“it is always the amount of capital, or _quantity of accumulated labour_, and not the sum of accumulated and immediate labour expended on production, which determines the exchangeable value of commodities.”[106] No definition is given of “accumulated labor” other than that implied in the sentence above: “the amount of capital, _or_ accumulated labour.” Torrens defines capital to be the raw-material, machinery, and _subsistence of labor_ necessary to production. The “accumulated labor” which a product costs must be, in his view, the labor[107] which its raw-material costs, plus that which the machinery used up for it costs, plus—_not_ the labor actually employed on it in connection with this machinery, but plus the labor which the _subsistence_ of this labor has _cost_![108] Since Torrens himself offers no explanation why this newly-defined quantity of labor cost should regulate value, we are not in duty bound to go very far into his fantastic conception. Just as labor theorists generally proved that labor cost is the regulator of value by the simple process of showing that utility is not, so, I suppose, Torrens shows that labor cost, as he defines it, regulates value because as defined by Ricardo it does not. An astonishing thing about his conception is that the labor _directly applied_ to a thing is not a part of its labor cost! To the money outlay of the capitalist-entrepreneur for machinery and for raw-material corresponds the labor cost of these goods. To the money outlay in wages must correspond either the labor cost directly remunerated by these wages, _or else_ the labor cost of the things the laborers buy with their money wages (subsistence). Both of these cannot be represented by these wages, otherwise some labor cost would be counted twice over. For instance, the labor employed directly on a pair of shoes would be part of the labor cost of production of the shoes; but somewhere else it would be counting as the cost of production of hats, or what not, according to the employment of the laborer of whose subsistence these shoes became a part. Having made the absurd choice to count this labor (the direct labor cost of the shoes) as part of the cost of hats, upon the general principle thus adopted, Torrens cannot count labor, directly applied, as part of the cost of the thing to which it is applied. From this it follows that the direct labor cost of a machine or a piece of raw-material cannot be counted as a part of the real labor cost which determines its value. As a consequence, it is interesting to note that none of the labor directly applied to a commodity, or directly applied to any production goods used up in it, is a part of its labor cost. If Mr. Torrens had been pressed by a critic in his day, he could well have defied any man to locate the Torrens labor cost of an article to show that it _does not_ correspond to the value of the article. CHAPTER VII THE EMPIRICAL THEORY AS DEVELOPED BY MALTHUS.