on the page 459:
It is with this in mind and for want of a better term, that I have come to think of society as a 'set of sets', the sum of all the things that historians encounter in the various branched of our research. I am borrowing from mathematics a concept so convinient that mathematicians themselves distrust it; and I am perhaps using rather a grand word (in French the word for ^set^ is ^ensemble^, which also means 'whole') to underline the obvious truth that ^everything^ under the sun is, and cannot escape being, social. But the point of definition is to provide an approach to a problem, to lay down some guidelines for preliminary observation. If it makes that obervation easier, both at the beginning and in later stages, if it helps to produce an acceptable classfication of the material and to develop the logic of the argument, then the definition is useful and has justified itself. If we use the expression 'set of sets' or 'ensemble des ensembles', does this not usefully remind us that any given social reality we may observe in isolation is itself contained in some greater set; that as a collection of variables, it requires and implies the existence of other collections of variables outside itself?
- Jean-Francois Melon, secretary of John Law, was already saying in 1734: 'There is such an intimate connection between the different parts of Society, that one cannnot strike one part without its having repercussions on the other.'
- Which is much the same as saying today, 'the social process is an indivisible whole' or 'the only possible history is general history', to quote only one or two of a hundred similar remarks. (...)