I do not mean to deny that it is a good thing to discover causal sequences in history when it is possible, but I think the possibility exists only in rather limited fields. Gresham’s law that bad money drives out good is an example of one of the best established of such causal sequences. The whole science of economics, in so far as it is valid, consists of causal laws illustrated by historical facts. But as everybody now recognizes, supposed laws of economics have a much more temporary and local validity than was thought a hundred years ago. One of the difficulties in searching for such laws is that there is not so much recurrence in history as in astronomy. It may be true, as Meyers maintains in his little book on The Dawn of History, that on four separate occasions drought in Arabia has caused a wave of Semitic conquest, but it is hardly to be supposed that the same cause would produce the same effect at the present day. Even when historical causal sequences are established as regards the past, there is not much reason to expect that they will hold in the future, because the relevant facts are so complex that unforeseeable changes may falsify our predictions. No historian, however scientific, could have predicted in the fourteenth century the changes brought about by Columbus and Vasco da Gama. For these reasons I think that scientific laws in history are neither so important nor so discoverable as is sometimes maintained.
出典：History as an art (1954)
There is another sense in which history attempts to be scientific, and this sense raises more difficult questions. In this sense history seeks to discover causal laws connecting different facts, in the same sort of way in which physical sciences have succeeded in discovering interconnections among facts. The attempt to discover such causal laws in history is entirely praiseworthy, but I do not think that it is what gives the most value to historical studies. I found an admirable discussion of this matter in an essay which I had read forty years ago and largely forgotten: I mean George Trevelyan’s Clio, a Muse. He points out that in history we are interested in the particular facts and not only in their causal relations. It may be, as some have suggested, that Napoleon lost the Battle of Leipzig because he ate a peach after the Battle of Dresden. If this is the case, it is no doubt not without interest. But the events which it connects are on their own account much more interesting. In physical science, exactly the opposite is true. Eclipses, for example, are not very interesting in themselves except when they give fixed points in very early history, as is the case with the eclipse in Asia Minor which helps to date Thales and the eclipse in China in 776 B.C. (Some authorities say that it was in 775 B.C. I leave this question to historians and astronomers.) But although most eclipses are not interesting in themselves, the laws which determine their recurrence are of the very highest interest, and the discovery of these laws was of immense importance in dispelling superstition. Similarly, the experimental facts upon which modern physics is based would be totally uninteresting if it were not for the causal laws that they help to establish. But history is not like this. Most of the value of history is lost if we are not interested in the things that happen for their own sakes. In this respect history is like poetry. There is a satisfaction to curiosity in discovering why Coleridge wrote “Kubla Khan” as he did, but this satisfaction is a trivial affair compared to that which we derive from the poem itself.
出典：History as an art (1954)
歴史(学)は科学であると（人が）言うとき，それには二つの非常に異なった意味があります。（まず)科学は歴史的事実を確認することに関係しているという，比較的陳腐な意味があります。このことは,証拠が乏しくて明瞭でない,初期の歴史（歴史時代の初期）において,特に重要ですが，証拠の間にありがちな対立（矛盾）が生じた場合にはいつでも，もっと最近の時代でも，起こります。我々は，プロコピウス(Procopius, 326-366， ４世紀のローマ帝国における東方の帝位簒奪者(さんだつしゃ)，在位：365-366)をどの程度信ずべきでしょうか？ セントヘレナ（島）でのナポレオンの労作（incubration）から何か歴史的価値を発見する（見出す）ことができるでしょうか？（注：セントヘレナでナポレオンが執筆した『回顧録』のことを言っているのか？） このような問いは，ある意味で科学的です。なぜなら，異なった証拠のそれぞれの情報源に対して付与される重み付け（重要性）に関係しているからです。それらは，それに関する考察が不明瞭かつ専門的になりがちなので，歴史家が他の歴史家に，当然なこととして､話しかける事柄です。この種の仕事は，大規模な歴史書を執筆しようとする時に前提とされるものです。歴史は，芸術としてどれほど追求されようとも，事実に忠実であることによって制御されなくてはなりません。事実よりも真実を大事にすること（truth to fact 事実に対する真実？）は芸術の原則ですが，その原則はそれ自身で芸術の卓越性を与えるものではありません。それ（その原則）は,必ずしも出来上った結果をよいものとしないソネットの規則のようなものです。しかし，歴史学は，歴史家が事実への忠実性を維持するためにできるだけのことをしなければ，最大限,純粋に芸術的な観点からみても，賞讃に価するものであることはできません。この意味での科学は，歴史の研究において，絶対的に必須のものです。
When people speak of history as a science, there are two very different things that may be meant. There is a comparatively pedestrian sense in which science is involved in ascertaining historical facts. This is especially important in early history, where evidence is both scarce and obscure, but it arises also in more recent times whenever, as is apt to be the case, there is a conflict of testimony. How much are we to believe of Procopius? Is there anything of historical value to be made out of Napoleon’s lucubrations in St. Helena? Such questions are in a sense scientific, since they concern the weight to be attached to different sources of evidence. They are matters as to which the historian may justifiably address himself to other historians, since the considerations involved are likely to be obscure and specialized. Work of this sort is presupposed in any attempt to write large-scale history. History, however much it may be pursued as an art, has to be controlled by the attempt to be true to fact. Truth to fact is a rule of the art, but does not in itself confer artistic excellence. It is like the rules of the sonnet, which can be scrupulously observed without conferring merit on the result. But history cannot be praiseworthy, even from the most purely artistic point of view, unless the historian does his utmost to preserve fidelity to the facts. Science in this sense is absolutely essential to the study of history.
出典：History as an art (1954)