Mill on Kinds and Types

A while back I excerpted some Whewell on classification by types. Here is John Stuart Mill disagreeing with him, and, I think, starting off the modern literature on natural kinds.

Kinds are Classes between which there is an impassable barrier; and what we have to seek is, marks whereby we may determine on which side of the barrier an object takes its place. The characters which will best do this should be chosen: if they are also important in themselves, so much the better. When we have selected the characters, we parcel out the objects according to those characters, and not, I conceive, according to resemblance to a type. We do not compose the species Ranunculus acris, of all plants which bear a satisfactory degree of resemblance to a model-buttercup, but of those which possess certain characters selected as marks by which we might recognise the possibility of a common parentage; and the enumeration of those characters is the definition of the species.

The question next arises, whether, as all Kinds must have a place among the classes, so all the classes in a natural arrangement must be Kinds? And to this I answer, certainly not. The distinctions of Kinds are not numerous enough to make up the whole of a classification. Very few of the genera of plants, or even of the families, can be pronounced with certainty to be Kinds. The great distinctions of Vascular and Cellular, Dicotyledonous or Exogenous and Monocotyledonous or Endogenous plants, are perhaps differences of Kind; the lines of demarcation which divide those classes seem (though even on this I would not pronounce positively) to go through the whole nature of the plants. But the different species of a genus, or genera of a family, usually have in common only a limited number of characters. A Rose does not seem to differ from a Rubus, or the Umbelliferae from the Ranunculaceae, in much else than the characters botanically assigned to those genera or those families. Unenumerated differences certainly do exist in some cases; there are families of plants which have peculiarities of chemical composition, or yield products having peculiar effects on the animal economy. The Cruciferae and Fungi contain an unusual proportion of nitrogen; the Labiatae are the chief sources of essential oils, the Solaneae are very commonly narcotic, &c. In these and similar cases there are possibly distinctions of Kind; but it is by no means indispensable that there should be. Genera and Families may be eminently natural, though marked out from one another by properties limited in number; provided those properties are important, and the objects contained in each genus or family resemble each other more than they resemble anything which is excluded from the genus or family.

After the recognition and definition, then, of the infimae species, the next step is to arrange those infimae species into larger groups: making these groups correspond to Kinds wherever it is possible, but in most cases without any such guidance. And in doing this it is true that we are naturally and properly guided, in most cases at least, by resemblance to a type. We form our groups round certain selected Kinds, each of which serves as a sort of exemplar of its group. But though the groups are suggested by types, I cannot think that a group when formed is determined by the type; that in deciding whether a species belongs to the group, a reference is made to the type, and not to the characters; that the characters “cannot be expressed in words.” This assertion is inconsistent with Dr. Whewell’s own statement of the fundamental principle of classification, namely, that “general assertions shall be possible.” If the class did not possess any characters in common, what general assertions would be possible respecting it? Except that they all resemble each other more than they resemble anything else, nothing whatever could be predicated of the class. [System of Logic IV.vii.4 (Mill 2006)]

The bolded passage (my emphasis) is rather crucial. As we know, Mill thought that induction was the only way in which truths are ascertained, but that classification is, as the section title has it, "subsidiary to induction". Hence the definitional aspect of classification comes to the fore for him. For Whewell, however, classes are formed inductively, and hence types have no necessary and sufficient properties. The modern literature on natural kinds takes the Millian tine of the fork, while the Whewellian, and I think cladistic, approach to classification is that it is the outcome of inductive generalisation. Mill thinks that to make generalisations, a set of properties have to be shared by all (and only) the members of the class. Whewell thinks that to make generalisations, types serve as anchor points for our taxonomic descriptions. Modern cladistics, especially in the way it was set up by Rosen, Patterson, Nelson and Platnick, holds that a phylogeny permits inductive generalisations without the clades being definitionally tight, and so they fit in between Whewell and Mill. More on this later...

Reference

Mill, John Stuart. 2006. A System of Logic Ratiocinative and Inductive, Being a Connected View of the Principles of Evidence and the Methods of Scientific Investigation (Books IV-VI and Appendices). Edited by J. M. Robson. 33 vols. Vol. VIII, The Collected Works of John Stuart Mill. Toronto, London: University of Toronto Press, Routledge and Kegan Paul. Original edition, 1974. [NB: This complete critical edition of Mill's works is available online - legitimately! - from Online Library of Liberty. Take advantage of it!