the morally bad course of action is the former course of action (which benefits long-term expected individual utility).
It has been stressed that judgments of moral goodness and
badness are very subtle and difficult because the whole purpose of social organization and indoctrination is~ in one sense~ to achieve as close a correspondence as is possible between the self-interested behavior which maximizes
long-term
beneficial
behavior
welfare.
expected which
individual
maximizes
utility
long-term
and
the
expected
socially social
Any reward of any form for a morally good act makes it more
difficult to impute moral goodness to people who act in the same way under the same circumstances at a later point in time.
Similarly~ any
punishments for morally bad behavior makes it more difficult to impute
55 moral badness to similar behavior at a later point. This is not meant to suggest the limited relevance of the concepts of good and evil, however, it does suggest limited relevance for the concepts of moral good and moral evil.
For example,
it is now a
virtually universal judgment that Adolf H i t t e r was extremely evil:
that
his activities from 1920 to 1945 had a severely adverse effect on worldwide social welfare. difficult
But was he morally evil?
question.
Hitter
apparently
sincerely
This is a far more believed
that
his
activities served social interests as he defined them. We are tempted to label Hitler's version of the social interest insane, although this would have the uncomfortable implication that a very large proportion of the German population during Hitler's regime were in fact insane. [t may be more satisfactory simply to hold that Hitler drastically and disastrously misperceived the true social welfare function and also the constraint system relevant to its maximization.
13ut to the extent that we grant
such a misperceptton~ it makes a judgment of tmmora|ity on H i t l e r more problematic. Tt
has
utilitarianism
been
stressed
throughout
the
discussion
that
broad
does not provide pat solutions to complex problems of
social policy and individual
morality.
To many non-utilitarians, the
stereotype of a utilitarian is a wide-eyed and somewhat simple-minded enthusiast for various rather cockeyed schemes for social improvement (e.g., solar Dower as the solution to the energy problem). conjecture
is
that
simple-mindedness
utilitarians than among non-utilitarians.
is
no
more
My own
endemic
among
There is room for an infinite
amount of sophisticated disagreement among utilitarians. The only thing that
utilitarians
have
in
common
is their
explicit
admission that
everything has to be evaluated--ultimately--in terms of its effects on individual u t i l i t y and thence on social welfare.
But what is the overall
effect of a given social policy or individual behavior pattern on social welfare?
We may realize that there is only one correct answer to this
question, but many different
people propose many different answers.
For example, in an article in Philosophy and Public A f f a i r ,
Douglas P.
Lackey presents a utilitarian argument to the effect that the United
56 States should commence unilateral nuclear disarmalnent.
At the risk of
doing considerable injustice to Lackey's elaborate argument, it appears to offer little or no substantive improvement over the "better dead than red" argument.
But we are not being confronted with a forced choice
between Redness and deadness. There is a third alternative: being alive and non-Red. It is true that we are risking our lives in order to preserve this third possibility.
But we risk our lives whenever we get into an
automobile or an airplane,
the problem may be set out as follows.
The
three possible outcomes, in the order of decreasing value of social welfare, are: world
([) peace, mutual disarmament, and mutual respect; (2)
suzerainty
holocaust.
of
Communist
social
systems;
and
(3)
nuclear
Two possible policies open to the United States are-
(a)
maintenence of nuclear weaponry equivalent to that of the tJ.S.S.R4 (b) unilateral nuclear disarmament.
Relative to policy and (b), policy (a)
increases the probability of outcome (t), decreases the probability of outcome (2), and increases the probability of outeome (3). [n Lackey's judgment, the social welfare consequences and probability factors are such that policy (b) has a higher expected long-term social welfare value than policy (a).
[n my judgment, the social welfare consequences and
probability factors are. sueh that the first policy (maintenenee of nuclear weaponry equivalent to that of the U.S.S.R.) has a higher expected longterm social welfare value than policy (b). strictly
utilitarian
evaluation--yet
opposed conclusions.
Both Lackey and [ apply
we come
I take this conflict,
up with
diametrically
along with myriad other
similarly based conflicts about proper social policy, to be evidence that utilitarianism does not tend to lead to ~at, simple-minded solutions to complex problems. However, the advantage of utilitarianism is that it appears very plausible
as
formulation
a u~ful
problem
formulation,
and
correct
would appear to be a necessary, although
sufficient, condition for correct problem solution.
problem
clearly
not
Conscious, explicit
utilitarian formulation of social policy and individual morality problems will in all probability increase the reliability of our judgments. Owing to the complexities of the issues which continue to force a large degree of
57 reliancu on i n t u i t i o n rega~'dless of our desire and intention to employ rational cognition, the increase in the r e l i a b i l i t y of our judgments w i l l not be substantial. b e t t e r than no gain.
Still, i t would seem a x i o m a t i c t h a t a small gain is
58 REFERENCES Copter, Robert, "Optimal Tax Schedules and Rates: Mirrlees and Ramsey," American Economic Review (December 1978) pp. 756768. Eisenberg-Berg, Nancy and Neal, Cynthia, "Children's Moral Reasoning About Self and Others: Effects of Identity of the Story Character and Costs of Helping," Personality and Social Psycho_Lo_gy__B_ulle.tin (March 1981) pp. 17-23. Green, Edward, "The Reasonable Man: Legal Fiction or Psychological Reality?" Law and Society Review (February 1968) pp. 241-257. Hamilton, V. Lee, "The Effect of Roles and Deeds on Responsibility Judgments: The Normative Structure of Wrong-doing," Social P_sycholoqy Quarterly (September 1981) Do. 7_37-254. Howe, Roger E. and Roemer, John E., "Rawlsian Justice as the Core of a Game," American Economic Review (December 198[) pp. 880895.
b(aneko, Mamora and Nakamura, IKenjiro, "The Nash Social Welfare Function," Econometrica (March 1979) Io1~.423-435. Lackey, Douglas P., "Missiles and Morals: A Utilitarian Look at Nuclear Deterrence," Phi!osophy and Public Affairs (Summer 198Z) pp. 189-23i. Mirrlees, James A., "Optimal Taxation: An Introduction to the Literature," Journal of Public Economics (.November 1976) pp. 327358. Ng, Yew-Kwang, "Bentham or Nash? On the Acceptable Form of Social Welfare Functions," Economic Record (September 1981) pp. 238-250. Ordover, James A., "Redigtributing [ncomes: Economic Inquiry (April I98i) pp. 3)3-349.
Ex Ante or EEx Post,"
Rest, James R., Developments in Oudqinq Moral Issues, Minneapolis, I979. Staub, Ervin, "Social and Personal Influences," Positive Social Behavior and Morality, Volume I, New York, I978. Staub, Ervin, "Socialization and Development," Positive Social Behavior and Morality, Volume If, New York, i979.
59 Tare, C. Neal, "Personal Attribute Models of the Voting Behavior of U.S. Suoreme Court 3ustices: Liberalism in Civil Liberties and Economic Decisions, 1946-I978," American Political Science Review (3une i981) pp. 355-367. Yaari, Menahem E., "Rawls, Edgeworth, Shapely, Nash: Theories of Distributive 3ustice Re-Examined," 3ournal of Economic Theory (February I981) pp. 1-)9.