Every historian must have a thesis, some principle of illumination to guide him, and the value of his work will largely depend upon the sanity and profundity of that thesis. But I would suggest that, the subject-matter of history being what it is, we should be chary of becoming too dogmatic about any principle of interpretation which we put forward. For history works under conditions wholly unlike those of the natural sciences, and historic truth must be something very different from mathematical truth, or even from biological truth. […]
The older school, of the type of Buckle and Guizot, believed that they had established historical laws of universal validity, and provided a clockwork uniformity of effects and causes. It would appear that they misunderstood the kind of material with which they had to deal. M. Bergson has shown us that half the blunders of philosophy are due to the application of the methods and ideals of physical science to spheres of thought where they are strictly inapplicable. In the kaleidoscope of the past we cannot, as a rule, sort out effects and causes with any precision, nor can we weigh events in the meticulous scales which science demands. Even when causes are reasonably plain, their classification eludes us. We cannot tell which is the causa causans, which are proximate, or efficient, or final. We must be content with generalisations which are only generalisations and not laws, with broad effects and massed colours. The weakness of the scientific historian is that he underrates the complexity of human nature. He would turn mankind into automata, motives into a few elementary emotions, and the infinitely varied web of life into a simple geometrical pattern. Order and simplicity are great things, but they must be natural to the subject and not due to the blindness of the historian.
(John Buchan, The Causal and the Casual in History. Cambridge: Cambridge University Press, 1929, pp. 10–12).