296 



SCIENCE. 



[Vol. XXI. No 539 



the god Nin-ghir-sou has erected." We read, also, in another 

 inscription behind the head of the king: "Nina-Our, King of 

 Sirpoula," and a little above his knees: "from Magan in the 

 mountain quantities of wood he has ordered," and after the last 

 personage figured : " the temple of the goddess Nina he has 

 erected."' We are evidently in presence of a very pious prince, 

 and we know from other sources that he erected or repaired a 

 certain number of temples dedicated to his gods. All the figures 

 and inscriptions show a most primitive art, inferior to that dat- 

 ing from the days of E-anna-dou. They are nevertheless of high 

 value as historical and genealogical records. 



The importance of these discoveries cannot be overrated. That 

 importance resides not only in the insight they bring on the cus- 

 toms, wars, and religions of nations whose very names were un- 

 known but a few years ago, but also in the greater antiquity we 

 must now accept for the origin of manhimself. The dates given 

 for the creation must be amended, as we know now with certi- 

 tude, that in those days men and nations already existed in num- 

 bers, towns were built, monuments were erected, arts were flour- 

 ishing, kingdoms already powerful were in existence, and we 

 find both in Asia and in Africa traces of a civilization which 

 centuries alone could have reared and maintained. 



AN EXPERIMENTAL BASIS FOR LITERARY CRITICISM. 



BY CONWA.Y MAC MILLAN, UNtVERSITT OF MINNESOTA, MINNEAPOLIS, 

 MINN. 



The volume, entitled "Analytics of Literature." just pub- 

 lished by Ginn & Co. of Boston, seems to the writer so epoch- 

 making a work that he takes advantage of the courtesy extended 

 by Science to direct the attention of scientific men in general and 

 biological students in particular to the new and brilliant applica- 

 tion in it of the familiar methods of research which they have 

 themselves used in other departments of investigation. The 

 truth is that there is the emergence of a new science — the sci- 

 ence of experimental criticism, or, if one likes, the science of 

 style-morphology, embryology, and physiology. It is a most 

 noteworthy volume, and though unpretentious and perhaps 

 marred by departures here and there from the strict scientific 

 method, it will take its place with such a work as that of 

 Fechner, in which he brought recalcitrant psychology under the 

 laws of empiricism, and banished the intuitional and closet- 

 metaphysician in the ratio in which he introduced the laboratory 

 method of psycho-physics and the experimental psychologist of 

 the school of Wundt. No more far-reaching scientific work 

 has been done in America than the reduction, in this book by Dr. 

 L. A. Sherman, of so mysterious a matter as literary style to the 

 basis of a department of experimental science. But after an ac- 

 quaintance with the method and an application of it, during the 

 ten years past in which it has been laboriously and carefully devel- 

 oped by its originator, I have no hesitation in pronouncing the 

 work an extraordinary and inspiring advancement of biological 

 methods into a field where, oddly enough, they have not before 

 been employed. 



The new point of view is simply this: style is an institution. 

 It may be considered apart from the message which the writer 

 wishes to convey. Under such an analysis style is found to obey 

 the laws of other institutions or organisms. It is a matter of 

 evolution. It is in any case a structure of which the phylogeny 

 and the ontogeny may be calculated. In the child, one may 

 study the ontogeny of a style, and of children's phases of sen- 

 tence development the author of ' ' Analytics of Literature " 

 gives some valuable examples. And in the literature of the 

 English-speaking peoples there is a vast storehouse of palseonto- 

 logical material from the study of which, after comparison with 

 the ontogenetic development, it is possible to determine some of 

 the laws of style- evolution. Thus a foundation for a style taxo- 

 nomies is laid and one finds that, as one should expect, all the 

 well-known laws of heredity in general and of progress, degen- 



' Amiaud, " Records of the Past," T. I., p. 64 ; Jensen, " Kellenschriftllche 

 BlbUothek," T. Ill , p. 10. 



eration, variation, reversion, or atavism, persistence of type and 

 modification of type in particular, apply to literary styles pre- 

 cisely as to organisms. It becomes possible to determine a style, 

 not in the old intuitional manner of literary art as indicated in 

 Sainte Beauve, Arnold, or Lessing. but in the precise manner of 

 the zoological monograph. It becomes possible to establish 

 genera, species, tribes, orders, if one will, of literary style, and 

 the whole matter of literary criticism at one touch passes over 

 into the domain of natural science, just as music so passed in 

 the thought of Schopenhauer and Wagner, metaphysics under 

 the hand of Wundt, biology by the genius of Aristotle. Bacon, 

 and Darwin. 



The genesis of such a work must be of interest. As indicated 

 in the preface it was a development, not an inspiration. The 

 first published paper that pointed out the objective method in 

 criticism, so far as known to the writer, was that of Sherman in 

 the University Studies, October, 1888.'' Here the matter of en- 

 quiry was the changing length of the sentence in English prose 

 and a number of statistics were presented. It was shown that 

 there has been a progressive shortening of the sentence from 

 early pre-EIizabethan prose to the present. Some data are added 

 here by way of illustration. They are taken from both the 

 Studies article and from the recent volume. 



Average number of words to the sentence in various English 

 writers, computed from prose, on the basis of five hundred 

 sentences. 



These averages once established may be tried in other parts of 

 the works of any author and will be found practically constant. 

 For example, in Macauley's '"Essays " the average length of the 

 sentence is 23.+. Testing by the " History of England " it was 

 found that in this the average of the 41,579 periods counted is 

 23 43 words per sentence. Thus it can be shown in any author 

 who has acquired a style that five hundred or a thousand sen- 

 tences taken at random will establish a sentence-norm for that 

 author and from this norm the variation will be slight. Dispari- 

 ties, too, are greatest in more ancient styles, indicating their less 

 complete organization. For example, in Chaucer the average of 

 Melibceus is 48. -I- , while that of the Parsonn's tale is 36. -F, an 

 almost unparalleled discrepancy. 



It is possible, then, for any author to plot a curve of sentence- 

 length, and when this is done the surprising fact stands forth 

 that the average is brought about by " evening-up " a compara- 

 tively large number of sentences only a little shorter than the 

 mean with a comparatively few sentences excessively long. 

 Since the long sentence is clearly shown by palseontologic inves- 

 tigations to be the older type in any literature, it appears that in 

 modern stylists, even, there is an atavistic tendency, and this is 

 capable of beautiful and instinctive comparison with the per- 

 sistent style.s of low type that can be picked up anywhere — in 

 newspaper-advertisements or in cheap novels — where, if there 

 is an independent style at all. it will be one of older and lower 

 organization. 



Even in the preliminary analysis of sentence-length, singular 

 and unintelligible facts have been discovered that demand further 

 investigation before their import can be known. DeQuincy is 

 peculiar in the number of prime-sentences, those in which the 

 number of words is indivisible by any quantity but the number 

 itself and unity. Curious lapses into ancient manner in mod- 

 erns and astonishing forecasts of modern manner by ancients 



^ On the Sentence-length in English Prose, pp. 119-130. 



