To think out a problem is not unlike drawing a caricature. You have to exaggerate the salient point and leave out that which is not typical. "To illustrate a principle ," says Bagehot , "you must exaggerate much and you must omit much." As to the quantity of absolute truth in a thought : it seems to me the more comprehensive and unobjectionable a thought becomes, the more clumsy and unexciting it gets. I like half-truths of a certain kind they are interesting and they stimulate.