In order to simplify this analysis it is convenient to adopt in advance a premise (about our normal usage of the words ‘same’ and ‘different’) whose truth is confirmed in the course of the analysis but which will I think be accepted in any case without demur. Unfortunately in talking about the words ‘same’ and ‘different’ I have to use them as well as mention them; and I cannot be precise without misleadingly anticipat­ ing the terms in which their function is later analysed. (As pointed out in Chapter I, I have to use the method I am advocating while I am advocat­ ing it.) I shall therefore state the premise at this stage as vaguely and as unquestion-beggingly as is consistent with its being clear to a co­ operative mind. With this proviso I can best suggest its general import by saying I am assuming the reader’s agreement with the following remarks:

It would not be contrary to ordinary usage to say that the sentences “ Cast iron is brittle” and “ Glass is brittle” say the same thing about cast iron and glass; nor to say that, if the two sentences express true statements, then cast iron is the same as (or similar to) glass in a certain respect-in being brittle. From this it follows that, if we were to say “ Cast iron is the same in breakability as other things correctly called brittle” , we should be conveying-however pedantically-the same in­ formation as is conveyed by saying simply “ Cast iron is brittle” . Analogously, “ This is the same in colour as things ordinarily called blue” would convey the same information as “ This is blue” and “ This is different in colour from things ordinarily called blue” as “ This is not

In fact, when an empirical statement conveys new information, at blue” .