¹ Cf. Derk Bodde, JAOS 62.8, fn. 1 (1942). A reference to the transcription used by H. H. Dubs, JAOS 61.215–21 (1941).

² Denzel Carr, Japanese Romanization Again, JAOS 61.188–90 (1941).

³ For the results contained in this paper I am indebted to Yuen Ren Chao and George L. Trager for their help and suggestions during many hours of discussion. Without their advice on many points of theory and detail this paper could never have appeared. The material used is taken largely from phonetic descriptions of the Peiping dialect made by Professor Chao. I wish to thank also Francis W. Cleaves for his help in collecting examples of syllables ending in the phoneme /r/. I am also indebted to Bernard Bloch for numerous suggestions about terminology and other details in the preparation of the typescript.

⁴ J. Marshman, Elements of Chinese Grammar 87 (Serampore, 1814).

⁵ Henri Frei, Monosyllabisme et polysyllabisme dans les emprunts linguistiques, avec un inventaire des phonèmes de Pékin et de Tokio, Bulletin de la Maison Franco-Japonaise 8.1.130 ff. (1936).

⁶ Tchen Ting-ming, Les phonèmes de la langue chinoise, Bulletin de la Société de Linguistique de Paris 40.109 (1939).

⁷ B. Bloch and G. L. Trager, Outline of Linguistic Analysis 40 (Baltimore, 1942).

⁸ These statements of stress distribution are true in general for minimal free forms occurring in isolation. They are approximations rather than complete descriptions. There is in addition a contrastive stress, while C. F. Hockett reports a fourth stress-level between loud and secondary. The whole subject needs a thorough investigation.

⁹ The phonetic transcription used here follows as closely as typographical circumstances will permit the system presented by Bloch and Trager, OLA, Chap. 2. Affrication is denoted by two symbols in juxtaposition, e.g. [ts] [tś], the position of articulation being indicated by a diacritic on the second element. Nasalization is indicated by the subscript [_n], retroflexion by the subscript [_r]. The symbols [z·] and [ẓ·] are admittedly makeshifts; they denote syllabic, frictionless, voiced spirants, homorganic with the preceding consonants, and having mid-central vowel timbre. The vowel denoted by [ӡ] is lower-mid, central, unrounded.

¹⁰ The vowels [a] and [α] are respectively not as far front and back as the cardinal positions for these vowels. In actual speech the vowel [ë] tends to shift during the utterance through the whole mid vowel range, while remaining back unrounded, and this symbol serves merely to define a norm. The vowels in [juw] and [wij] are centered slightly and somewhat lower than the positions indicated. The vowels in [in], [yn], and [iŋ] are slightly lower than the cardinal position. In addition to these finals there also occurs a final [m]. It is rare in absolute final position, one example, however, being [t’ȧm], a contraction of the more usual [t’ȧ-mən] (first tone on first syllable, second syllable weak) ‘they, them’. Otherwise it occurs internally and remains a problem to be investigated, its solution being related to the matter of juncture.

¹¹ B. Karlgren, Études sur la phonologie chinoise 295–97 (Leiden and Stockholm, 1915): a thorough discussion of these two finals and a summary of attempts to transcribe them.

¹² The numbers refer to the class headings in B. Karlgren, Analytic dictionary of Chinese and Sino-Japanese (Paris, 1923).

¹³ Adapted from a table prepared by Professor Chao.

¹⁴ Although it has never been reported heretofore, an additional final cluster /-rw/ should be mentioned for the sake of completeness. I have heard it in the pronunciation of three native Peiping speakers. It was lacking in the speech of one informant of Manchu derivation. With the three informants who distinguished it from /-wr/ it occurred after both /a/ and /e/ in syllables beginning with both the aspirated and unaspirated velar and alveolar stops. No minimal contrasts were found. An example is [tə_rw] ‘a small pocket’. The final [w] is quite lax. Information on this point is by no means complete.

¹⁵ No allophone is listed for /ŋ/ in initial position, although with many speakers [γ] may occur initially in syllables containing /a/ or /e/. It is in free variation with no initial consonant and is thus analogous in this respect to the initial glottal stop in English. In Tientsin [ŋ] is heard initially in this same role. It is concluded here that this onset of the vowel is not significant and it is hence excluded.

¹⁶ The principal sound features of /r/ in final position are voice and retroflexion. This and the positional complementation are sufficient to permit the inclusion of initial [ẓ] as an allophone of the /r/ phoneme. Note that [ẓ] is usually modified by slight lateralization.

¹⁷ The result that the palatals and alveolars are allophones of the same phonemes was implicitly stated by Morris Swadesh in A Short Account of Mandarin Phonetics, Travaux du Cercle Linguistique de Prague 8.213–16 (1939), where, however, it appears to be assumed rather than proved. The length of his paper permitted only the listing of conclusions. It is guilty, however, of a number of inaccuracies, considered at least from the standpoint of Peiping phonetics.

The suggestion that the cacuminals might be analyzed as clusters with /r/ as the second member was first made by Trager in 1941.

¹⁸ This complementary distribution may be illusory. I have heard the pronunciation /wîŋ/ used several times by a Peiping informant for a surname. This surname has been reported universally to be /wêŋ/. Whether this syllable will be found to have more general occurrence or not, it serves nevertheless to illuminate further the phonemic difference between these two finals.

¹⁹ A practical system of transcription, one which has been based upon this analysis and which has been put into actual use in class, involves the one-to-one substitution of unit symbols for the symbols based upon the phonetic analysis and presented above. Thus /p, t, k, ts/ are written b, d, g, z respectively, while /p', t', k', ts'/are written p, t, k, c. This of course does not affect the system and serves merely to introduce a neater orthography.