^* Presidential address, read at the regional meeting of the Linguistic Society of America in New York, December 31, 1945.

¹ Z. S. Harris, Simultaneous components in phonology, Lang. 20.181-205 (1944).

² Sometimes, an ancient Chinese morpheme with an initial consonant cluster became a dissyllabic morpheme through the appearance of an extra vowel between the parts of the cluster. But the separate syllables thus resulting usually acquired the status of separate morphemes, with separate meanings. See P. A. Boodberg, Some proleptical remarks on the evolution of Archaic Chinese, Harvard Journal of Asiatic Studies 2.329-72, esp. 336 ff.

³ Such as Sinologists, missionaries, and Chinese students studying abroad.

⁴ The learned term for it is tz'ŭ² (not related etymologically to tzŭ⁴).

⁵ B. Russell, Our knowledge of the external world as a field for scientific method in philosophy 89, 107 (London, 1914).

⁶ B. Russell, Analysis of mind 189 (London, 1921).

⁷ This manner of expression can be compared with the situation when a person uttering or hearing an allophone is said to be uttering or hearing a phoneme. When a person utters or hears a word, he is really uttering or hearing one of its allologs, so to speak.

⁸ This illustration was given by I. A. Richards in connection with another subject.

⁹ See Homer Dudley, R. R. Riesz, and S. S. A. Watkins, A synthetic speaker, Journal of the Franklin Institute 227.739-64 (1939).

¹⁰ The emphasis on distributional characteristics of speech has lately tended to displace Bloomfield's idea of ‘significant features’. Some recent experiments, however, have shown a striking correspondence between phonetic quality and rather prominent acoustic features. See R. K. Potter, Visible patterns of sound, Science 102.463-70 (1945), reprinted as Monograph B-1368 of the Bell Telephone System Technical Publications.

¹¹ Because the low tone of a high-pitched person (or even of a higher-pitched key of the same person in a different mood) may be physically identical with the high tone of a low-pitched person (or key), the assignment of membership of pitch movements to tones cannot be treated on a simple quality level. It seems that an approach from the gestalt point of view will fit the case very well, but it is probably also possible to define the classes of qualities in such a way as to allow for shifting of the scale and overlapping of membership.

¹² According to L. Bloomfield in speaking of Andrade's unpublished studies.

¹³ C.W. Luh of Yenching University in 1940 compiled (in MS) a list of 5000 monosyllabic morphemes in the Peiping dialect. Since the dialect has only about 1300 different syllables (not counting syllables with retroflex endings), there are on the average four homonyms to each syllable. It is unlikely, however, that another linguist recording the same dialect independently would duplicate this list.

¹⁴ It may be noted in passing that a play on words based on different extensions of meaning of what is considered the same word is usually more effective than a pure pun based on what are regarded as unrelated homonyms.

¹⁵ A study of etymons on a large scale is B. Karlgren, Word families in Chinese, Bull, of the Museum of Far Eastern Antiquities, No. 5 (1933).

¹⁶ The most important exception is the very active non-syllabic retroflex diminutive suffix. But even here, those speakers of the Peiping dialect who are of Manchu origin tend to pronounce the retroflex suffix as a separate syllable. In Hangchow, the corresponding suffix is a syllabic l.

¹⁷ For example in B. Karlgren, Analytic dictionary of Chinese and Sino-Japanese; Paris, 1923.

¹⁸ Because of this latitude of variation, Lin Yutang is able in designing his typewriter to standardize the left-hand and the right-hand parts of a vertically divisible characte in the ration of 2:3; yet no one will notice this unless he has his attention drawn to it.

¹⁹ Modern cases are limited to a few characters for dialect words, like the one that represents words for ‘fling’ (Mandarin shuai³, Wu dialects hwεⁿ').

²⁰ Mencius 7-B.16.

²¹ According to Jones, the word probably appeared first in H. E. Palmer, Principles of romanizat ion 60, 61 (Tokyo, 1931). By analogy, we should have to call a general word in several dialects a dialog.

²² Letters or words in pointed braces indicate diaphones.

²³ Palmer, loc.cit., attributes this term also to Daniel Jones. Jones, however, denies it in a recent letter to the writer, and thinks that it may have been invented by N. W. Thomas, an anthropologist and linguistic investigator in Africa.