Be it noted further, however, that even if we grant a fixed ratio, on the basis of Fisher's argument, between M and M', Fisher has offered no jot of proof that the causation runs from M to M'. He simply assumes that point outright. "Any change in M, the quantity of money in circulation, _requiring as it normally does a proportional change in M'_, the volume of deposits subject to check." (_Ibid._, p. 52, Italics mine.) For this, no argument at all is offered. A fixed ratio, so far as causation is concerned, might mean any one of three things: (a) that M controls M'; (b) that M' controls M; (c) that a common cause controls both. Fisher does not at all consider these alternative possibilities. I shall myself avoid a sweeping statement as to the causal relations among the factors in the equation, because I do not think that any of the factors is homogenous enough, as an aggregate, to be either cause or effect of anything. But if a generalization concerning these magnitudes were required, I should be disposed to assert that the third alternative is the most defensible, and that to the extent that M and M' vary together it is under the influence of a common cause, namely, PT! That is to say, that the volume of bank-deposits and the volume of money tend to increase or decrease in a given market--and Fisher's theory is a theory of the market even of a single city[165]--_because of_ increases or decreases in PT (considered as a unitary cause rather than as two separate factors) in that market. But I shall not put my proposition in quite that form, as I find the factors in the equation of exchange too indefinite for satisfactory causal theory.