of ‘profits’ to be roughly equal in different but similarly conditioned lines of business. 4 We derive a similar proposition from the principle of maximizing net returns and associate it with the principle of substitution. The ‘classics,’ it has been held, 5 were not in possession of the latter principle. 6 This is true and so it is that this constitutes one of the most serious shortcomings of their analytic apparatus. But if they did not formulate it explicitly and did not apply it systematically, neither were they entirely unaware of it. They used it in individual cases. And it is implied in some of their propositions. (a) Ricardo and Marx. By theories of value we mean attempts at indicating the factors that account for a thing’s having exchange value or—though this is not strictly the same—the factors that ‘regulate’ or ‘govern’ value. Let us begin with Ricardo. A.Smith, we remember, may be credited with three different theories of value: the labor-quantity theory illustrated by his beaver and deer example; the labor-disutility theory conveyed by his reference to ‘toil and trouble’; the cost theory he actually used in the central part of his analysis. We also know that in addition he recommended labor (along with ‘corn’) as a relatively stable unit by which to express commodity values (numéraire). 7 Ricardo, starting his theoretical work by a study of the Wealth of Nations, was displeased with what he rightly felt to be a logical muddle and came to the conclusion that the labor-quantity 8 theory of value as conveyed by the beaver and deer example was the one to adopt, not only for ‘primitive’ conditions in which there was no scarce factor other than labor, but generally for all cases, even where there were also other scarce factors. His first chapter is an attempt to carry out this idea. A.Smith’s cost theory he evidently thought logically unsatisfactory (perhaps circular). The labor-disutility theory he neglected, probably because it did not occur to him that it was different from the labor-quantity theory. And throughout, he mixed up his argument against A.Smith’s lapse from the labor-quantity theory of value with an argument against A.Smith’s (and Malthus’) choice of labor as a measure of value. 9 Before going on, I shall first try to remove this difficulty from our path. 4 Their concern with differences in the rates of return earned, at the same time and place, in different occupations—discussion of which was, since A.Smith, part of the stock in trade of every textbook—was chiefly motivated by a desire to protect the fundamental assumption of equality. 5 See, e.g., G.J.Stigler, ‘Stuart Wood and the Marginal Productivity Theory,’ Quarterly Journal of Economics, August 1947, p. 647. 6 As has been noted above, Senior made the maximum principle explicit. But it has also been noted that neither he nor anyone else knew how to make full use of it. 7 It cannot be repeated too often that choosing labor for this role—for instance on the ground that the significance of a man-hour is less subject to change than is the significance of an ounce of gold, no matter whether this is so or not—has nothing whatever to do with adopting a labor theory of value. Malthus, for instance, was an opponent of the latter. But he recommended labor days for the purpose of expressing values (for ‘measure of value’). Though this should be quite clear, it is worth emphasizing again because these two things have so often been confused even by first-flight theorists such as Ricardo. 8 Meaning the quantity of labor that a commodity ‘embodies.’ History of economic analysis 562 9 Meaning the quantity of labor a commodity ‘commands’ in the market, which differs in general from ‘labor embodied.’ Two things must be distinguished. On the one hand, Ricardo, like everybody else, was of course aware of the fact that there can be no commodity (labor no more than any other), the exchange value of whose unit could serve as an invariant standard by which to measure the variations in the exchange values of other commodities (Principles, ch. 1, 6). On the other hand, his labor-quantity theory of value seemed, subject to qualifications that will be discussed presently and are neglected for the moment, to provide a method by which to measure these variations all the same: where the exchange value of a unit of labor was bound to be unsatisfactory, the unit of labor itself—since according to this theory the amount of labor embodied in a commodity ‘governs’ its value—really was what was needed in order to have a measure of exchange values after all. Subject to the qualifications we are now neglecting, all that was necessary in order to have a commodity of at least theoretically invariant value was to imagine one that always embodied the same quantity of labor. Such a commodity would then provide a stable yardstick with which to measure the variations in the relative prices of all the others. The pounds and shillings of his numerical examples must be understood to stand for such a commodity. 10 It is very important to grasp the implications of this logical tour de force. By virtue of it, commodities acquired absolute values, which were capable of being compared, added up, and of increasing and decreasing simultaneously, the very thing that was impossible so long as exchange value was defined simply as exchange ratio. This is what pleased Marx so much about Ricardo’s theory of value. But the latter failed to work out the idea completely. Moreover, he created much unnecessary confusion by adopting for his concept the term Real Value. Our own meaning of this term, which refers to the value of a monetary quantity in terms of the goods it will buy, was gaining currency at that time, and people were puzzled by Ricardo’s use of it according to which, for example, ‘real’ wages might be falling (if the quantity of labor em- bodied in the goods that constitute real wages in our sense was decreasing, owing, for instance, to technological improvement) when everybody else would say that they were rising (if the quantities of those goods themselves were increasing). Another point must be mentioned which is of considerable importance for understanding Ricardo’s theory of distribution—which was primarily concerned with relative shares—and in particular his famous theorem that ‘there can be no rise in the value of labour [real wages in his sense] without a fall of profits’ (see, e.g. 4, Principles, ch. 1). The true significance of this theorem will be discussed later. But, in the place referred to, Ricardo reduced it to a triviality by explaining that, if the product be divided between capital and labor, ‘the larger the proportion that is given to the latter the less will 10 Ricardo derived some satisfaction from thus fulfilling Destutt de Tracy’s precept that values should be expressed in units of values as lengths are expressed in units of length. In this, however, he erred. For, whatever we may think of Destutt de Tracy’s precept, a little reflection will show that Ricardo did not satisfy it or rather that he satisfied it only by means of a verbal trick: the values he measured in terms of physical labor hours were not themselves (though they were for Marx) just labor hours. General economics 563 remain for the former’—which is in fact how James Mill and many later interpreters (e.g., A.Wagner) understood the theorem. How was this possible? Evidently Ricardo, when he penned this passage, thought that relative shares are always rendered by the relation between the labor hours embodied in the absolute shares. This, however, is not true generally but only if the total quantity of labor applied is kept constant (On this tangle, see Cannan, op. cit. pp. 341 et seq.) Ricardo, then, tells us on the first page of his work that utility is a necessary condition for the emergence of exchangeable value and that ‘possessing utility, commodities derive their exchangeable value from two sources: from their scarcity, and from the quantity of labour required to obtain them.’ Illogically identifying scarce commodities with commodities, the quantity of which cannot be increased by labor, and setting them down as rare exceptions, he turns to the category of those that may be increased by human industry. I cannot stay—but the reader should—to point out all the shortcomings of this start and shall proceed at once to state the central theorem of the Ricardian theory of value: in conditions of perfect competition (which Ricardo failed to specify) the exchange values of commodities will be proportional to the quantities of labor contained or embodied in them. The first thing to be observed about this proposition, which hails from the Wealth of Nations (Ricardo referred specifically to Book I, ch. 5), is that it is not in itself a theory of value in the sense defined above. Such a theory is contained in Ricardo’s next sentence, ‘that this [i.e. labor applied or embodied, J.A.S.] is really the foundation of the exchangeable value of all things.’ The proposition in question is a theorem on values intended to be valid in perfect equilibrium only. Of this Ricardo was perfectly aware. In Chapters 4 and 30, he therefore dealt with the Cantillon-A.Smith concept of market price, which he made dependent, like the price of monopolized commodities, on supply and demand as if determination of price by supply and demand were entirely different from, and incompatible with, determination of price by quantity of labor embodied. But, not being in full possession of an explicit perfect-equilibrium concept, he expresses this by saying that his labor-quantity law applies to natural prices, that is, to the relative prices that will ultimately prevail when fluctuations due to temporary disturbances shall in each case have subsided. This is the reason why interpreters and indeed Ricardo himself spoke of his law—and of his reasoning in general—as ‘abstract’ and as envisaging fundamental or long-run tendencies only. He did not use the Marshallian term Long-Run Normal but he had got the idea. The second thing to be observed is that our theorem would be true (for perfect equilibrium in perfect competition) if labor—and labor of one kind and quality—were the only requisite of production. In fact, it would then follow as a special case from the more general marginal-utility theory of a later time. 11 The third thing, then, to be observed about Ricardo’s labor-quantity law is the manner in which he tried to overcome the difficulties that stand in the way of generalizing a result 11 In order to show this, it is sufficient to refer to a theorem that is rationally deduced within the marginal utility theory, though it has been frequently implied in ‘classical’ pieces of reasoning, especially where the ‘classics’ made use of the ‘law’ of the uniform rate of profit. This theorem reads that, in equilibrium, all factors will be allocated to all their possible uses in such a way that the last increments of each factor employed in all these uses produce increments of products that are of equal value. If the products are beavers and deer, and if labor is all that is needed in order to History of economic analysis 564 kill them, beavers killed per hour of hunting must be worth as much as deer killed per hour of hunting and beavers will hence exchange for deer in inverse proportion to the time it normally takes to kill them. But this is the Ricardian theorem, which, by the same token, cannot be true if there are also other scarce factors. that holds—though he never proved it—in a special case. The rest of his first chapter ( 2–7) is devoted to an attempt to show that his labor-quantity law of equilibrium values, though not generally true, yet constitutes an acceptable approximation throughout the range of perfect competition. But this chapter does not deal with the fundamental difficulty that arises from the existence of scarce natural factors: their elimination from the problem is left for the second chapter. Following suit, we also neglect them for the moment. Ricardo saw, of course—what Marx was to elaborate—that the labor whose quantity is to ‘govern’ or ‘regulate’ values must be of the quality a laborer normally does in any given time and place, not more or less efficient than that, and that it must be applied according to the prevailing standards of technological rationality: to use Marx’s term, it must be socially necessary labor. The time employed in acquiring skills, including the labor of the teacher, must be counted in 12 and so must be ‘the labour also which is bestowed on the implements, tools, and buildings with which such labour [the directly applied labor, J.A.S.] is assisted’ (sec. 3). But what about natural skills or those elements in skills that are not themselves acquired by labor? Following the eighteenth-century tradition noticed above, Ricardo did not think much of their importance. For the rest, he relied, as had A.Smith, on the market mechanism to determine a scale for the evaluation of different (natural) qualities of labor by means of which an hour of superior labor may be expressed as a multiple of the normal labor hour: if a ‘working jeweller’ is being paid twice as much per hour as is a ‘common labourer,’ one hour worked by the former will simply be counted as two hours worked by the latter. Since such relations do not vary greatly from year to year, they have ‘little effect, for short periods, on the relative value of commodities.’ 13 This may or may not be so. But it should be noticed that this appeal to market values—that are evidently not determined by any quantity of labor—in the course of an argument that is to expound the labor-quantity law spells in strict logic the surrender of the latter, no matter whether this be acknowledged or not. But acknowledgment of failure of the labor-quantity principle came in Sections 4 and 5. There, Ricardo faced the facts that relative values of commodities are not ‘governed’ exclusively by the quantities of labor embodied in them but also by ‘the length of time which must elapse before’ they ‘can be brought to market’ For this is what his argument amounts to: unequal proportion between that part of capital which ‘is to support labour’ and that part which ‘is invested in tools, machinery and buildings,’ and unequal durability 12 This Ricardo did not say explicitly. But it is only fair to interpret him in this sense. 13 It is of some interest to note that Ricardo, while professedly arguing about long-run phenomena, displayed in this case no compunction at using a short-run argument—another instance of his extreme carelessness. General economics 565 of the latter or unequal rate of turnover of the former—which are the facts discussed— are relevant to the relative values of the products only because of that time element which they bring into the picture of the productive process. 14 They simply mean different periods of investment of the (possibly) equal quantities of labor embodied in the capital goods or (to put quite bluntly the common-sense business fact of which Ricardo was thinking) different amounts of carrying charges that, logically, are on a par with quantity of labor in influencing ‘natural,’ that is, equilibrium, values. So the murder is out. To be sure, Ricardo tried to minimize the damage to his fundamental construction by pointing out that quantity of labor still remains the most important determinant of relative value, which is why above we have described his theorem as an approximation. This seems to do more justice to his thought than does the interpretation that appeals to other historians: these, following a lead of Marshall’s, prefer to say that Ricardo had ‘really’ a cost theory of value. It is true that in effect Ricardo ended up by co-ordinating the element of accrued profits with the element of quantity of labor. It is also true that sometimes (see ch. 30, first sentence) he did make Cost of Production (evidently including the former element) the ‘ultimate regulator’ of values. But if this were all, his exposition would reduce simply to a roundabout way of stating a view that was current in his time: it would be difficult to see what it was he fought for with so much insistence and what the ensuing controversies were about. Only if we recognize that he believed, wrongly of course, that labor applied is something more fundamental or important than are accrued profits, shall we understand why he first introduced his theory of values under the assumption that capital structures were exactly similar in all industries. The comfort is of course quite illusory that he drew from the fact that then (if we accept his elimination of the influence of natural agents) the relations of quantities of labor applied would ‘regulate’ relative values. Logically it would be just as admissible to say that, with equal quantities of labor applied, it is capital structure or ‘time’ which regulates relative values. Therefore, he must have thought that the former proposition is true in some sense in which the latter is not. And our interpretation—the interpretation that is characterized by the word approximation— seems to me the most obvious one in the case of a writer who was quite free from either emotionalism or philosophical preconceptions. Another point must be mentioned, however. What Ricardo did in Chapter 1, Sections 4 and 5 was to recognize the fact that carrying charges do influence relative values. He also formulated some of the consequences of this fact. But he did so, as it were, with a shrug of his shoulders and did not make the slightest attempt to explain it unless we accept the phrase ‘just compensation for the time that the profits were withheld’ as a token of such an explanation. Here as elsewhere he was content to remain on the surface of things. But he did worry about what his admission would do to his pet proposition that ‘no alteration in the wages of labour could produce any alteration in the relative value of…commodities,’ which is, throughout the book, the practical spearhead of his theory of 14 Let us notice once more that this constitutes an important link between Ricardo and Böhm- Bawerk. History of economic analysis 566 value. On principle, it must also be given up, of course (see 5, last paragraph). But actually it is retained, again, as I like to say in order to be as fair as possible to him, as an approximate truth. The effect of the admission is confined to a particular theorem: if wages, say, rise, the relative prices of goods into the production of which ‘fixed capital’ or ‘fixed capital’ of high durability enter largely will fall, and the relative prices of goods ‘which are produced chiefly by labour with less fixed capital or with fixed capital of a less durable character than the medium in which price is estimated 15 will rise,’ a proposition that has been dubbed in our time, the Ricardo Effect—and a curiously devious way of admitting something of which one does not wish to admit the implications. It is not worth our while to stay to describe the way in which the inner-circle Ricardians, James Mill, De Quincey, and McCulloch, handled Ricardo’s theory of value and the spurious problems it created. 16 But it will be con-venient, before we take up the contributions of Ricardo’s opponents and then the half-way house position of J.S.Mill, to consider with the utmost brevity some essentials of the doctrine of Ricardo’s only great follower, Karl Marx. Marx’s theory of exchange value is also a labor-quantity theory, perhaps, if we neglect such stepping-stones between Ricardo and Marx as W.Thompson, the only quite thoroughgoing one ever written. At first we are, indeed, struck by the similarity of Marx’s argument to Ricardo’s. Marx asked himself what it is that makes commodities, so heterogeneous as to value-in-use, comparable at all, and emerges with the conclusion that is the fact that they are all the products of labor. Having established to his own satisfaction this highly debatable proposition—for the fact that all commodities have value-in-use is not only as true but more general—Marx proceeded to deal with the difficulties that beset this approach at the threshold almost exactly as Ricardo had dealt with them. He added precision and elaboration here and there—I have already mentioned the ‘socially necessary labor’—but failed, like Ricardo, to notice the danger that lurks behind the assumption that the market prices of labor of different nonacquired qualities may be used in order to reduce labor hours of superior quality to multiples of standard ones. I take this opportunity to mention a point of technique that Marx considered one of his most important contributions to economic theory: his distinction between labor, the quantity of which is measured in hours, and ‘labor power’ (Arbeitskraft), the value of which is given by the quantity of labor that enters into the goods the workman consumes (including the goods and services used up in bringing him up and training him) and in a sense ‘produces’ his labor power. These goods and their real value are, of course, essential elements also of Ricardo’s analysis. But he did not identify explicitly this real value with the real value of the commodity labor power. Senior, as we know, took a step toward doing so. Marx, however, not only completed this step but also, in his exploitation theory (see below, sec. 6b), put the labor-power concept to a use of which neither Ricardo nor Senior had thought or would have approved. 15 This is correct. Marx slipped when he replaced this by the average composition of total capital. 16 However, we may note in passing a device that McCulloch used to generalize Ricardo’s labor- quantity theorem. Recognizing that, looking at the matter from Ricardo’s standpoint, the main trouble was connected with the element of time, he simply took the view that the labor quantity General economics 567 embodied in durable capital goods goes on to do further labor during their lifetime. A harsh critic might call this a purely verbal expedient and an inept one at that. But it is also possible to see in it a particular way toward renunciation of the labor-quantity theory and toward recognition of a multiplicity of ‘factors’ or services of production, all of which help to create product value. If we look at his reasoning in this light, it amounts to generalizing the concept of labor. In itself But even non-Marxist historians should have realized—though, mostly, they have not—that there is a much more fundamental difference between the labor-quantity theory of Marx and the labor-quantity theory of Ricardo. Ricardo, the most unmetaphysical of theorists, introduced the labor-quantity theory of value simply as a hypothesis that was to explain the actual relative prices—or rather the actual long-run normals of relative prices—that we observe in real life. But for Marx, the most metaphysical of theorists, the labor-quantity theory was no mere hypothesis about relative prices. The quantity of labor embodied in products did not merely ‘regulate’ their value. It was (the ‘essence’ or ‘substance’ of) their value. They were congealed labor. Lest non-metaphysically disposed readers should refuse to be very much impressed by this does not do much for us. But it points in the direction of a more fruitful theory all the same. this, let me at once point out the practical difference that this made for our two authors’ analytic structures. When Ricardo recognized that the element of time—or of the carrying charges that accrue in the course of the productive process—entered into the determination of values or relative prices, this meant for him the necessity of admitting that his hypothesis was contradicted by the facts and that, in the manner described above, it had to be reduced to a mere approximation. But Marx had recognized from an early stage of his thought— certainly before he published the first volume of Das Kapital (1867) 17 —that exchange ratios do not, not even as a tendency, conform to Ricardo’s equilibrium theorem on values, which accordingly forms no part of Marx’s teaching. This, however, was no reason for him to modify his value theory: value was always, for every commodity as well as for output as a whole, identical with labor embodied, however relative prices might behave, and his problem was precisely to show how, in consequence of the mechanism of perfect competition, these absolute values without being altered came to be shifted about in such ways that in the end commodities, while still retaining their values, were not sold at relative prices proportional to these values. For Ricardo, deviations—other than temporary—of relative prices from his proportionality theorem spelled alterations of values; for Marx, such deviations did not alter values but only redistributed them as between the commodities. This is why we may say that Marx actually went through with the idea of an absolute value of things, 18 whereas Ricardo, 17 This fact is evident from the material published in the Theorien über den Mehrwert (1905–10) and hence was not evident at all before the publication of these volumes. In consequence, even the greatest of nineteenth-century Marx critics, Böhm-Bawerk, took the view that Marx expounded a labor-quantity theory in the first volume of Das Kapital which Marx’s later thought convinced him to be hopelessly at variance with facts, so that he was driven to shifting his ground in those writings that were, after his death, published by Engels (1894) as the third volume of Das Kapital—Marx’s aversion to continuing publication of his work was interpreted as a confession of failure. In other History of economic analysis 568 words, the value theory of the first volume was interpreted too much in Ricardo’s sense. This was an error, and one that spelled missing the essential point of Marx’s value theory. This is not to deny, of course, that some of the criticisms proffered retain validity in spite of this error. Nor do I wish to assert that Marx carried out the program appropriate to his point successfully. This is sufficiently obvious from our text, although it is impossible to clear up the matter fully within the space at our command. 18 He was the only author who ever did. although his argument implies this idea in spots, never made it the pivot of his analytical structure. Or, to put it differently: whereas for Ricardo relative prices and values were essentially the same thing and whereas hence the economic calculus in terms of values was the same thing as the calculus in terms of relative prices, values and prices were not the same thing for Marx, so that he created for himself an additional problem that apparently does not exist for Ricardo, namely, the problem of the relation between the two calculi or the problem of Wertrechnung und Preisrechnung. 19 Some of the implications and applications of this value theory will be discussed later on. But three points about it should be made before we temporarily leave the subject. First, for us it is nothing but a construction devised for purposes of analysis and to be judged in the light of considerations of analytic usefulness and convenience. For orthodox Marxists it may indeed be sacred truth in some extra-empirical realm of Platonic ideas, where the ‘essences’ of things are exhibited. And it may have been something of the kind for Marx himself. Actually, however, there is nothing mystic or metaphysical about the Marxist theory of value. Its central concept in particular, absolute value, has nothing to do with the meanings we attach to this word in some parts of philosophy. It is nothing but Ricardo’s real value fully worked out and fully made use of. Second, if readers have followed the argument, they will realize that the objections that may be made against Ricardo’s use of the concept of real value do not apply to Marx’s theory. Even if we do not admit that labor embodied is the ‘cause’ of exchange value in the ordinary sense, there is no logical rule to prevent us from defining labor embodied as exchange value, though this gives another and perhaps misleading sense to the latter term. For, on principle, we may call things what we please. 20 Third, whereas Ricardo simply recognized the actual existence of carrying charges and then stopped, Marx made at least an attempt, successful or not, to absorb them into his schema. For him, the carrying charges were also part of the labor embodied in total output. Ricardo had to add them to labor cost and ought to have explained them. For Marx there was no problem of explaining why these elements of product value exist. His only problem was to explain how they come to be clipped off from a total of value that exists independently of them. At this we must let the matter rest for the moment. At a later stage of our argument we 19 See on this Ladislaus von Bortkiewicz (1868–1931), ‘Wertrechnung und Preisrechnung im Marxschen System,’ three articles in the Archiv für Sozialwissenschaft(1906 and 1907) and the same author’s ‘Zur Berichtigung der grundlegenden theoretischen Konstruktion von Marx im dritten Band des “Kapital,”’ Jahrbücher für Nationalökonomie und Statistik (1907). 20 But Marx would no doubt have avoided much confusion and futile controversy had he named his absolute-value concept differently. The word ‘value’ was not at all well chosen to express its actual analytic significance. But much agitatorial glamour would have been lost by choosing a different one. Also, he may have wished to join up with Ricardo’s real value, which was not less misleading. General economics 569 shall see that it was after all the same difficulty, namely, the influence of time, which, owing to their different approaches, presented itself to them in the guise of different problems. To use a Marshallian phrase, for Ricardo time was the great disturber of his analytic pattern. But it was also, though less overtly, the great disturber of Marx’s. (b) The Opponents of the Labor-Quantity Theory of Value. Remember: the Ricardians were always in the minority, even in England, and it is only Ricardo’s personal force which, as we look back, creates the impression that his teaching—his coinage of Smithian metal—dominated the thought of the time and that the other economists were just opponents of what was then called the New School— opponents, too, that were not quite up to the latter’s doctrines. The opposite is nearer the truth, in the matter of value as in others, although our impression is no doubt reinforced by the fact that nearly all those opponents, whatever else they may have been, were with but few exceptions inferior to Ricardo as controversialists. Discussion of the problems of value on non-Ricardian lines that carried over from the eighteenth century collided with the Ricardian forces and blazed into controversy about 1820, the year in which Malthus’ Principles appeared. The active phase of this controversy lasted little more than ten years and notwithstanding the testimony of a very few stalwart defenders—McCulloch and Marx stood arm in arm at this point—and of some historians ended in the defeat of Ricardianism. It had its full complement of mutual misunderstanding and of logical errors, but on the whole it moved on a creditable level. The peak performance was Bailey’s 21 (see above, ch. 4, sec. 3c), the influence of whose criticism was much greater than appears on the surface. He showed up the weaknesses of Ricardo’s analytic structure forcefully, in particular, the futility of Ricardo’s method of eliminating natural agents from the value problem, the arbitrariness involved in calling quantity of labor ‘the sole determining principle of value,’ the defects of the concept of real value and of the Ricardian theory of profit, and so on. The discourteous reply from some Ricardian in the Westminster Review (1826) was pathetically inadequate, and though but few contemporaries did justice to him, it became clear in time that he had in fact turned the tide and dealt a fatal blow. Considerations of space forbid the describing of the controversy in detail. 22 Instead, we shall confine ourselves to what was, for the time being, the main point at issue between Say, Malthus, and Ricardo, noticing other names and aspects only so far as this purpose requires. 23 21 Samuel Bailey, A Critical Dissertation on the Nature, Measures, and Causes of Value; Chiefly in Reference to the Writings of Mr. Ricardo and His Followers (1825). 22 Three other main contributions, however, should not go unmentioned: first, a tract entitled Observations on Certain Verbal Disputes on Political Economy… (1821), which displays sound sense for the spurious or fictitious nature of part of the issues involved; second, another tract, entitled An Essay on Political Economy… (1822), notable for an early recognition of the logical weakness of any explanation of value by cost and of the fact that cost affects value only through affecting supply; and a third, already mentioned, C.F.Cotterill’s Examination of the Doctrines of Value (1831), forgotten because of his defense of (most of) Bailey’s tenets. [The two anonymous tracts are mentioned by Seligman, Essays in Economics, pp. 81–2.] History of economic analysis 570 To get at that main point, we must first recall that the pioneer performances of the marginal utility theory, which the period produced, failed indeed to exert any perceptible influence, but that many writers perceived that utility was more than a mere condition of exchange value, in the sense in which Ricardo meant this phrase, and that in fact it was the ‘source’ or ‘cause’ of exchange value. Only they were no more able to do anything with this idea than were the Ricardians, who, precisely for this reason, refused to accept it. And so this approach came to nothing. J.B.Say, for instance, following the French tradition (Condillac, in particular), made exchange value dependent upon utility but, failing (like Condillac) to add scarcity, stumbled over the fact, so often explained before him, that such ‘useful’ things as air or water normally have no exchange value at all. He said that actually they do have value; only this value is so great, infinite in fact, that nobody could pay for them and hence they go for nothing. 24 It is true that he did not stop at this ineptitude. He did rise to the imperfect (yet so significant) statement that price is the measure of the value of things and value the measure of their utility, a statement that heralds Walras’: les valeurs d’échange sont proportionnelles aux raretés [marginal utility, J.A.S.]. Mostly, however, he merely used a rather primitive supply-and-demand analysis. The same applies to Hermann (see above, ch. 4, sec 5). As in France, perhaps in part under French influence, a utility-theory tradition had developed in Germany. But it was equally inoperative: it stopped at recognitions of the utility element that are difficult to distinguish from the Ricardian way of assigning to utility the role of a condition of value. Hermann went further than others but he also confined himself substantially to working with supply and demand. Some English economists, such as Craig 25 and Senior, did better. As regards the latter, there is truth in the common view, shared by Walras, that credits him with the notion of marginal utility. But I can only repeat: he did not go through with it and, after a glimpse, it practically vanished behind mere supply and demand. Lord Lauderdale and, more elaborately, Malthus went straight to the supply- and-demand apparatus and concentrated entirely upon it. Thus, for Ricardo, the main point at issue was from the first labor quantity versus supply and demand. The utility theory of value, which he had glanced at and rejected (as the ‘source’ or ‘cause’ of exchange value), was not really in the picture, though he criticized it in his chapter on ‘Value and Riches.’ The cost theory of value was not wholly an enemy. For he looked upon his own theory as a reformulation of it and frequently himself invoked cost in terms of labor and capital. The true enemy was the supply-and- demand theory, which ‘has become almost an axiom in political economy, and has been the source of much error’ (ch. 30, third paragraph). The reader should observe 23 Say and Malthus were, so far as value is concerned, as much on one side as they were on opposite sides in the question of saving and general gluts. But there was no perfect agreement between the two in the first case, and no perfect agreement between Say and Ricardo in the second. Ricardo’s position in the controversy on value may be pretty completely understood from his Letters to Thomas Robert Malthus, 1810–1823 (ed. J.Bonar, 1887), his Notes on Malthus’ ‘Principles’ and chs. 20 and 30 of his own Principles. 24 Condillac, on the contrary, states that such things have a price—which consists in the effort to appropriate them, for instance, by breathing, drinking, and so on. 25 John Craig’s Remarks on Some Fundamental Doctrines in Political Economy… (1821) is a performance of free or absorbing money income, influence other prices. He also understood, like Say, that (marginal) value in use must be ‘accurately measured’ by (must, in equilibrium, be proportional to) value in exchange. If we may indulge in what is no doubt a mistake and read into his statement all that our bracketed reformulations convey, we find a whole Marshall in nuce. General economics 571 . volume of Das Kapital—Marx’s aversion to continuing publication of his work was interpreted as a confession of failure. In other History of economic analysis 568 words, the value theory of the. arbitrariness involved in calling quantity of labor ‘the sole determining principle of value,’ the defects of the concept of real value and of the Ricardian theory of profit, and so on. The discourteous. forgotten because of his defense of (most of) Bailey’s tenets. [The two anonymous tracts are mentioned by Seligman, Essays in Economics, pp. 81–2.] History of economic analysis 570 To get at that