( Pages 1 to 298 )


( Pages 1 to 260 )





-^ ,

Second Reprint 1967 Price Rs, 40.00












(Of Balliol College Oxford)







£ T";,.

>hanti Lai am

Sundar Lai Jam

Motilal Ilium rNhlusft, flirl J»*»c»«™ ******

, u _ I Buncalow Road, Jawahar

Bungalow Road, ^ ^ ^

Jawahar Nagar, Delhi-6

Copies available at :

i. Motilal Banawldass, Bungalow Road, Jawaharnagar, Delhi -6 a. Motilal Banarsidass, Nepali Khapra, Varanasi. 3. Motilal Banaraidass. Bankipur, Patna.

II %ft:




On my return home to India, a systematic and exhaustive search for the hitherto unutilised mss. of the Nirukta was made. I myself undertook an extensive tour in Rajputana, Baroda, Tihri-Garhwal and Cashmere and examined the state collections of Sanskrit mss. I also visited Benares, Patna, Madras, Madura, and Tanjore. I wrote to scholars, and librarians in charge of Sanskrit mss. throughout the country. I was therefore able to secure the use of several mss. hitherto not utilised for the constitution of the text. These mss. are as follows :

Bk.1 This ms. belongs to the state Library, Bikaner. The loan was secured for the University of the Panjab through the courtesy of Maharaja S'ri Sir Bhairon Singhji, K. C. I. E., Vice- President, Council, Bikaner.

Contents. The Nirukta of Yaska in the shorter recension containing the two parts i. e. the piirvdrdha and the uttardrdha in in leaves. It is a badly preserved ms. and full of mistakes.

Size: 9J"x4|" Material: Paper. No. of leaves: in No. of lines per folio: 8. Characters: Devanagarl.

Date : on f. inr. ( sic. ) «ft OTntTsrH*?^ II *N^ *i*rrftj*r§?5*Trif I i sre^ri fovprft % feftkii <ra* i. e. I7o5 Vik.=l67i) A. D.

The colophon ends thus : ( sic. )

BK.8 Contents : the S'-iksd Cat'iiMaya written without a break. The IIIH. is injured in many places. Niyhantu is given from f. 9 r. to f. 18r.

Size : Og* x 4f . Material : Paper. No. of leaves : 18.

No. of linos: 8. Charaetern: Devaniigarl. Dato: cm f, 18 P:

MIC: $*? «VH9It || «o ^«^ ^rj^ 1*\^ q»i W^A ^

*ft i^f'HnT fef^^f i *?t iTiinfvrn^ %ft si^jiftrf & fonwA u ftr^ir^i 11 K.1 A am. written on paper in Uevauaguil eharucU

obtained through the courtesy of the Curator, Central Library, Baroda. The contents are the 1st half of the Nirukta. It is full of mistakes and belongs to the longer recension.

Size : 8 ^ x 3§". Number of leaves : 69. Number of lines: 9. Date: sic. ^r% 11 n u ^ u

pJi u «<tqi«ft %%*K^fi Scribe : sic. w^Hflw^TO x **** * *rnr a u

Peculiarities: The colophon at the end of the 1st chapter runs thus : ll ^% %?5% ^^wt«rnr: ll ; at the end of the 3rd chapter thus ; sic. u ^r 5*^fc u i^% u ?jrforhan*T: ^n^rr ^r u VL u & M ?i% in \ n ^^ u i4wwft«iuw«wi<1 n «nTi5^[^te*^i ft«n ^jttr'rlftr HIT?: u ; at the end of the 4th chapter thus : u ^fa %^ i^r^ ^m«Tni: w\i\ft*i ?&fo *n \\

Evidently ayam refers to the 4th chapter and not to the Nirukta as the 5th chapter is immediately continued. But the use of vd does not seem to be correct in this case.

Sandhi and spelling : The visarga is retained but at the same time euphonically combined i. e. a double process is introduced, e. g. f. lv: 5^n^n?r^r:^8r«T ...... *wiwn:«& etc. The avagraha is not

marked, e. g. f. 2r. sn^fa f. lv. 3ST<nT5*nftft etc. The dental

nasal is avoided in conjunction, being reduced to unusvdra, e. g. f. lv. Pkn*i«iil*t«<6<i ^«JS =50«iifiio

F. 2r. «f%nftu : f. Gv.

R.a A neat, well-written, woll-[>roscrvcd and complete ms. of the Nirukta in Devanagarl characters.

Sixe: 8i/;x4J". N urn ber of leaves: 69 + ?7 + i blank=147. Three leaves 75-77 are wrongly numbered 85-87. -Number of linew : 9.

No date given. The name of the scribe not known. The ins. lookn alxjut 200 year.s old. It belongs to the longer recciiHion and doe« not give any new variants.

B*. It contaiiiH the 2nd half of the Nirukta, written on paper in Devanagari character*.

Size: 8£"x3|-". Number of leaves : 83 + i blank. Number of lines: 9.

Date : sic. i ^ i 31% I

The owner seems to be one Ramakrsna. He is saluted like a god which is rather unusual : sft i *m$wnTrfe<snr «nr. I Or the word Mcilika does not refer to the owner of the ms. but to God, who is the master of all.

The ms. belongs to the longer recension. It represents a very late stage of textual expansion. The colophon at the end runs

thus I s[Kf «i^»Tf> >3THM^j ^8*n^lT« I .

R*. A fragment of the 1st half of the Nirukta, written in Devanagarl characters on paper. It looks old. It is illegible in many places. Ink is bleached by age. It belongs to the shorter recension.

Size : 9£" x 3J". Number of lines : 7. Number of leaves : 61- Eight leaves are supplied in a different handwriting.

Kn. A palm leaf ms. written ia old Canarese characters, presented to the Panjab University Library, Lahore. It belongs to the shorter recension and sometimes gives important readings.

Size: lU"xiy. Number of leaves: 94+i blank. Number of lines : 7.

No date is given, but as it is written in old Canarese characters, it must bo at least ;500 years old, and might bo older.

K6. The text of the Nirukta is made up by 2 different inss. Tho 1st half is given in GO+i blank leaves.

Size: i)i"x3.y'. Number of lines: 9. Date is not givoii. Tho name of the surilx) is also unknown. The colophon ends thus:

n wrft u sr«^ u i^%«?fn|owrg i The ms.

belongs to the longer reeeiisiun.

o o

Tho tind halt' is given in i + C5 leaves. Si/e: <J" x il". Nuinhor -)f lines: 9.

Colophon ends thus: ( sic. )

<rtonra*ta; u u

u V u wj u ** u v u ^ n $wm*nw *ro: « s *rr^<*u« m \\

The text belongs to the longer recension, both the parisl^tas being given in full.

K.8 An incomplete ms. of the 1st half of the Nirukta. The 1st leaf is missing. It looks old and has preserved the old spelling. The text belongs to the shorter recension. In many parts it is illegible.

Si«e : 9" x 4". Number of leaves : 78.

Number of lines: 8. Characters : Devanagarl. Material : paper.

The text is given up to the 34th section of the 6th chapter. The last leaf is missing.

R.T A fragment of an old ms. gives the text of the Nirukta in tne shorter recension from the words srorw^f trranrer ^nrroir ...... of

the 4th section of the 5th chapter up to : a^faf *ff : xrf^ir sfowrr of the 9th section of the 8th chapter. The 7th ch. begins thus : ( sic. )

Size: 9"x4£". Number of leaves:... 54-97... missing. Nu- mber of lines : 9. Characters : Devanagarl. Material : paper. Spelling : old. Date etc : unknown.

R8. A ms,, the contents of which arc tho 2nd half of the Nirukta.

Size : 9" x 4". Number of leaves: 62.

Number of lines per page: 9. Characters : Dovanagarl, Material: paper.

Date; sic.

uwfort vnftat i<nnfi feftwfe^ u v^ u ^? *?^3 u

The Parift'iijtas are given together as one chapter, being bodily separated from the 12th ch. of tho Nirukta.

Spelling is old. The paper is very much worn. There is no ground to suHpect the date. Tho external appearance, tho state of the paper, which unmistakably bears the stamp of old ago, auc|

the internal evidence of old spelling support the above mentioned date. It belongs to the shorter recension. The following case of accidental omission is to be noted. F. 29r. 1. 6 from top:

The eye of the scribe wandered from the 1st line to the similar words in the 3rd line with the result that the intervening passage «nr^HTT...^5n7: was omitted.

D. A ins. brought for me by my friend P. Bhagavaddatta. B. A.. It is a neatly written, well preserved ms., and belongs to the longer recension.

Size: 6j"x2j". Number of leaves: 112+128+i+21=261. There are 3 parts. The 1st two parts contain the 1st and the 2nd half of the Nirukta respectively. The contents of the last 21 leaves are the Nighantu. Number of lines: 7. Characters: Devanagarl. Material : paper.

Date on f. 112r. (sic.) sr%

Scribe:— (sic.) ^im<S*«t>< ^3^^^ *?$WT&T feftiH u Date on f. 128r. ( = 240 ): (sic.) u *n% i w <Mhmr*i««ft ^nh

Scribe i >sn*r&T^T ^^nTPSf r ^wi^^Twrs^T^VT^ fefe^i I Colophon ends with the usual statement : *m( 5^1% ijr etc.

Date on f. 21r ( =261r ) : sic. u sr

Scribe : ^fr^

There is a good ms. of the Jfiruktq in the library of H. H. the Maharaja of Alwar. All my efforts to secure a loan of this ms. were fruitless, as the authorities refused to lend the ms. to the Panjab University on any terms. Even a copy of the ms. could not be obtained. But H. H. the Maharaja was kind enough to let me see the ms. On examining a few test passages, I found that no new variants were forthcoming. Its collation was therefore unnecessary.

Besides, six ms.«. of the Rnghunatha Temple Library, Jammu were collated by Principal Raghubar Dayal M. A; M. O. L. of the S. D. College, as far as the 12th section of the 1st chapter of the Nirukta. He WMS good enough to place the result of this collation at my disposal. On carefully examining the critical notes supplied by Principal Raghubar Dayal, I did not find any new variants. I did not therefore feel justified in collating the mss. afresh\

Th& evidence supplied by the Indian mss.- further supports the conclusion, deduced from the collation of European mss.

Thfc evidence of the European mss. was discussed at length in my Introduction to the Nirnkta, published by the Oxford University Press in 1920. The Introduction was sold out within a few years of its publication and is now out of print. New of the Nirukta will require information with regard to the principles of the constitution of the text. For their benefit, the relevant part of the Introduction is reproduced.

The Relationship of the MSS. : two recensions.

The manuscripts* foil into two groups, and for the sake of convenience and brevity, may be called A and B- A representing the longer and B the shorter recension. None of the manuscripts grouped in these two families is earlier than A. D. 1479. Although they have been copied form earlier manuscripts often with great labour and trouble as some of the scribes remark neither of them transmits the text of the Nirukta in an uninterpolated state. Both recensions add the paris'ista which can be proved to be an interpolation by independent testimony as an integral part of the text, and cannot, therefore, be the faithful repre- sentatives of the archetype. Moreover, both have besides the paris'ista, an entire section or the equivalent of a section added on to them. These additions are meaningless. The commentary on the Vedic stanzas quoted therein is very poor, and written in a style quite different from that of Yaska. For instance, there can hardly be any doubt as to the interpolated character of ix. 2, which is given as a constituent part of the text by the manuscripts of both recensions. Further, the commentary on the Vedic stanza in xi. 7 is meaningless and written in a different style. The Vedic stanza, being quite easy, requires no explanation.

Yaska generally does not comment on easy Vedic stanzas, simply remarking: iti sd niyada-vy.(lkhydtdl9 i. e. 'this stanza is explained by the mere reading'. In all such cases, this note of Yaska comes after easy Vedic stanzas only. It would thus be intelligible, if it had followed immediately the Vedic stanzas in xi. 7. But as the text now stands, it is placed just after a very difficult Vedic stanza in xi. 8. This is contrary to Yaska's method. It is clear that the words : iti sd niyada-vydL'hydtd were originally placed immediately after the Vedic stanzas in xi. 7. The intervening passage is an interpolation, and rather a clumsy one, for it can be easily detected. This is further proved by the fact that Durga, who repeats every word of Yaska in. his commentary, ignores them. How these additions gradually found their way into the text is illustrated by the following example. There i? an easy quotation in xii. 2, and Yaska, as usual, simply adds : iti sd diyada-vydJchydtd. Some interpolators have endeavoured to add after these words a short comment. Thus some manuscripts IICTO subjoin lliu following remark: ^rr%g ^r ^r<ft srerrrRT "a^ff 3*??% ...... g*i<n?h n

Further, each recension contains passages, which, being super- fluous, are omitted by the other, or are amplified versions of those in the other. For example, B adds, between vii. 19 and 20, one entire section, which is omitted by A. It is clearly an interpolation as the commentary on the Vedic stanzas is identical with that of xiv. 33 with slight alterations.

Again, in B the commentary on the Vedic stanza quoted in v. 27, reads as follows: fj^Kc^r s^niTRt TO <re ^ 13$

ti n

A's version of this is greatly amplified:


Further, A contains a long passage in 6. 5: : omitted by B.

*0miss,io ex homoeoteleuto' in Sanskrit Manuscripts. It is clear, therefore, that both the recensions cannot faithfully

1 Cf. N. x, IS. 24; xi. 3, 45; xii. 31,

represent the archetype. Hence the question arises which of them adheres more closely to the original ? Koth adopted the text as given by the longer recension in his edition, without, however, assigning adequate reasons for his preference. The same text is also adopted by most of the editors of the Nirukta. This text, as has been shown above, does not represent the original. It is true that often the longer recension preserves the better text, for sometimes passages are omitted by accident. The eye of the scribe wanders from a particular word to the same or to a similar word, occurring further on in the text, with the result that the intervening words are omitted. This phenomenon known as omissio ex homoeoteleuto is universal and of very frequent occurrence. The following example illustrates this kind of omission. In copy- ing the lines: 'The book, which is rather scarce, was till very lately of absolute necessity for the Student of the Christian Hynmology, above all for the Student of Adam of St. Victor's hymns',1 the eye of the copyist wandered from the student of the first to the same word in the second line and the words 'of the Christian... for the* were left out. The same thing happened to the scribe of ms. C 3. In copying the sentence : *?fifaH*{*H SKfjTTOt ^jp&Wi I swftg-

his eye wandered from the word

in the first line to the same word in the second line, with the result that the words 3*?mi T^r« I srcfaug were left out.

Again,, in copying N. vi. 22: ^ ?re: smni $3f*ff $foftg I RV. VIII. 4. 19. r<p: snrrfaKnrraV srsi^nrft i the eye of the scribe wandered from the **jr of the first line to the similar word **p: in the second line, consequently the intervening words <rr«r: 9RTW... were omitted in ms. C 3.

Further in N; ii. 26 :

: is the first pdda of the second hemistich of the stanza of RV. III. 33. 6. Unconsciously the scribe remembered the second pdda ?rar ^ sm% *rnr ssff: and wrote it down immediately after finishing the first pdda with the result that the intervening words I <nf&T: <TOfT^...^3rqi?cf are missing in ms. C 4. It cannot

therefore be concluded that the shorter recension is always the best, for sometimes omissions are accidental.

1. Clark, Descent of Manuscripts, p. 1.

Dittography in Sanskrit Manuscripts.

On the other hand, there is also the phenomenon called dittography, i. e. the repetition or addition of a few words or sentences. An excellent example of dittography is furnished by The Globe on July 9, 1915.

The Echo departs publishes a message from Cettinje announcing the message form Cettinje announcing the appointment as Governor of Scutari of Bojo Petrovitch.'1 The part of the second line is a verbatim repetition of a part of the first line.

In N. ii. 28, ^srwft ffcqfa gr^ffa rftaror ^ arft ......

the eye of the scribe wandered by chance after *% to the Vedic stanza, and he mechanically copies the whole of the first line except 3<T *r in ms. C 5.

Again, in N. vi. 8, the scribe of the ms. Mi. repeats

Further there are some passages whose omission by B is absolutely unjustifiable. Yaska explains every word occurring in the fourth chapter of the Nighantii. The omission of the passages containing the explanation of any of these words is therefore inconsistent with Yaska's plan. Examples of such omissions are the following. Yaska explains ftsprr: ( Ngh. 4. 3. 12. ) in N. vi. 3, but the passage : ferersn ftsr«Tfrftoi: is omitted by B. Again, Yaska explains sffow ( Ngh. 4. 3. 28 ) in N. vi. 6 as sfrro sraror, which is omitted by B. This omission makes the following Vedic quotation meaningless.

FurtKer, in commenting upon a Vedic stanza, Yaska always starts from the very beginning of the stanza. To leave out the first few words and to begin from somewhere in the middle of the stanza is altogether foreign to his practice, yet if the text of B be followed. the omission of the passage: qrafcR f^onfoft Rcj^r ( N. vi. 26 ) would involve Yaska in an inconsistency. All this shows that B is not absolutely reliable.

1 Cl*rk, op. cit., p. 6.

Now let us examine A. The majority of the manuscripts of A belong to a period later than those of B. Thus not one of them lias preserved the old spelleing, while most of the B mss. retain this peculiarity, i. e.

of writing 15 as I / as ^rt for ^

I /'I


$ for

Again, some of the A MSS. divide the paris'ista into the so-called thirteenth and the fourteenth chapters, while those of B put the whole of the paris'ista into one chapter only, which is numbered the thirteenth.

It has already been pointed out that A contains an obvious interpolation in N. vi. 5, and an amplified version of B's comment in N. v. 27. Besides these there are shorter passages scattered throughout the book which are omitted by B and are suspected to be interpolations. Ona very fertile and insidious source of interpo- lations is supplied by. Yaska's own method of giving etymological explanation. He does not content himself with one derivation, but goes on adding derivation after derivation of a single word till the whole list of probable, possible, and even fanciful etymologies is exhausted. In many cases, interpolators found it quite easy to add new derivations and attribute them to Yaska. A contains a considerable number of such additions, while B has only two.

The following are a few samples :

N. ii. 6. A reads : %$ft Jwmj 1 t*9TT «frf ftre^tft 3T I qn fy^foi-

B reads: The two derivations are omitted. N. ii. It). A reads:

B reads:

N. ii. 13. A reads: B reads :

N. ii. 20. A reads : B reads :

N. ii. 22. A reads: sr«w ^fa g&rara B reads :

N. iii. 8. A reads : B reads :

N. iii. 10. A reads: B reads:

N. iii. 15. A reads: ^t ^t

3F&& I

Breads: ^t wt ?r^ fr>gfcr ^j; i

In this particular case it is obvious that the passage ^rc: ^^TT^, &c., is an interpolation, for as the words stand in the first line, Yaska would naturally give the etymological explanation of &w first and then of ^r, not vice versa. As a matter of fact he does so ; after explaining ftrvrer he says :

This would have been absurd if the reading of A represents the original.

N. iii. 16. A reads : -su®«n ^ l^r 0fa I i^

*r ii u ll B reads : sro^n er ^55T %fir u i $ u

N. iii. 19. A reads: fofifens^nihr...^ I iWfir ^wr^ i B reads: ft

N. iv. 2. A reads: JT^^T...^!^ I ^pn^r 4fiUi<&<kfr I

B reads :

N. iv. 10. A reads : B reads:

N. iv. 13. A reads: B reads :

N. iv. 15. A reads :

B reads : «Erqr ^wftar ?ra% i tfesf ^r«^fcr 31 i

N. iv. 19. A reads: 3^ s* 3^ I ^for 571*353^ I 3^% i

B reads : sr^i 5? <r^s% i <»)«t»i

N. v. 3. A reads : <TRftfrT

B reads : 7i%ftfflr N. v. 12. A reads:

B reads : N. v. 26. A reads:

B reads : N. vi. 8. A reads :

B reads: N. vi. 16. A reads :

B reads : N. vi. 33. A reads : ?ft^ $& i

B reads : ?ft?T% ft% i

N. vi. 32. A reads : f^ 5$&rf<r fe^ srr f*??^1 ^rr

B reads : N. vi. 33. A reads:

B reads : *R ^r^sfiifftni 'Wf^nfiRf «rr

Instances might be multiplied, but the above examples suffice to show that A has been much more tampered with than B.

Fortunately, as has been said above, Durga repeats every word of the Nlrukta in his commentary, so that the text of the Nirukta in toto can be reproduced from his commentary alone. This commentary therefore serves the purpose of a manuscript of the Nirukta and supplies valuable information about the condition of the text in its author's time. Durga does not recognize the paris'ista as an integral part of the Nirukta, as in fact he is even unaware of its existence. Thus his commentary preserves the text of the Nirukta as current before the addition of the paris'ista.


Further, it derives great value from the fact that Durga displays critical judgment in the adoption of readings in the text, while giving variants and adding critical notes on them. For example, in N. i. 2, he reads STJJITT^ but gives ^TT^ as a variant, adding :

Again, in N. i. 12, he reads tfte^Mift but gives qfcflRift as a variant, adding scorer tffti*?Rriv|f

Again, in N. iii. 15, he remarks : srfari *r %&**


Again, in N. iii. 21, he reads 3i£Mifcr but gives «rfe^Rf as a variant, adding :

Again, in N. iv. 19, he reads *g: but gives «RT^ as a variant, adding : ^TT^sf^1 ^ftc^f TTS: i sr^rf^^^f MHI^^I^: i

Again, in N. vi. 2, he remarks : £fr% *TT

Again, in N. vi. 4, he reads sRsrnrr^ but gives 3H3«inq and as variants.

Again, in N. vi. 6, he reads snrorer but gives ifam as a variant, adding : $ftop$r3r^% w**§ I acg^^qq^^ I g^»^)fft fe *TPST Again, in N. vi. 21, he remarks: 5^^%: sremCT^tri i

Again, on N. vi. 33, he remarks:

This shows that Durga took pains to ascertain the correct readings and has handed down a sort of critical edition of the Nirukta, as it existed in his time.

Three stages of interpolations.

We have thus manuscript materials which belong to three distinct periods.

(1) D, i. e. the commentary of Durga, written before the addition of the paris'istas and embodying the whole text of the Nirukta, represents the earliest period, i. e. about the thirteenth century A. D.

(2) B, i. e. the manuscripts of the shorter recension, represents a period later than D, when the paris'istas were added, but not divided as yet into different chapters, and when the old orthography was still prevalent.

(3) A, i. e. the manuscripts of the longer recension, represents a still later period when the paris'istas had been divided into chapters and the old orthography had gone out of use.

A collation of these three different recensions indicates that three distinct stages of interpolations in the Nirukta can be clearly traced. For example, let us take a passage in N. i. 4. On collating D, B, and A, we find that the reading of D has been expanded in B, and that of B in A.

N. i. 4. Dreads: arrarifaii^ ^rfefa

B reads : «iHN!(M3ft rc%H n^f*" I


A reads : sr

i ^Mimifiv^i^^lcud i

Another example for these three stages is supplied by N. ix. 2, as follows:

D reads : arer «gr^iq; i <refar *rar% u ^ u

B reads : «r4t S^R^M: i 3^Nr ^r% u ? u

wit €te^r g^ ...... v^i%<^ 'ift ^RT u

>£( ^ ^®^f t


*rr 'rt film A reads : «r^ S^T^T

"ift ^Rl II

I g>< ^l*<l<d ^T I I W^ «4i<JHin: I ^HT *T^l?r U ^ U

It has been shown above that the list of etymologies increases as one passes from the text of B to that of A. In the same manner the list of etymologies increases in B as compared with D. The following are some examples.

N. i. 4. D reads: fgHiqlfoTO&HfEafr I 3 f^r4t...

B reads : ^cr+imifar^i^'^q^foM^ I ^C^IM:: $%$ tft^cf I «J

T> reads : TOT; ^THST %%: I B reads : TOT: SFTT^T i^

N. i. 7. D reads : <&$& & %& \ ts fa^r

B reads : if ?%TT ^% %^^ I *mV •**&: \

N. ii. 22. D reads: B reads:

N. ii. 26. D reads : B reads:

: n

N. iii. 18. D reads : rer^: B reads : wrf :

: u ^rr


N. v. 4. D reads : wfi

B reads : N. v. 23. D reads: ^^n oft

B reads :

u ^ ii

MS. C 1 agrees with B except that the last line *ro is omitted. N. vi. 3. Dreads:

B reads:

N. vi. 8. D reads : %$t ^n?w<iMi fjqr ^f^rMi u ^ u

B reads : ^t ^i*ii<HTtt«{i ^TT i f^n $?te^ +^3f«fi u * II

N. vi. 24. D reads : A reads : B omits it altogether.

N. vi. 28. D reads :

B reads ; srwrr fa®... ...

Parallel instance of Servius, commentator of Virgil.

Thus the stages of interpolation at different periods can be traced. The principle of the ' best mauuscript ' is obviously inappli- cable in this case, for none of the manuscripts can be called the best. All that is available is the best manuscript of each family, and the best plan, under the circumstances, would be to place all the three families side by side. Fortunately it is possible to do so, for the successive interpolations from one family to another are invariably the amplifications of the text of a shorter recension, and are thrust between sentences wherever the text could be so enlarged with impunity, as, for instance, in multiplying the number of etymologies and attributing them all to Yaska. I have, however, distinguished the evidence of Durga's commentary from that of the manuscripts of the Nirukto, although Durga's commentary is very important for supplying such valuable evidence for the history of the text of the Nirulcta, it cannot, strictly speaking, be called a manuscript of the NiruJcta. The relation of the shorter to the longer recension is shown by the use of square brackets, which contain the additional passages of the longer recension, while the relation of the shorter recension to the text preserved by Durga is indicated , by foot-notes. An analogous example is furnished by Latin literature. The text of Servius, commentator of Virgil, shows a similar threefold amplifica- tion ; the three stages of interpolations being pointed out by Thilo in his edition. I think the text of the Nirukta reproduced from Durga represents the archetype as closely as it is possible to restore

it with the help of the present materials. I have collated thirty- seven manuscripts myself, and in addition have taken into account the evidence of fourteen manuscripts collated by Roth, eight by the editor of the Nirukta in Bib. Ind., and six by Principal Raghubar Dayal as stated above. Thus, directly and indirectly, the evidence of sixty five manuscripts is available for this edition. I doubt if any useful, hitherto unutilised ms. of the Nirukta will now be forthcoming. The text may, therefore, be regarded as more or less settled.

The present text is in the main identical with the text, which I constituted at Oxford, and which served as the basis of my English translation. But as a result of the collation of Indian mss., this text is somewhat further developed than that used for the translation. There are certain variations in detail, see for example, line 3 on page 35. Thus the present text differs from the basis of my translation although the difference is not considerable. This may serve to show the existence of the differnce until such time as I may be able to publish a revised edition of the translation based on the final text.

The text is followed by Appendix I. Parallel passages from the Sarhhitds, the Brdhmanas, the Prdtis'dkhyas, the Brhaddevatd, the Astddhydyl, the Mahdbhdsya, the Arthas'dstra of Kautalya and other works of Vedic and Classical Sanskrit are compared with the text of the Nirukta. The Appendix I will be useful for the history of the Nirukta. One could see at a glance the extent to which Yaska is indebted to his predecessors and the influence, exercised by him, on his successors. It will also be useful in enabling one to estimate the originality of Yaska's contribution.

It was formerly proposed to add Appendix II, containing the hitherto unknown and unpublished commentary of Mahes'vara on the Nirukta. But as the text of the Nirukta is already very much delayed and the addition of Appendix II would require consider- able time, the publication of the commentary is being withheld for the present.

Commentators of Yaska.

Although, from an early period, Yaska's work has been recognized as one of the most important veddngas by the orthodox

tradition of literary India, he, unlike Panini, has not had many commentators. This does not mean that he had few followers or that his speculations did not dominate the thought of succeeding generations. On the contrary, he has been acknowledged to be the pre-eminent authority on etymology. Hence, at first sight, it seems rather inexplicable that his work should have been com- mented upon by so few people. One reason of this paucity is that Yaska's work itself is a commentary and not an independent treatise, hence it did not stand in need of much elucidation. Secondly, it is written in classical Sanskrit prose, and, notwithstand- ing its somewhat archaic 'and terse style, is easily intelligible to the reader as compared, for instance, with the aphorisms of Panini ; consequently there was not much demand for further comment. Yet four commentators, at least, are known to have elucidated Yaska's work.

(1) Ugra is mentioned as a commentator on the Nirukta by Aufrecht in his Catalogus Catalogorum.1 But no other information, about his personality, the character of his work, and the time when he lived, is available. No reference is made to him by any of the other writers in the same field.

A ins. in the Library of the Asiatic Society of Bengal is entitled 'Ugra's commentary on the Nirukta '. It is, however, not Ugra's but Durga's commentary. In writing the name of Durga, the letter D was accidentally omitted by the scribe, i. e. Bhagvad- durga was written Bhagvad-urga. This Urga became Ugra by metathesis. The cataloguer never looked at the commentary. He did not even read other colophons, otherwise he would not have committed such a blunder. This misspelt name of Durga appears as Ugra. I suppose it was this ins., which served as the source of Aufrechts' information.

(2) Another commentator is Skaridasvamin, mentioned by Devarajayajvan in his commentary on the Nighanlu :

i tt §.

| ^^TTmrr ^ WHWdl J...

1 Vol. i, p. 297.

It is clear that Devaraja was well acquainted with .the commentary of Skandasvamin on the NiruJcta, and utilised the same in writing his own commentary on the Nighanfa.. No ms. of Skanda's commentary on i<he Nirukta, has yet come to light. He is anterior to Devaraja.

Date of Devaraja.

(a) Devaraja quotes Bhoja frequpqtly, see pp. 20, 21, 29, 85, 37,43,55,69,77,93,117, 130, 145, ftf, 173, 175, 181, 182, 183, 184, 187, 193, 197, 198 etc. of the first volume of the Bib.Ind. edition of the Nirukta. Devaraja 'is therefore later than Bhoja. *

(6) Devaraja quotes the Daiva, a work on grammar by Deva:—(sic.) gft: $nref*?r OTSWT &^t $ni^ ^r %^ i This occurs in the Daiva9 as follows:—^: 5^n7% OTTOT Sr^ ^T^i^% I 136. The quotation is almost identical. The difference may be attributed to the faulty reading of the mss. But even accepting the identity of the passage, the quotation does not lead to any definite result foiS the date of Deva is still subject to controversy.

(c) Devaraja twice quotes a passage from a Dhdtuvrtti. The passage is the following: (sic.) $n^ sWf { srft syran^ 5% wrf^s irf&ftsft «ig<»flafir<ft1«i ft trol qug<«k3 H ^rr^f ^ q^* I The same quotation occurs a second time as follows: (sic.) $rr sfcnft I . %«lild[c<i

TlT H«Mn I

The only extant Dhdtuvrtti is that of Sayanacarya and the passage is not found therein. Nor is it likely to occur in Sayana's Dhdtuvrtti because Devaraja is anterior to Sayana as the latter quotes the former, the quotation being the following:

1. See Bib. Ind. ed., vol. i, pp. 2-4.

2. See, op. cit. p. 43. commentary on $pn ^gh. I. 7.

3. Trivandrum Sanskrit Series No. 1. p. 95,

4. The Nirukta, Bib. Ind. ed. voL I. p. 43.

5. Op. oit. vol. I. p. 109. 4

1 i Devaraja's comment on the word Ngh. II. 11. is the -following: <

I Samas'ramfs edition of the commentary of Devaraja is capable of improvements. Max Muller's ms. of Devaraja's commentary reads 3^rr for 3^r of Samas'rami. It is clear however that Sayanaearya is posterior to Devaraja, who therefore could not have quoted from the Dhdtuvrtti of the former,

The above mentioned quotation of the Dhdtuvrtti also occurs in the Purusakdra, a commentary on the Daiva by Krsnallldduka- muni, as follows: qfa qn*3mi> \ snr sR$r I

I This comment is written on verse 136 of the Daiva, quoted by Devaraja. It is therefore very probable that Devaraja's quotation of the Dhdtuvrtti is borrowed from the Purmakdra. Devaraja will therefore be later than the author of the Purusakdra. The lower limit of the Purusa- kdra can be easily fixed for Hemacandra is quoted three times :—

(1) (2)


The upper limit of the Purusakdra can also be fixed with certainty for it is quoted by Sayanacarya in his Dhdtuvrtti:—




The second quotation is found in the published text as follows: ^3T ^rf% ^4^ I

1. Sayana's commentary on RV. I. 62. 3. Also, see, Max Muller's 2nd edition, IV, CXXXIII.

2. The Nirukta, Bib. Ind. ed. I. 230.

3t Trivandrum Sanskrit Series No. I. p. 95.

4. Trivandrum Sanskrit Series No. I. p. 22.

5. Op. oit. I. 24.

6. Op. cit. I. 37.

7. Quoted by Ganapatis'aatri in op. cit, p. III.

8. Op. cit. p. 61.

(d) Devaraja quotes the Padamanjarl in his commentary on the word ^^nwr1 (sic),

Haradatta, the author of the Padamanjarl is also mentioned8. Haradatta was the son of Padmakumara, a younger brother of Agnikumara, and a pupil of Aparajita. The Padamanjarl is a commentary on the well-known Kds'ikd and later than the Mahd- bhasyapradlpa of kaiyyata, who is mentioned by the author of the Sarvadars'anasamgraha. The Padamanjarl is assigned to c. 1100 A. D. by Prof. Belvalkar in his Systems of Sanskrit Grammar. Devaraja therefore must be later than the llth century A. D.

(e) Devaraja also quotes Bharatasvamin: #«-C[|JIMJ

In his introductory remarks, Devaraja mentions a Bharata- svamin as a commentator of the Veda. The quotation shows that Bharatasvamin belonged to the Samaveda and must have therefore written a commentary on that Veda. A ms. of the commentary of Bharatasvamin on the Samaveda is mentioned by Burnell in his Sanskrit mss* in the Palace at Tanjore5. The commentary of Bharata was written in the reign of king Rama of the Hosala dynasty. King Rama reigned at Devagiri from 1272,3 1310 A. D. The commentary is therefore to be assigned to the end of the 13th century. Devaraja is therefore later than the 13th century. But as he is quoted by Sayanacarya, he is earlier than the middle of the 14th century A. D. He may therefore be assigned to the beginning of the 14th century.

Devaraja also quotes one Durga8. This Durga however is not the commentator of Yaska but a commentator of the Kdtantrasutrapatha, the standard work of the Katantra School of grammar. This Durga is quoted by Hemacandra and is assigned to the 8th century A. DT.

1. Ngh. I. 14.

2. The Nirukta, Bib. Ind. ed. I. 147.

3. Op. cit. pp. i. 174, 240, 245, 246 etc.

4. The Nirukta, Bib. Ind. ed. I. 95.

5. Vedic and Technical Literature, Part I. p. 11, ed. 1879.

6. The Nirukta, Bib. Ind. ed. p. i. 112.

7. Belvalkar, Systems of Sanskrit Grammar, p, 87.

(3) But the most important of all these commentators is Durga. He seems to be later than Devarajayajvan who is familiar with the then extant commentaries on the Vedas, the Nighantu, and the Nirukta, and who does not mention Durga in the long list of tha authorities used by him for the purpose of his own work. Although a conclusion based on the argument of silence is not cogent, yet in this particular case, it is justified to assume that Durga is not refer- red to because he was posterior to, or a contemporary of, Devaraja. The latter made an exhaustive study of the commentaries on the Nighantu aud the Nirukta and could not have ignored the very important work of the former. Durga would also be later than Skandasvamin. Durga's commentary is published, and has super- seded the works of his predecessors. His work is important for two reasons: (1) he is a later commentator, and therefore represents a fuller development of the traditional interpretation of the Nirukta; (2) the very fact that it has survived at the cost of earlier commentaries indicates its importance. We shall therefore examine his work somewhat in detail.

Date of Durga.

It has already been pointed out that in all probability Durga is posterior to, or a contemporary of, Devarajayajvan, and therefore later than the beginning of the 14th century A. D. However, Durga's upper limit can be determined almost with certainty. A manuscript * of his commentary in the Bodleian Library is dated 1387 A. D. The date is genuine and is accepted as such by Professor A. B. Keith.2 The manuscript was copied at Bhrguk§- etra in the reign of Maharana Durgasirhhavijaya. Thus he could not be later tHan 1387 A. D. It is not definitely known as to which particular site was represented by Bhrguksetra but probably it is to be identified with the present Broach.3 As Durga wrote his commentary in a hermitage near Jammu, a place not easily accessible in the absence of modern. means of communications, the migration of the ms. of his commentary to Bhrguksetra

1. MS. Wilson 475.

2. See Catalogue of Sanskrit Manuscripts in the Bodleian Library, Toi ii, p. 108.

3. See, The Imperial Gazatteer of India Vol., IX. p. 16.

presupposes the lapse of half a century at least in order to account for the spreading of his fame as a commentator from the isolated heights of Jammu to the plains of Bhrguksetra. It will riot be far from the truth, therefore, to place Durga about the </