may not come, and which yet is wide enough to embrace a real self. Hence, if the test is only sameness for an outside observer, it seems clear that sometimes a man’s life must have a series of selves. But at what point of difference, and on what precise principle, that succession takes place seems not definable. The question is important, but the decision, if there is one, appears quite arbitrary. But perhaps, if we quit the view of the outside observer, we may discover some principle. Let us make the attempt.
We may take memory as the criterion. The self, we may hold, which remembers itself is so far one; and in this lies personal identity. We perhaps may wish also to strengthen our case by regarding memory as something entirely by itself, and as, so to speak, capable of anything whatever. But this is, of course, quite erroneous. Memory, as a special application of reproduction, displays no exceptional wonders to a sane psychology, nor does it really offer greater difficulties than we find in several other functions. And the point I would emphasize here is its limits and defects. Whether you take it across its breadth, or down its length, you discover a great want of singleness. This one memory of which we talk is very weak for many aspects of our varied life, and is again disproportionately strong for other aspects. Hence it seems more like a bundle of memories running side by side and in part unconnected. It is certain that at any one time what we can recall is most fragmentary. There are whole sides of our life which may be wanting altogether, and others which will come up only in various degrees of feebleness. This is when memory is at its best; and at other times there hardly seems any limit to its failure. Not only may some threads of our bundle be wanting or weak, but, out of those that remain, certain lengths may be missing. Pieces of our life, when we were asleep, or drugged, or otherwise distempered,
