Pramāṇa · Search · About · API

The Religion and Philosophy of the Veda and Upanishads v1.c23: Chapter 15. The Demons

The Religion and Philosophy of the Veda and Upanishads v1.c23

Language
English

Contents

Chapter 15. The Demons

§ 1. The Enemies of the Gods

v1.c23.s1.1

THE enemies of the gods *par excellence* throughout the Yajurveda, the Atharvaveda, and all the subsequent Vedic literature are beings called Asuras, but this connexion can be traced only in the latest parts of the Rigveda and even there but occasionally. In the singular the meaning is found but three times certainly, used of Varcin,¹ Pipru,² and the wolfish Asura;³ a fourth case is very doubtful, and may instead refer to Varuṇa.⁴ In the plural there are at most eight cases, predominantly in the tenth book: in them the gods as a body appear opposed to the Asuras as in the later texts, or Indra scatters them.⁵ The term Āsura is applied also to Namuci, and to Svarbhānu, while Indra, Agni, and the sun are called slayers of Asuras. In the Atharvaveda,⁶ on the other hand, the singular is used in the hostile sense three times, the plural thirty times, and the application of the term to the gods, which is found very occasionally in the later literature,⁷ is confronted with a regular application to the enemies of the gods. In the Rigveda, on the contrary, it is the normal attribute of Varuṇa and, more rarely, of other high gods.

v1.c23.s1.2

¹ RV. vii. 99. 5.

v1.c23.s1.3

² RV. x. 138. 3.

v1.c23.s1.4

³ RV. ii. 30. 4.

v1.c23.s1.5

⁴ RV. x. 124.

v1.c23.s1.6

⁵ RV. viii. 96. 9; 97. 1; x. 157. 4; 53. 4; 82. 5; 124. 5 (dubious); 151. 3; *Āsura*, v. 40. 5, 9; x. 131. 4; *Asurahan*, vi. 22. 4; vii. 13. 1; x. 180. 2. Cf. Macdonell, *JRAS.* xxvii. 168-77.

v1.c23.s1.7

⁶ Von Bradke, *Dyāus Asura*, pp. 101 ff.

v1.c23.s1.8

⁷ Ludwig, *Rigveda*, iv. p. xvii; TS. i. 6. 6.

v1.c23.s1.9

The theory of Haug⁸ that the change of meaning of Asura as between Asura in India and Ahura Mazdāh in Iran was due to a divergence in religion in the Indo-Iranian people, which ended in the schism of the two nations, is hopelessly opposed to the fact that the change of meaning takes place in India itself, and, since Darmesteter,⁹ the theory has prevailed that the change by which Asura became the name of demons in India, while in Iran the Devas became demons, is an internal change of meaning in the two languages, brought about by causes which can be made more or less clear. In the case of India the development of Asura into a hostile sense is traced to its use in connexion with the word Māyā, ‘wile’¹⁰ or ‘occult power’, assisted by the popular derivation which saw a negative in the first letter of the word, and vaguely conceived the meaning as ‘not heavenly’, this view leading to the creation of the term *sura* for god in the Upaniṣads, and perhaps also by the absence of any collective name for the enemies of the gods proper. The change of Deva was by Darmesteter attributed to the misunderstanding under the régime of Zoroaster of the old phrases ‘wrath of gods and men’, and ‘trouble made by gods or men’, but this argument is of very problematic value indeed.

v1.c23.s1.10

⁸ Criticized by Justi, *GGA.* 1866, pp. 1466 ff.; Ludwig, *Rigveda*, iv. pp. xvii f.

v1.c23.s1.11

⁹ *Ormazd et Ahriman*, pp. 266 ff.; Geldner, *Ved. Stud.* i. 142.

v1.c23.s1.12

¹⁰ Hillebrandt, *VOJ.* xiii. 320. The derivation from *mī*, ‘injure’ (Geldner, *Glossar zum RV.*, p. 135), is clearly wrong; it is from *mā*, ‘fashion’, RV. v. 85. 5; i. 159. 4; iii. 38. 7; ix. 83. 3; cf. Neisser, *Festschrift Hillebrandt*, pp. 144 ff.; Schayer, *Mahāyān. Erlösungslehre*, pp. 22 f.

v1.c23.s1.13

Hillebrandt¹ has recently opposed the prevalent doctrine, and asserted the opinion that the difference of view is due to religious relations with the early Iranians before the reform of Zoroaster, but after the period of the Rigveda in its main portions. He points out that the real cleavage in view is considerable: that the fall of the Nāsatya, Indra, Vāta, and Çarva to be demons as Nā̊ṅhaithya, Indra, Vātō, and Çauru (Saurva) is significant; that the Vedic Kavis and the Uçij family of priests fell in rank; and that there is no trace of the gradual change of sense, which would be expected, if the prevailing view were correct. He also points to the fact that among the names of Asuras, who appear in the accounts of the Brāhmaṇas, there are some with an Iranian aspect: namely Çaṇḍa and Marka, the latter being Avestan Mahrka,² Kāvya Uçanā, who is comparable with Kai Kāōs, Prahrāda Kāyādhava,³ perhaps Avestan Kayadha, and Sṛma,⁴ Iranian Salm, son of Thraētaona. The evidence is, however, clearly inadequate to prove the thesis, and the efforts of Hillebrandt to show that in the Rigveda occur names of Persian princes who patronized the singers, a fact which would indicate the possibility of close intercourse during the late Rigveda period natural, must be definitely taken to have failed to produce conviction.⁵

v1.c23.s1.14

¹ *Ved. Myth.* iii. 430 ff.; cf. Moulton, *Early Zoroastrianism*, pp. 115, 140; von Bradke, *Dyāus Asura*, pp. 108 f. Greek influence for Vedic times (V. Smith, *JASB.* 1889, p. 133; 1892, p. 60) is implausible.

v1.c23.s1.15

² TS. iv. 4. 10. 1; MS. iv. 6. 3; ÇB. iv. 2. 1. 16. Hillebrandt treats him as Death; these demons receive the Çukra and Manthin cups at the Agniṣṭoma, invented by the gods to rid themselves of them, and the Manthin cup brings disease. Çaṇḍikas in RV. ii. 30. 8 are foes of Indra. The Avestan Karapans who are in evil odour he connects with the *kalpa* of the Vedic priests.

v1.c23.s1.16

³ TB. i. 5. 9. 1.

v1.c23.s1.17

⁴ MS. iv. 2. 9.

v1.c23.s1.18

⁵ Macdonell and Keith, *Vedic Index*, i. 29, 349, 450, 509, 518; ii. 63. Jackson (CHI. i. 319 ff.) adds nothing new to Hillebrandt’s conjectures. It is noteworthy that Bloomfield’s evidence (*Rig-Veda Repetitions*, p. 645) shows book vi as by no means early as it should be if it deals with events in Iran before the invasion of India.

v1.c23.s1.19

Another suggestion is made by von Schroeder,⁶ based on von Bradke’s view⁷ of the term Asura, as applied to the gods, as meaning not ‘spirit’ as usually held, but ‘master’ or ‘lord’, and cognate with Latin *erus*. The name would thus normally have applied specifically to Dyaus, as the great sky god, or rather the god created by the moral sense of men, and also regarded as a nature deity, but as usual with the Vedic poets would have been extended by henotheism to other gods, such as Varuṇa, Parjanya, Indra, Agni, and Savitṛ. The degradation of such a term seems unlikely, and accordingly von Schroeder believes that there originally existed two distinct words, *asura*, ‘lord’, and *asura*, ‘spirit’. In India the latter use prevailed, and the older or contemporaneous *asura*, ‘lord’, was given up, because perhaps of risk of confusion with the other term, while the Iranians retained the use of Ahura as ‘master’. This suggestion is ingenious, but purely hypothetical, and hardly necessary to explain the facts. Moreover, the idea that *asura* as ‘spirit’ is naturally applied to evil beings is decidedly fantastic. The precise sense which *asura* had in the minds of those who used it is unknown to us, but there is nothing to show that it had any connexion with the worship of the spirits of the dead, as frequently suggested.¹ On the contrary, even from its possible connexion with *asu*, ‘breath’, the word may rather have meant that which is essentially alive and possessed of power and strength.²

v1.c23.s1.20

⁶ *Arische Religion*, i. 317 ff.

v1.c23.s1.21

⁷ *Dyāus Asura*, pp. 29 ff.

v1.c23.s1.22

¹ Thus Moulton (*Early Zoroastrianism*, p. 150; *Early Rel. Poetry of Persia*, pp. 34 ff.) argues that the high Ahura worship comes from ancestor worship as contrasted with the inferior worship of Daēvas, nature deities. But he himself (*Early Zoroastrianism*, pp. ix, 344, &c.) renders Ahura by ‘lord’, which disposes of any connexion with ancestor worship. Varuṇa sufficiently refutes for India and Ahura for Iran Schroeder’s doctrine that morality and ancestor worship are more closely related than the worship of the gods and conduct.

v1.c23.s1.23

² Cf. Oldenberg, *Rel. des Veda*², p. 160, n. 2. On the whole, the sense ‘lord’ seems adequate for Veda and Avesta alike. On the Gaulish Esus, see Rhys, *Celtic Heathendom*, pp. 60 ff.

v1.c23.s1.24

The view of these personages taken by the Brāhmaṇas is that they are equally sons of Prajāpati, though born of a less worthy part of the god, from the descending breath, not the mouth, and that they are in constant conflict with the gods, and have to be defeated by some ritual performances. They are associated with the darkness, untruth, and error, as opposed to the gods. The term ‘pertaining to the Asuras’ is freely given to any ritual performance which the priests do not approve. Thus the Asuras are said to have only a morning and afternoon pressing,³ not three as the Vedic Indians, a point which has some affinity to the Avesta, and the mode of making the grave differs.⁴ We have one fairly clear proof that Aryan enemies are included, for the Çatapatha Brāhmaṇa⁵ has preserved to us two barbarous (*mleccha*) words used by the Asuras, *he ’lavo*, which, despite various efforts to interpret them as Assyrian or something equally implausible, seem merely a Prāktitic version of the Vedic *he ’rayaḥ*, ‘O enemies’, pointing to an eastern dialect as in Māgadhī.

v1.c23.s1.25

³ TS. vi. 2. 5. 3.

v1.c23.s1.26

⁴ ÇB. xiii. 8. 2. 1.

v1.c23.s1.27

⁵ iii. 2. 1. 23, 24; Macdonell and Keith, *op. cit.* ii. 181, 279, 517.

v1.c23.s1.28

Among names of Asuras we learn of Kirāta and Ākuli,⁶ Araru (with which may be compared the Arurmaghas or Aruñmaghas, foes of Indra), Aru, Kustā,⁷ in opposition to Aditi, Etadu, from the use of *etad u* in a formula, Daivya as their messenger as opposed to Agni, Dābhi as the address of the Gāyatrī metre,¹ in place of Viçvakarman, Asita Dhānva, their king,² Pūru, the name of a Vedic prince, Vibhinduka and Viṣād. More important are the Kālakañjas,³ who figure in a story similar to that of Otos and Ephialtes, the bricks of the altar taking the place of mountains.

v1.c23.s1.29

⁶ PB. xiii. 12. 5; ÇB. i. 1. 4. 16; JB. iii. 168 ff.

v1.c23.s1.30

⁷ MS. iv. 2. 3.

v1.c23.s1.31

¹ TS. ii. 4. 3; MS. ii. 1. 11.

v1.c23.s1.32

² ÇB. xiii. 4. 3. 11.

v1.c23.s1.33

³ TB. i. 1. 2. 4–6; Eggeling, SBE. xii. 286.

v1.c23.s1.34

In comparison with the Asuras the Paṇis are unimportant demons, who play no great part in the literature after the Rigveda. They are the demons who withhold the cows, or, as it is differently put, the ghee in the cow, from Indra or his allies, Agni, Bṛhaspati, the Aṅgirases or Soma, and are overthrown by him. The word is normally plural, but a single Paṇi four times represents the group. The most probable explanation⁴ is undoubtedly that which sees in the Paṇis the personification of the enemy, who will not sacrifice to the gods or bestow gifts on the priests, and who is therefore an enemy of the gods and men alike, and has been brought into the old myth of the winning of the light in cow shape. Hillebrandt⁵ seeks to show that the Paṇis are an historical tribe, comparing the name Parnians, but the suggestion is most improbable.

v1.c23.s1.35

⁴ Oldenberg, *Rel. des Veda*², p. 143; Macdonell, *Ved. Myth.*, p. 157.

v1.c23.s1.36

⁵ *Ved. Myth.* i. 83 ff.; ZDMG. lxx. 512 ff.

v1.c23.s1.37

The Dāsas or Dasyus are also made into enemies of the gods, though, like the Paṇis, their primitive function was doubtless different; in their case it probably was that of aborigines, who opposed the Aryan advance, though Hillebrandt⁶ changes them into Dahae. That in many cases historic men may be meant when Dāsas are overthrown, is true; but gods of the defeated aborigines may also be denoted, and more generally powers of the air opposed to the gods: Dasyus seek to scale heaven,⁷ Indra vanquishes them from birth,⁸ he wins the sun and waters after defeating them,⁹ a Dāsa is husband of the waters,¹⁰ and the Dāsas have seven autumnal forts, doubtless in the air, not on the steppes.

v1.c23.s1.38

⁶ *Op. cit.* i. 95; (Kl. Ausg.), pp. 96 f.

v1.c23.s1.39

⁷ RV. viii. 14. 14; i. 33. 7.

v1.c23.s1.40

⁸ RV. i. 51. 6; viii. 77. 1–3.

v1.c23.s1.41

⁹ RV. i. 100. 18; x. 73. 5.

v1.c23.s1.42

¹⁰ RV. i. 32. 11.

v1.c23.s1.43

Of the individual names of the enemies of the gods Vṛtra ranks first; he is a serpent with power over the lightning, mist, hail, and thunder, when he wars with Indra; his mother is Dānu, apparently the stream or the waters of heaven, but he bears that name himself as well as Dānava, offspring of Dānu. His abode is hidden in the waters, but is also on a summit or on lofty heights, which suggest the waters of the air. He is by name the encompasser of the waters, rather than the holder back by congealing them: the cloud mountain is, therefore, said to be within his belly. He has ninety-nine forts which Indra shatters as he slays him. From the single Vṛtra the Vedic conception, as often, produces many Vṛtras, and we find also the plural used of foes who must be clearly human, perhaps, however, never without a sub-reference to Vṛtra, though, as the neuter is used in this way, it may be that the use is not simply a direct generalization of Vṛtra as a demon. The Brāhmaṇas, which tell many tales about Vṛtra,¹ make him out to be the moon, swallowed at new moon by Indra as the sun.

v1.c23.s1.44

¹ Hillebrandt, *Ved. Myth.* iii. 239 ff. Usener (*Götternamen*, p. 206) finds the Danaoi in the Dānava as demons, despite the change of quantity. Implausible is the view that Vṛtra is a creation of the imagination, *vṛtrahan*, as in the Avesta Verethraghna, denoting ‘assault repelling’ (Bartholomae, *Air. Wb.* 1420).

v1.c23.s1.45

Vala is a pale figure compared to Vṛtra: he is mentioned in the myth of the Paṇis: Bṛhaspati or Indra takes from him the cows which he had in his forts, his fences are burst by Indra, or his hole is opened. The word means literally covering, and is found often in this sense or in some cognate meaning. The fold of Vala is also mentioned, and he is clearly simply the personification of the pen in which the cows are supposed to be kept, as is indicated by the fact that he is not said to be slain, but to be pierced, broken, or cloven. In the post-Vedic mythology he appears in the epithet of Indra, ‘piercer of Vala’, and is deemed brother of Vṛtra, with whose myth, however, he had, it is clear, little originally to do.²

v1.c23.s1.46

² Hillebrandt, *Ved. Myth.* iii. 260–6.

v1.c23.s1.47

Arbuda appears seven times as a beast whose head Indra struck off: he seems to be cognate to Vṛtra with whom or Ahi he is mentioned.³ Svarbhānu,⁴ the Asura, is more interesting: he is clearly the demon who eclipses the sun, and who has to be overthrown by Atri or the Atris, and by Indra. Though he is several times mentioned in the Brāhmaṇas, his place is in post-Vedic mythology regularly taken by Rāhu. The Atharvaveda⁵ also knows of Grahas who affect the moon. Of Uraṇa who had ninety-nine arms we know no more.⁶ But we hear a good deal of Viçvarūpa, the three-headed son of Tvaṣṭṛ, slain by Trita and by Indra for the sake of his cows, and who in the Brāhmaṇas⁷ appears as the Purohita of the gods, though akin to the Asuras. Tvaṣṭṛ, it is said, sought to punish Indra for the death of his son, but his effort to exclude him from the Soma sacrifice was defeated by Indra, who insisted on taking a share in it. The legend tells also of the origin of three birds from the three heads of the demon, as they were struck off by Indra, and explains the use of each of the three heads for drinking respectively the Soma, the Surā, and taking other forms of nourishment.⁸ To Hillebrandt⁹ the moon seems to be meant by this form, and he lays just stress on the fact that the Brāhmaṇas are aware of the hostile character of the moon, which they often equate with Vṛtra, and which seems once regarded in a hostile light in the Rigveda; but this view cannot be regarded as probable as that which insists on the identity of the legend with those of Herakles and Geryoneus and Hercules and Cacus.¹⁰

v1.c23.s1.48

³ RV. ii. 11. 20; 14. 4; i. 51. 6; viii. 3. 19; 32. 2; x. 67. 12. Cf. the snake priest Arbuda Kādraveya, ÇB. xiii. 4. 3. 9.

v1.c23.s1.49

⁴ RV. v. 60; Rāhu occurs in AV. xix. 9. 10, a late passage.

v1.c23.s1.50

⁵ *Loc. cit.*

v1.c23.s1.51

⁶ RV. ii. 14. 4.

v1.c23.s1.52

⁷ TS. ii. 5. 1. 1. For the three heads, cf. Hopkins, *Origin of Religion*, pp. 297 ff.

v1.c23.s1.53

⁸ ÇB. i. 6. 3. 1 ff.; v. 5. 4. 2; TS. ii. 4. 12. 1.

v1.c23.s1.54

⁹ *Ved. Myth.* iii. 531 ff.

v1.c23.s1.55

¹⁰ Oldenberg, *Rel. des Veda*², p. 142, n. 6.

v1.c23.s1.56

Of the Dāsa enemies of Indra the chief is Çambara, the son of Kulitara,¹ who had 90, 99, or 100 forts and was the great foe of Divodāsa Atithigva. His historical reality seems reasonably assumed.² Pipru is styled a wild beast, an Asura, and a Dāsa, and is defeated for the sake of Rjiçvan: it seems unnecessary to see in him a spirit of the air.³ Dhuni, the roarer, and Cumuri, whose name is not connected with any known root, are apparently chiefs defeated for Dabhīti. Varcin is both Asura and Dāsa, but his 1,100 or 100,000⁴ warriors do not seem wholly mythical, and the names Anarçani, Ilibiça, Dṛbhīka, Rudhikrā, and Sṛbinda, which are those of other foes, seem human enough. On the other hand, Namuci, as we have seen, seems to be a demon, and Çuṣṇa has been explained not as a human enemy of Kutsa, but as a demon of drought from his name interpreted as scorcher.⁵ But this seems unlikely: he is child of the mist and moves in the water and a Dānava; if, therefore, he is a demon, he is rather the hisser, for Vṛtra’s hissing drove Indra away in fear, and this applies to other enemies than drought demons.

v1.c23.s1.57

¹ In this name Brunnhofer (*Arische Urzeit*, pp. 71, 72) finds as usual an ethnic reference without any conceivable ground.

v1.c23.s1.58

² He considered himself a godling, RV. vii. 18. 20.

v1.c23.s1.59

³ Macdonell, *Ved. Myth.*, p. 161.

v1.c23.s1.60

⁴ RV. ii. 14. 6; iv. 30. 15.

v1.c23.s1.61

⁵ Macdonell, *op. cit.*, p. 160.

§ 2. The Enemies of Man

v1.c23.s2.1

It is impossible, when we deal with the lesser demons of the Vedic religion, to ascertain with any reasonable certainty the origin of the different conceptions exhibited, for the lack of clearness of the notices given in our sources is in accordance with the fact that the demons are objects of aversion, and that therefore they are not minutely described, but only indicated in vague terms. It is probably true that in many cases the idea of such spirits is born of the idea of the hostile dead,⁶ or again the demons are the product of independent creative thought, corresponding to the ‘abstract’ deities, but they may also often be more naturalistic in their origin. It must be remembered that there are poisonous plants, that waters are often regarded as cruel and dangerous as well as kindly, as in the widespread belief that a river must be propitiated by the life of man at least once a year, that the savage character of the forest creates the view that it is inhabited by a dread spirit, that the hail is regarded as evil, and that trees are awe-inspiring⁷ and sometimes deadly in their fruits. The animatistic growth of such ideas as that of hostile spirits is as natural as the animistic, or the spiritist, but in the case of the demons the distinctions cannot be made with certainty in the confused figures which we find in the literature. Nor in the case of the demons can we overlook the fact that actual men may be included as well as spirits of the dead, and actual animals also be considered a very natural fact.

v1.c23.s2.2

⁶ As in Babylonian religion, Farnell, *Greece and Babylon*, p. 206. Cf. Arbman, *Rudra*, p. 169.

v1.c23.s2.3

⁷ Cf. Hopkins, *Epic Myth.*, p. 7, n. 3.

v1.c23.s2.4

The forms of the demons are conceived either anthropomorphically or theriomorphically, or, and this is a distinction between them and the gods, as compounded of both forms. They are enumerated in groups, but the distinctions between the groups are not carefully drawn, and in addition we have many names of individuals, derived from their activity or appearance, or fancy names, or names perhaps of real enemies.

v1.c23.s2.5

The Rakṣases are the most famous of the classes of demons, and occur in the Rigveda upwards of fifty times, nearly always in connexion with the mention of a god who is desired to deal with them. In two hymns¹ the Rakṣases seem to be more precisely defined as Yātus or Yātudhānas, words which denote wizard or sorcerers, and the latter may be regarded as a subdivision of the Rakṣasas, a fact of the utmost importance as showing how important even in the Rigveda was the belief in such beings, many of them probably real men. The Rakṣases are often, as is natural, theriomorphic; they appear as dogs, vultures, owls, and other birds flying about at night;² they can assume the form of husband, brother, or lover,³ and they are a constant peril to the woman in child. In dog or ape form they are ready to attack her;⁴ at the wedding service they prowl around, and small staves are flung in the air to pierce the eyes of the demons.⁵ When they have human forms, they are often hideously deformed, with two mouths, three heads, four eyes, five feet, with feet turned backwards, without fingers, or with horns on their hands, and bear-necked.⁶ Blue, yellow, and green are their colours, and they are not without social organization, having both families and kings:⁷ unlike the gods, they are mortal, and have not won immortality.⁸ They are essentially blood-suckers who seek to enter men, especially by the mouth in the process of eating and drinking, but also by other inferior passages which the gods are therefore besought to protect.⁹ When within they eat the man's or beast's flesh and cause disease;¹⁰ they bring about madness and destroy the power of speech;¹¹ they invade human dwellings¹² and dance around houses in the evening; they make a noise in the forest, pray aloud, or laugh and drink out of skulls as cups,¹³ a point which shows that the ghouls of the places of burial must have contributed an element to the conceptions: in such spirits the idea of the spirits of the dead, especially the dead not duly buried or burnt, and the spirits, which are, like Vāstoṣpati in the house, resident in the place of the dead, must inevitably inextricably combine.

v1.c23.s2.6

¹ RV. vii. 104; x. 87. In JB. iii. 266 Ṛkṣa appears in the sense of Rakṣas.

v1.c23.s2.7

² RV. vii. 104. 18, 22.

v1.c23.s2.8

³ RV. x. 162. 5.

v1.c23.s2.9

⁴ AV. iv. 37. 11.

v1.c23.s2.10

⁵ MGS. i. 4. 10.

v1.c23.s2.11

⁶ AV. viii. 6; HGS. ii. 3. 7.

v1.c23.s2.12

⁷ AV. v. 22. 12; HGS. ii. 3. 7.

v1.c23.s2.13

⁸ AV. vi. 32. 2.

v1.c23.s2.14

⁹ AV. v. 29. 6–8; viii. 6. 3.

v1.c23.s2.15

¹⁰ RV. x. 87. 16, 17; viii. 60. 20; AV. vii. 76. 4.

v1.c23.s2.16

¹¹ AV. vi. 11. 1, 3; HGS. i. 15. 5.

v1.c23.s2.17

¹² Kauç. cxxxv. 9.

v1.c23.s2.18

¹³ AV. viii. 6. 10, 11, 14; HGS. ii. 3. 7.

v1.c23.s2.19

Like most evil spirits, the Rakṣases love the night,¹⁴ especially the night when there is no moon, for in the east the sun disperses them, and in this connexion with night they are akin to the souls of the dead. Hence, it seems, a falling meteor¹ is considered to be the embodiment of a Rakṣas.

v1.c23.s2.20

¹⁴ RV. vii. 104. 18.

v1.c23.s2.21

¹ Kauç. cxxvi. 9.

v1.c23.s2.22

As is natural in a priestly collection, we hear much of the attacks of the Rakṣases on the sacrifice: with the Yātus they taint it and throw it into confusion.² On the other hand, the sorcerer can use a spell, and by the Rakṣases and the Yātudhānas ruin the sacrifice of his enemy.³ In order to obtain the advantage of sacrifices, the evil spirits often assume the form of souls of ancestors and come to the offerings for the dead.⁴ But the demons can be yoked by a skilled man to his own ends; in the Atharvaveda we find that demons are invited to attack the man who sends them, and the later texts recognize the need of providing against the attacks of sorcerers.⁵ In these cases and for the protection of the offering the essential aider is Agni, the burner, who therefore represents the primitive and universal belief in the power of fire to repel hostile demons. An isolated mention is made of Kubera Vaiçravaṇa as their king.

v1.c23.s2.23

² RV. vii. 104. 18, 21; x. 182. 3.

v1.c23.s2.24

³ AV. vii. 70. 2.

v1.c23.s2.25

⁴ AV. xviii. 2. 28; Caland, *Altind. Ahnenkult*, pp. 3 ff.

v1.c23.s2.26

⁵ RV. vi. 62. 9; vii. 104. 23; viii. 82. 20; AV. ii. 24; TS. ii. 2. 3. 2.

v1.c23.s2.27

The meaning of the term is doubtful: the root ‘protect’ is the obvious one, in which case the word must apparently mean that which is to be guarded against, unless we suppose that they are given a good-sounding name, in order to make them good. In the sense ‘injure’, there is little authority for *rakṣ*. Bergaigne⁶ thinks that they are named as guardians of celestial treasure, who being greedy are hated, for which view the parallel of the Gandharvas and the Soma may be cited, as well as the fact that Kubera is their king. Bartholomae connects Rakṣas with the root seen in Greek ἐρεχθεῖν.

v1.c23.s2.28

⁶ *Rel. Véd.* ii. 218.

v1.c23.s2.29

The Piçācas are only once mentioned in the Rigveda,⁷ where Indra is invoked to crush the yellow-peaked, watery, Piçācī and every Rakṣas. In the later Saṁhitās they figure, however, in the plural, and they are opposed to the Fathers, as the Asuras to the gods, and the Rakṣases to men, but not consistently.⁸ They may, however, have been supposed to be specially of the ethnic type of theriomorphic ghouls of the dead, as they have the name *kravyād*, ‘eaters of raw flesh’.⁹ But they also are treated as eating away the flesh of a sick man, while again they appear as infesting human dwellings and villages, and even as will-o’-the-wisps.¹⁰ The view that these Piçācas are really, or originally, a special tribe who were addicted to cannibalism, and who were the speakers of the Prākrit known to the grammarians as Paiçācī, has been put forward of late, but has no probability, the reverse process being much more likely.¹¹ Similarly, the idea that the Rakṣases are originally conceived as hostile aborigines cannot be accepted as explaining the class generally, but merely as one element in the conception.¹ Charpentier² finds in the Piçācas the souls of the dead conceived as glow-worms. The Brāhmaṇas³ tell a curious tale of a Piçācī who married the Ikṣvāku King, Tryaruṇa, and dulled his fire, until the priest Vṛṣa by a rite had her burned up.

v1.c23.s2.30

⁷ i. 133. 5.

v1.c23.s2.31

⁸ TS. ii. 4. 1. 1.

v1.c23.s2.32

⁹ AV. v. 29. 9.

v1.c23.s2.33

¹⁰ AV. iv. 36. 8; 20. 9; 37. 10.

v1.c23.s2.34

¹¹ Grierson, ZDMG. lxvi. 68.

v1.c23.s2.35

¹ Macdonell and Keith, *Vedic Index*, ii. 516.

v1.c23.s2.36

² *Kleine Beiträge zur indoiranischen Mythologie* (1911). But see Winternitz, *VOJ.* xxvii. 229–32.

v1.c23.s2.37

³ PB. xiii. 3. 11–13; JB. iii. 94–6; Sieg, *Die Sagenstoffe des RV.*, p. 64; Oertel, *JAOS.* xviii. 21 ff.; Oldenberg on RV. v. 2. 1.

v1.c23.s2.38

The Arātis, who occur about a dozen times in the Rigveda and frequently later, are clearly abstract deities of illiberality, but clothed with quite a real life in the imagination of the indignant Vedic poet, who asks the gods to overthrow them. The Druhs are found about as often in the Rigveda,⁴ but they are Indo-Iranian, the Avestan being *druj*, and, like all the older conceptions, are not living features of the religion, being merely vaguely conceived as injurious spirits. The Kimīdins as a pair of demons occur already in the Rigveda,⁵ contrary to the more normal practice of grouping the demons in sets, perhaps owing to the influence of the dual deities. The names which we have of such Kimīdins, Mroka and Anumroka, Sarpa and Anusarpa, show, however, their entire distinction from the dual deities which have distinct personalities and quite different appellations.⁶

v1.c23.s2.39

⁴ Macdonell, *Ved. Myth.*, p. 8.

v1.c23.s2.40

⁵ RV. vii. 104. 23; x. 87. 24; Weber, *Ind. Stud.* xiii. 183 ff.

v1.c23.s2.41

⁶ AV. ii. 24.

v1.c23.s2.42

Among the homes of these spirits, especially probably those of the dead, one is especially noteworthy, the cross way, which is the scene of various magic rites. It is there that evil spirits and disease are banished by the sorcerer; there is performed the spell to find what has been lost, there Rudra is said to dwell, after a death the fire which becomes thus impure is deposited there.⁷ It is doubtful what motive in each case can be seen for the superstition which is world-wide: unquestionably the spirits of the dead are thought to live there, especially perhaps evil spirits, as evil dead are often so buried, but it is possible to find other motives: the cross-roads is the place whence all ways deviate, so that it is the proper centre for a spell to seek for what is lost: or again by depositing there what is impure, the idea may be that the impurity is induced to go one way, the former owner of it another, or the evil which is laid aside may enter into and be taken away by one of the many wayfarers who pass that way.

v1.c23.s2.43

⁷ Kauç. xxvi. 30; xxvii. 7; xxx. 18; lii. 14; ÇB. ii. 6. 2. 7; HGS. i. 16. 8; AGS. iv. 6. 3; Oldenberg, *Rel. des Veda*², pp. 269 f.

v1.c23.s2.44

One set of demons seems still to show its essential connexion with a natural object, those which are conceived as embodied in plants. Agni and Indra are implored to destroy the demons together with the root,⁸ and the point, or three points, of demons is alluded to: it is natural to assume that the root is the incorporation of the deity, and this view is strengthened by the reference⁹ to those ‘whose god is the root’, an expression which surely denotes that the root was the actual demon, rather than merely that the demon was supposed to have taken up his abode¹ in the root, though the former conception might easily be passing into the latter.

v1.c23.s2.45

⁸ RV. iii. 30. 17; x. 87. 19.

v1.c23.s2.46

⁹ RV. vii. 104. 24; x. 87. 2, 14.

v1.c23.s2.47

¹ Oldenberg, *op. cit.*, p. 268, n. 4; the natural origin of such spirits is seen in Celtic religion; MacCulloch, *Rel. of Anc. Celts*, pp. 173, 185; cf. Farnell, *Greece and Babylon*, pp. 110, 111, 43.

v1.c23.s2.48

In the case of the disease demons it is not easy precisely to determine what conception was formed of them. The idea formed of them certainly hovers between the conception of a spirit like Takman, 'fever', who brings the disease, and a more materialistic, or scientific, conception of some substance which carries the disease itself. The constant efforts to drive out diseases by means of spells and sympathetic magic may, according to Oldenberg, fairly be held to contain a more advanced conception of the nature of disease than the more simple concept of a demon disease. But this is not at all clear, unless we accept Oldenberg's animistic views. The real question is whether the demon disease was conceived in the first place abstractly, the disease being caused by a demon, and the demon identified with the disease it caused, or whether the disease was conceived as something real and material, with a life of its own. The latter view seems to accord most with primitive thought, and with the material, if less personal, character of the disease as attacked by the means of the sorcerer in the Atharvaveda, in which, however, constant hints of the personal character are still to be found. The actual signs of the disease, the symptoms visible to the eye, are in the primitive mind the disease itself, which is, as the symptoms come and go, clearly a living entity.

v1.c23.s2.49

The demons are not, however, merely repelled by the use of fire, and by many spells; they are actually occasionally propitiated with offerings, as in the case of the Rakṣases and Yātudhānas who are thus treated in the Taittirīya Brāhmaṇa,² a fact which may stand in relation with the use which sorcerers were able to make of these spirits. Tricks are also tried to deceive them: thus in the period³ of continence immediately after marriage, there has been seen by Oldenberg⁴ an ingenious device to prevent any hostile action on the part of the demons, whether by attacking the bride or seeking to enjoy her,⁵ and in any case the sympathetic magic of the device is obvious. The practice in the acts of driving away fiends of enumerating as many names as possible is parallel to that of enumerating as many gods in the offerings of the Vaiçvadeva rite in the domestic ritual, and it has preserved to us several lists. To protect the woman in child-birth a fire is lighted,⁶ and the spirits are banned under the names of Çaṇḍa and Marka, Upavīra, Çauṇḍikeya, Ulūkhala, Malimluca, Droṇāsa, Cyavana, Ālikhant, Animiṣa, Kiṁvadanta, Upaçruti, Haryakṣa, Kumbhin, Çatru, Pātrapāṇi, Nṛmaṇi, Hantrīmukha, Sarṣapāruṇa, Cyavana.

v1.c23.s2.50

² iii. 4. 1. 5.

v1.c23.s2.51

³ Winternitz, *Altind. Hochzeitsrituell*, p. 87; J. J. Meyer, *Das Weib im altindischen Epos*, pp. 235 f.

v1.c23.s2.52

⁴ *Rel. des Veda*², p. 273.

v1.c23.s2.53

⁵ AV. iv. 37. 11; ÇB. iii. 2. 1. 10.

v1.c23.s2.54

⁶ PGS. i. 16. 23; ApGS. xviii. 1; MP. ii. 16. 1 ff. A longer list in HGS. ii. 3. 7; Pramṛçant, Kūṭadanta, Vikleça, Lambastana, Uraspeça, &c.

v1.c23.s2.55

Still more interesting is the treatment of a child which has a barking cough;¹ the demon within him is exorcized to let loose the child, being addressed as Kūrkura, good Kūrkura, and doggy; he is assured that Saramā is his mother, Sīsara his father, and the brindled dogs of Yama his brothers: the idea of the demon cough in dog form is perfectly clear: its name is also given as Kumāra. When the man performs the Vaiçvadeva offerings of the domestic ritual, his wife outside the house offers food² to the man, to the woman, to every age, to the white one with black teeth, the lord of bad women, to those who allure her children, whether in the wood or the forest. On the day of the final bath the pupil banishes from him all the evil forms of Agni, ten in number, which he enumerates.³ Demons figure among those to whom he is given in charge by the teacher on his reception.⁴ In the ritual we find in the Taittirīya Āraṇyaka⁵ a list of hideous forms to be driven away from the offering, Viçīrṣṇī, Gṛdhraçīrṣṇī, and so on. The sacrifice is full of all sorts of magic devices to repel the evil spirits, who, on the one hand, are allowed to have offerings of the blood, though only with inaudible words of offering lest the speech of the speaker should become the voice of a Rakṣas,⁶ but on the other hand are kept off by the use of the enclosing sticks, by the free employment of fire, by the drawing of magic circles, and by the sound of the pressing stones.⁷

v1.c23.s2.56

¹ PGS. i. 16. 24.

v1.c23.s2.57

² PGS. i. 12. 4.

v1.c23.s2.58

³ PGS. ii. 6. 10.

v1.c23.s2.59

⁴ HGS. i. 6. 5.

v1.c23.s2.60

⁵ i. 28.

v1.c23.s2.61

⁶ AB. ii. 7, 1.

v1.c23.s2.62

⁷ TS. ii. 6. 6, 2; RV. x. 36. 4; ÇB. vii. 7. 1. 2; ApÇS. xi. 11. 6. Araru is a demon banned from the place of sacrifice, TS. i. 1. 9; MS. i. 1. 10; iv. 1. 10; VS. i. 26; TB. iii. 2. 9. 4; possibly in RV. x. 9. 9, 10.

v1.c23.s2.63

The great gods of the Vedic period are not reduced to demons, nor do they approximate to them: it is true that their names are invoked in very trivial occasions such as that of Indra in an idiotic spell to induce a slave not to run away from his master,⁸ or to secure connexion,⁹ but that merely denotes that the gods were familiar enough to their votaries not to despise aiding them in all their actions. Hillebrandt¹⁰ indeed sees in some passages the tendency to reduce Rudra to the rank of little more than a demon, but, while Rudra is a terrible god and is in some ways akin to the demons, the evidence adduced by Hillebrandt is not altogether in point. Rudra is to be addressed when a man is on a path, at cross-roads, crossing a river, at a mountain, a forest, a burial-ground and a stable, but the fact that he is so to be invoked is expressly explained by his omnipresence,¹¹ and we have here not so much a primitive idea as the extension to every sphere of activity of the great god Rudra, the sign of an advanced theistic view of the god rather than a degradation of his nature. More kinship to legends of demons is to be seen in such a legend as that of Indra becoming a horse's tail;¹² the gods are able to change form at will, a fact of which Indra makes free use in his amourettes, but it is rarely that they adopt anything but a noble form such as the eagle form of Indra, but that god has a tendency to assume popular traits. Kubera, later the god of wealth, is a Rakṣas and lord of robbers and evil-doers in the Çatapatha Brāhmaṇa;¹ in the Sūtras² he is invoked with Īçāna for the husband in the marriage ritual, and his hosts plague children. Comparison³ with Greek Kabeiros and explanation as a mountain spirit of hiding propensities are most doubtful.

v1.c23.s2.64

⁸ PGS. iii. 7. 3.

v1.c23.s2.65

⁹ PGS. i. 9. 5.

v1.c23.s2.66

¹⁰ *Ved. Myth.* iii. 425.

v1.c23.s2.67

¹¹ PGS. iii. 15. 7 ff.; HGS. i. 16. 8 ff.; AGS. iv. 8. 40; MP. i. 13. 8; ApGS. ix. 3; Oldenberg, *Rel. des Veda*², pp. 220 ff. Arbman's objections (*Rudra*, pp. 222 ff.) are irrelevant.

v1.c23.s2.68

¹² Geldner, *Ved. Stud.* ii. 183; perhaps an ant, RV. i. 51. 9; Hillebrandt, *op. cit.* iii. 172, 173.

v1.c23.s2.69

¹ xiii. 4. 3, 10; cf. AV. viii. 10. 28.

v1.c23.s2.70

² ÇGS. i. 11. 7; HGS. ii. 1. 3. 7.

v1.c23.s2.71

³ Hopkins, *JAOS.* xxxi. 55-70.

v1.c23.s2.72

A certain interest attaches to the very late Nejameṣa, who appears apparently as banned in the ritual of the parting of the hair of the bride,⁴ for he reappears in the form of Naigameṣa in the medical work of Suçruta, and as Nemesa in a Mathurā inscription.⁵ It is dubious whether the name is found in the epic, though possibly the Naigameya there recorded may be an error for it.⁶

v1.c23.s2.73

⁴ ÇGS. i. 22. 7; AGS. i. 14. 3.

v1.c23.s2.74

⁵ Bühler, *EI.* ii. 316; Winternitz, *JRAS.* 1895, pp. 149 ff.; Hillebrandt, *Ved. Myth.* iii. 424.

v1.c23.s2.75

⁶ Hopkins, *Epic Myth.*, pp. 103, 227, 229.

v1.c23.s2.76

In the somewhat late Mānava Gṛhya Sūtra⁷ we have a rite prescribed for one who suffers from possession by the Vināyakas, Çālakaṭaṅkaṭa, Kūṣmāṇḍarājaputra, Usmita and Devayajana, in which a strange variety of deities are invoked, Vimukha, Çyena, Baka, Yakṣa, Kalaha, Bhīru, Vināyaka, Kūṣmāṇḍarājaputra, Yajnāvikṣepin, Kulaṅgāpamārin, Yūpakeçin, Sūparakroḍin, Haimavata, Jambhaka, Virūpākṣa, Lohitākṣa, Vaiçravaṇa, Mahāsena, Mahādeva, Mahārāja. Mahādeva is doubtless Rudra, Mahāsena appears elsewhere⁸ as a disease demon and is an epithet of Skanda, Vaiçravaṇa is Kubera. In Yājñavalkya⁹ we find a single Vināyaka, who is here son of the goddess Ambikā, and appointed by Brahman and Rudra to be the overlord of the Gaṇas, troupes, perhaps akin to those assigned to Rudra’s entourage in the Yajurveda.¹⁰ Doubtless we have here something akin to the later Gaṇeça, who is not Vedic, but it is by no means certain that we are to interpret Vināyaka as denoting ‘leader’, instead of ‘remover of obstacles’,¹¹ or the epithet Vighneça, which later is found applied to Gaṇeça, as lord of Vighnas, conceived as destroyers. The term Yakṣa,¹² which occurs in this list, is far better known in the Buddhist form of Yakkha; we find the term applied to a wondrous thing in the Jaiminīya Brāhmaṇa;¹³ the Ṛṣis seek to behold something of this kind, and Indra reveals to them the tortoise Akūpāra, of boundless dimension, who is clearly the cosmic tortoise who finds the earth in the ocean. The specification of the term to mean a species of spirit, usually associated with Kubera, is not found until the period of the Gṛhya Sūtras.

v1.c23.s2.77

⁷ ii. 14; Arbman, *Rudra*, pp. 57 f. 219 ff.

v1.c23.s2.78

⁸ PGS. i. 16. 24 is cited by Arbman, but wrongly. Skanda appears in BhGS. iii. 9; AV. Par. xx; BG. Par. iv. 2.

v1.c23.s2.79

⁹ i. 271 ff.

v1.c23.s2.80

¹⁰ TS. iv. 5. 4.

v1.c23.s2.81

¹¹ Jacobi, *ERE.* ii. 807.

v1.c23.s2.82

¹² Boyer, *JA.* 1906, i. 393 ff.; Oldenberg, *RV.* ii. ff.; Hopkins, *Epic Myth.*, pp. 30 f.

v1.c23.s2.83

¹³ iii. 203, 272. *contra*, Winternitz, *JRAS.* 1898, p. 383.

Linked texts (21)