.4~CARYA-, N.

"MARVEL"
by
F. B. J. KUIPER

Leiden
1. The word agcarya- occurs for the first time in the passage Ka.th. Up. 2.7 agcaryo vakta kugalo 'sya labdha "gcaryo j~ata kugalanugi.stah "Merveflle qui l'explique et le poss~de, l'habile! Merveille qui le connait, instruit par un habile!" (Renou). Here the word need not be taken as an adjective, as it may be the substantive used as a predicate. 1 Although of unknown origin, it is evidently an Indo-Aryan word. As such it is likely to belong either to the class of verbals abstracts in -ya- of the type v.rtrat~rya(Wackernagel-Debrunner, Altind. Gramm., 11/2, p. 830), or to that of the adjectiva necessitatis of the type g~r.sa-haryd- "was sich auf dem Kopf tragen lasst" (op. l., p. 793). The earliest attempt at an explanation of the word is met with in Pan.ini 6.1.147, who mentions it among the words in which an s is supposed to have been inserted. Since this explanation fails to account for the meaning, Wackernagel-Debrunner, op.l., p. 829, took agcarya- as a formation like brahmacdrya- (AthS.): "in rig scheint die Interjektion kl. d.h zu stecken, danach ware die urspriingliche Bedeutung 'ein Tun, zu dem man ch ausruft" '. To this Edgerton objected that, though the etymology is ingenious, "unfortunately it is not clear that this interj, ever connotes surprise (ordinarily either joy or annoyance)". ~ It cannot be denied, however, that this etymology has the merit of explaining the word as a specifically Indian formation instead of having recourse to such farfetched connexions as that with OIcel. skars, n. "giantess" and Lith. k~ras "magic". 3 The last word, indeed, belongs to kurift, k~rti "to build" and shows the same semantic development as Skt. k.rtyd- "act, deed, magic, enchantment". ~ The connexion of agcarya- with the Lithuanian word 1 Wackernagel-Debrunner, Altind. Gramm., II/2, p. 830. 2 JAOS, 75 (1955), p. 64. 3 Walde-Pokorny, Vergleichendes WSrterbuch der idg. Sprachen, I, p. 518. 4 This is evidently what Mayrhofer meant by his reference to krtyd- (Kurzgef. etym. WSrterb., I, p. 83).

.g.,SCARYA-, N. "MARVEL"

137

leaves the exact mode of formation of dicarya- unexplained and disregards the fact that dgcarya-, unlike k.rtyd- etc., does not denote any kind of act. Wackernagel's etymology is however open to two far more serious objections than that raised by Edgerton. Firstly the Indian interjections are never used with carati, but always with the root k.r-: el. Rigvedic akhkhalfkftyd, hlh~-akrn, ot, hihk.r.nvatt, kikird krvu, cited krn. oti, and from the epics downwards huthkaroti, hftrhkdra-, hYtrhk.rti-, hdhd-k.rtvd, hdhdkdra-, hahdk.rta-, etc. In the later language occur such formations as aholdbhakdra- Mhbh. V. 131.25 "one who says 'aho! labha.hI', who is content with a small gain". Secondly, this explanation is implicitly based on the assumption that Bdhtlingk was right in positing a Sanskrit interjection ds, while it is evident that he was so only with some important restrictions. 2. From a lexicographical point of view we can distinguish two different interjections d.h in Sanskrit - a fact, which was already duly recognized by the Indian lexicographers. The first denotes not so much a "recollection" (Monier-Williams, cf. "sm.rtau, smaran,e" lexx.), as rather the sudden recognizance of a fact, the Aha-Erlebnis. in this function it is more or less synonymous with hanta, aye, etc. and occurs in a triple form, viz. as d, d.h, and drh. In many passages two or three of these forms are handed down as variant readings of the manuscripts. Cf. a sa e.sa dharmdran.yanivasi sarhghilako ndma du.st.a~akyabhik~u.h Padmapr~bh. 23.4 (v.1. 8.h), vddavighat.t.itend 'nena bhavitavyam 16.10 (v.1. d.h), d sai.sd bhdn..d~rasendyd duhita kumudvat~ndma 28.1 (v.1. d.h; cf. aye, iyarh s a . . . M.rcch. 1.53.1), 8, ayarh tdvad v.rk.savat.ikdpak.sadvdren, d 'tikrdmati 35.9, 8, tasya malatikadfttTsvayarhgrdhavinaya atmaAahkdm utpadayati "I see, his misbehaviour in violating the messenger of M a l a t i k ~ . . . makes him selfconscious" 21.12, a, devasena roditi 41.19 (v.1. a.h), a, yathaivam . . . vdyurh paryupaste 37.1 (v.1. a.h), (vicintya) d~h, idarh yaj~opav~tarh pramdn.asfttrarh bhavi.syati M.rcch. III. 15.2, drh, idarh tdvad asti mama ca X. 17.1, (dhy6tva) ~, tasminn eva surapan, e vism.rtam iti tarkayami Mattavilftsaprah. 10.9, (cirdt sm.rtvd) a.h, asti janasthdne pa~cavat.~ ndma goddvar~tat.oddega.h Mahav. V.9.2, d.h, divaukaso 'pi rajaviruddham anumodante 1.42.3 (v.1. aha!), 8 . . . pre.sito . . . ratha.h Abhi.sekan. VI. 11.2, 8, atra ca sam~pagho.se mama vayasyo nandagopa.h prativasati B~lacar. 1.18.4, a asty etan nimittam Uttarar. 1.2.12, drh, atthi khu edarh III.12.12, 5 6, asty etat Pratijfiayaug. II.9.3, ath idam Mudr. II.10.3, M.rcch. IX.34.6, arh evvarh M.rcch. II.0.10, 8, evam etat P~dat~5 Most edd. read hurh or omit it; arh (ed. Stchoupak) is incorrect since arh is not altered in Pr~krit, cf. e.g. ~ak. IV.0.22, VI.20.2 Pischel, M.rcch. II.0.10.

138

y.B.J. KUIPER

.ditaka 19.2 (v.1. dO), d evath kila tat K~w ad P~.n. 1.1.14, d eso ddsfe putto ~ak. VL18 (140).3 Capp. (dth Pischel VI.20.2), also v.1. VI.13 (135).2 (B/Shtlingk), (vicintya) dth, ajja un.a asathnihida hiaen, a IV.0.22 Pischel, (vicintya) dm (sic), n f m a m . . . VI1.14.7 Pischel (ore. Cappeller, Brhflingk, etc.), (prakagam) ath, ku~alath kumarasya Uttarar. Iii.25.6 (a.h, ed. Bomb. 1899, ed. Chakhamba Skt. Ser.), d.h, pun.yavdn asmi, d.r.st.a.h priyasuh.rn me Citragr~va.h Hitop. 1.29q- (p. 12,3 Peterson). Sometimes this function of a, ath, a.h is expressed more explicitly by adding j~atam, etc. More or less synonymous are such expressions as (vilokya) bhavatu vij~dtam Padmaprfibh. 21.7-9, (vicarya) bhavatu vij~atam Padat~.d. 113.1-2 (cf. 67.1-2), (vitarkya) hanta vij~atam 30.9-10. In Mudr. VI.14.1 a v.1. of (vicintya) athjhatam is dth bhavatv (evam); in Vikram. IV.9(21).2 (vicintya) bhavatu viditam etat is a variant of (vicintya) dthj~dtam. The v.1. dth (d.h) bhavatujgdtam Mudr. 111.23.2, VL14.1 may be a contamination of two variant readings. This usage is particularly characteristic of Kfifid~sa and of Mudr~r~ksasa, M.rcchakat.ika, Caturbhfi.ni and Prabodhacandrodaya. Cf. e.g. (sm.rtvd) athjhatam gak. III.2.2 Capp. (v.1. d), 6 (vicintya) dthjhdtam Sak. V.1.2 (Pischel ~ am!), Vikram. IV.21.2, M.rcch. 1.7.2, Mudr. 1.18.89, HI.23.2, VI.14.1 (v.1.), dth jhdtam Mudr. III.17.12 (vv. 11. d, d.h), Uttarar. VI.3.1, inscr. (CII, III, p. 59), (vilokya vicintya ca) a.h jhdtam Prabodh. III.4.17 (d Brockh.), 7 (vicintya) djhdtam Pfidat~.d. 32.4-5, (nirvarn.ya) dj~dtam 41.16-17, d.hj~dtam 110.2, (vilokya) kith ca tdvad ayam dvigna iva? d jgdtam 78.7-9, d jfzdtam Pratimhn. Vii.13.13, (in a stanza!) comm. ad Prabodh. 1.27, d vijgdtarn Padmaprfibh. 28.5 (v.1. d.h), (vimr~ya) d viditam 17.5 (v.1. a.h), P~dat~.d. 117.1-2, dth sm.rtam Mudr. IV.7.22 (vv. 11. d sm.rtirh k.rtvd, aye sm.rtamI), (dhydtvd) d d.r.st.am Dhfirtavit.asarhv~da 40.6, d g.rh~tam Padmapr~bh. 11.10 (v.1. a.h). This interjection is of course mostly met with in monologues of the dramatical literature. From the evidence adduced it is clear that a, ath, d.h are free variants, optionally used in the same function. They may be taken, accordingly, as three different realizations [a:], [~:], [a:h] of the same interjection. 6 Cf. the Latin variants a and ah. Phonemically they could only be rendered as a, ath, a.h. Although B/Shtlingk for practical purposes interpreted them as the sandhi forms of d, am, as (PW. i, col. 6 Pischelin his first edition gives a variant reading d (see his note ad p. 47, line 3), but omits this in the second ed. All edd. read drh, e.g. Cappeller, p. 27, line 19. Quotations from the Prabodhacandrodaya refer to the edition by K. Samba~iva ~fistri (Trivandrum Sanskrit Series, No. CXXII, Trivandrum, 1936). s Note dth.h, v.l. in Mudr. VI.14.1.

~.SCARYA-, N. "MARVEL"

139

581,667, 728, cf. also Amara iii.3.241 ds), he appears to have been fully aware of the purely theoretical character of these forms (see s.v. ds). The sandhi rules did not, indeed, apply to these words, which probably were followed by a short pause (cf. d.h jgdtam, d.h g.rMtam, and d prag.rhya sm.rtau Amara IIi.3.239). it is impossible, therefore, to derive adearyafrom this interjection, which besides does not express astonishment. 3. A different interjection d.h is used to express a strong feeling of annoyance or indignation (khede, pLddydm or krodhe in the commentaries). The Indian lexicographers rightly took this as a separate word, cf. e.g. Amara III.3.241 d prag.rhya sm.rtau vdkye 'pi, ds tu sydt kopapidayo.h, with K.sirasv~min's comment kope: dO rak.sas ti.st.hati (cf. p. 142); pf.ddydm: vidydmdtaram d.h pradardya n.rpaddn bhik.sdmahe nistrapd.h. That it is a separate word can also be inferred from the following facts: 1. The contexts in which it occurs (which will be discussed below) differ from those mentioned in the preceding section. 2. It expresses a much stronger emotion. 3. Its phonetical realization seems to have been [a :11] without variation. Sandhi spellings like ds ti.st.ha (M.rcch. p. 132,16 Stenzler, p. 363,556 Godbole), d a.najje (p. 3,16 Stenzler )were rightly objected to by Brhtlingk, PW, 1,728. The same objection must be raised against d atithiparibhdvini Sak. IV.0.28 Brhtlingk, 27 Cappeller (cf. d.h, katharh mdrh paribhavasi IV.0.25 Pischel), d k.satriyabat.o Mahfiv. Ii.48.2 Todar Mall (cf. also (h)d durdtman in a spurious passage, p. 289), d, apadhvarhsa DOtav. 1.5.7, Bfdacar. II.3.2, etc. ~ It is accordingly equally impossible to base the assumption of an interjection *ds on this word d.h. Since the interjection is exclusively used in a few stereotyped kinds of phrases, it can best be studied by considering more closely these categories. A) Only comparatively seldom is it followed by a sentence which states the exact cause of the feelings of annoyance or anger. Cf. &h, daktir asti kasya cid videhardjasya... Mahhv. iII.16.1, dO, dakyarh ndma... IV.54.9, d.h, mayy eva bhruku!idharah sarhv.rtta.h II.46.2 (cf. aye, mayy eva bhrukut.Mhara.h sarhv.rtta.h Uttarar. V.35.3), d.h, pratikfdika.h sarhv.rtta.h V.58.6, d.h, ka.st.dm daddm anuprapanno 'smi Uttarar. V.20.8, d.h ka.st.am Mrcch. IV.24.8, (sahasotthdya) dO, mama 'pi sattvair abhibhftyante g.rhd.h Sak. VI.30.1, d.h, abhd.syas tram DQtav. 1.36.2, d.h, tena ruddhd.h? Dfltagha.t. 1.16.3, d.h, atirdk.sasarh khalu te vacanam Madhyamavy. 1.24.4, d.h, v.rddhas tram, apasara 1.15.7, d.h, kadarthito 'ham ebhir.., pdpai.h Uttarar. 9 Prakrit ~ for &h in ~ avehi Svapnav~s. V.6.68; 8.1 (against d.h...apehi M~latim. V.25.1), ~ paoasiddhim puccha M~lav. 1.8.13 (v.1.) is probably incorrect. Cf. e.g. ~h a.nayye C~rud. 1.0.36, &h a.najje Mrcch. 1.8.31, ~7.h,kith .na dfkadi IX.5.28.

140

F.B.J. KtaVER

V.10.5, a.h, svayarh mrto 'ham Hitop. 1.59q- (p. 20,18 Pet.), a.h (v.1. arh) . . . ity atra "pi ka gastragrahan, apek.sa Prabodh. IV.10.17, a.h, sarvatra . . . ~trfigadhara Paddh. 177c, a.h, durbuddhavilasitarh narapagfmam Prabodh. IL21.3, d.h, prasiddhamahaprabhdvd sd ... II.25.11, d.h, rdvan,asya cak.survi.sayam dgatd kva ydsyasi Pratimhn. V.16.2. B) More frequently the cause of the indignation is expressed by a sentence introduced by ka- or katham (cf. a, quotiens... Ovid, Metam. 2.489, 491, 15.490, a, quam es similis patri 6.621). Cf. e.g. (sonmdtham) d.h, katham anukampate ndma Uttarar. V.12.1, (conmdtham) d.h, katham adyd'pi rdk.sasatrdsaO I. 50.1, d.h, katham atithirh mdrh paribhavasi ~ak. IV.0.25 Pischel, a.h, katharh katharh nirjaghana Abhi.sekan. III.7.1, a.h, katham etan marhsarh sulalitarh bhak.sayami Hitop. 1.4+ (p. 15,4 Pet.), a.h, katham asmdkam api kula~ladikam idan~th pariksitavyam Prabodh. II.6.12, a.h, ka e.sa Dhfirtasamfgama 1.18.6 (p. 75,8 Lassen), a.h, ka e.sa p a p o . . , sathv.rtta.h M~datim. V.26.1, a.h, ka e.sa kauverya.h . . . Mahfv. VI.27.1, a.h, ko 'yam agnir nama Uttarar. IV.10.13, a.h, kas tara 'ryaputra.h VII.5.2, a.h, ko n.rtyati, a m a m a . . . Mattavilfsaprah. 14.31 (read a.h, mama), a.h, ko mat ta.h, kasya va mada.h Pratij fifly. IV. 1.71, a.h, kasya mahgtraja.h B flacar. V.7.2, d<.h>, kasya vijhdpyam Dfitagha.t. 1.50.6, d.h, kasya daivatam Madhyamavy. 1.48.23, a.h, kim uktath bhavati Uttarar. IV.I.16, a.h, kim ucyate Hitop. 1.16q- (p. 9,14 Pet.), a.h, kim evam ucyate III.3-t- (p. 88,8 Pet.), (sakrodham) a.h, kim evam atimugdhau ganter api vibh?ta.h Prabodh. II.26.33, a.h, kim evarh garvayase? (sakrodham) are, ka i r a . . . II. 10.3, a.h, kirh n.a d?gadi mama vavahale Mrcch. IX.5.28, a.h, kena Mudr. III.10.6, (sakrodham) a.h, kena kathitam 111.17.9. Cf. a.h, kas, ti.st.ha/ kas, ti.st.ha! v.1. ~ak. VI.32.1. C) Still more frequent is the use of a vocative after a.h, somewhat in the sense of re, are (el. ~ ~eLZ~ Horn., quofugis ah demens Prop. 2.30.1, a! miser Hippomene Ovid, Met. 10.632); e.g. (savegam utthaya) a.h papa . . . kva yasi Uttarar. III.43.3, a.h papa Prabodh. 1.9.2, II.6.6; 8.4, Ill.9.20; 9.45; 14.3; 18.8, Uttarar. IIL43.3 (cf. re papa Mflatim. V.30.1, are papa Hitop. III.9 +), a.hpape, ti.st.ha, tist.ha Mahfv. 1.38,4. (sakrodham) a.h papa pigacak.rte . . . Prabodh. III.9.20, a.h papa pi~aca malapalikadhara IIi.9.39, a.h papa pakhan..ddpagada IiI. 13.5, a.h papa kamacan..dala IV.9.1, a.h papa mahamohahataka IV.5.14, a.h papa duratman 1.21.1, Mflatim. VI.0.2 (v.1.), a.h kapalikapasada duratman, apehi Mflatim. V.25.1, a.h tatapavada Uttarar. V.35.1, a k.satriyabat.o Mahfv. II.48.2 (of. are k.satriya.dimbha II.46.2), a.h asarhbaddhapralapin M.rcch. IX.18.22 (but in the Pr~krit passage IX.33.4 the edd. of Stenzler, Godbole and Parab read arh vu.d.dhakut.t.ani; w . l l . d.h, d), d.h purobhdgini ~ak. V.27.6 Pischel, d.h

.~SCARYA-, N. "MARVEL"

141

daivopahaM.h Mudr. 1II.1.5, d.h svastivdcanika dus.t.a Mah~v. 111.19.1, a.h du.st.a durmukha n . 19.3, a.h dhftrta Mattavilasaprah. 11.40 & 57, a.h vacale Prabodh. Vi.21.4, (saro.sam) a.h an.ayye Cfirud. 1.0.36, (sakrodham) ii.h an.ajje M.rcch. 1.8.31, a.h duratman Malatim. V.31.1, VI.0.2, d duratman Mah~tv. p. 289 Todar Mall, a.h duratman Carudattahataka M.rceh. IV. 17.1, a.h dur~trnan rak.sah. Vikram. IV.5.3 ( = 0.53), a.h duratman pakhan. .dacan..dala M~latim. V.29.1, a.h duratman ks.atriyapasada Mah~tv. 1II.31.1, a.h duratman nair.rtapasada V.18.1, a.h duratman samarabMro Abhi.sekan. V.11.1. Although the word with which the person is addressed has mostly, of course, a depreciatory meaning, this is not necessarily so, cf. e.g.a.h priye, kirn ucyate ekam utpattisthdnam iti. Nanu janaka eva 'srndkam abhinna.h Prabodh. I. 16.5. More important is the fact that, while are can introduce all sorts of depreciating terms, a.h seems as a rule to demand a following vocative which explains the emotion expressed by the interjection, it cannot be matter of mere chance that these vocatives are nearly always adjectives. As a rule other particles are used before substantives, e.g. are (which, of course, is also allowed before adjectives). There is however this additional difference between ah and are that the former seems to have expressed indignation, 1~ while the second rather expressed a mere contempt. The last fact accounts for the circumstance that we never meet with the association a.h mf~rkha instead of are mftrkha M.rcch. Ii. 12.9, re mftrkha Prabodh. II.8.5, aye mftrkha Mudr. I. 17.8,11 whereas the first tendency may explain, why mostly papa or duratman is inserted before substantives. The inverse word order, it is true, is occasionally met with, e.g.a.h svastiracanika du.st.a Mah~tv. II.19.1, a.h kapalikapasada duratman M~latim. V.25.1, but it is less common than, e.g., a.h duratman k.satriyapasada Mah~tv. III.31.1. Very rare is the use of a single substantive without a qualifying adjective, el. d < . h > k.satriyabat.o Mahhv. I1.48.21~ (but are k.satriya.dimbha Mah~tv. II.46.2, are brahmana.dimbha Malatim. V.28.1, sakrodham: are ana.dvan Mah~tv. III.18.4), (sakoparn) a.h dasie putta M.rcch. 1.8.49, a.h dasre dhre, kith bhan.asi III.21.11; 21.14 (but e.g. sakrodham : are re ka.nelrsuda IX.30.12). D) The passages where a.h is followed by an imperative are closely connected with the preceding category. Cf. a.h, ti.st.ha jalma M.rcch. 10 Cf. e.g., a~t ak.sepe, P.rthvidhara ad M.rcch. 1.8.49. 11 However are papa Hitop. III.9, re papa Mfilatim.V.30.1 occur beside a.h papa Prabodh. III.9.20, etc. as In the genitive the adjective is added inahoduratmana.h ks.atriyabat.oranatmaj~ata II.27.2, aho duratmanab kcatriyabat.ob sarvati~ayi caritam VI.3.1.

142

F . B . J . KUIPER

VIII.43.4, ah., ti.st.ha ti.st,ha kau.napdpasada Sak. VI.32.1 with a.h, kdpalikapasada duratman, apehi M~latim. V.25.1, a.h duratrnan, prahara, prahara V.31.1, a.h pape, ti.st.ha, ti.st,ha Mahav. 1.38.4, a.h duratman nair.rtapasada, ti.st.ha, ti.st.ha V.18.1, a.h duratman rak.sas, ti.st.ha, ti.st.ha Vikram. IV.5.3. Some more instances are: a.h ti.st.hedanfm Dfltav. 1.40.1,2, Pratiman. V.21.4, d < . h > apadhvarhsa Dfitav. 1.5.7, Balacar. II.3.2, a.h, kathaya, kathaya Mudr. 1.6.6, d.h, hanyatam ayarh vanara.h Abhi.sekan. III.15.5, a.h, nirvdsyatdm ayarh vdnara.h III.23.1, a.h, apan[yatam ayam M.rcch. X.52.10, a.h, apanfyatdm e.sa citrapat.a.h DQtav. 1.18.1. E) A particular usage is met with in Bfdacar. 1.16.4, III.12.1, IV.6.5, Madhyamavy. 1.33.2, where a(.h) introduces a stanza (cf. M.rcch. IX. 18.22), and in a.h, ita ita.h Abhi.sekan. III. 13.11. 4. With a single possible exception to be discussed below, both interjections are not attested before the classical literature. Since however such interjections may have existed at any period (cf. Hindi aha, aha, ahaha, ahaha, a, etc.), their absence from the Vedic texts is sufficiently explained by the particular character of the earlier literature. On the other hand, one Vedic text shows a peculiar usage of a particle am (arh?) before vocatives, which is not known from any other Vedic or classical text and, therefore, seems to have been a dialectal feature. Cf. sa hovaca '.sa.dha am aru.ne Jaim. Brahm. 1.2714, tarh hovaca 'rh ke~in 1.285 ~, tarh hovaca 'rh somya sutaputra II.43111, tarh hovaca 'rh rahasyo IIi. 1905. See Caland, JaiminfyaBrahman. a in Auswahl, p. 105 n. 3. It may be added that the Vedic interjection am, supposed to occur in the mantra drn agf.so ddhakama.h MS. 1.4.1 (p. 47, 10)13 is probably non-existent. The interpretation of dmag~so as dm a~Lso in the Padapa.tha is obviously at variance with the explanation of this mantra given in 1.4.5 (p. 53, 4f.), viz. ag(so vaid6hakamaydjamanam abhisarpanti. The mantra must apparently be read (against the Padapat.ha) d mai(so ddhakama.h, in accordance with KS. V.3 (p. 46,1). Cf. also KS. XXXII.3 (p. 21, 5), where it is explained as follows: d magi.so doha-

kdma ity adi.so vai dohakama yajamanarn upati.st.hante.

5. For the sake of completeness it should be observed that these interjections are distinct from am or area "yes" (ahg[kare), which is sometimes used in dramatical literature instead of ba.dham, atha kim, etc. The final consonant of am seems never to become an anusvara: the reading arh in Malavik. 1.3.20 (Tullberg, Calc. ed. of 1871, against am in modern edd.),
13 See Vedic Concordance, p. 177b, Vedic Variants, II, pp. 386, 387. The Aundh ed. reads dm a~f.solike Von Schroeder.

ASCARYA-, N. "MARVEL"

143

M.rcch. II.20.16 (Stenzler, Godbole, Parab) is apparently due to an incorrect application of the sandhi rules by the editors. Equally incorrect, on the other hand, is the reading am for [~ :] "I see" in Motichandra and Agrawala's edition of Padmaprabh. 16.10 la (in accordance with the ed. princeps; the only MSS. readings recorded in Loman's ed. are however a, a.h). Its use may be illustrated by the following quotations: (Vid~.saka.h) bhodi, kith mam? (ce.ti) ama C~trudatta III.16.9, (ce.ti) ka.nn,aflraa, sohadi ajj'aae pavarao? (Kar.nap~raka.h) arh, sohadi ajjaSe pavarao M.rcch. 11.20.16, kith brav~.si? etad api mfdadeviyarh dat.hyam iti. am(a) Padmapr~bh. 12.5-6, dma, d.r.st.a iti prokte BKSS. V. 114, ame 'ti ca maya prokte IX.70, area. saumya, sa eva 'ham XX.284, ama Svapnav~ts. IV.3.43, VI. 13. 13, C~rud. 1.0.50, Pratijfiay. ii.13.3, 111.1.7, ama bho Pratijfi~y. 1II.1.94; 3.2; 3.8, ama, bha!t.a Balacar. 1.19.40 and 19.42; III.3.1; 3.5, am Abhi.sekan. 11.0.21, Malavik. 1.3.20, ama piyasahi M~latim. II.0.42. Amara Ill.4.17 records am "evam" (K.sirasv~tmin: ahg~kSrarthe: arh kurma.h "evarh kurma.h"). Cf. Hemac. PariL III.93.95 and ama in Buddh. Hybrid Skt., Pali, and Prakrit. To take am(a), like Hindi, Kann. hh "yes" as a mere exclamation might perhaps seem the most natural explanation. In support of this idea it might pointed out, indeed, that in the Vedic language, where this word am(a) is unknown, om is used instead. Cf. e.g. sa hovaco 'm arun.e .SvB. 1.6.13 'and he [the monkey] answered "so be it, ~ru.ni"!' (The divergent reading sa hovaca om arun.e, Oertel, JAOS. 26, p. 194, is apparently incorrect.) This use was still known to Amara III. 4.13 om "evam" (K.sirasv. : orh kuru, evath kuru . . . evarh bad.ham ity adi ca), whose explanation recurs in the commentary on .SvB. om ity ahg?k.rtyai 'vain. In spite of the wellknown statement in AB. VII. 18.13 om iti vai daivarh, rathe 'ti manu.sam this use of om is quite normal, e.g. om iti hovaca gB. X.6.1.8, BAU. III.9.1, VI.2.1, etc. Now, since it cannot be doubted that om was in origin a mere exclamatory sound (cf. P W , I, 1122), a similar origin might seem plausible for am(a). There is however this difference between om and am that, while the first word regularly becomes orb before consonants, am remains unaltered. The disyllabic form ama (amma v.1. M~lavik. 1.3.20) may even owe its origin to the care taken to pronounce the full m-sound in am. If so, this case is not fully parallel to that of a.h, exclamation of indignation, for which an isolated v.1. aha occurs only once in one MS. in Mah~tv. 1.42.3 (although it must be admitted that a.h too is treated as a pausa form). Since the common expression of assent in Tamil is am "yes, 14 Thus also Pischel in Sak. V.1.2 and VII.14.7.

144

F . B . J . KOII'ER

so", we cannot avoid considering the possibility that am is a Dravidian loanword. Our quotations from the literature prior to Bhavabh~ti are restricted to the Bh~tsa plays, K~tlidasa, Mrcchakat.ika and Padmapr~tbh.rtaka. In the other works of the Caturbha.ni am does not occur. Now M.rcch. and Padmaprabh. (which both are ascribed to SQdraka) have some more colloquialisms in common, e.g. bhava "respected sir ''x5 (which is the Dravidian word for "elder male cross-cousin", Emeneau, Kotami, p. 262, nr. 2026) and kapotapalika-, which is a Dekkhan word according to Prthvidhara (ad M.rcch. 1.51.2). A Dravidian origin may be suggested for the interjections ammo M.rcch. 1.35.3, IV.4.2, etc., Sak., ammahe Mrcch. 1.32.12, IV.8.8., Abhi.sekan. VII.10.1, etc., ~ak. Vikram.), of. Tam. amma, an exclamation of surprise or wonder, ammaiyr, an exclamation of astonishment, amm~yr, an exclamation of grief, ammak6, ammavr, an exclamation of pity. Now, regionalisms proper to M.rcch> e and Padmapr~bh, may be supposed to have been Ujjayinisms, and the use of Drav. am "yes" in K~lid~sa, M.rcch. and Padmaprabh. might be explained as such an Ujjayinism (although its use need not have been restricted to that town). It is clearly impossible to propose a definite solution, but the possible historical implications of the use of am "yes" would seem to merit our attention. 6. Two conclusions emerge from the preceding discussion. I] agcaryacontains an interjection (and this is so far the most plausible suggestion that has been made), this can only have been d (not *ds), whereas the verb in that case must have been k.r- "to do". Now a as an exclamation of astonishment is admittedly unknown in Sanskrit, but such exclamations consisting of a single vowel may have existed at any time in colloquial speech. In the Vedic language its use, though not to express astonishment, is attested in b.rhati prastuta a ity uktva 'dim a dadita 'after the Prastava of the B.rhat (-S~tman) has been finished he should commence the ~,di having

15 Cf. e.g. C~rud. 1.17.3 and Bharata, Na~ya~astra17.72 (19.10) manyo bhaveti vaktavyah, ki~cid anas tu mari~a.h(freely quoted by K.sirasvamin ad Amara 1.6.12 as manye bhavo 'pi vaktavyah, kirhcid ~ne~u mari.sab), which corresponds to the use of these terms by the pariparAvaka and the satradhara in Abhi.sekan. I. 1.6,1.2.1 and 17/rubhafiga 1.1.7, 1.2.1. The interpretation as vidvan (Amara 1.7.12) or budhab (e.g. in R~ghava's N~narthamafijari 1736, thus supplementing Mafikha 855, ~a~vata, etc., where this meaning is lacking) must have been conjectured from the context and shows that the word was not known at the time or in the region of these lexicographers. le Cf. e.g. Pusalker, J. Andhra Hist. Res. Soc., XI (1937-38), p. 41, Wilh. Printz, KZ, 44 (1911)~ p. 92 ft.

. 7 ~ C A R Y A , N. "MARVEL"

145

said "~.h"' (JB.I.1305)37 That the exclamation is here used because it is suggestive ofadi-, a-da-, does not imply that the word was not in common use. So the sole possibility of explaining agcarya- would be to analyse it as a- and dcar-, the palatalized variant of(s)krn.oti. TM This would involve that a~carya- must necessarily have been created in the oldest period of Vedie Sanskrit, for verbal forms with skr- are already rare in the Rigveda and palatalized forms of k.r- occur only in a few fossilized survivals. But many antique words of a more colloquial character are not attested in literature before a comparatively late date. The proper meaning of agcarya-, accordingly, may have been "something at which the sound a must be uttered" (cf. aphfttkarya- Kath~s.) and nothing prevents us from assuming that in its earliest occurrence (see p. 136) a last trace of its original adjectival use has been preserved. In spite of all the difficulties involved this is so far the most acceptable explanation that can be suggested.

Additional Remarks: The explanation of Skt. bhava- as a Dravidian loanword (p. 144) has been abandoned in A Dravidian Etymological Dictionary by T. Burrow and M. B. Emeneau. Indeed, just as bhavuka- "sister's husband" is a Prakrifism for bhrat.rka- (Wackernagel, Album Kern, p. 152 = KI. Schr. p. 404; Altind. Gramm., II/2, p. 483), so bhava- must stand for Pkt. bMu-. In Sanskrit both words seem only to occur in the specialized meaning, which is also found in bhdma- (Bh~g. Pur.). On the other hand mama- "maternal uncle" is a Dravidian word, see BurrowEmeneau, Nr. 3945. - For d, drh, d.h see also J. K. Balbir, ABORI, 40 (1959), p. 95.

1~ Oertel, The Syntax of Cases in the Narrative and DescriptiveProse of the Brahmatu~, p. 260; Minard, Trois Enigmes, w418, suggests that the pragrhya interjection a may have been used here. xs Thus also, e.g., Pokorny, Idg. etymol. W6rterb., p. 642.