550 J. M. EIGG : observed uniformities, the impotence of the experimental methods to establish laws of nature becomes patent. Obser- vation, eked out by memory and experiment, may establish an uniformity, but such uniformity only furnishes science with a datum on which to found an hypothesis. The expe- rimental methods so elaborately analysed by Mill and Lotze are essentially methods of elimination. The same pheno- menon is observed to be sequent, now upon one, now upon another, combination of antecedent phenomena. If then the several sets of antecedent phenomena have a common and only one common element, it is presumable that in the absence of the common element the sequent phenomenon would not occur ; and if by observation or experiment we obtain a case in which, no new element being present in the antecedent, the common element is wanting and the sequent phenomenon no longer occurs, we infer what? Simply that as often as the case is repeated without change in the condi- tions, the result will be the same. Into the question of the warrant for this inference I do not here enter. The reader is entitled to assume that the warrant is the uniformity of nature, or the principle of identity, or any other principle that he may prefer. All that I am concerned to maintain is that the relation thus established does not amount to a law of causation, even if causation be no more than uncondi- tional invariability of antecedence and sequence 1 between phenomena. The experimental methods cannot establish the existence of any unconditional relations between pheno- mena. All that they can establish is the existence of rela- tions, the invariability of which is contingent upon the conditions remaining constant. They establish, in fact, conditional invariabilities of relation. Nay more : science distinctly negatives the existence of unconditionally invariable relations between phenomena. Geology has unrolled the records of past ages during which the forces of nature, though identical in kind with those now operating on the earth, were nevertheless so differently com- pounded as to produce widely different effects. Most of the sequences of events which were then observable on the planet must have been in striking contrast to any that are now observable ; and whether we accept the nebular hypo- thesis or no, it is clear that the sequences of events now observable in the solar and sidereal systems are far from similar to those which might once have been observed, and
I say sequence because I hold that consequence is more than the
empirical theory of causation is entitled to.
