Now in the twentieth century the inverse of what happened in the eighteenth was taken. It became known - and that is the theory of chaos which I will come back to later on for some details - that under certain conditions systems which obey totally well-described, well-understood laws, in which one can predict with certainty the progress of a trajectory once one knows initial conditions, that those systems were so sensitive to initial conditions that in practice they could predict nothing. That is, if you know fifteen decimals of initial conditions you could go on for a certain time and then essentially have no prediction whatsoever. If you knew more decimals for every initial condition you could go on longer but eventually, again, predictability is lost. Therefore, mechanics/dynamics - the same thing - became split into two parts: the nice, well-behaved dynamics which, if you know a few decimals you know pretty well how things behave; and the chaotic dynamics in which you must know an infinite number of decimals to have exact predictions. In theory you could predict everything, but in practice you can predict very, very little.