You no longer say, indeed, as in your previous essay: 'L'économie pure n'est pas seulement semblable à la méchanique: c'est, à proprement parler, un genre de méchanique.' But you still say that 'Pure economics employs the same methods as rational mechanics, and has many points of contact with this science.' Although you do not pause over the mechanical considerations, it is not from a clear conviction that a datum in economics, as such, is quite different from a datum in mechanics; but merely because it seems to you _convenient_ to omit such considerations, of which you do not deny, but rather admit, the possibility. Now I on the contrary, say decisively that the data of economics is not that of mechanics, or that there is no transition from the mechanical aspect of a fact to the economic aspect; and that the very possibility of the mechanical point of view is excluded, not as a thing which may or may not be abstracted from, but as a contradiction in terms, which it is needful to shun. Do you wish for the simplest and clearest proof of the non-mechanical nature of the economic principle? Note, then, that in the data of economics a quality appears which is on the contrary repugnant to that of mechanics. _To an economic fact words can be applied which express approval or disapproval._ Man behaves economically _well_ or _ill_, with _gain_ or _loss_, _suitably_ or _unsuitably_: he behaves, in short, _economically_ or _uneconomically_. A fact in economics is, therefore, capable of _appraisement_ (positive or negative); whilst a fact in mechanics is a mere fact, to which praise or blame can only be attached metaphorically. It seems to me that on this point we ought easily to be agreed. To ascertain it, it is sufficient to appeal to internal observation. This shows us the fundamental distinction between the mechanical and the teleological, between mere fact and value. If I am not mistaken, you assign to metaphysics the problem of reducing the teleological to the mechanical, value to mere fact. But observe that metaphysics cannot get rid of the distinction; and will only labour, with greater or less good luck, at its old business of _reconciling_ opposites, or of _deriving two contraries from one unity_.