Yau," has as a phonetic usually the value kiau or hiau. But h and k come from g, and Yau being in the lower series of tone classes, should also come from g, to complete the natural connexion of the sounds. So 472ayo "mountain," is in the dictionaries ngak. But this is no other than the root used in the name of the Emperor Yau. It is also identical with kau, 1, kok, “high,” and several other words with like meaning, and sprung from the same root, which may be called gok. So also yen, 5, ngem, 1035, "a precipitous and abrupt cliff," is the same with 815 hiem.

That the direction of change has been from g to ng, and not from ng to g, may be supported by the fact that in the language at present ng is a new initial in many words, e.g. inngan or an, "rest," we have a word which, when Roman visitors went to China in the reign of Marcus Aurelius Antoninus, was selected to spell the first syllable of that monarch's name. Hence ng was not an initial of the word in the latter half of the second century of our era.

Many examples occur of k and ng meeting as initials in the same phonetic; e.g. kwei "demon," with R. hill, feu, is ngwei “lofty.”

Final k has been extensively dropped in consequence of a law by which all the final surds, k, t, and p, have disappeared.

When the mediaeval dictionaries give two sounds to characters of which one is in the fourth tone class (ju) and the other in the third (c'hü), this means that the dropping of k, t, or p final was taking place at the time. In A.D. 500, for example, tak "to weave" was pronounced tik (Ky) and ti, the former in the ju tone class and the latter in the c'hü.

T'i “to end,” “destroy," "wait," is in Kwy, Tsy, Yh, pronounced t'i in c'hü sheng. But in Tsy, Yh, Chy, it is also pronounced t'iet "loosen," "relaxed." Final t then was being lost during the period A.D. 600 to 1200.

#799, composed of R. spear, kwo, and yin "sound," "adhesive earth," is in Kwy chik (tik) and in Tsy chik, shi 3, and chi 3. Here we see the k losing its ground in the time of Tsy, while in the earlier period of Kwy it was strongly rooted in the language. During the same time ch (ti) became sh or ch'.

Final k sometimes changed to t before being lost altogether. Thus in Kwy tsio "a bird," 626a, is tsak and dzit. The final t is here a transitional sound destined shortly to disappear. In P. 974

tsie "moderation," "joint," we have final t well established, and in 575 tsi "immediately," final k and t are in Kwy much mixed. Final k is found by Tyt in the Odes as the final of this phonetic, and is the one final of both. In hiue "blood" 281, hit and hik both occur in Kwy. Probably where k occurs, as with R. water, "canal,” R. blood, is used ideographically in the sense of reticulated.

At present in Amoy and Tie chiu lik "strength," is also called lat, a modern instance of an ancient law.

In some phonetics commencing in m, where k does not occur as a final in the dictionaries or dialects, it may be shown by the meanings to have formerly existed. Thus mi "rice," "deceive," "bewilder," has k, because (1), 627 mik has the sense "dark;" (2), 546 meu, mok, has that of "stupid," "dark," "dim vision;" (3), 788 mok means "cover with the hands," "stupid," "a veil," "evening."

So also mau, "hair" with yü "feathers," has the sound mok in the dictionaries, and interchanges with 546. In the case of 597 mau, the final k is found in Kwy alone and with R. boat, and kia "a coat of mail." In 589miau "young grain," a picture of a field with grass growing above, the k is likely to have existed by the meaning miau, to draw a copy of pictures or characters, which probably means originally "make black strokes," and is the same word as me "ink," mei "coal," both of which have final k. Also miau has the sense "seek," which is found in phonetic 773a mek.

The phonetic jo, nok, "if," is found in Kwy twice with the sound nit, viz. with RR. sun, heart. This helps to show that the course of change was from k to t.

In the case of phonetics beginning with ki, tsi, si, j, examples are more numerous than with others. Thus sie, 4, sit, "fragment," is so spelt in Kwy. The phonetic J、 siau, sok, “small," would lead the incautious to believe that t could not be the final; but wrongly so, for Sw has pa instead of siau. The latter of these came into use because its sense suited the meaning of sie.

F

In the Tie chiu dialect we find niet, pronounced ngiak, as also with RR. wood and rice in place of son. In the same dialect niet is ngiak. Probably here k is old, and on this the dictionaries give us some light. The meanings also render final k the more probable ancient owner of the position of final letter in these words. Thus ngai "to stumble," was ngak. The stump of a tree left after cutting

down the trunk was written with wood underneath instead of son, and is called in Kh nget and ngak. The last is in a poem of Su tung p'o, A.D. 1000.

The chronology of letter changes of this sort may be determined in the following manner. When in the classical period we find a word for "thou," written in the Yi li by jok, and in the Shu by R. water, P. ju, the respective phonetics of these words being in more than one particular closely related, we conclude that in the Cheu dynasty, at the time when the Yi li was written, the final k of this word for "thou" was not yet lost. Else why should the writer use for it a character which kept its k down to the T'ang dynasty?

The finals ng and n are confused at present in western and southern Mandarin, and partially in the Old Middle Dialect. At Nanking, e.g. lan and lang are sounded alike.

Tai tung was a native of Si chwen. He says in explaining the Fan t'sie or syllabic spelling, that joiningshi "a model," and

shang "merchant" is spelled by kan "a pole." Being himself a

speaker of western Mandarin as it existed 700 years ago, he did not notice that an and ang are not one sound.

Final ng has been lost from a few words. T ta "to beat," with R. hand, is called tang in Shanghai. The dialect there has preserved the old sound. Such examples are rare.

4. PALATAL LETTERS.

Among the palatal letters is j. It comes from ni and is modern Chinese. At Sucheu and the cities lying west of it towards Nanking there is to some words both a reading and a colloquial sound. Among the words having in the reading sound are jen "man," jen "patience," je "hot." In the colloquial dialect these words are sounded with ni. In transcriptions from Sanscrit, such words represent Sanscrit syllables beginning with ni. This change has taken place within the last thousand years.

Among the 36 initials j ranks as the last. It is vowel ï, as in rï 2 “ear,” rï 5 “and,” “son," ri 3 "two."

before the

That is to

say, the old ni has become separated into two initials j and r in the modern language.

Ch comes regularly from t in Chinese and neighbouring languages. Thus chi "to know" is ti in Amoy. Since ch was found by the Hindoo authors of the syllabic spelling already existing in the language, but not so widely spread as now, this is a change that took place within a time reaching back to at least the beginning of the Christian era. Thus then the Chinese ch has a dental origin. The tongue has slipped back from the upper teeth to the palate.

Since the whole of the words commencing with t did not together make this change, but only a part of them, there was some law of limitation. That law was probably based on the nature of the following vowel. At present when I becomes ch, it is necessary that the vowels i orü follow. The same necessity may have existed when some of the words beginning with t took ch instead of it, while others kept t unaltered. In Amoy cho "to take fire," is tio, and c'heng "a city," is siang. If, however, it was necessary that i should follow t in order that t might become ch, it was not a necessity ruling in all words, for t'in "heaven" still kept its t unaltered. Yet this word may very well have been t'en and so escaped change.

The old initial d became ch in the same way. Yet there was a difference. D first changed to dj, and afterwards, when the sonants all became surds, was altered to ch or c'h. Thus ch'wen "boat," once dun, assumed the form jön in the Sucheu dialect and the region west of that city. The next step, as shown in the case of the Nanking dialect, will be to change j to c'h.

C'h aspirated occurs in words which belong to the fifth tone class, or hia ping. Ch occurs in words belonging to the three tone classes known as hia p'ing, hia c'hü, hia ju.

There are words beginning with ni which do not in Mandarin change that initial for j. These are such as niang "woman," "wife," "mother," "lady," nü "woman,' daughter."

99.66

Kwo pu tells us that in his time, A.D. 350, wives of brothers in Kwan si (Shan si) addressed one another as dok lik. The characters here are. The orthography cheu li, now expresses the sound of the words that he had in view. In spelling the sound of cheu by the syllabic method, he used dok. Hence d was the initial and

k the final in his time.

1 These words are placed by Kh under the ch series and not under j. Kh's reason is that they never become j ar r.

5. TOOTH LETTERS.

The old t, t', s, ts, sh, t's, n, in the case of many words remain as they were. The old d, z, zh, have become t, t', ts and t's, sh and ch or c'h. Many words having formerly t and ť have taken ch or c'h instead.

In Chinese, as in other languages, s has quite commonly sprung from t. Thus in Cochin Chinese and dialects further south, sim "heart" is tim. When Cochin China was made a province in the Han dynasty, if the initial t was common in China, its appearance in the Cochin Chinese vocabulary is accounted for. That it was so is to be shown by examples which throw light on the law of sibilization.

In the south Fu kien and eastern Canton dialects we have the same phenomenon. This is not by change from ch to t, which is unprecedented, but by change from t to ch in Mandarin and in those parts of China where ch exists. In Tie chiu, chung "middle" is tong, c'ha "tea" is te, chui "to follow" is tui, chi "straight" is tit for dik.

The title of the Emperor of the Hiung nu, corresponding to the Chinese "Son of heaven," is given in Panku's History of the Han Dynasty. The Turkish word for "son" is there said to be ko do. This is probably the same as ugli, the word now used. The Turks of the time would say for "son of heaven," T'ingri gudu or ugudu. Since that time d has changed to in Turkish. Panku's word ko do is now read ku t'u by the speaker of Mandarin.

In regard to the epoch when t became s and ch, we may conclude that it was in the case of very many words a little after the Han dynasty. The Cochin Chinese transcription was made in that dynasty. Later by a few centuries we have the early tonic dictionaries, which contain a very large number of examples of words int and d now pronounced with s or ch. The change was spread over several centuries, and appears to have been completed in the T'ang dynasty a thousand years ago. We must except the province of Fu kien, where the change here referred to is not yet concluded.

D has in many words changed to sh, through the medium of dj. There is no room for doubt that this change has taken place extensively. According to the spelling of shi with R. metal in the dictionaries, we have the sounds ti in Kwy and Tsy, a sharp edge," di in Tsy, "an instrument for pricking blood," dji in Tsy, “key.”

66