________________ REVIEWS 235 contain well-organized lists of examples which should prove useful to students. Only the Tamil volume contains a bibliography, and its references are mainly to works on phonetics rather than Tamil. $2. Upadhyaya's Kannada specifies that it deals with "formal speech by the educated people of the southern part of Mysore state", he also gives some information on colloquial variants. (The other three volumes, regrettably, never state what style or regional dialect they are describing; but they too seem to concentrate on formal varieties.) Comments on specific passages now follow. P.4: something behind the "teeth ridge" is said to be "known as alveolum"; there is, of course, no such word. P. 14: in a, "the back of tongue is slightly lowered from its neutral position": how is the student to know what position is "neutral"? P. 15: the term "gemination" is introduced here with no explanation to the student. P. 21:j is normally an affricate, as U. says; but the cluster js, as in his example jnaana, is commonly pronounced with a palatal STOP. P. 23: aspirates occur not only in loans from Sanskrit, as stated, but in many words borrowed from Hindi-Urdu, e.g. khaali "empty' (p. 29). P. 28: n is said here to occur "only before palatal affricates"; but again note jnaana "knowledge and some related words. Pp. 28-9; it should be noted that f in loanwords is replaced by ph or p for many speakers; and that another borrowed fricative, z, is used by many speakers (alternating with j), e.g. in dazan 'dozen'. P. 30: U. notes that many speakers replace s by s; but it might be added that this tends NOT to happen in the cluster st, as in kasta 'difficult'. P. 50: the graphemes r and rr are said to be "rarely used"; in fact, rr is NEVER used (except when writing the alphabet!) Something should be said about the pronunciation of r, which varies between [ri] and [ru]. U. mentions visarga, "a glottal fricative", but doesn't say whether it is pronounced the same as h.8 83. Syamala Kumari's Malayalam deals with a more complex system, one which is of considerable interest to general phonology because of its surface contrast of dental, alveolar, and retroflex stops, and of nasals in six articulatory positions. The following comments apply to her description." Pp. 7-8: in a, the tongue is said to be "as low as possible"; in aa, "the tongue position is lower and backer than [a]"! P. 8: in the production of initial ["0), SK says, "the tongue starts to produce a fronter sound like [w] but soon retracts back and produces the sound [o]". This makes no sense, but perhaps a reduction in lip-rounding is what the author had in mind. P. 10: the Malayalam vowel transcribed here as (U) is probably not high BACK, but high CENTRAL unrounded, differing in this respect from the corresponding Tamil sound. The statement that it is in "free variation" with [u] when medial, but in contrast when final, is an oversimplification, at least for some dialects.10 Many examples throughout the book seem to be misprinted with (u) instead of [U]. Pp. 12 ff.: SK lists a set of phones (PTCK), called "weakly voiced fricativised stops", occurring only intervocalically, and occasionally after y and I, in both Dravidian and loan vocabulary. On p. 49 they are described as allophones of ptck. But fully voiced plosives [b dj g) also exist, said to occur medially after nasals in Dravidian vocabulary (tumbi 'beetle'), initially and medially in loans (baabu 'a name'); SK con * For details see W. 'Bright, An Outline of Colloquial Kannada (Poona, 1958), pp. 5 and 73-74. * Cf. Sreedhar, p. 110. 10 Cf. W. Bright, "The enunciative vowel", IJDL 1 (1972), pp. 37-39; and S. Velayudhan, "Discussions on Bright's enunciative vowel", IJLD 2 (1973), pp. 50-53.
