from S. with the suffix ro, ri (ro, ri), we have, as non-diminutives, tukiro 'a piece,' ćhokarī ‘a girl,' etc., or, as diminutives, jinduro 'a short life,' dhiari 'a small daughter,' etc.; in It., with olo, as non-diminutives, lenzuolo, figluolo, etc., or, as diminutives, bagnuolo, sassuolo, etc. Such suffixes are thus used in all the languages of both groups. For its extensive use of diminutive suffixes S. is most conspicuous in the Gaurian group, as It. is in the Romance; nearly every substantive and adjective in these languages being capable of receiving a diminutive suffix.

(To be continued.)

317

ART. XIII.-A Comparative Study of the Japanese and Korean Languages. By W. G. ASTON, Assistant Japanese Secretary H.B.M.'s Legation, Yedo.

IN comparing Japanese and Korean with each other and with other languages, there are three things to be considered: -1st, their phonetic systems; 2nd, the functions of their grammar; and 3rd, the character of their grammatical procedures.

The comparative poverty of the Japanese vowel system is incompatible with the action of the principle of the harmony of vowels, which is accordingly almost unknown in this language. Korean is much richer in vowel sounds, but in so far as I have been able to observe, it agrees with Japanese in rejecting that principle. Perhaps é and è, which, although simple sounds, are represented by the letters for ůi, ai, should be excluded from the above enumeration of Korean vowels. They are no doubt of more recent development. There are indications that the Japanese e is also more recent than the other vowels.

The consonants ph, th, chh, and kh (pronounced as in Sanskrit), are wanting in Japanese. They are, however, comparatively little used in Korean, and the aspiration disappears at the end of a syllable. The difference made in some parts of Japan in pronouncing such homophonous words as kaki 'a persimmon,' and kaki 'an oyster,' is perhaps owing to an original distinction between aspirated and unaspirated consonants. One may have been kaki, the other khaki, though they are written alike, and even in speaking the difference is barely appreciable.

The consonants b, d, j, z, and g are wanting in the Korean alphabet. But this want is only apparent. To the ear, the soft consonants are unmistakably present. As in the Dravidian languages, there is no doubt some rule by which the letters for p, t, ch, s, k, are in certain circumstances pronounced b, d, j, z, g.1 1 The Japanese alphabetical character originally neglected this distinction, and even in old printed books the nigori, or mark by which the soft consonants are distinguished, is usually omitted. There is some analogy between the Tamil and Japanese rule as to this distinction.

1 "Le mandchou possède des consonnes fortes, et des consonnes demi-fortes figurées par des caractères distincts, mais dans le corps et à la fin des mots, ces dernières sont sujettes a dégénérer en douces sans que l'écriture manifeste cette dégénérescence."-Lucien Adam's Manchu Grammar, p. 13.

In Tamil a consonant is soft, except at the beginning of a word, or when doubled. In Japanese the initial consonant of the second element of a compound word is usually softened, and doubled consonants are always hard.

In classical Japanese there is no p. It is probable, however, that the Japanese language was not always without this letter, and that the present h or ƒ represents an original p. P has re-appeared in the modern colloquial Japanese. H and ƒ are written with the same letter in Japanese. It is read before u, h before any other vowel. The Korean h belongs to the guttural series of consonants. In comparing the two languages, we are therefore prepared to find that a Japanese ƒ or h corresponds to a Korean p, and a Korean h to a Japanese k or g.

Where European languages have two letters r and l, Japanese and Korean have only one. The Japanese r sometimes closely resembles the English sound. At other times, and especially before i, it approaches our d, except that in forming it, the tip of the tongue is directed more backward and upward, as in the so-called 'cerebral' sounds of Sanskrit. The Korean sound fluctuates between r and l. It has been written throughout this paper, but it corresponds very closely in character to the Japanese r. The form given to it by the inventor of the Korean alphabet shows that he considered it to be related to t, and this view is corroborated by other evidence. As in several other Turanian and Dravidian languages, r (or 1) cannot begin a word either in Japanese or Korean. Japanese tolerates it in this position in words derived from Chinese and other foreign languages, but Korean has a strong tendency to substitute an n. A vowel is occasionally prefixed for the same reason, as in the word 'Russia,' which is in Japanese Oroshiya, in Korean Arasa. In Dravidian derivatives from Sanskrit, the same means is taken to avoid an objectionable r at the beginning of a word.

In Japanese, it depends entirely on the vowel which follows whether a consonant of the Dental or Palatal series is used. A, o, e are preceded by t, d or s; i by ch, j or sh; u

VOL. XI.-[NEW SERIES.]

21

by ts, ds or s. This rule does not hold good in Korean, but a similar tendency is indicated by the fact that before i and y, it is indifferent whether t or ch, th or chh, are written. The particle chi is often written ti; the word for 'good' is either chyota or tyota.

The Korean sibilant character probably represents both s and sh. The form shows that it is considered to be allied to ch and chh, and Mr. Ross,' who has a practical acquaintance with the language, represents it sometimes by s, sometimes by sh. At the end of a syllable it is frequently, like the Japanese (tsu), only a sign that the following consonant is doubled.

The Japanese ng is not distinguished in writing from g, and indeed is only found in the Eastern dialect. It cannot begin a word. The Korean ng is only found at the end of a syllable. When the same alphabetical sign occurs at the beginning of a syllable, it represents a spiritus lenis ('), and is prefixed to every syllable which does not begin with a consonant. The use of a letter which is properly a guttural to represent a spiritus lenis, seems to indicate that the loss of a guttural at the beginning of a word is frequent in Korean, and there are other facts which suggest the same inference. The Korean h is closely related to ng, as appears from the circumstance that the Koreans have the same name for both, viz. heng, and that the form of the letters representing them is similar.

In Japanese, a syllable consists of a vowel, or of one consonant followed by a vowel. No syllable contains more than one consonant, or ends with a consonant. Two consonants can never come together.2 This characteristic, together with the comparative poverty of its vowel system, lends Japanese a certain superficial resemblance to the languages of the Polynesian group, but there is no evidence that any real connexion exists. The mechanism of the syllable in Korean

1 Author of a Corean Primer, Mission Press, Shanghai, 1877. It may be procured from Trübner & Co.

2 The modern Japanese language presents exceptions to this statement and to others made in the course of this paper, but the old classical language is always meant, except when otherwise indicated.