But though there was more unity than in Kant’s system, still, as the subject stood, there was not as much as might be desired. Why should there be three principles of reasoning, and what have they to do with one another? This question, which was connected with other parts of my schedule of philosophical inquiry that need not be detailed, now came to the front. Even without Kant’s categories, the recurrence of triads in logic was quite marked, and must be the croppings out of some fundamental conceptions. I now undertook to ascertain what the conceptions were. This search resulted in what I call my categories. I then named them Quality, Relation, and Representation. But I was not then aware that undecomposable relations may necessarily require more subjects than two; for this reason Reaction is a better term. Moreover, I did not then know enough about language to see that to attempt to make the word representation serve for an idea so much more general than any it habitually carried, was injudicious. The word mediation would be better. Quality, reaction, and mediation will do. But for scientific terms, Firstness, Secondness, and Thirdness, are to be preferred as being entirely new words without any false associations whatever. How the conceptions are named makes, however, little difference. I will endeavor to convey to you some idea of the conceptions themselves. It is to be remembered that they are excessively general ideas, so very uncommonly general that it is far from easy to get any but a vague apprehension of their meaning… .