It might seem that we are now thrown back on what may be called naive history—on the belief that the facts will speak for themselves, and presumably even select themselves, that all historical events are unique, and that generalisation is the crime against history. Of course, this belief is nonsense. All coherent thought requires general ideas: classificatory terms are necessary for rational discourse. A unique event could only be a miracle. Historical research begins with the discovery of ‘facts’ it is true, but, to be understood, these require to be given some meaning, just as do the facts that impinge on us, or that we observe, in everyday life. The task of reacting to common experience provides us with a set of rough concepts for dealing with these, and naive history gets little farther.

A more sophisticated kind of history calls for a more sophisticated analysis. This is not to say that it needs a general theory, but it does need a language. In some respects it already has this. Thus, most of the history that has been written is political history, and since a long series of political scientists, from the time of Aristotle to the present day, have analysed political structures and events, there is no lack of a coherent and comprehensible political vocabulary, though even this requires to be continually redefined and brought into relation with new usages and changing political conditions. The language of economics was a much later invention. It did not exist when Hobbes, wanting to describe what we would call consumption and production, had to entitle his chapter ‘Of the Nutrition and Procreation of a Commonwealth’.