NOTES
[1] Whitehead, Science and the Modern World. Eddington comes near to the same idea in an essay in Science, Religion, and Reality, 1925. See also Weyl, Was ist Materie? 1924, p. 84. It has also been expressed by others quite independently, though I do not know of other published references.
[2] E.g. the irreversible motion of an electron in the field of a bar magnet is rendered formally reversible by the assumption that the magnetic field is due to moving electrons. Yet this assumption is highly artificial since it postulates electronic movements that have never been observed. In other cases irreversibility is eliminated by the choice of special co-ordinate systems. Some physicists now hold the view that irreversibility may be inherent in atomic as it is in organic processes.
[3] Internal Constitution of the Stars, 1926, p. 56. Compare note on p. 44.
It may be convenient here to summarize the processes that give at any rate superficial evidence of their irreversibility: processes involving heat changes, or the radiation of light, or mass; the production of energy in a star, the motions of electrons in magnetic fields, certain types of atom-ion collision in mixed gases, processes dependent on retarded potentials, radioactivity, organic growth and evolution, and consciousness itself. Eddington deals only with the case of the emission and absorption of light, but suggests that the direction of time can only be deduced from statistical processes. This is the orthodox view, though it is very doubtful if it is valid now that the quantum processes are receiving formulation. In this connection, see note 4.
[4] Einstein. Berlin Akad., Sitzungsberichte, 1925, p. 418. But Einstein’s view must be revised in view of recent experimental results (e.g. Harnwell, Phys. Rev., vol. 29, 1927, pp. 683 and 831), if these have been correctly interpreted. See Born, Zeitschr für Physik, vol. 40, pp. 177-8; and Jordan, Naturw. 1927, p. 792.
[5] The idea that time may be an active factor in causation has the mathematical significance that ‘t’ (for the system in question) must appear explicitly in the formulation of the law, and not merely as the square of a time-differential found convenient for the correlation of a standard clock with a reversible process which is being observed. A law whose mathematical formulation involves ‘t’ measured from some moment in the history of the system, gives an entirely new meaning to ‘t’, though one consistent with the properties of the reversible Newtonian differential ‘dt’. Such a law may claim to express the fact of historic, irreversible, duration, a feature in nature which is neglected by laws involving only ‘dt’ squared.
[6] Einstein, Annalen der Physik, vol. 49, pp. 776-7, 1916.