VI
1. Developments on an entirely different track sug-
gest that economists have never abandoned the hope
of proving that utility is after all cardinally measurable.
A remarkably elegant scheme for determining a cardi-
nal scale for utility was developed by Frank P. Ramsey
(1926) and, independently, by O. Morgenstern and John
von Neumann (1944). The scheme follows Bernoulli's
idea in reverse. Bernoulli determined the stake of the
gambler from the knowledge of the gambler's utility
function. The authors just mentioned proposed to de-
termine the utility function of an individual with the
aid of the odds which that individual would be just
willing to accept. They reasoned that if an individual
is just willing to pay five dollars for a lottery ticket
that gives him one chance in two of winning twelve
dollars, we can infer that for that individual the utility
of the additional seven dollars is equal to the utility
of the five dollars that he may lose. By experimenting
with various gambling propositions, we may thus de-
termine step by step how great is the utility of the
individual's n-th dollar in comparison with all his other
dollars. This is tantamount to constructing a utility
scale which, like that of temperature, is completely
determined once the origin and the unit of measure-
ment are chosen arbitrarily.
The idea was received with enthusiasm as well as
with strong reservations. Some authorities on the sub-
ject of utility openly doubted their ability to construct
a cardinal scale even for their own utility. Various
suggestions have been made as to what piece of the
theoretical apparatus may be at fault. Most probably,
the culprit is the assumption that an individual may
be perfectly indifferent between a dollar in hand and
the probable prospect of winning ten dollars. Indeed,
this assumption overlooks the fact that risk adds a new
and irreducible dimension to man's choice (Georgescu-
Roegen, 1954a).
Whatever the fault, the operational feasibility of the
project has never moved beyond the paper-and-pencil
stage and Ramsey's new vision of a psychogalvanom-
eter had no better fate than Edgeworth's hedonimeter.
2. The conclusion of a recent approach to the proc-
ess of choice is that utility is not even ordinally meas-
urable (Georgescu-Roegen, 1954a). The point of de-
parture is Menger's framework: man chooses in
response to his varied wants and in accord with the
hierarchy of these wants. But the new approach notes
that at the bottom of the hierarchy are the most urgent
wants, which are grounded in the biological nature of
man and consequently are ordered alike for all human
beings. These are followed by the social wants, which
have the same order for all persons belonging to the
same culture. Lastly, there are the personal wants;
these vary irregularly from one individual to another.
Moreover, only some of the personal wants may possi-
bly compensate for each other. That is, a person might
be just as happy with more records and fewer movies.
But biological and social wants are irreducible. He who
does not have enough to eat cannot satisfy his hunger
by wearing more shirts.
In this framework, choice is determined by the least
important want that can be satisfied in the given situa-
tion. For example, a person who does not have enough
food will prefer the grocery basket with the greatest
food value. But between two baskets with the same
food value, the same person will choose the one with
tastier foods, taste thus becoming the least important
want. Should the baskets differ only with respect to,
say, packaging, the next want will come into play.
Choice, in this case, completely orders all possible
baskets, but there are no completely indifferent baskets.
The result is that choice no longer yields indifference
lines and hence cannot be represented by an ophelimity
index.
Several criticisms may be levelled against this ap-
proach. There is, first, the fact that wants are not
sharply defined notions and hence fit poorly into a
quantitative framework. Blurred as the hierarchy of
wants may be, it offers a legitimate, objective basis for
the interpersonal comparison of welfare—without
which taxation must remain a completely arbitrary
operation. It also justifies, for example, such useful
analytical tools as the distinction between wage goods
and luxury goods.
The preceding picture is foreign to the utility theory
launched by Gossen, Jevons, and Walras. This theory
has instead accumulated an impressive mathematical
arsenal around the idea of the complete reducibility
of wants, which is tantamount to the assumption of
complete substitutability among commodities. The ul-
timate product needs unparsimonious stressing: the
modern utility theory reduces all wants to one general
abstract want called “utility.” In line with this reduc-
tion, one need not say “these people need more shoes”;
instead, “these people need more utility” should suffice.
The reduction is responsible for the fact that the same
theory teaches that there is no objective basis for inter-
personal comparison of utility. All this may again be
due to a particular feature of the economies in which
the builders of the modern theory of utility lived. Those
were not economies in which a low income kept basic
wants in front of everybody's eyes; they were econo-
mies where most people were able to satisfy even many
personal wants. Modern utility theory is a theory of
a consumer who has a relatively ample income and
whose economic choice is guided only by the quantities
of commodities.
