ऋग्वेदे विश्वासटिप्पन्यः

ABORI

Annals of the Bhandarkar Oriental Research Institute

Altindische Grammatik

ĀgnGS

Āgniveśya-Gṛhyasūtra

Ait.

Aitareya

AitĀ

Aitareya-Āraṇyaka

AitB

Aitareya-Brāhmaṇa

Ān

Āndhra (rec.; see MNārU)

ĀpDhS

Āpastamba-Dharmasūtra

ĀpGS

Āpastamba-Gṛhyasūtra

App.

Appendix (in references to passages in the MBh or the Rām., App.

designates Appendix I of the respective critical editions) ĀpŚS

Āpastamba-Śrautasūtra

Altindische Syntax

Asiat. Soc.

Asiatic Society (Calcutta; from 1832: The Asiatic Society of Bengal; from 1936: The Royal Asiatic Society of Bengal; from 1951: The Asiatic Society) Āśv.

Āśvalāyana

ĀśvGS

Āśvalāyana-Gṛhyasūtra

ĀśvGSPar

(Āśvalāyana-)Gṛhyasūtra-Pariśiṣṭabhāga

ĀśvŚS

Āśvalāyana-Śrautasūtra

Āth

Ātharvaṇa (rec.; see MNārU)

Atharvaveda (the abbreviation is used both for the text tradition and the Atharvaveda-Saṃhitā; see Primary Sources and Translations → AV) AVP

(Atharvaveda)-Paippalāda(-Saṃhitā)

AVPar

Atharvaveda-Pariśiṣṭa

BaudhDhS

Baudhāyana-Dharmasūtra

BaudhGS

Baudhāyana-Gṛhyasūtra

BaudhGŚS

Baudhāyana-Gṛhyaśeṣasūtra

BaudhŚS

Baudhāyana-Śrautasūtra

BEH

Brill’s Encyclopedia of Hinduism

BhagG

Bhagavad-Gītā

BhāgP

Bhāgavata-Purāṇa

BhārGS

Bhāradvāja-Gṛhyasūtra

BhārŚS

Bhāradvāja-Śrautasūtra

BNP

Brill’s New Pauly

xx ∙ gāyatRī: mantRa and motheR of the vedas BrahmP

Brahma-Purāṇa

BṛhĀU

Bṛhad-Āraṇyaka-Upaniṣad (Kāṇva rec.)

BṛhĀUM

Bṛhad-Āraṇyaka-Upaniṣad, Mādhyandina rec.

Brill

Koninklijke Brill NV (Leiden)

Brockhaus

F. A. Brockhaus (Leipzig/Mannheim)

Calcutta ed. (see Primary Sources and Translations → TaittĀ) CarS

Caraka-Saṃhitā

ChāndU

Chāndogya-Upaniṣad

Clarendon Press Clarendon Press (imprint of OUP)

CUP

Cambridge University Press (Cambridge)

Drāviḍa (rec.; see MNārU)

ed(s)./éd.

editor(s), edition(s), edited (by) / éditeur(s), édition, édité (par) EFEO

École française d’Extrême-Orient (Paris)

Egbert Forsten

Egbert Forsten Publishing (Groningen)

EJVS

Electronic Journal of Vedic Studies

The Encyclopedia of Religion

EWA

Etymologisches Wörterbuch des Altinodarischen

feminine

GarP

Garuḍa-Purāṇa

GautDhS

Gautama-Dharmasūtra

Gāyatrī-Mantra = ṚV III 62.10 (see Table 2 on p. 2 below)

GobhGS

Gobhila-Gṛhyasūtra

GopB

Gopatha-Brāhmaṇa

Har.

Harivaṃśa

Harrassowitz

(Otto) Harrassowitz (Wiesbaden)

Harṣ.

Harṣacarita

HirGS

Hiraṇyakeśi-Gṛhyasūtra

HirŚS

Hiraṇyakeśi-Śrautasūtra

History of Religions

Hrsg(g).

Herausgeberin(nen)/Herausgeber (“editor(s)”)

HUP

Harvard University Press (Cambridge, Mass.)

IPS

Index of Primary Sources

Indo-European

IIJ

Indo-Iranian Journal

JaimB

Jaiminīya-Brāhmaṇa

JaimGS

Jaiminīya-Gṛhyasūtra

JaimUB

Jaiminīya-Upaniṣad-Brāhmaṇa

JAOS

Journal of the American Oriental Society

JayS

Jayākhya-Saṃhitā

JIPh

Journal of Indian Philosophy

Kāṇva (rec.; see Primary Sources and Translations → BṛhĀU and ŚatBK) Karm.

Karmapradīpa

KaṭhB

Kaṭha-Brāhmaṇa

KāṭhGS

Kāṭhaka-Gṛhyasūtra

KaṭhS

Kaṭha-Saṃhitā

abbReviations ∙ xxi

KātyŚS

Kātyāyana-Śrautasūtra

Kauṣ.

Kauṣītaki

KauṣB

Kauṣītaki-Brāhmaṇa

KauṣGS

Kauṣītaki-Gṛhyasūtra

KauśS

Kauśika-Sūtra

KauṣU

Kauṣītaki-Upaniṣad

KhādGS

Khādira-Gṛhyasūtra

KūrmP

Kūrma-Purāṇa

KYV

Kṛṣṇa (“Black/Dark”) Yajurveda

LiṅgP

Liṅga-Purāṇa

Mādhyandina (rec.; see Primary Sources and Translations → BṛhĀU, ŚatB, and VājS)

masculine

MaitrS

Maitrāyaṇī-Saṃhitā

MaitrU

Maitrāyaṇīya-Upaniṣad

MānGS

Mānava-Gṛhyasūtra

MānŚS

Mānava-Śrautasūtra

MārkP

Mārkaṇḍeya-Purāṇa

MatsyP

Matsya-Purāṇa

MBh

Mahābhārata

MCI

Mahābhārata–Cultural Index

MER

Macmillan’s Encyclopedia of Religion

MLBD

Motilal Banarsidass Publishing House (Delhi)

MNārU

Mahānārāyaṇa-Upaniṣad

MuṇḍU

Muṇḍaka-Upaniṣad

NārSm

Nārada-Smṛti

OUP

Oxford University Press (Oxford)

Pune ed. (see Primary Sources and Translations → TaittĀ)

Paris ed. (see Primary Sources and Translations → MNārU)

PañcB

Pañcaviṃśa-Brāhmaṇa

PañcBh

Pañcārthabhāṣya

PañcT

Pañcatantra

PAPhS

Proceedings of the American Philosophical Society

PārGPar

Pāraskara-Gṛhyasūtra-Pariśiṣṭa

PārGS

Pāraskara-Gṛhyasūtra

ParSm

Parāśara-Smṛti

PāśS

Pāśupata-Sūtra

PIE

Proto-Indo-European

pos.

position(s)

PraśnU

Praśna-Upaniṣad

Petersburger Sanskrit-Wörterbuch

Rām.

Rāmāyaṇa

rec(s).

recension(s)

Routledge

Routledge (Milton Park), Taylor & Francis (London)

xxii ∙ gāyatRī: mantRa and motheR of the vedas ṚV

Ṛgveda (the abbreviation is used both for the text tradition and the Ṛgveda-Saṃhitā; see Primary Sources and Translations → ṚV) ṚVidh

Ṛgvidhāna

s.v( v).

*sub verbo/verbis *(“under the word(s)”)

ṢaḍvB

Ṣaḍviṃśa-Brāhmaṇa

ŚāṅkhĀ

Śāṅkhāyana-Āraṇyaka

ŚāṅkhGS

Śāṅkhāyana-Gṛhyasūtra

ŚāṅkhŚS

Śāṅkhāyana-Śrautasūtra

Sarv.

Sarvānukramaṇī

ŚatB

Śatapatha-Brāhmaṇa (Mādhyandina rec.)

ŚatBK

Śatapatha-Brāhmaṇa, Kāṇva rec.

SāvU

Sāvitrī-Upaniṣad

SBE

Sacred Books of the East

SDN

Sammlung De Nobili (Vienna)

SGB

Sanskrit Grammar for Beginners

SGS

Sanskrit Grammar for Students

ŚivDhŚ

Śiva-Dharmaśāstra

SkandP

Skanda-Purāṇa

Sanskrit Syntax

SUNY

State University of New York (New York)

Sāmaveda (the abbreviation is used both for the text tradition and the Sāmaveda-Saṃhitā; see Primary Sources and Translations → SV) ŚvetU

Śvetāśvatara-Upaniṣad

SVJ

Sāmaveda(-Saṃhitā), Jaiminīya rec.

ŚYV

Śukla (“White/Bright”) Yajurveda

Taitt.

Taittirīya

TaittĀ

Taittirīya-Āraṇyaka

TaittB

Taittirīya-Brāhmaṇa

TaittS

Taittirīya-Saṃhitā

TaittU

Taittirīya-Upaniṣad

TAK

Tāntrikābhidhānakośa

tr(s).

translator(s) / traducteur(s) / translation(s) / traduction(s) / translated (by)

/ traduit (par); also used to refer to translations (Übb.) into German Trübner

Verlag von Karl J. Trübner (Strasbourg)

Üb(b).

Übersetzerin(nen)/Übersetzer (“translator(s)”) / Übersetzung(en) (“translation(s)”)

UVC

Updated Vedic Concordance

V&R

Vandenhoeck & Ruprecht (Göttingen)

*v.l. *

*varia lectio *(“different reading”)

VādhŚS

Vādhūla-Śrautasūtra

VaikhGS

Vaikhānasa-Gṛhyasūtra

VaikhŚS

Vaikhānasa-Śrautasūtra

VaiṣṇDhŚ

Vaiṣṇava-Dharmaśāstra

Vāj.

Vājasaneyin

VājS

Vājasaneyi-Saṃhitā

abbReviations ∙ xxiii

VārGS

Vārāha-Gṛhyasūtra

VasDhS

Vasiṣṭha-Dharmasūtra

Vedic Grammar

VGS

Vedic Grammar for Students

VIA

Verba Indoarica

ViṣṇDh

Viṣṇudharmāḥ

ViṣṇSm

Viṣṇu-Smṛti

VÖAW

Verlag der Österreichischen Akademie der Wissenschaften (Wien) /

Austrian Academy of Sciences Press (Vienna)

WRV

Wörterbuch zum Rig-Veda

WZKS

Wiener Zeitschrift für die Kunde Südasiens

YājñSm

Yājñavalkya-Smṛti

Yajurveda (divided into KYV and ŚYV)

ZDMG

Zeitschrift der Deutschen Morgenländischen Gesellschaft

∙:∙

Introduction

The short mantra on page v above is among the most frequently recited religious texts of mankind.¹ Since its composition by a poet called Viśvāmitra Gāthina²

more than 3,000 years ago somewhere in the northwest of South Asia, generations of people have repeated it on a daily basis, often more than one hundred times in a row. Over time, it has developed into what has variously been called the most important, most efficacious, and holiest mantra of all,³ and has been placed on a par with other eminent religious texts, such as the Christian Lord’s Prayer,⁴ or the Buddhist mantra oṃ maṇipadme hūṃ ⁵.

The mantra is popularly called Gāyatrī or Sāvitrī, ambiguous terms possessing several denotations (see Table 2 below). First, *gāyatrī * is a Vedic poetic meter. Second, there is a specific verse set in the *gāyatrī * meter, often designated by its source reference, “ṚV (Ṛgveda) III 62.10,” or a variant of it. Mentioning the deity Savitṛ, the verse is a so-called sāvitrī, one among several “Savitṛ verses”

that are found in the Vedic texts. Over the course of time, the verse came to be viewed as the epitome of the *gāyatrī * meter and as the most important sāvitrī.

As a result, both *gāyatrī * and *sāvitrī * came to be used interchangeably as names of the verse: *the * Gāyatrī and *the * Sāvitrī, so to say.⁶ Hereafter then, the verse will be designated the **“Gāyatrī-Mantra” **(GM). Only occasionally will I also call it

“the sāvitrī” (or “the *sāvitrī * verse/mantra”), but always with the definite arti-1

The translation given above is based on the text of the Ṛgveda (ṚV) and is my own; for a collection of other translations, see Appendix 1 (pp. 275–283).

For Viśvāmitra, see generally Sathaye 2015; see also n. 575 on p. 135 below.

The excellence and pre-eminence of the mantra is routinely emphasized in both the primary and secondary literature. The following characterization, taken from a contemporary encyclopedia of religion, is typical: “Within the set of sacred scriptures, a single passage may stand out as the holiest of all, and therefore the most efficacious. Hinduism recognizes the mystic syllable *oṃ * as the essence of all the Vedas, and the hymn known as the Gāyatrī ( *Ṛgveda * 3.62.10), has achieved a place of preeminence among all mantras.”

*MER * VIII: 5304.

See, for instance, Basham 1959: 162 and Anonymous 1902: 115.

Martinengo-Cesaresco 1902: 102. For this mantra, see Studholme 2002.

See Chapters 2 and 4. Because Sanskrit does not use a definite article (or a similar device) to designate unique entities (as for instance “*the * Scripture” is used in English to refer to the Bible), in the source texts it is not always immediately apparent whether these words are used in that sense or not.

2 ∙ gāyatRī: mantRa and motheR of the vedas cle.⁷ Table 2 offers an outline of the various designations that will be used here (including some that have not yet been mentioned, but will soon become clear).

gāyatrī

(1) the *gāyatrī * meter

(2) a verse set in the *gāyatrī * meter

= the Gāyatrī-Mantra = the verse ṚV III 62.10 (as a

mantra)⁸

Gāyatrī

(1) the personification/deification of the GM (occasion-

ally, the GM itself)

(2) the personification/deification of the *gāyatrī * meter sāvitrī

(1) a “Savitṛ verse”

(2) anything relating to Savitṛ (in the feminine grammat-

ical gender; e.g., an *iṣṭi * or “oblation”)

the sāvitrī

= the GM

Sāvitrī

(1) the personification/deification of the GM (occasion-

ally, the GM itself)

(2) Savitṛ’s/Prajāpati’s/Brahmā’s daughter

(3) Aśvapati’s daughter

**Table 2: **Disambiguation of designations used in this study Originally, the GM was simply the last verse in a short and unsophisticated hymn dedicated to Savitṛ, a god who is responsible for various types of motion in the universe (literally, his name means “Impeller”).⁹ During its long 7

The most popular name nowadays is Gāyatrī or Gāyatrī-Mantra. It is due to its popularity and widespread use that these are the designations that will be used throughout this study, even in Vedic contexts. This is undeniably anachronistic (see Chapter 2), but it has the great advantage of consistency. Moreover, in a study where the mantra is referred to on virtually every single page, “GM” is preferable over the much more cumbersome

“ṚV III 62.10.” (Yet another option would be “Gāyatrī-Sāvitrī,” which is somewhat more precise, but rarely used; see below p. 79).

“GM” is used for the verse in its redacted form, both with and without the accents, which in practice are often neither pronounced nor written: t ( )

át savit ( )

úr v ( )

áreṇyaṃ bh ( )

árgo dev ( )

ásya

*dhīmahi / dh( )*íyo y ( )

ó naḥ pracod ( )

*áyāt *//. In ritual practice and recitation, this verse is often supplemented by the syllable *om * and three special words, the so-called Vyāhṛtis ( bhūr bhuvaḥ svaḥ), which are combined with the mantra in various ways to form a new mantra in itself (most often, they are recited before the verse). However, these are not part of the GM proper, as is sometimes assumed (see Rao 2019: 3, n. 2 [see p. 11]). Probably the earliest text to combine the GM with the Vyāhṛtis is the Vājasaneyi-Saṃhitā (VājS

XXXVI 3); see below p. 100 (#13). For *om *(or oṃ, oṁ, etc.), see Gerety 2015 and 2016; for the employment of the Vyāhṛtis, see, for instance, Gonda 1980a: 226.

See below pp. 43–44.

intRoduction ∙ 3

life, however, it was employed for a wide array of functions and ascribed a variety of meanings. The mantra was used as a typical Savitṛ verse in several Śrauta rituals, where it often served as the first mantra in a sequence.¹⁰ Towards the end of the Vedic period, it was chosen to be the primary initiation mantra, and was imparted to the young students as part of the **Upanayana **ritual.¹¹ Recitation of the GM also became an essential component of the Sandhyā, a composite ritual performed in the morning and in the evening (sometimes also at noon).¹² Due above all to its employment in the Upanayana and the Sandhyā, the GM became in many ways the “first and foremost” Vedic verse. As Brahminism evolved, it even came to be seen as the epitome or essence of all the Vedas, and as such became a hallmark of Vedic or Brahminical orthopraxy.¹³ Around the second or third century ce, the author of the famous Mānava-Dharmaśāstra (MānDhŚ) concluded that “nothing is higher than the sāvitrī.”

Brahminical culture continued to exert a lasting influence on the emerging Hinduism¹⁴ throughout the first half-millennium ce. The Vedic system of elaborate rituals, however, declined. New forms of religion developed that often placed more emphasis on personal devotion and adherence to a single deity.¹⁵

The worship of Śiva and Viṣṇu in particular spread widely throughout all strata of society, with kings and rulers officially declaring themselves as the “supreme devotee of the Lord” (i.e., of Viṣṇu), or “of the Great Ruler” (i.e., of Śiva), respectively.¹⁶ An interesting side effect of these developments was the creation of a number of adapted or strongly modified versions of the GM.¹⁷ While these modified GMs are inspired by the form of the original GM, they are directed not at the god Savitṛ, but at a deity revered by the tradition that adapted the mantra, 10

See Chapter 3.

See Chapter 4.

See below pp. 146–152.

See Chapter 5.

With a view to its very diverse manifestations in antiquity and the medieval period, I pragmatically define “Hinduism” as a heterogeneous umbrella category comprising all known religious traditions and systems of the inhabitants of the Indian subcontinent, excluding, however, Buddhism and Jainism. This definition is derived, on the one hand, from an emic view (some) Hindus had of their own traditions around the tenth century ce (see Sanderson 2015: 198), and, on the other, from an etic understanding of non-Islamic religions cultivated by Muslim authors between the tenth and fifteenth centuries ce (see Sanderson 2015: 156, n. 2). Notwithstanding its late origin, I believe it can be usefully applied to earlier phases of South Asian religious history.

Cf. Mirnig & Rastelli forthcoming.

*paramabhāgavata * or *paramamāheśvara *(cf. Sanderson 2013b: 223, who translated the latter as “completely devoted to Maheśvara”); see Mirnig & Rastelli forthcoming. See also generally Sanderson 2009.

These modified GMs have attracted even greater scholarly attention than the GM itself; see Kajihara 2019: 17–24, Bisschop 2018a: 2–4, Beck 1994, Mirashi 1975: 56–59, and Krishan 1990.

4 ∙ gāyatRī: mantRa and motheR of the vedas such as Rudra or Viṣṇu.¹⁸ Even so, the modified GMs drew most of their “potency” from the fact that they were similar to the GM, thus testifying to the high status of the original version.

With the ascendancy of new deities and shifting religious landscapes, the GM evolved as well. While remaining a prominent Vedic mantra, during the first centuries of the Common Era it gradually also came to be considered a **goddess **itself. The history of this deity, called Sāvitrī as well as Gāyatrī, is quite convoluted. Personifications of the mantra are first found in the late or early post-Vedic period.¹⁹ However, an actual “mantra goddess” literally appears for the first time in the famous Sāvitrī story in the Mahābhārata (MBh),²⁰ in which the *sāvitrī * seems to be identified with the epic successor of the Vedic goddess Sūryā Sāvitrī, here only called Sāvitrī. In later texts, the Purāṇas in particular, Sāvitrī regularly appears as a mantra goddess. Already in the MBh she came to be seen as the consort of Brahmā and was soon also identified with Sarasvatī.

Being the personification of the first mantra of the Vedas, she also received the epithet vedamātṛ: the “Mother of the Vedas.”²¹

Little is known about the history of the GM and its manifestation as a goddess throughout the medieval and the early modern period, but it – and she –

seem to have retained their status.²² In the nineteenth century, the GM came to be used by Hindu reform movements to convert people to Hinduism or even transform them into Brahmins, a practice started by Swami Vivekananda.²³ The so-called All World Gāyatrī Parivār, a religious movement founded in 1958 by Śrīrām Śarmā Ācārya, believes the mantra to be the core of Hindu civilization.

In a “Vedic” ritual invented by its founder, the mantra is chanted by thousands of people, many of them converts.²⁴

The modified GMs are generally based on the pattern *tat * X vidmahe, Y *dhīmahi */ *tan naḥ * Z

*pracodayāt *//, with X and Y being divine names in the dative or genitive and Z in the nominative, as for instance in tat puruṣāya vidmahe, *mahādevāya dhīmahi */ tan no rudraḥ

*pracodayāt *//. The translation of these mantras, especially of the first two pāda s, has proven difficult. They are frequently translated as “may we know X, let us contemplate Y” (e.g., in Beck 1994: 51); “let us know X, let us contemplate Y” (e.g., in Gonda1963a: 292);

“we know X, we meditate on Y” (e.g., in Bühnemann 1988: 178); or in a similar fashion.

Brereton & Jamison (2020: 214) translated them as “we know this [mantra, dedicated] to X, we contemplate it to Y” (square brackets in the original). Bisschop & Griffiths (2003: 330) translated “we strive for X, we meditate for Y.” (The translation variants listed here are not to be understood as direct quotations.)

See Chapter 6.

See Chapter 7. While the MBh’s earliest contents possibly go back to the fourth century bce, many of its components only came into being in the course of the following eight or more centuries; see Fitzgerald 2018.

For a portrait of the mantra goddess, see Chapter 8.

Among the few studies known to me are Gupta 1972, Chemburkar 1976, and Goldman

& Sutherland Goldman 2016.

Cf. Larios 2017: 192.

See Bechler 2013 and Heifetz 2021.

intRoduction ∙ 5

As a consequence of the popularization by Vivekananda and reform movements such as the Arya Samaj, the GM has become “a mantra that in contemporary Hindu practice may be chanted by anyone, including those traditionally forbidden even from hearing Vedic mantras, such as women, Śūdras, and Dalits.”²⁵ It has even become part of some cultural currents outside South Asia: Numerous spiritual manuals in various languages are devoted exclusively to explaining what the mantra means and how it is to be used.²⁶ It is printed on clothes, set to music, chanted in yoga studios, taught in spiritual seminars, and explained and interpreted on numerous websites.²⁷

In all these modern contexts, the deification of the mantra is mostly taken for granted. While she is often depicted in a similar fashion as Sarasvatī, there is also a form peculiar to the deification of the mantra, showing her with five heads and ten arms. With its distinct symbolism, this form is reserved for the mantra goddess. In spite of her popularity, however, there are only few temples where the goddess is worshipped nowadays. Most important among them is Pushkar, a town located near Ajmer, which houses one of the few temples dedicated to Brahmā.²⁸ In a number of places in the area, his consorts, Sāvitrī

and Gāyatrī (here they are discretely distinguished) are worshipped as well. The temple of Sāvitrī, who in this case is Brahmā’s first wife, is located on a hill behind Brahmā’s temple, while that of his second wife Gāyatrī (who is a shepherd girl, not the personified meter) is situated on a lower hill on the opposite side of the lake. Since Gāyatrī is much dearer to Brahmā, however, she also has a place by Brahmā’s side in his own temple.²⁹

Sathaye 2015: 247.

To name but three books in English: Iqbal Kishen Taimni, Gāyatrī. The Daily Religious *Practice of the Hindus *(Chennai: The Theosophical Publishing House, repr. ¹⁰2014

[¹1978]); Sadguru Sant Keshavadas, *Gāyatrī. The Highest Meditation *(repr. of the 3rd rev.

ed.: MLBD, 2000 [New York: ¹1978]); S. Viraswami Pathar, *Gayatri Mantra *(Chennai: Sura Books, 2006).

For one such website, see https://siddhayoga.org/gayatri-mantra/exposition/invoking-the-divine (retrieved on February 2, 2022). A number of other examples testifying to the global popularity of the GM in modern times can be cited. Thus, there is an annual “Gayatri Festival,” organized by Miten and Deva Premal ( *https://devapremalmiten.com/holidays/gayatri-festival-corfu-greece * retrieved on October 13, 2018). The GM is also engraved on the belt of the statue of former Beatle George Harrison, located at the Pier Head in Liverpool. It also gained further popularity as a result of occurring in the introduction of *Battlestar Galactica *(2003–2009), a successful science-fiction television series.

Cf. Bailey 1983: 21–26.

An interesting aspect of the story behind this cult is that Gāyatrī, a shepherd girl, could only be married to Brahmā after she had been purified by being put into the mouth of a cow and then being pulled out from its anus. For the various versions of this story, see Malik 1996.

6 ∙ gāyatRī: mantRa and motheR of the vedas 1. The present study

1.1 Aims and state of research

Among many others, the question arises of how the GM could rise to such prominence, even to the point of becoming a deity itself. What caused this mantra to attain the status of one of the most important Hindu texts? How could this mantra – basically a sequence of sounds – come to be revered as a deity? How did a text that is unambiguously addressed to a male god become a goddess? How was this goddess imagined and visualized? During the entire history of their development, the various entities called Gāyatrī or Sāvitrī – the meter, the mantra, the goddess(es), the literary character(s) – were conflated in a multitude of ways, but also continued to exist independently on their own.

Most of this entangled history, however, is unknown.

The present study is an attempt to contribute to filling this research gap by focusing on two key aspects of the history of the GM: its early development and rise as a mantra, and its personification and deification. The two **aims **of the study are:

**(I) **to find out how and in what sense the mantra gained prominence as a religious text;

**(II) **to investigate how it was personified and became an anthropomorphic goddess.

The study thus aims to disentangle the histories of the following elements: (1) the *gāyatrī * meter; (2) the GM as a mantra; (3) its ritual applications; (4) its personification and (5) deification; (6) the goddess Sūryā/Sāvitrī; and (7) her

“offshoot,” the princess Sāvitrī.

Although a large number of casual citations are found in secondary literature, very little attention has been paid to the history of the mantra itself or to its deification. For this reason, the **current state of research **directly relevant for these topics is easily surveyed. Only a few scholars have dealt with aspects of the mantra in more detail, among them Krishna Lal (“Sāvitrī – from Saṁhitās to Gṛhya-Sūtras” [1971]) and Harry Falk (“Savitṛ und die Sāvitrī” [1988]). In recent years, research on the GM has been resumed by Mieko Kajihara (“The Sacred Verse Sāvitrī in the Vedic Religion and Beyond” [2019]); Joanna Jurewicz (“Cognition Begins in the Morning: An Analysis of *Ṛgveda * 3.62” [2021]); and

intRoduction ∙ 7

Joël P. Brereton (“How the Gāyatrī became the Gāyatrī” [2022]).³⁰ However, the process of its ascendancy, especially in the post-Vedic period, has remained largely unexplored.

Even less research has been published on the GM’s deification. Only rarely are more than a few lines dedicated to the mantra goddess.³¹ Even studies on goddesses or literary characters that have certain ties to the mantra (such as Sarasvatī or the princess Sāvitrī) do not usually go any further than mentioning it. To be cited here are a chapter and a long essay by Asko Parpola (“Sāvitrī and Resurrection” [1998], “The Religious Background of the Sāvitrī Legend” [2000]); Renate Söhnen-Thieme’s overview article on Sarasvatī (2018); and Catherine Ludvik’s book *Sarasvatī. Riverine Goddess of Knowledge *(2007).³²

1.2 Scope and sources

While making extensive use of the available secondary literature, including translations, the present study is directly based on primary sources. These comprise a great variety of Vedic and Sanskrit texts that were composed in, and often even across, several historical periods. The point of departure was naturally determined by the GM’s first attestation in the ṚV, whose hymns were composed before the twelfth century bce.³³ However, many Vedic texts (including the ṚV

itself)³⁴ only attained their final form in the following, first millennium bce. The end point – or better end period – was determined by the time when the mantra began to be regularly and widely worshipped as an anthropomorphic goddess, which began only after the mantra had already been elevated to a certain status.

While the origins of this development go several centuries further back, it is not until the third or fourth century ce that we can speak of a true mantra goddess.

The “career” of this goddess reached an initial peak in the last third of the first millennium ce, when Tantric elements were integrated into her worship – a significant and lasting innovation.

The **scope **of the study thus essentially covers the period between 1000 bce and 1000 ce (neither end point, however, is to be understood as a sharp boundary).³⁵ The number of Vedic and Sanskrit texts from this long time-frame is con-30

These studies are discussed in detail in the individual chapters; see Section 1 in Chapter 3

and Section 1 in Chapter 4. For general overviews of the GM, see also Kane II(1): 302–

304; Gonda1963a: 259–261, 274, 276, and especially 284–294; and Brereton & Jamison 2020: 213–217. Cf. also Hatcher 2019.

See, for instance, Hopkins 1915: 86 (§41); Leeming 2001: 150; Varenne, EU; and Timalsina 2018.

Parpola’s and Ludvik’s studies are addressed in Chapters 7 and 8, respectively.

See Witzel 1997b: 263.

See Witzel 1997b: 324–326.

For reasons of time as well as for the sake of a concise presentation, I have only briefly examined some of the later sources or have dispensed with textual analyses altogether.

8 ∙ gāyatRī: mantRa and motheR of the vedas siderable. In total, a selection of more than one hundred texts have been taken into account, many of them mentioning the mantra more than once. Most of them can be roughly divided into the following categories:³⁶

• Vedic literature: e.g., Saṃhitās, Brāhmaṇas, Āraṇyakas, Upaniṣads

• Ritual manuals: e.g., Śrautasūtras, Gṛhyasūtras, Gṛhyapariśiṣṭas; including also Tantras as well as Tantricized and Smārta texts

• Dharma literature: e.g., Dharmasūtras, Dharmaśāstras, Smṛtis

• Epic and (early) Purāṇic literature: e.g., the MBh and its appendices; Purāṇas

The sources and the criteria for their selection are treated in more detail in the individual chapters; for their approximate dates, see Table 1 on page xviii above.

1.3 Approach and methodology

In view of the great variety of sources and the aims of this study, the following three primary tasks emerged for its implementation:

**(a) **locating relevant text passages in the available Vedic and Sanskrit literature

**(b) **reading and interpreting them against their textual and historical backgrounds

**(c) **analyzing and evaluating their role in the history of the GM and its deification

The three tasks required the deployment of a variety of methods. To accomplish the first task (a), a range of research techniques were utilized. Many passages could easily be found by means of searching keywords or collocations of keywords such as gāyatrī, sāvitrī, tat savitur, and others in digitized texts. The number of digitized Vedic and Sanskrit texts has grown exponentially since the 2010s, and a number of tools were employed to search through them. To be mentioned here are, above all, *GRETIL *( Göttingen Register of Electronic Texts in Indian Languages; http://gretil.sub.uni-goettingen.de/); *SARIT *( Search and Re-

→

The latter applies in particular to the Devī-Bhāgavata-Purāṇa (see Śrīkṛṣṇadās 1983 and Vijnananda 1922) and the Gāyatrī-Tantra (= Devī-Bhāgavata-Purāṇa XII), two long texts that deal extensively with the mantra goddess but lie at the extreme edge of the scope of the study (c. tenth century ce or later).

For a general and concise overview of the identity, role, and function of the texts belonging to these categories, see Leach 2014.

intRoduction ∙ 9

trieval of Indic Texts; https://sarit.indology.info); and *DCS *( Digital Corpus of Sanskrit; http://www.sanskrit-linguistics.org/dcs/index.php).

However, since there are also many texts that are not digitally accessible, indices in (and of) primary and secondary texts were used as well. Moreover, due to the fact that digitized texts are often unreliable and/or do not include secondary (but relevant) information, many passages had to be looked up in their original – that is, printed or scanned – editions. Due to the fact that slight differences in wording can have an impact on how a single word or phrase is understood – and often passages concerning the GM do not exceed one or two sentences – different manuscript readings, when included in the editions, were also taken into account. The results of this research have been recorded in several internal databases and reference lists.

As regards the second task (b), the collected text passages were subsequently assessed by a **close, philological-historical reading **of those sources.

Since the GM is more often simply cited or quoted and only rarely the sole subject of longer explanations or expositions, taking the context and the immediate textual environment thoroughly into account was crucial. Sometimes the use of certain words would indicate the esteem in which the mantra was held, or whether it was just viewed as one mantra among many others. Especially in the case of liturgical texts, it is the combination of the GM with other mantras as well as its placement among them that gives access to the meaning or significance of the mantra itself. From a more distant vantage point, the statistical frequency of citations and reuses in certain sub-corpora also provided some clues as to the relative prominence of the GM, especially when compared with other, similar mantras.

In the case of the goddess, attributes, epithets, and qualifiers (such as *vedamātṛ * or devī) naturally play an important role. Despite the apparent clarity of such words, however, in many cases it is difficult to distinguish between the personification of the mantra and outright deification.³⁷ For these reasons, it became even more important to give complete attention to textual details in order to fan out the full range of possible interpretations of the texts. This, in turn, made it possible to draw inferences about the religious realities reflected in the texts.

Retracing the paths of the GM and its deification **(c) **involved, above all, **establishing the chronological order **of the relevant text passages by means of wide reading and the use of whatever secondary literature proved available.

As is well known, establishing the absolute dates of Vedic and Sanskrit texts is often extremely difficult and tracing of precise chronological developments 37

I will return to this issue in Section 3.4 (pp. 30–34) below.

10 ∙ gāyatRī: mantRa and motheR of the vedas is often nigh impossible.³⁸ In view of the fact that many of these texts contain passages from different time periods, awareness of implicit assumptions about the date of *a * text was essential: not infrequently, later developments are reflected in textual interpolations that are tacitly added to the original material.³⁹

In many cases, however, it is indeed feasible to distinguish between earlier and later phases, at least among different texts; this allowed for deliberations concerning intermediate developments.

In order to understand the GM’s rise to a prominent mantra (even so far as its becoming a deity) against the wider religious background (as opposed to its being an isolated phenomenon), it was above all necessary to gain a good understanding of two crucial topics: mantras and deification. The background knowledge concerning these topics that has informed my analyses is introduced in the next two sections, where I also delve deeper into theoretical and methodological aspects. Following that, I briefly outline the structure and contents of this study.

Because both mantras and deification are very large topics, I have selected a number of key aspects that I discuss in greater detail (especially for the former). In the first section, which is devoted to mantras, I focus on four aspects: the relationship between mantras and language (2.1); how Vedic mantras are employed in Gṛhya and Śrauta rituals (2.2); the rise of several individual mantras to so high a status that they are given names (2.3); and the deification of mantras in Tantric contexts (2.4).

Continuing from the last point, I turn to the topic of deities and deification.

Here, I focus on the nature of deities, the conditions of their emergence, and the deification of humans, natural objects and phenomena, and abstractions (3.1); the meaning of the term personification in the context of deities and deification, including the question of the perceived fictitiousness or realness of such personifications, and their path to becoming actual deities (3.2); then I present a compact scheme to explain the process of deification (3.3); and finally offer a few methodological considerations (3.4).

In developing my theory of deification, I have to some extent followed an **interdisciplinary approach **that combines philological research with perspectives and insights from religious studies. Because the deification of mantras –

and, indeed, the process of deification in general – has been little researched, I have also drawn on research on personification and deification in Greek, Roman, and other ancient religions. It should be emphasized that the present study 38

As dates and the divisions of texts into earlier and later “layers” continue to be a subject of debate, writing about them is a very difficult task. As Feller (2004: 47) aptly puts it, “at times the feeling is rather like that of walking in a mine-field, where one’s smallest step should be watched.”

The prime example for this is the development of the “Gāyatrī passage” in the various versions of the Mahā-Nārāyaṇa-Upaniṣad (MNārU); see below pp. 184–193.

intRoduction ∙ 11

does not claim to offer a full-fledged analysis of the deification of the GM from the perspective of religious studies. By tentatively situating the deification of GM within a broader framework, however, I hope to have facilitated such an approach.

2. Mantras

2.1 Mantras and language

Being a salient feature in almost every major religious tradition originating in South Asia, the concept of mantra has become widely known.⁴⁰ The word

“mantra” has even become a part of the English language. Many definitions have been given, but none has found general acceptance. One of the most universal is that of André Padoux:

a mantra is a formula or a sound with a fixed and prescribed form, to be used according to certain rules and in prescribed circumstances, and em-powered with a general or a specific efficacy acknowledged by the tradition wherein it is used.⁴¹

Most mantras could indeed be called “potent formulae,”⁴² in the sense that they are fixed sequences of words that, when recited or written, are expected to yield a certain result. Many of them are used like – or as – prayers, spells, incantations or commands. The relationship between mantras and linguistically meaningful utterances, however, is by no means straightforward. A mantra may contain meaningful words or a meaningful sentence (or sentences), but it may also be a combination of completely meaningless – that is, non-lexical – sounds, as for instance in the case of hsṣmlraṃ, known variously as the “Vyāpaka-Mantra,”

“Saptākṣara-Mantra,” or “Piṇḍa.”⁴³

But even when they are made up of “normal” words, one often has the impression that mantras are not used in the same way as language as it is normally 40

For general introductions, see Gonda 1963a, Alper 1989: 1–14, and Burchett 2008: 813–

818; for Vedic mantras in Śrauta/Gṛhya ritual, see generally Gonda 1977: 502–508 and 565–581; for Tantric mantras, see generally Padoux 2011.

Padoux 1990: 379. At the same place, Padoux also points out that the origin plays a crucial role in defining whether such a formula is a mantra or not: “a mantra is what is pronounced as such by the revealed tradition and the teaching of the masters: it is a formalized utterance declared to be a mantra, ‘revealed’ by those texts and masters who are entitled to do so, that is, who are recognized as holding authority in this matter by the group to which they belong.”

Attributes such as “potent” or “powerful” are preferable to the much more elusive and loaded adjectives “sacred,” “religious,” or “magic”; cf. Burchett 2008.

For more on this mantra, see Rastelli 1999: 137–139.

12 ∙ gāyatRī: mantRa and motheR of the vedas understood, that is, as a means of communication. For instance, the three words bhūr bhuvaḥ svaḥ, known as “the Vyāhṛtis,” taken together constitute a mantra that is used in a great number of ritual contexts.⁴⁴ While they can be translated as eaRth, inteRspace, sKy, this literal meaning plays no role at all; the mantra is not meant to “tell” anyone anything, not even the reciters themselves.⁴⁵ The question is therefore: What in fact is the linguistic status of mantras?

The scholarly debate regarding this question has above all been driven by Frits Staal, who argued that rituals as well as the mantras recited in them are essentially meaningless.⁴⁶ His fundamental observation concerning mantras is that while they may *consist * of language or language elements – and most often do – they *are not * language (as, for instance, prayers or poems are,⁴⁷ though they, too, may be used *as * mantras). The linguistic utterances a mantra contains undeniably have a meaning of their own: a mantra may even consist entirely of perfectly intelligible semantic content, which can also be translated into other languages. One may even try to determine the “linguisticality” of a certain mantra in terms of the degree to which it contains intelligible content.⁴⁸

The function of the mantra or the meaning attributed to it, however, do not necessarily depend on this content directly (which may also be incomprehensible due to its being enigmatic or archaic). A certain effort or even special knowledge is very often needed to construe the meaning or purpose of a mantra in a given context. However, in most contexts known from South Asian religions, the reciters of a mantra are by no means required to always make this effort, nor are they obliged to have such knowledge. Meaning, therefore, is not necessarily an essential feature of many mantras, and in this sense they may indeed be called “meaningless.”⁴⁹

This as well as many of Staal’s other theories have been the subject of con-troversy, with most scholars feeling the urge to argue against them.⁵⁰ Indeed, it 44

See, for instance, Gonda 1980a: 226.

It is also impossible to interpret all mantras as speech acts; cf. Staal 1989b: 66: “[A]ll speech acts involve intention; since all mantras do not, mantras cannot be speech acts.”

Cf. also Wilke & Moebus 2011: 406–407.

For a summary and references, see Staal 2008: 191–241. Staal was not the first to make this claim: “Even some within the Sanskrit tradition, including Kautsa in Vedic times and the fifth century Buddhist philosopher Vasubandhu, argued for the meaninglessness of mantras. In modern Bengal as well, the phrase *mantra-tantra *(or tantra-mantra) is frequently used in the pejorative or dismissive sense of ‘mumbo jumbo’” Yelle 2003: 15.

Cf. Padoux 2011: 9.

As proposed by Alper 1989: 6–8.

Nevertheless, it should also be mentioned that the Sanskrit tradition in general strongly associates mantras with vāc, that is, “speech” or “language.” Mantras may even be considered the highest form of language. According to Padoux (1989: 299–300), this is above all to be explained by the peculiar theories and notions about language in this tradition, rather than by what language actually is.

See, above all, Penner 1985; see also the various contributions in Thompson & Payne 2016

as well as Alper 1989: 10–12; cf. also Patton 2005: 61–62.

intRoduction ∙ 13

is no exaggeration to say that Staal overemphasized the formal – in themselves semantically meaningless – aspects of mantras. In my view, however, one of his basic insights – namely that mantras, like songs, should be considered *containers * rather than *content *– remains a good starting point for further reflections.⁵¹

Above all, it allows us to think about their semantic content independently from their use and purpose – and the other way round too. Mantras belong to a category of their own. Just as it would be insufficient to study songs *only * as texts, mantras must not simply be analyzed as manifestations of language. Because most of them consist of language, however, linguistic concepts are indispensable to fully understanding them.⁵² In fact, in the following it will become clear that the content of a mantra usually does play a major role in its application.

2.2 Mantras in Gṛhya and Śrauta ritual

Since the end of the nineteenth century, there have been repeated efforts to understand how mantras come to be used in certain contexts, with early scholars focusing especially on Vedic mantras. Frequently, the relationship between mantras and the ritual acts they accompany was viewed more in terms of quan-tity rather than quality. The perceived gap between semantic meaning and ritual practice often gave rise to judgements about the degree of their applicability –

or, much more often, inapplicability – in certain contexts.⁵³ In the following survey, I will review some of the most important theories regarding the application of mantras. In doing so, I will focus on “practically oriented” theories concerned with Vedic mantras (as the GM is Vedic), dealing first with Gṛhya and then with Śrauta ritual.⁵⁴

Among the first to study the application of mantras was Edwin W. Fay.

In his 1890 dissertation, which dealt with the occurrences of Ṛgvedic mantras in Gṛhya ritual and mantras, he distinguished above all between “general”

and “specific” applicability.⁵⁵ A mantra belongs to the first category if it “has a merely general applicability, and would serve on almost any conceivable occasion as well as for the one in which we find it employed.”⁵⁶ Specific applicability, 51

Gerety (2015: 25), too, has recently made an effort to rehabilitate Staal’s theory; cf. also Lubin 2016c: 146–148.

Cf. Davidson 2014: 5–10.

Cf. Apte 1939: 14–15.

For similar surveys, see Patton 2005: 76–83 and Lamers 2012: 4–9.

See especially Fay 1899: 26. As he studied the occurrences of mantras in a Sūtra, he further distinguished “homonymous citations” (“Here the m a n t r a cited is utterly out of relation to the ritual, but lugged in because the m a n t r a accidentally contains some word inherent to the S ū t r a” p. 22) and “warranty citations” (“Sometimes the m a n t r a is cited as a warranty for a belief, much like legal citations now, or like proof-texts in the doctrinal study of the Bible” p. 25).

Fay 1899: 14.

14 ∙ gāyatRī: mantRa and motheR of the vedas on the other hand, is given if the content of the mantra has a specific link to the ritual act it accompanies.

This distinction turned out to be problematic. Directly referring to Fay’s work, Vinayak M. Apte remarked that “such a classification is, in my opinion, too general, too superficial to do justice to the *definite principles * underlying the citation of a mantra in a particular rite.”⁵⁷ In fact, Fay himself already noted that

[b]etween the opposing poles […] of general and specific applicability, lies a class of quotations hard to refer absolutely to either extreme […] One difficulty that will meet us in testing the specific applicability of a m a n t r a is of this sort: a verse of a purely general sense may contain some word that has suggested a specific rite to accompany it.⁵⁸

Fay tried to solve the problem by creating a catch-all category (entitled “A”), which includes mantras with either general or specific applicability.⁵⁹ This category, however, only indicates that a certain mantra is applicable – an insight already established by the fact that the mantra evidently *was * used in a ritual.

As Apte recognized, applicability is more a matter of kind rather than degree. In his study of the application of mantras in the Āśvalāyana-Gṛhyasūtra (ĀśvGS), he proposed five distinct categories:

• The **sacramental **class: “When a Ṛgveda mantra has the same sacramental setting or context in the RV. [= ṚV] itself, as it has in the ĀG. [= ĀśvGS]

where it is cited, I call that citation ‘ *sacramental. *’”

• To the **invocational **class belong “[v]erses or hymns containing prayers for blessings in general […] or for some special blessings connected with the particular rite under description.”

• The **mythological **class “of citations […] become appropriate in a rite mainly because they are addressed to a deity who is associated with that rite.”

• **Oblational **citations “become appropriate to a ritual act through oblations *to Agni * or in other words because the act is accompanied by oblations to the domestic fire […] These are naturally all ‘Agni verses’.”

• The **superficial **class: to this class belong “citations […] whose only link with the ritual context is some superficial resemblance in the form of a common word or phrase without any relationship to its meaning […] Even in these few cases […] the choice of the mantra is not purely arbitrary, 57

Apte 1939: 15.

Fay 1899: 17.

Another category, “B,” includes mantras with general or specific applicability as well as homonymous citations.

intRoduction ∙ 15

but is dictated either by earlier liturgical employment or some striking metaphor.”⁶⁰

In a very similar manner, Laurie L. Patton (2005) distinguished four general functions:

• **Consecratory **function: “mantras that make sacred a particular act, such as [a] wedding or a funeral.”

• **Oblational **function: “mantras that refer to the power of Agni as the oblation is poured into the domestic fire.”

• **Purposeful **function: “mantras that comment briefly on the larger purpose, or significance of the act they are to accompany, such as the gaining of progeny of wealth.”

• **Benedictions **or aversions: “mantras that are expressions of wishes, such as for future health, as well as for avoidance of an evil spirit.”⁶¹

For Śrauta ritual, the second type of Vedic ritual, a number of categories have been proposed as well. As a rule, mantras are not used here in an isolated fashion, but forged together in litanies. In a single ritual, these litanies can comprise hundreds of different mantras, all taken out of their original contexts and arranged in a new order. In some cases, mantras or their components are also expanded, compressed, taken apart, or rearranged.⁶²

Among the first to study this subject was Viman Ch. Bhattacharyya (1953), who analyzed the application of mantras in the Aitareya-Brāhmaṇa (AitB).⁶³

In contrast to Apte, Bhattacharyya’s basic assumption was that most mantras recited in Vedic rituals were actually misapplied, and that the aim of the commentarial literature, the Brāhmaṇas, was to explain this misapplication away by means of certain “devices.” The devices identified by him are the following:⁶⁴

• The **single word **device: the occurrence of a certain word (or sometimes a number of words) in a mantra or hymn is said to justify its recitation.

• The **pattern **device: a prominent refrain entails the employment of the entire hymn.

All quotations are from Apte 1939: 16–17; cf. also Gonda 1977: 568–571.

All quotations are from Patton 2005: 66–67.

For examples, see Staal 1989b: 48–58.

Summarized in Gonda 1979a: 236–237; see also Bhattacharyya 1955.

See Bhattacharyya 1953: 300–305.

16 ∙ gāyatRī: mantRa and motheR of the vedas

• The analogy, simile, or **symbol **device: an analogy or correspondence is assumed between certain elements mentioned in a mantra and ritual or real-life elements.

• The “sage” or “**school **device” (as it could be called): the use of a hymn or verse is explained to be authoritative in another Vedic school or to have been effective in the case of a well-known Vedic sage.

• The “Ākhyāna” or narrative device: mantras are related to a myth or legend that is in some way connected with a ritual detail.

• The “Ṛk” or “**gāthā **device”: an authoritative Ṛgvedic text or popular saying is drawn upon in order to show the prevalence of a certain ritual practice.

Bhattacharyya’s main aim was to “dismantle” these devices and to show their ineptitude to explain the connection between the recited texts and ritual events.

I would argue, however, that the devices may in many cases reflect the principles according to which the mantras were selected and put together in the first place.

Rather than assuming a general failure on behalf of the commentators, I think it is more probable that they often followed a ritual logic whose foundations had been laid long before.

I would even suggest adding two further principles to Bhattacharyya’s list.

As Jan Gonda noted, “mantras are often said to be suitable for reasons that are foreign to their contents, for instance their metrical structure.”⁶⁵ Second, the “deep structure” of the ritual or litany itself may determine the need for a mantra calling upon a particular god.⁶⁶ The fact that a mantra is addressed to a certain deity can be decisive as well, even if that deity is not explicitly named in the mantra.⁶⁷ Thus, the list could be extended by the “metrical” and the “deity”

principles (the latter, in fact, corresponds to the “mythological class of citations”

Apte postulated for Gṛhya ritual and mantras).⁶⁸

Only some of the principles, functions, devices, and categories already identified – most prominently, Bhattacharya’s “analogy device” – presuppose a connection between the semantically meaningful elements of a mantra and actors, actions, and objects, in the world. As scholars such as Laurie L. Patton have shown, however, it is often not too difficult to find such connections if the ritual context is thoroughly taken into account – especially in the case of Gṛhya ritual. These generally follow an associative logic.⁶⁹ As one of many examples, 65

Gonda 1979a: 236.

As shown, for instance, by the structure of the Vaiśvadeva-Śastra, see below p. 93.

As is the case of ṚV V 50.1 in the Vaiśvadeva-Śastra, see below p. 93 (#6).

The deity principle was also recognized by Yāska, the author of the Nirukta (c. fourth–

third century bce); see Visigalli 2016.

Patton (2005: 75) has also called them “metonymic connections.” For a review of her study, see Smith 2006.

intRoduction ∙ 17

one may look at the use of a certain Ṛgvedic hymn, ṚV III 33, in Gṛhya ritual.⁷⁰

This hymn, which is also known as the “River Hymn,” is essentially a dialogue between the poet, Viśvāmitra, and the two rivers Vipāś and Śutudrī, which are asked to cease their flow in order to allow the Bharata forces to cross. According to the Ṛgvidhāna (ṚVidh) and the Śāṅkhāyana-Gṛhyasūtra (ŚāṅkhGS), the hymn or, in the latter case, only its last verse, should be recited as a mantra by someone who is crossing a river. As Patton pointed out, the reciter of the hymn is in these cases associated with its poet; the river to be crossed, in turn,

“is likened to the gracious primordial rivers, Vipāś and Śutudrī, who acceded to the sage’s request.”⁷¹ Thus, the reason for which this mantra is recited is not just a loose thematic appropriateness; rather, it is possible to establish very concrete, associative linkages between the elements in the text of the mantra and the elements of reality.

The way in which a creator or performer of a ritual mentally establishes such linkages, however, may vary.⁷² Many reciters may not mentally engage with the mantras they utter and the ritual acts they perform at all, either because they do not try to or simply cannot understand their purport. But although these linkages may not be as obvious as other principles, looking for them is not superfluous. As Bhattacharya remarked, the analogy device – or, as I would suggest calling it, the “**associative **principle” – “is the most commonly accepted device of all.”⁷³

The associative principle can indeed be found both in Gṛhya and Śrauta rituals. This fact as well as the various examples to be discussed in this study suggest that the use of mantras in the two related ritual systems are often guided by the same, or similar, principles. This means that, when analyzing mantras in one system, it would be inappropriate to rule out the presence of a trait of the other system. Rather, it is advisable to take into account as many aspects as possible. I propose to pay attention to the following four aspects in particular: the meter; the mention of possibly relevant words; the deity to whom the particular mantra is dedicated (it is important to note that the name of this deity is not always explicitly mentioned); and any associative linkages that may be established between the actors, actions, etc. mentioned in the mantra and those of the real/ritual world. Moreover, it should also be taken into account who composed the mantra, what position it occupies in its hymn of origin, and how it was used in other contexts. Based on this information, it is usually possible to explain why a mantra was used in a particular context.

For an analysis, see Patton 2005: 161–164.

Patton 2005: 164.

Cf. Patton 2005: 62.

Bhattacharyya 1953: 303. In fact, this principle also plays a major role in the system of the Pūrvamīmāṃsā; see Lamers 2012.

18 ∙ gāyatRī: mantRa and motheR of the vedas As the above survey illustrates, mantras may be “legitimately” used in a variety of contexts (the recognition of this fact even led to the creation of the term “hyperapplicability”⁷⁴). The application of Vedic mantras is clearly not hap-hazard, but follows certain principles. An important aspect, especially in Vedic ritual, is that mantras are most often *selected * according to these principles (or a combination of them) and are then adaptively reused⁷⁵ rather than modified and adapted to a particular ritual context.⁷⁶ Although variations are not at all uncommon, most mantras are used as they are. The process of their selection, however, is not always straightforward – especially when considering that there are so many mantras to choose from.

2.3 Pre-eminent mantras

The total number of Hindu mantras in existence is traditionally given as 70 million.⁷⁷ The actual number is certainly much smaller, but still very large. The Updated Vedic Concordance, for instance, contains close to 90,000 entries, each of which can theoretically be called a Vedic mantra.⁷⁸ The number of Tantric mantras is far more difficult to count, but certainly does not fall behind that figure. In practice, any particular tradition (Vedic, Tantric, or other) preserves and uses only a subset of mantras, but always acknowledges them as being parts of limited, yet vast corpora. To give an overview of the composition and structure of even one of these corpora – that is, identifying groups or subsets according to their similar wordings or applications – is not an easy task. However, when looking at the bulk of mantras from a distance, one thing quickly becomes clear: a very small number of them clearly stand out from the crowd. Within and also across the various traditions, we can observe that a few mantras have become far more prominent or popular than the others.

Patton 2005: 67.

Freschi & Maas (2017: 14) distinguish “adaptive” from “simple” reuse as follows: “In contrast to simple re-use, adaptive reuse is not merely the repetition of a previous use; it implies more than an item just being used again. In adaptive reuse, the reuser expects his or her audience to recognize the reused elements in order to achieve a well-defined purpose, as for example adding prestige, credibility, etc., to the newly created item. Adaptive reuse may involve a more substantial change in the usage.”

Nevertheless, mantras were modified in some cases. There are even special rules for this procedure; see, for instance, Lamers 2012 and Bronkhorst 2007: 188–191.

Padoux 1989: 310.

The *UVC * can indeed be called “a comprehensive index of all mantras” (p. xxii); it should be noted, however, that not all of them were actually used as individual mantras. Thus, the majority of b- pāda s – all of them indexed as separate entries – were only recited together with the rest of the verse. According to Gonda1963a: 266, “The number of Vedic mantras included in the ritual handbooks for the performance of the domestic rites ( Gṛhyasūtras) comes, for instance, approximately to 2,500.”

intRoduction ∙ 19

This is shown, for instance, by some of them being given proper names.

A number of examples are easily given. The verses ṚV X 9.1–3, for instance, are called “Abliṅgas,” because they are addressed to the personified Waters –

*abliṅga * literally means “whose characteristic are the Waters.” The verses are prescribed by a number of Dharmasūtras as a means for purification.⁷⁹ While there are of course many verses in the Vedic corpus that mention the Waters, these three verses were used so frequently that the designation *abliṅga * was established as their name.

A similar example is that of mantras that are typically associated with a certain deity, which may then be directly named after that deity. Perhaps the most prominent example for this is, indeed, “the sāvitrī” (i.e., the GM), which mentions Savitṛ.⁸⁰ In other cases, a prominent ritual application played a role in naming. The verse ṚV VII 59.12, for instance, is often used in rituals whose purpose is to overcome death. Accordingly, it is not only called “Tryambaka”

(after its first word, which is also its deity), but also “Mahāmṛtyuṃjaya-Mantra,”

the “Great Death-Conquering Mantra.”⁸¹

The practice of giving prominent mantras names was continued throughout all historical phases, both for Vedic and Tantric mantras. One of the most important mantras among the devotees of Viṣṇu is the “Twelve-Syllable” or

“Dvādaśākṣara-Mantra”: oṃ namo bhagavate vāsudevāya, “om – obeisance to Lord Vāsudeva.”⁸² While there are also other mantras of twelve syllables, this designation is most often used for this mantra. Similarly, the “Five-Syllable”

or “Pañcākṣara-Mantra” usually designates one of the most important mantras dedicated to Śiva:⁸³ namaḥ śivāya, “obeisance to Śiva.” While both terms are also used to designate other mantras, in most contexts (especially, of course, within the respective traditions) they are effectively used as proper names.

In several cases, pre-eminent mantras also play a crucial role in how a tradition defines itself. A good contemporaneous example for this is the International Society for Krishna Consciousness (ISKCON), which is also known as the “Hare Krishna movement,” after the mantra to which they attach supreme 79

*UVC * I *s.v. *“abdaivatam”; cf. also n. 237 on p. 62 below.

Note, however, that this only became a proper name in the course of time; originally, the word could also refer to other verses; see Chapter 2.

For this mantra, see especially Einoo 2005b; see also Wilke & Moebus 2011: 723–729. Its application is in consonance with its semantic content: “We worship Tryambaka, the fragrant increaser of prosperity. Like a cucumber from the stalk, may I come loose from death, not from deathlessness!” tríyàmbakaṃ yajāmahe, *sugándhim puṣṭivárdhanam */

*urvārukám *( i) va bándhanān, *mṛtyór mukṣīya mā́mṛ́tāt *// (originally, it must have been urvārukáṃ va, which was later normalized to urvārukám iva); cf. the translation by Jamison & Brereton 2014: 954.

Cf. Rastelli 2006: 205.

Cf. Rocher 1989: 180. As it is frequently preceded by om, the mantra is also called “Six-Syllable” or “Ṣaḍakṣara-Mantra”; see Sanderson 2013a: 88; cf. Bisschop 2018b.

20 ∙ gāyatRī: mantRa and motheR of the vedas significance, the “Hare Krishna mantra” or “Mahāmantra.”⁸⁴ The GM, likewise, has always strongly been associated with Brahminical Hindu traditions, and in particular with the Vedic tradition.

The status of individual mantras varied over time and across traditions, and not every mantra that was once deemed worthy of a name continued to enjoy pre-eminence. Conversely, not every significant mantra received its own name.

What the practice of giving names to mantras very well illustrates, however, is that mantras as a rule do not form a homogeneous, anonymous mass. Rather, a select few mantras are recited much more often than all the others or are praised as especially powerful or sacred. As I will argue throughout this study, this status tended to perpetuate or even increase itself: once a mantra was reused more often and associated with certain functions and purposes, it also was more likely to be selected for the same or similar purposes in other ritual contexts, and indeed, beyond them too.⁸⁵

2.4 Deification of Tantric mantras

So far, I have almost exclusively dealt with Vedic mantras. While the scope of this study does not allow me to elaborate on Tantric mantras in the same way, there is a specific aspect that pertains directly to the subject of this study, and therefore must not be overlooked: the deification of mantras, something that is known to be a particularity of Tantric traditions.⁸⁶ In contrast to the Vedic case, mantras are here generally considered to be deities themselves. Accompanying their recitation, they are often visualized, usually in an anthropomorphic form.

Frequently the individual body parts of a mantra deity have a mantric manifestation as well. The so-called aṅgamantra s or “limb mantras,” in particular, form a predefined set of elements associated with a deity or mantra.⁸⁷

The intermediate goal of the repetitive recitation of a Tantric mantra is to manipulate or even “master” ( sādh, siddhi) it, that is, to gain control of the deity it represents, or rather, the deity that it is.⁸⁸ This is above all done by repeated recitation (generally known as japa), combined with visualization and offerings.

The ultimate goal of the practice is to achieve and obtain whatever one desires.

The texts of the mantras used to this end are, as a rule, addressed to the deity that is being invoked. Let me offer one example to illustrate this. In the Pañcarātra 84

hare kṛṣṇa hare kṛṣṇa, kṛṣṇa kṛṣṇa hare hare, hare rāma hare rāma, rāma rāma hare hare; see Neubert 2018.

See especially Chapter 4.

For an overview, see Timalsina 2018; see also Brunner 2001.

See below p. 250.

Interestingly, the gāyatrisiddhi, the “mastery of the Gāyatrī,” is already mentioned in one of the oldest extant Śaiva Tantras (c. seventh century ce), the Niśvāsamukha (NiśvMukh IV 13; see Kafle 2020: 338).

intRoduction ∙ 21

tradition, the Jayā-Mantra is used to address the goddess Jayā, one of the four śakti s or “powers” of Vaikuṇṭha. The mantra can be translated as: “oṃ jāṃ –

obeisance to Jayā! To the one who is fixed in the abode of the Unconquered One [i.e., of Vaikuṇṭha] – jāṃ jRīṃ svāhā.”⁸⁹ When the Jayā-Mantra has been visualized and finally mastered, it appears to the reciter in anthropomorphic form,⁹⁰ saying: “You have mastered me well, Son. Free from fear and affliction perform the action that is desired with my mantra.”⁹¹ This account is exemplary of other practices in which Tantric mantras are used: the mantra, an audible entity, is mastered and subsequently appears – or is expected to appear – in a visible form.

The relationship between mantras in their sonic forms and their divine aspects is explained in various ways. On the one hand, mantras are frequently said to be that which expresses or signifies the deity (i.e., mantras are the vācaka s of their vācya s, the deities),⁹² or they are thought of as manifestations or the powers ( śakti s) of deities. On the other hand, they are also imagined as consisting of or even being identical with the mantra’s sound, in which case the mantras really *are * deities themselves.⁹³ However, if a mantra has a name, this name is not necessarily that of its divine aspect. The Dvādaśākṣara-Mantra mentioned above, for instance, is not the manifestation of a god “Dvādaśākṣara,” but of Vāsudeva. In contrast, the GM can be understood as the sonic manifestation of the goddess Gāyatrī or Sāvitrī, but not of Savitṛ, who is actually mentioned in the text of the mantra.

How exactly the deification of mantras began to hold sway has barely been explored. As the GM was one of the first mantras (possibly even the first) that were consistently thought of as being a deity, studying its history promises to contribute to our understanding of the process of mantra deification. It should be emphasized, however, that the GM is not a Tantric mantra, but a Vedic one.

It would be ill-advised to view its deification as a mere foreshadowing or prefiguration of much later developments, nor would it be permissible to study it with the help of Tantric categories. Deification in the Vedic and Early Hindu periods must be understood in their own frameworks. Before this can be done, however, it is necessary to offer a few preliminary considerations on the nature of gods, and on the implications of what it means to become one.

*oṃ *⁺ jāṃ jayāyai namaḥ, *ajitadhāmāvasthitāyai *⁺ *jāṃ *⁺ *jrīṃ svāhā *(Krishnamacharya 1967: 31 [no. 4]); see Rastelli 1999: 130, n. 496.

She is said to look like Lakṣmī; cf. Rastelli 2000: 360, where the following attributes are listed: “no definite colour, white garments, beautiful, various adornments, with a crescent on the forehead, two arms, holding a noose and a hook J[ay]S 6.85–91b.”

JayS 27.119: *susiddhāsmi ca te putra manmantreṇa samācara */ yad abhīṣṭaṃ tu vai kāryaṃ

*niśśaṅko vigatajvaraḥ *//; tr. Rastelli 2000: 335.

Cf. Rastelli 1999: 120.

Padoux 2001: 398–399.

22 ∙ gāyatRī: mantRa and motheR of the vedas 3. Deification

3.1 The nature of deities

Across the world’s religions, countless deities are worshipped. Some of them, such as Yahweh, have continued to be the object of veneration for thousands of years, and some have even procreated. In other cases, two or more of them have merged into a single deity, as was the case with Skanda and Murukaṉ in South India. Some, like Kronos, no longer attract much attention. Occasionally, gods and goddesses are also revived or even invented anew, as is the case with Odin in Neo-Paganism, or with Bhārat Mātā, the personification of the Indian nation.⁹⁴ Historically, pantheons around the globe have been in a constant flux.⁹⁵

The question is: how do gods emerge, and how do they develop? Do they come into existence by themselves, or are they “invented” – or perhaps, both?

From the perspective of the human sciences, a key condition for the

“production of divinity,” as Gustavo Benavides has called it, is the ability to recognize other beings of whatever sort as conscious and intentional agents.⁹⁶

Another factor is the general disposition in humans to understand reality in a

“teleological manner,” that is, to perceive it as consisting of entities that have purposes. These qualities may be attributed to all kinds of entities, be they human or non-human, animate or inanimate, objects or phenomena, real or imagined. As a result, all kinds of agents may be suspected behind reality: gods as well as spirits, ghosts, demons, and so on. In the case of gods in the classical sense of the word, another disposition can be observed: the tendency to think of them as more or less human-like beings, or persons, in the widest sense.⁹⁷

Indeed, most of them are imagined to be endowed with bodies, minds, desires, thoughts, etc.; in many cases, they are anthropomorphic, which includes having a gender.⁹⁸ Many of them even demand offerings and sacrifices for their sustenance, or at least for their benevolence.

On the other hand, gods are also generally “believed to be largely, but not fully, free from the physical, psychological, and moral constraints that limit the agency of humans.”⁹⁹ Being clairvoyant, omniscient, unageing, immortal, etc., they are in many respects superior to ordinary humans, and in many ways super- human. By way of their divinity, they often belong to, or partake in, an-94

For Kronos, see Baudy 2006. For Murukaṉ, see Clothey 2018. For the revival of Norse paganism (e.g., Ásatrú and Odinism), see Schnurbein 2016. For Bhārat Mātā, see McKean 1996.

For an overview of Hindu deities, see Narayan 2018.

See Benavides 2016: 564–566.

Guthrie (1993) even argued that anthropomorphism is the key feature not only of gods in particular, but of religion as a whole.

This was the case, for instance, in the PIE pantheon; see West 2007: 138.

Benavides 2016: 561.

intRoduction ∙ 23

other, “transcendent” domain. Gods are thus simultaneously – and often also contradictorily – defined both by their strong similarities to humans and by their otherness from human beings, generally in terms of superiority.¹⁰⁰

This, however, does not mean that gods are necessarily non-human or supernatural, nor are they always only products of imagination in the way that literary characters are. From the historical perspective, it is clear that many of them have their origin in very real beings and phenomena that then become subject to deification. This term, together with its near-synonyms “divinization” and “apotheosis,” usually refers to the elevation of heroes, emperors, and eminent religious persons.¹⁰¹ Famous examples of this kind of deification are found in ancient Egypt, where Pharaohs were considered living gods, or ancient Rome, where emperors were often posthumously deified by their successors.¹⁰²

Cases of deification of the dead are also known to come from South Asia, both ancient and modern.¹⁰³

In most polytheistic religions, however, the range of subjects or objects susceptible to deification is even wider: virtually anything can become a deity. Besides humans, natural and especially celestial objects and phenomena have been particularly prone to deification. The Vedic god Agni, for instance, is generally called “the god *of * fire” – while in fact his name itself means “fire”

(accordingly, it is sometimes also rendered as “Fire”).¹⁰⁴ The deification of the sun in many religions is well known; the Egyptian god Ra, whose name also means “sun,” is a prime example.¹⁰⁵

Deification may also affect much more abstract entities. In the ancient Greek and Roman religions, for instance, the seasons, the hours of the day, and even virtues and principles such as Justice (Dike/Justitia) or Fortune (Tyche/Fortuna) were also deified, many of them having their own cult¹⁰⁶ – even including diseases and negative powers such as Fever, Mildew, and Misfortune, all of which were deified and worshipped.¹⁰⁷ Other entities were personified in a similar manner, but were not necessarily worshipped or considered deities: in 100 Cf. also Barrett 1998.

101 Cf. *ER * IV: 259–262 and Benavides 2016: 570–571. For the divinization of kings in particular, see Strathern 2019: 155–218.

102 See, for instance, Lipka 2009: 129–132.

103 See, for instance, the various contributions in Dimitrova & Oranskaia 2020; see also Blackburn 1985.

104 Words such as *agni * can thus function as common nouns and theonyms at the same time; cf. Elizarenkova 1995: 105.

105 See, for instance, Kahl 2007.

106 For a general overview of personification in ancient Greek and Roman religion, see Bendlin & Shapiro 2006. For the deification of impersonal notions in Roman religion, see Lind 1974, Feeney 1998: 85–92, Stafford 1998: 59–65, and Lipka 2009: 127–129; for personification (and deification) in the Greek world, see the collected articles in Stafford

& Herrin 2017.

107 Febris, Robigus, and Mala Fortuna; see Bendlin & Shapiro 2006.

24 ∙ gāyatRī: mantRa and motheR of the vedas ancient Rome, for example, Death (Mors) was personified, but not considered a god; his Greek counterpart Thanatos, on the other hand, was considered a deity (albeit one that “had practically no cultic significance”¹⁰⁸).

3.2 Personification

The above examples raise the question as to how real entities (such as fire and death) and their deifications (Fire and Death) – or personifications, as they are often called – relate to each other. Let me begin with the terms “person” and

“personification,” which in the descriptions of polytheistic religions are usually meant to imply a “human-like being.”¹⁰⁹ Superficially, using these terms seems justifiable in most cases: While deities such as Viṣṇu, Athena, or Odin have the ability to act in superhuman ways, they are generally anthropomorphic (or even

“super-anthropomorphic” – the many hands and heads of Hindu gods can be interpreted as a manifestations of their superhuman abilities). They could very well be called “divine persons.” Deities such as Agni and Fortuna, being divine personifications, inevitably take on human features too (see below).

A crucial characteristic of real human persons, however, is that they are tied to their corporeal, carnal existence, that is, to their spatially and temporally limited bodies. Human individuals can be distinguished from each other first and foremost by their corporeality. While the body is obviously liable to change and modification, seen from a pragmatic perspective, it remains the anchor point in establishing a person’s identity, that is, their sameness-with-themselves through time. Since gods generally do not have such a locus, individuality and self-identity are often much less pronounced than in the case of humans.

Aspects of divine individuality and identity may significantly vary or change. In ancient Greece, for instance, different manifestations of Zeus were worshipped at different places (e.g., Zeus Aetnaeus on Mount Aetna or Zeus Lykaios on Mount Lykaion). The same was the case with his Roman equivalent, Jupiter,¹¹⁰ with whom he was essentially considered to be identical (an impossible relationship among human individuals). Another deity, Yahweh, in the course of time developed from a local Semitic storm-god into the almighty god of Judaism, Christianity, and Islam, keeping his name (as YHWH, Jahweh, or Jehovah) in the former two religions. Whether he should be considered to be one and the same in all three monotheistic religions – let alone be identified with the ancient god Yahweh – is debatable, however.

108 Kunz 2006; cf. Bendlin & Shapiro 2006. For the deification(s) of death in the ancient Greek religion, see also Burton 2017.

109 For a different approach to the concept of personhood and especially personality in the context of gods, see Hick 2004: 264–266.

110 See Lipka 2009: 132–133.

intRoduction ∙ 25

The individuality of gods, their self-identity, and the bonds to their manifestations or bodies are subject to change. This is the reason why the terms

“person” and “personification” can and indeed have been used only somewhat loosely. In most studies concerned with deification, these terms appear to denote any being that, in its intentionality, agency, and abilities, and often, but not necessarily also in its form, bears some resemblance to a human being.¹¹¹ In this definition, what is deducted from the concept of being human is primarily its physical or corporeal aspect.

If we accept this usage then,¹¹² a number of questions arises. When characterizing the god Agni as the personification of fire, we have to ask to what extent physical fire is actually being personified, that is, made or thought of as a person. Is the fire its body, or is “Fire” only the name of the deity governing fire? Similarly, we could ask about the relationship between luck as we experience it and its personification as the goddess Fortuna. If this goddess only works through luck, to what extent can she herself “be” luck? Turning to the subject of this study, we might ask: To what extent is the goddess Sāvitrī the audible sounds of her mantra? Are they her manifestation, or even her body? If so, how can this be reconciled with her visual manifestation?

When looking at the host of deifications – or “personification deities,” as we might also call them – it soon becomes clear that answers to these questions vary from case to case. The personal “constitution” of a deified entity may be determined by a number of factors. In cases where the deification is based on an entity that is perceptible to the senses (especially a visible one), this entity is indeed often described as the body of the deity. Agni’s flames, for instance, are frequently described as his body parts.¹¹³ If a language has a gender system, the gender of the deity is usually determined by the grammatical gender of its name: Fortuna is female, Mṛtyu is male, and so on. This often results in a deity being conceived of as a woman or a man, especially in cases of personifications of abstractions. Also, the features of such deities are often determined by their 111 Cf., for instance, Stafford’s (1998: 25–26) definition of “personification” as “an anthropomorphic representation of any non-human thing.”

112 One alternative would be to introduce other terminology, such as “agent” and “agenti-zation.” “Person” and “personification” are preferable because, as explained above, the frame of reference for the definition of deities most often is the human being, even in the case of theriomorphic deities (who, such as Garuḍa, have a name, quite human emotions, use language, etc.). It should however be noted that my definition of “person” does not automatically entail anthropomorphism, which consequently also affects my definition not only of “personification,” but also of “deification.” Cf. Strathern (2019), who uses the term “personhood” (pp. 29–30) “to refer simply to motivated agency,” while also noting that “emic theories of what exactly personhood is and how exactly matter relates to non-matter, may vary enormously and are often composite.”

113 See Feller 2018. Rochberg (2011) discusses in detail how gods in ancient Mesopotamia may be either thought of as manifesting by means of celestial bodies, or as being the celestial bodies themselves; cf. also Pongratz-Leisten 2011b.

26 ∙ gāyatRī: mantRa and motheR of the vedas nature or function: Dike, the goddess (of) Justice, for instance, is depicted as a woman carrying a scale.

In several cases, gods originally came into being by way of personification, but then developed in such a way as to no longer be recognizable as such.

This was probably the case with Zeus, whose predecessor in PIE religion was the personification of the bright diurnal sky – that is, he was the sky (* di ̯éu̯) itself.¹¹⁴ Zeus continued to be associated with the sky. Over time, however, he became much more than a mere personification deity, as he was much more often envisaged and worshipped as the lightning-flinging ruler *of * the sky than as the sky itself. On the other hand, gods may also acquire new – personified –

aspects. The Vedic god Savitṛ, for instance, first appears to have been a more or less independent, anthropomorphic, functional or agent god.¹¹⁵ One of his many activities included impelling the sun to rise. Probably due to this function, he later came to be almost completely identified with the sun – that is, he became its personification.

It is thus often difficult to sharply distinguish between personification deities and “normal,” independent deities. The relationship of both types of deities to their respective effects, manifestations, or embodiments (e.g., visible phenomena or special occurrences, ranging from events such as equinoxes up to

“wonders” of all sorts) is often equally difficult to determine.¹¹⁶ Transitions and overlaps are the rule rather than the exception, and essentialism is best avoided (see Figure 3).

ordinary aspect

deity aspect

**“substrate” **

+ divinity

deification

e.g., the sky

e.g., the PIE divine sky

+ personhood

effects, manifestations

divine

independent deity

e.g., thunder

agency

e.g., Zeus

**Figure 3: **Deifications and independent deities

114 Cf. Jackson 2002: 71–73 and West 2007: 166–173. For personification deities in PIE religion, see also West 2007: 135–136. Witzel (2004a: 586) rightly emphasizes that even

“deities of nature,” such as the divine sky, “had acquired, in PIE and even in pre-PIE times, their own ‘personal’ biographies, as seen in a number of more or less inter-connected, common IE myths.”

115 See Haas 2020. Being an agent god, the early-Vedic Savitṛ can be characterized as a deification of a certain force, namely his “impulse” ( prasavá), which is frequently mentioned.

116 For the different ways in which deities make themselves present or become embodied in the world, see Descola 2013; cf. also Werner 1982.

intRoduction ∙ 27

By definition – theoretical as it may be – personification deities are strongly tied to their “substrates” (i.e., that which is being personified), which as a rule are firmly rooted in everyday human experience. (In contrast, a divine person such as Athena may be the goddess *of * wisdom – somewhat impre-cisely, she has even been called the *personification * of wisdom – but this does not necessarily delimit her being.) There is always a component whose existence is beyond doubt, even in the case of abstractions. At the same time, there also is “something extra” – the divine person – that ultimately exists only on the mental plane.¹¹⁷ However, insofar as this divine person often stands out clearly from and transcends what has been personified, the question arises as to what extent it is perceived as being real or fictitious. Is the being behind a personification, even if divine, always felt to be as real as its substrate?

Jon Whitman proposes that two general kinds of personification need to be distinguished, “fictional personification” and “real personification.”¹¹⁸ Fictional personification – also called “prosopopeia” – is a rhetorical, literary device. This kind of personification may be limited to a single sentence (“the sun smiles”), or have the form of a “personification characterization” (as James J. Paxson termed it), if the trope is employed for “the narrational invention of actual characters, objects, or places that occupy the material space-time of the fabular, or ‘story’

level of a narrative text.”¹¹⁹ This is the case, for instance, in the Late Antique Latin allegory Psychomachia, where vices and virtues are personified and battle each other; or in Kṛṣṇamiśra’s Prabodhacandrodaya from the eleventh century ce,¹²⁰ a Sanskrit play in which many types of entities (including not only virtues, but also texts like the Bhagavad-Gītā [BhagG]) are personified.

In both cases of fictional personification, the speaker or author and the recipients (readers, viewers, etc.) are aware that the personification is just a product of fantasy. Real personification, on the other hand, according to Whitman refers to the practice of giving an *actual * personality to an abstraction. This practice has its origins in animism and ancient religion, and is called “personification” by modern theorists of religion and anthropology.¹²¹

This type of personification would be given in the cases of deities such as Agni or Fortuna mentioned above.

In the case of religion, however, one wonders where “reality” ends and

“fiction” begins,¹²² and whether the theoretical distinction between real and fic-117 Depending on the perspective, this may also be called the transcendental or divine plane.

118 Whitman 1987: 271–272; cf. Paxson 2009: 6. For the history of the term “personification,”

see Whitman 1987: 269.

119 Paxson 2009: 35.

120 See Śāstrī 1936 and Kapstein 2009.

121 Whitman 1987: 271.

122 The phenomenon that fictional stories can contribute to and, in fact, inspire real religious beliefs has been mostly studied in the context of contemporary religions; see, for

28 ∙ gāyatRī: mantRa and motheR of the vedas tional personification is always valid. For most people, gods do not become known by way of direct experience. Rather, knowledge about their character and appearance is often transmitted by way of verbal testimony in stories, myths, religious texts or scriptures, and the like. Religious and literary texts, however, are clearly not two categories opposing each other in the sense that one provides truthful (religious) information, while the other is only meant to entertain and not to be taken seriously. It hardly needs to be argued that works of fiction, such as the Iliad or the MBh, have often served – or have even been explicitly proclaimed – as sources for religious imagination and belief.¹²³

In the ancient Greek and Roman religions, it is indeed the case that “[t]he boundary between literary personification and those personifications which were the objects of cultic adoration is indistinct and permeable.”¹²⁴ It is even possible to collect entire lists of such “in-between deities.”¹²⁵ Hinduism, too, is known for a host of personifications, not all of which are consistently considered real, divine persons.¹²⁶ (I will return to the subject of deification in Hinduism in the next section.) A deity, be it a personification or a “normal” deity, can have a life in the realm of “fictional” literature and, at the same time, be perceived as a real being and be worshipped. The two modes of existence may easily influence each other – in fact, it is most often impossible to disentangle them in the first place, especially in hindsight. Due to their nature, personification deities in particular may in many cases be called “fringe deities”: their status as deities can increase or decrease depending on a variety of circumstances.¹²⁷

3.3 The deification process

Continuing from and building on the observations made above, I propose to understand the production of deifications as a complex, entangled interplay of intra- and interpersonal creation, appropriation, and modification of information and ideas. While deification may start as a “theophany,” or a one-off vision of the divine (e.g., in a dream), there are usually many preconditions and circumstances that help to form the idea of a deity, and also contribute to filling

→

instance, the contributions in Davidsen 2016, especially Petersen 2016. At this point mention should also be made of Max Müller’s outdated theory that mythology originates from a “disease of language,” in which personifications originally meant metaphorically were later misunderstood as real; see Yelle 2013: 50–55.

123 For the role of (South Asian) fictional literature in conveying information about, or valid in, the real world, see Ferstl 2020: 19–34.

124 Bendlin & Shapiro 2006; cf. Paxson 2009: 7: “the distinctions among gods, ghosts, genii, fantastic creatures, and personification characters were not always clear even to the principal theorists of classical, medieval, or Renaissance rhetoric.”

125 At least in the case of the ancient Roman religion; see, for instance, Lind 1974: 111.

126 See Michaels 2006: 225–226.

127 Cf. Lipka 2009: 127: “At any rate, all deifications in Rome were in fact partial, with a (new) divine aspect added to a (hitherto) profane notion, rather than replacing it.”

intRoduction ∙ 29

this idea with content. Throughout an individual’s life, a plurality of sources, factual as well as fictional, may contribute to the production of a deification.

If, for instance, a person has the conviction that a certain mountain is divine and begins to envision its deification, information about its name, its identity, its powers, and so on is often gathered or derived from an external source (another person, a text, an anecdote, etc.). Similarly, if someone comes to believe that death is a personal entity, their imagination might very easily be influenced by the way death is personified in literature, or art in general.¹²⁸ The existence or prevalence of deification in a particular culture is another important factor.

Simply put, in most cases personification deities do not arise out of thin air.

On the other hand, when it comes to religion, individual creativity plays a much greater role than is often assumed. Arguing against a somewhat mechanical, but quite widespread understanding of religion, D. Jason Slone emphasized that people do not simply believe and reproduce what they have learned as children, but “are continuously engaged in the construction of novel thoughts and in the transformation of culturally transmitted ideas.”¹²⁹ Thus, what people actually think, believe, and do may differ strongly from the ideas or ideals prevalent in their own culture. This variability may consequently also affect the conviction that a certain entity is (or is not) a deity, especially in religions where there is no centralized power that determines what is a correct and what is a wrong belief. Deification, therefore, is not always simply decreed by certain authorities, such as priests or prophets, and then acknowledged by those who accept those authorities.¹³⁰ The origin of a deification is rather often to be located in a non-isolated, but nevertheless largely intrapersonal, creative process.

The further development and, most importantly, *establishment * of a deification, on the other hand, has to be understood as an interpersonal process. The idea that a particular entity is a deity is, as mentioned above, often retrieved from an external source. In the interpretation of such a source – be it a human being or a text – “misunderstandings,” or rather, different understandings, may play an important role. One may imagine, for example, one person referring to a certain entity (such as fate) as a deity in a purely rhetorical manner, while another person might understand such a statement quite literally.¹³¹

128 Perceiving death as a personal entity has indeed always been common, with the Grim Reaper being a quite popular personification in modern times; see, for instance, Tamm 1996.

129 Slone 2004: 121.

130 In the Roman religion, the divinity of certain abstractions was evidently questioned by some, or was even joked about. For the analysis of deified abstractions in Roman religion by the ancients themselves, see Stafford 1998: 56–59; see also Feeney 1998: 87–88. I am not aware that any such survey of the opinions of ancient South Asian thinkers regarding the status of personifications and deifications (be they abstractions or other) exists.

131 The proverb audentes fortuna iuvat, “Fortune favors the bold,” has been repeatedly used as a slogan or motto, for instance in the United Statess military. But while a modern soldier

30 ∙ gāyatRī: mantRa and motheR of the vedas Such processes are naturally difficult to document and reconstruct. What we can observe are only snapshots of what is in in reality a highly complicated and not necessarily linear or continuous process. However, when analyzing deification as a historical and cultural process, we can forage for traces in the products of authors, artists, or teachers, and try to determine how they may have been understood by their recipients, their audience, students, etc. Indeed, in the case of South Asian religious history, the abundance of textual witnesses and especially too of reworkings over several centuries offers a comparatively good basis for such an investigation.

3.4 Methodological considerations

In this final section, I offer some thoughts on how the observations made above can be applied in practice, with a particular focus on examples from ancient South Asia. In order to be able to trace deification processes on the basis of the sources, I believe it makes sense to orientate oneself toward the result.

Deities have been defined here as superhuman persons, that is, divine human-like agents. If this definition is accepted, deification can be analyzed as a combination of two processes: one that involves divinization, and another involving personification.

I define divinization as the attribution of superhuman or “supernatural,” divine qualities, such as (among others) indestructibility, the power to heal instan-taneously and to bestow blessings of all kinds. Divinity can thus also be ascribed to the manifestations or effects of a deity (such as thunder, or an earthquake), and does not necessarily entail personification.¹³²

Following Whitman and Paxson, personification can be divided into metaphorical personification and “personification characterization.” In both cases, personhood is attributed to a non-personal entity. In contrast to the as-cription of divinity, this quality is frequently fictitious, that is, personification is often only an artistic device. Regardless of whether it occurs in fictional or factual contexts, however, it may very well become real, both in the story-world and in the real world. Literary and actual personification should therefore be understood as non-exclusive, and in fact often interacting categories.

→

will most likely only understand it as an exhortation to be bold, a soldier in ancient Rome (whom we, for the sake of the argument, might imagine standing in front of Fortuna’s temple near the Tiber), might have taken it quite literally. Stafford (1998: 28–33) lists many examples of metaphorical personifications bordering on deification.

132 The divinization of human beings (especially while still alive) is a case sui generis. Clearly, not every person who is attributed superhuman powers is also considered a deity – limited, as they are, by their physical bodies and other earthly constraints. However, there are many shades between “divine” and “non-divine,” and at least in ancient South Asia, some persons were considered (or considered themselves) “divine enough” so as to be called *deva * or devī; cf. my remarks below pp. 32–33.

intRoduction ∙ 31

Thinking about divinization and personification as two aspects of deification is not as theoretical as it may sound. In the following, I will give examples of divinization and personification from the context of ancient and medieval South Asia. The examples will illustrate that divinity and personhood are not always equally attributed. What they will also show is that the boundary between reality and fiction can be quite permeable, and can shift in the course of time. Moreover, they will demonstrate that there are many shades, on the one hand, between the divine and the non-divine and, on the other hand, between personal beings and inanimate objects or notions.¹³³

Most personifications that were evidently believed to be real may be easily characterized as deifications, but not all of them partook in divinity to the same degree.¹³⁴ While trees and other plants have been regularly worshipped and asked for blessings in rituals up to the present day,¹³⁵ their categorization as animate beings has not always (and not by everybody) been taken for granted.¹³⁶

As such, they did not belong to same category of divine persons as Indra or Śiva.

The same can be observed of purely fictional personifications. Kālidāsa’s famous Meghadūta, for instance, is about a personified cloud who is asked by a *yakṣa * to convey a message to his wife. In Hinduism, individual clouds are not usually recognized as persons, let alone worshipped as deities.¹³⁷ The personification of the cloud messenger is thus purely fictional, its personhood fictitious.

Just as personification does not automatically lead to divinization, divinity can also be attributed to non-personal entities.¹³⁸ In several Gṛhya rituals, for 133 Perhaps the most prominent example for this in the context of Hinduism is that of Brahmā vs. brahman; cf. Bailey 1983: 6: “Making clear distinction [sic] between personal and impersonal in relation to a concept like *brahma * was not the Indian way of doing things. It is more accurate to say that in some ‘idealistic’ circles *brahma * was completely divested of any personal attributes; whereas in other circles (represented in many passages of the oldest Upaniṣads) personal portrayals of *brahma * were interwoven with impersonal ones.”

134 For personifications/deifications in the ṚV, see Elizarenkova 1995: 83–105.

135 See Smith 2018.

136 See, for instance, Hara 2003 and Ferrari & Dähnhardt 2016.

137 The cloud’s non-divine status in the poem may be subject to debate. After all, it has the superhuman power to fly, and to change its shape. However, the cloud itself is asked to worship Śiva on its way and does so in a very human way; cf. Feller 2012: 321: “Of course, due to its nature, the cloud can (or is at least asked to) perform certain extraordinary deeds of worship which are inaccessible to the common mortals (like turning into a cloud of flowers or a flight of steps). On the other hand, playing the drum, bathing, bowing, showering with flowers, circumambulating, are ordinary acts of worship, and we see that the cloud is actually performing a complete *pūjā * of Śiva and his family on the way to its destination.” For a similar personification of a cloud in a MBh story (XII 263), see Bedekar 1960.

138 Divinity may, of course, also be ascribed to fictitious entities. In a story of the Pañca-Tantra (PañcT VIII; tr. Ryder 1925: 89–104) a carpenter constructs a mechanical bird made of wood, similar to Garuḍa. Disguised as Viṣṇu, the carpenter’s friend uses this machine to seduce a princess. In the course of the story, the real Garuḍa causes the apparatus to actually fly, which thus becomes divine.

32 ∙ gāyatRī: mantRa and motheR of the vedas instance, a great variety of entities – such as insight, intelligence, meters, and melodies – are offered oblations, that is, they are “satisfied” ( tṛp). In these rituals, powerful and divine, non-personal entities may stand side by side with proper gods such as Indra. While they are worshipped and to some extent considered divine, they generally do not appear as full-fledged persons. They are abstract divinities rather than actual gods.

Nevertheless, from a historical perspective we observe that divinization indeed often has entailed personification. A good example is given by Puṣpaka, Rāma’s divine chariot. While generally described as a magnificent vehicle in the earlier strata of the Rāmāyaṇa (Rām.), in the comparatively late Uttarakāṇḍa, Puṣpaka also becomes an animate being, having the ability to speak, to bow down, and so on.¹³⁹ Similarly, nowadays the Vedas themselves even have dedicated temples (most importantly the Ved Mandir in Nashik, Maharashtra), where they (taken together as “the Veda”) are worshipped as an anthropomorphic *bhagvān ved */ bhagavān vedaḥ, or “Lord Veda.”¹⁴⁰

As we can see, divinity and personhood are often, but not always, intertwined. If an entity is both divinized and personified – say, in a story or a myth (written or told), an inscription, a ritual manual – it is appropriate to speak of a personification deity, that is, a deification. But how can we know if this deification was really believed to exist, or was perhaps even worshipped? Of course, if there is an actual temple or shrine dedicated to a certain being who is regularly worshipped there in the form of an effigy (perhaps even as an anthropomorphic one), then there can be little doubt that this being is understood as an actual deity by a certain community. If, however, the only evidence to work with are some passages from ancient texts, then determining the reality or factuality of deification becomes much more difficult.¹⁴¹

First of all, it should be noted that the terms defined above – personification, deification, and so forth – have no (or no exact) equivalents in Vedic or Sanskrit. Moreover, the presence of words such as *devī *“goddess,” *deva *“god,”

*devatā *“deity/divinity,” or *daivya *“divine” does not allow the conclusion that the entities qualified by them actually were considered *fully * identical with deities.

They may, for instance, simply be used for the purpose of (self-)praise, as in the case of kings or Brahmins.¹⁴² For this reason, individual text passages have to 139 See Feller 2020.

140 See Larios 2011 and 2021; cf. also below pp. 236–237.

141 Cf. Gombrich’s (1971: 255) enlightening illustration: “Don’t we all pretend every year that Christmas personified will arrive in a sledge pulled by reindeers, or at least that our children believe that he will? What will a future historian make of our beliefs when he examines our Christmas cards? And what, to return more closely to the core of our problem, would be the conclusions of a visitor from Mars whose data are confined to our language habits but who could not know where figures of speech end and figures of thought begin?.”

142 For the origins of this practice, see Falk 1994: 313–324.

intRoduction ∙ 33

be assessed on a case-by-case basis. The theoretical considerations made above can only be applied after, and not before, a close reading and analysis of the texts.

Second, the nature of the sources hardly allows for straightforward conclusions. One may be tempted to assume that, for instance, the authors of ritual manuals presupposed that deities exist. While this was certainly the view of most of them, it would not necessarily have been shared, for instance, by the proponents of the Mīmāṃsā, the most important intellectual tradition dedicated to the interpretation of rituals and the ritual texts. For the Mīmāmṣakas, the efficacy of Vedic rituals did not depend on the gods. Gods were only secondary –

in fact, it was even thought that “it ultimately does not matter if the deities exist apart from the Veda, their ‘linguistic reality’.”¹⁴³ While this may be an extrem-ist view, it very well illustrates that we should not rashly accuse the ancient South Asians of being credulous: even those who were entirely dedicated to the worship of gods did not necessarily take their existence for granted.

On the other hand, stories that would nowadays be most often categorized as fiction were (and are), in some cases, considered authoritative in religious matters. The MBh and the Rām. in particular have risen to this status.¹⁴⁴ Danielle Feller observes that

the Epics never make explicit truth-claims with regard to the tales they narrate (as far as this can be affirmed with any certainty of such voluminous texts). Not, however, because these were assumed to be fictitious, but rather because everything narrated in the Epics is implicitly true. The idea that any narrative could be fictitious never occurs at all.¹⁴⁵

At the same time, it is more than doubtful that the recipients of these texts –

as well as their composers – would have believed *everything * the Sanskrit Epics contain to be truthful statements about reality. Nor would they have understood all the rituals described therein as actually feasible and worthy of imitation.

Rather, we have to assume that beliefs and convictions are more often backed up by, rather than mechanically derived from, the texts.

As we can see, ascertaining the perceived fictitiousness or reality of deities on the basis Vedic and Sanskrit literature is a difficult task. What one frequently ends up with is a *range of possibilities * rather than a definitive conclusion about the ontological status attributed to an entity. It is therefore vital to heed Emma Stafford’s advice: “Rather than attempting to define a figure’s exact status on an imaginary scale, it might be more helpful to take a broad overview of her/his 143 Clooney 1988: 284.

144 Cf. below p. 202.

145 Feller 2004: 34–35.

34 ∙ gāyatRī: mantRa and motheR of the vedas incarnations in various media to gain a general indication of more or less widespread recognition.”¹⁴⁶

But even if there are cases where we can be quite certain that the divinity of an entity was established and widely recognized, further difficulties arise from the historical dimension. Writing about the development of a single deity over the millennia always involves major risks. Stephanie W. Jamison succinctly summarized these in her review of Ludvik’s book (2007) on Sarasvatī (a figure who was widely recognized as a goddess and was indeed very close to the deified GM):

There is, on the one hand, an understandable tendency to select and overemphasize aspects in one stratum that can be related, one way or the other, to material in another stratum, and in particular to “read back” into earlier texts what is going to be prominent in later ones. On the other hand, given the concentrated focus on a single issue (in this case, a single divinity) in a text that treats manifold matters, there is the danger of interpreting things found associated with that divinity as significantly unique to her, rather than being broadly characteristic of divinities in general.¹⁴⁷

Heeding these warnings, we should not assume any unbroken “biographical”

(or “theographical”) coherence for a particular deity. Especially in the case of

“fringe deities” – such as the Gāyatrī/Sāvitrī – it is better to assume that their identity was reconstituted to a not inconsiderable extent in the various historical phases. Moreover, when adopting a historical perspective, it is imperative to take countermeasures against the natural tendency to regard everything that has the same name as somehow belonging together (even to the point of being identical). Accordingly, it may be useful to pay more attention to differences than to similarities, perhaps even if one thereby runs the risk of undermining the (perceived) coherence of historical developments.

4. Content outline

In presenting the results of my research, I have generally tried to outline specific historical developments rather than to narrate a continuous, linear “biography.”

Simply dealing with all text passages one after the other in a chronological order would have led to an essentially unreadable study. Instead, I have chosen to bundle them in thematically concentrated chapters, each devoted to a certain aspect, or sometimes process or phase, in the development of the GM and its 146 Stafford 1998: 28.

147 Jamison 2009: 354.

intRoduction ∙ 35

deification (see Figure 4 below). However, since many passages are relevant to several topics, I have sometimes had to go beyond the respective topic. Occasionally, therefore, I have chosen a more text-oriented approach.

In accordance with its aims, this study is divided into two parts. In each part, there are two main chapters dedicated to analyzing historical processes, while the final chapters focus more on the results of those processes.

**Part I **(“The Mantra”) is dedicated to the early history of the mantra. It follows this history through the first millennium bce up to around the third century ce. Notwithstanding certain variations, after the third century ce the ritual uses and functions of the GM remained largely stable, as did its status as the first and most important Vedic verse. The most salient and significant innovation in the first millennium ce is clearly the deification of the mantra, which is treated in **Part II **(“The Mother of the Vedas”).

The first two chapters of Part I deal with two basic issues: first, the meaning of the text of the mantra (Chapter 1); and second, its various designations, or names (Chapter 2).¹⁴⁸ Both chapters are mainly based on pre-medieval sources.

However (to my knowledge), there is nothing to suggest that either the meaning(s) of the text – in the linguistic sense – or its designations have changed significantly up to today.

The first main **Chapter 3 **(“Adaptive Reuse in Śrauta Ritual”) deals with the ritual employment of the mantra in the large and “solemn” Śrauta rituals during the mid-and late-Vedic period (c. eleventh–fifth centuries bce) and analyzes the motivation behind its recitation in specific Śrauta rituals. How these adaptive reuses might relate to each other against the chronological and geographical background is also explored here.

In **Chapter 4 **(“Selection as Initiation Mantra”), I show how the functions already associated with the mantra in the Śrauta rituals influenced its selection as the primary Brahminical initiation mantra. This selection possibly took place as early as the eighth century bce. However, for the most part, the principles that guided this selection have to be inferred from Gṛhyasūtra passages, which as a rule come several centuries later. By moving backwards in time up until the ṚV, and by reconstructing earlier developments on the basis of later sources, I deviate (to an extent) in this chapter from the usual, linear direction of argumentation.

In the final chapter of Part I, **Chapter 5 **(“Status in Early Hinduism”), I explore how and in what sense the GM became one of the most important mantras of Hinduism. My focus here lies on the Gṛhya and Dharma literature produced in the period between c. 500 bce and 300 ce (for simplicity’s sake, I will refer to 148 The various names and identities of the mantra goddess are dealt with below pp. 239–240.

36 ∙ gāyatRī: mantRa and motheR of the vedas this period as “Early Hinduism”).¹⁴⁹ This chapter, therefore, deals with a period before which the GM had already been in use as an initiation mantra for some time.

Part II, which is dedicated to the goddess, has a somewhat less linear structure. In studying how the mantra was transformed into a goddess, I have identified two distinct, developmental strands: first, from personification and divinization until finally deification of the mantra itself; and second, its association with the goddess Sūryā/Sāvitrī.

The first developmental strand is dealt with in **Chapter 6 **(“Personification, Divinization, Deification”), which tracks down how, over the course of time, various texts contributed to the emergence of a personified and divine mantra.

A number of passages in these texts are ambiguous and allow for readings with or without a “mantra goddess,” so to speak. For this reason, a large part of this chapter is devoted to the philological analysis of the sources and their various reworkings and recensions.

The second strand is discussed in **Chapter 7 **(“Identification with Sāvitrī”), which is above all concerned with the prehistory of the goddess(es) called

“Sāvitrī,” and with the role this goddess and the homonymous mantra came to play in the Sāvitrī story. My core argument here is that the deification of the mantra was significantly advanced by its identification with a pre-existing goddess.

Both chapters 6 and 7 focus on developments that took place in the centuries around the turn of the Common Era. In order to explain those changes however, Vedic and medieval texts are taken into consideration as well.

The last chapter of Part II, **Chapter 8 **(“The Mantra Goddess”), aims to outline a portrait of the mantra goddess as she appears in the Sanskrit literature up to the end of the first millennium ce. The chapter looks at three different aspects of the goddess: her role as the “Mother of the Vedas”; her close relationship with the famous goddess Sarasvatī; and her worship in the Sandhyā, where she came to be treated (almost) like a Tantric mantra, or mantra deity.

In the **General Conclusion **(“From Verse to Deity”), I summarize the results of the two parts with reference to the individual chapters and the relevant sections of the Introduction.

149 In this periodization, Hinduism is followed by what has been called “Classical Hinduism,”

which is especially associated with the Gupta period (c. 350–550 ce). For my definition of Hinduism, see n. 14 on p. 3 above.

intRoduction ∙ 37

–

1000

–

→

–

900

–––

ddess

800

––

–

700

–––

Mantra

600

––

500

–

8:The

–

400

–

eification

–

chapters

–

300

–

Sāvitrī

200

–

individual

–

the

–

with

100

–

Divinization,

Hinduism

vie

-100

Mantra

over

Early

-200

esignations

sonification,

7:Identification

onological

-300

2:D/

Initiation

6:Per

––

Chr

5:Status

––

–

-400

–

aning

Ritual

–

ction

–

––

–

Figur

-500

–

1:Me

–

Śrauta

–

4:Sele

–

-600

–

––

–

-700

–

Reuse

–

-800

–

daptiv

–

-900

–

3:A

–

-1000

←

bce

Part

I I

Part

∙:∙