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.