The Internal Logical Structure of the Comparative Method
In the above discussion there has been an implicit assumption apparently shared by all historic linguists that there exists either a set of sound laws or a valid reconstruction that is, as it were, timelessly true. We need to distinguish here two things. One is the real history of investigations by historical linguists as they take place over time. The other is logical priority of certain steps in the process in relation to other, We have already seen something of the notion of logical priority in the discussion of the relation of sound correspondences to reconstruction. In fact, the logically prior is likely to occur first historically. Sound correspondences are noted before proto phonemes and reconstructed forms are set up. The two kinds of priority both figure in the following statement of Delbrück (1888:47), the collaborator with Brugmann in the great comparative grammar of Indo-European.
Since obvious etymologies are the materials from which sound laws are drawn, and this is material can always be increased or changed, therefore new sound laws can continually be recognized and old ones transformed.
First, Delbrück is obviously portraying the development of linguistics over time. New sound laws arise and old ones are changed. The initial sentence, however, contains the key notion that sound laws are in the first place drawn from obvious etymologies. But such obvious etymologies have then a kind of logical priority and must somehow be recognized by properties not involving sound laws since sound laws are derived from them.
Others have made essentially the same point. Goddard (1975:25) says,
In general, the establishing of phonological correspondences goes on within a family of languages known to be related…, quoted with approval by Campbell and Mithum (1979:52).
Somewhat earlier Newman (1970:39) had similarly stated,
The proof of genetic relationships does not depend on the demonstration of sound laws. Rather, the discovery of sound laws and the reconstruction of linguistic history normally emerge from careful comparison of languages known to be related.
Regarding Newman’s observation, Watkins, a distinguished Indo-Europeanist, has said more recently (1990:292), ‘As to the mystique of sound laws on the other hand, Greenberg is quite right to quote with approbation the Africanist Paul Newman (1970)’.
None of these citations face the problem of the initial step, namely how we recognize ‘a family of languages known to be related’ (Goddard) or ‘languages already presumed to be related’. (Newman). What I believe does emerge here is the recognition of at least three stages, which are in order of logical and usually historical priority the following: (1) Recognition of a family of related languages; (2) The discovery of sound laws; (3) The application of the comparative method, starting with sound laws, leading to the reconstruction of linguistic history and the proto-language. In this process, sound laws are often revised or abandoned and new ones found. A core of basic etymologies furnishes the starting point (cf. Delbrück above) and it is rare for any of these to be abandoned. New ones are found, and some which are proposed become matter of debate with some linguists accepting and some rejecting. Etymology will never cease to exist as a field of study in any linguistic stock simply because more etymological problems will always exist.
