The term bhrigvangirasak is, as far as the evidence of the accessible literature goes, found only in Atharvan texts. Though bhvzgu takes in this compound the place of atharvan, the terms bhyzgava or bhviguveda do not occur. The term bhvzgvangirasah, however, is the favourite designation of the Veda in the Atharvan ritual texts?: it makes a show, in fact, of crowding out the other designations. Thus the Kausika 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 (59,253; 73,123
* Episches im vedischen Ritual, Proc. of the Royal Academy at Berlin, July 23, 1891; nr. xxxviii, p. 785 ff. (especially 787, top); Ragastiya, pp. 4, 23, note.
* 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 (salasava, IX, 3); or a goat with five messes of porridge, five cows, five pieces of gold, and five garments (agaudana, IX, 4). There are twenty-two kinds of these sava, and the eighth book of the Kausika is devoted to their exposition (Kesava 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.
* In the Samhita the stem bhv7gvangiras is never employed as the name of the Atharvan writings ; in AV. V, 19, 1. 2 the terms bh7zgu and Aagirasa occur as the names of typical Brahman priests.
INTRODUCTION, XXVIl 125, 2.) The term also occurs in Vait. Su. 1, 5; Gop. Br. I, 1, 39; 2, 18 (end); 3, 1. 2. 4, and it is common in the Parisish¢as (see Weber, Omina und Portenta, p. 346 ff. ; Verzeichniss der Sanskrit und Prakrit Handschriften, I], 89 ff.), and the Anukramazi. No valid reason appears why the term bhvzgu has succeeded in encroaching so far upon the term atharvan. The following may, however, be remarked. The three words atharvan, angiras, and bhvzgu are in general equivalent, or closely related mythic names in connection with the production or the service of fire. Occasionally in the mantras (RV. X, 14, 6) they are found all together +, or bhyzgu is found in company with atharvan (RV. X, 92, 10), or with angiras (RV. VIII, 43, 13). This interrelation of the three names continues in the Yagus and Brahmava-texts, but in such a way that the juxtaposition of bhvzgu and angiras becomes exceedingly frequent ?, broaching in fact on complete synonymy. The latter is reached in Sat. Br. IV, 1, 5, 1, where the sage A yavana is designated either as a Bhargava or as an Angirasa*. It is conceivable that the frequency of this collocation sug- gested to the Atharvavedins a mode of freshening up the more trite combination atharvangirasah ; of any reason for a conscious preference of the word bhvigu the texts show no trace *.
The term brahma-veda whose origin is discussed below (p. lxv) likewise belongs to the sphere of the Atharvan ritual. Outside of the Atharvan there is to be noted only a single, but indubitable occurrence, Sankh. Gvzh. I, 16, 3.
' Cf, Weber, Verzeichniss, II, 46.
? E. g. Tait. S. 1, 1, 7, 2; Maitr.S.1,1,8; Vag. S.1,18; Tait. Br.I, 1, 4,8; Ill, 2, 7,6; Sat. Br. I, 2,1, 13; Katy. Sr. II, 4, 38; Apast. Sr.I, 12, 3; 23,6; YVaska's Nigh.V, 5; Nir. XI, 18. The juxtaposition of bhy7gu and atharvan is decidedly rarer in this class of texts (e. g. Apast. Sr. IV, 12, 10); that of bhvigu and afigiras continues in the Mahabharata, and later; see Pet. Lex. s.v. (col. 364, top). ;
? Cf. similarly Dadhya#é Atharvama, Tait. S. V,1, 4, 4, with Dadhyaas Aagirasa, Pavtk. Br. X11, 8, 6.
* A statement like that of the late AUlik4 Upanishad to, that the Bhrzgu are foremost among the Atharvans (atharvamo bhrigiittama’), if it is taken seriously at all, reflects rather the result than the cause of the substitution of the name bhrigu for atharvan.
XXVIil HYMNS OF THE ATHARVA-VEDA..
Even in the Atharvan Upanishads the term is wanting’. The earliest occurrences of the word, aside from Sankh. Grth., are Vait. Sd. 1, 1; Gop. Br. I, 2,16. The word is common in the ParisishZas.
We may note finally the terms pav#kakalpa and pa#ka- kalpin. They do not refer directly to the Samhitas of the AV., but are both bahuvrihi-compounds designating ‘one who practises with the five kalpas of the AV.,’ i.e. Atharvan priests. Thus the words were first explained by the author, Journ. Amer. Or. Soc. XI, 378; Kausika, Introduction, p. lvii. Cf. also Magoun, The Asuri-kalpa, Amer. Journ. Phil. X,169. They are very late: they do not occur in the Siatras or Brahmavza of the AV., nor, as far as is known, in the literature proper of that Veda. They appear as the titles of scribes of Atharvan texts, see Kausika, Introduc- tion, p. ix; Weber, Verzeichniss der Sanskrit und Prakrit Handschriften, II, 96. But they are sufficiently attested outside of the Atharvan, in the expression, pazkakalpam atharvazam, Mahdbh. XII, 342, 99 = 13258, and in the Mahabhashya (Ind. Stud. XIII, 455).
II. THE POSITION OF THE ATHARVA-VEDA IN Hinpu LITERATURE IN GENERAL.
In addressing oneself to the task of characterising the
estimate which the Hindus placed upon the Atharvan
Statement texts and practices, it is especially needful to
of the take a broad, if possible a universal view, of
problem. the statements of the Vedic and mediaeval
texts bearing upon the question. The Atharvan is
' The word occurs in certain doubtful variants of the text of the Musdaka Up.; see Ind. Stud. I, 301, note. In Ram I, 65, 22 brahmaveda is contrasted with kshatraveda, just as at Mahabh. VII, 23, 39=988 brahma veda with dhanurveda. In such cases the word brahma is not to be referred pregnantly to the fourth Veda, but to Brahmanic religion in general represented by the first caste, the science of war being in the hands of the second, or warrior-caste. Cf. below, p. xlii,. The word brahmavid, Mahabh. III, 2625 (Nala 14, 18, brahmarshi),
however, seems to mean ‘skilled in sorcery,’ and may contain an allusion to the AV. -
Introduction. Xxix
a sacred text in more than one respect: aside from the materials which it shares with the Azg- and Yagur- vedas, many of its hymns and practices are benevolent (bheshaga) and are in general well regarded, though even these, as we shall see, do not altogether escape the blight of contempt. Many hymns of the AV. are theosophic in character: on whatsoever ground they found shelter in the Atharvan collections they cannot have been otherwise than highly esteemed. The class of charms designed to establish harmony in family and village life and reconciliation of enemies (the so-called sdsmanasy4ni, p. 134 ff.), and the royal ceremonies (ragakarmavi), are obviously auspicious in their nature. Even the sorceries of the Atharvan neces- sarily show a double face: they are useful to oneself, harmful to others. According as they are employed objectively and aggressively, they are a valuable and forceful instrument for the benefit and aggrandisement of him that employs them ; according as one suffers from them subjectively and passively, they are dreadful and contemptible. This con- flict of emotions lasts throughout the history of the recorded Hindu thought ; the colaur of the Atharvan remains change- able to the end, and is so described in the final orthodox and stereotyped view that it is used ‘to appease, to bless, and to curse?.’ The fact, however, is that there must have arisen in the long run a strong wave of popular aversion against the Veda, whose most salient teaching is sorcery. This appears from the discussions of the Hindus themselves as to the orthodoxy of that Veda*; from the conscious efforts of the later Atharvan writings to vindicate its char- acter and value; from the allegorical presentation of the Atharvan as ‘a lean black man, sharp, irascible, and amorous *;’ and many occasional statements of the Vedic and classical texts. The history of the relation of the Atharvan to the remaining Hindu literature is, however,
* santikapaush/ikabhi4aradipratipadaka, Madhusiidanasarasvati (Ind. Stud. I, 16); Kesava to Kaus. 1,1 ; Deva to Katy. Sr. XV, 7, 11, and elsewhere.
* According to Burnell, Vamsabrahmama of the Samaveda, p. xxi, the most influential scholars of Southern India still deny the genuineness of the Atharvan.
* Kagendralalamitra in the Introduction to the Gopatha-brahmawa, p. 4.
XXX HYMNS OF THE ATHAKVA-VEDA.
still unwritten, and the following pages aim to supply the necessary data. In the hymn to the Purusha; the primaeval cosmic man (RV. X, 90, 9), the three Vedic categories, x7kak samani a! ... yaguh, are mentioned; a fourth term, an ‘e khandamsi, is generic, embodying the metrical the Rig- canons, or the metrical compositions as a whole, but the opportunity to mention the Atharvan is neglected’. The names atharvan, angiras, and bhvigu, which occur frequently elsewhere in the RV., designate mythic personages, intimately connected with the production of the fire, and the soma-sacrifice ; nowhere do they seem to refer to any kind of literary composition. Even the expression brahmavzi, used in connection with atharvan, RV. I, 80, 16, can claim no special interest, be- cause, as will appear later (p. Ixvi), the word brahma is never used as a specific designation of Atharvan charms. No great importance is to be attached to this silence; the praises to the gods in connection with the great soma- sacrifices, with their prevailing mythical colouring, darkened very often by priestly mysticism, offer but scant occasion for the mention of sorcery, or the plainer practices of every- day life. Yet sorcery and house-practices there were in India at all times?. The failure of the Rig-veda to mention any systematic redaction of charms by a collective name like atharvangirasak must be gauged by the slenderness of its opportunities to mention the Veda as a generic name (cf. VIII, 19, 5), or Vedic collections or redactions in par- ticular (X, 90, 9)°. There is no proof that even the oldest
* Kor RV. X, 71, £1, which also hints at the three Vedic types, and the brahma that embraces them all, see the full discussion below, p. xiv ff.
? Cf. e. g., RV. I, 191 ; VII, 50, and especially VII, 104, 16.
* The familiar mention of compositions called ri#, siman, uktha, stotra, sastra, &c., does not, it is important to note, refer to collections at all, but to types of poetic productivity; they are moreover all of them such as were dis- tinctly connected with the soma-sacrifice. Their presence simply accentuates the preoccupation of the body of the Rig-vedic collection with the great priestly sacrifices, and the consequent absence of the more general terms for Vedic classes of writings. The stem yaguA/, in the sense of collection of formulas of the Yagur-veda, occurs only in the above-mentioned passage, X, 90, 9.
Introduction. Xxx1
parts of the RV., or the most ancient Hindu tradition accessible historically, exclude the existence of the class of writings entitled to any of the names given to the Atharvan charms; there is no evidence that these writings ever differed in form (metre) or style from those in the existing Atharvan redactions; and, finally, there is no positive evidence—barring the argumentum ex silentio— that the names current in other texts as designations of Atharvan hymns (bheshagani, atharvazak, angirasak, &c.) were unknown at the earliest period of literary activity. On the other hand, the existing redactions of the AV. betray themselves as later than the RV. redaction by the character of the variants in those mantras which they share with the RV. As regards the AV., the stanza X, 7, 20 presents the four Vedic categories, rzkah, yaguh, samani, and atharvan- Position of girasa#, the last the traditional name of the the AV.in Saunakiya-version. The same tetrad is intended we i. ©at XI, 6, 14, where the narrower term bhesha- Saunakiya- gA(ni) takes the place of atharvangirasaz. At samhita. ‘
XIX, 54, 5 the mention of atharvan and angiras, though not directly referable to the AV., certainly suggests it, because stanza 3 speaks in the same strain of the rzkak and yaguh; and in XIX, 22,1; 23, 1 (parisishfa in character; cf. above, p. xxii), the angirasani and Athar- vazani (sc. siktani) are mentioned separately. Otherwise this text also fails to present a fixed name for the type of literature known later as Atharvanic'. The Atharvan is very much in the same position as we shall find the Yagus- texts: the three Vedas are mentioned, often in connection with other more specific forms and designations of prayer and sacerdotal acts, but the Atharvan is omitted. The impression left in both cases is by no means that of con- scious neglect or contempt, but rather of esoteric restriction to the sphere of the great Vedic ritual (srauta)*. Thus