mueller a sanskrit grammar for beginners in devanagari and roman letters throughout

8/12/2019 MUELLER a Sanskrit Grammar for Beginners in Devanagari and Roman Letters Throughout

1/323

SANSKRIT GRAMMAR

FOR BEGINNERS,

IN

DEVANAGARI and ROMAN LETTERS THROUGHOUT,

BY

F. MAX MtJLLER, M.A.,vossiev mniBSB ov thb vbsvch ivititutb,

STC

SECOND EDITION, REVISED AND ACCENTUATED.

LONDON:

LONGMANS, GREEN, AND CO.1870.


2/323

^

xforlr:

T. COMBE, M.A., B. B. QABDNER, AND E. PICKARD HALL,

PBINTXB8 TO THE UNIVEB8ITY.


3/323


4/323

iv PREFACE TO THEcould vie with the wonderfiil mechanism of his eight books ofgrammatical rules. But um-ivalled as that system is,it is notsuited to the wants of English students,least of all to the wants .of beginners. While availingmyself therefore of the materialscollected in the grammar of PS^nini and in later works, such as thePrakriySrKaumudl, the SiddhSnta-Kaumudl, the Sirasvatl Prakriyft,and the MMhavlya-dh tu-vritti,I have abstained, as much as pos-ible,

from introducingany more of the peculiarsystem and ofthe terminologyof Indian grammarians* than has already foundadmittance into our Sanskrit grammars; nay, I have frequentlyrejected the grammatical observations supplied ready to handin their works, in order not to overwhelm the memory of thestudent with too many rules and too many exceptions. WhetherI have always been successful in drawing a line between whatis essential in Sanskrit grammar and what is not, I must leaveto the judgment of those who enjoy the good fortune of beingengaged in the practicalteaching of a language the studentsof which may be counted no longer by tens, but by hundreds f.I only wish it to be understood that where I have left out rules

* The few alterations that I have made in the usual terminology have been madesolelywith a view of facilitatinghe work of the learner. Thus instead of numberingthe ten classes of verbs, I have called each by its first verb. This relieves the memoryof much unnecessary trouble,as the very name indicates the character of each class ;and though the names may at firstsound somewhat uncouth, they are after all the onlynames recognizedby native grammarians. Elnowingfrom my experiences an examiner,how difficultit is to remember the merelynumerical distinction between the first,econd,or third preterites,r the firstand second futures,I have kept as much as possibleto theterminologywith which classical scholars are familiar,callingthe tense corresponding tothe Greek Imperfect,mperfect;that corresponding to the Perfect,ReduplicatedPerfect;that correspondingto the Aorist,Aorist; and the mood corresponding to the Optative,Optative. The names of PeriphrasticPerfect and Periphrasticuture tell their ownstory; and if I have retained the merely numerical distinction between the First andSecond Aorists,it was because this distinction seemed to be more intelligibleo aclassical scholar than the six or seven forms of the so-called multiform Preterite. If itwere possibleo make a change in the established grammatical nomenclature, I shouldmuch preferto call the Fii-stthe Second, and the Second the First Aorist; the formerbeing a secondary and compound, the latter a primary and simple tense. But Firstand Second Aorists have become almost proper names, and will not easilyyieldtheirplace to mere argument.

t In the University of Leipzig alone, as many as fiftypupilsattend every yearthe classes of Professor Brockhaus in order to acquirea knowledge of the elements ofSanskrit,previousto the study of ComparativePhilologyunder Professor Curtius.


5/323

FIEST EDITION. v

or exceptions,contained in other grammars, whether native orEuropean, I have done so after mature consideration, deliberatelypreferringhe less completeto the more complete,but, at the sametime, more bewilderingstatement of the anomalies of the Sanskritlanguaga Thus, to mention one or two cases, when giving therules on the employment of the suflixes vat and mat { 187),I haveleft out the rule that bases ending in m, though the m be precededby other vowels than a, always take vat instead of mat. I did sopartlybecause there are very few bases ending in m, partlybecause,if a word like kimr-vdn should ' occur, it would be easy to discoverthe reason why here too v was preferredto m, viz. in order to avoidthe clashingof two m's. Again, when giving the rules on theformation of denominatives (J495), I passed over, for very muchthe same reason, the prohibitiongivenin P4n. in. i, 8, 3, viz. thatbases ending in m are not allowed to form denominatives. It istrue, no doubt, that the omission of such rules or exceptionsmaybe said to involve an actual misrepresentation,nd that a pupilmight be misled to form such words as kimr-mdn and kim-yati.But this cannot be avoided in an elementary grammar ; andthe student who is likelyto come in contact with such recon-iteforms, will no doubt be sufficientlydvanced to be able toconsult for himself the rules of P^nini and the explanationsof hiscommentators.

My own fear is that, in writingan elementary grammar, I haveerred rather in giving too much than in giving too little. I havetherefore in the table of contents marked with an asterisk allsuch rules as may be safelyleft out in a first course of Sanskritgrammar*, and I have in different places informed the readerwhether certain portionsmight be passed over quickly,or shouldbe carefullyommitted to memory. Here and there,as for instancein 103, a few extracts are introduced from F nini, simply inorder to give the student a foretaste of what he may expect inthe elaborate works of native grammarians, while lists of verbslike those contained in J 332 or J 462 are given, as everybodywiU see, for the sake of reference only. The somewhat elaboratetreatment of the nominal baaes in t and 4, from $ 220 to $ 226,

* In the second edition all these paragraphsare printedin smaller type.


6/323

vi PREFACE TO THEbecame necessary, partlybecause in no grammar had the diflFerentparadigms of this class been correctlygiven, partly because itwas impossibleto bring out clearlythe principleon which thepeculiaritiesnd apparent irregularitiesf these nouns are basedwithout enteringfullyinto the systematicarrangement of nativegrammarians. Of portionslike this I will not say indeed, fia n^ r TaLTty fiaWov ^ fiin^(r Taiut I feel that I may say, ^ ^ ^ ^ ftwflr'ft'W ^: ; ' and I know that those who will take the trouble toexamine the same mass of evidence which I have weighed andexamined, will be the most lenient in their judgment, if hereafterthey should succeed better than I have done, in unravellingtheintricate argumentations of native scholars *.

But while acknowledging my obligationsto the great gram-ariansof India, it would be ungrateful were I not to

acknowledge as fullythe assistance which I have derived fromthe works of European scholars. My first acquaintancewith theelements of Sanskrit was gained from Bopp's grammar. Thoseonly who know the works of his predecessors,of Colebrooke,Carey, Wilkins, and Forster, can appreciatethe advance madeby Bopp in explainingthe difficulties,nd in lightingup, if Imay say so, the dark lanes and alleysof the Sanskrit language.I doubt whether Sanskrit scholarshipwould have flourished asit has, if students had been obligedto learn their grammar fromForster or Colebrooke, and I believe that to Bopp's little grammaris due a great portionof that success which has attended the studyof Sanskrit literature in Germany. Colebrooke, Carey,Wilkins,and Forster worked independently of each other. Each derivedhis information from native teachers and from native grammars.Among these four scholars,Wilkins seems to have been the firstto compose a Sanskrit grammar, for he informs us that the firstprinted sheet of his work was destroyed by fire in 1795. The

* To tbose who have the same faith in the accurate and never swerving argumen-ationsof Sanskrit commentators, it may be a saving of time to be informed that in

the new and very useful edition of the Siddh nta-Kaumudi by S^rt TMnitha-tarkav rchaspatithere are two misprintswhich hopelesslyisturb the order of the rules on theproper declension of nouns in t and il. On page 1 36, 1. 7, read *fl Cnstead of ^pfNl^;this is corrected in the Corrigenda,nd the right readingis found in the old edition.On the same page, 1. 13, insert after f^RT,or joinf^i^mpfHftV^Vi^.


7/323


8/323

viii PREFACE TO THEregard to doubtful or difficultforms, of which there are many inthe grammar of the Sanskrit language, not one of them can beappealed to as an ultimate authority. Every grammar contains,as is well known, a number of forms which occur but rarely,f ever,in the Uterarylanguage. It is necessary, however, for the sake ofsystematiccompleteness,o give these forms ; and if they are to begiven at all,they must be given on competent authority. Now itmight be supposed that a mere reference to any of the numerousgrammars alreadypublished would be sufficient for this purpose,and that the lists of irregularr unusual forms might safelybecopiedfrom their pages. But this is by no means the case. Evenwith regard to regularforms,whoever should trust implicitlyn thecorrectness of any of the grammars, hitherto published,would neverbe certain of having the right form. I do not say this lightly,rwithout being able to produce proofs. When I began to revise mymanuscript grammar which I had composed for my own use manyyears ago, and when on points on which I felt doubtftil, I con-ulted

other grammars, I soon discovered either that,with a strangekind of sequacity,they all repeatedthe same mistake, or that theyvaried widely from each other, without assigning any reason orauthority. I need not say that the grammars which we possessdiffer very much in the degree of their trustworthiness ; but withthe exception of the first volume of Colebrooke and of ProfessorBenfey'slargerSanskrit grammar, it would be impossibleto appealto any of my predecessorsas an authorityon doubtful points.Forster and Carey, who evidently depend almost entirely onmaterials supplied to them by native assistants,give frequentlythe most difficultforms with perfectaccuracy, while they go wildlywrong immediately after, without, it would seem, any power ofcontrollingtheir authorities. The frequent inaccuracies in thegrammars of Wilkins and Wilson have been pointed out byothers ; and however useful these works may have been forpracticalurposes, they were never intended as authorities oncontested pointsof Sanskrit grammar.

Nothing remained in fact,in order to arrive at any satisfactoryresult,but to collate the whole of my grammar, with regard notonly to the irregularut likewise to the regularforms, with P4niniand other native grammarians, and to supplyfor each doubtful case,


9/323

FIBST EDITION. ixand for rules that might seem to differ from those of any of mypredecessors, reference to P4nini or to other native authorities*This I have done, and in so doing I had to re -write nearly thewhole of my grammar ; but though the time and trouble expendedon this work have been considerable,I believe that they have notbeen bestowed in vain. I only regret that I did not give theseauthoritative references throughoutthe whole of my work*, because,even where there cannot be any difference of opinion,ome of myreaders might thus have been saved the time and trouble oflookingthrough F nini to find the Stitras that bear on everyform of the Sanskrit language.

By this process which I have adopted,I believe that on manypointsa more settled and authoritative character haa been impartedto the grammar of Sanskrit than it possessedbefore ; but I do byno means pretend to have arrived on all points at a clear anddefinite view of the meaning of P4nini and his successors. Thegrammaticalsystem of Hindu grammarians is so peculiar,hat ruleswhich we should group together,re scattered about in differentparts of their manuals. We may have the generalnile in the last,and the exceptionsin the first book, and even then we are by nomeans certain that exceptionsto these exceptionsmay not occursomewhere else. I shall give but one instance. There is a rootffP|jdgriiwhich forms its Aorist by adding ^ isham, ^ A,fi( t.Here the simplestrule would be that final ^ ri before ^ ishambecomes ^ r (P n.vi. i, 77). This,however, is preventedby anotherrule which requiresthat final^ fi should take Guna before ^ isha/m(P4n.vn. 3, 84). This would give us wirmfic?iajdgar-isham. Butnow comes another generalrule (P n. vn. 2, i) which prescribesVriddhi of final vowels before ^ isham, i.e. v^uiiM ajdgdrisha/m.Against this change,however, a new rule is cited (P n.vn. 3, 85),and this secures for irnjjdgri a specialexceptionfrom Y^iddhi,andleaves its base again as wn^jdgar. As soon as the base has beenchangedto mn^jdgar,it Mia under a new rule (P n.vn. 2, 3),andis forced to take Vjiddhi,until this rule isagainnullified by PAn. vn.2, 4, which does not allow Vfiddhi in an Aorist that takes interme-iate

^ i,like iiHN(M cydgarisham. There is an exception,owever.* They have been givenin the second edition.

b


10/323

X PKEFACE TO THE

to this rale also,for bases with short i a, beginningand endingwith a consonant, may optionallyake Vriddhi (P n. vn. 2, 7). Thisoptionis afterwards restricted,nd roots with short m a, beginningwith a consonant and ending in ^r, like wx^jdgaVy have no optionleft,but are restricted afresh to Vjiddhi (Pfijgi.ii. 2, 2). However,even this is not yet the final result. Our base ymn^jdgar is afterall not to take Vriddhi, and hence a new specialrule (P4n. vn. 2, 5)settles the pointby granting to WPlJdgri a specialexceptionfromVriddhi,and thereby establishingts Guna. No wonder that thesemanifold changes and chances in the formation of the First Aorist of^OJiJdgrihould have inspireda grammarian, who celebrates themin the followingcouplet:

Guna, Vriddhi, Guna, Vriddhi, prohibition,ption,again Vriddhiand then exception,these, with the change of ri into a semivowelin the first instance,are the nine results.'^

Another difficultyonsists in the want of critical accuracy inthe editions which we possess of Pd^nini,the Siddhinta-Kaumudl,the Laghu-Kaumudl, the SArasvatl, and Vopadeva. Far be it fromme to wish to detract from the merits of native editors, likeDharantdhara, Kdiin tha, T4rd.n^tha, still less from those ofProfessor Boehtlingk,who published his text and notes nearlythirtyyears ago, when few of us were able to read a singleline ofP nini. But during those thirtyyears considerable progress hasbeen made in unravellingthe mysteriesof the grammaticalliteratureof India. The commentary of Sd.yanato the Rig-vedahas shown ushow practicallyo apply the rules of P nini ; and the translation ofthe Laghu-Kaumudl by the late Dr. BaUantyne has enabled evenbeginnersto find their way through the labyrinthof native grammar.The time has come, I believe,for new and critical editions of Pdiiiniand his commentators. A few instances may suffice to show theinsecurityf our ordinaryeditions. The commentary to Pftn. vn. 2,42,as well as the SArasvatl n. 25, i, givesthe Benedictive Atmanepada^^Nif vartshtshta and n6^ startshishta ; yet a reference to P n. viL2, 39 and 40, shows that these forms are impossible.Again, ifP nini (vni.3,92)s right and how could the InfalliUe be wrong?


11/323

FIKST EDITION. xiin using ^npnfNftr agragdmini with a dental n in the last syllable,it is clear that he extends the prohibitiongiven in vin, 4, 34,with regard to Upasargas,to other compounds. It is useless toinquirewhether in doing so he was rightor wrong, for it is an articleof faith with every Hindu grammarian that whatever word is usedby P nini in his SAtras, is eo ipso correct. Otherwise, the rulesaffectingompoimds with Upasargas are by no means identical withthose that affect ordinarycompounds ; and though it may be righttoargue a fortiorifrom wmftfftfpragdmini to ^iirnfMH agragdmini, itwould not be rightto argue from ^Rpvpr agraydna to hutt praydna,this beingnecessarilynmrr praydna. But assuming ^npnftrftfagrongdmini to be correct, it is quiteclear that the compounds laiHifttiflsvargakdminau, yjufaiutrisha^dminaUy ^rmHifti harikdmdni, andgfil^iit arikdmena, given in the commentary to viii. 4, 1 3, are allwrong, though most of them occur not only in the printededitionsof P nini and the Siddh nta-Kaumudi, but may be traced back tothe MSS. of the Prakriy4-Kaumudl,the source, though by no meansthe model, of the Siddhd^nta-Kaumudi. I was glad to learn frommy friend Professor Groldstlicker,ho is preparingan edition of theK Sik4 -Vritti,and whom I consulted on these forms, that the MSS.of Yimana which he possesses, carefrdlyvoid these faultyexamplesto P n. vni. 4, 1 3.

After these explanationsI need hardlyadd that I am not so san-uineas to suppose that I could have escapedscot free where so many

men of superiorknowledge and talent have failed to do so. All I cansay is,that I shall be trulythankful to any scholar who wiU takethe trouble to pointout any mistakes into which I may have fallen ;and I hope that I shall never so far forgetthe regard due to truthas to attempt to representsimplecorrections,touchingthe declensionof nouns or the conjugationof verbs, as matters of opinion,r so farlower the character of true scholarships to appeal,n such matters,from the verdict of the few to the opinionof the many.

Hearing from my friend Professor Btihler that he had finisheda Sanskrit Syntax,based on the works of Pdnini and other nativegrammarians,which will soon be published,I gladlyomitted thatportionof my grammar. The rules on the derivation of nouns, bymeans of Krit, TJn di,and Taddhita suflBxes,do not properlybelongto the sphereof an elementarygrammar. If time and health permit,

b2


12/323

xii PEEFACE TO THE FIRST EDITION.

I hope to publish hereafter, as a separate treatise, the chapter ofthe PrakriyfirKaumudl bearing on this subject.

In the list of verbs which I have given as an Appendix,pp. 244-285, I have chiefly followed the PrakriyA-Kaumudl andthe S rasvati. These grammars do not conjugate every verb thatoccurs in the Dh tup tha, but those only that serve to illustratecertain grammatical rules. Nor do they adopt, like the Siddh nta-Kaumudl, the order of the verbs as given in P nini's Dh4tup tha,but they group the verbs of each class according to their voices,treating together those that take the terminations of the Parasmai-pada, those that take the terminations of the Atmanepada, and,lastly, those that admit of both voices. In each of these subdi-isions,

again, the single verbs are so arranged as best to illustratecertain grammatical rules. In making a new selection amongthe verbs selected by B machandra and Anubh{itisvardp ch iyaI have given a preference to those which occur more frequentlyin Sanskrit literature, and to those which illustrate some pointsof grammar of peculiar interest to the student In this mannerI hope that the Appendix will serve two purposes : it will notonly help the student, when doubtful as to the exact formsof certain verbs, but it will likewise serve as a useful practicalexercise to those who, taking each verb in turn, will try toaccount for the exact forms of its persons, moods, and tenses by areference to the rules of this grammar. In some cases referenceshave been added to guide the student, in others he has to find byhimself the proper warranty for each particular form.

My kind friends Professor Cowell and Professor Kielhom haverevised some of the proof-sheets of my grammar, for which I beg toexpress to them my sincere thanks.

F. MAX MULLER,PARIS,

5th April, 1866.


13/323

PEEFACETO THE SECOND EDITION.

i.HE principalalterations in the new edition of my Sanskrit grammarCQnsist in a number of additional references to F nini^ in all caseswhere an appeal to his authority seemed likelyto be useful, and inthe introduction of the marks of the accent I have also been ableto remove a number of mistakes and misprints which, in spite of allthe care I had taken, had been overlooked in the first edition. Mostof these I had corrected in the German translation of my grammar,published at Leipzig in 1868; some more have now been corrected.I feel most grateful to several of my reviewers for having pointedout these oversights, and most of all to Pandit B j rftmaiiftstrl,whose list of notes and queries to my grammar has been of thegreatest value to me. It seems almost hopeless for a Europeanscholar to acquire that familiarity with the intricate system ofP nini which the Pandits of the old school in India still possess ;and although some of their refinements in the interpretation ofP nini's rules may seem too subtle, yet there can be no doubt thatthese living guides are invaluable to us in exploring the giganticlabyrinthof ancient Sanskrit grammar.

There is,however, one difficultyhich we have to contend with,and which does not exist for them. They keep true throughout toone system, the system of P4nini ; we have to transfer the facts ofthat system into our own system of grammar. What accidents arelikelyto happen during this process I shall try to illustrate byone instance. B j4r mai^trl objects to the form fi^ pun^u as thelocative plural of jin9( pumdn. From his point of view, he is perfectlyright in his objection,for according to P nini the locative plural hasAnusv ra, |f purhsu. But in our own Sanskrit grammars we firsthave a general rule that ^^ is changed to i^^^ after any vowelexcept m and wi d, in spite of intervening Anusvfija (see 100);and it has even been maintained that there is some kind ofphysiologicalreason for such a change. If then, after having laid


14/323

xiv PEEFACE TO THE SECOND EDITION.down this rule,we yet write ^ pvmsu^ we Biinplycommit a gram-atical

blunder ; and I believe there is no Sanskrit grammar, exceptColebrooke's, in which that blunder has not been committed. Inorder to avoid it,I wrote fi^g[punsu, thus, by the retention of thedental ( n, making it grammatically and physically possible forthe ^^ to remain unchanged. It may be objectedthat on the sameground I ought to have written Instr. j^m punsd^ Gen. 5^^ punsdh^c. ; but in these cases the \s i radical,and would therefore not beliable to be changed into 1^sh after a vowel and Anusv ra (PAn. vin.3, 59). Professor Weber had evidently overlooked these simplerules, or he would have been less forward in blaming Dr. Kellerfor having followed my example in writing 511 punsu^ instead of^ pumsu. In P4nini's grammar (as may be seen from my noteappended to $ icx))the rule on the change of ^5 into 1^sh is so care-ully

worded that it just excludes the case of jj purhsu^ although the^su of the loc. plur.is preceded by an Anusv^. I have now, bymaking in my second edition the same reservation in the generalrule, been able to conform to P nini's authority,and have written^ pumsuj instead of fif punsu, though even thus the fact remainsthat if the dot is reallymeant for Anusv ra, and if the ^ ^t^ is thetermination of the locative plural,the ^ s would be sounded as1^ shy according to the general tendency of the ancient Sanskritprommciation.

I have mentioned this one instance in order to show the peculiardifficulties which the writer of a Sandcrit grammar haa to contendwith in trying to combine the technical rules of P nini with themore rational principlesof European grammar ; and I hope it mayconvince my readers, and perhaps even Professor Weber, that whereI have deviated from the ordinaryrules of our European grammars,or where I seem to have placedmyself at variance with some of thenative authorities, I have not done so without having carefullyweighed the advantages of the one against those of the othersystem.

F. MAX MtJLLER.PARKS END, OXFORD,

August, 1870.


15/323

TABLE OF CONTENTS.

CHAPTER I. The AlphabetThe DevanSgari letters

.

I. The Devandgar alphabet .2. Direction of Sanskrit alphabet3. How to write the letters .4. Sounds represented hj the De-

yandgari alphabet5. Number of letters .6. The letter It

.

f Jihydmiilfja and UpadhmSntja8. Signs of nasals and their substi

tute....

9. The three nasal semivowels10. Consonants without correspond

ing nasals.

1 1. Anusv4ra before /, sh, b, hl^i2. Names of letters

13. Vowel signs, initial,medial, andfinal

....

14. Consonants followed by vowels15. Yir ma ....16. Combination of consonants1 7^ The sign for r18. The Virdma used as a stop-gap19. The signa for a pauae .ao. The Avagraha. List of com

pound consonants21. Numerical figures

.

22, Rules of pronunciation .

PAOS

2344

4555

56

667

7777888

89

10

CHAPTER II. Rules of Sandhi. 23. Object and use of Saudhi . .1124. Distinction between External

and Internal Sandhi.

.112$. Classification of vowels, long,

short, protracted.

.12

26. Monophthongs and diphthongs .27. Nasalized vowels28. Light and heavy vowels 29. Acute, grave, and circumflexed

vowels. . . .

30. Guna and Yriddhi .31. Guna of X, d .32. Combination of vowels at the

end and beginning of words.No hiatus

....

33. Vowels meeting the same vowels34. Vowels and i, followed by

di ferent- vowels.

35. Vowels d and d, followed hydiphthongs. .

36. Voweb i,A, ri,followed by dis-imilarvowels

37. Vowels e and o, followed by anyvowel except d .

38. Voweb at and cm, followed byany voweb . . .

39. Treatment of final y and v40. The hiatus occasioned by Sandhi41. Vowels e and 0 before d 42. Unchangeable or Pragrihya

voweb....

43. Irregular Sandhi; prepositionsending in X or ^, followed by6 or o

44. Prepositions ending in X or ^followed hj fi ,

^ 45. The 0 of oahfha^ and otu^^ 46. Irr^^lar compounds:^ 47. The final 0 of indeclinable words:)t 48. Monosyllabic indeclinable words:^ 49. Sandhi of the particle d

PAOI

121213

131313

1313

14

14

15

16

16161717

17

18

181819191919


16/323

XVI TABLE OF CONTENTS.

:^ 5o. Particlee unaffected by Sandhi:^ 51. Protracted vowels unaffected by

Sandhi ....g2. Table showing the combination

of final with initial vowels .53. Combination of final and initial

consonants54. The eleven final consonants55. No word ends in two consonants56. Classification of consonants, ac-ording

to their place57. Classification of consonants, ac-ording

to their quality,.e.contact^ approach,opening .

58. Surd and sonant consonants .59. Aspiratedand unaspiratedon-onants

....

60. Changes of place,and changesof quality. . . .

61. Changes ofplaceaffect Dentals,Anusvdra, and Yisarga

62. Final t before Palatals ch,chh,

63. Final n before /^A, fl,/64. Final t before f, fA, ^, 4K ?

(notah) ,6g. Final n before d^^k^n (notsK)66. Changes of quality67. Final ^ f, ,jp before nasals

^ 68. Final k, f, ,jp before maya ornMra

^ 69. Initial h after final kft,t,p .f o. Final t before I . *f I. Final n before I . .fa. Final , n, n after a short vowel73. Final n before the firsts and

seconds ....74. Final n and n before i,eh, a .75. Final n before i or a (notah) .76. Final ( before ...77. Anusv^ and final m78. M in pcmady and before conso-ants

....

79. Final mheioTehn,hmfhf/,hi,hv80. Sam before fyi,acmakjri81. Sam heiore rdf,aamrdij .8a. Yisargaand final or r


17/323

TABLE OF CONTENTS. xvu

3-114.

116-

116.

* 7-

:yb ii2.Final eyai,o, m changed toay, dy, av, dv; roots endingin diphthongs .

Final consonants, only elevenTwo consonants at the end of

a word impossibleSonant and surd initialsrequire

sonant and surd finalsFinal aspirateslose their aspi-ation

. .

Final gh, dh, dh, hh, followedby tjth,lose their aspirationand change t,ih into dh

^ 118. FmsA gh,dh,dh,bh,foWowedhjdhv, hh, and 8, or final,osetheir aspirationnd throw itback on initialg, d, d,b .

Final chj,jhchangedto ^ or ^Final $h changed to f .Final $h before 8 changedto kFinal sk before t,th,changes

them to t,th ,Final 8h changed to t before

other consonantsFinal^'n certain roots treatedlike 4

. . . .

Final i,chh, Juh, 4ch treatedlike 8h .

Final 4 changed to A; .Final h before 8 treated likeghFinal h treated like gh or dh.Final A optionaUytreated likegh or 4h .

Final h of nah treated like dhFinal 8 changed into t in cer-ainnominal bases

^ 132. Final 8 before 8 changed into tin verbal bases ; 8 dropt be-ore

dhi; optionallyhangedinto t , . . .

^ 133. Final n or m before sibilantschanged to Anusv ra

4t 134. ilTunchangedbefre semivowels:y^ 135. If imchanged before y,r, I .^ 136. If changed to n 137. The five nasals abbreviated

into the Anusvdra dot

FAOI

5353

54

54

55

55

55565656

56

56

57

57575757

5858

58

58

59595959

59

1 38. Annsv ra before i,8h, 8, h .4t 139. i\rafter ch orj changed into H

140. Chh changed to chchh .:)t 141. (7AA before norm changedto/^ 142. Final y and v dropt before

consonants, except y .i/t 143. Final iv,ir,wr lengthened if

followed by consonants% 144. Final vr and ur lengthenedif

ending a word .% 145. Radical if or t^ at the end of

nominal bases lengthened .^ 146, 147. Doubling of consonants

148. Explanationof some gramma-icalterms used by nativegrammarians

FAOI60606060

60

60

60

6061

**^

**

61

CHAPTER m. Declension.49. Gender, number, and case . 6450. I. Bases ending in consonants;

II. bases ending in vowels . 6451. I. Bases ending in consonants;

no bases mn,fi,y . . 6452. Terminations . . .6453. Li. Unchangeable and L 2.Changeable bases . 6554. I.i.Unchangeablebases; u^ais555. S vrva4dk . . .6656. Chitralikh . . . .6657. HarU,agnimcAh,8uhfid,hvdh,

gup, hakvhh . .6758. Jalaanuch . . . .6759. Specialbases in ch; hnmch,

prdfich,vfUch . .6860. Prdchh

....68

61. Ruj, Hrj . . .6862. Bases in j,changeableto ^y

8a/mrdj,vibhrdj,eoej,vUv(j^8iij,par{vrdj,vUvardj,hfijj 68

63. Irregularouns vaj; kkamj,cwaydj .69

64. Bases in r; gvr, vdr,pur,dvdr,kir 70

65. Bases in / A.basesformedbycu, 18, U8; sumcmtM, 8ujyott87166. Jaras ajudjard . . .72

67. Nifjarcaand nvrjara 73c


18/323

XVlll TABLE OF CONTENTS.


19/323

TABLE OF CONTENTS.

^ 227.PAOB

Compounds ending in poly-yllabicfeminine bases in t

and Hy bahuireyast . . 107Strt 108AtisM . . .108Bases in t and u, masc. fem. neut 109Kati

..111

Skikhi IllPad 112Akahl, asthi,dadkif sakihi . 112Bases in fi,masc. fem. neut.,

naptfiypUfi .112Krashfu .113NH 114

IL 2. Bases ending in a and d,kdntah, tdf tarn . .114

Bases in d, masc. and fem.,vUva/pd .... 116

Hdhd 116CHAPTER IV. Adjectives.

241. Declension of adjectives . 116242. Formation of feminine base . 117243. Priyah, fem, priyd .117

:)t 244. Fdchakah, pdchikd .117245. Feminines formed by $ . . 117

^ 246. Exceptionalfeminines in i . 117ij/t247. Irregularfeminines .117;yb 248. Formation of feminine sub-tantives

. ..118

249. Degrees of comparison . . 118250. Tara and tama, how added . 118251. lycM and ishfhafhow added . 118252. Exceptional cpmparativesand

superlatives . .119CHAPTER v. Numerals.

253. Cardinals and declension ofcardinals,eka . . . 120

254. Dvi , . . . - . 124255. Tri, tisfi .... 124256. Chatwr, chaUuri . . . 124257. Fanchan, Bhash,ctahfan . 124

2^ 258. Construction of cardinals 124259. Ordinals . . .125260. Numerical adverbs and other

derivatives. .

.126

PAOICHAPTER VI. Pbohouhs.

261. Personal pronouns . . 127262. Sah, sdytat . . . .128

:yb 263. Syahj syd, tycU . . . 128264. Possessive pronouns . .128265. Reflexive pronouns, waycan . 129266. Atman

. ..129

267. Svahy ivd, svcMn . 129268. Demonstrative pronouns^M^/^,

e8hd,etat . . .129269. Ayamy iyam, idam .129

;|^ 270. nam, endm, eruU .130271. AsaUf cacMy adah . . . ISO272. Tah, yd, yat . 131273. Xah, kd, him . . .131

:i^ 274. Pronouns modified by o^ . 131275. Compound pronouns, f ie^p/ bc.132276. Tdvat c. . ... 132277. KcUchU Ac. .... 132278. Pronominal adjectives,ortw,

vUvtty Ac. .... 133279. Anyah, anyd, cmycU .134280. Ubhau, ubhsy tubhe . .134281. Ubhayah, yt,yam .134

:^ 282. FUrva and itsoptionalforms . 134;yb 283. Fraihar}$a and its optional

nominative plural .134:yb 284. Bvittya and its optionalorms 1352f: 285. Adverbial declension . .135

CHAPTER Vn. CoHJUGATiON.286. Active and passive .137287. Parasmaipadaand Atmanepada 137288. ParasmaipadaandAtmanepada

in derivative verbs . .138289. Passive . . . .138290. The thirteen tenses and moods 138291. Significationf tenses and

moods. . .

.139292. Numbers and persons . .140

CHAPTER Vin. The Ten Classes. 293. Specialand generaltenses, in

the ten classes. . .140

294. Specialor modified,generalrunmodified tenses.

.140295. Division of verbal bases . .141

C 2


20/323

TABLE OF CONTENTS.PAas

296. I. First division; Bhft,Tud,Div, Chur classes . ,141

297. II. Second division,and sub-ivisions.... 142

298. Ha, Sn, Tan, Eri classes . 143299. II 5. Ad, Hu, Rudh classes . 143

CHAPTER IX. ^Augment, Reduplica-ion,AND Terminations.

300. Augment and reduplication. 145301. Augment a . . . 145302. Reduplicationin the perfect,

and in the Hu verbs. .145

303. Qeneral rules of reduplication146304. Aspirated initials . . .146305. Guttural initials . . .146306. Double initials . . . 146307. Initial sibilant followed hj a

tenuis.... 146

308. The vowel of the reduplicativesyllableis short . . .146

309. Medial e and ai are redupli-atedby t, 0 and cmhj u , 146

310. Final e, oi, 0 are reduplicatedby a 146

311. Irregular reduplication bySamprasftrana . .147

312. Short initial a . . . 147313. Initial a followed by two con-onants

. .148314. Initial ri . . . .148315. Short initial i and u . . 148316. Specialrules of reduplication.148

^ 317. Nij,vij,vUh . .148^ 318. MA,hA .148^319. Hcm,hifji,chi , . 148

320. Terminations . . . 149321. Terminations of first and se-ond

divisions. . .149

322. Regular conjugation . .150CHAPTER X. General Tensed

323. General or unmodified tenses . 159324. Reduplicatedperfect . , 169325. Verbs which may form the re-uplicatedperfect . .159326. The periphrasticerfect . 159

PAOS327. Strong and weak termina-ions160

328. Weakening of base . . 160329. Bases ending in d and diph-hongs,how changed . . 161330. Bases ending in t, C,ft, w, ^ fi,

how changed . . .161CHAPTER XL Intermediate i.

331. When it nrnat be omitted,when it may be omitted,when it rrmst be inserted

.162

^332. List of verbs in which the inter-ediatei mitst be omitted . 163

^t 333. Verbs in which the interme-iatei mtist be omitted in

certain tenses. .165

^ 334' Special rules for the redupli-atedperfect . . .167

^ 336* Specialrules for the 2nd pers.sing.Par. of the red. perf. . 167

^ 33^- Table showing when interme-iatei must be omitted

.168

^ 337 Optionalinsertion of t . . 168^ 338. Necessaryinsertion of i .170

339. The intermediate % never liableto GuQa .... 171

340. Insertion of long i . .171^341. Optionalinsertion of long t .171

34a. Periphrasticperfect . . 172^ 343 Periphrasticperfect of inten-

sives and desideratives. 172Paradigms of the reduplicated

perfect .... 172CHAPTER XII. Steengthening and

Weakening.344. Two classes of terminations,

strengtheningor weakeninga verbal base

. .175

^ 345- Specialforms of strengtheningand weakening certain bases 177

.CHAPTER XIIL AoRiST.

346. First and second aorist . . 179347. Four forms of the firstaorist . 179348. Rules for the first form

. .180


21/323

TABLE OF CONTENTS. xxiPAOS^ 349- Stales for desideratives,nten-

Bives,c .... 181350. Bules for the second form . 181351. Terminations bînning with

St or 9th . . . . 181352. Boots in d and diphthongs . 181

^353* Mt,mi,dtyU . . .181^ 364. E(m ' 181^ 355. Gam 181^ 356. 7am 181

357. Bules for the third form . 182ifc358. Mt, mi,ll. . . 182^ 359* 7a^ny ramiy nam . . . 182

360. Bales for the foarth form . 182jyt361. iSUah 1822^ 362. Btth, dihflihygvh 182

Paradigms .... 182363. Second aorist . . . 186364. Boots ending in 4, 0, i,ri; dfii 187365. Boots with penaltimate nasal . 187366. Irrâlar forms . . 187

iff367. Verbs which take the secondaorist

.... 187^ 368. Verbs which take the second

aorist in the Par. only . 188^ 369. The Tan verbs . .188

370. Beduplicatedsecond aorist . 188iff 371. SriydrUf sruy ham; M^ dhe . 189

372. Shorteningof bases ending inay 189

373. Bases that cann6t be shortened 189374. Compensation between base

and redaplicativeyllable . 189375* Vowelsof redaplicativeyllable190

% 31^' Verbs beginning and endingwith doable consonants 190^377. Verbs with penaltimate H, r^ . 190

378. Verbs beginningwith vowels . 191:||lr79. Irregularreduplicatedorist . 191

Paradigm .... 191380. When the diiOTerent forms of

the aorists are used . . 191

CHAPTEB XIV. Future, Conditional,Pesiphsastic Futube, and Benedictivb.381. Future .... 192382. Changes of the base . . 192

PAOS383. Conditional . . . .193384. Periphrasticuture^ . .194385. Benedictive . . . .195

;yb 386. Bases ending in ay . . 195387. Weakening in benedictive Pa-

rasmaipada^ strengtheningnbenedictive Atmanepada . 195

388. Intermediate i . .195^ 389. Weakening of base before y . 196^ 390. Verbs ending in i,u, ft, ft .196:)t391. Verbs ending inn . , 196^ 392. Verbs ending in d. . 196^ 393* êrbs which take Samprasi-

rana 197^ 394. Other verbs which take Sam-

prasirana .... 197^ 395. iSdschanged to Hsh . 197

396. Benedictive Atmanepada . 198CHAPTEB XV. Passive.

397. Atmanepada terminations . 198398. Special tenses of passive . 198

% 399 Causative, denominative, in-ensivebases

. . .198400. Weakening of base. Paradigm 199401. General tenses of passive . 199402. The aorist passive. . . 200403. The 3rdpers.8ing.orist passive 200

^ 404. Aorist of verbs ending in d . 200:yb 405. Aorist of verbs ending in ay . 200;yb 406. Aorist of intensive and desi-

darative bases. . . 200

^ 407. Irregular forms . .201% 408. Verbs ending in am .201

409. Paradigm . . . .201410. Future, conditional,and bene-ictivepassive . . .201

411. Their optional forms . .201:jl412. Aorist passiveof intransitive

verbs.... 203

^413* Optionalforms . . . 203CHAPTEB XVI. Participles, Qbbunds,

AND Infinitive.414. Participlepresent Parasmai-

pada. .... 203415. Participleuture Parasmaipada 204


22/323

zxu TABLE OF CONTENTS.

FAOB416. Participlef reduplicatedper-ectParasmaipada . 204

^ 41'j, Participlef reduplicatedper-ectwith i,

. . .205

418. Participleof reduplicateder-ectAtmanepada . .205

419. ParticipleresentAtmanepada 205420. Participleuture Atmanepada 206421. Participlepresent and future

passive .... 206422. Past participlepassive and

gerund .... 206423. Gerund in tvd . . . 206

:)t424. I. The terminationB tah and^df with intermediate i . 207

4t 425. Penultimate u with optionalGuna

.... 207:)t426. Tvd with intermediate i and

Guna.... 207

4t 427. Tvd with intermediate i andwithout Guna

. . .207:^ 428. Nasal lost before Â,j9A;vancA^

Itmch.... 207

:)^429. II. The terminations tah andtvd,without intermediate i . 207

2^ 430. Final nasal dropt before tahand tvd . . , . 207

:)^ 431. Final n dropt and vowel length-ned; final chh,v, rchh, and rv 20S

:)^432. Boots changing t^ to 1^ . . 2082^ 433. Final at changed to d or i . 2083/c434. Bo, 80, md, athd,dhd, M change

their final into i. , 208

^ 435. JSoand chho take iot d. . 208;|t436. Exceptionalforms . 208^ 437. Verbs which take Samprasi-

rana 208

3/c438. Verbs which lose penultimatenasal

....209

439. Causal verbs . . . 209440. Desiderative yerbs . . 209441. Intensive verbs . . . 209442. Participlesin TiaA . . . 209

;|t443. Adjectivalparticiples . .210444. Vat added to participles . 210445. Gerund in â . . . 210446. Gerund in tya . .210

^ 447. Gerund of causatives .2f^ 448. Ghu verbs, md, sthd,gd,

hd, 80, take final d;yb 449. Verbs ending in nasals^ 450. Verbs ending in rt^ 451- Ve,jyd,vi/e,% 452. Mt, mi, di.It

FAOB211

211211211211211

CHAPTER XVII. Verbal Adjectives.453' Verbal adjectives,JTrifya .211454. Adjectives in tavya . .212455. Adjectivesin aaitya . .212456. Adjectives in â . .212

^ 457. Exceptional verbal adjectivesin ya and tya . . . 214^ 458. Verbs changing final ch and^

into k and g . . . 214459. Infinitive in inim . .214460. Verbal adverbs in a/m . .214

CHAPTER XVIII. Causattvb Verbs.461. Causal bases,how formed . 215

;yb 462. Gupa or V|iddhi . . . 215463. Exceptionalcausative bases,I.

^11 217464. Conjugationof causative verbs 219465. Passive of causative verbs . 219466. General tenses of the passive . 219

CHAPTER XIX. Desiderative Verbs.467. Desiderative bases,how formed 220468. Desiderative bases, how con-ugated

.... 220469. Desiderative bases, with or

without intermediate i.

220^ 470. Strengtheningof base . . 220;yb 471. Exceptional strengtheningor

weakening. . . . 220:^ 472. Desiderative bases,treated as

BhA verbs.. .

.221473. Reduplicationof desiderative

bases.... 222

if^474. Bases in av and dv . . 222^ 476 Sru, iru, dm, pru, plu, ckyu . 222

476. Internal reduplication . .222:)^477. Exceptional forms .222


23/323

TABLE OF CONTENTS. XXlllPAGE

CHAPTER XX. Intbsksivb Vbbbs.478. Meaning of intensive or fre-uentative

verbs .223479. Verbs which may form inten-ive

bases .... 223480. Two kinds of intensive bases.

Atmanepada . . .223481. Intensive bases in ya, how

formed and conjugated .223482. Parasmaipadabases,how form-d

and conjugated . .224483. Conjugation of Parasmaipada

bases....

224484. Beduplicationf intensive bases 224

if^485. Verbs which insert nt . . 225ifc486. Verbs ending in nasals . .225iy^487. Jap, jabh, dah, dcmUy hharij,

pai 225He 488. Char, phal .... 225ifc489. Verbs with penultimate ri . 225^ 490. Verbs ending in ft .226:^ 491. Exceptional inten3ive bases . 226

492. Secondary and tertiarybases . 226

CHAPTER XXI. DbnominativbVebbs.493* Character of denominative

verbs.... 227

494, 495. Denominatives in ya, Pa-asmaipada. .

227;yb 496. Changes of base . . . 227

497. Denominatives in ya, Atmane-pada 228

% 498. The Kandvfidi verbs . . 228:yb 499. Denominatives in sya . .229ifc500. Denominatives in kdmya . 229

501. Conjugationof denominatives 229s/^ 502. Denominatives in aya . . 229^ 503. Denominatives without affixes 230

CHAPTER XXIL Pbbpositions andParticles.

504. Prepositions,paaon-ga . 230505. Prepositions,ati, . . 230506. Prepositions,armapravacha

niya. .... 231507. Adverbs . . . .231

508. Conjunctions509. Interjections

FAOl. 233. 233

CHAPTER XXIIL Compound Words.510. Manner of compounding nomi-al

bases.... 233

^511. Treatment of feminine bases . 234512. Six classes of compounds . 234

I. Tatpurusha, determinativecompounds . . . 234

lb, Earmadhdraya, apposi-tional determinative com-ounds

-. 234

Ic. Dvigu, numeral deter-inativecompounds . 234

II. Dvandva, collective comp. . 235III. Bahuvrihi, possessive com-ounds

. .235

IV. Avyayibhdva,adverbialcom-pounds . . 235

513. I. Determinative compounds . 235;)t 514. Exceptionaldeterminative

compounds . . .237i)^515. Inverted determinative com-ounds

.... 237lt516. Determinative compounds end-ng

in verbal bases .237517. Ih. Appositionaleterminative

compounds . . 237:yb 518. Inverted determinative com-ounds

. . ..238

519. Ic. Numeral determinativecompounds . .238

:)^520. Modifications of the final lettersof determinative compounds 238

521. 11. Collective compounds, Ita-retara and Sam^ira

. . 2402ft:522. Precedence of words . . 240:)^ 523. Nouns ending in ri . . 240:yb 524. Names of deities c. . . 240:it 525. Modifications of the finalletters

o^ collective compounds inthe singular . .241

:)^ 526. Idiomatic expressions . .241527. III. Possessive compounds . 241

:yb 528. Modifications of the finallettersof possessivecompounds . 241


24/323

XXIV TABLE OF CONTENTS.

PAOS

242529. lY. Adverbial compounds4t 530. Exceptional compounds

.243

i)^ 531. Modifications of the final lettersof adverbial compounds

.243

APPENDIX I

epada

epada

List of Verbs.

Bhii Class (Bhv di, I Class)L Parasmaipada Verbs

IL Atmanepada Verbsin. Parasmaipada and Atman

VerbsTud Class (TudSdi, VI Class)

L Parasmaipada and AtmanVerbs

IL Parasmaipada Verbsm. Atmanepada Verbs

Div Class (Dividi, IV Class)I. Parasmaipada Verbs

IL Atmanepada VerbsIII. Parasmaipada and Atmanepada

VerbsChur Class (Churddi, X Class)

Parasmaipada Verbs onlySu Class (SvSdi, V Class)

I. Parasmaipada and AtmanepadaVerbs

IL Parasmaipada VerbsIII. Itmanepada Verbs

PAGB

244-285

245245260

263265

265266267267267269

. 269.

270

.270

.270

270271271

FAasTan Class (TanvSdi, VILE Class)

.272

Parasmaipada and AtmanepadaVerbs

. . ..272

Eri Class (Kry^, IX Class). .

273I. Parasmaipada and Atmanepada

Verbs. . .

.273IL Parasmaipada Verbs

. .274

III. Atmanepada Verbs.

.274Ad Class (Ad^, II Class)

. .275

I. Parasmaipada Verbs .275IL Atmanepada Verbs

. .279

III. Parasmaipada and AtmanepadaVerbs

. . .

.280

Hu Class (Juhoty di, LEI Class).

281L Parasmaipada Verbs

.

.281II. Atmanepada Verbs

. .

282III. Parasmaipada and Atmanepada

Verbs. . .

.283Rudh Class (Rudh di, VII Class)

.

284I. Parasmaipada and Atmanepada

Verbs. . .

.284n. Parasmaipada Verbs

. .284

nL Atmanepada Verbs. .

285

APPENDIX ILPAQB

Ok the Accent in Sanskjkit.

286-292

Index of Notjnb

Index of Vebbs

293-297

.297-300


25/323

SANSKRIT GRAMMAR.


26/323

THE DEVANiGARt LETTERS.

^ Sometunes represented in the Veda by ^S SS* \ ( ' 0*3 Sometimes represented in the Veda hy oSS^ 59^9 1^ (o^ ^)'


27/323

CHAPTER I.

THE ALPHABET.

I. Sanskrit is properly written with the Devanftgarialphabet ; but theBengali, Grantha, Telugu, and other modem Indian alphabets are commonlyemployed for writing Sanskrit in their respective provinces*

Note Devandgari means the Ndgari of the gods, or, possibly,of the Br hmans. A moreouirent style of writing, used by Hindus in all common transactions where Hindi is the lan-uage

employed, is called simply Ndgari. Why the alphabet should have been called Ndgari,is unknown. If derived from nagara, city,it might mean the art of writing as first practisedin cities. (P 9. iv. 2, 128.) No authority has yet been adduced from any ancient author forthe employment of tlieword Devandgari, In the Lalita-vistara (a life of Buddha, translatedfrom Sanskrit into Chinese 76 a. d.), where a list of alphabets is given, the Devandgari \anot mentioned, unless it be intended by the Deva alphabet. (See History of AncientSanskrit literature^ p. 518.) Albiruni, in the nth century, speaks of the Nagara alphabetas current in Malva. (Reinaud, M^oire sur I'Inde, p. 298.)

Beghr m {bhagdrdma, abode of the gods) is the native name of one or more of the mostimportant cities founded by the Greeks^ such as Alexandria ad Caucasum or Nicsea. (SeeMason's Memoirs in Prinsep's Antiquities, ed. Thomas, vol. i. pp. 344-350.) Ck uMDevanftgart have been meant as an equivalent of BeghrlUni ?

No inscriptionshave been met with in India anterior to the rise of Buddhism. Theearliest authentic specimens of writing are the inscriptionsof king Triyadarii or A4oha, about

-^'250.C. These are written in two different alphabets. The alphabet which is found in theinscriptionof Kapurdigiri, and which in the main is the same as that of the Arianian coins,is written frt m right to left. It is clearlyof Semitic origin, and most closelyconnected withthe Aramaic branch of the old Semitic or Phenician alphabet. The Aramaic letters,how-ver,

which we know frt)m Egyptian and Palmyrenian inscriptions,ave experienced furtherchanges since they served as the model for the alphabet of Kapurdigiri, and we must haverecourse to the more primitive types of the ancient Hebrew coins and of the Phenicianinscriptionsin order to explain some of the letters of the Kapurdigiri alphabet.

But while the transition of the Semitic types into this ancient Indian alphabet can beproved with scientific precision,the second Indian alphabet,that which is found in theinscriptionof Gimar, and which is the real source of all other Indian alphabets, as weU as ofthose of Tibet and Burmah, has not as yet been traced back in a satisfactoiyanner to anySemitic prototype. (Prinsep's Indian Antiquities by Thomas, vol. 11. p. 42.) To admit,however, the independent invention of a native Indian alphabet is impossible. Alphabets werenever Invented, in the usual sense of that word. They were formed gradually, and purelyphonetic alphabets always point back to earlier,syllabicor ideographic, stages. There are nosuch traces of the growth of an alphabet on Indian soil; and it is to be hoped that newdiscoveries may still bring to light the intermediate links by which the alphabet of Gimar,and through it the modem Devan gart, may be connected with one of the leading Semiticalphabets.

B 2


28/323

THE ALPHABET. 2-^ 2. Sanskrit is written from left to right.Note Samskrita (li^il)eans what is rendered fit or perfect. But Sanskrit is not called

so because the BriLhmans, or still less,because the first Europeans who became acquaintedwith it,considered it the most perfectof all languages. Samskrita meant what is renderedfitfor sacred purposes ; hence purified,acred. A vessel that is purified, sacrificialvictimthat is properlydressed,a man who has passedthrough all the initiatoryites or samskdrasjaU these are called samskjrita. Hence the language which alone was fit for sacred acts, theancient idiom of the Vedas, was called Samskrita, or the sacred language. The local spokendialects received the generalname of prdkfita. This did not mean originallyulgar,butderived, secondary,second-rate,literallywhat has a source or type/ this source or type(prakfiti)eing the Samskfita or sacred language. (See Vararuchi's Pr krita-Prak sa, ed.Cowell, p. xvii.)

The former explanationofprdkritain the sense of the natural,originalontinuations ofthe old language {hhdshd),'is untenable, because it interpolatesthe idea of continuation.If prdkfitahad to be taken in the sense of ' originalnd natxiral,' language so called wouldmean, as has been well shown by D'Alwis (An Introduction to Kachch yana'sGrammar,p. Ixxxiz),he originallanguage,and samskrita would then have to be taken in the sense ofrefined for literaryurposes.' This view, however, of the meaning of these two names, isopposed to the view of those who ^med the names, and is rendered impossibleby thecharacter of the Vedic language.

^ 3. In writing the Devan gari alphabet,the distinctive portion of eachletter is written first,hen the perpendicular and lastlythe horizontal line.Ex. , 4, W*/ ^, 9 ^kh; I, *i, ilff; , M, ^ffh; ^, V^i, c.

Beginners will find it useful to trace the letters on transparent paper, tillthey know them weU, and can write them fluentlyand correctly.

4. The followingre the sounds which are representedin the Devanagarialphabet:

Unmodified Nasal or Anusv ra, * m or ^ iJi.Unmodified Sibilant or Yisarga,: i^.

^ In the Veda 7 4 and 7 4^, if between two vowels, are in certain schools written35 / and oS^ fh.

^ 7 A is not properly a liquid,ut a soft breathing,' 7 1? is sometimes called Pento-labial.* The signsfor the gutturaland labial sibilants have become obsolete, and are replaced

hy the two dots t J.


29/323

-$ 8. THB ALPHABET. 5Students should be cautioned againstusing the Roman letters instead of

the Devan gari when beginning to learn Sanskrit The paradigms shouldbe impressed on the memory in their real and native form, otherwise theirfirst impressions will become unsettled and indistinct. After some progresshas been made in mastering the grammar and in readingSanskrit,the Romanalphabet may be used safelyand with advantage.^ 5. There are fiftyletters in the Devanagari alphabet, thirty-seven

^consonants and thirteen vowels, representingeveiy soimd of the Sanskritlanguage.

^ 6. One letter the long i{ ft,is merely a gnunmatical invention ; itnever occurs in the spoken language.

^ 7* Two sounds, the guttural and labial sibilants^re now withoutdistinctive representativesin the DevanSgari alphabet. They are calledJihvdmUliya,the tongue-root sibilant,formed near the base of the tongue ;and Vpadhmdniya^ i.e. afflandus,the labial sibilant. They are said to havebeen representedby the signs X (calledVajrdkjitiâving the shape of thethunderbolt) and % (calledGajakumbhdkriti^ having the shape of anelephant'stwo firontal bones). [SeeYopadeva's Sanskrit Grammar, i. 18 ;Histoiy of Ancient Sanskrit Literature,p. 508.] Sometimes the sign X called Ardha-visarga, half-Visarga,is used for both. But in commonwriting these two signs are now replaced by the two dots, the Dvivindu, :,{dvi two, vinduj dot,)properly the sign of the unmodified Visarga. Theold sign of the Visarga is described in the Katantra as like the figured 4 ;in the Tantr bhidhftna as like two 7 fA's. (SeePrinsep,Indian Antiquities,vol. T. p. 75.)^ 8. There are five distinct letters for the five nasals^^'f^ ^n,Janf^nî(^97 ,s there were originallyive distinct signsfor the five sibilants. When,in the middle of words^ these nasals are followed by consonants of their ownclass,{h by *, *A, ^, gh ; n by ch, chh, j\jh ; n by /, /A, ^, 4h ; n by t, th^dy dh ; m by p^ ph^ by bhy)they are often,for the sake of more expeditiouswriting,replacedby the dot, which is properly the sign of the unmodifiednasal or Anusvfira. Thus we find

iffvilTinstead of ^rf^piTnkitd.ivf^HT instead of ^rfvWT anchitd,jAnrrinstead of ^fÎTTkuiiditd.fff^TTnstead of ^ffi^^nanditd.^ftniT instead of ivfînrrkampiid.

The pronunciation remains unaffected by this styleof writing, itf^nrrmust be pronounced as if it were written ^rf^piTitkitOyc.

The same applies to final n m at the end of a sentence. This too,


30/323

6 THE ALPHABET. J 9-though firequentljritten and printed with the dot above the lineîs tobe pronounced as H m. ^^9 1îs to be pronounced ^v^ aham. (SeePrefaceto Hitopadesa,in M. M/s Handbooks for the Study of Sanskrit,p. viiL)

Note According to the KaumlLras final ^ m tn pausd may be pronounced aa Anusvfira;cf. Sarasvatt-Prakriyll,d. Bombay, 1839*, pp. la and 13. W\ IlVI^TOm 5 ^^Hi Xi^HM^H I^^Rn% ^ I WTOT% V^R^T }^li) fw ^^. I ^ T ^^'^H The Kaumftraa are thefollowers of Kumftra^ the reputed author of the K tantra or Kalipa grammar. (SeeColebrooke, Sanskrit Grammar, Preface ; and page 315, note.) S'ftrvavarman is quoted bymistake as the author of this grammar, and a distinction is sometimes made between theKaum ras and the followers of the Kalllpagrammar.

^ 9. Besides the five nasal letters,expressingthe nasal soimd as modifiedby guttural palatal,ingual dental,and labial pronunciation,there are stillthree nasalized letters,the ^^ 7^, , or t^,, ^, y, 7,v, which are used torepresent a final 1^972,f followed by an initial ^^9 c^ /,^t;, and modifiedby the pronunciationof these three semivowels. (Pa^, viii. 4, 59.)

Thus instead of TT infirtaTh ydti we may write nxqrfirtay ydti;instead of w cwi tarn labhate we may write l^k^ ted labhate ;instead of 7 ^^flTtarn vahati we may write ll^^flcav vahati.

Or in composition^'îjlffamydnam or m6\\'Asayydnam;Îb^ samlabdham or ^r^Tfsallabdham;HH^rm sathvahati or ^rl^^fwavvahatu

But never if the 1^ m stands in the body of a word, such as 1KI9^ kdmyah nor if the semivowel represents an originalvowel, e.g. Rig-veda x. 13a, 3.^ ^ 'WR'^^sam u drauy changed to ||%|K ^8am vdran.^ 10. llie only consonants which have no corresponding nasals are t r,^ ,T^Â, ^ , 7 A. A final i^ , therefore,before any of these letters atthe beginning of words, can only be representedby the neutral or immodi -fied nasal,the Anusv^.

A xs(hitarn rakshati. Or in composition, t^firsamrakshati,t ^mftfirarn Srinoti. ^BWlfrfiramiriw î*W ^nirt tarn shakdram, ^Whlfw samshthivati.a ^Rfir tath sarati. ?mflr sarhsarati.a fxflrtarn harati. llîOlsamharati.

J II. In the body of a word the only letters which can be preceded by* This edition,which has latelybeen reprinted,ontains the text ascribed either to Vi^l

herself,.e.Sarasvatt, the goddess of speech(MS. Bodl. 386),ortoAnubhAti-svarApa-Ach rya,whoever that may be and a commentary. The commentary printedin the Bombay editions iscaUed ^tl^, or in MS. Bodl. 383. thfPa .e.î^h^nft.n MS. Bodl. 382. Mahtdhara orMahidftsabhatta is said to have written the Sdrasvata in order that his children might read it,and to please fs'a,he Lord. The date given is 1634,the place Benares, (Sivarljadhant.)


31/323

-$ l6. THE ALPHABET, 7Anusvftra are îJ'^sh, s^ ^ h. Thus wy: athia^,v^^ dhandmshi,if^^tftraid^m,tifp simha^. Before the semivowels i^y, \r,T9l9\v^the i m, in the bodyof a word, is never changed into Anusvfira. Thus '^m^ gamyatey tf^ namral^^hit: amlal^. In iîfhiamyoh (Rv. i. 43, 4, c.)the rh stands 'padlnte/butnot in ^rnrflridmyaii, (See 9.)^12. With the exceptionof Jihvdmitltya ;((tongue*rootletter),padh*mdntya X f{tohe breathed upon),Anusvdra * fh (after-sound),isarga : i^(emission,ee Taitt.-Br^m. iii. p. 23 a),and Bepha r (burring),ll lettersare named in Sanskrit by adding kdra (making) to their sounds. ThusV a is called WîlTt cAdraff. w ka, 'WWTKl kakdrahy c.

13. The vowels, if initial,re written,^, ^5 x %y ^1 ^K ^ (^ '. ^f ^f ^, ^ * ^ ;a, 4, , r, ri K / (/0 * * ^ ^ ;if they foUow a consonant, they are written with the foUowing signs-

a, a, f, /,fi,rU Ih {lf)y dj . * o 'There is one exception. If the vowel ^ ft follows the consonant ^ r, itretains its initial form, and the r is written over it. Ex. f^^fln nirfiii^.

In certain words which tolerate an hiatus in the body of the word, thesecond vowel is written in its initial form. Ex. nt^RQ ffoagra, adj.precededby cows, instead of ntsu go 'gra or T^lfr gavdgra ; ^ftil goaivam, cows andhorses ; Tfm praiigayyoke ; f)nr9 iUaii,sieve.^ 14. Every consonant, if written by itself,s supposed to be followed bya short a. Thus ii is not pronounced k,but ka ; iK not y, but ya. But li kor any other consonant, if followed by any vowel except a, is pronouncedwithout the inherent a. Thus

-mkdy fmki, irtHf *r ,^ *r/, f *? ,(]fo, f * , 7| w, **î%*fl ,ij^koy 4t Aatf .

The only peculiaritys that short f t is apparentlywritten before the con-onantafter which it is sounded. This arose from the fact that in theearliest forms of the Indian alphabet the long and short V% were bothwritten over the consonant, the short $ incliningto the left,the long %incliningto the right Afterwards these top-marks were, for the sake ofdistinctness,drawn across the top-line,o as to become f% and iA, insteadof % and 4. (SeePrinsep'sIndian Antiquities,d. Thomas, vol. 11. p. 40.){ 15. If a consonant is to be pronounced without any vowel after it,theconsonant is said to be followed by Virdma^ i.e. stoppage, which is markedby ^^. Thus ak must be written W| ; kar^ nr^ ; iky^.j 16. If a consonant is followed immediately by another consonant, thetwo or three or four or five or more consonants are written in one group


32/323

8 THE ALPHABET. 17-{samyoga). Thus atka is written ^cm i alpa is written 'W^ ; kdrtmya iswritten iEn^ These groups or compound consonants must be learnt hjpractice. It is easy, however, to discover some general laws in their forma*tion. Thus the perpendicularand horizontal lines are generallydropt inone of the letters: ^ + ^=' **a; W + ^ = ^ rffl/ i^4-^=W/va;. ^+5irmktva; ^^h\+'^ = m ^^y^'^17. The t r following a consonant is written by a short transversestroke at the foot of the letter; as 1(^ ^=111 or '%kra; T + T = gÂ;1^+T:=?r orir/ra; ^ + T:='5rfra; H^+T + x:=I Â^ra.

The ^ r preceding a consonant is written by *^ placed at the top of theconsonant be re which it is to be sounded* Thus ^+ 9= V% arka ; ^+i^+iT=:^r arshma. This sign for ^r is placed to the right of any othermarks at the top of the same letter* Ex. ^ arkam ; iR^^r arkena ; ^sXxarkendH.

^ k followed by i^ sh is written t^ or ^ ksha,i^j followed by If ^ is written Tfjna.^jh is sometimes written ff^jh.^ r followed by 7 ti and 9 1 is written ^ ni, ^fH.Z d followed by 7 1 and if^His written 7 du, \^^'lî,particularlyn combination with other letters,s frequentlywritten '^

^18. The sign of Virdma ^ (stoppage),hich if placed at the foot of aconsonant, shows that its inherent short a is stopped,is sometimes, when itis difficultto write (orto print)two or three consonants in one group, placedafler one of the consonants : thus W^ instead of ^ yunkte.

^19. The proper use of the Yir ma, however, is at the end of a sentence,or portion of a sentence, the last word of which ends in a consonant.

At the end of a sentence, or of a half-verse,the sign \ is used ; at theend of a verse, or of a longer sentence, the sign 11.{ %o. The sign s (Avagraha or Arddhdkdra) is used in most editions tomark the elision of an initial V a, afler a final ^ V0 or ute. Ex. ^sf^ so 'pifor irt^rfl?0 apiy i.e. H^ irf^laa apt ; irsftlte 'pifor i^mfq te api.

List of Compound Consonants.II *-*a, mr k-kha, nm k'Chay is k-ta^ ^ k-t-ya, ^ k-t-ra, ^ k-t-r-ya,

K k-t-va,n k-na, ^ k-nrya^ W k-ma, m k-ya, n or n k-ra^ gsf orifSf k-r-ya^9 *-to, V k'Vaj fsr k-v-yGf ^ k-sha^ W k-sh-ma^ ^ k-sh-yay w k-sh-va ; W kh-yay ?f khâ; nj g-ya, u g-ra, ijf g-r-ya ; K gh-na, 9 gh-n-ya,TR gh-ma^ Tflrgh-ya^ ir gh-ra ; ^1 -* , |f w-*-/a, |p[ n-k-t-ya, ^ ii^k-ya,


33/323

THE ALPHABET.

^^ ^ ^ ^9

nr 9-ya, ^ 9-tâ.

T ^dOy II cT-dAa, d-bhryaîr f-ma, v f-ya ^ â, S d^-Ô'yT ^-^îir rf-t^-ya M dh-nOf iffdhr-nrya,nr dh-ma^ m dh-ya, H rfA-ra, 9 dh-r-â^9 dhâ; ^ n-/a, iw n-tâ^ w n-t-ra, ^ nwfo, Hf n^d-ra, ^ n-dha,^ n-dhrtaj w - , ^ â, 5p n-jHray 'm n-ma, a| n-yo, ^ n-ra, iir -â.

H p-ta^ m p-t-ya, jf p-nCy tq ji-pa, ^ j7-ma, p-ya, m p-ra, jf p4apvf p-va^ ^ p^Qy 1^ p-a-va ; âi b-gha^ m Ihfa, ^ 4-da, v bÂOf |f 4-na,V d-ia^ ^ ft-Ma, 99 b-bhryOjv| d-ya, n â, V A-va ; libh'-nayv^ bhrya^V iA-ra^ )^ bhrva; ^V m-mi^ iT m-â,' IV m-pâ, if m-ia, nr m^bha^m m-ma, vq m-ya, ^ m-ra^ j[ tih-la,f m-va.

l y-ya, ^f y- a ; ^9 l^ka,W l-pa, m IrmOy -gi ^ya, V ^to, ^ f-ra ;U tMta, | t^â, tf v-ra, if tM?a.

^ i-choy VI i-ch-ya, v ^na, ^ â, W i-ra ^ i-r-yOy ^Ja^ iff â| f-v-â, 9 -ia ; ^? A-â, w sh-f-ya^ 1 sh-t-ra, w sh-f-r-Ôyf sh-f-â,

W ih-thay W sh-^y iqn sh-n-yay tq A-â, ig sh-p-ray ^ M-ma^ v| Ââ,H Aâ '~r:?? *:*S * *-*Aa, ii| #-/a, iRq ^/-ya, ^ ^/-ra, ^ ^/-ra,W s-thay m s-noy W( *- -ya, W *â, n #Âa, 9 s-ma, W{ s^m-yay ^ -ya,

Ci ^*9^ K ^ ^^^ V A-mo, W A-ya, f A-ra, n A-fo, 3( A-i;a.

Numerical Figures.^ 21. The numerical figuresin Sanskrit are

1:134567890These figuresere originallybbreviations of the initial letters of the Sanskrit numerals.The Aiabs, who adopted them from the Hindus, caUed them Indian figures in Europe,where they were introduced by the Arabs, they were caUed Arabic figures.

c


34/323

10 THE ALPHABET. J 2*Thus ^ stands for^ e of ^Wt ekah, one.

^ stands for W dv ai ^ dvau, two.^ stands for I' tr of IRt trayaft,three.d stands for ^ cA of î IIO chatvdrah, four.M stands for î? of '^^ paHcha, five.

The similarityecomes more evident by comparing the letters and numerals as used inancient inscriptions.See Woepcke, ' M^moire sur la Propagation des Chif es Indiens/in Journal Asiatique,i s^rie,tome i ; Prinsep's Indian Antiquitiesby Thomas, vol. ii.p. 70; Chips from a German Workshop, vol. 11. p. 289.

Pronunciation. %%. The Sanskrit letters should be pronounced in accordance with the

transcriptiongiven page 4. The following rules, however, are to beobserved :I. The vowels should be pronounced like the vowels In Italian. The short W a,

however, has rather the sound of the English a in * America/%. The aspirationof the consonants should be heard distinctly.Thus W Kh

is said,by EngUsh scholars who have learnt Sanskrit in India, to soundalmost like kh in 'inkhom ;' ^ th like th in * pothouse,- ph like/Âin ^ topheavy ;' ^ gh Uke gh in ' loghouse ;^ V dh Uke dh in* madhouse ;' ^ bh like hh in * Hobhouse.' This, no doubt, is asomewhat exaggerated description,ut it is well in learning Sanskritto distinguishfrom the first the aspiratedfrom the unaspirated lettersby pronouncing the former with an unmistakable emphasis.

3. The gutturall? it has the sound of n^ in ' king.'4. The palatalletters ^ ch and ^j have the sound of ch in 'church' and

of J in 'join/5. The lingual letters are said to be pronoimced by bringing the lower

surface of the tongue againstthe roof of the palate. As a matter offact the ordinarypronunciation of /, f,n in English is what Hinduswould call lingual,nd it is essential to distinguishhe Sanskrit dentalsby bringing the tip of the tongue against the very edge of the upperfront-teeth. In transcribing EngUsh words the natives naturallyrepresent the English dentals by their Unguals, not by their owndentals ; e. g. OfX^^TiirekXaT Tn$f?| avarnmet^ti c. *

6. The Yisaiâ, Jihvdmdliya and Upadhmdniya are not now articulatedaudibly.

7. The dental ^ 8 sounds Uke s in ' sin,'the Ungual 19 sh Uke sh in ' shim,'the palatal i like ss in ' session.'

* Biihler,Madras LiteraryJournal, February,1864. RigendralalMitra, ' On the Originof the Hindvi Language,' Journal of the Asiatic Society,Bengal, 1864, p. 509.


35/323

-j 24. RULES OF SANDHL IIThe real Anusv ra is sounded as a very slightnasal,like n in French ^ bon/

If the dot is used as a graphic sign in placeof the other five nasals itmust, of course, be pronounced like the nasal which it represents*

CHAPTER II.KULBS OF SANDHI OR THE COMBINATION OF LETTERS.

^ 23. In Sanskrit every sentence is considered as one unbroken chainof syllables.Except where there is a stop, which we should mark byinterpunction,he final letters of each word are made to coalesce with theinitial letters of the followingword. This coalescence of final and initialletters,(of vowels with vowels, of consonants with consonants, and ofconsonants with vowels,)s called Sandhi.

As certain letters in Sanskrit are incompatiblewith each other, i e.cannot be pronounced one immediately after the other,they have to bemodified or assimilated in order to facilitate their pronunciation. The rules,according to which either one or both letters are thus modified, are calledthe rules of Sandhu

As accordingto a generalrule the words in a sentence must thus be gluedtogether,he mere absence of Sandhi is in many cases sufficient to mark thestops which in other languageshave to be marked by stops. Ex. VMO^illflPT^^g^ ^^PTT 'nnnT: astvagnimdhdtmyam^ indrastu devdndm mahattanui^, Letthere be the greatness of Agni ; nevertheless Indra is the greatest of thegods.

JHstinciion between External and Internal Sandhi.^ 24. It is essential,in order to avoid confusion, to distinguishetween

the rules of Sandhi which determine the changes of final and initial lettersof words (pada) and between those other rules of Sandhi which apply tothe final letters of verbal roots (dhdtu)nd nominal bases (prdtipadika)henfollowed by certain terminations or suffixes. Though both are based on thesame phoneticprinciplesand are sometimes identical,their applicationisdifferent. For shortness^ sake it will be best to apply the name of External

* Accordingto Sanskrit grammarians the real Anusv ra is pronounced in the nose only,the 'five nasals by their respective organs and the nose. Siddh.-Kaum. to P n. i. i, 9.The real Anusyftra is therefore ndtikya,nasal; the five nasals are anuttdsikaynasalized,i.e. pronounced by their own organ of speech and uttered through the nose.

c 2


36/323

12 RULES OF EXTERNAL SANDHI. *5-Sandhi or Pada Sandhi to the changes which take place at the meeting offinal and initial letters of words, and that of Internal Sandhi to the changesproduced by the meeting of radical and formative elements.

The rules which apply to final and initialletters of words {pada)apply,withfew exceptions,o the final and initial letters of the component parts of com-poimds, and likewise to the final letters of nominal hsu^s (prdtipadika)hen fol-owed

by the so-called Pa^fa-terminations (hitbhydm, fin bhiJ^ n bhya^, ^ su)ôr by secondary(taddhita)uffixes beginning with any consonants except i^y.

The changes produced by the contact of incompatibleletters in the bodyof a word should properly be treated under the heads of declension,conjugation,and derivation. In many cases it is far easier to rememberthe words ready-made fi*om the dictionary,r the grammatical paradigmsfirom the grammar, than to acquire the compHcated rules with theirnumerous exceptions which are generallydetailed in Sanskrit grammarsimder the head of Sandhi. It is easier to learn that the participlepassiveof f^ lih,to lick,is cÂt: IHha^^ than to remember the rules according towhichiE + TÂ+ are changed into ^-hWÂ + /,^+^ / + rfA,and ^+^^ + ^;^^ is dropt and the vowel lengthened : while in îftf^ IK parivfih + ta^,the vowel, under the same circumstances, remains short ; parivfih + tai =parivri4h + tai^ypanvfi4+dJuiijt:=parivri4-^4âi^=:paHvri4hai, In Greekand Latin no rules are given with regard to changes of this kind. If theyare to be given at all in Sanskrit grammars, they should, to avoid confiision,^be kept perfectlydistinct from the rules afiectinghe final and initialletters*of words as brought togetherin one and the same sentence.

Classificationof Vowels.^ 25. Vowels are divided into short (hrasva),ong {dirgha\and protracted

{pluta)vowels. Short vowels have one measure {mdira)ylong vowels two,protractedvowels three. (P 9. i. 2y 27.) A consonant is said to last halfthe time of a short vowel.I. Short vowels : V ^9 ^ t, 7 ti, ^ ft, 7l( li.%. Long vowels : W a, ^ ^,is tl, f^,1^ e, ^ ai, ^ o, ^ au.3. Protracted vowels are indicated by the figure1 3 ; v| ^3y^^\^3iW^3y

%\^ i* ^\^3i^\^'î' Sometimes we find V ^ ^, a 3 i,instead of^ % ^ 3 ; or ^ ^ 7, 4 3 1 , instead of ^ |, au 3.

^ %6* Vowels are likewise divided into1. Monophthongs {samdndkshara) V a, ^ a, i^t, f,W, 9 12, fj, ff, l^ li.2. Diphthongs {iandhyakshara) 1^ ^ ^ ai, ^ 0, ^ au.^ 27. AU vowels are liable to be nasalized,r to become anundsika : ^2,^d.


37/323

- 33- BULBS OF EXTERNAL SANDHI. 13 28. Vowels are again divided into light{laghu)and heavy {guru). This

division is important for metrical purposes only.I, Light vowels are Vo, l^t^^tiy^ft^Tif/tîfot followed by a double consonant,a. Heavy vowels are ^ d^ ^ f,^ it, f^,^ e, ^ ot^ lit o, ^ at , and any

short vowel, if followed by more than one consonant.j 29. Vowels are, lastly,ivided according to accent, into cicuie (udfttta),

grave (anudfttta),nd circumflexed(svarita).he acute vowels are pronouncedwith a raised tone, the grave vowels with a low, the circumflexed with an eventone. (P 9 i If 29 32.) Accents are marked inVedic literature only.

Qw^ and Vriddhu 30. Ouna is the strengtheningof 1^{ ^ i 7 u^ 9 l2, fj ^ f^,IE /i,by

means of a precedingv a, which raises ^t and^ ttoTie,'9u and 'mitto'ô.If ri and ^ K to 1 ^ ar, in fi to mt^ al. (P n. i. i, 2.)

By a repetitionf the same process the Viiddhi (increase)owels areformed, viz. ^ ai instead oiê^âu instead of ^ 0, IIT dr instead of 1V^ arând trn^di instead of ^ a/. (P 9.i. i^ i.)

Vowels are thus divided again into :1 . Simplevowels : wa^ WT(4,\if%iy Tf w, ifid, ^f , ^rf, ij/i.2. Guâ vowels: ^c(a + ), ô(a + tt), iRar, iTcâ/.3. Vriddhi vowels : WTd ^diâ-â-î) dtt(fl+(i+tt),Ti^dr,in7â/.^ 31. V a and ^ d do not take Ouna, or, as other grammarians say,

remain unchanged after taking Guâ. Thus in the first person sing,of thereduplicatedperfect,which required Guna or Vriddhi, Âan forms withGuna wnjc^ghana^ or with Vriddhi '^^^J9 jaghdnay I have killed.

Combination of Vowels at the end and beginning of words.^32. As a general rule, Sanskrit allows of no hiatus (viv^tti)n asentence. If a word ends in a vowel, and the next word begins with a

vowel, certain modifications take place in order to remove this hiatus. 33' ôr the purpose of explaining the combination of vowels, they

iDAy be divided into two classes :I. Those which are liable to be changed into semivowels, jî, ^,V fi, is ^,

^rit'^rt ; also the diphthongs,ve,^ ai,^ 0, ^ au.%. Those which are not, V a, VT d.Calling the former liquid*, the latter hard vowels, we may say : If the

* The Pr tis khya calls them n^mtn, for a different reason ; see Rig-yeda-pr tis khya,ed. M. M., p. xziii.


38/323

14 RULES OF EXTERNAL SANDHI. $ 34-same vowel (long or short)ccurs at the end and beginning of words^ theresult IS the long vowel. (P n. vi. i, loi.) Thus

^ or ^ + W or l T = ^ S'^d = d.Iôr l+ôr |=s|i + t=i?.V or 4-^ or ifi='ai ^-f u = tf. ijor ^+ôr ^=^rr + rf = rf . |

Ex. 9W VMÎ^ni =3 ^WM'l^ril ukivd'\'apagachchhaH:=zukivdpag u:hchhaH, ihaving spoken he goes away. j

'^ ^^^ = H^l ^9l1nadt + IdfiSi nadidrUU such a river. ii|| ^ = ^shjkartri + riju= kartrijudoing (neuter)right. jf)|i|i^f)r=fiil|^tf)rintu + udeti = kintildeti,ut he rises. i

Or in compounds, ^^ + ^|l^ ^T^)^ mahi + tia^ = mahiiahi lord of the earth. ;^ 34. If hard vowels (longor short)occur at the end of a word, and thenext begins with a liquidvowel (exceptdiphthongs),the result is Guâ of

the liquidvowel. (Pan. vi. i, 87.) Thusôr^ + ôr^=^fi + i = c (at).ôrWT + orifl=^5 + S = o (au).^ or WT + ôr ^r=iR d-^-ft^âr.(P p.1. 1,51.)

Ex. ir^ 1^ = W^^t tava + indra^ = tavendral^ thine is Indra. IHT 9W=^fNn 9d + uktvd = 8oktvdf she having spoken.

t HT ^flr = irflF:d + fiddhii^ sarddhi^, this wealth.1 ^ mmKl = IHRVnc: âva + likdrahk tavalkdraJ^,thy letter fi.Or in compounds, mwi + ^fte ^nv^fr: kdmya + MA/ii^= kdmyeshtih^ an

offeringfor a certain boon,f^ + 9^1^^ = f^lfhl^^hita + upadeiah = hiiopadeiah^good advice.

J 35. If hard vowels (longor short)ccur at the end of a word, and thenext beginswith a diphthong, the result is Yiiddhi. (P n.vi. i, 88.) Thus

^ or ^ + 1? = % a + c==ai.^ or ^ + ^=^ a + at = ai.^ or ^+^=^ a + o = ai .^ or ^H-^=^ a + dtt = ai .

Ex. Tf^ u^ = n^ iava + eva = tavaiva^ of thee only.HT Â^ s %f^ 9d + aikshishfa = saikshishfa she saw.

* The letter Ifft is left out, because itis of no practicaltility.It is treated like ^ r* ^ IsubstitutingJSl for ^ r in Gupa and Vriddhi. Thus '^-f-V^^hlS+anuhandhah becorc9H^V lanuhandhahf i.e. ha^dng /t as indicatoryetter.

t Some grammarians consider the Sandhi of a with ft optional,but they requireshortening of the long d. Ex. WVT-h^V^ brahmd'\'rishih:=:WWffttrahmarshifi or^qf^Vtrahma jrishih,rahma, a Rishi.


39/323

-$37- BULES OF EXTERNAL SAKDHI. 16m îtWl= ir^lv:ava + oshfhah = iavatuhiha^f thj Up.WÎ^W^rf) = ^fyw^^ sd + atdsukyavati = sautsukyavaii he desirous.

Or in compouDdsy T7ir + îg^=tiîn rama-^ aiharyam^rdTnaiivaryafn^the lordshipof Rlboia.

^ftrrr Ilhf4 = ^^^^Ht^îiBUdaupamyam = sttaupamyam, similarityithStt ,the wife of Rfima.36. If a simple liquidvowel (long or short) occurs at the end of a

word, and the next be^ns with any vowel or diphthong, the result ischange of the liquidvowel into a semivowel. (P n. vi. i^ 77.) Thus

Hr 7?r = î^ A:ar/ft uta = kartrttta,doing moreover.f^1[7= fftiRnadhu + it;a = madhviva, like honey.1^^ ^1^ = îdr^ na^f ^ + ai4asya = nadyai4asya,the river of Aiâ.

In compounds, 7|^ + ^=: fv4 nadi-hartham = nadyarthain, for the sakeof a river.

Note Some native grammarians allow^ except in compounds, the omission of thisSandhi, but they requirein that case that a long final vowel be shortened. Ex. ^V^ HHchaM atra may be ^nfKpf chakryatra or ^fn H?| ehakri atra,

^37. If a Guna-vowel occurs at the end of a word, and the next beginswith any vowel or diphthong (excepta),the last element of the Guna-vowelis changed into a semivowel. If a follows a is elided,and no change takesplace in the diphthong; see 41. (P n. vi. i^ 78.) Thus

I? (e)4- any vowel (excepta)= in^ (ay).lit (0)+ any vowel (excepta)= 1R( (av).


40/323

16 RULES OP EXTERNAL SANDHL #38-Ex. ^ Wf9 = ^sraRPRK sakhe Agachchha = sakhaydgachchhay Friend,come 1

?R^ ^= ^nrftf sakhe iha = aakhayiha, Friend, here nnt ^^^zjfui^f prabho hi=prabhavehi^ Lord, come near m^ ^Fhni 1PTi$M prabho ttU8Aadham=iprabhavau8hadham, Lord,

medicine.In compounds, Tft+ t^r:='nJh|r:go-\-iid(^=^g(wlia1^.here are variousexceptionsin compounds where 7ftgo is treated as ir? gava. (^41.)

^ 38. If a Ypddhi-vowel occurs at the end of a word, and the nextbegins with any vowel or diphthong, the last element is changed into asemivowel. (P 9.vi. i, 78.) Thus

% (at)- any vowel = ^nn^ (ay).^ {au)+ any vowel = in^ (dv).Ex. f^ v9: = finrnr9: triyai arthahîriydyarthab

ftrt^ = ftnn^ iriyaifite= iriyayriie.T^ iv^ffln^ssiqi l%jDi'iiavau astafnite^ravdvasiamite^ after sunset.W^ X^=1ITflfftrau iti=ztdvUi.

In composition,i04-^=f|T^ nau-bartham^^ndvcartham^ for the sake ofships.

39. These two rules,however, are liable to certain modifications :1. The final y and ^v o(%nay, iraav, which stand according to rule for

^ e, ^ 0, may be dropt before all vowels (except, ^ 41) ; not, however,in composition. Thus most MSS. and printed editions change

^ VTSRK sakhe dgachchha^ not into ?nnnTaK sakhaydgachchha, but into^l?r miNA. sakha dgachchha.

ig^ ^ 'Sakhe ihOy not into Tlirftlâkhayiha, but into ^m ^ sakha iha.wit 'f^ prabho e t,not into Wi(^f prabhavehi, but into inr wf^prabha ehi.nnt Â^rl prabho aushadham^ not into iPT^vM prabJiavaushadham, but into

IPT ^Fhniprabha aushadham.2. The final i^y of ivpi dy, which stands for ^ di,may be dropt before all

vowels, and it is usual to drop it in our editions. Thusftl^w4: iriyai arthaf^ is more usuallywritten f^m W$: iriyd arthab instead

of ftinrnr^ Mydyarthab.3. The final ^ r of ^BT^ dv, for ^ du, may be dropt before all vowels, but is

more usuallyretained in our editions. Thuslit1^ iau iti is more usuallywritten irrf^^ tdviti,and not iTT jflfid Hi.

Note Before the particle u the dropping of the final ^y and ^ v is obligatory.It is without any reason that the final ^y of Guiâ andVpddhi and the final ^9 of

Gupa are generallydropt,while the final ^v of Vfiddhi is generallyretained. It would bemore consistent either always to retain the final semivowels or always to drop them. SeeKg-veda-pr tisftkhya,d. M. M., Siitras 129, 133, 135 : P 9. vi. i, 78 ; viii. 3, 19.


41/323

-$ 42. RULES OP EXTBRITAL SAITOHI. 17f 40. In all these cases the hiatus^ occasioned by the dropping o{^y and

^ Vf remains, and the rules of Sandhi are not to be applied again. 41. ^ e and ifV 0, before short v a, remain unchanged, and the initial

^ a is elided. (P 9.vi. i, 109.)Ex. f^ iV9=f^s^ Hve atra = iive'trayin Sîva there.

wit' V |J|^llU mt*'^^JfTO prabho anugfihdna^prabho 'nugrihdi^.Lord, please.

In composition this elision is optional. (Pftii.i. i 122.)Ex. ift+ WW: = 'ft 'W: or nhnBt: ffo-âivd]^=go'ivdiL or go aJhdf^ cows

and horses.In some compounds iHT gava must or may be substituted for vf^go, if a

vowel follows ; ^nrnvt gcwdkshaf^ya window, lit.a builds eye ; 'iH^.gavendraiylord of kine, (a name of Krishna) ; imftpj or iftsftptâm/tnam or go ^jinamâ bull's hide.

Unchangeable Vowels (Pragrihya).{ 42. There are certain terminations the final vowels of which are not

liable to any Sandhi rules. These vowels are called /TroÂya (Pd^.1. 1, 11)by Sanskrit grammarians. They are,

I. The terminations of the dual in ^ (^9 t2,and ^ tf,whether of nouns orverbs.

Ex. V7t ^ kavi imau^ these two poets.flTTft?rr giri etau, these two hills,^nv ^ sddM imau, these two merchants.W^wnr bandhH dnaya, bring the two fiiends.(91^^ late ete, these two creepers.f^ ^ vidye imey these two sciences.^nni^ P^^ Saydte arbhakau^ the two children lie down.^nm% ^TPft iaydvahe dvdm^ we two lie down,in^ ^1^ ydchete artham, they two ask for money.

Note Exceptions occur, as 'Hlfhfmanivay i.e. *Wft 1{^matî wa, like two jewels.; ^Hin^dampativa, i.e. ^Mff1 dampatiiva, like husband and wife.

a. The terminations of ^nftamf and ix^âmd the nom. plur.masc. and thenom. dual of the pronoun ^r^ adas. (Pfin.. 1, iz.)

Ex. îiftWiVTt ami aivdi^ these horses.^fft JJW* ami ishavaby these arrows.m^ ^pSift am4 arbhakauy these two children, (Thisfollows from rule i.)


42/323

18 RULES OF EXTERNAL SANDHI. 43-IrregularSandhi,

43. The followingare a few cases of irregularSandhi which require tobe stated. When a prepositionending in v or ^ a is followed by a verbbeginning with T[ or wt 0, the result of .the coalescence of the vowels is1? c or ^ 0, not ^ ai or ^ au. (Pan. vi. i, 94.)Ex, IT + îfT^ îf? pra + ejate=:prejate,

Tff + ^^ = ^ ^^W upa H- eshaie = upeshate.1T-i- ^^^f?r= iTR^fifpra + eshayati^preahayati *.'TCr + iwfir= ^^rfif para + ekkati =parekhatL'9^ + ^ftilfif ^M^MPri upa + oshati = uposhati,iRT + ^ft^ = MO^Pripara + ohati =parohatu

This is not the case before the two verbs ^^ edh, to grow, and ^ t, to go, if riused by Guiâto 1? e. (P 9. VI. I, 89.)Ex. W + ^Wn = ^hMW upa-êdhate=:upaidhate.

In verbs derived from nouns, and beginning with 1? or ^ ^ or o, the elision of the finalV or W a of the preposition is optional.

44. If a root beginning with ^ ft is preceded by a prepositionendingin W a or ^ a, the two vowels coalesce into ^btt dr instead of ^R ar. (Pan.VI. I, 91.)Ex. Wi| + ^Ôf = im aini apa -f-richchhati = apdrchchhati.

'R + ^5^DflfkWnSlHl ava + rindti = avdrndti.H + 1^ = HT^ pra + fijate prdrjate.TOT + '^îftr^TO^firpard + rishati =pardr8hati.

In verbs derived from nouns and beginning with ^ fi,this lengtheningof the V a ofthe prepositionis optional. (P^. vi. i, 92.)

In certain compounds ^^ riii^mydebt, and ÎH litah,affected,take Vriddhi instead ofGuna if precededby âj H + ^'^ =^V[t9pra-^rinamznprdrnam, principal debt ; ^^ +^pir= ^î3 f na-|-n?am=rtiuiniam, debt contracted to liquidatenother debt ; 5[hlf^ÎR = ^H^lflt ioka-^-fitahzîokdrtah,ffected by sorrow. Likewise ^ 2A,the substitutefor ?TO vdh, carrying,forms Vriddhi with a preceding^ a in a compound. Thus PhB +Vf viiva-\-iihah he ace. plur.of r ini iiîâvdh, is frin ^vUvauhah. (P n. vi. i, 89,vlirt.)

45. If the initial ^ o in wtVt osklhah,lip,and Wt^ otuh,cat, is preceded in a com-oundby V or ^c, the two vowels may coalesce into w au or wo. (P n. vi. i, 94,

vlirt.)Ex. ^WC + ^ftWl= 1WOTJ or ÎV^t adhara + oshfhah = adharaushlhak or adharoshthafi,

the lower lip.^?y + 'f^ ^'^J^S*r ^[f^tj:thiUa'{-otuh=zsthiUautuhor sthiUotuh,a big cat.

* In nouns derived from r1^presh, the rule is optional. x. h^ or h^ preshya orprtti$kya,messenger, fmpresha, gleaner,b derived from Itpra and ^ 9^.


43/323


44/323

20 EXTEBNAL SANDHI. $5*-

.2

I

3

If

'8

Ik'

UK

1^

^

fur

^

9

'8

mueller a sanskrit grammar for beginners in devanagari and roman letters throughout

Documents