It may be open to question how much of this virtual knowledge of both past and present we ever could know directly in any sense comparable to the way in which we know the fact actually presented at some given moment, however perfectly we might succeed in ridding ourselves with our intellectual pre-occupation with explaining instead of knowing; but, if reality forms an unbroken whole in duration, we cannot in advance set any limits, short of the whole of reality, to the field of virtual knowledge. And it does really seem as if our pre-occupation with discovering repetitions in the interests of explanation had something to do with the limited extent of the direct knowledge which we ordinarily enjoy, so that, if we could overcome this bias, we might know more than we do now, though how much more it is not possible, in advance, to predict. For in the whole field of virtual knowledge, which appears to be continuous with the little scrap of fact which is all that we usually attend to, present bare sensation and such bare sensations as resemble it, form very insignificant elements: for purposes of abstraction and explanation, however, it is only these insignificant elements that are of any use. So long, therefore, as we are preoccupied with abstraction, we must bend all our energies towards isolating these fragments from the context which extends out and out over the whole field of virtual knowledge, rivetting our attention on them and, as far as possible, ignoring all the rest. If Bergson’s theory of virtual knowledge is correct, then, it does seem as if normally our efforts were directed towards shutting out most of our knowledge rather than towards enjoying it, towards forgetting the greater part of what memory contains rather than towards remembering it.