a consumer prefers some specific quantity of a good or service to some specific quantity of the same or another good or service. One alternative is considered to satisfy a want that is more urgent than that which could be satisfied by the rejected alternative. The relatively greater want-satisfying power of the first good or service is called its greater utility. Of course, “want-satisfying power” springs from some quality, real or imagined, associated with the use or enjoyment of the good or service. The utility of coal springs from its heating powers, that of a painting from its artistic merits; the utility of a shoeshine is associated with an appropriate glossiness, the utility of a textbook with the knowledge it confers. But all these are the specific qualities that characterize goods or services on the basis of which one good is preferred over another. Acting man considers these “objective” qualities of the goods among which he chooses; he weighs, with more or less expert knowledge, the relative objective merits of the goods and then arranges them on one scale—the scale of preference.
Man cannot “objectively” compare the glossiness of a newly shined pair of shoes with the thermal capacity of a quantity of fuel, but he must sometimes choose between them. When he chooses he is arranging them in order of “importance.” There is a homogeneous common denominator that makes it possible to compare them: that of their relative positions on the utility scale. The one is more urgent, significant, and important than the other. The economist, concerned exclusively with the logic of choice, needs only to be indirectly conscious of the “objective” qualities of goods. It is not the intensity of these qualities, but the degree of subjectively felt significance with which the law of diminishing marginal utility is concerned, and from which demand analysis takes its start.
Several corollaries follow immediately from the establishment of the subjective character of utility. Most important is the implication that the utilities of the same good for two different people cannot be compared. This, it is noted, is saying considerably more than that it is possible for the same good to have different utilities to two different people. It is even saying more than that there is no conceivable way of discovering for which of two people a given good has more utility. The impossibility of interpersonal utility comparisons implies that no meaning at all can be attached to a statement comparing the utilities of the same (or different) goods to two people.
Utility refers to relative position on a value scale. A good of greater utility is higher on the scale and thereby preferred over a good of lower utility. There is no single value scale on which a specific “good-for-A ” can take up a position relative to a “good-for-B ”; there is no conceivable act of “choice” that should “prefer” a good for A rather than for B.
The impossibility of comparing the utility of a good for two people does not affect, of course, the fact that each of us frequently engages in comparisons concerning the relative “usefulness” of a good to different people. We say that a hungry man “needs” food more than one who has just dined. We try to give charity “where it will do most good”; we distribute gifts among our friends or children where we think they will be the most useful or pleasurable. All this is quite in order, but it does not involve any comparisons of that utility demand analysis depends upon. An outsider C is entitled to his opinion, however irrational, as to how a quantity of a good “ought” to be shared out between two other people, A and B. Frequently he does so by placing himself mentally in the positions of both these people simultaneously. But it is always his choice, always his assessment of relative “urgency,” which operates in such decisions.
Another, and a related, implication of the subjective character of utility is that utility must be clearly distinguished from both ethical values and psychological pleasure-pain sensations. As far as ethics is concerned, the matter is straightforward. In studying demand, we are interested in the patterns of action that follow from given tastes, no matter what these tastes may be. Utility refers to the importance attached by man to possession of goods. What degrees of importance a man attaches to different goods is indeed a matter determined in part by his ethical values. But just as the economist analyzes the demand for coal not by reference to its technological thermal capacity but to the subjective significance that men attach to coal (of course chiefly on technical, objective grounds), quite similarly the economist analyzes the demand for goods (flowers or bullets, knowledge, or liquor) by starting out in a quite “positive” way, and without the need for any moral evaluation, from men’s demonstrated preferences.
The distinction between the utility used in demand theory and pleasure-pain sensations should be equally clear cut. The distinction must be especially stressed because many of the earliest expositions of utility analysis did fail, in fact, to recognize such a distinction or were phrased as if they failed to do so. This failure was both unfortunate and unnecessary. The utility of a loaf of bread, insofar as demand theory is concerned, is not to be identified either with the hunger pangs suffered for lack of the bread or with the sensation of satiety experienced upon its consumption. These sensations may be “real” and important enough, but like ethical values, underlie the preferences that men reveal in their actions. A man’s value scale and the utility to him of given commodities are doubtless dependent on the intensity of these sensations. But the economist must be satisfied to commence from the colorless fact of preference.
Utility as a Relative Concept
We conceive of utility as a purely relative notion. In saying that a good has utility to a man, we mean that it possesses importance, or significance, to him because of its power to remove uneasiness. As we have seen, “importance” and “significance” take on meaning only in the context of a comparison with other goods. Utility reveals itself only in acts of choice when two or more goods are being compared. Thus, it is quite meaningless to conceive of the utility of a loaf of bread, as it were, in a vacuum. All we can say is that a loaf of bread may have either more or less utility than a glass of beer, a news magazine, or twenty cents.
If utility could be identified with some “objective” property of a good, say its mass, calorific value, or even moral worth, then the concept would not depend on the relationship between one good and another. But the utility of demand analysis refers to none of these objective qualities and does, therefore, by its very definition, imply a comparison with other goods or services. Utility refers to position on a scale of values. Without other goods or services, there is no scale of values and hence no utility concept at all.
The relative character of utility means that men’s preferences can be the subject of interpersonal comparisons. There is, as we have seen, no value scale upon which the relative positions of a loaf-of-bread-for-A and a loaf-of-bread-for-B can be observed. But it may be possible for an observer to discover whether a loaf of bread bears the same relationship to twenty cents on A’s scale of values as it does on B’ s; and it may be possible to assert that a loaf of bread has greater utility to A than twenty cents, but that for B the situation is reversed. In fact, this kind of assertion is, as we shall discover, the foundation of market theory.
The Ordinal Character of Utility
Two conflicting approaches to utility theory are met in the literature. The older (but by no means extinct) approach was to treat the utility of a good for an individual as a magnitude to which, in principle, a cardinal number could be assigned. An apple has, let us say, 10 units of utility; a shirt, 50 units; and so on. Such an approach involves the postulation of a numerical scale of utility against which the utilities of goods might—again only in principle—be measured with precision.2 The theoretical concept of numerical quantities of utility involves, again, the notion that a larger “quantity of utility” (one, that is, comprising a larger number of “units” of utility) is built up through the addition of smaller quantities of utility or of units of utility. A good with utility of 10 possesses 10 times the utility of a good with unit utility;
