Such a picture, I submit, avoids the circle which is presented in all other formulations of the Austrian theory of value. I wish, however, to indicate its limitations as a theory of value, and the impossibility of any application of it to the problem of the value of money. (1) Its data are inaccessible: nobody could possibly know all the utility-curves and all the time-preference curves (and disutility of labor-curves, etc.) of all goods to all individuals in, say, the United States. To explain market ratios by utility-curves is a case of _ignotum per ignotius_, so far as practical application is concerned. Moreover, the scheme is so difficult to visualize that it is useless as a tool of thought--as one will find who tries to think it through, without the aid of higher mathematics, for ten goods, and ten persons, with unequal distribution of wealth, and different utility curves, time-preference curves, and disutility-curves for each kind of good to each individual. (2) The scheme must assume smooth curves and infinitesimal increments in consumption, which is a fiction so far as the individual psychology is concerned. Without this assumption, the point-for-point correspondence between individual and market ratios does not exist. It is only in social-value curves, or in demand-curves in the big market (which are social-value curves, expressed in money),[79] that you have, as a matter of fact, the right to smooth out your curves. (3) The theory must assume the frictionless static state, in which marginal adjustments are perfectly accomplished, and equilibrium really reached. Without this assumption, again the point-for-point inverse correspondence of market ratios and individual ratios fails. But this makes it quite impossible to apply the doctrine to any functional theory of the value of money, or to bring money in any realistic way into the scheme. As will be shown more fully in later chapters, money functions in bringing about just the absence of friction which static theory assumes. That is what money is _for_. The functional theory of money, therefore, cannot abstract from friction and dynamic change.[80] It is, of course, possible, on this scheme to pick out any one of the goods in the system, say the 1-1000th part of a horse, call it the "money-unit," and determine a set of money-prices. These "money-prices" are already given in the scheme in the ratios between the abstract numbers of this unit and the abstract numbers of the units of all other goods. But this is meaningless, so far as a theory of money is concerned. It abstracts entirely from the _differences_ in _salability_[81] of goods, on which the theory of money must rest. It gives us no clue to that part of the value of the money-article which comes from its money-functions.