Robinson (1980) is one of her last economic papers and an interesting read.
The three types of time that we have in reality are past, present and future.
Robinson notes that, while many events that have already happened have antecedent causes in the past that at least in principle can be understood, we cannot necessarily predict events of the future arising from present actions: the future is unknown or at least not known with certainty (Robinson 1980: 219).
But there is another type of time: the hypothetical logical time of economic theory (Robinson 1980: 220). For example, the “stationary state” of equilibrium theory, though it is supposed to persist between two points of time, actually has no distinction of real time, only logical time. A movement from one static equilibrium position (supposedly at a given point in time) to another does not happen in real time either.
So many economic models reduce to logical time, instead of real time, and therefore rob the model of the ability to deal with human decision making under uncertainty through real time. Curiously, Robinson charges that even Sraffa’s fundamental model is subject to this problem (Robinson 1980: 223).
But, most importantly, Robinson discusses the “long run” or “long period” concept in neoclassical economic theory. At first, it seems that the “long run” might be in real time, rather than logical time, but on closer inspection this idea evaporates.
The problem goes back to at least as far as Marshall. The Marshallian “ceteris paribus” clause used in long run models and the very assumption that some economic phenomenon happens “in the long run” render such models dubious as a description of reality (Robinson 1980: 226).
Robinson, Joan. 1980. “Time in Economic Theory,” Kyklos 33.2: 219–229.