Ricardo elaborated the labor theory of value, and tried to think it through. He was too keen a logician to shift view-points with Smith's facility, and he tried to make a completed system.[55] There is some shifting from the theory of labor as a cause of value to labor as a measure of value, as in the following passage: "If the state charges a seigniorage for coinage, the coined piece of money will generally exceed the value of the uncoined piece of metal by the whole seigniorage charged, because it will require a greater quantity of labour, or, which is the same thing, the value of the produce of a greater quantity of labour, to procure it." (_Works_, McCulloch ed., 213.) In general, however, Ricardo developed a causal theory of value, quantity of labor being the basis of the absolute values of goods, their _relative_ values depending on the relative amounts of labor involved in the production of each. I shall not go into the matter fully, but shall call attention to the rock on which the system split, as Ricardo himself admits. A greater or less proportion of capital works with labor in producing different things, and the value of product, in that case, varies not merely with the labor, but also with the amount of capital, and the length of time the capital is employed. How say, then, that labor alone governs value? How reduce labor-cost and capital-cost to homogeneous terms? James Mill tried to do it for him by making capital merely stored up or petrified labor, which gives up its value again in production. But this doesn't meet the difficulty, because there is a _surplus_ value, over and above that explained by all the labor, including the labor which produced the machine, and the labor which produced the raw materials which entered into the machine, etc. The case of wine is a particularly obstinate case. Wine increases in value merely with the passage of time, at a rate which corresponds to the profit on capital. Ricardo finally, in correspondence with McCulloch, definitely abandons the case, stating that there are many exceptions to the proportionality between exchange value and labor-cost. "I sometimes think that if I were to write the chapter on value again which is in my book, I should acknowledge that the relative value of commodities was regulated by two causes instead of one, namely, by the relative quantity of labor necessary to produce the commodities in question, and by the rate of profit for the time that the capital remained dormant." (Davenport, _Value and Distribution_, p. 41.) But this is a "dualistic" rather than a "monistic" explanation--one element is a money-expense, or at all events a pecuniary item, while the other is a "real cost" item. The two are incommensurate and incommensurable.