Is It Possible to Prove Any General Statements about Historical Fact?

^ 7 zURING THE past eleven years it has been my habit to read alterJnately books on history and books on statistical method. The contrast , between these two disciplines has provoked the following remarks, which were also stimulated by a broadcast discussion on 4 Januaty, I948, between Professor P. Geyl of Utrecht and Professor A. J. Toynbee. Examples of generalizations concerning historical fact will be taken chiefly from four works: Pitrim Sorokin's Social and Ct4ltural Dynamics; Quincy Wright's A Study of War; my own microfilm Statistics of Deadly Quarrels; and A. J. Toynbee's A Stmhdy of History, as abndged by D. C. Somervell. On the literary grace, poetry, philosophy and religion which pervade Toynbee's book I shall not presume to comment, except that they make it delightful to read, and that the reader may find it difficult to maintain his critical judgment while thus enchanted. As to Toynbee's particular historical illustrations, so impressively numerous and gathered from all over the world, I shall respectfully assurne them all to be accurately ascertained facts. The question at issue is whether his endeavours to prove generalizations have succeeded. The statisticians have agreed on certain proprieties of method which they regard as important. Some of the simplest of these will be mentioned in turn, and the question will be raised whether they have been, or can be, attained when the subject matter is history. In this paper, intended for historians, it would be inappropriate to go into the mathematical elaborations in which professional statisticians delight. Suffice it to say that by statistical methods I am alluding to techniques more fully described in such books as Yule and Kendall's An Introduction to the Theory of Statistics, R. A. Fisher's Statistical Methods Jor Research Workers, Harold Jeffreys' Theory of ProbabZity, and M. G. Kendall's Advanced Theory of Statistics. Detailed references are collected in a list at the end of this article. Statisticians, like other specialists, while showing a united front to laymen, have their internal controversies, the most notable of which relates to the 77