Angiosperm classification is to be understood in relation to its development against a particular historical and philosophical background—that of seventeenth- and eighteenth-century Europe. The selection of the world's flora and the botanical literature available to Linnaeus and his immediate successors determined the main shape of the classification which we use today.
Willis's generalization about the relationship between the age of a taxon and its size (part of his so-called ‘A.S.A. hypothesis’) is valid enough, both for families (if size is judged by number of genera) and for genera (if size is judged by number of species). Willis, however, thought of the age of the taxon in evolutionary terms; all that is necessary to validate the generalization is to interpret ‘age’ quite literally as the number of years since the particular taxon was created in the mind of the taxonomist.
A corollary of this thesis is that we have no reason to think that Angiosperm classification would be substantially the same if botany had developed in, say, New Zealand in the nineteenth century instead of medieval and post-medieval Europe. The stability of Angiosperm families, essentially unchanged since de Jussieu, is no evidence of their ‘correctness’; family boundaries might be seriously different, and the main reference purposes of the classification still be adequately served.
Taxonomists should accept that their function is to provide a convenient reference system for the science as a whole; they would then be very conservative in suggesting changes to the existing system, which is convenient enough, and could devote their efforts more practically to making available in a reasonably accessible form the enormous body of taxonomic information already accumulated in the specialist literature.