The usual work of the intellect consists in analysis and classification: if you have anything presented to you which you do not understand the obvious question to put yourself is, “what is it?” Suppose in a dark room which you expected to find empty you stumble against something, the natural thing to do is to begin at once to try to fit your experience into some class already familiar to you. You find it has a certain texture which you class as rather rough, a temperature which you class as warm, a size which you class as about two feet high, a peculiar smell which you recognise and you finally jump to the answer to your question: it is “a dog.” This intellectual operation is a sample of the way in which it comes natural to us to set to work whenever we find ourselves confronted with any situation which we are not able to classify off hand, we are not easy till we can say _what_ the situation is, and saying _what_ consists in hitting upon some class with which we are already familiar to which it belongs: in this instance the question was answered when you succeeded in describing the situation to yourself as “stumbling upon a dog.” Now you were only able to class what was stumbled upon as a dog after you had recognised a certain number of properties as being those shared by dogs—the rough texture, the size, the smell. You analysed the situation as containing these qualities and thereupon classified what had been stumbled upon as a dog. Analysis and classification are the two methods which we are accustomed to rely upon for improving our knowledge in unfamiliar situations and we are accustomed to take it that they improve our knowledge of the whole situation: anyone who said that after you were able to say what you had stumbled upon you knew less of the whole situation than you knew before would find it difficult to get you to agree. And yet this is very much the position which Bergson takes up. Analysis and classification, he would admit, are the way to get more knowledge, of a kind; they enable us to describe situations and they are the starting point of all explanation and prediction. After analysis and classification you were able to say, “I have stumbled upon a dog,” and having got so far you could then pass on to whatever general laws you knew of as applying to the classes into which you had fitted the situation, and by means of these laws still more of the situation could be classified and explained. Thus by means of the general law, “dogs lick,” you would be furnished with an explanation if perhaps you felt something warm and damp on your hand, or again knowledge of this law might lead you to expect such a feeling. When what we want is to describe or to explain a situation in general terms then Bergson agrees that analysis and classification are the methods to employ, but he maintains that these methods which are useful for describing and explaining are no use for finding out the actual situation which we may