________________
86
M. A. MEHENDALE
similar to each other as far as their plural endings are concerned. then the formations plural' has assimilated 'kine' to it.
All analogies, however, are not proportional. In Sanskrit, for example, if *ekadaša has given place to ekadaša on the analogy of dvadaša no such proportion seems to be at work. But this change in the lexical item can also be looked upon as 'assimilation. Two forms *ekadaša and dvadaša which were dissi milar with regard to the vowel length in the syllable preceding -daša are made similar to each other by analogy. In a way, again, 'dvadaša 'assimilates' *ekadaša to give rise to ekadaša.
It is true that, strictly speaking, analogy is not sound change. But this is true to some extent also of assimilation in that assimilation is not a sound change in the sense that a phonemic shift, a split, or a merger is. Both analogy and assimilation affect the phonemic shape of a morpheme but do not seem to affect the phonemic stock of a language. But whereas assimilation occurs within a 'word', analogy works across them.
(III) Regularity of Phonetic Changes
Bloomfield looks upon sound change as reflecting a change in the speaker's manner of articulation. He therefore is of the opinion that a sound change 'affects a phoneme or a type of phonemes either universally or under certain strictly phonetic conditions, and is neither favoured nor impeded by the semantic character of the forms which happen to contain the phoneme" (1933: 364), This view has been endorsed by Hockett (1965: 190-191) while saying that in a statement of the following type,
Parx Dau y (in the environment z)
the environment must be a sound or a finite combination of sounds in the parent language.
The above view has been challenged by (Robert D. King (Historical Linguistics and Generative Grammar p. 119 ff). He finds that there are certain types of sound change exceptions to which cannot be explained in strictly phonetic environments and hence the hypothesis regarding the regularity of phonetic changes should be modified and stated as follows: Phonological change is regular, but its environment cannot always be stated in strictly phonetic terms.
The examples which King gives to refute the Bloomfieldean view of sound change are as follows: (1) The Middle High German e [a] in the word final unaccented position is lost in Standard Yiddish. E. g. tage> teg 'days', erdeerd 'earth'. But in some cases, especially when e is an adjective inflectional ending, it is not lost: di groyse shtot' the big city', dos alte land "the old country. The environments, in the latter examples, are not phonetic but morpho logical (retention of word final -e as an adjective ending).
(2) The sequence [kw] from proto-Mohawk to Mohawk undergoes epenthesis: "kwistos> kewistos 'I am cold', But when the kw sequence arises out Madhu Vidya/369
Jain Education International
For Private & Personal Use Only
www.jainelibrary.org