Introduction
As outlined in the Introduction, the verse ṚV III 62.10 is known under two names: sāvitrī (Sāvitrī, Savitri, Savitree, etc.) and Gāyatrī (other spellings include gāyatrī, Gayatri, Gaayatree, or even Gayutree). As will be seen, sāvitrī is the older of the two names. The history of this name is closely connected to the early history and rise of the mantra itself, which will be discussed in detail in the remainder of Part I.²³⁵ In the following, I will only briefly summarize some of the basic facts about the designation sāvitrī. For reasons that will soon become clear, the history of the designation “Gāyatrī” is a much more contested topic, and reconstructing it will indeed be the main task of this chapter.
sāvitrī́ is the nominalization of sāvitrá/ ī́, a relational adjective derived from the word savitṛ́, the name of a Vedic god. As such, this adjective simply means
“belonging/relating/related to Savitṛ,” as in the case of the sāvitrá gráha, a ritual
“cup” containing a portion of Soma “for Savitṛ.” Apart from ritual parapherna-lia, sāvitrī́ and sāvitrá also became names for mythical and literary characters.
Most importantly, the name sāvitrī́ was given to the goddess Sūryā, when the idea that she is Savitṛ’s daughter took hold.²³⁶ Since Savitṛ became more or less identified with the sun already during the time of the Brāhmaṇas, sāvitra could also become an epithet of the epic character Karṇa, whose father is the sun god.
Typically, and especially in its feminine form ending in ī, sāvitra/ ī became the designation for any ṛ́c (f.) or “verse” addressing or mentioning the deity Savitṛ (a “Savitṛ verse” or “Impeller verse”) even without the addition of the word ṛ́c. This basic meaning of the word was never really lost. Unless the context indicated it, it was therefore sometimes necessary to specify which of the many sāvitrī verses was meant. Especially in ritual texts, this is usually done by way 235 See especially below pp. 119–120.
236 See below pp. 204–212.
62 ∙ paRt i ∙ the mantRa
of citing the beginning of the verse (a part known as the pratīka),²³⁷ in the case of ṚV III 62.10, “‘tat savitur’ iti” or “‘tat savitur vareṇyam’ iti. ”²³⁸
Already towards the end of the Vedic period, however, ṚV III 62.10 virtually became the most frequently recited and most important of all sāvitrī s.²³⁹ As a consequence, it was in most cases no longer necessary to specify the verse, but rather to indicate if any other sāvitrī than ṚV III 62.10 was meant. As we will see in Chapter 7, this also made it possibly to associate it with the deity Sāvitrī
or Sūryā Sāvitrī, Savitṛ’s daughter, who originally had nothing to do with the mantra of the same name.
Let us now turn to the second name of ṚV III 62.10, its “cognomen,” as it might be called. The word gāyatrī́ literally means “belonging to the song,” or
“song-related,” but in Vedic literature is more often used as the name of a meter.²⁴⁰ As we know, at some point the word also became the proper name of a specific verse set in the gāyatrī meter. The questions are then how, why, and when was this use of the word introduced. My main aim in this chapter is to answer these questions. In order to do so, I analyze the source texts chronologically, beginning with the earliest and working my way forward.
• **Sections 1.1 **and **1.2 **(pp. 63–72) review the Vedic passages that have been thought to use gāyatrī as a designation of the verse. In fact, only few Vedic texts were read in this way, above all the Atharva-Veda (AV) and the Bṛhad-Āraṇyaka-Upaniṣad (BṛhĀU). In these texts the word gāyatrī is always used to denote the meter, and never the GM itself.
• **Section 2.1 **(pp. 72–78) analyzes the usage of the words sāvitrī and gāyatrī in the MBh. I demonstrate that the early portions of the Great Epic generally only use the word sāvitrī when speaking about the mantra.
237 In the Vedic context, mantras are generally cited by means of giving their pratīka, that is by quoting their first few words. Sometimes, a suffix īya is added, as in āpohiṣṭhīya = ṚV
X 9.1, which begins with the words ā́po hí ṣṭhā́ mayobhúva-; cf. Quillet 2011: 359–360. The rest of the mantra is then inferred by the reciter, who is expected to know the relevant mantras by heart.
238 Or possibly even tad; see n. 377 on p. 91 below.
239 See Chapter 4, especially pp. 130–133.
240 The derivation of the word gāyatrī́ can be explained in two ways: (a) it is the feminine form of *gāyatrá, a relational adjective with the meaning “belonging to the song” that is derived from the noun gāyatrá (m./n.) “song” (cf. AG II,2: 402 [§250ζ]). In its feminine form, this adjective came to be used as the proper name of a specific meter, the gāyatrī́
meter. Alternatively, (b) it is a direct feminine derivative of gāyatrá “song” ( AG II,2: 383
[§247]), motivated by the fact that meters are generally feminine. Either way, from the proper name gāyatrī́ a relational adjective was derived by strengthening ( vṛddhi) the first vowel. Since this vowel was strong already, the resulting adjective was homophonous with the original adjectival from which gāyatrī́ had been derived: gāyatrá. This adjective qualifies something that is composed in the gāyatrī meter. It can also be easily used as a noun ( gāyatrī) in the meaning of a “gāyatrī verse.” As I argue below, however, such an understanding of the word can lead to far-reaching problems; see pp. 71–72.
chapteR 2 ∙ designations ∙ 63
Only later additions, such as the Vaiṣṇava-Dharmaśāstra (VaiṣṇDhŚ), also use the word gāyatrī for the mantra.
• **Section 2.2 **(pp. 78–81) argues that the new designation gāyatrī became popular as a consequence of the introduction of modified GMs. The creation of these mantras led to the revival of an obsolete category, that of gāyatrī verses, among which the GM immediately emerged as the typical representative.
**1. Meter or mantra? **
1.1 Atharvaveda-Saṃhitā
As a simple word search quickly reveals, the text of the mantra cannot be found in any form in either of the two recensions of the AV. The pratīka of the mantra is not mentioned either. While the word sāvitrī́ does appear two times, it is used exclusively for the goddess Sūryā, to whom we shall return in Chapter 7. The word gāyatrī́, the other possible designation of the mantra, is mentioned more often. This means that if the mantra is present in the Saṃhitā, it is referred to either implicitly or by means of the name of its meter.
Several scholars²⁴¹ were inclined to see references to the GM in four AV
passages: AV XIII 1.10; X 8.10 and 41; and IX 10.19. Further, there is one late passage – XIX 71 – which conspicuously calls upon a vedamātṛ́, a “Mother of the Vedas,” one of the most common epithets of the mantra deity in later times.
As the interpretation of X 8.10, 41, and IX 10.19 more or less depends on AV
XIII 1.10, I will discuss this verse first. The discussion of AV XIX 71, however, is found in Chapter 6, where I will show that this verse is considerably younger than most of the other parts of the AV.²⁴²
The first verse to be discussed is located in kāṇḍa XIII of the AV. This kāṇḍa was probably composed during the period of the later Saṃhitās (such as the Taittirīya-Saṃhitā [TaittS]),²⁴³ and is dedicated to róhita, the “reddish,”
rising sun (or “raised” sun – this would be the literal translation of the almost homophonous rohitá). Throughout this book, the red morning sun symbolizes 241 Louis Renou (1956), Franklin Edgerton (1962), Subhash Anand (1988), and Thomas Oberlies (1998, 2012).
242 See below pp. 193–197.
243 Witzel 1997b: 281.
64 ∙ paRt i ∙ the mantRa
royal power.²⁴⁴ róhita also appears in our first verse, AV XIII 1.10, where it qualifies a “calf”:
yā́s te víśas tápasaḥ saṃbabhūvúr, vatsáṃ gāyatrī́m ánu tā́ ihā́guḥ /
tā́s tvā viśantu mánasā śivéna, sáṃmātā vatsó abhíyètu róhitaḥ //
Your folk (subjects), which have assembled out of fervor, have come here, following the Calf, the gāyatrī́.
Let them enter into you with well-disposed thought; let the Ruddy Calf come hither with his mother.²⁴⁵
The overall context makes clear that the person spoken to in this verse is the king. His subjects have united under him, following a “calf” or “young one”
( vatsá), which, pāda d suggests, is Rohita. While the king is sometimes even identified with Rohita in the remainder of the kāṇḍa, he and Rohita / the Ruddy Calf are clearly discrete in this case. Rohita is probably characterized as “young”
or a “calf” because he is the new-born, rising sun.
Besides the king, his subjects, and Rohita, mention is also made of Rohita’s mother. The juxtaposition of vatsá and gāyatrī́ in pāda b is probably best read as an asyndeton, that is, the connector “and” can be supplied: “the calf and the gāyatrī́.” The word sáṃmātṛ should probably be understood as “together with the mother” – as opposed to the more common and differently accented saṃmātṛ́, which means “having the same mother”²⁴⁶ or, in the dual, “two mothers together.”²⁴⁷ If the verse is construed in this way, it stands to reason that Rohita’s mother is called gāyatrī́.
Following a suggestion of Louis Renou, Franklin Edgerton argued that the gāyatrī́ is the GM, recited at dawn in the Sandhyā, which is regularly performed before sunrise (and after sunset).²⁴⁸ As the gāyatrī́ (mantra) thus precedes the rise of the sun god in the daily cycle, she could be conceived of as his mother:
“This sacred stanza was the necessary beginning of the morning service, the start of the ritual day. It was recited at dawn, just at or just before the sunrise.
What more natural than to call it (as elsewhere the dawn is called) the sun’s mother?”²⁴⁹
244 For further literature, see Witzel 1997b: 267, 279–280.
245 AV XIII 1.10; tr. Edgerton 1962: 56. Cf. the translations by Whitney 1905: 711; Renou 1956: 205; and Oberlies 2012: 481, n. 419.
246 ṚV X 117.9 (tr. Jamison & Brereton 2014: 1587).
247 AV XIII 2.13 and VIII 7.27 (tr. Whitney 1905: 721 and 501).
248 Renou 1956: 205, n. 16 (see p. 271); followed by Edgerton 1962; in turn followed by Oberlies 1998: 313, n. 802, and Oberlies 2012: 301, n. 419 (see p. 481).
249 Edgerton 1962: 58. Following the same logic, the night is said to be Rohita’s mother in AV
XIII 3.26; cf. Edgerton 1962: 57, n. 3. A similar case would be that of Savitṛ, who originally was a more or less nocturnal deity and came to be called the father of Sūryā, a morning goddess; see below p. 204.
chapteR 2 ∙ designations ∙ 65
In many AV passages, however, the gāyatrī́ is mentioned alongside other meters such as the triṣṭúbh or the jágatī.²⁵⁰ Although this does not preclude the word gāyatrī́ from being used not only for the meter but also for the mantra, I would argue that in the passage under discussion, it only denotes the meter. In fact, the verse can be very well understood in this way.
Since Edgerton’s paper, several scholars have demonstrated that in the system of Vedic religion the gāyatrī was often thought to be the “first” and
“primary” among meters. The gāyatrī is the shortest of all Vedic meters and the most frequently used meter in the ṚV after the triṣṭubh. It was regularly associated with other first or most important things: On the level of social strati-fication, with the Brahmins; on the cosmic and ritual levels, with fire and its deification, Agni; on the textual level, with the ṚV, and so on.²⁵¹ On the level of time, there existed a rather strong association between the gāyatrī́ meter and the morning, the beginning of the day.²⁵² According to Thite, this meter is said to be the “excellence” or greatness of the morning pressing of Soma, and
“[t]herefore, whatever meter may be used at the time of that pressing, it is called, mystically, Gāyatrī.”²⁵³ As opposed to the other pressings, the morning pressing could be “legitimately called gāyatra i.e. belonging to the Gāyatrī-metre.”²⁵⁴ We will come across similar associations throughout this study.
Understood against this background, this association with the morning perfectly explains the verse: the gāyatrī́ meter precedes the rising of the sun, as it were, and can therefore be conceived of as its mother. A parallel passage in the Taittirīya-Brāhmaṇa (TaittB) supports this reading wherein we read gāyatráṃ vatsám instead of vatsáṃ gāyatrī́m. Adapting Edgerton’s rendition, I translate its first half: “your folk (subjects), which have assembled out of fervor, have come here, following the Calf belonging to the gāyatrī́ ( gāyatrá).”²⁵⁵ Unlike gāyatrī́, gāyatrá is never used as a name of the GM – at least nobody has ever argued that it does. Rather, gāyatrá in this case is probably an adjective qualifying the calf, “the calf belonging to the song-meter.”²⁵⁶ In this passage, it is impossible to detect the presence of the GM, indicating that it might not be present in AV XIII 1.10 either.
250 E.g., in AV VIII 9.14,20: gāyatrī́ṃ triṣṭúbhaṃ jágatīm anuṣṭúbhaṃ-, gāyatrī́ … triṣṭúp- …
jágatī … anuṣṭúp- , XVIII 2.6: triṣṭúb gāyatrī́ chándāṃsi, XIX 21.1: gāyatry ùṣṇíg anuṣṭúb bṛhatī́ paṅktís triṣṭúb jágatyai.
251 For further details and references, see below p. 120.
252 Cf. Thite 1987: 438–441, Smith 1993: 80, and Fujii 2010: 3–4.
253 Thite 1987: 438.
254 Thite 1987: 439.
255 TaittB II 5.2.2: yā́s te víśas tápasā saṃbabhūvúḥ / gāyatráṃ vatsám ánu tā́s ta ā́guḥ / tā́s tvā́viśantu máhasā svéna / sáṃmātā putró abhyètu róhitaḥ //.
256 Cf. n. 240 on p. 62 above.
66 ∙ paRt i ∙ the mantRa
In his paper, Edgerton also refers to three other AV verses that possibly refer to the GM.²⁵⁷ Two of these verses are part of the eighth hymn in kāṇḍa X, a long and loose collection of “mystic” verses.²⁵⁸ In the first passage, AV X 8.41ab, Mātariśvan, a mythical, winged being that brought the fire to the earth, is said to have “strode out almost higher than the gāyatrī́, upon the immortal.”²⁵⁹ This refers to the story of the flight of the gāyatrī́ meter, which flew to heaven in order to procure Soma.²⁶⁰ Against the backdrop of this story, it becomes clear that Mātariśvan does not “surpass” the GM, but really reaches a place almost a bit higher than the gāyatrī́ meter.
The other two relevant verses mentioned by Edgerton both revolve around a similar topic: **AV X 8.10 **poses a riddle, asking, in brief, which one of the ṛ́c s or “verses of praise” can be universally applied.²⁶¹ Again following a suggestion made by Renou, Edgerton suggests that this verse could be the GM. The second verse, AV IX 10.19, speculates about the pāda s constituting a verse. Three “feet”
( pad) are mentioned, so the verse most likely is a verse in the gāyatrī́ meter. The verse asserts that when the pāda s are put together to form a verse, the entire living world is created. As only one verse is mentioned, Edgerton concluded that a specific ṛ́c must be meant: the GM.²⁶²
These two passages could, at best, only offer additional confirmation for a result that would have to be reached by other means. The question posed in the first passage, X 8.10, might even be a rhetorical one. There are, as far as I can see, no indicators allowing us to conclude that the verse in question is the GM: it is, for instance, not clear how the GM could be said to “extend the ritual forward”
( yáyā yajñáḥ prā́ṅ tāyáte). The second passage, IX 10.19, probably implies a verse in the gāyatrī́ meter, but there is little reason to assume that any specific verse is meant: the passage simply adds the word ṛcás “of a verse of praise” to the word pad in order to clarify that by pad, “foot,” the pāda of a verse is meant.
257 Again following suggestions made in Renou 1956: 167, n. 11 (see p. 264) in the case of AV
X 8.10 and 41.
258 Tr. Whitney 1905: 595–601.
259 úttareṇeva gayatrī́m amṛ́té ’dhi vícakrame; cf. Whitney 1905: 601.
260 See below p. 191.
261 In Edgerton’s (1962: 57) translation: “(The Rigvedic stanza) which is employed in front and behind, which is employed in all cases and in every case, by which the sacrifice is extended forward (or perhaps, with Renou, in the east), that I ask you: which of the stanzas ( ṛcā́m) is it?” yā́ purástād yujyáte yā́ ca paścā́d yā́ viśváto yujyáte yā́ ca sarvátaḥ /
yáyā yajñáḥ prā́ṅ tāyáte tā́ṃ tvā pṛcchāmi katamā́ ⁺ sárcā́m (← °ā́ + ṛ°) [instead of ° á ṛ° in Roth & Whitney 1856: 234; see Whitney 1905: 597] //; cf. the translations by Renou 1956: 167 and by Whitney 1905: 597.
262 “Fashioning by measurement (one) (verse-) “foot” of the stanza, with a half-stanza (i.e., two verse-“feet”) they fashioned all that stirs; with (all) three (verse-) “feet” the bráhman (here Holy Word, Vedic utterance = ṛ́c) spread out many-formed; by that the four directions live.” ṛcáḥ padáṃ mā́trayā kalpáyanto ’rdharcéna cakḷpur víśvam éjat / tripā́d bráhma pururū́paṃ vítasthe téna jīvanti pradíśaś cátasraḥ //; tr. Edgerton 1962: 57.
chapteR 2 ∙ designations ∙ 67
To summarize: Both AV XIII 1.1, the verse about the reddish calf, and its variant in the TaittB can be explained if gāyatrá/ ī́ is taken to refer to the meter.
It is difficult to show that the other three AV passages, the “Mātariśvan verse” X
8.41, the riddle posed in X 8.10, and the speculation about the pāda s of a verse in IX 10.19, refer to the GM. Edgerton concluded that they did. He did this because he assumed that the GM was, at that time, already as “famous” as it would later become. However, considering the general use of the words sāvitrī́ and gāyatrī́-
in the AV and other Vedic texts as well as the role of the GM in Śrauta ritual,²⁶³
this assumption can no longer be upheld.
1.2 Bṛhad-Āraṇyaka-Upaniṣad
While scholars obviously did not consider it self-evident that the GM was already important in the AV, this has been taken for granted in the case of the BṛhĀU. In this text, most of whose contents were probably composed in the seventh and sixth centuries bce,²⁶⁴ the GM occurs in two passages – or rather, in only one (to be explained momentarily). The first passage, BṛhĀU VI 3.6, will be dealt with in Chapter 4.²⁶⁵ It prescribes the preparation of a mixture of fruit and herbs for a man who wants to attain “greatness” ( mahat); the man has to drink the mixture while reciting a combination of the pāda s of the GM, the ṚV verses I 90.6–8, and the Vyāhṛtis. As I will argue, the use of ṚV V 82.1
(another gāyatrī sāvitrī) in similar rituals indicates that the category to which these verses belong was more important than the textual content of the verses.
It is difficult to elicit much more information about the GM from this passage (which is in fact the only one in the Upaniṣad specifying the use of ṚV III 62.10), and indeed, few scholars have attempted to do so.²⁶⁶
The second passage, BṛhĀU V 14,²⁶⁷ on the other hand, has been interpreted more than once as a full-fledged exposition of the GM. Jan Gonda, for instance, stated “that this famous and important mantra had already at an early moment become the object of esoterical speculation and ‘mystic’ explanation.
In the Bṛhadāranyaka-upaniṣad (5, 14) the sacred Gāyatrī is esoterically explained.”²⁶⁸ Similarly, Mieko Kajihara, who otherwise carefully distinguished between the gāyatrī meter and the sāvitrī verse(s), concluded that “the words 263 For the role of the GM in Śrauta ritual, see Chapter 3.
264 For the date of the Upaniṣads, see generally Olivelle 2018b.
265 See below p. 129.
266 E.g., Gonda 1963a: 285–286 and Lal 1971: 226.
267 BṛhĀU(K) V 14 ≈ BṛhĀUM V 15 = ŚatB(M) XIV 8.15 (see Olivelle 1998: 35); cf. Kajihara 2019: 7, n. 24. Cf. also ChāndU III 12 (tr. Olivelle 1998: 207), which also deals with the gāyatrī meter.
268 Gonda 1963a: 287; cf. also Gonda 1963a: 248, n. 6, and Gonda 1975: 52. This reading has been accepted time and again and has been the basis for a great number of misinterpre-tations; see, for instance, Slaje 2009: 523, n. 8.
68 ∙ paRt i ∙ the mantRa
Sāvitrī and Gāyatrī came to function to represent one and the same verse,”²⁶⁹
specifying that this “tendency is seen already in the mid-Vedic texts.”²⁷⁰ As I will show in the following, this assessment needs to be adjusted: as in the AV, the word gāyatrī here only denotes the meter.
In the first half of BṛhĀU V 14, there is a lengthy speculation concerning the pāda s of the gāyatrī (no word indicating a “verse” is mentioned). The three pāda s of the gāyatrī are associated with different sets of words, each having eight syllables in total.²⁷¹ In addition, there is a fourth (imaginary) pāda that is beyond the sky. In a typically Upaniṣadic fashion, the text then elucidates what this fourth pāda is “based” upon, then on what this basis is based, and so on. Thus, the fourth pāda is based on “truth” ( satya), which in turn is based on
“strength” ( bala), which in turn is based on “breath” or the “vital force” ( prāṇa).
Tied to the last link of this chain is an explanation of the word gāyatrī, which also mentions two kinds of sāvitrī s.
For the sake of the following discussion, I will not give as the base text the standard version of the BṛhĀU, which is that of the Kāṇva (K) recension, but that of the **Mādhyandina (M) recension **(the variant readings in the K text are given in brackets):
sā́ haiṣā́ gáyāṃs tatre | prāṇā́ vái gáyās tát prāṇā́ṃs tatre | tád yád gáyāṃs tatre tásmād gāyatrī́ nā́ma sá yā́m evā̀mū́m ( evāmūṃ sāvitrīm) anvā́-
haiṣáivá sā́ sá yásmā anvā́ha tásya prāṇā́ṃs trāyate |7|
tā́ṁ háike ( tāṃ haitām eke) | sāvitrī́m anuṣṭúbham ánvāhur vā́g anuṣṭúb etád vā́cam ánubrūma íti | ná táthā kuryāt | gāyatrī́m evā́nubrūyāt ( gāyatrīm eva sāvitrīm anubrūyāt) |
yádi ha vā́ ápi ( apy evaṃvid) bahv ìva pratigṛhṇā́ti ná haivá tád gāyatryā́
ékaṃ caná padáṃ práti |8| sá yá imā́ṃs trī́ṃl lokā́n | pūrṇā́n pratigṛhṇīyā́t sò ’syā etát prathamáṃ padám āpnuyād-It [i.e., the gāyatrī́] protects ( tatre [← trā]) possessions ( gaya). Clearly, one’s possessions are the vital forces, so it protects the vital forces. And because it protects one’s possessions, it is called gāyatrī. That (K adds
“Savitṛ verse”) which he recites, that is exactly the one [that has just been mentioned]. Because he recites it, he/it protects his [i.e., another person’s]
vital forces. |7|
Some recite this Savitṛ verse as an anuṣṭúbh, saying: “the anuṣṭubh is 269 Kajihara 2019: 11.
270 Kajihara 2019: 11, n. 37.
271 First pāda: bhūmi, antarikṣa, d( i) yau; second pāda: ṛcas, yajūṁṣi, sāmāni; third pāda: prāṇa, apāna, v( i) yāna.
chapteR 2 ∙ designations ∙ 69
speech – in this way we teach speech.”²⁷² One should not do so, one should recite only a gāyatrī́ (K adds “Savitṛ verse”).
Even if one (K adds “who knows it in this way”) gains quite a lot, it is nothing against even a single foot of the gāyatrī́. |8| If someone were to gain these three worlds in full, he would [only] obtain its first foot.²⁷³
As mentioned, the occurrence of both sāvitrī and gāyatrī in one and the same passage has been the cause of some confusion. Patrick Olivelle, for instance, variously translated the word gāyatrī in this passage with “Gāyatrī” and
“Gāyatrī verse.”²⁷⁴ Gonda and several other scholars understood it in this way, too. I would argue, however, that there is little reason to assume that the word gāyatrī in this passage refers to anything other than the meter, which, in my view, was of primary interest to the authors and redactors. In contrast to the gāyatrī meter, the sāvitrī verse is only mentioned in a brief digression. As I will argue in the following, there are good reasons to assume that this part is an insertion in an otherwise monothematic text.
The passage beginning with tā́ṁ háike up to evā́nubrūyāt (the middle paragraph in the translation above) has a parallel passage elsewhere in the **Śatapatha-Brāhmaṇa **(ŚatB) proper. The passage is placed within a section dealing with the Upanayana:
tā́ṁ haitā́m éke | sāvitrī́m anuṣṭúbham ánvāhur vā́g vā́ anuṣṭúp tád asmin vā́caṃ dadhma íti ná táthā kuryād yó hainaṃ tátra brūyā́d ā́nvā́ ayám asya vā́cam adita mū́ko bhaviṣyatī́tīśvaró ha táthaivá syāt tásmād etā́ṃ gāyatrī́m evá sāvitrī́m ánubrūyāt |
Some recite this Savitṛ verse as an anuṣṭúbh, saying: “Clearly, the anuṣṭúbh is speech – in this way we put speech into him [i.e., the student].”
One should not do so. If, in that case, anyone was to say of him: “Indeed, he [i.e., the student] has taken his speech – he will become dumb!” the master [i.e., the teacher] might become so. Therefore, one should recite it only as a gāyatrī́ Savitṛ verse.²⁷⁵
The passage in the BṛhĀU, which I will henceforth call the “Upanayana passage,”
reuses some of the formulations of this ŚatB passage. In the following I will show that the Upanayana passage is only loosely connected to its surrounding text.
First, in the M version of the BṛhĀU passage that is translated above, the word sāvitrī́ is mentioned only once; the K text, on the other hand, inserts it 272 Cf. below p. 120 and 123.
273 BṛhĀUM V 15 = ŚatB(M) XIV 8.15.7–9 ≈ BṛhĀU(K) V 14.4–5; cf. the translation by Olivelle 1998: 139 and 141. The text proceeds in a similar fashion with the other feet.
274 Cf. his note on the translation of BṛhĀU V 14.1 (see Olivelle 1998: 525).
275 ŚatB(M) XI 5.4.13; cf. the translations by Kajihara 2019: 8 and Eggeling, SBE XLIV: 89.
70 ∙ paRt i ∙ the mantRa
two times. Without the insertion in the first paragraph, nothing indicates that
“that which he recites” ( sá yā́m evā̀mū́m) should be understood to refer to the sāvitrī́ in the context of the Upanayana. Rather, in the M text the sentence simply clarifies that the gāyatrī́ meter that one recites in practice really is the same as the gāyatrī́ that protects a person’s property and that is ultimately based on the “mystic” fourth foot beyond the sky and the vital force within oneself, as the preceding text elaborates. The next sentence says more or less the same (though using other words): because he recites it, it protects his vital forces. Only then does the Upanayana passage (the middle paragraph) begin in the M text too.
Moreover, the paragraphs surrounding the inserted Upanayana passage can be well read together (above all in the case of the presumably earlier M
reading): The gāyatrī́ meter protects one’s property, but however much one might gain (even if it is “quite a lot”), nothing can be greater than the four feet of the GM. In this context, mention of the potentially even dangerous practice of teaching an anuṣṭubh sāvitrī appears to be little more than a brief digression.
While it is difficult to establish which one of the recensions of the BṛhĀU is older or more original,²⁷⁶ in this case, it is more likely that the variant readings are the result of K additions rather than M omissions. The M text is coherent even without the additions. Upon close examination, it turns out that the insertion of the word sāvitrī in the explanation of the meaning of the word gāyatrī́ in the sentences preceding the Upanayana passage interrupts the flow of argumentation. The addition was motivated only by the fact that the teaching of a sāvitrī́ in the Upanayana is the most famous case in which the recitation of a verse in the gāyatrī́ meter is preferable.
But what does this somewhat microscopic analysis tell us about the GM
itself? As mentioned above, it would be a grave mistake to equate the gāyatrī́
with the GM, and to read the passage as a speculation about the GM. While the role and significance of the gāyatrī meter was certainly important in the selection of the initiation mantra in the context of the Upanayana, that is a separate topic. The meter, after all, was only one of the criteria used for this selection.²⁷⁷
Second, the passage does not corroborate the view that the expression amūṃ sāvitrīm – in which sāvitrīm is probably secondary – suggests “that no identification is needed, because it was established well enough as to which verse is the Sāvitrī.”²⁷⁸ Admittedly, given its frequent employment in Śrauta ritual and its predominance in later times, there is little reason to assume that at the time of the BṛhĀU the gāyatrī sāvitrī used in the Upanayana was not ṚV III 62.10. Nevertheless, it is the category “gāyatrī sāvitrī” that is emphasized, even 276 Caland 1926: 103.
277 See below pp. 126–128.
278 Kajihara 2019: 7, n. 24.
chapteR 2 ∙ designations ∙ 71
in the case of the K recension. The evidence does not warrant the conclusion that any special significance was (yet) attached to a specific verse in this category. The corollary of this is that the BṛhĀU does not use the word gāyatrī to designate ṚV III 62.10.
As we can see, passages dealing with the gāyatrī (meter) are frequently mistaken to be passages about the GM, no doubt because it is often assumed that the GM had always been a central component of Vedic and Hindu culture. But while the system behind the associations of meters with times, deities, etc., was studied in detail only in the second half of the twentieth century, the peculiar nature of meters in the Vedic culture had already been well known before that time. A question that is not entirely incidental to this study is, therefore: How could this mix-up have happened so easily?
In order to clarify the confusion, I suggest taking a step back to reconsider the relationship of verses and meters. Although already in Ṛgvedic times, relational adjectives such as tráiṣṭubha were introduced for texts set in the respective meters,²⁷⁹ an individual verse was rather said to be a triṣṭúbh than a tráiṣṭubha or tráiṣṭubhī. In Vedic poetry, a verse is always set in a meter, and a meter can only become manifest as a verse. Just like rhythms, which one can clap or drum, for example, one can also make meters audible with the help of a variety of verses. Contrary to what one might expect, there is, to my knowledge, no indication that Vedic meters are ontologically subordinate to verses.
It must not be forgotten that in an oral and aural culture, a meter is primarily a sonic entity, not merely an abstract sequence of long and short syllables,
“revealed” by scansion and symbolic mark-up such as – u u. As sonic entities, meters belonged much more to the domain of experience rather than abstraction, they were heard rather than thought or analyzed, a fact that possibly also facilitated their reification, or even personification and deification, as in the case of the gāyatrī bird.²⁸⁰ In Vedic poetry, it is more the case that a verse is a manifestation of a meter, than that a meter is merely the underlying pattern of a 279 In the case of the meter triṣṭúbh we also start with a word that means “song”: According to Grassman ( WRV: 560) the word originally meant “praising three times [i.e., very much]” (“ursprünglich wol [sic]: dreifach (d.h. sehr) preisend, stúbh”; possibly by way of
* triḥṣṭubh [?]). By way of metonymy, the word for “song” simply came to be used as the word for a meter as well, and no derivation (as in trá → trī́) was necessary. In tráiṣṭubha, however, the accent is on the first syllable, whereas in gāyatrá it remains on the last; cf.
AG II,2: 134 (§40c–d). The placement of the accent on the first syllable is quite frequent in relational adjectives, and it is not entirely clear why gāyatrá/ī́ (← gāyatrī́) is not analogously formed (* gā́yatra/ī). Possibly, the accent remained on the final syllable because barytonic derivatives ending in tra usually denote a means or instrument, whereas abstract nouns derived from verbal roots are regularly oxytonic (i.e., they end in trá); see AG II,2: 701–705 (§517a–b). *gā́yatra/ī (either as adjective or noun) may have given the impression that the word does not mean “song,” but “a means to sing,” and was therefore avoided. For the root gā/gī(/gai) “to sing,” see VIA I: 283 (no. 278); cf. also EWA I: 482–483.
280 See below p. 191.
72 ∙ paRt i ∙ the mantRa
verse. While in practice, reproducing a meter generally involves the recitation of a text set in that meter, the quite common practice of supplying the word
“verse” when translating from Vedic or Sanskrit can be somewhat misleading: the verse to be recited might not be that important.
The importance and peculiar nature of meters must always be borne in mind, especially, of course, in the case of the GM. That the gāyatrī and the GM
could be confused was, in part, a consequence of ignoring this special role. On the other hand, the name of the mantra really is somewhat bewildering. How could this gāyatrī verse become known as the Gāyatrī verse? I will deal with this question in the second part of this chapter.
2. From meter to mantra
As we saw above, the word gāyatrī was not used as a name of the GM in the AV and in the BṛhĀU. In fact, these are the only Vedic texts for which this use of words was either argued or taken for granted. To the best of my knowledge, all other texts from the same period as well as the (older parts of the) Gṛhya- or Dharmasūtras generally use the word sāvitrī or the pratīka, but never the word gāyatrī, to refer to the mantra. As it turns out, the first texts that unambiguously use the word in this sense are found in the late strata of the MBh and the Dharmasūtras. Above all due to the ample material it provides, the MBh therefore becomes the primary source for the study of the development of the name
“Gāyatrī.”
**2.1 sāvitrī ****and gāyatrī **in the Mahābhārata In the voluminous MBh the words sāvitrī and gāyatrī taken together occur more than 180 times. Depending on the context, they may refer to the GM, to the gāyatrī meter, to the personification or deification of either the mantra or the meter, to the literary character Sāvitrī, or to her eponym, the goddess Sāvitrī.²⁸¹
In several cases, the boundaries are either unclear or blurred. An analysis of all of the relevant passages reveals that in the main text of the MBh, the word gāyatrī in all probability always denotes the meter.²⁸² The mantra, on the other 281 See Appendix 2 (pp. 285–286). Cf. MCI I: 218–219. Note that in this index, the referents of the words sāvitrī and gāyatrī are not (clearly) differentiated.
282 Occurring seven times in the main text and nineteen times in the star passages and appendices.
chapteR 2 ∙ designations ∙ 73
hand, usually goes by the name of sāvitrī.²⁸³ Since the relevant passages are only few, they can all be dealt with here.
In the first passage, VI 32.35b (= BhagG X 35b), Kṛṣṇa proclaims that
“among the meters, I am gāyatrī” ( gāyatrī chandasām aham). As outlined above, the gāyatrī was considered the first and best of meters.²⁸⁴ In the present context, this passage therefore requires no further explanation.
In the second passage, VI 5.9–21, Sañjaya presents a categorization of all living beings in the world. He counts nineteen of them “within the five elements” ( mahābhūteṣu pañcasu). Somewhat unexpectedly, he then asserts that The twenty-four that have been shown are known as a gāyatrī by the world. Whoever truly knows this meritorious gāyatrī possessed of every virtue, does not lose the world, O best among the Bharatas.²⁸⁵
The mention of the number suggests to me that, in all likelihood, the meter is meant: the main characteristic of the gāyatrī meter is that it has twenty-four syllables in three lines (the GM at that time was already reduced to twenty-three syllables, but this was certainly not taken too strictly).
In the third passage, XIV 44.1–16, the god Brahmā gives a long list of beings and things that are “first, foremost,” or “highest.” In the middle of this list, he mentions both the sāvitrī and the gāyatrī. The passage is more or less self-explanatory:
After that, I shall explain next that which is first and highest among things: the sun is the first of lights, fire is taken to be the first among all elements; /4/ among the knowledge-branches / the mantras,²⁸⁶ the sāvitrī [is the first]; among the deities, the Lord of Progeny; among all Vedas, the sound om; and among the utterances, it is breath. If it is fixed in this world, it is said that it all belongs to Savitṛ [?]. /5/ The gāyatrī is first among meters; among cattle, it is said to be the goat. Among the quadrupeds [i.e., animals], cows are first; among humans, the twice-born. /6/²⁸⁷
A special case is given if a text states that a gāyatrī should be “muttered”
or “recited” ( jap or paṭh). In the Tīrthayātraparvan, Pulastya mentions a place 283 MBh II 11.25; III 177.29, 277.9; III 80.4; V 106.10; XII 326.7, 36.33, 43.14; XIII 24.25, 24.28, 85.6, 92.14, XIV 44.5.
284 See above p. 65.
285 MBh VI 5.18cd–19: caturviṃśatir uddiṣṭā gāyatrī lokasaṃmatā // ya etāṃ veda gāyatrīṃ
puṇyāṃ sarvaguṇānvitām / tattvena bharataśreṣṭha sa lokān na praṇaśyati //.
286 For the translation of vidyā, see below p. 192.
287 MBh XIV 44.4–6: ataḥ paraṃ pravakṣyāmi bhūtānām ādim uttamam / ādityo jyotiṣām ādir agnir bhūtādir iṣyate /4/ sāvitrī sarvavidyānāṃ devatānāṃ prajāpatiḥ / oṃkāraḥ sarva-vedānāṃ vacasāṃ prāṇa eva ca / yady asmin niyataṃ loke sarvaṃ sāvitram ucyate /5/
gāyatrī chandasām ādiḥ paśūnām aja ucyate / gāvaś catuṣpadām ādir manuṣyāṇāṃ dvijātayaḥ /6/; cf. the translation by Deussen & Strauss 1906: 970–971.
74 ∙ paRt i ∙ the mantRa
of pilgrimage holy to the gāyatrī. After describing the place, Gokarṇa, he continues:
From there to the place of the gāyatrī, renowned in the three worlds. Having stayed for three nights, one obtains the reward of a thousand cows.
There is an obvious indication of Brahmins, O lord of men: if someone of mixed parentage recites a gāyatrī, it turns into a gāthā or song for him, O
king.²⁸⁸
The mention of the term gāthā, which is especially used for non-Vedic metrical texts, and gītikā or “song” in this passage probably suggests that gāyatrī refers to the meter. The point is that the meter, if recited by someone who is not a full-blooded Brahmin, does not sound like a proper Vedic one, but more like a popular stanza. In this context, it must not be forgotten that like gītikā and gāthā, the word gāyatrī is derived from the root gā/gī(/gai) and could also be translated as “the song-meter.” Nevertheless, the passage easily makes one think of one specific verse, the GM, and it is certainly one of those passages where the meter and the verse can easily be confused. As a matter of fact, some manuscripts of the Northern recension also add the line: “but if a non-Brahmin recites the sāvitrī, he perishes.”²⁸⁹ While this clarification suggests that the passage might imply the recitation of the GM, it would be going too far to assert that the word gāyatrī actually denotes the verse.
These, then, are all the passages in the main text where the word gāyatrī occurs, and as we can see, in all of them the word most likely simply denotes the meter, and not the verse. This situation quickly changes, however, in the star passages and appendices. These contain textual material that is found in one or more manuscripts, but was not included in the main text. These passages are dealt with in the following.²⁹⁰
In several cases it is clear that the word gāyatrī again simply denotes the meter. In II 35.25*358.1, for instance, it is plainly stated that “the gāyatrī is chief among meters” ( gāyatrī chandasāṃ mukham).²⁹¹
In another passage, VIII 24.66–84,²⁹² we find the description of a divine chariot made by (and at least partly even of) the gods for Mahādeva (the horse-288 MBh III 83.26–27: tata eva tu gāyatryāḥ sthānaṃ trailokyaviśrutam / trirātram uṣitas tatra gosahasraphalaṃ labhet /26/ nidarśanaṃ ca pratyakṣaṃ brāhmaṇānāṃ narādhipa /
gāyatrīṃ paṭhate yas tu yonisaṃkarajas tathā / gāthā vā gītikā vāpi tasya saṃpadyate nṛpa /27/; for another and similar translation, see van Buitenen II: 394.
289 III 83.27*439.1: abrāhmaṇasya sāvitrīṃ paṭhatas tu praṇaśyati /.
290 These are II 35.25*358.1; VI 40.78*113.5, App. 3A36, 3B111, 3B36; VIII 24.76*262.1; XII 274.60 App. 28.283; XIII 107.62*491.3; 113.13*569.4; 80.45 App. 9B146–147; XIV 96.15
App. 4.1544, 1552, 4.3121, 4.3126, 4.3201, 4.494 (the latter six passages are dealt with below).
291 Interestingly, this passage has a parallel in Buddhist texts; see Shults 2014: 119.
292 Tr. Bowles 2006: 323–329 (including the appendix).
chapteR 2 ∙ designations ∙ 75
horses of this chariot are Indra, Varuṇa, Yama, and Kubera).²⁹³ This passage is part of the main text. The list of the chariot’s components (elaborated in the appendix) goes on and even includes elements that are not part of the chariot itself. The equipment that comes with it includes, for instance, a goad (in the case of this divine chariot; this is the ritual utterance vaṣaṭ) at whose tip a strap is fixed. This strap is called gāyatrī.²⁹⁴ Already the next verse specifies that the bow-string used by Mahādeva is the sāvitrī (the bow itself is the
“year,” saṃvatsara), thereby showing that (as in XIV 44.4–6 translated above) the two are distinct. A parallel case is found in VII 173.56 App. 25.10, where the gāyatrī and the sāvitrī are both made the reins of the chariot.²⁹⁵
In IV 5.31 App. 4G1–66, Yudhiṣṭhira praises the goddess Durgā in a durgāstava, or “eulogy of Durgā.” In doing so, he calls upon her by many different names, among them gāyatrī sāvitrī.²⁹⁶ While it is most natural to read this expression as a combination of the meter and the mantra, it cannot be ruled out completely that they were intended to be synonymous.
In several other passages, the context suggests that gāyatrī can only denote the mantra, and not the meter. The first passage, XIII 113.13*569, contains a list of meritorious deeds that save a twice-born from crimes causing a loss of caste (known as pātaka s), among them readings of the Vedas, satiating a thousand cows or the recitation of 100,000 GMs.²⁹⁷ The repetition indicates that the GM
is meant, as it is typically recited in this way as a means of purification.²⁹⁸
The passage XIII 107.62*491.3 states that “one should regularly exercise reflection on the Gāyatrī, concentrating on the Sandhyā(/ Juncture worship).”²⁹⁹
In this case, it is the mention of the Sandhyā that suggests that one should reflect on the primary mantra used in this ritual.
In a third passage, the “beginning and the end of the gods are the Gāyatrī
and the sound om.”³⁰⁰ The typical combination with om again suggests the GM.
In other cases, determining the meaning of the word is less easy. At the end of the BhagG (MBh VI 40.78), a number of manuscripts add a passage in which the Gaṅgā, the Gītā, the Gāyatrī, and Govinda are presented together as a set.³⁰¹
293 Bowles 2006: 556.
294 VIII 24.76*262.1–3: vaṣaṭkāraḥ pratodo ’bhūd gāyatrī śīrṣabandhanā / yo yajñe vihitaḥ pūrvam īśānasya mahātmanaḥ / saṃvatsaro dhanus tad vai sāvitrī jyā mahāsvanā /. Cf. VII 173.56*1457.2 and XIII 145.27.
295 gāyatrīṃ pragrahaṃ kṛtvā sāvitrīṃ ca maheśvaraḥ /.
296 IV 5.31 App. 4G22: namo gāyatri sāvitri namas te jātavedasi. For another durgāstava and early Durgā worship in general, see Lubin 2020, especially pp. 41–43.
297 pārāyaṇaiś ca vedānāṃ mucyate pātakair dvijaḥ / gāyatryāś caiva lakṣeṇa gosahasrasya tarpaṇāt //.
298 See below pp. 164–166.
299 gāyatrīmananaṃ nityaṃ kuryāt saṃdhyāṃ samāhitaḥ /. For the Sandhyā or “Juncture (worship),” see below pp. 146–152.
300 XII 274.60 App. 28.283: ādiś cāntaś ca devānāṃ gāyatry oṃkāra eva ca /.
301 VI 40.78*113.5: gaṅgā gītā ca gāyatrī govindeti hṛdi sthite /.
76 ∙ paRt i ∙ the mantRa
A similar statement is found in the so-called Gītāsāra, a text of fifty-one verses preserved only in Kashmiri manuscripts (App. I 3 of MBh VI).³⁰² This “summary”
of the BhagG also has another passage, stating that “this should be recognized to be the highest Gāyatrī: the one known as the unmuttered one ( ajapā).”³⁰³ The line could refer to the idea that the GM is most effective when repeated only mentally.³⁰⁴ As we saw above, in the BhagG itself the word gāyatrī is only used for the meter. Considering, however, that the word ajapa is mentioned (mental japa, too, is a kind of recitation), it is most likely that these passages follow the new trend of using the name of the meter for the mantra.
The last passage is XIII 80.45 App. 9b146–147, which is part of an explanation of the six types of so-called kapilā cows.³⁰⁵ Each cow is associated with a certain deity, depending on her characteristics. In view of the other associations, it seems likely that gāyatrī here refers to a divinity as well. As I will argue in Chapter 8,³⁰⁶ this divinity is most likely the deification of the mantra, and not of the meter.
Looking at the bulk of the epic evidence, we see that the transition from the meter to the mantra may have been, in many cases, fluid. In general, however, it is clear that the use of the word gāyatrī for the mantra only becomes common in the later strata of the MBh (I will return to the issue of the date below).
This question then arises: Why was this alternative designation introduced in the first place? Why did the mantra ṚV III 62.10, already well known as ( the) sāvitrī, additionally receive the name ( the) Gāyatrī?
What this survey reveals is that most sources use either sāvitrī or gāyatrī to denote the mantra, but they do not as a rule replace one with the other. In other words, the two are rarely used in one and the same text as synonyms in the sense of an “elegant variation.” The only exception to this rule that I have found is in the MBh, in Appendix I 4 of the Āśvamedhikaparvan, a rather long text of roughly 1,700 verses. This little-studied text is known under the name VaiṣṇDhŚ,³⁰⁷ and is in many ways similar to the so-called ViṣṇDh (or Viṣṇudharmāḥ).³⁰⁸ In the VaiṣṇDhŚ, the mantra is generally called sāvitrī, fourteen times 302 App. 3A110/B111: gītā gaṅgā ca gāyatrī govindo hṛdi saṃsthitāḥ /.
303 MBh VI 40.78 App. 3B36: gāyatrī sā parā jñeyā ajapā nāma viśrutā /.
304 In medieval texts, ajapa gāyatrī is also used to designate the mantra haṃsa (or so ’ham); see Mallinson & Singleton 2017: 134. The present passage can also be understood in this way (in this case the word definitely refers to a mantra); however, the mention of the Sandhyā in the preceding line, VI 40.78 App. 3B35, probably indicates that the GM is meant.
305 gāyatryāś ca vṛṣaṇayor utpattiḥ ṣaḍguṇā smṛtā / evaṃ gāvaś ca viprāś ca gāyatrī satyam eva ca //.
306 See below pp. 239–240.
307 See generally Rastelli forthcoming-b; cf. Rastelli 2019: 181–182 and 2017.
308 See Grünendahl 1984: 51–54.
chapteR 2 ∙ designations ∙ 77
in all. Four times, however, the author chooses to use gāyatrī instead of sāvitrī.
Let us look briefly at these four passages.
In the first passage, the context is the Sandhyā:
Those who in the morning and in the evening correctly and regularly perform the Juncture (worship), pass and lead across [that is, they attain and bring salvation] by making a boat consisting of the Veda. If someone softly recites the purifying goddess Gāyatrī, the Mother of the Vedas, he does not sink down while taking possession of the earth and the sea.³⁰⁹
We will return to this passage, which presents one of the clearest examples of deification of the mantra, in Part II of this study.³¹⁰ For now, it will be sufficient to observe that gāyatrī here essentially refers to the GM recited in the Sandhyā.
That it is really not the meter, but the verse that is intended is corroborated by the second and third passages that mention the gāyatrī. In both of them, the Lord gives prescriptions for bathing. Among other texts, the gāyatrī should be recited, “accompanied by the Vyāhṛtis and the Humming [i.e., the syllable om].”³¹¹ This leaves little doubt that the GM, which is frequently accompanied by this introductory formula (generally in the order oṃ bhūr bhuvaḥ svaḥ), is meant.
The fourth and last passage of the VaiṣṇDhŚ to be taken into account here mentions that the GM should accompany the drinking of a cow’s purifying urine.³¹² The mention of other Vedic verses in this context confirms that nothing other than the GM is meant. As it thus turns out, in the Śāstra the word gāyatrī always refer to the mantra, and is not once used for the meter.
But the text goes even further. In three passages we learn that Lord Viṣṇu has his own GM.³¹³ This Gāyatrī is, with little doubt, a so-called Viṣṇugāyatrī, a modified GM dedicated to Viṣṇu.³¹⁴ As already mentioned, a great number of these modified GMs exists, and their history has attracted even greater scholarly attention than that of the GM itself.³¹⁵ For our current purposes, we shall stay 309 MBh XIV 96.15 App. 4.492–495: sāyaṃ prātas tu ye saṃdhyāṃ samyaṅ nityam upāsate //
nāvaṃ vedamayīṃ kṛtvā tarante tārayanti ca // yo japet pāvanīṃ devīṃ gāyatrīṃ vedamātaram // na sīdet pratigṛhṇānaḥ pṛthivīṃ ca sasāgarām //; cf. ViṣṇDh 51.1: sāyaṃ prātaś ca yaḥ saṃdhyām upāste ’skannamānasaḥ / japan hi pāvanīṃ devīṃ gāyatrīṃ vedamātaram //.
310 See below p. 239.
311 XIV 96.15 App. 4.1544: savyāhṛtiṃ sapraṇavāṃ gāyatrīṃ ca japet punaḥ /, 4.1552: savyāhṛtiṃ sapraṇavāṃ gāyatrīṃ vā tato japet /.
312 XIV 96.15 App. 4.3201; cf. BaudhDhS IV 5.12 (tr. Olivelle 2000: 337) and ParSm II 11.32.
313 XIV 96.15 App. 4.2778: gāyatrīṃ mama vā devīṃ sāvitrīṃ vā japet tataḥ /, 4.3121: bhagavaṃs tava gāyatrī budhyate tu kathaṃ nṛbhiḥ /, 4.3126: japtvā tu mama gāyatrīm atha vāṣṭākṣaraṃ nṛpa /.
314 See Beck 1994: 53; cf. below p. 79.
315 See n. 17 on p. 3 above.
78 ∙ paRt i ∙ the mantRa
with the topic by only asking what these modified verses might reveal about the GM and, more specifically, about the development of its designations.
**2.2 The modified Gāyatrīs and the **Gāyatrī
The earliest sources for the modified GMs are generally understood to be, first, the Maitrāyaṇī-Saṃhitā (MaitrS II 9.1), and, second, prapāṭhaka X of the Taittirīya-Āraṇyaka (TaittĀ X 1.5–7), which is also known as the Mahā-Nārāyaṇa-Upaniṣad (MNārU).³¹⁶ In none of the available editions of these texts are the modified verses called gāyatrī s, as they are in the MBh. In fact, they are not called by any name at all. One of the earliest non-epic texts do so is probably the Pāśupata-Sūtra (PāśS), which must have been composed before the fourth century ce.³¹⁷ The text mentions a modified GM, the Rudragāyatrī, and in doing so uses the expression raudrī gāyatrī.³¹⁸ The parallel usage in texts like the VaiṣṇDhŚ ( mama gāyatrī, said by Viṣṇu) suggests that the verses modelled after the GM were indeed called gāyatrī s,³¹⁹ even though not all of them strictly follow the rules of the gāyatrī meter.³²⁰
Calling the new verses gāyatrī s made sense: sāvitrī, after all, means “Savitṛ verse,” and as such was not well suited for the purpose.³²¹ The fact that sāvitrī was the name of a goddess and a princess might also have played a role.
Metrical accuracy was, in any case, of little interest to the creators and the pre-servers of the modified GMs. Moreover, it has to be taken into account that the gāyatrī meter had already gone out of fashion by the end of the Vedic period.
At the time of the great Sanskrit Epics, virtually no one composed verses in the gāyatrī meter any more. For most people, every verse that resembled the most renowned verse in the gāyatrī meter could also be called gāyatrī.³²²
316 There also exists an independent recension of this text associated with the AV; cf.
Bisschop 2018a: 4. For the various recensions of the MNārU, see below p. 184.
317 PāśS I 17; see Kajihara 2019: 17–18. The earliest commentary on the PāśS has been dated to the fourth/fifth century ce (Bisschop 2014: 27), which suggests that the Sūtra already existed before that time.
318 For the Rudragāyatrī in particular, see Kajihara 2019: 17–22.
319 The Baudhāyana-Gṛhyaśeṣasūtra (BaudhGŚS), for instance, generally uses sāvitrī for the GM, but occasionally also gāyatrī. However, it also knows of other gāyatrī s, which it specifies as rudragāyatrī (II 18.9, 19.2) and vaiṣṇavī gāyatrī (V 3.6).
320 Three of the eleven modified GMs in the MaitrS are hypermetric (II 9.1: 119.7–120.15; the fifth, eighth, and eleventh); among the twelve modified GMs of the TaittĀ, five are hypermetric (X 1.1–7; the first, fifth, sixth, seventh, ninth, and tenth) and one is hypometric (the eleventh). At least in the ṚV, hypermetric gāyatrī s are extremely rare; see Arnold 1905: 161–162.
321 Nevertheless, designations such as rudrasāvitrī were used as well, if only rarely; see, for instance, AVPar XL 2.5–6 (cf. Kajihara 2019: 19). See also n. 579 on p. 139 below.
322 Cf. Kajihara 2019: 20: “Within the terms Raudrī Gāyatrī and Rudra-Gāyatrī, the word gāyatrī refers not only to ‘a verse in the gāyatrī meter’ in general, but also more specifically to ‘the particularly sacred verse Gāyatrī,’ just as in the case of the Vedic Gāyatrī.”
chapteR 2 ∙ designations ∙ 79
I would argue that this development in turn influenced how the original GM was called. Clearly, the GM was the best-known gāyatrī verse, and merely mentioning the word would have easily brought the GM to mind. But its use as an actual synonym must have been pushed, or perhaps even triggered, by the introduction of new gāyatrī verses. While Sanskrit does not generally make use of a definite article (in the strict sense) to indicate that a word denotes a specific entity, it must have been clear that the gāyatrī verse par excellence is ṚV III 62.10.
In a few cases, authors nevertheless felt the need to clarify that it is really the famous GM that is meant by combining the words sāvitrī and gāyatrī. The Vaikhānasa-Gṛhyasūtra (VaikhGS), for instance, uses both sāvitrī and, less frequently, gāyatrī, for the GM. One time it also designates it as gāyatrī sāvitrī.³²³
Similarly, the TaittĀ uses the word gāyatrī for the GM without further specification,³²⁴ but also uses sāvitrī gāyatrī.³²⁵ It is, in my view, no coincidence that both texts also recognize the existence of modified GMs: the VaikhGS mentions a Viṣṇugāyatrī,³²⁶ and the TaittĀ is even known as a source of the modified GMs. Only under this circumstance does it make sense to distinguish between several, apparently easily confused mantras.³²⁷
This raises the issue of chronology. The prevalent view is that the Vedic texts were the sources for later, post-Vedic traditions, in which the modified GMs, such as the Viṣṇugāyatrī or the Rudragāyatrī, became quite prominent.
Peter Bisschop, however, argued that some of the mantras used in Pāśupata Śaivism, among them the Rudragāyatrī, were not taken from the ancient Vedic texts, but rather inserted into them long after the end of the Vedic period.³²⁸
If this is correct, it shows that in the centuries around the Common Era, texts like the TaittĀ were not yet fixed.³²⁹ Indeed, considering their language, the faulty (one is even tempted to say: “amateurish”) accentuation, and their content, the conclusion that many passages in the TaittĀ are post-Vedic is un-323 VaikhGS VI 1 (tr. Caland 1929: 151; omitting sāvitrī).
324 TaittĀ II 2; cf. below p. 149.
325 TaittĀ II 16; see below p. 160.
326 VaikhGS X 10 (tr. Caland 1929: 222).
327 The use and the order of words do not seem to directly depend on the immediate presence of modified GMs; cf. MBh IV 5.31 App. 4G22, see above p. 75. In the case of the Gopatha-Brāhmaṇa (GopB I 1.31), which generally uses the word sāvitrī to designate the GM, the expression sāvitrī gāyatrī is once used in order to turn one’s attention to the meter: “Teach
[me], sir, the sāvitrī gāyatrī, which has twenty-four wombs, twelve couplings” ( adhīhi bhoḥ sāvitrīṃ gāyatrīṃ caturviṃśatiyoniṃ dvādaśamithunām; also translated by Patyal 1969: 33); cf. also BaudhDhS IV 4.6 (tr. Olivelle 2000: 335), where several manuscripts add the word gāyatrī most often before, and sometimes also after sāvitrī; see Olivelle 2000: 626.
328 Cf. Bisschop 2018a: 2–5; cf. also von Schroeder’s (1900) n. 8 on KaṭhS XVII 11: 253.20–21.
329 Bronkhorst (2016: 32, n. 96) remarks that the TaittĀ “may date in its present form from the beginning of the Common Era.”
80 ∙ paRt i ∙ the mantRa
avoidable.³³⁰ While most of the MaitrS is no doubt ancient, there is little reason to assume that the case of MaitrS II 9, which contains the modified GMs, is much different.³³¹ The MNārU, in turn, has been placed by Doris Srinivasan between the fourth/third century bce and the third/fourth century ce.³³² As Bisschop concluded: “Not everything that is found in the Vedas is necessarily old.”³³³
The testimony of the Epics and the Dharmasūtras, too, points to a relatively late date both for the modified GMs and, most importantly in the present context, for the designation “Gāyatrī” itself. As we saw, in the constituted text of the critical edition of the MBh, the word gāyatrī is practically never used to denote the mantra. This only happens in passages relegated by the editors to the footnotes and appendices. Their placement within the edition does not allow to directly infer their age: it is very well possible that an older text had been preserved outside the epic manuscript tradition, but was then found suitable by a redactor, who then chose to insert it. It is, at least in theory, also possible that an older text was pushed back or replaced by a younger one.³³⁴ Fortunately, however, we do not have to rely on the epic manuscript tradition alone.
Looking outside the epics, we find that the earliest datable texts calling the GM “Gāyatrī” are the metrical portions of the Baudhāyana-Dharmasūtra (BaudhDhS) and the Vasiṣṭha-Dharmasūtra (VasDhS).³³⁵ The metrical portion of the BaudhDhS “are probably not earlier than the third to fourth centuries ce.”³³⁶ The VasDhS, on the other hand, is the youngest Dharmasūtra, and can possibly be dated even to the first century ce;³³⁷ its metrical portions, however, are probably younger.
Interestingly, while the word gāyatrī is used as a name of the mantra in the late strata of the Dharmasūtras, no modified GMs are ever mentioned (this 330 Regarding language and content of the TaittĀ, see below p. 148 and 160 as well as Srinivasan 1973: 173, n. 63. The word sandhyā, for instance, is to my knowledge not attested in any (other) Vedic text.
331 Further research on this issue is needed, and I have to confine myself to a few remarks and references. Gods like Gaṇeśa (mentioned in the MaitrS; cf. Krishan 1990) do not even play an important (if any) role in the MBh itself. The language, too, is only pseudo-Vedic: in the modified Gāyatrīs of the MaitrS, pracodáyāt is always accented, despite the fact that it is now the verb of a main clause (cf. Mirashi 1975: 58); the accent of the preceding word is always on the final syllable regardless of the correct accentuation ( viṣṇú instead of víṣṇu, vahní instead of váhni, sṛṣṭí instead of sṛ́ṣṭi). Similarly, the ending °ā́ya always bears the accent before dhīmahi, even in Bahuvrīhis such as hastimukhá “elephant-faced.”
332 See generally below p. 184 and Cohen 2018.
333 Bisschop 2018a: 5. In a similar vein, Bronkhorst (2007: 198) argued that “much of Vedic literature was still in a state of flux” even in the time of Pāṇini, concluding (p. 206) that
“Vedic texts were still being modified, perhaps even produced, down to the time of Patañjali, and perhaps beyond.”
334 Regarding the improbability of this case, see Mehendale 2001: 193–194.
335 BaudhDhS IV 1.27–28, V 12, 31, VI 1 (tr. Olivelle 2000: 329, 337, 341, 341); VasDhS XXV 9, 12–13 (tr. Olivelle 2000: 451).
336 Olivelle 2000: 7, n. 10.
337 Olivelle 2000: 10.
chapteR 2 ∙ designations ∙ 81
is true for all strata). A number of even later Dharma texts do mention a mantra called durgāsāvitri,³³⁸ durgāsāvitrī, or simply durgā. ³³⁹ It is unknown, however, to which verse or verses these words refer. According to several commentaries, it is ṚV I 99.1 or a combination of ṚV I 99 and the sāvitrī “devasya tvā savituḥ…”³⁴⁰ In any case, the durgāsāvitrīs – if they actually were modified GMs, which is doubtful³⁴¹ – only appear in late strata of the Sūtras. Considering that these Sūtras are not “sectarian,” however, the absence of modified GMs is not really telling. In fact, it would be surprising if they prescribed the recitation of any other than the original, Vedic gāyatrī sāvitrī.
Lastly, it is also possible to adduce a piece of negative evidence. Contrary to what one might expect, the MānDhŚ, presumably composed in the second or perhaps third century ce,³⁴² uses only the word sāvitrī to denote the GM.³⁴³
While this might be a coincidence (or a corollary of the fact that it is a strictly Brahminical text whose allegiance lies with the Vedic tradition), it is perfectly in line with the development outlined in this chapter.
Conclusion
As I have tried to show, little points to the existence – let alone the widespread use – of modified GMs before the second century ce at the earliest, the time when Pāśupata Śaivism (possibly) came into existence.³⁴⁴ Most texts making use of the modified GMs are, in fact, younger. The VaiṣṇDhŚ, for instance, was probably composed at least three centuries later. While it is difficult to establish 338 The transfer of a stem ending in ī to one ending in i is quite common in Epic Sanskrit; see Oberlies 2003: 79–82.
339 See Kajihara 2019: 24–26.
340 This formula is quoted in many texts; see, for instance, VājS I 24.
341 BaudhDhS IV 3.8 (tr. Olivelle 2000: 335) has durgā( -), which is followed by vyāhṛtayo- or mahādoṣavināśanāḥ (depending on the manuscript) and could therefore be either singular or plural. The form durgāḥ is attested in an edition of the Parāśara-Smṛti (ParSm; see Kajihara 2019: 25, n. 82) and explained by a commentator to refer to the hymn (and not only the first verse of) ṚV I 99. In other instances, only ṚV I 99.1 is meant. VasDhS XXVIII 11 (tr. Olivelle 2000: 459) has durgāsāvitrir eva ca. Despite these uncertainties, for Kajihara (2019: 25–26) the existence of the terms “suggests that, at the time of the Dharma texts, some people worshipped Durgā, and called their sacred formula by the name Sāvitrī.
The most probable reason why it was called by the name Sāvitrī would be that it could express the formula’s sacredness by invoking that of the Vedic Sāvitrī proper.” For the early use of the term durgā to designate the (Warrior) Goddess, see Yokochi 2004: 16–18.
342 See Olivelle 2018a: 24 and Bronkhorst 2012.
343 Here again, the designation Gāyatrī for the verse is found in additional passages in several manuscripts; see MānDhŚ II 83 in the critical edition (Olivelle 2005: 418). In the constituted text itself, the word is not mentioned.
344 See Sanderson 2013a: 8.
82 ∙ paRt i ∙ the mantRa
the date of the first creation of the modified GMs, the belief that they must be
“ancient” merely because they have been transmitted in Vedic texts, is questionable. Their designation as “Gāyatrīs,” in any case, cannot be shown to be earlier than the third or fourth century ce.³⁴⁵
The introduction of the designation “Gāyatrī” for the verse ṚV III 62.10
occurred, therefore, more or less simultaneously to the emergence and spread of the so-called modified GMs. In my view, this is not coincidental. Rather, I would argue that the creation of the modified forms of ṚV III 62.10 led to the revival of, or renewed the attention to, a category that had, by that time, become obsolete: the category of gāyatrī verses. Among this group, the Vedic GM was naturally considered the original one – the Gāyatrī, as it were – but was nevertheless sometimes specified as a sāvitrī gāyatrī or gāyatrī sāvitrī. Considering that already in the early Gṛhyasūtras the use of the word sāvitrī was sufficient to denote ṚV III 62.10, the addition of the word gāyatrī can best be explained by the fact that this category had regained significance.
The cumulative evidence therefore suggests that the practice of calling the verse ṚV III 62.10 or its modified forms “Gāyatrīs” only became common after c.
200 ce. This means that the GM acquired its popular name more than a thousand years later than was previously thought. As we have also noted in passing in Section 1.2, the word sāvitrī was not always used as a name for the GM either.
(I will come back to this issue in Chapter 4.) That the mantra literally made a name (or rather, two names)³⁴⁶ for itself only centuries after its composition suggests that its career was not preordained, as traditional and even scholarly texts often seem to assume. This impression will be further reinforced in the following chapter, which deals with the early reuses of the mantra.
345 The new name was apparently in need of an explanation. Kauṇḍinya (fourth/fifth century ce [Bisschop 2014: 27]), for instance, made the following comment: “And why Gāyatrī?
Because the song ( gītā) saves ( trāyate) the singer ( gātṛ), or because the Gāyatrī[-Mantra!]
is set in the gāyatra meter.” PañcBh I 17.10–12: gāyatrī ca kasmāt | gītā gātāraṃ trāyata iti | gāyatre vā chandasi vartata iti gāyatrī.
346 The development of the two designations of the GM is, in a way, similar to that of the designations of the Bible, which is also known as “the Scripture.” The designation “Bible”
ultimately goes back Koine Greek τὰ βιβλία – “the books.” The Latin word scrīptūra (from which “scripture” is derived), on the other hand, could originally be used for any kind of written text, but is nowadays restricted to sacred texts – primarily that of Christianity, but also of other religions.