Degrees are no longer measured in terms of increasing perfection or a differentiation and increase in the complexity of the parts, but in terms of differential relations and coeffecients such as selective pressure, catalytic action, speed of propagation, rate of growth, evolution, mutation, etc. [TP: 48]
[T]he degrees are not degrees of preexistent development or perfection but are instead global and relative equilibriums: they enter into play as a function of the advantage they give particular elements, then a particular multiplicity in the milieu, and as a function of a particular variation in the milieu. [TP: 48]
Relative progress, then, can occour by formal and quantitative simplification rather than by complication, by a loss of components and syntheses rather than by acquisition (it is a question of speed, and speed is a differential). [TP: 48]
A quotient of differentials is indeed calculable if it is a matter of the limit of variation of the production flows from the viewpoint of a full output, but it is not calculable if it is a matter of the production flow and the labour flow on which surplus value depends. Thus the difference is not canceled in the relationship that constitutes it as a difference in nature; the "tendency" has no end, it has no exterior limit that it could reach or even approximate. [AO: 230]