place of origin remains, wherever that may be; meanwhile the center of distribution shifts like the center of low barometric pressure, while the center of intensive development is likewise as liable to shift as the center of a cyclone area. Cataclysms are more startling than slow infiltrations, but the latter are none the less effective in making significant changes. Culture seeps as well as flows, but does not seep equally in all directions; and when it flows, often the stream is largest far from its source. We cannot infer age from distribution unless wise enough to make the inference at the proper historical moment, and, unless we have history at our service we shall not know the proper moment for making the inference. In the absence of history, then, we can make no such inference, for we are as likely to make the wrong one as the right one. The earliest features have the first opportunities for diffusion, but later ones appear and submerge the earlier. As soon as this occurs and it is constantly occurring -the more recent features are the more widely distributed. Is this not the manner in which development, advance, change, take place? When change takes place, some recent traits become more widely distributed than certain older ones. The assumption that distribution is an index of age is based on the implicit assumption that, other things being equal, the more widely distributed feature is the older one, since culture traits have a tendency to spread from place of origin to adjoining regions. But other things are never equal, being, on the contrary, egregiously unequal. Differences in culture areas must be taken into account. One culture is ready for the trait and receives it; another is unprepared, and rejects it. There is a dynamic in culture traits which further their distribution in unequal degree. Agriculture, for example, may spread with far greater rapidity than does pottery or basketry; canoes may be adopted with much greater readiness than dog transportation, or, of course, the other way round, depending on a number of factors of which usually we are ignorant. The Ghost Dance religion of aboriginal North America spreads rapidly over an area in which we find no other trait of as great distribution, though we happen to know that almost all other traits of the area in question are older. Objection may be made that analogy from our own culture is misleading, inasmuch as the possibilities of diffusion are greater in civilization than in savagery. But the difference is merely an intensification of the factors at work in savagery, rather than the introduction of new ones. Fads spread with great rapidity in civilization, but fads spread also in savagery. The barrier of language is greater in savagery, but is no insurmountable barrier, and culture influences break through. The savage is slow to perceive the utility of a new device. In time, however, he perceives it, provided the new device is not too far removed from his culture setting. This holds also for civilization: witness the distribution of the telephone, spreading with more rapidity in America than in England, more rapidly in the commercial districts than in the non-commercial, more rapidly where science is developed and applied than where this is not the case. If we are not to make application of the law of diffusion in civilization where history can check up inference, where are we to test it? Without testing it, how can we have confidence in it? To sum up: the conditions which lead to intensive cultivation of a trait are various: the region in which a trait originates is not always the region in which the trait attains intensive development. Comparative distribution is not evidence of comparative age. The tendency to spread differs with traits and differs with culture areas. That traits disappear is as well established as that traits appear. The disappearing trait shifts the bounds of distribution, giving negative correlation of age with geographical distribution. If certain isolated similar traits are to be interpreted as due to diffusion rather than to independent origin, then the sparsely distributed trait is indicative of greater age than one uniformly distributed over the area in question. As the trait drops out in area after area, the age of the trait becomes correlated with sparseness of distribution rather than with extensiveness of dis-. tribution. But if the one criterion is as good an index to age as the other, then neither can be relied upon. UNIVERSITY OF MINNESOTA, MINNEAPOLIS, MINNESOTA. MAYA, NAHUATL, AND TARASCAN KINSHIP TERMS THE BY PAUL RADIN HE MAYA, Nahuatl, and Tarascan kinship terms presented below were extracted from the works of Spanish authors, as the citations will show. This work was financially supported by the New York Academy of Sciences. |