[29] The current marginal-utility diagrams are not of much use in this connection, because the angle of the tangent with the axis of ordinates, at any point, is largely a matter of the draftsman's taste. The abscissa and the ordinate do not measure commensurable units. The units on the abscissa are units of frequency, while those on the ordinate are units of amplitude; and the greater or less segment of line allowed per unit on either axis is a matter of independently arbitrary choice. Yet the proposition in the text remains true,--as true as hedonistic propositions commonly are. The magnitude of the angle of the tangent with the axis of ordinates decides whether the total (hedonistic) productivity at a given point in the curve increases or decreases with a (mechanical) increase of the productive agent,--no student at all familiar with marginal-utility arguments will question that patent fact. But the angle of the tangent depends on the fancy of the draftsman,--no one possessed of the elemental mathematical notions will question that equally patent fact. [30] A similar line of argument has been followed up by Mr. Clark for capital and interest, in a different connection. See _Essentials_, pp. 340-345, 356. [31] _Cf. Essentials_, pp. 83-90, 118-120. [32] _Cf._ chap. xxii, especially pp. 378-392. [33] _Cf._ p. 391. [34] _Cf. Essentials_, p. 39. [35] _Essentials_, p. 39. [36] _Cf. Essentials_, chap. iii, especially pp. 51-56. [37] This difficulty is recognized by the current marginal-utility arguments, and an allowance for intensity is made or presumed. But the allowance admitted is invariably insufficient. It might be said to be insufficient by hypothesis, since it is by hypothesis too small to offset the factor which it is admitted to modify. [38] The limit to which the intensity rises is a margin of the same kind as that which limits the duration. This supposition, that the intensity of application necessarily rises to such a pitch that its disutility overtakes and offsets the utility of the product, may be objected to as a bit of puerile absurdity; but it is a long time since puerility or absurdity has been a bar to any supposition in arguments on marginal utility. THE LIMITATIONS OF MARGINAL UTILITY[1]