² Karlgren, B., ‘Word families in Chinese’, BMFEA, No. 5, 1933.Google Scholar

³ Other work in this field includes: H. Maspero, ‘Práfixes et dárivation en chinois archaiique’, MSLP, XXIII, 1935; Yu Miin , ‘Word derivation in Archaic Chinese through the annexing of the suffix ‘d’, YCJ, No. 34, 1948, 29–8; B. Karlgren, ‘Cognate words in the Chinese phonetic series’, BMFEA, No. 28, 1956.

⁴ Karlgren, The Chinese language, New York, 1949, 96, 97. Maspero too came to this conclusion (op. cit., 327), and so, later, did Yu Miin. See Shiannday hannyeu yeufaa Pt. 1, Peking, 1954, p. 32, by Luh Tzongdar and Yu Miin.

⁵ Unless otherwise stated, throughout this paper words are given in Karlgren's Ancient Chinese transcription, as found in his ‘Grammata Serica recensa’, BMFEA, No. 29, 1957. Tone is indicated by a small circle placed at the corner of the reconstructed form thus: No sign is needed for ruhsheng words, which are always characterized by a final stop consonant.

¹ Jou Tzuumu , ‘Syhsheng byeyih shyhlih’ .Oriinally in Fuuren Shyuejuh , XIII, 1–2, 1945. I have seen only the reprint in his Hanyeu inyunn luennwenjyi , Shanghai, 1957. Jou Fahgau ‘Notes on Chinese grammar’, CYYY (Taiwan), No. 24 1953, 197–212.

² Jou Fahgau (op. cit., 211), who treats the ruh/chiuh and pyng, shaang/chiuh contrasts separately, explains the appearance of chiuhsheng in both contrasts by assuming that in Archaic Chinese chiuh and ruh differed in final but were alike in tone, whereas chiuh, pyng, and shaang differed in tone but were alike in final; thus, pyng, shaang, and ruh all tended to ‘interchange’ with the chiuhsheng. This does not explain why pyng and shaang do not interchange in the same way.

³ Wang Lih , Hannyeu shyygao, Peking, 1958, pp. 213 ff.

⁴ Maspero, op. cit., p. 327, n. 1.

⁵ A. Haudricourt, ‘Comment reconstruire le chinois archaique”, Word, x 2–3, 1954, 364.

⁶ e.g. Jou Tzuumu, op. cit., p. 54, under yeu ; theshaangsheng member is taken as ‘basie pronunciation” the chiuh as ‘modified pronunciation’ .

⁷ Not all his readings were used, only those in the Liijih with Jenq Shyuan's commentary, and in the Chuenchiou and Tzuoojuann , with the commentary of Duh Yuh .

¹ Examples may be seen conveniently in Wang Lih's lists (op. cit., pp. 213 ff.).

¹ The marking of non-chiuh members as regular is probably connected with the fact that the chiuh readings generally were becoming obsolescent in Luh's time. See below, p. 266.

² An exception, in another context, seems to be the occurrence of tones in final particles. See Kennedy, ‘A re-examination of the Classical pronoun forms ngu and ngo”, CYYY, XXVIII, 1, 276, 277.

¹ See The Chinese language, 96. A good example is to be found in Modern Japanese, where verbs ending in -eru are intransitive/passive when the simple verb is transitive (e.g. miru-mieru, toru-toreru, yaku-yakeru, kiru-kireru, etc.), and transitive/causative when the simple form is intransitive (e.g. komu-komeru, tatu-tateru, aku-akeru, iru-ireru, etc.); thus analogous to Groups C and F of chiuhsheng derivation. (See the lists on pp. 281–3 and 287–8.)

² See Group H, and below, p. 268.

³ op. cit., 364. Maspero, op. cit., 326, has suggested the influence of a prefix to account for the change in tone. This seems less likely.

⁴ See below, pp. 269–70.

⁵ This is especially likely in the last four examples on p. 270, which belong to the anomalous chiuhsheng rimes in the Chiehyunn. There is abundant evidence (from the poets' use of rimes) that the words of these rimes retained their final stop right down to the Six Dynasties. See Wang Lih, Nanbeeichaur shyren yonqyunn kao pp. 49–53, in his collection Hannyeushyy luennwen jyi Peking, 1958.

¹ The last two, according to shyesheng, etc, had final stop consonant in Archaic Chinese. It is necessary to posit some difference from other syllables with final stop; one way is simply to project the tone back into Archaic Chinese. See also below, p. 266, n. 2.

² For longer lists, showing examples from all tones, see Karlgren, ‘Cognate words in the Chinese phonetic series’, BMFEA, No. 28, 1956, 9.

³ For the relationships between these finals, see below, p. 265.

⁴ Compare the simple chiuhsheng derivative piuan° ‘to give relief’ (D.8, p. 284).

¹ For references to Guh and Chyan see Jou Tzuumu, op. cit., 51, 52, and Jou Fahgau, op. cit., 197. Duann Yuhtsair's attitude is clear from his remarks in his Shuowen jieetzyh juh under the characters , etc.

² Yanshyh jiashiunn

³ Introduction to the Chiehyunn.

⁴ Introduction to the Shyyjih jenqyih

⁵ e.g. C.2.

⁶ See Wang Lih, Hannyen shyliuhshyue , Shanghai, 1958, pp. 133–42.

⁷ All examples may be found in Luh Jyhwoei Beeijinghuah dan'intsyr tsyrhuey , Peking, 1956.

¹ Jou Tzuumu, op. cit., 52 ff.

² Jou Tzuumu, on this evidence, concluded that tne derivation process began in Hann times (op. cit., 52). Jou Fahgau (op. cit., 209) has already pointed out the non sequitur involved.

³ The Archaic and Ancient readings are taken from Grammata Serica. In the few cases where Gram. Ser. does not include the derived form, the chiuhsheng pronunciation may be seen in homonyms.

⁴ Gram. Ser., No. 784, notes that several characters in this series are irregular in Ancient Chinese. Here the derived form is regular, the basic form is irregular.

⁵ There seems to be no comparable form from which to adduce the Archaic reading.

⁶ This reading, also found in the Goangyunn , seemsvery aberrant in the Chiehyunn system.

¹ Karlgren reconstructs an Archaic labial semivowel in the chiuh form of A.20 and D.6. This is also found in a few other similar pairs (not included because of the irregular initials):

In all these cases it is possible to take the labial semivowel in the form ending in -d as the last trace of an original final labial stop, and thus not reconstruct the semivowel in the original form, i.e. nab > nwad/nuai.

However, the labial semivowel does not survive as a trace of the original labial stop in all cases, as A.32 shows. Whether its occurrence is random or whether it can be determined from other data remains to be seen. (Note the occasional appearance in Mandarin of a herkoou vowel in a few words, as the only trace of an Ancient final labial consonant, e.g. shymn, liun, ruh.)

² The writer realizes that the ascription of chiuhsheng derivation to Archaic times raises the difficulty that in the cases listed above the contrast, according to Karlgren's reconstruction, is not between tones but between voiced and voiceless final stops. This would not affect the argument for a system of derivation, but would add to the complexity of the phonological description of it. However, the writer believes that even in Archaic times these words may be better explained as cases of tonal contrast, but reserves discussion of this problem to a later article.

³ See e.g. the lists in Karlgren, The Chinese language. Another indication of the relative lateness of chiuhsheng derivation is found in the use of the same character for both simple and derived forms, where other pairs of cognate words are usually written with different characters.

¹ See above, p. 264.

² The fullest lists may be found in the Chyunjing inbiann by Jea Changchaur (finished by Baoyuan 2 , A.D. 1039), really a collection of characters with two or more readings, taken from the Jingdean shyhwen. Other lists of these words, from the Sonq and Yuan periods, are to be found in the Iayunn shyhyi compiled by Ouyang Derlong revised by Gwo Jenqjii in jeatzyy year of Jiingdinq (A.D. 1264). See especially the Introduction. Also in the Chueijiann luh , by Yuwen Baw , Peking, 1958, p. 75.

³ e.g. A.9, A.2, and A.51 (1)

⁴ e.g. A.10, A.15, A.42.

⁵ e.g. A.7, A.31, A.29, and A.50.

⁶ e.g. A.19, A.47, and A.64.

⁷ C(huenchiou Tzuoojuann) Ai 1, comm., j.57, 1b (1180).

¹ C.2, C.5, C.6, C.10, and C.13.

² ef. Latin jaciō/jactō, volō/volitō, etc.

¹ See above, p. 266.

² i.e. those marked ‘alternative reading’ or ‘Shyu Moh’ in the lists.

³ e.g. .

⁴ ‘They might combine their functions’, L, Neytzer, comm., j.8.14a (735).

⁵ Jou Tzuumu, in another context, takes the ruhsheng reading (given by Shyu Moh) as an artificial reading to agree with a rime. In the Chuenchiou reference, however, there are no riming words involved, so the reading is probably legitimate. See his Tarngbeen Maushy' in juannren kao (p. 1 in his Hannyeu inyunn luennwenjyi).

¹ . C.22.

² Hwainantzyy, j.11, p. 13b (SBBY) edition).

¹ E.1 and E.9. The Chyunjing inbiann (see above, p. 267, n. 2) has many more examples.

² Japanese is an example.

³ e.g. C, Jau 27, j.52.10a (1154), (also commentary), where both words are used in their basie tone as ‘handful’.

⁴ e.g. C, Jau 16, j.47.11b (1126), where it means ‘several’. Here modern usage differs from Luh's.

⁵ Many of his readings are very puzzling. The character , for example, is often given the reading ngák when it would seem that the meaning is clearly ‘joy’, not ‘music’.

⁶ The references are to be read as follows:

L—Liijih Jenqjuh (SBBY edition), followed by name of the chapter, jiuann and page numbers.

C—Chuenchiou Tzuoojuann jenqyih (SBBY ed.), followed by reigning duke and year, jiuann and page.

Comm.—in the commentary.

Jing—in the Chuenchiou, not the Tzuoojuann.

The figure in brackets refers to the page number of the Jingdean shyhwen (TSJC ed.).

Other references are given in full.

Where the usage seems regular, I have generally given only one reference. In more problematical cases, two or three references may be given. When both members of a pair lack references, it may be taken that the contrast is commonly found in all texts, i.e. it is part of general Chinese.

‘Alternativ’ means that Luh gives both basic and derived form. ‘Shyu Moh’ indicates a reading of Shyu Moh's quoted by Luh.

The characters in each group are arranged in order of the traditional 36 tzyhmuu .

¹ See referenees in Jou Tzuumu, ‘Syhsheng byeyih shyhlih’, p. 52.

¹ In the Wuchaursheaushuo dahguan Bk. 336 (Shanghai, 1926).

² See Jou Tzuumu, op. cit., p. 65.

¹ See also Koong Yiingdar's comment on this passage (same reference).

¹ Grammaia Serica gives miuk for this character, following the Chiehyunn. Luh Derming in the above reference says it is pronounced like , which seems to agree with the Chiehyunn. However, elsewhere (e.g. L, Wangjyh, j.4.3b (684)) he says it is pronounced like muk. This would agree better with the derived word.