In order to establish the identities and differences existing between all natural entities, it would be necessary to take into account every feature that might have been listed in a given description. Such an endless task would push the advent of natural history back into an inaccessible never-never land, unless there existed techniques that would avoid this difficulty and limit the labour of making so many comparisions. It is possible, a priori, to state that these techniques are of two types. Either that of making total comparisons, but only within empirically constituted groups in which the number of resemblances is manifestly so high that the enumeration of the differences will not take long to complete; and in this way, step by step, the establishment of all identities and distinctions can be guaranteed. Or that of selecting a finite and relatively limited number of characteristics, whose variations and constants may be studied in any individual entity that presents itself. This last procedure was termed the System, the first the Method.
Michel Foucault, The Order of Things. (Random House, 1970) p140.