________________
125
Semantic Randomness and the Comparative Method
set of words of randomly different meaning, A is the innovator in that particular respect.
6. In spite of its tenuous character and its precarious standing, this proviso is of uncanny importance. An accident of history has allowed us to lift the veil just a little from a fundamental relationship between the two layers of double articulation, narrow phonological and wide-ranging morphemic structure which will never cease to obsess us. Distinguishing what is distinct is fine, but absolute Manichaean dichotomy is not, and when all is said and done, language is an integral whole.
Annotations :
1. There are others.
2. Something very much like this was in effect said long ago, though it was since
forgotten. See Morpurgo Davies 1998 : 169.
“Dialect borrowing is only half the story. There is another body of material to be taken care of before sound-changes can be formulated : analogic formations that look like sound-changes but aren't. These may in theory and under the same lucky circumstances be recognized in the same way as dialect loans, provided there is an added finding that the discrepancy between the doublets they create recurs also in alternants in paradigms (most of them originally created by sound-changes). The requirement of preliminary winnowing-out of non-sound-changes is, in other words, not always just a pious wish." See Hoenigswald 1983.
REFERENCES
Henry M. Hoenigswald, “Doublets" in F. Agard et al. (ed.), Essays in honor of Charles F. Hockett, 1983 pp. 167-171.
Anna Morpurgo Davies, Nineteenth-Century Linguistics (= Giulio Lepschy (ed.), History of Linguistics, Volume IV). London and New York 1998 : Longman.
