COMPENSATORY LENGTHENING: IN DEFENSE OF THE CONCEPT 'MORA'

1. In a 1979 paper on compensatory lengthening (CL), DE CHENE and ANDERSON (D & A) make two major claims: (a) CL is one of the many terms of traditional historical linguistics which label diachronic correspondences without explaining them; (b) the correct phonetic explanation is that 'such changes always involve the loss of a consonant immediately adjacent to the vowel in which length subsequently appears, and . . . the change can be analysed quite generally into the two independently-motivated processes of weakening of a consonant to a glide (either semivocalic or laryngeal) and subsequent monophthongization of a complex syllable nucleus. (505) That is, CL is a misnomer. A first step toward a rebuttal of this position is found in CLEMENTS (1982; cf. also 1984). While more or less sharing D & A's negative evaluation of traditional linguistics (3), he proposes a phonological explanation of CL äs serving to 'maintain the quantitative integrity of the syllable.' (6) Unfortunately, however, most of CLEMENTS' examples are of doubtful cogency: A fair number involve consonant gemination which, äs shown in HOCK (1976a) and MTJKRAY and VENNEMAN (1983), may result from a variety of other, non-CL developments. The Bantu example of mun$tu to munSntn (my notation), with CL aUegedly resulting from the shift of n from the first to the second syllable, can be attributed also to lengthening before sonorant + consonant which, äs OE cild > clld, NE child shows, may take place without a shift in syllable boundary. The dialectal American English Variation between [kamt] and [kse:t] is of doubtful usefulness, since äs far äs I can teil, the vowel is long not only in the n-less form, but also in the form with n. The regional American English Variation between forms like [ka:t] and [kart] is