Imatges de pàgina

cerebral, and dental classes respectively, according to the organ principally concerned in their modulation. But * e, ^ ai, %h o, ift au, are diphthongal or compound vowels, as explained above at 18. c* So that e and ai are half guttural, half palatal; o and au half guttural, half labial.

b. The vowels are of course considered to be soft letters.

ao. A consonant is not the modulation, but the actual stoppage, of the vocal stream of breath by the contact of one or other of the five organs, and cannot be enunciated excepting in conjunction with a vowel.

a. All the consonants, therefore, are arranged under the five heads of gutturals, palatals, cerebrals, dentals, and labials, according to the organ concerned in stopping the vocal sound.

b. Again, the first two consonants in each of the five classes, and the sibilants, are called hard or surd, because the vocal stream is abruptly and completely interrupted, and no murmuring sound (aghosha) allowed to escape: while all the other letters are called soft or sonant, because the vocal sound is less suddenly and completely arrested, the effect of stopping it being attended with a low murmur (ghosha).

e. Observe, that as the palatal stop is only a modification of the guttural, the point of contact being moved a little more forward from the throat towards the palate t; so the cerebral (murdhanya) stop is a modification of the dental, the difference being, that whereas in the dental consonantal sound the tip of the tongue is brought into direct contact with the back of the front teeth; in the cerebral it is kept more back in the mouth and curled slightly upwards, so as to strike the gums or palate above the teeth, thus producing a more obtuse sound.

d. The name cerebral is retained in deference to established usage. Perhaps a more correct translation of murdhanya would be supernal, as murdhan here denotes the upper part of the palate, and not the head or brain, which is certainly

* If the two vowels a and i are pronounced rapidly they naturally form the sound e pronounced as in prey, or as a and i in sail; and so with the other diphthongs. The sound of ai in aisle may readily be resolved into a and i, and the sound of on in out into d and u.

t The relationship of the palatal to the guttural letters is proved by their frequent interchangeableness in Sanskrit and in other languages. See 17. b. and 176, and compare church with kirk, Sanskrit tatvdr with Latin quatuor, Sanskrit 6a with Latin que and Greek Km, Sanskrit ja'tiu with English knee, Greek yovv, Latin genu. Some German scholars represent the palatals ^ and H by *' and g.

not the organ of enunciation of any letter. But the inaccuracy involved in the word cerebral hardly justifies a change of name. As these letters are pronounced chiefly with the help of the tongue, they are more appropriately called Unguals.

21. A nasal or narisonant letter is a soft letter, in the utterance of which the vocal stream of breath incompletely arrested, as in all soft letters, is forced through the nose instead of the lips. As the soft letters are of five kinds, according to the organ which interrupts the vocal breathing, so the nasal letters, are five, guttural, palatal, cerebral, dental, and labial. See 15.

22. The semivowels y, r, I, v (called antahstha because in the first arrangement of the alphabet they stand between the other consonants and the sibilants) are formed by a vocal breathing, which is only half interrupted, the several organs being only slightly touched by the tongue. They are, therefore, soft or sonant consonants, approaching nearly to the character of vowels; in fact, half vowels, half consonants.

a. Each class of soft letters (excepting the guttural) has its own corresponding semivowel to which it is related. Thus the palatal soft letters 5 i, t i, E e, $ ai, *(j, have V y for their kindred semivowel. (Compare Sanskrit yuvan with Latin juvenis &c.) Similarly ^ r is the kindred semivowel of the cerebral soft letters f ri, ^ ri, and T d; c5 t of the dentals ^ Iri, ^ Iri, and ^ d *; and 9t)of7«, 31 a, ^ 0, ^rt au, and ^ b. The guttural soft letters have no semivowel in Sanskrit, unless the aspirate if h be so regarded.

23. The sibilants or hissing sounds (called winds by the native grammarians) are hard letters, which, nevertheless, strictly speaking, have something the character of vowels. The organs of speech in uttering them, although not closed, are more contracted than in vowels, and the vocal stream of breath in passing through the teeth experiences a friction which causes sibilation.

a. Sanskrit does not recognise any guttural sibilation, though the palatal sibilant is really half a guttural. See 17.0. The aspirate ^ h might perhaps be regarded as a guttural flatus or wind without sibilating sound. The labial sibilation denoted by the letter/, and the soft sibilation denoted by z, are unknown in Sanskrit.

b. In the S'iva-siitras of native grammars the letters are arranged in fourteen

* That 75 I is a dental, and kindred to T d, is proved by its interchangeableness with d in cognate languages. Thus lacryma, OMCpvpuz. Compare also m with groups: thus, ai it nri Iri k—e o n—ai au 6h yvr t—l nh mn- nn m—jk bk ngk dk dk she bg dds—kk ph 6k tk th ( t t t—k p y—f sh s r—h I. By taking the first letter of any series and joining it to the last of any other series various classes of letters are designated; thus al is the technical name for the whole alphabet; hal for all the consonants; ad the vowels; ait all the simple vowels; an the vowels a,i, u, short or long; ed the diphthongs; yan the semivowels; jai the soft consonants g, j, d, d, b; has the same with their aspirates; jhask the soft aspirates alone; yar all the consonants except k; jhal all the consonants except the nasals and semivowels; jhar all the consonants except the aspirate, nasals, and semivowels.


34. Accentuation in Sanskrit is only marked in the Vedas. Only three names for the accents are generally recognised by grammarians; viz. 1. Vddtta, raised,' i. e. the elevated or high tone, marked in Roman writing by the acute accent; 2. Anuddtta,'not raised,' i. e. the low or grave tone; 3. Svarita,' sounded,' i. e. the sustained tone, neither high nor low, but the combination of the two (samdkdra, Pan. I. 2. 32) which is thus produced. In pronouncing the syllable immediately following the high-toned syllable, the voice unable to lower itself abruptly to the level of the low intonation, is sustained in a tone not as high as the uddtta, and yet not so low as the anuddtla. A syllable uttered with this sustained mixed intonation is said to be trarita, sounded.' These three accents, according to native grammarians, are severally produced, through intensifying, relaxing, and sustaining or throwing out the voice (dydma pUrambha dkshepa); and these operations are said to be connected with an upward, downward, and horizontal motion (tiryag-gamana) of the organs of utterance, which may be illustrated by the movements of the hand in conducting a musical performance *.

But although there are only three recognised names for the accents, there are in reality four tones. This may be proved (as Prof. Roth observes) by any one who tries to adjust the exact relationship between the sounds of the three accents above described. If they are arranged in regular musical series or progression, one link will be found wanting. The uddtta and svarita are names for (so to speak) positive sounds, and the anuddtta for negative; but the neutral, general, accentless sound, which may be compared to a flat horizontal line, and lies as it were between the positive and negative, remains undesignated.

Those grammarians, such as Panini, who recognise only three names for the accents, apply the name anuddtta to this neutral accentless sound also. Hence this name becomes unsuited to the low tone, properly so called, i. e. the tone which immediately precedes the high and is lower than the flat horizontal line taken to represent the general accentless sound. The fact is that the exertion

* In native grammars the uddtta sound of a vowel is said to result from employing the upper half of the organs of utterance, and the anuddtta from employing the lower half.


required to produce the high tone (udulla) is so great that in order to obtain the proper pitch, the voice is obliged to lower the tone of the preceding syllable as much below this flat line as the syllable that bears the uddtta is raised above it; and Panini himself explains this lower tone by the term sannatara (for which the commentators have substituted the expression anuddttatara), while he explains the neutral accentless tone by the term eka-sruti (called in the Pratisakhyas praiaya or pra&ta), i. e. the one monotonous sound in which the ear can perceive no variation. We have therefore really four tones in Sanskrit, and four expressions are now usually adopted to correspond. The name anuddtta is confined to the neutral, indifferent, accentless or monotonous tone represented by the flat horizontal line. The expression anuddttatara has been adopted to designate the lowest sound of all or that immediately preceding the uddtta, while the svarita (which in some respects corresponds with the Greek circumflex) denotes the mixed sustained sound which follows the uddtta.

25. The three accents are thus marked in the Rig-veda.

When a syllable having a horizontal mark underneath (anuddttatara) is followed by one bearing no mark, the one bearing no mark is uddtta; and when followed by two syllables, bearing no mark, both are uddtta.

The svarita accent is denoted by a small perpendicular stroke above the syllable. Thus in the word 1*li. the syllable ^ is anuddttatara, qt is uddtta, and T is svarita.

In the Pada text (if anuddttatara be admitted) the horizontal stroke under a syllable may mark both the anuddtta or neutral tone, and the anuddttatara or low tone; and if it extend under all the syllables of the same word, the whole word is anuddtta accentless, thus TtW.. In the Samhita, the stroke underneath marks the

anuddttatara and all such anuddtta syllables as precede the first anuddttatara syllable, but in the remainder of the sentence the absence of accent (anuddtta) is denoted by the absence of all mark after the svarita until the next anuddttatara.

In fact all the syllables (in a word or sentence) which follow the svarita are supposed to be pronounced in the accentless tone until the anuddttatara mark under a syllable appears again; so that the absence of mark may denote both uddtta and anuddtta. Properly, therefore, the anuddttatara mark is the beginning of a series of three accents, of which the svarita is the end; the appearance of this mark preparing the reader for an uddtta immediately following, and a svarita. The latter, however, may sometimes be retarded by a new uddtta syllable. Moreover, the svarita mark does not always imply an anuddttatara mark preceding, as in the word ffl^! at the beginning of a line, where the svarita merely shows that the first syllable is uddtta. Again, in the Pada, where each word stands separately, there may be no svarita following an uddtta, as tTBTT I ^TTrT. It must also be borne in mind that where a svarita is immediately followed by an uddtta syllable, the svarita becomes changed to anuddttatara: thus in firm ^nWfi' the svarita of 1 becomes so changed, because of the uddtta following.

Again, as to the svarita mark, it may either indicate a dependent svarita, or an independent, i. e. either a svarita produced by an uddtta immediately preceding, or a svarita produced by the suppression of a syllable bearing the uddtta, as in Iraqi contracted from 'W'", where the middle syllable is properly uddtta. In the latter case, if the syllable bearing the svarita is long, and another word follows beginning with an uddtta, then that syllable and all preceding syllables in the same word receive the anuddttatara mark, and the figure 3 is inserted to carry the svarita, having also the anuddttatara mark beneath; thus n«ii 3 T.

If the syllable bearing the independent svarita be short, then the figure ^ carries the svarita, with an anuddttatara under it; thus '8)4n"i^.

Observe—The accent in Sanskrit is not confined to the last three syllables of a word, as in Greek and Latin. Observe also, that although the Sanskrit independent tvarita is in some respects similar to the Greek circumflex, it should be borne in mind, that the latter is confined to long syllables, whereas the svarita may also be applied to short *.


26. According to Hindu grammarians every syllable ought to end in a vowel f, and every final consonant ought to be attracted to the beginning of the next syllable; so that where a word ends in a consonant, that consonant ought to be pronounced with the initial letter of the next word. Hence in some Sanskrit MSS. all the syllables are separated by slight spaces, and in others all the words are joined together without any separation. Thus the two words wnftf XXS(l asid raja would in some books be written WT^ft'JTin and in others «l*fl$Ull. In Sanskrit works printed in Europe, the common practice is to separate only those words the final or initial letter of which are not acted on by the rules of combination. In such books asid raja would be written together, wnfljlHI, because the final e is the result of an euphonic change from \, caused by the following t r. There seems, however, but little reason for considering the mere spaces left between the words of a sentence to be incompatible with the operation of euphonic laws; especially as the

• See on the subject of Vedic accentuation, Roth's preface to the Nirukta: two treatises by Whitney in the Journal of the American Oriental Society, vol. IV. p. 195 etc., and V. p. 387 etc.: Au/recht, de accentu compositorum Sanscriticorum, Bonnae, 1847; reviewed by Ben/ey, Giittinger Gelehrte Anzeigen, 1848, p. 1995—


t Unless it end in Anusvara or Visarga h, which in theory are the only consonantal sounds allowed to close a syllable. That Anusvara is not to full consonant i» proved by the fact that it does not impede the operation of rule 70.

« AnteriorContinua »