then, that the product of a given quantity of ‘labor’ increases with every increase in Jevons’ ‘amount of investment of capital.’ But he also postulated that this increase proceeds at a decreasing rate. 37 This amounts to setting up a law of decreasing (physical) returns that is formally analogous to the law of decreasing marginal productivity of any other factor: in the arithmetical tables by which Böhm-Bawerk illustrated his ideas (see, e.g., Kapital und Kapitalzins, II, p. 463 et seq. of the 3rd ed.), he ‘applied,’ as it were, successive units of time to a given amount of resources (actually: a month’s labor). The restriction involved is stronger than necessary but made the going much easier for Böhm- Bawerk than it would have been without it. In appraising the standing of this postulate we must, however, never forget this lesson from the practice of the physicist: a postulate may be justified not only by establishing observationally the facts it asserts but also by its results. Finally, Jevons’ ‘amount of investment of capital’—which has a time dimension— divided by his ‘amount of capital invested’—which has no time dimension—gives Böhm-Bawerk’s famous Period of Production. This quantity is to characterize, by a single figure, the structure of production, if possible of total national production, and to serve as the fundamental variable of capital theory. Formally, it represents a gravitational center. Think of n particles of the masses m 1 , m 2 ,…m n that lie on a straight line. Choosing this line for axis and denoting the co-ordinates of the particles on this axis by x 1 , x 2 ,…x n , we find that the co-ordinate X of their center of gravity is let us, but only in order to simplify the argument, further assume that productive resources are given, constant, and (this is really a pleonasm) optimally allocated. The only reason why in these conditions there could be methods of production that are known to be ‘superior’ to those in use, and yet are not being used instead, evidently is that these superior methods cannot be ‘financed’ in the Jevons-Böhm sense. But, since under the conditions assumed, the wage-good capital is being fully and optimally used, the only reason, in turn, for that impossibility is that the superior methods would ‘lock up’ too much capital for too long a time. Now let the wage-good capital increase, all other requisites being kept constant. Everyone would agree that this will tell primarily in favor of new investment in productions involving a longer ‘period.’ This is all that is necessary. That part of the increase will be absorbed by wages, hence go toward financing production for immediate consumption, is so far from being an objection as to be even an essential part of Böhm-Bawerk’s theory of wages. In extenuation of the sins of the critics it should be admitted, first, that with the printer’s devil waiting at his door, in 1888, for new batches of manuscript, he put his argument unsatisfactorily from the start, and, second, that when criticisms poured in upon him, he frequently mismanaged his defense (particularly in the matter of inventions). 37 Denoting physical product by p, means of production by a, b, c,…and time by t, we have then p=f(a, b, c…t), that is to say, the idea really reduces to the introduction of some kind of time factor into the production function (see below, ch. 7, sec. 8). Böhm-Bawerk’s postulates were (1) that and (2) that History of economic analysis 872 Now let the m’s represent, instead of n masses of particles, n quantities of physical resources that are being successively applied, at n points of time, t 1 …t n , to the production of a consumers’ good that, after another spell of time of storage, 38 is sold and consumed. This set-up inevitably imposes upon us either the necessity of identifying these physical resources with a single homogeneous agent—Böhm-Bawerk chose, like Jevons, homogeneous units of labor 39 —or else the assumption that these resources consist of doses of invariant composition. For axis we now choose time instead of distance, and for zero point on this axis the point of time at which the consumers’ good is sold. Evidently the t’s are all to the left of this zero point, hence negative, and decrease numerically as we proceed from the first act of investment in t 1 to the right toward the zero point. The expression 40 which has only a time dimension (since the resource dimension cancels out), is Böhm- Bawerk’s period of production. The phrase is most infelicitous—it is difficult to think of a more infelicitous one—and to a considerable extent responsible for the flood of adverse criticisms. But the meaning of the thing itself is clear: it is the average of the time distances from the sale of the products of all the units of ‘invested labor.’ 41 The need for comments on this theory of capital (in addition to those that have been made already) is much reduced by the fact that Professor Knight, by far the most eminent of its critics, has admitted that, under all the assumptions made by Böhm-Bawerk, it is valid. 42 First of all, it must be emphasized 38 This is a slight deviation from Böhm-Bawerk’s set-up. 39 Strictly, it should be ‘subsistence’ considered as a homogeneous quantity. Obviously, however, he did not like to go as far as this. 40 The expression is made positive by prefixing a minus sign, since the t’s enter negatively. The reader will see that it is quite reasonable to consider ‘investment’ as an essentially negative quantity from the standpoint of the ‘investor’: it is something he gives away. T itself should be positive, however. 41 In his Rate of Interest, Irving Fisher asked the question why that weighted average should be considered as the ‘correct’ method of measuring the period of production. This question evidently embarrassed Böhm-Bawerk greatly (see Exkurs III in the 3rd and 4th eds. of Kapital und Kapitalzins, II) but it should have been easy to answer. In fact it should never have been asked: for the formula simply defines something that Böhm-Bawerk chose to call the period of production. 42 This short cut has, of course, its dangers; in addition it deprives the reader of the benefits of exercise in the art of theorizing that he conceivably might reap from a fuller discussion. In order to make up for this, I refer the reader to Mr. Kaldor’s survey article, ‘The Recent Controversy on the Theory of Capital,’ Econometrica, July 1937, the first three footnotes to which present a bibliography for the 1930’s, including of course the papers of the protagonist, Professor Knight, to which I have only to add (except Knight’s ‘Reply,’ ibid. January 1938, and Kaldor’s ‘Rejoinder,’ ibid. April 1938) F.Burchardt, ‘Die Schemata des stationären Kreislaufs- General economics 873 again that here we are not concerned with Böhm-Bawerk’s theory of interest or with the bearings upon it of any of the elements of his capital theory. This makes a lot of difference. As an example, consider the argument that in a synchronized process in which production and consumption run along continuously, with all its elements perfectly co- ordinated, the idea of periods of production ceases to have any importance or even meaning and production may safely be treated as timeless. Now, it may be true that in such a process the period of production ceases to have any importance in the explanation of interest. 43 But that is not the same as saying that the concept has no use or even meaning in such a process. Even in Clark’s waterfall, assuming it to be perfectly steady, we may, for example, try to define the time which a drop of water takes on the average to get from top to bottom, which would be a method, though a very imperfect one, of describing some of its properties. Similarly, Böhm-Bawerk’s period of production would, if his assumptions be accepted, express one of the most meaningful characteristics of an economic process, however ‘cycleless’ it may be. This has been shown, in one of the none too numerous constructive contributions to Böhm-Bawerk’s capital theory, by Professor Marschak. 44 In the second place, we must bear in mind Böhm-Bawerk’s technical disabilities, which resulted in his idea’s being much more open to formally successful attack than it would have been if presented in a stronger armor. This armor has been strengthened, however, by several writers, especially by Gifford 45 and Marschak. In the third place, and independent of technical blemishes, it must not be forgotten that the concept of period of production, as framed by Böhm-Bawerk, was only a device to express an aspect of the economic process and that it neglected all others, which is what Wicksell meant when he said that Böhm-Bawerk’s capital theory was so ‘abstract’ as hardly to bei Böhm-Bawerk und Marx’ (Weltwirtschaftliches Archiv, October 1931 and January 1932); W.Eucken, Kapitaltheoretische Untersuchungen (1934); and J.M. Thompson, ‘Mathematical Theory of Production Stages in Economics,’ Econometrica, January 1936. F.A.von Hayek’s Pure Theory of Capital (1941) not only presents Professor Hayek’s latest views but also sheds interesting light (in Part I) on the capital controversy. Of earlier criticisms I shall mention only Fisher’s in his Rate of Interest and von Bortkiewicz’ ‘Der Kardinalfehler der Böhm-Bawerkschen Zinstheorie’ (Schmoller’s Jahrbuch, 1906). The interesting thing about the latter is its spirit of uncompromising hostility that differs so strikingly from the spirit he displayed in his famous critical pieces on Marx. Böhm-Bawerk’s not wholly felicitous reply to both is in the 3rd and 4th editions of Kapital und Kapitalzins. 43 Though I raised this point myself 40 years ago, I do not consider now that it is well taken. 44 Jacob Marschak, ‘A Note on the Period of Production,’ Economic Journal, March 1934. His reasoning seems to run in value terms, but in the ratio of total value of existing stocks of commodities to the value of the flow of finished consumers’ good the value dimension cancels out. 45 C.H.P.Gifford, ‘The Concept of the Length of the Period of Production,’ Economic Journal, December 1933. Marschak, op. cit. History of economic analysis 874 constitute even a first approximation to reality. For both reasons the whole construction no doubt looks gaunt, not to say freakish. Some of the features that account for this impression can be removed without great difficulty. Jevons knew better than to let labor be added to a growing intermediate product until a finished consumers’ good emerges that is consumed at once: as stated already, he included the process of ‘uninvestment’ so that his period was not simply a period of production. Böhm-Bawerk himself, inspired by Rae, added the gradual using up of durable consumers’ goods. Wicksell showed how services of natural agents might come in along with labor. His pupil, Professor Ākerman, also inspired by Rae, dealt in one of the most important works in this field, with the problems of fixed capital, which is so curiously absent from Böhm-Bawerk’s schema. 46 One of those features of Böhm- Bawerk’s schema that seemed most ridiculous to critics—that his period of production seems to start from a state in which all production proceeds without any tools or materials at all and people catch fish with their bare hands—can be removed so soon as we realize that all economic theory is a theory of planning and inevitably has to accept the results of the past—plant, equipment, and stocks all included—as data. We shall then cease to try to construct an economic process ab ovo and, looking forward only, consider instead of the ‘amount of investment of capital,’ the ‘amount of investment to be done.’ 47 Incidentally, this would also remove one of the motives for ‘resolving’ all capital goods into ‘labor and land’ or labor alone. Similarly, we can get rid of the ‘linearity’ of Böhm- Bawerk’s schema of production—of the idea that all products emerge as the result of processes during which, in each intermediate stage, nothing is added to the result of the previous stage but labor. Nor does it seem impossible to me to derive from the ‘periods’ of individual firms the social period of production that is needed in Böhm-Bawerk’s theory. But no satisfactory answer that I can see exists to another objection. 46 Gustaf Ākerman, Realkapital und Kapitalzins (1923–4). See on this Wicksell’s comments, republ. in the 2nd appendix to his Lectures, and Ākerman’s own partial reformulation in: Om den industriella rationaliseringen…(1931). Erik Lindahl’s contribution (available in English under the title, ‘The Place of Capital in the Theory of Price,’ as Part III of his Studies in the Theory of Money and Capital, 1939) complements this from a different standpoint, located nearer to Walras. 47 It is a question of some interest why Böhm-Bawerk did not do this. I think that the answer is to be found in a curious attitude that was common to all the Austrians. They were never content with explaining a state of a process by the preceding states of the same process. They suspected circular reasoning in any argument that did this—or at least a begging of the fundamental question. Any truly ‘causal’ explanation had to be ‘genetic.’ It had to uncover the (logical) origins of things. Hence the capital concept had to be developed from conditions in which there was no capital. As it was, Böhm-Bawerk relied, for the purpose of getting a reasonably short period, on an ad hoc postulate, namely that, as we go back into the history of a given industrial process, the quantities of resources that have been applied in the past but are still present in it (iron mined in Roman times that may be still present in a modern pocket knife) decrease with time much more rapidly than the time increases with which they have to be multiplied. General economics 875 In order to make Böhm-Bawerk’s conception of the structure of capital serve his analytic intention, this structure must be a physical fact; and the different quantities of product that different time-structures turn out must be comparable physically. In order to secure the first requisite, we need indeed a physically homogeneous resource, the elements of which differ in nothing but the time dimension; in order to secure the second requisite, the products that enter Böhm-Bawerk’s tables must all be the same in kind and quality, differing in nothing but physical quantity. Neither requirement can be fulfilled except in special cases. And it is this which reduces the analytic value of Böhm-Bawerk’s capital theory, so far, to such value as may still attach to the non-operational illustration of an aspect of reality. 48 But, so the reader might well ask, if we recognize all this and if we introduce all those corrections, what is left of Böhm-Bawerk’s capital theory and in particular of his period of production? Well, nothing is left of them except the essential idea. And this keeps on proving its vitality by every piece of criticism and every piece of constructive work it evokes. 49 3. THE REVOLUTION IN THE THEORY OF VALUE 1 AND DISTRIBUTION In this section, we shall try to formulate, in an entirely elementary manner, what this so- called revolution consisted in and what difference it made to economic analysis. For this purpose, we shall adopt the language of the marginal utility theory in its original and most uncritical form. And we shall use primarily the Austrian edition of it, because the Austrians (Menger, Wieser, Böhm-Bawerk), in spite of their defective technique, succeeded in bringing out certain fundamental aspects more clearly than did either Jevons or Walras. Marshall’s teaching will come in, as an instructive contrast, both in this section and in the next chapter, where we shall move on the higher level of Walras. 2 48 Situations like this are more frequently met with in economics than one might think. Marx’s system offers several examples. Another is Professor Pigou’s Pound of Resources, which he first introduced (1st ed. of Wealth and Welfare) and then dropped. Still another is Marshall’s baskets of goods that he used in his theory of international trade. Perhaps the problem is not insoluble. (See, e.g., W.Leontief’s article, ‘Composite Commodities and the Problem of Index Numbers,’ Econometrica, January 1936.) In any case, the examples adduced show that arguments that labor under the difficulty alluded to are not necessarily valueless. 49 A recent constructive reinterpretation of Böhm-Bawerk that has not been mentioned yet is presented in J.R.Hicks, Value and Capital (1939), ch. 17. It is not in Böhm-Bawerk’s spirit. But it proves that Böhm-Bawerk’s ideas worried Professor Hicks. Professor Douglas unintentionally erected a monument to Böhm-Bawerk’s memory in Chart 9 of his Theory of Wages (1934), p. 128. 1 Instead of using this traditional phrase, I might also have used ‘exchange ratios’ or ‘relative prices.’ For most of the purposes of that period’s theory, these three terms all mean the same thing. 2 Only a few essentials will be discussed and no systematic attempt will be made to reproduce and criticize the phrasing of individual authors. For more complete analysis History of economic analysis 876 The history of the marginal utility theory itself and of its successors will be treated in Chapter 7. But we need a few elements of it right now. Menger started from what he supposed were the obvious facts of human wants, which he formalized in this way: first, there are different categories of wants or tastes or desires (Bedürfniskategorien), such as the desire for food, for shelter, for clothing, and so on, which define the concept of Goods and can be arranged in a definite order of (subjective) importance; second, within each of these categories of wants, there exists, given as a psychic reality, a definite sequence of desires for additional increments of each good (Bedürfnisregungen), which we experience as we go on consuming successive increments. Menger illustrated this by a numerical table that has been reproduced by Professor Stigler (op. cit. p. 144), and dealt carefully with the large number of questions that arise in connection with the schema— such as how far wants may be treated as data in spite of their expansibility and malleability. Neglecting these questions, we proceed at once to state the postulate—or ‘law’—that was fundamental to the ‘new’ or ‘psychological’ theory of value: as we go on acquiring successive increments of each good, the intensity of our desire for one additional ‘unit’ declines monotonically until it reaches—and then conceivably falls below—zero. Or, replacing Menger’s discrete figures by a continuous curve or function, and the phrase ‘desire for one more unit’ 3 by Marginal Utility: ‘The marginal utility of a thing to anyone diminishes with every increase in the amount of it he already has’ (Marshall, Principles, p. 168). Waiving various objections, we may define from this (as a sum or integral) the concept of Total Utility and then also say that the total utility of a thing to anyone increases, up to the point of satiety, with every increase in the amount of it, but at a decreasing rate. In either form this is what Marshall called the Law of Satiable Wants and what the Austrians called Gesetz der Bedürfnissättigung. In honor of the most important ‘forerunner’ it is also called Gossen’s First Law. 4 We add immediately the proposition which is—or should be—called Gossen’s Second Law. Unlike the first it is not a postulate but a theorem: in order to secure a maximum of satisfaction from any good that is capable of satisfying different wants (including labor or money), an individual (or household) must allocate it to these different uses in such a way as to equalize its marginal utilities in all of the reader is referred to Professor Stigler’s book, Production and Distribution Theories (1941). 3 Instead of unit, Menger said Teilquantität, by which, as Professor Stigler points out, he meant a small but finite increment. If, for the sake of analytic convenience, we use continuous and analytic functions, we mean infinitesimal increments. The word ‘unit’ is never strictly correct. The phrase Marginal Utility is Wieser’s (Grenznutzen). Jevons said Final Degree of Utility, Walras rareté. The latter term, which we will surely translate by scarcity, suggests that there is little to the case of objectors who, like Cassel, wished to discard utility but nevertheless to retain scarcity. The turn of phrase ‘desire for one more unit’ is Fisher’s, who also spoke of ‘wantability.’ Pareto introduced the term ophélimité élémentaire; Clark, the term Specific Utility. 4 See below, Appendix to ch. 7, ‘Note on Utility.’ General economics 877 them. 5 At first sight both statements are nothing but somewhat technical renderings of sad trivialities. But we must not forget that the proudest intellectual structures rest on trivialities that are entirely uninteresting in themselves. What could be more trivial than that a body at rest will remain at rest unless something (a ‘force’) acts to set it in motion (Newton’s First Law)? Let us, then, look at the structure that was erected on those trivialities. (a) The Theory of Exchange Value. The first problem that Jevons, Menger, and Walras—Gossen too—tackled by means of the marginal utility apparatus was the problem of barter. Like their ‘classic’ predecessors, they realized the central position of exchange value although, also like these predecessors, they did not make it sufficiently clear to their readers and perhaps did not sufficiently realize themselves that exchange value is but a special form of a universal coefficient of transformation on the derivation of which pivots the whole logic of economic phenomena. 6 Their barter theories or, to use Whately’s term once more, their catallactics, differed greatly as regards technical perfection and correctness: the peak achievement of the period is contained in leçons 5–15 of Walras’ Éléments. 7 But they all—also Gossen’s—aimed at the same goal, which was to prove that the principle of marginal utility suffices to deduce the exchange ratios between commodities that will establish themselves in competitive markets and also the conditions under which ranges of possible exchange ratios must be substituted for uniquely determined ones. In other words, they established what A.Smith, Ricardo, and Marx had believed to be impossible, namely, that exchange value can be explained in terms of use 5 Menger spoke of the different wants that a good is capable of satisfying as being ‘satisfied to an equal level of urgency.’ This is all right so far as it goes. But it is interesting to note how hesitant many writers were about this Second Law. 6 In consequence, history-minded or sociology-minded critics understood still less what these economic theorists were about. Taking at face value the simplified schemata by which the latter introduced their subject, and finding, e.g., that these schemata dealt with the bartering of consumers’ goods that were present in given quantities, these critics wondered what could be the relevance of such analysis, not only to the great problems of social life but also to the really interesting purely economic problems of production and distribution. Veblen’s essay, ‘The Limitations of Marginal Utility’ (republ. in Place of Science in Modern Civilization, 1919) exemplifies this attitude to perfection. 7 Walras was the only one of the three to deal with the case of three and more commodities— involving indirect exchange—and to state anything approaching satisfactory equilibrium conditions in terms of excess demand. Jevons’ chapter on the ‘Theory of Exchange’ is far inferior. Menger’s treatment of the problem is all right so far as it went. But it did not go very far. When Böhm- Bawerk tried to elaborate his theory, defective technique told immediately, and his famous horse market duly came in for easy criticism from Edgeworth: the latter’s most important contributions are in the Mathematical Psychics and, incidentally, in many of his papers. I mention specially the one in the Giornale degli Economisti, March 1891, where the reader also finds an interesting paper by Arthur Berry in the June 1891 number. Marshall’s Principles contain all that the reader needs to know about the barter theory as it had shaped before the century was out (see mainly pp. 414–16, and Appendix, Notes I, II, and XII). History of economic analysis 878 value. 8 Jevons, Menger, and Walras would all of them have approved of this statement. It is this which they meant when they claimed to have discovered the ‘cause’ of (exchange) value. However, even if granted, this would not in itself have amounted to a great deal, especially since the ‘paradox of value’ had been, as we know, resolved a dozen times before. 9 It is more important that the ‘new’ theory of exchange was more general than had been the old ones 10 and that it proved more fertile in results—many of them due to Edgeworth—even in the cases covered by the old ones. But this is not the essential point either. The essential point is that, in the ‘new’ theory of exchange, marginal utility analysis created an analytic tool of general applicability to economic problems. 11 This will become clearer as we go on. (b) Cost, Production, Distribution. The concepts of marginal and total utility refer to consumers’ wants. Hence they carry direct meaning only with reference to goods or services the use of which yields satisfaction of consumers’ wants. But Menger went on to say that means of production— or, as he called them, ‘goods of higher order’—come within the concept of economic goods 8 In Marxist terms, this means that fundamentally the ‘exchange economy’ (Tauschwertwirtschaft) is also a ‘use economy’ (Gebrauchswertwirtschaft), which Marxist orthodoxy denied on principle. Nothing, of course, is gained or lost for socialism whichever view we adopt. But both parties, or at all events the Marxists, thought that there was practical significance to this issue. It may not be superfluous to point out that this issue has logically nothing to do with the issue of modern popular economics, production for use vs. production for profit. 9 See, e.g., above, Part II, ch. 6, sec. 3. But all the same there is point in emphasizing, against one particular form of unfair criticism, that it is not true that A.Smith or Ricardo or J.S.Mill scorned this approach to the economic phenomenon because of its obviousness. The truth is that they did not see how ‘value in use’ could possibly be made to explain ‘value in exchange.’ They saw no further than that the former was a condition of the latter. 10 This can best be realized by comparison with the labor-quantity theory of exchange value. The latter, so we have seen, is not ‘wrong’ as the revolutionaries, especially the Austrians, were in the habit of saying. It is much more enlightening to say that it covers only a special case. Even where correct, the theorem that ‘prices’ of commodities tend to be proportional to the labor quantities embodied in them does no more than state a property of equilibrium prices. It does not give a description of the process that establishes these prices and therefore cannot be called a theory of prices at all, any more than the statement that in certain conditions the price level will be proportional to the quantity of money, even where correct, can be called a theory of money; or the statement that in certain conditions real wages will equal minimum-of-existence requirements, even where correct, can be called a theory of wages. So far as this goes, the revolutionaries did not revolutionize an existing theoretical structure, but they erected one where none had stood before. 11 This was felt but not properly brought out by the revolutionaries themselves. In part this was due to the fact that the tool of general application—the theory of maximizing behavior—appears in the theory of exchange in the guise of a special case. It had never been made to stand out, stripped of all non-essentials (including marginal utility itself), and reduced to its logical fundamentals before the publication of P.A. Samuelson’s Foundations of Economic Analysis (1947; chs. 1, 2, and 3). General economics 879 by virtue of the fact that they also yield consumers’ satisfaction, though only indirectly, through helping to produce things that do satisfy consumers’ wants directly. Let us pause for a moment to consider the meaning of this analytic device that looks so simple or even trite and was nevertheless a genuine stroke of genius. 12 It enables us to treat such things as iron or cement or fertilizers—and also all services of natural agents and labor that are not directly consumed—as incomplete consumable goods, and thereby extends the range of the principle of marginal utility over the whole area of production and ‘distribution.’ The requisites or factors or agents of production are assigned use values: they acquire their indices of economic significance and hence their exchange values from the same marginal utility principle that provides the indices of economic significance and hence explains the exchange values of consumable goods. But those exchange values or relative prices of the factors constitute the costs of production for the producing firms. This means, on the one hand, that the marginal utility principle now covers the cost phenomenon and in consequence also the logic of the allocation of resources (structure of production), hence the ‘supply side’ of the economic problem so far as all this is determined by economic considerations. And it means, on the other hand, that, in as much as costs to firms are incomes to households, the same marginal principle, with the same proviso, automatically covers the phenomena of income formation or of ‘distribution,’ which really ceases to be a distinct topic, though it may, of course, still be treated separately for the sake of convenience of exposition. The whole of the organon of pure economics thus finds itself unified in the light of a single principle—in a sense in which it never had been before. Most of the problems that arise from this set-up can be discussed only on a level on which Walras rules supreme. But, though I believe that Jevons should be credited with a vision of the facts above and if so holds priority, the credit for having worked out that theory systematically, on the plane on which we are moving now, should go to the Austrians and particularly to Menger, whose Grundsätze contain all the essentials. Professor Stigler, indeed, pointed out many a ‘hiatus’ in Menger’s treatment and rightly attributed them to his preoccupation with the threshold problems of the valuation of directly consumable goods. This accounts in fact for the impression that he was neglecting cost aspects. But on Stigler’s own showing, Menger had all the essential results. Nor must we forget that the Grundsätze was, in a sense quite different from that applicable to Marshall’s Principles, intended to be but an introduction. Actually, it was left for Wieser to work out the Austrian theory of cost and distribution explicitly. But he was the worst technician of the three great Austrians. And objections to methods that were peculiar to him crowded upon his readers—and especially Wicksell—so as to impair the effect of what was really a great performance. Böhm-Bawerk expounded, developed, and defended the Mengerian theory of value. But in this field he neither had nor asserted claims to originality. The best formulation of the Austrian doctrine was presented later on by Wicksell. 12 In an embryonic form this device had already been used by Gossen. History of economic analysis 880 If the explanation of the exchange value of means of production is based upon their indirect utility or use value to the consumers of their final products, that is to say, if their economic significance is to be derived from the contribution which they severally make to consumers’ satisfactions, the problem naturally arises how the contribution of each of them is to be isolated, seeing that all ‘factors’ are equally ‘requisite’ for the final product and that complete withdrawal of any one of them will in most cases result in a zero product. The very fact that some German critics continued to urge that this problem was insoluble and that, because it was, the marginal utility theory was inapplicable to the evaluation of any goods other than consumers’ goods present in given quantities, hence inapplicable to production, should suffice to show that here was in fact a real and non- trivial difficulty, removal of which was the prerequisite for the fundamental idea’s becoming analytically operative. Menger removed it by applying the analogue of the method that he had used to resolve the value paradox. He accepted the impossibility of separating the contributions of ‘factors’ to the product that results from their co- operation. But he observed that in order to remove the difficulty it was sufficient to determine their marginal contributions (Wieser’s Grenzbeitrag). 13 And these can be very simply found by withdrawing successively small quantities of each requisite of production, keeping the others constant each time, and ascertaining the loss of satisfaction this will cause to the consumers of the product or products. Some technical points about this procedure 14 will be discussed in the next 13 This should, however, have convinced both the marginalists and their critics that the marginal utility theory of income formation was constitutionally incapable of ‘defending’ the capitalist method of distribution. For it is obvious that the merits—moral or other—of, e.g., the labor factor are not affected if, relatively to the available quantities of other factors, laborers are so numerous that their marginal contribution is small. 14 It is interesting to note some of the troubles that arose both for the Austrians themselves and for their critics from their lack of experience in handling the relevant concepts. Thus, there was a discussion within the Austrian circle whether or not Menger had been correct in taking as guide the loss of consumers’ satisfaction incident to the loss of a small quantity of some factor: some held that we should observe instead the gain that a small increase in any factor might cause. A problem of secondary importance arises indeed in the case of discontinuities. But so long as we are concerned with a first formulation of the fundamental principle, we are within our rights in not bothering about these, and then all qualms on that score are simply due to failure to understand the logic of infinitesimals—of which examples could be cited even from the 1920’s. Again, the objection turned up fairly early that, if you withdraw a small quantity of some factor from a technically adjusted going concern, you will cause a disturbance that will not be smaller than any that would be caused by further withdrawals but, on the contrary, so upset the plan of production as to render the rest of that factor next to useless and the schedule of decreasing productivities of the ‘successive units’ of a factor imaginary. And a curious situation arose in which this and similar objections remained unanswered because some of the marginal utility theorists did not know how to reply and because others who knew did not think it worth while to do so. General economics 881 . criticize the phrasing of individual authors. For more complete analysis History of economic analysis 876 The history of the marginal utility theory itself and of its successors will be. ratio of total value of existing stocks of commodities to the value of the flow of finished consumers’ good the value dimension cancels out. 45 C.H.P.Gifford, ‘The Concept of the Length of the. introduction of some kind of time factor into the production function (see below, ch. 7, sec. 8). Böhm-Bawerk’s postulates were (1) that and (2) that History of economic analysis 872 Now