Let us return to our imaginary world and ask ourselves if its inhabitants could not, without renewing the history of the sleepers of Ephesus, perceive this evolution. No doubt, no matter how perfect the heat conductivity on their planet, it would not be absolute, so that extremely slight differences in temperature would still be possible. They would long escape observation, but perhaps there would come a day when one would imagine more sensitive measuring devices and where a physicist of genius would bring out these almost imperceptible differences. A theory would be constructed, one would see that these differences of temperature have an influence on all the physical phenomena, and finally some philosopher, whose views would seem hazardous and reckless to most of his contemporaries, would affirm that the average temperature of the universe could have varied in the past and with it all known laws. Could not we do something similar? For example, the fundamental laws of Mechanics have long been considered absolute. Today some physicists say that they must be modified, or rather enlarged; that they are only true for the speeds to which we are accustomed; that they would cease to be so for speeds comparable to that of light; and they support their view of certain experiments made with radium. The old laws of Dynamics are none the less virtually true to the world around us. But could it not be said with some semblance of reason that, owing to the constant dissipation of energy, the velocities of bodies must have tended to diminish, since their living force tended to become heat; that going back far enough in the past, one would find a time when speeds comparable to that of light were not exceptional, where consequently the classical laws of Dynamics were not yet true?
Suppose, on the other hand, that the observable laws are only results, depending at the same time on the molecular laws and the arrangement of the molecules; When the progress of science has made us acquainted with this dependence, we will no doubt be able to conclude that, by virtue of the molecular laws, the arrangement of molecules must have been formerly different from what it is today, and by consequently, observable laws have not always been the same. We would therefore conclude with the variability of the laws, but, as we remark it, it would be by virtue of the very principle of their immutability. We would argue that the apparent laws have changed, but that would be because the molecular laws, which we would now see as the true laws, would be proclaimed immutable.