136 Psychology and Eth7iology.
is too much chance in it and too much psychology. Two chess players in presence of the same disposition of the pieces may make different moves; we cannot foresee exactly which, though we can expect one, yet whichever is made we can deduce it from the situation and the end in view. It is perhaps a weakness of the historical sciences that in them we are wise after the event. Yet, in the history of a custom, a more rigorous sequence and prediction may become possible within narrow limits, because individual aberrations compensate one another and leave us with nothing but averages to trouble about, because a mass of conflicting psychological processes are eliminated and only a few broad principles remain. Nor are we studying merely the average conduct of large numbers, but that conduct spread over extensive periods of time, during which the fluctuations compensate one another, so that only constant tendencies survive. Each woman may dress differently, but out of all these differences we can abstract fashion. Fashions may vary erratically from year to year, but, if we take a long period, we find a steady progress from the cumbrous, staid, and elderly, to the light, frivolous, and youthful. Our modern records are so full and minute, and they are so near that we cannot see the wood for the trees ; but we study savages from a distance which blurs the details, and the absence of records leaves us only the outlines to work upon. That is doubtless one reason why ethnology has begun with savages, instead of beginning, like charity, at home.^^
If out of all these fluctuations there results in any period of time, not a standstill but a continuous progression in one direction, it is evident that there must be some factor or factors at work which are constant in their operation ; when they fail we have stagnation. What they may be it will be the task of the future to discover ; it is something if, in the meantime, we have determined their direction.
^^Poincare, op. cil., p. 211.