» REPRESENTATION OF ‘ % 3 . es Py f ‘ ea) or) 4 . vie ee ) e215 Pa & , ? ~ a + < " ' yy esbenteeronnne Mena: : ve tememsieie Sa i, _ On the Representation of Tones _ in Oriental . Languages t | v4 :: : | | Sir G. A. GRIERSON vee = PRINTED BY STEPHEN AUSTIN & SONS, HERTFORD. mer G Pace ts, Ode csr” ee att On the Representation of Tones in Oriental Languages y/ By Sir GEORGE GRIERSON, K.C.I.E., M.R.A.S. N offering the following notes I would especially ask for consideration on the part of students of Sinitic languages. T in no way pretend to be familiar with these forms of speech, and I must explain that, while I shall be grateful for criticisms on their part, my remarks are not directly addressed to them, but rather to those students of philology who have not made a special study of Chinese, but who, for purposes of comparison, are compelled to acknowledge a bowing acquaintance with all, or nearly all, the tonic languages of Asia. First of all let me state quite clearly that in this paper I do not pretend to lay down a final statement as to the natures of the various tones in the languages of which examples are offered. The accounts given by me are in many cases mere compilations from the descriptions found in standard grammars. I am fully conscious that I may have misunder- stood the statements found there, and, moreover, that differences of opinion exist in regard to some languages. Of the latter, Chinese, Burmese, and Tibetan are examples. The correctness of my accounts of the tones is, indeed, hardly relevant to the object of my paper, which is merely to devise a system of representing all possible tones, and not to describe accurately the tones of any particular form of speech. The examples are given only to show that such a system is possible, and if the tones are incorrectly shown by me, it 1s of little importance for my main object. If any particular symbol chance to be wrong another and more correct symbol can always be substituted for it. The system suggested by me will at once supply the correct symbol. Hitherto, in working at the Linguistic Survey of India, I have had very little to do with tones, and their indication has offered no difficulties. None of the languages with which 2 TONES IN ORIENTAL LANGUAGES I have had to deal has had more than two or, at most, three, tones. Or, perhaps, it would be more correct to say that, in regard to less-known languages, the existence of tones has been ignored by my reporters, or has been imperfectly noticed by them. | Now that the General Survey is completed, and that I am busy with the Introductory Volume, the question of com- parison with languages of the Far Hast, such as Siamese, Annamese, and Chinese—not to speak of Burmese and Shan— has arisen. In most of these languages a system of tones is in vigorous existence, and I am faced with the necessity of indicating them all on one uniform system. So far as I can ascertain, no such system, suitable for my requirements, exists at present. Even for Chinese, different writers employ different systems. Those who follow Wade employ numbers. Others employ diacritical marks, and others, again, employ a modification of the signs employed by the Chinese. For Siamese, European writers employ diacritical marks; for Shan a system of numbers is employed; for Annamese our French fellow-workers have an independent system inherited from Portuguese and Spanish Missionaries ; and for Burmese it is most customary to use the signs found in the native character. As an example take the word for “horse”, common to many of these languages. In Pekinese it is written ma, in Cantonese £ md, in Siamese md, in Shan made, in Burmese mrang8, and in Kachin by ma. In all these, except Kachin, the tone is indicated, but each language employs a different method. A further difficulty arises from the necessity of tracing a tone from one language or dialect to another. Hach writer gives his description of the tones of his language in different terms, and sometimes these descriptions are so indefinite that it 1s not easy to know what is meant. For instance, if we take, say, the low level tone of Cantonese, it is not at all easy to trace, from the accounts in the Grammars, if this tone occurs in Shan, and, if so, what it is there called. TONES IN ORIENTAL LANGUAGES 3 I think that the best way of settling the difficulty is first of all to devise a framework which will permit of the repre- sentation of all possible tones, and then to fit into it the existing tones of each language. In preparing such a framework one or two points must be borne in mind.! First, the tone indications must be simple and readily intelligible. Words with tones, as quoted by me, will be isolated words, scattered through words of many more non- tonic languages, and their quotation is intended for the information of people who are not necessarily acquainted with the languages to which they belong. Hence a system of indicating tones by numbers, necessitating a reference to a key whenever a word is met, will not do. We want something that will appeal at once to the eye. Secondly, the general framework proposed cannot go into minute differences. It may indicate that a tone is high, or that it is low; but it would introduce great complications if an attempt were made to indicate, say, how high the high pitch is compared with a tone in a low pitch, or vice versa.: So also it may indicate a rise or fall, but it cannot indicate the extent of the rise or fall. Nor can it indicate absolute pitch, as between languages. For instance, the same mark may be employed to indicate the high level tone in Cantonese and in Siamese. But this will not mean that the pitch of these two tones is identical. It will only mean that, in the case of each language, this level tone is pitched high in com- parison with other level tones in the same language. Thirdly, as regards the position of signs. For my purposes it will not do to put them over or under any letter or a word. The words are already overloaded with other diacritical marks indicating length, stress, and what not. They must, _therefore, come either before or after the word to which they belong, and, from practical experience, after several 1 T find from experience that the system employed by Lepsius and that devised by P. Schmidt for use in Anthrop@s do not meet my requirements. 4. TONES IN ORIENTAL LANGUAGES experiments, I find that it is easiest to put them before it. Such a mark catches the eye of the reader, and warns him how to pronounce the following word. It is first necessary to state what I mean by the word “tone”. After consultation with expert friends I believe that the following definition meets the requirements of the case :— “A tone is a relatively fixed musical pitch or change of pitch, inherent in a word, and necessary for its significance.” The pitch is “relatively ” fixed, because it is fixed, with reference to other pitches, relatively to the range of the speaker’s voice. But, for the same speaker, the intervals between the pitches of different words are approximately the same as the intervals between the pitches of the same words as uttered by another speaker. To quote an example from Mr. Daniel Jones’s Cantonese Phonetic Reader, he gives the following example of the different musical pitches as uttered in Cantonese by a male voice :— and adds, “for ladies’ voices this tune might be transposed thus”: 1 It will be seen that, though the pitches of these two tunes are different, the relative intervals are the same in each. In other words, everyone sings the same tune, though not necessarily in the same key. The above definition holds good as a general rule, but one exception may be noted, although of no importance for our + As suggested by Mr. Jones, I have pitched the female register two notes lower than as given in his book. TONES IN ORIENTAL LANGUAGES 5 present purpose. The definition says that the tone is inherent in a word and is necessary for its significance. Take, for example, the Siamese word for “come”. It is ma, and to have this significance it must be uttered with a mid level tone. If it is uttered with any other tone it means something else. Thus, with a low level tone, it means “soak ’’, with a high level tone followed by a fall it means “a horse”, with a falling tone it means “beautiful”, and with a rising tone it means “a dog”. While a particular tone is thus usually associated with a word to give a particular signification, it is to be observed that in some, and perhaps in all, tonic languages, in certain col- locations a word occasionally changes its pitch or has no definite pitch. In such cases the de-tonation occurs with the less emphatic. words of the sentence. For instance, the Siamese word _cu7, then, has a low level pitch ; but in the phrase ld bida veur _bain +k‘av \hai, and the father then divided the property, cuz has no emphasis, and is in this collocation uttered with a falling tone. If the “then” were: emphasized, we should then have the regular ecw. Similarly, for Chinese, Monsieur Guernier, in his Notes sur la Prononciation de la Langue Mandarine de ‘Pekin (p. 11), says :— Comme dans toutes les langues, la division phonétique de la phrase chinoise peut s établir en groupes de souffle et en groupes de force. Ces groupes sont composés de syllabes faibles et de syllabes fortes dont l’alternance détermine un principe rythmique d’un caractere particulier en chinois parlé du Nord. La syllabe forte ou accentuée d’un groupe consiste, en pékinois, dans la reproduction, avec une sorte d’insistance, du ton individuel de cette syllabe, alors que les tons des syllabes faibles ou moyennes sont si peu modulés, que, pratiquement, la note chantée propre & chacune d’elles ne se fait pas entendre; il ne subsiste que des sons émis normalement. It will be observed that a tone is defined as “a musical pitch or change of pitch”. There is nothing in this definition 6 TONES IN ORIENTAL LANGUAGES referring to the length of the musical note or progression uttered. A toned syllable may be long or short; and just as in music a note may be tenuto or staccato, so a toned syllable may be held on, smoothly connected with the following syllable, or may be cut off abruptly at its termination. When a toned word has a long vowel, it is sometimes said that the tone is prolonged. But tone is pitch, and it is the note on which the pitch is held, not the pitch itself, that is prolonged. A pitch may be high or low, but it cannot be long or short. Hence it is wrong to talk of a prolonged tone. When a word is cut off abruptly it is said by Chinese scholars to possess the “ entering tone”; but, according to the above definition, this abrupt ending is not a tone at all, any more than a staccato check in music is a part of the musical pitch with which it appears. In fact a word may have any real tone—rising, falling, level, or what not—and at the same time have, or have not, this abrupt ending. For instance, Professor Parker, whose authority on the subject is universally admitted, informs me that in Southern Mandarin the “ entering tone ” has usually the same cadence as the lower level tone, and that in Cantonese the so-called “ upper entering tone” has the same cadence as that of the upper level tone, while the middle entering tone ”’ has that of the upper departing tone, and the “lower entering tone”’ has that of the lower departing tone. On the other hand, in Hakka, the “ upper entering tone’’ has the cadence of the lower level tone, while the “lower entering tone” has that of the upper departing tone. Translating this into the terms hitherto used by me, and rendered necessary by my definition of the word ‘“ tone ’’, we find that :— (1) In Southern Mandarin, when a syllable is checked, it usually has the low level tone. (2) In Cantonese, when a syllable is checked, it may have the _ high level tone, the high falling tone, or the low falling tone. Ped eee at Mr. Jones informs me that in colloquial Cantonese checked syllables also have the high rising tone. Thus “ko_jat, the sun, is pronounced ko‘jat in colloquial. TONES IN ORIENTAL LANGUAGES 7 (3) In Hakka, when a syllable is checked, it may have either the low level tone or the high falling tone. Tt will be convenient to clear off this question of the abrupt check before proceeding to the consideration of the tones proper. Not being a tone, what is it? It corresponds, as I have said, to the staccato of music, and, just as either a long musical note or a short musical note may be staccato, so there is no question of length in connexion with this check. A vowel or syllable may be long or it may be short, and in either case its final utterance may be abruptly checked. For practical purposes we may say that when a syllable ends in. a vowel or nasal immediately preceded by a vowel, the abrupt- hess is caused by a glottal check. If the word ends in a consonant (usually k, ch, p, or t), that consonant is pronounced without the off-glide or release. It is necessary to distinguish these two cases, and I propose to indicate the glottal check by ° after (not before, as in the case of signs indicating tones) the checked syllable, as in the Southern Mandarin Sap one. When a final consonant is sounded without the off-glide, I follow the example of Mr: Drake in his Karki Grammar, and write the consonant small, and above the line, as in Cantonese ~yat ‘pat, one hundred. The sign used by phoneticians for the glottal check is’. I do not employ it, because a sign closely resembling this is used by some Chinese scholars for other tonal purposes, and because in French Indo-China it is employed for the “ question ”’, i.e. rising, tone. Its use here for the glottal check would therefore give rise to confusion. Having thus disposed of the so-called “ entering tone ”’, we may now discuss the question of the true tones. A musical pitch or change of pitch may be either constant or variable, If it is constant, it may continue level on the same note, or it may rise, or it may fall. When it is thus constant I call it a “simple tone”. Or a change of pitch may be variable. It may first be level, and then rise or fall, or it may rise and then fall, or vice versa. Such a variable change of pitch I call a “ compound tone”. 8 TONES IN ORIENTAL LANGUAGES The simple tones are of three kinds, level, rising, and falling. I represent the level tone by a straight horizontal line before the syllable; thus, -ka; a rising tone I indicate by a line + sloping upwards, as in /ka, and a falling tone by a line sloping downwards, as in \ka. The indication of a compound tone can be based on the above. Thus a rising-falling tone would be indicated by 1, a falling-rising tone by v, a level- falling tone by ~, and a level-rising tone by ~. I claim no originality for this device. I do not know who first suggested it, but I owe the general principle to Mr. Daniel Jones, and I observe that it is also followed by Mr. Grant Brown in his works on Burmese languages. A syllable may further be pitched high, or in the middle register of the voice, or low. Tones occur in all these three rezisters, and I propose to indicate the register (or, as Chinese scholars call it, the series) by its position on the line. Tones in the high register will be indicated by marks above the line, in the middle register by marks on the line, and in the low register by marks below the line. We may thus illustrate the nine simple tones as follows :— Level. Rising. Falling. High Register. ka ‘ka ka Middle Register . -ka ka \ka Low Register. _ka ka ka The register of the compound tones can be illustrated in the same way. Thus, “ka, “ka, Yka, and so on for these and others. There is one objection to the above scheme—that the mid level tone is liable to be mistaken for a hyphen. I propose to avoid this difficulty by, in this case, omitting the tone- mark. Thus, ka, not -ka. In this way the hyphen can be used for its proper purpose. In Siamese and other languages it is also customary to omit the sign for this tone, so that in this I am only following established practice. The middle series is in the ordinary natural register of the voice, and it TONES IN ORIENTAL LANGUAGES 9 is a simple matter to explain that when a syllable is pro- nounced in a level tone in the natural register of the voice it is given no tone-mark. The whole nine simple tones do not, so far as I am aware, all occur in any one language. A great many languages have only two registers, a higher and a lower. In such cases I would omit the tone-signs for the middle register altogether. This will be in accord with the general proviso already mentioned, that my system does not pretend to indicate how great is the interval between any two registers. The low series may quite possibly be on the middle register, and, if it were required to be particularly accurate, it might be shown as such. But, for my purposes, I wish only to show that one series is higher than the other, and I accordingly put one series above and the other below the line. For instance, _ Southern Mandarin has only two level registers, and I indicate them on this principle. Thus (Douglas’s spelling): ,wo “kin “tien tung shén, smitao sche ili _ pur chi shwei \lu, I start to-day. Having lost myself I came here and don’t know my way back. ; I now proceed to give examples of the application of this system of indicating tones in various languages. The first example is in Southern Mandarin Chinese. It is the first few verses of the parable of the Prodigal Son, and has been transcribed for me into the International Phonetic Script by Dr. H. B. Morse. This Chinese dialect has (excluding the “ entering tone ”’) four tones, viz. the high level, the low level,’ the rising, and the departing tone. For the rising and departing (i.e. falling) tones there is no distinction of higher and lower. I therefore represent the four tones as follows : 1 This is the conventional name and description. But Professor Bradley’s researches with the Rousselot apparatus (Journal of the North China Branch of the Royal Asiatic Society, vol. xlvi (1915), pp. 40 ff.) show that in Pekinese this tone is really low rising. For our present purposes I follow convention. 10 TONES IN ORIENTAL LANGUAGES “ka, _ka, ka, \ ka. Swe iye. 0) Rawo, yu: \Ord /dzi. Again say suppose have (= there are) two son. _tfi: \d31: Reba uivaaenisashsiecupycnpeinee e144 He (or this) younger son talk father say, “request \fu: i: 180 Moen yer. “d3i father take whatsoever must obtain (possessive or Sjey5 ry WO,’ \fu: \SuL relative) patrimony, to I, — father according(ly) take property daide he (or merely euphonic). “1 tfan “fan dz; The next specimen is the same passage read in Cantonese. It also has been transcribed into phonetic spelling by Dr.- Morse. Omitting the so-called “ entering tone”, regarding which see above, p. 458, this form of Chinese has six principal ~ tones, viz. high and low level, high and low rising, and high and low falling, represented severally as follows: “ka, _ka, ‘ka, ,ka, ‘ka, and \ka. | VJausaciy’ vi wakic jeu. “Mami tna om ical a ec scun ee Again say suppose have two son. This younger AWAY Hn ah vaste EN gay ip A pk manna | an PHRF tS) son talk father say ‘request father take whatsoever “ton ~— tak 31: jup ay. must obtain (possessive, or relative) patrimony to BIOS a awit: Sul ib. jam fan I’. Father according(ly) take property divide “d3i:. he (or merely ewphonic). From the time of Pallegoix the tones of Siamese have been variously described, and authors have differed. I have had the advantage of discussing the question with Mr. Daniel Jones in the company of a Siamese gentleman, and our results closely correspond with those given by Professor Bradley pmy TONES IN ORIENTAL LANGUAGES 11 in vol. xxxi (1911), pp. 282 ff. of the Journal of the American Oriental Society. I can therefore, omitting minor details, give them with considerable confidence. Siamese has six tones, as follows :— (1) A mid level. This is Pallegoix’s Tonus rectus. As explained above, I give it no sign. Thus, md, come. (2) A low level. This is Pallegoix’s Tonus curcumflecus. I represent it thus, _md, soak. (3) A high level, preceded by a slight rapid rise, and ending with a rapid fall through a considerable interval. This is Pallegoix’s Tonus gravis. Wershoven wrongly calls it ‘der eingehender Ton’’, for it has nothing to do with the so-called “entering tone” of Chinese. In indicating it I omit reference to the slight initial rise, and represent it thus: ~md, a horse. (4) A high level. It occurs only in certain syllables ending (1) in a short vowel with a glottal check, or (2) in a checked consonant. The grammarians look upon it as a variant of No. 3, and give it no special name. I indicate it byt, is in ma, often. - (5) A mid falling. This is Pallegoix’s Tonus demissus. I represent it thus, vd, beautiful. (6) A mid rising. This is Pallegoix’s Tonus altus. I indicate it thus, md, a dog. As an example I give a few verses from the parable of the Prodigal Son. The system of spelling requires a few explanations. The letter © represents the sound of the a in the German ‘“‘ Mann”, and a (long or short) is a variety of unrounded 0, not far from the a of “ America’. The sound represented by w (long or short) is a variety of unrounded uw, the w of the International Phonetic Association. The other letters call for no remarks. Kons snun ~mi,:* but. \efai’) san .| kon, {a Person one have son male two person. And _but = ~néi ~nan wa ka bi-da \wa, “bi-da son little that =say to father that ‘father 12 TONES IN ORIENTAL LANGUAGES \cau-\k‘a, ka _suion \saP tT _tok_ yt Sur, please-give share (of) goods which falls ki \k‘a-p‘o-\cau mae la bi-da \cun to I (i.e.me) (imperative)’. And father then _ bain /k‘an chai sk& )6©but tan =san =-Nnan. divide property gwe to son all two that. x * x * * * x ‘kK Gan Wce, “luk . ,k‘un pai ha bi-da I, will stand-up arise go see father (of) vKay da c& wa ka tan wa, etc. I, and will’ say to he that, ete. The tones of Shan have been discussed by me with Mr. Daniel Jones and a native of the Southern Shan States, and the following tones have been fixed :— (1) A mid rising, as in md, a dog. This is Dr. Cushing’s first tone. (2) A low level, as in _mda, the shoulder. This 1s Dr. Cushing’s second tone. (3) Amid falling, asin wd, tobe mad. This is Dr.Cushing’s third tone. In connected speech it is mid level, as in md. (4) A high level, as in “ma, come. This is Dr. Cushing’s fourth tone. (5) A high falling, as in ‘ma, a horse. This is Dr. Cushing’s fifth tone. As an example, I give a few sentences from Dr. Cushing’s Elementary Handbook of the Shan Language. The letters a and wu are pronounced nearly as in Siamese. KOR nal belkin ti). ae Person this get money did. ie. This man _ has obtained money. LIMAN At 8: es aa a a. He buy horse’ that does. i.e. He buys that pony. aman aa als one eal: He sell goods ¢ i.e. Does he sell goods ? TONES IN ORIENTAL LANGUAGES 13 “mau vhan “kon ‘nan “hi. Thou see person that ~? i.e. Dost thou see that man 2 /kau-\k‘4 chan “man 4. I see him do. I see him. Annamese has six tones, as follows :— (1) A high rising, as in ‘ma, the cheek, written md in Quocangii. (2) A low falling, as in ma, but, written ma in ‘Quocangii. (3) A low level, as in _ma, a rice-seed plot, written ma in Quocangii. (4) A midrising, as in sma, a grave, written ma in AQuocangii. (5) A rising-falling, as in ama, a horse, written ma in Quocangti. This tone is variously described by different writers. Aubaret describes it as falling-rising. Others again differ. I follow the description given by a learned Annamese, P. J. B. Triténg vVinh ‘Ky, kindly communicated to me by Monsieur Cabaton. The different descriptions of this tone may be due to differences of dialect. (6) A mid level, for which as in‘Quoenngii I give no mark, as in ma, a phantom. As examples, I give a few sentences from pp. 17 and 40 of Aubaret’s Grammar. The spelling is that usually employed for AQuocangii, except that i and 6 respectively are used instead of “ bearded ” uw and o, and that the “ barred” d is represented by d. anrhiing- ngiiort shai ‘chét, men must die. ngidr ndy la chu téi, this man is my father. gan day ‘co _mét ngiiii _ho la Viidng, near here is a man (whose) first-name is Viiéng. khong ‘co ngiidi ‘hiéw ‘dam :bo cha me, there is not a pious man (who) will dare (to) abandon (his) father mother (i.e. parents). 14 TONES IN ORIENTAL LANGUAGES As an example of the Man languages of Indo-China, we may take Miao-tsii. It has four tones, as follows :— (1) A high level, as in “dav, a manger. (2) A high rising, as in ‘dan, large. (3) A high level-falling tone, like the third Siamese tone, as given above. An example is ~da7v, yellow. (4) A low rising, as in_ ,dav, to tell falsehoods. The glottal check may be heard sometimes in this language, together with the high level tone, as in “daiv°, to wallow. I regret that I have been unable to procure any text, or even a single connected sentence in this lantuage. The above examples are taken from Monsieur M. F. Savina’s very complete Dictionary (BEFEO., vol. xvi). I am indebted to Mr. L. I. Taylor, of the Burma Education Service, for very full information regarding the tones of Sga and Pwo Karen. Taking first the tones of Sga Karen, it is to be observed that this dialect has received a certain amount of literary cultivation, and is the official dialect of the schools. What is here described is the language so taught. There are several variant dialects amongst the uneducated. This form of Sga Karen has five tones, each of which is indicated in writing the Karen alphabet, according to the system of spelling introduced by the missionaries. These are as follows :— (1) A mid level, as in ka. As in the native script, it is not marked. (2) Another mid level, slightly higher than No. 1. For want of a better sign I mark it wka. In the native script it is marked 1. It is sometimes accompanied by a glottal check, a; in ~ ka°, and in the native script is then marked 4 . (3) A high level, usually accompanied by a glottal check, asin ka®°. In the native script it is marked g. (4) A mid falling, as in ka. In the native script it is marked $. TONES IN ORIENTAL LANGUAGES 15 (5) A high falling, as in ‘ta. In the native script it is marked p. As an example I give a short sentence from Mr. Gilmore’s Grammar. As the mid level first tone is not indicated, I use the hyphen for its proper purpose. ta-bla-bla ~Sa-Wa “6a°-\,bya ‘né “GU th teaca went Sometimes Saw-Wa old (that) build up house k1é-k]é ~~ la. quickly does. : i.e. Sometimes old Saw-Wa builds a house quickly. The tones of Pwo Karen are also marked in writing in the native character. In his Manual Mr. Duffin mentions no less than ten, but four of these are simply nasalizations of four of the others. Mr. Taylor tells me that there are two chief dialects of this language, that of Tenasserim, and that of Bassein. These differ considerably in tones. Taking Bassein Pwo for our present purposes, Mr. Taylor gives me the following . information. The tones vary: considerably, the same written ~ sign sometimes indicating what may be called the standard of the tone, and at other times used with a syllable which exhibits considerable variation from the norm. For instance, the first tone is usually a mid level, but the syllable ya, although marked as having this tone is pronounced with a high level tone, as if it were yd. Omitting these irregularities, the normal tones in Bassein tones are as follows :— (1) A mid level, as in ka. As in the native script, it is not marked. (2) A mid rising, as in -ka. In the native script it 1s marked 1. (3) A mid level, slightly higher than No. 1. For want of a better sign I mark it ~ka. In the native script it 1s marked J. (4) A low level, as in _ka. In the native script it 1s marked 1. 16 TONES IN ORIENTAL LANGUAGES (5) A high level, as in “ka. In the native script it is marked $. This tone is sometimes accompanied by a glottal check, as in “ka®, and is then marked t in the native script. As an example of Pwo Karen, with Bassein tones, I give a few sentences from Mr. Duffin’s Manual. As the mid level first tone is not indicated, I use the hyphen for its proper purpose, ya ~@é- ya 16 a-ka-s-ma-s‘6-~na . 6°. I know — that he-will-do-it not. i.e. I do not know that he will do it. na ka-—§6 “ru mwai-»ba-~ lai, Your horse shy why a i.e. Why does your horse shy ? mwal-l6 a-_wé a- mé°- ~§a-~ bla-ka-_ ka a-/kyan 1a. be- he eye-blind-one -cause (is). le. Because he is blind of one eye. As regards Burmese, according to Mr. Grant Brown, in his Half the Battle in Burmese, there are three tones for words ending in vowels (except the unaccented short a) or nasals. Words ending in other letters have no special tones, though final consonants are sounded without the off-glide. The three tones are :— (1) A mid level, as in -ka. Words with this tone are left without indication in the native script. (2) A high falling, always accompanied by a glottal check as in ‘ka°. In the native script it is indicated by the sign >: Some people describe this as a rising tone, and the facts are doubtful. Whether there is a rise or a fall is obscured by the check. I follow Mr. Grant Brown in describing it as falling. Mr. Taylor thinks it is rising, falling, or level, according to the context or the speaker. Its essential element is its height. (3) A low falling tone, as in ‘ka. In conversation it is often mid falling, or even a little higher, the height being TONES IN ORIENTAL LANGUAGES 17 affected by sentence-stress. Theoretically, however, it is low falling. The following sentences from Mr. Grant Brown’s book will serve as an example. Words without tone have no tone- mark, and the hyphen-marks consequently indicate the mid level tone. min‘do® Sne°® dain \pyadd-lo -#in-dyin-de. -sa de-hma You (pl.) daily speak-as — learn-want. —-writing-in \yeda-lo ma-din-dyin Se bu. -sa ye‘de°d tain ma pya‘hnin® writes-as not-want-learn-yet. writing-writes-as not-s peak (umpve.) -ni. na-léda Ja. na-lé-ba-dé. ‘ne° dain — pyada-lo will-you? understand? (I) understand. Daily speak-as kaun gaun tat pi‘hma°® -sayedi-lo — gin-me. good-good know-how-complete-after writing-write-as learn-will. That is to say :— I want to learn the language you use every day, not the written language. Don’t use the written language, now, will you? Do you understand ? I understand. ee When I have thoroughly mastered the language of conversation | will learn the written language. I think it will be found that the many Tibeto-Burman languages spoken in North Burma and in Assam all have tones, but no information on the subject is available for most of them, and, as for the others, with one exception, what we are told is either misleading or incomplete. The one exception is the language of the Kadus of Upper Burma. In the Bulletin of the School of Oriental Studies ! Mr. Grant Brown has lately given us full particulars of the tones of this form of speech. It has three tones—a mid level, as in -ka, a low falling, as ka, and a high falling, always accompanied by a glottal check, as in ‘ka°. The high falling tone is heard only when a word is pronounced deliberately. In unstressed syllables 1 Vol. I, pt. iii (1920), pp. 1 ff. 18 TONES IN ORIENTAL LANGUAGES in ordinary conversation it is neglected. These three tones are used only with syllables ending in a vowel or nasal. A final ¢ or k is pronounced without the off-glide. It will be observed that the tones are the same as those of Burmese. The following example of Kadu, written in the alphabet of the International Phonetic Association, I owe to the kindness of Mr. Grant Brown. It consists of the first three verses of the parable of the Prodigal Son. Where no tone- marks are shown the syllable has no significant tone. te-mi-s‘a@ ho -wa be -s‘a ke-ley (hu -na -ma. Man person one at son two person be (past). “S‘a -s‘j@ hek o-wa -ba -nan ban na, ‘o-wa -a Son small ... father to go having, ‘Father I gins jel gabe -ou-ne o-mwe et JUp Ja,’ for get have-to as-much-as inheritance gue (umperative),’ “Sen -ma o0-wa hek havy. -de -man -i -ma. ask (past). Father ... goods (object) divide give (past). As an example of one of the best available accounts of the tones of others of these languages. I here quote from Mr. Henson’s Kachin Grammar what he says about the tones of that form of speech. He says: (a) The tones of the Kachin language, although not of 50 great importance as in Chinese or Shin, must, however, be carefully mastered and practised, as accuracy in speaking depends to a great extent on the power to enunciate them correctly. (0) The tones often give to a vowel naturally long, a shorter sound than it otherwise would take, and vice versa. (c) Five tones are easily distinguished in Kachin, and may, for lack of better tonal marks, be indicated by the numerals 1, 2, 3, 4, 5. (1) The first tone is the natural pitch of the voice, with a slight rising inflection at the end. It may be called the natural tone. TONES IN ORIENTAL LANGUAGES 19 _ (2) The second is a bass tone; it may be called the grave tone. (3) The third is a slightly higher tone than the second, being pronounced with an even prolonged sound ; it may be called the riseng tone. (4) The fourth tone is very short and abrupt; it may be called the abrupt tone. (5) The fifth tone is somewhat higher than the third, and — is uttered with more emphasis; it may be called the emphatic tone. He then gives, as examples, the tones of three different words. Not another tone is marked in his grammar. If I make the following remarks upon his statements it is not with the intention of fault-finding. It is rather with the object of showing how tones, as I define them, are, even by those most familiar with a language, confused with prosodial length, and with checked sounds. Let us take his statement (6). He says that a tone may give a shorter sound to a vowel naturally long, and vice versa. It is difficult to understand what exactly is meant here. I do not see how any vowel in a Tibeto-Burman language can be short and at the same time “naturally ” long. In these languages a syllable has once for all either a long or a short vowel. The length of this vowel cannot be altered by any tone. The tone is as much inherent in the syllable as the length or shortness of the vowel. If the tone is different the word is another word, and means something else. Similarly, if the quantity of the vowel is altered, it again becomes another word, and again means something else. No tone can possibly alter the quantity of any vowel. A tone is a musical pitch, not a musical note, and can have no effect on quantity. No musical pitch can lengthen or shorten any sound. The most that we can say is that such and such a tone is usually accom- panied by such and such a vowel, but the fact of company in no way identifies the one with the other. It is, however, dangerous to say even so much. Let us take as an example 20 TONES IN ORIENTAL LANGUAGES two such cases as -wa and -wa. Here we have two different words with the same mid level tone. The fact that in one the vowel is short and in the other the vowel is long does not affect the tone at all. A definition like Mr. Hanson’s statement would make out that there are two different tones. The definition of the first would be that the tone was the “ natural pitch of the voice’’. The definition of the second would be that the tone was the natural pitch of the voice, but was “pronounced with a prolonged sound”. Such a division is unnecessary and confusing. Taking Mr. Hanson’s five tones in order, we find :— (1) If the account of the first tone is correct, it may be represented by ~wa, a hut. Perhaps, however, what is intended is merely the ordinary mid level tone. (2) It is not stated whether this tone is level, rising, OF falling. Assuming that it is level, we may represent it by _wa, to return. (3) Here the tone is said to be “rising”, and at the same time is said to be “even”. If ‘even’ means “level ’— the usual signification of the word in this connexion—the two words “ rising’ and “ even” are incompatible. Perhaps “even ” here means “ with a sustained voice’’. It is further said to be “ prolonged”. As explained above, this is simply a statement of prosodial length, and has nothing to do with tone as I have defined it. The prolongation of the sound must be indicated by some sign other than and additional to that employed for indicating the tone. If the tone really is mid- rising, we might indicate it and also the prolongation by, say, swa, a father. (4) It is said that the tone is short and abrupt. It is probably meant that the word ends with a glottal check or an unexploded consonant, but nothing is said about the real tone that accompanies this. It may be level, rising, or falling, low, middle, or high. I am, therefore, unable graphically to represent this tone, as I do not know what it is. If I guess, and say that the glottal check is accompanied by a high TONES IN ORIENTAL LANGUAGES 21 falling tone, as in Burmese and Kadu, I can write it ‘wa®, to weave. (5) It is impossible to define this tone. It is apparently on the high register. But is it level, rising, or falling? Assuming, as in No. 3, that it is rising, it may perhaps be indicated by ‘wa, to bite. It is further said that it is uttered with emphasis. But this has nothing to do with tone. It is a question of stress- accent, which should be indicated by some special appropriate mark, e.g. by the acute accent of ‘wd. It is, however, not clear what exactly is meant by “ emphasis ”’. For other languages of this class we have far less information. What can be more tantalizing when we are told in the only grammar that we possess of Miri that the language “abounds in intonations ”’, and that “it must be understood that a slight modulation of the voice 1s required ”’ in uttering each syllable. No information is given as to what these “ intonations ” are. Even for Tibetan, our information is incomplete. The fullest description is Amundsen’s. He mentions six tones, viz. : Tone 1. High pitched, often nasal, and short as if butted against something. Tone 2. High like tone 1, but long. Tone 3. Medium pitched and short, like tone 1. Tone 4. Medium pitch and long. Tone 5. Curved tone ; deep but gradually raised to medium 99 pitch, like saying “two” in a surprised questioning tone. Tone 6. Descending long tone. It is unnecessary to go into this list in detail. It will be sufficient to point out that, under the head of “tone” are included not only glottal checks and prosodial length, but even nasalization. Moreover, even the true tones cannot all be identified. For instance, it is not stated where the descending (long) tone begins. Is it high, medium, or low ? We may tentatively indicate Amundsen’s six tones as follows :— 22, TONES IN ORIENTAL LANGUAGES (1) ~“ka®, often “ka? (4) -ka 24 Paste er (5) {ka (3) -ka® (6) \ka If the above is correct there are really only four tones, a high level, a mid level, a low rising, and a mid falling, but some of these, especially the last, are doubtful. To sum up :— There are several different factors which attend the utterance of a syllable. For our present purposes we may mention four :— (1) Musical pitch or change of pitch. When this is significant it is a “ tone”, (2) Quantity, ie. the shortness or length of the sound. This is a question of time, and not of pitch. It has therefore nothing to do with “tone”. A syllable with the same tone may be short, long, or prolonged. This fact, if significant, should be indicated by some device distinct from the tone- mark, (3) Smoothness or abruptness. A syllable may be either tenuto, i.e. held throughout the whole time of its utterance, so as to be connected with the next s yllable, or it may be staccato, i.e. held throughout its whole time but abruptly checked at the end, so as not to be connected with the next syllable. This, again, has nothing to do with pitch, and neither the tenuto utterance nor the staccato utterance can be called atone. The fact, if significant, should therefore be indicated by some appropriate mark, distinct from the tone-mark. (4) Stress. This, again, has nothing to do with pitch, any more than a sforzando affects the pitch of a note in music. It must, therefore, if significant, be indicated by a special, independent mark. Many writers mix up these four conditions and group them all together as “tones”, thus giving rise to the greatest confusion, and rendering it impossible for the best-intentioned reader to understand what is meant. On behalf of the TONES IN ORIENTAL LANGUAGES 93 many students who are not familiar with Indo-Chinese languages, but who have to deal with words occurring in them, I would urge all writers on the subject to avoid confusing these four independent factors. and, when describing a tone, to say clearly what it is, i.e. on what register—high, middle, or low—it is uttered, and whether it is level, rising, or falling, or, possibly, a combination of two of these so as to form a compound tone. I put forward the above scheme with all diffidence. If a better is devised I shall be the first to welcome it. I will only add that the devices suggested by me are easily under- stood, and that they are within the powers of any decent printing office without the necessity of casting many new types. APPENDIX Tones in Inpo-EvuropEAN LANGUAGES In the foregoing pages I have dealt only with Indo- Chinese tones ; but tones also occur in Indo-European forms of speech. Norwegian offers a familiar example. For our - present purposes, it is important to notice that Vedic “ Accents ” are merely tones. The fact is obscured by the difference of name, but is a fact nevertheless. Professor Macdonell’s account (Vedic Grammar, p. 448) is as follows :— In the Rigveda the Udatta, the rising accent, has secondarily acquired a middle pitch, lower than the initial pitch of the Svarita. The Svarita is a falling accent representing the descent from the Udatta pitch to toneless- ness. In the Rigveda it rises slightly above the Udatta pitch before descending: here, therefore, it has something of the nature of a circumflex ... The Anudatta is the low tone of the syllables preceding an Udatta. Translating this into the language used in the preceding pages, we find that in the Rigveda :— The Udatta is a mid rising tone, as in ka. The Svarita is a mid rising-falling tone, as in aka. The Anudatta is a low level tone, as in _ka. 24 TONES IN ORIENTAL LANGUAGES The principles followed for Indo-Chinese languages can also be employed for indicating these Vedic tones, although the matter is slightly complicated by the fact that Indo-Aryan languages are polysyllabic, while the others are, in the main, monosyllabic. I may add that the Sanskrit Visarga is merely a glottal check, and is the same as the so-called “entering tone ”’ of the Chinese. As an example, I give the first two verses of the Rigveda, with the tones marked on these principles :— _avgmiamilé _pu/rdshitarh _ya‘jiia,sya _dé/va_mrvtviajain shoatarath ra_tna/dha,tamaih. 1h _avenl? /ptiarvé bhir -rasi_bhir Wid miatanai_ru-ta sa désva /é/ha avaksati 1 2. Of the modern Indo-Aryan languages, Panjabi and Lahnda have tones. These tones do not appear in every word, being significant only in a certain number, much as we have observed in the case of Burmese. In both languages the tones are the same, being three in number. Dr. Grahame Bailey describes them as follows 2 :— All the three tones are compound in character. They are :— 1. Low level-rising (~). This begins on a low register and rises a little, not necessarily more than a tone or a semi- tone. 2. High level-falling (~). This begins on the high register and falls a little. 3. Rising-falling (-). This begins on the low register, rises to the high register, and then falls. In a way, it combines | and 2. When a tone-word is closely followed by another word, the _ second part of the tone falls on the second word. That is to * I have to thank Professor Macdonell for being so kind as to check and correct the above representation of the Vedic accents. * I have taken the liberty of slightly altering Dr. Bailey’s wording so as to bring it into accord with the language used in the preceding pages. TONES IN ORIENTAL LANGUAGES 25 say, that normally the tone requires two syllables to complete it. If there is a pause after the tone syllable, the whole tone will be finished on that one syllable, which will have two notes; but if the tone-syllable is followed by another syllable in the same word, or if, being the final syllable or only syllable of a word, it is followed by another word with no pause between, the second part of the tone will go to the non-tone word. Thus, in reality, it is syllables, and not words, which have the tone. A non-tone word will never have any particular tone except in so far as it accidentally receives one from a tone-word as just described. If two or more syllables with the high falling tone follow closely after one another, the tone will be omitted from some. Hence, in rapid talk, many high falling tones are not sounded. Again, they may be sounded in different ways. Thus neth laihndi in the first Panjabi sentence may have (1) no tone on neih and the whole tone on laihndi (high level part on laihn- and falling part on -di), or (2) tone on neih, in which case the falling part will come on to laihn- of the next word, unless, in a very rare case, when neth is emphasized and has a pause after it, in which case lachndi will get a tone also. Unless — a pause follows, the tone is never completed in one syllable. This division of a tone amongst two syllables is, in its essence, also the case with Vedic “accents”. For instance, the Udatta (mid rising) is immediately preceded by the Anudatta (low level), and the two together would be repre- sented by ~, which is the same as Dr. Bailey’s first tone, and, like it, is distributed over two adjacent syllables. So also the Udatta plus Svarita (, + 4) corresponds to his third tone. : In Panjabi and Lahnda, although this is not recognized by the speakers, the tones are indicated by the spelling of the words, and for this reason in the two following examples— one in standard Panjabi and the other in the Kagani dialect of Lahnda—I give the texts in strict transliteration and also in the International Phonetic Script, with the tone-marks added to the latter. For both I am indebted to the kindness of Dr. Grahame Bailey. 26 TONES IN ORIENTAL LANGUAGES PANJABI TRANSLITERATION.—Mai ikk kahani suni_ si, akhe PHONETIC Script.— Me 1k ~kani suni si, Axe By-meone story heard was, — that os jane di bhukkh neih laihndi jihda sakkhna os jane di ~puk:h ~Nnei Nlendi jida sak:hna that man of hunger not descends whose envpty e dhiddh te jihdi nazr e bhukkhi; bhawe 6) etid:eite yiqidl).. nary e opus ws stomach and whose regard is hungry ; although asi ohntu bahawiye hétha te tarkari te caul te Asi on’ti wbavie w-hethad te terkari te cool te we hm seat below and meat and rice and khand nal rajaiye, ta wi sabbho kujjh khand .nal )/rejaie; ta. 4uvi \8Ab:0 kvj sugar with satisfy, then even every something khake akhega pai mena dhahna khake akhega pAl ment atana _having-eaten —he-will-say _ that me popcorn caba. caba. cause-to-masticate (v.e. feed). KAGANI hikke jane de do puttar ahse. Ohnai bicco ~hik:e jane de do put:er ~ase. ~On'& - bic:o One man of two sons were. Them from-in nikre puttar thorea diharea picche apna nikre putor thored drivhare& pic:he apna small son Sew days after —_his-own mal sara kharab kar churea. Oh hun mal sara xerab kar _ churea. ~O hun property all bad - doing was-left. He = now TONES IN ORIENTAL LANGUAGES bhukkha hundai; hikke jane usko ~pvtk:ha ~hunde; vhik:e jane vsko hungry being-is; one man-by him indhe caranne wase chohrea. “unde ceran:e yase ~chorea. pigs feeding for was-left. Qitbe di jhanht da te dihe da jhagra Qutbe di ,janu da te Ndie da ~jAgra North of gale of and sun of quarrel jehra patti utto ulharsi, oh “jera pAtiu ut:0 ul—arsi 0) who cloak over-from will-take-off he hosi. ~hosi. will-be. 2,7 bhéde pede sheep hoea ; ~hoea ; became ; dahda ~dada strong 2 vv ! Binder Gaylord Bros., Inc. Makers Pamphlet Syracuse, Pat. No. 877188 Cd S 5 i) E 3 ten! ° m- A,