ABSTRACT
A fuller answer, then, than any which Kant gives is called for to the question how we should describe that condition of the significant employment of concepts which is loosely expressed by saying that they must be so employed as to have application in a possible experience. I shall not attempt to answer the question. But the main conditions which a fuller description would have to provide for can be indicated. We should begin by noting that there is little difficulty in explaining what we mean by observational criteria for the application of many kinds of established, unproblematic concepts. Then we may remark that in order for a newly introduced or problematic concept (or concept-extension) to have significant employment, it is necessary that it should be possible to state or indicate types of observable situation in which it has application - which is not to say that its objects must be observable. It is further necessary that its application in such a situation should have consequences or implications which do not merely duplicate those of other, established, non-problematic concepts. The temptation here is to say "testable consequences or implications"; and that addition, though calling for further elucidation, is no doubt on the right lines in so far as we are concerned with the concepts we call scientific. But perhaps we can also be said to extend our knowledge of the world by learning to see it afresh, to extend or modify our classifications and descriptions, in ways and directions with which natural science has little to do. And then the consequences will relate more to the nature of our experience itself and to the connexions we make in it than to the possibility of confirmatory or disconfirmatory tests which we might undertake.
