^{page 320 note 1} I am indebted to one of my colleagues for pointing out to me that my use of the term ‘ exponent ’ in this article is not always consistent with accepted phonological usage. The regular use is exemplified in the stating of the phonematic units, and the irregular in the stating of the prosodic systems. In these latter what are in this article termed ‘ exponents ’ are not in fact the sum total of statable exponents but only such of them as are phonetically differentiated in the context in question. The exponents of a given term are not therefore stated in those contexts in which they are phonetically identical with those of other terms in the system.

^{page 320 note 2} The syllable-structure of the trisyllabic verbal phrase (Verb + Particle 1 + Particle 2) may be stated as:—

Initial syllable (verb): CVC

Medial „ (Particle 1): C₁V₁

Final „ ( „ 2): C₂V₂

This statement differs in one respect from that of Part I; it has been found preferable to treat the structure of the final syllable as C₂V₂ and not C₂V₂C₂.

^{page 321 note 1} In ‘ close words ’ (cW) the vowel of the final syllable (-V₂) may be characterized by short, medium, or long duration. These three degrees of duration have been stated in Part I for the sentence (S) as exponents of two-term prosodic systems of intonation, hence termed ‘ intonation one ’ (1S) and ‘ intonation two ’ (2S):—

(a) Sentence of one word: 1S.: medium duration of vowel

2S: long duration of vowel.

(b) Sentence of more than one word: 1S: medium duration of vowel.

2S: short duration of vowel.

Since it is assumed that both of these intonation systems are valid for all the cW examples given below, strictly speaking the three degrees of duration should be symbolized for every example. In practice except where all three degrees are explicitly symbolized in the stating of the vowel (-V₂) of the final syllable, all cW examples have been symbolized for convenience as with (I), with the implication that all three degrees are possible in the appropriate prosodic context.

^{page 323 note 1} The symbols ‘ fS ’ and ‘ sS ’ refer to a two-term prosodic system of ‘ rate of utterance ’ stated in Part I for the sentence (S), hence ‘ fast sentence ’ (fS) and ‘ slow sentence ’ (sS). Examples have been designated fS or sS according as their phonetic form is appropriate to the ‘ fast ’ sentence or the ‘ slow ’. Any general-phonetic example not thus designated is considered to be valid for either. By means of this prosodic statement in terms of ‘ rate of utterance ’ it has been possible to make a single statement of syllable structure for two or more diverse phonetic forms, e.g.

^{page 323 note 2} It has not been found necessary to state this general-phonetic difference of duration in terms of a prosodic system of quantity. It suffices to state both the two degrees of duration and correlated features of aperture, as exponents of the c/o prosodic system of the word.

^{page 330 note 1} A phonemic account of the phonetic feature, frontness and backness of vowel, treated here in terms of the prosodic system of the syllable final y/w appears in ‘ The Phonemes of Tibetan (U-Tsang Dialect) with a Practical Romanized Orthography for Tibetan-speaking Readers ’ (Rev. Miller, P. M., Journal of the Asiatic Society, Letters, vol. 17, 1951, no. 3, pp. 191–216).Google Scholar This account of a Tibetan dialect similar to LT (the Rev. Mr. Miller's informant was from Shigatse) deals with this phonetic problem by assigning the frontness feature to three of the 27 contoids: ‘ … the dental phonemes, l and n, occurring in syllable-final position modify the preceding vowel in the same manner.… 1 may be voiced or ‘ zero ’ leaving the modified vowel of often increased length as the only evidence of its existence ’ (p. 198); ‘ t ’—a voiceless aspirated stop formed also at the dental point of articulation by the tip of the tongue [t'].… This phoneme in syllable-final position occurs only as a modification of the preceding vowel plus (in sentence final and often phrase final position) a glottal closure ' (p. 193).

There is much in this article that is novel and of interest both in the Tibetan material and in the type of phonemic statement, but the treatment of the feature in question seems unnecessarily tortuous.

^{page 331 note 1} In § 1 B i above long and short duration of vowel were stated not in terms of a prosodic system of quantity but as exponents of c/o. Similarly here it has not been found necessary to state this duration feature in terms of quantity, or, for that matter, of the c/o prosodic system but, with correlated features of backness and frontness of vowel, as exponents of the two-term F system, y/w.

^{page 334 note 1} In addition to being the initial-consonant system that takes into account the prosodie systems of the word h/ⁿh, v/ⁿv, the Ca- system comprises C- terms that have as their common exponent occlusion (vyT- has friction as its exponent in interverbal junction and occlusion + fricative release (dz) in intraverbal junction; there are no examples of vyT- in intraverbal junction in the trisyllabic verbal phrase). The deciding factor for the grouping of theco-articulations of this occlusion adopted here, a grouping that results in the Ca- system stated above, is the part played by the articulator. It is the dorsal area of the tongue that is concerned in the velar and palatal plosion stated as exponents of K-; the apical area is concerned in the alveolar affrication and dental plosion of T; it is, of course, impossible to establish palatographically whether the labial-plosion exponent of P- shares with its other exponent palato alveolar affrication an articulation with tongue-tip down behind the lower teeth. If, however, it is legitimate to cite evidence from other Tibetan dialects then the fact that has been recorded in Sikkimese for LT wәⁿvyPVT () is relevant, cf. also for LT (ⁿvyPETw), Sikkimese on the assumption that and are comparable. The alveolar affricate is not assigned to P-, T-, or K-, but, as rCa-, stands in the same relation to each of them.

P T K.

Ca-

It is possible that in some dialects a three-term prosodic system would need to be set up, for the syllable initial, i.e. r/y/w to account for the forms given in Jaeschke, (Tibetan-English Dictionary, London, 1934), pp. 18, 19,Google Scholar ‘ K'rug, gri, ‘pru-gu ’, etc. This would then give r/y/w P-/T-/K.-.

^{page 336 note 1} The N of yI is not, of course, to be identified with the N of the three-term system M/N/D of wI; it stands in the same relation to all three terms, as rCa stands in relation to P/T/K.

^{page 339 note 1} Lip-spreading characterizes -V- in yәW in certain contexts, and is perhaps better stated as an exponent of non w/әW, cf. әW § I, below.

^{page 347 note 1} For ‘ prehend ’ see Firth, J. R., The Tongues of Men, pp. 126–7.Google Scholar

^{page 347 note 2} In wW labial nasality was stated as an exponent of M-, one of a three-term system (M/N/D); here in yW it is stated altogether differently, as indeed it must be, where the co-articulations of nasality are not velarity, labiality, palatality, and dentality, but labiality and palatality alone. cf. also әW.