If the exchange ratio of cloth to linen alters, the doctrine is that the cause must be that some change has taken place in the quantity of labor required to produce cloth or linen, and _not_ that the “value” of a linen-maker’s day of labor has changed in ratio to the “value” of a cloth-maker’s day. In other words, if ten hours of a cloth-maker’s labor have earned the same wages (and thus occasioned the same cost to entrepreneurs) as twelve hours of a linen-maker’s labor, “we may safely conclude” that it is not alteration of this ratio that causes alteration of the exchange ratio of cloth to linen. It is this ratio between the wage-earning capacity of one kind of labor and another kind that “has long ago been adjusted.” To make this point perfectly clear, let us quote again: “The comparative degree of estimation [an equivocal expression which means comparative wage-earning power] in which the different kinds of human labour are held ... continues nearly the same from one generation to another, or at least ... the variation is very inconsiderable from year to year, and therefore can have little effect, for short periods, on the relative value of commodities.”[57] The question is this: In the labor-cost theory of value, does skilled labor count as _more_ labor per day than unskilled, and if so upon what principle? Ricardo’s argument, as just traced, avoids the question, and is faulty in two essential points. In the first place, it is not true, and was not true in Ricardo’s time, that the comparative skilfulness of labor employed in producing different commodities remains unchanged. Machine invention alone produces veritable revolutions in this field. But in the second place (a more important point as a matter of theory), in this argument Ricardo has shifted his ground with respect to the meaning of his labor-cost law. This principle is stated in italics at the head of his chapter: “The value of a commodity, or the quantity of any other commodity for which it will exchange, depends on the relative quantity of labour which is necessary for its production.” This means precisely that, if one A exchange for two B, it is because, _at this time, without reference to changes in time_, it costs twice as much labor to produce an A as to produce a B. But now Ricardo has virtually changed the principle to mean that _alterations_ in the exchange ratios of commodities will be due to _alterations_ in the comparative amounts of labor required to produce them. This is a different principle, and indeed one no stronger than the other. We are forced to the conclusion that Ricardo unconsciously avoided the real question in the case, and failed to explain away the difficulty of skilled labor in the labor theory.