Indirect, yet significant testimony that this double character of the AV. was clearly established in Brahman- ical times may be derived from the formation of the names of two apocryphal teachers. One is Bhisha^ Atharva^a, Ka//z. S. XVI, 3 (Ind. Stud. Ill, 459); the other is Ghora Angirasa, Kaush. Br. XXX, 6 ; Asv. Sr. XII, 13, 1 ; KMnd. Up. Ill, 17, 6 (cf. Ind. Stud. I, 190, 293). The formation Bhisha^* Atharva;/a is illustrated further by Fa.nk. Br. XII, 9, 10, bhesha^a;// ya atharva^ani ; and XVI, 10, 10, bhesha- gam vai devanam atharva^o bhesha^yayai*va*rish/yai2; cf. also the expressions samyu atharva^a, personified as a sage, Gop. Br. I, 2, 18, and atharvabhi/£ ^anta/^, Kaiu\ 1 25, 33. These names never, as far as is known, occur in inverted order : there is no Bhisha^ Angirasa, and no Ghora Atharva//a ; they reflect perfectly the individual character and the individual function of the two members of the compound atharvangirasa^.
It seems now, further, that the texts of the Atharva- sa/^hita mark this same distinction with no
angirasin uncertain touch. At AV. XI, 6, 14, four
the Atharva- Vedic mantra-categories are indicated by the
expressions, ri'kak, samani, bhesha^*a(ni), and
yagumshl. The choice of the word bhesha^a is certainly
eclectic and one-sided. The passage appeals to the auspi-
1 atharva^o veda^ . . . atharvawam ekam parva vya^aksha^a^ ; ahgiraso veda/& . . . angirasam ekam parva vya&iksha^a^. Elsewhere, aside from the Atharvan texts, the component parts of the dvandva atharvangiras are drawn asunder, But without accessory statements; thus Tait. Br. Ill, 12, 9, 1; NWsiwhapurvatapani Up. 5, 9.
2 A converse statement like bhesha^aw va angirasani, is, if we judge the matter aright, a counter-sense, and unheard of anywhere in Hindu literature.
3 So also 6anti, as the wife of Atharvan ; see Wilson's translation of the Vishwu-purawa, I, 110, 200; Bhagavata-pura^a III, 24, 24.
XXU HYMNS OF THE ATHARVA-VEDA.
1 angirasa'nam adyaiA pa^Hnuvakai/j svaha (XIX, 22, 1); sarvebhyo sngirobhyo vidaga/zebhya/fc sv£ha (XIX, 22, 18); atharvaraanaw/katurr^ebhyo/fc sva'ha (XIX, 23, 1).
2 Cf. Weber, lad. Stud. IV, 433 ff.
INTRODUCTION. XXlll
the subsequent emanation from these two of twenty Athar-
vanic and Angirasic descendant sages x, and finally, the pro-
duction by the Atharvans of the atharva;za veda, by the
Angiras of angirasa veda.
In another passage, I, 3, 4, the Gop. Br. also asserts the separate character of the Angiras and Atharvans ; the latter are again associated with bhesha^am, the former is made the base of a foolish etymology, to wit : bhuyisht/iam brahma yad bhrzgvangirasa/z, ye*ngiraso ye*ngiraso sa rasa^, ye*tharva/*o ye*tharva//as tad bhesha^-am.
As regards the chronology and cause of this differentia-
tion of atharvan and angiras the texts are
differentiation apparently wholly silent. The association of
of atharvan fo^k names (and later of the name bhrzgu
and angiras. v . r i
also) with the texts and practices of the fourth Veda may be sought in the character of these mythic beings. They are fire-priests, fire-churners 2, and the Atharvanic rites, as well as the house-ceremonies in general, centre about the fire, the oblations are into the fire. Fire-priests, in distinction from soma-priesfs, may have had in their keeping these homelier practices of common life. But whence the terrible aspect of the An- giras in contrast to the auspicious Atharvans ? In the hymn about Sarama and the Pa/zis, RV. X, 108, 10, Sarama threatens the Pa/zis with the terrible Angiras, angirasa.? fca. ghora^. This statement, wholly incidental as it seems to be, is, of course, not to be entirely discarded. More im- portant is the fact that Brzhaspati, the divine purodha (purohita), is distinctly angirasa. In Kaor. 135, 9, Brz'has- pati Angirasa appears distinctly as the representative, or the divinity of witchcraft performances. In the Mahabha- rata he is frequently called angirasa^ sreshtkak. In his function of body-priest of the gods it behoves him to
1 Doubtless by way of allusion to the twenty books in the existing redaction of the 6aunakiya-.yakha. The expression vims'mo * ngirasa^ is repeated Paw. V, 2, 37, as a designation of the twenty books of the ^aunakiya-^akha in its present redaction.
2 Avestan atar-, athra-van and Vedic athar-van may be derivatives from the root manth, math (mth) ' churn/ But the absence of the aspiration in atar- makes the doubtful derivation still more doubtful.
XXIV HYMNS OF THE ATHARVA-VEDA.
exercise against hostile powers those fierce qualities which are later in a broader sense regarded as Angirasic. Thus RV. X, 164, 4 = AV. VI, 45, 3\ certainly exhibits this function of the divine purohita, and the composer of AV. X, i5 6, when he exclaims, ' Prati^ina (;c Back-hurler") the descendant of Ahgiras, is our overseer and officiator (puro- hita) : do thou drive back again (prati'^i^) the spells, and slay yonder fashioners of the spells/ has also in mind the divine purohita 2. The stanza foreshadows the later forma- tion pratyangiras, discussed above. We look in vain, how- ever, for statements of the reason why the word atharvan should be especially associated with janta and bhesha^a, and must assume that this was accomplished by secondarily contrasting it with ahgiras after the sense of ghora, abhi^ara had incrustated itself over it 3. The uncertainty of all this does not endanger the result that at a comparatively early time the terms atharva^a^, in. the sense of * holy charms/ and angirasa^, in the sense of ' witchcraft charms/ joined the more distinctively hieratic terms rik&h, yag-uz/zshi, and samani, as characteristic types of Brahmanical literary performances. But this distinction was at a later period again abandoned ; in the end the name atharvan and its derivatives prevail as designations of the practices and charms of the fourth Veda without reference to their strongly diversified character.
The stem atharvan is modulated in a considerable variety of ways by derivative processes, the simple stem itself, or forms in the singular from it, being decidedly rare, and not at all early. I have noted Nr/si^hapurvatapani Up. I, 4, ngya^u^samatharvarupa^ surya/fc. Plural forms are less rare : atharva/zo veda^, 5at. Br. XIII, 4, 3, 7 ; atharvawam,
1 yad indra brahma;zas pate-bhidrohaw £aramasi, pra£eta na angiraso dvishata'ra patv awhasa^.
2 RV. IV, 50, 7-9 prescribes that kings shall keep in honour (subhrztam) a br/haspati, i. e. a Brahman purohita, in archaic language whose sense coincides completely with the later Atharvan ic notions. Barring the diction the passage might stand in any Atharva-ParLrish/a ; cf. below, p. lxviii, note.
3 A dash of popular etymology may have helped the process : a-tharvan, •not injuring;' cf. thurv in the sense of ' injure,' Dhatupa/^a XV, 62, and perhaps Maitr. S. II, 10, 1 ; also the roots turv and dhurv with similar meanings.
Introduction. Xxv
TB. Ill, is, 9, i ; atharvazza/z, Yank. Br. XVI, 10, 10. The derivative neuter plural atharvazzani (sc. suktani) is common, from AV.XIX, 23, 1 ; Tank. Br. XII, 9, 10 to W/ddhaha- rita-sa//zhita 111,45 (£ivananda, vol. i, p. 213), and later. The same stem, atharvazza, is used in the masculine singular, atharvazza.? (sc. veda/£) /£aturtha/z, Khand. Up. VII, 1, 2. 4 ; 2, 1 ; 7, 1 ; in the plural, mantra atharvazza/z, Ram. II, 26, 21. The stem atharvazza (without vrzddhi of derivation) is found Nrzsizzzhapurvatapani Up. II, 1, atharvazzair mantrai/i ; Mahabh. Ill, 189, 14 = 12963, atharvazza/^ (sc. veda/£). Still another derivative is atharvazza, in atharvazza-vid, Mahabh. XII, 342, 100=13259. The name atharva-veda appears about as early as the corresponding names of the other Vedic categories (rzgveda. &c), vSankh. 5r. XVI, 2, 10; Par. Grzh. II, 1, 7; Hir. Grzh. II, 19, 6; Baudh. Grzh. IV, 5, 1. The form employed in the Cainist Siddhanta is a(t)havvazza-veda (see below, p. lvi) ; that of the Buddhist scriptures is athabbazza-veda (ibid.).
In addition to the designations of the Atharvan discussed above there are still others, based upon different modes of 0th viewing this heterogeneous collection of Mantras.
designations A single passage, Sat. Br. XIV, 8, 14, 1-4 = Bz'zh. Ar. Up. V, 13, 1-4, seems to hint at the fourth Veda with the word kshatram. The passage is engaged in pointing out the merits of Vedic compositions, stated in the series uktham ( = rzk ; cf. vSat. Br. X, 5, 2, 2c), ya^u/k, sama, kshatram. Inasmuch as the first three ob- viously represent the trayi vidya, it is possible to view kshatram as epitomising the Atharvan 1. If so, the passage is of considerable interest, as it seems to view the fourth Veda as the Veda of the Kshatriyas. More precisely the passage substitutes the act of kshatra, i.e. the characteristic performances of the Kshatriya (through, or with the aid of
1 Cf. also Pra-ma Up. II, 6, where brahma and kshatra figure. Both together represent in the epics the best outcome of the life of a kshatriya, ' piety ' and * prowess.* It is possible to conceive the appearance of kshatra alone as an elliptic version of both brahma and kshatra, the two together being the out- come of the trayi preceding, rather than a supplementary statement of additional Vedic types of composition ; cf. Pra^na Up. II, 6. For brahma alone, see below, p. xxxi, note.
XXVI HYMNS OF THE ATHARVA-VEDA.
his purohita) as Atharvanic by distinction. Recently Pro-
fessor Weber x has emphasised the marked relation of the
Atharvan to the royal caste.
The text of the Sawhita abounds in ra^akarmam, ' royal practices,5 and Weber thinks that the name of Kaiuika, the author of the great Atharvan Sutra, points to a Kshatriya connection, since Ku^ika is identical with Visvamitra, and the latter, as is well known, stands forth among the ancient Vedic heroes as the representative of royalty. None of these points can be regarded as more than possibilities 2.
Two other designations of the AV. differ from all the
preceding in that they are the product of a later Athar-
vanic literary age, neither of them being found in the
Sawhita, and both being almost wholly restricted to the
ritual text of the Atharvan itself. They are the terms
bhrzgvangirasa// and brahma-veda.
The term bhrzgvangirasa^ is, as far as the evidence of the accessible literature goes, found only in Atharvan texts. Though bhrzgu takes in this compound the place of atharvan. the terms bhrzgava^ or bhrzguveda do not occur. The term bhrzgvangirasa^, however, is the favourite designation of the Veda in the Atharvan ritual texts3: it makes a show, in fact, of crowding out the other designations. Thus the Kaorika does not directly mention the Atharvan composi- tions by any other name (see 63, 3 ; 94, 2-4 ; cf. 137, 25 ; 139, 6), although vaguer allusions to this Veda and its adherents are made with the stem atharvan (59, 25 ; 73, 12 ;
1 Episches im vedischen Ritual, Proc. of the Royal Academy at Berlin, July 23, 1 89 1 ; nr. xxxviii, p. 785 ff. (especially 787, top) ; Ra^asilya, pp. 4, 23, note.
2 We may note also the prominence allowed in the AV. to the kind of performance called sava. These are elaborate and rather pompous bestowals of dakshiwa, rising as high as the presentation of a house (jalasava, IX, 3) ; or a goat with five messes of porridge, five cows, five pieces of gold, and five garments (a^-audana, IX, 4). There are twenty-two kinds of these sava, and the eighth book of the Kaurika is devoted to their exposition (Ke^ava 64-66 presents a brief catalogue of them). Revenues of this kind are not likely to have been derived from lesser personages than rich Kshatriyas, or kings.
3 In the Sawhita the stem bhrzgvangiras is never employed as the name of the Atharvan writings ; in AV. V, 19, 1. 2 the terms bhrz'gu and angirasa occur as the names of typical Brahman priests.