Now Bergson says that such errors are not confined to stray instances, such as we have noticed, in which the imposition of preconceived ideas can readily be detected by a little closer attention to the actual facts. He believes that a falsification due to preconceived ideas, runs right through the whole of our direct experience. He lays the blame both for this falsification and for our failure to detect it upon our intellectual habit of relying upon explanation rather than upon direct knowledge, and that is one of the reasons why he says that our intellectual attitude is an obstacle to direct knowledge of the facts. The intellectual method of abstraction by which we analyse and classify is the foundation of all description and explanation in terms of general laws, and the truth is that we are, as a rule, much more preoccupied with explaining the facts which we know than with the actual experiencing of them. This preoccupation is natural enough. The bare fact which we know directly is not enough to enable us to carry on our everyday lives, we cannot get on unless we supplement it with some sort of explanation and, if it comes to choosing between fact and explanation, the explanation is often of more practical use than the fact. So it comes about that we are inclined to use the facts which we know directly simply as material for constructing explanations and to pay so little attention to them for their own sakes that we simply take it for granted that they must be what our explanations lead us to suppose they are. Now according to Bergson the attitude of mind required for explaining the facts conflicts with that which is required for knowing them. From the point of view simply of knowing, the facts are all equally important and we cannot afford to discriminate, but for explanation some facts are very much more important than others. When we want to explain, therefore, rather than simply to know, we tend to concentrate our attention upon these practically important facts and pass over the rest. For in order to describe and explain a situation we have to classify it, and in order to do this we must pick out in it properties required for membership of some one or other of the classes known to us. In the situation which we originally considered by way of illustration, for instance, you had to pick out the qualities of roughness, warmth and so on, in order to classify what you had stumbled upon as “a dog.” Now the picking out of these particular qualities is really an operation of abstraction from the situation as a whole: they were the important features of the situation from the point of view of classifying what you had stumbled upon, but they by no means exhausted the whole situation. Our preoccupation with explaining the facts, then, leads us to treat what we know directly as so much material for abstraction.