evidence may be here adduced in support of this theory. The languages of the lower races use the sound pu to express an evil smell; the Zulu remarks that 'the meat says pu' (inyama iti pu), meaning that it stinks; the Timorese has poöp 'putrid;' the Quiché language has puh, poh 'corruption, pus,' pohir 'to turn bad, rot,' puz 'rottenness, what stinks;' the Tupi word for nasty, puxi, may be compared with the Latin putidus, and the Columbia River name for the 'skunk,' o-pun-pun, with similar names of stinking animals, Sanskrit pûtikâ 'civet-cat,' and French putois 'pole-cat.' From the French interjection fi! words have long been formed belonging to the language, if not authenticated by the Academy; in mediæval French 'maistre fi-fi' was a recognized term for a scavenger, and fi-fi books are not yet extinct.
There has been as yet, unfortunately, too much separation between what may be called generative philology, which examines into the ultimate origins of words, and historical philology, which traces their transmission and change. It will be a great gain to the science of language to bring these two branches of enquiry into closer union, even as the processes they relate to have been going on together since the earliest days of speech. At present the historical philologists of the school of Grimm and Bopp, whose great work has been the tracing of our Indo-European dialects to an early Aryan form of language, have had much the advantage in fulness of evidence and strictness of treatment. At the same time it is evident that the views of the generative philologists, from De Brosses onward, embody a sound
living sounds, are affected by phonetic changes such as that of Grimm's law, which act on articulate sounds no longer fully expressive in themselves, but handed down by mere tradition. Thus p and f occur in one and the same dialect in interjections of disgust and aversion, pub! fi! being used in Venice or Paris, just as similar sounds would be in London. In tracing this group of words from early Aryan forms, it must also be noticed that Sanskrit is a very imperfect guide, for its alphabet has no f, and it can hardly give the rule in this matter to languages possessing both p and f, and thus capable of nicer appreciation of this class of interjections.
