Differing Criteria for Temporal Symmetry
Temporal symmetry is looked at as time reversibility in dynamical processes. Classical mechanics is, thereby, generally postulated to show these symmetry aspects. The basic question is thereby: is it possible, for a mechanical system which has evolved from an inital state A to a final state B, that an evolution backwards from B to A occurs? Or, shortly, is the backward evolution possible? As the author points out, the crucial background lies in the word ``possible: possible in the sense of being allowed by the laws of classical mechanics, or additionally insisting that the reverse motion also conforms to the particular interactions that compose the dynamical system under consideration. Hutchison's view is the latter, and he sketches two examples for explanation: the ratchet and a steel ball dropped into a jar of honey, both motions ``(for obvious reasons) irreversible. In order to defend his viewpoint against his colleague Savitt, he cites the latter to say that it is not the laws of classical mechanics which prevents the ratchet from going backwards but the pawl. From this the author concludes that a) no one disputes that if the pawl is completely removed, a ratchet will readily turn backwards, and consequently b) Savitt does obviously not take the interactions between the material components into account. The main difference between these viewpoints is eventually characterized by a) Savitt's opinion: take the exisisting forces away, and look for a collection of forces that can be applied to the material components of the system that will allow a backward transformation from B to A, and b) Hutchison: do the interactions that already act within the system allow this transformation? In what follows, several arguments are presented like irreversibility in heat conduction and citations of some kinds of ``proof in the literature which presume that the reverse transformation is to be governed by the same interactions as the forward one and can thus motivate Savitt's opinion. However, these ``proofs already show a lack in the whole discussion. The one relies on the Lagrangian (hence, obviously, focusing on conservative systems), another one argues that \(F=m \cdot d^2x/dt^2\) yields the same for \(t=-t\) (hence neglecting the state dependence of \(F=F(x, dx/dt)?)\). To characterize the basic question, a comparison with control theory might be allowed, i.e. controllability: is there any possibility to bring back a ``disturbed state to its desired reference by means of a control vector -- no matter how? This would characterize Savitt's view (in Hutchison's interpretation, the original is not known to the reviewer). If, on the other hand (and which meets much more the sense of time symmetry), the question is: given a known state at a certain time -- how did the mechanical system get there? In this sense, Hutchison is totally correct, although his arguments of (real) physical irreversibility seem to be out of place. The ratchet cannot turn backwards, of course, but the problem under consideration is to find out (``backwards) the initial conditions leading to the actual state of forward turn, or, if need be, of backward turn with stop but looking at local deformations etc. as a consequence of (the unknown) initial conditions. In classical mechanics (which should be defined anywhere in order to make clear what one is speaking about) the ``dynamical laws are unambiguous and the phase curves of a mechanical system can never intersect, leading to time symmetry in any case. Nobody would doubt that the actual (but non-zero) state of a damped oscillator is caused by one unique initial state which can be calculated by looking backwards in time, although real motion is irreversible. Classical mechanics in this sense does not include loosing or missing informations during evaluation, whatever they may come from. However, if the latter occurs, one is not totally helpless in the forward evaluation using statistical methods and probabilistic assumptions (heat conduction, chaotic motions if need be), which obviously prevents from ``exact backward evolution. Here, Savitt's definition in the above sense would still apply (``bring the system back to its initial state anyhow). But what is the benefit? The initial state itself is unknown. One could think of making a certain amount of measurements (in forward time direction), and from these to derive the corresponding ``force laws. Although in practice in physics this is totally unsatisfactory, leading to vague and ambiguous results, and leading away from the basic question about the temporal symmetry in classical mechanics which is then obviously left.
This page was built for publication: Differing Criteria for Temporal Symmetry
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4940878)
