Ancient Sanskrit Online

Series Introduction

Jonathan Slocum

Please pardon our dust: This lesson series is currently under construction as we seek to remove errors and update the series. Karen Thomson contributed significantly to an earlier version of this lesson series. To see the original version of this series introduction, please click here.

By Ancient Sanskrit we mean the oldest known form of Sanskrit. The simple name 'Sanskrit' generally refers to Classical Sanskrit, which is a later, fixed form that follows rules laid down by a grammarian around 400 BC. Like Latin in the Middle Ages, Classical Sanskrit was a scholarly lingua franca which had to be studied and mastered. Ancient Sanskrit was very different. It was a natural, vernacular language, and has come down to us in a remarkable and extensive body of poetry. (We have intentionally avoided the use of the traditional word "Vedic" to describe the language of these poems for reasons which are described below; see Karen Thomson's other publications for the detailed arguments.)

1. The earliest Indo-European poems.

The earliest surviving anthology of poems in any of the Indo-European languages is in Ancient Sanskrit. Composed long before Homer's Iliad and Odyssey, it consists of over a thousand songs of considerable merit celebrating the riches of nature, whose forces are frequently deified. The relationship that the poets describe with their environment is a sophisticated one. Their hymns serve as talismans, ensuring that the natural world will continue to provide welfare and shelter for man. The power of poetry and song is their primary theme.

The circumstances of the original composition of these poems remain unknown. Believed to be of divine origin, this large body of material, in an archaic and unfamiliar language, was handed down orally, from generation to generation, by priests in ancient India. The highly metrical form of the poems, together with their incomprehensibility, made them ideally suited to ritual recitation by a religious elite. Faithfully preserved through the centuries as a sacred mystery, the text has come down to us in a state of considerable accuracy.

2. 'The Veda'.

Over time a body of dependent and scholastic material grew up around the poems, known loosely as 'the Veda'. Perhaps around 1000 BC (all dating in prehistoric India is only approximate), editors gathered the ancient poems together and arranged them, together with some more modern material, into ten books according to rules that were largely artificial (see section 4 below). They gave the collection the name by which it continues to be known, 'Rig-veda', or 'praise-knowledge'. Other collections came into being, based on this sacred material, and they were given parallel names. The editors of the 'Sāma-veda' arranged the poems differently, for the purpose of chanting, and introduced numerous alternative readings to the text. The sacrificial formulae used by the priests during their recitations, together with descriptions of their ritual practices, were incorporated into collections to which the general name 'Yajur-veda' was given. Later still, a body of popular spells was combined with passages from the Rigveda, again with variant readings, and was given the name 'Atharva-veda'. A continuously-growing mass of prose commentary, called the Brāhmaṇas, also came into being, devoted to the attempt to explain the meaning of the ancient poems. To the later Brāhmaṇas belongs the profusion of texts known as the Upanishads, which are of particular interest to Indologists, as Sanskrit scholars today often describe themselves, because of their important role in the development of early Indian religious thought.

2.1. The continuing influence of 'the Veda'.

This vast body of derivative material remains the subject of extensive study by Indologists. However, from the point of view of understanding the earliest Sanskrit text -- the Rigveda itself -- it has always been, and continues to be, crucially misleading.

Because the poems were put to ritual use by the ancient priests, much of their vocabulary was assumed by the authors of the later texts to refer in some way to ritual activity. The word paśú 'beast, cattle' came to designate a sacrificial victim in texts of the Brāhmaṇas, for example, and juhū́ 'tongue' was thought to mean 'butter ladle'. Abstract words of sophisticated meaning particularly suffered. The compound puro-ḷā́ś 'fore-worship' (from purás 'in front' and √dāś 'worship') acquired the specific sense 'sacrificial rice cake', despite the fact that the word vrīhí 'rice', found in later texts, does not occur in the poems of the Rigveda. The complex noun krátu 'power, intellectual ability', discussed in the introduction to Lesson 7, was misunderstood to mean 'sacrifice' by the authors of the commentaries. Similarly, a number of important verbs of abstract meaning were thought by the editors of the Sāmaveda to be related solely to the production of milk, and to refer to cows (see section 50 of Lesson 10). Indology has always used the word 'Vedic', 'of the Veda', to describe pre-Classical Sanskrit, and the poems to which the name 'Rig-veda' had been given are studied in the context of 'the Veda'. Many ancient mistranslations continue to be maintained with unshakeable conviction by Vedic scholars.

With major pieces of the jigsaw firmly in the wrong place, the rest, inevitably, refuses to fit, and the comparison of passages in the attempt to establish word meanings appears to be a fruitless exercise. Indology has concluded that the Rigveda is not only uninteresting, "describing fussy and technical ritual procedures" (Stephanie Jamison On translating the Rig Veda: Three Questions, 1999, p. 3), but that it is also intentionally indecipherable. "One feels that the hymns themselves are mischievous translations into a 'foreign' language" (Wendy O'Flaherty The Rig Veda. An Anthology, Penguin, 1981, p. 16). Stephanie Jamison vividly portrays the frustrations inherent in the indological approach for a conscientious scholar. "The more I read the Rig Veda, the harder it becomes for me -- and much of the difficulty arises from taking seriously the aberrancies and deviations in the language" (op. cit. p. 9). Viewed through the eyes of Vedic scholars, this most ancient of Sanskrit texts is by turns tedious, and unintelligible: "One can be blissfully reading the most banal hymn, whose form and message offers no surprises -- and suddenly trip over a verse, to which one's only response can be 'What??!!'" (Jamison, op. cit. p. 10). The sophistication of the earliest Indo-European poetry lies buried beneath a mass of inherited misunderstandings that overlay the text like later strata at an archaeological site. Not surprisingly, few Sanskrit scholars today are interested in studying the Rigveda.

2.2. Existing translations.

The poems of the Rigveda are nonetheless of considerable interest to scholars in other fields, in particular linguists, archaeologists, and historians. Linguists regularly refer to Karl Geldner's translation into German made in the 1920s, which is the current scholarly standard; it was reprinted by Harvard University Press in 2003. Geldner's attempt to translate all the poems was however in his own view far from definitive, and it remained unpublished during his lifetime. As he wrote in the introduction to a selection of passages published in 1923, his versions are 'only a renewed attempt to make sense of it, nothing conclusive... where the translation appears dark to the reader, at that point the meaning of the original has also remained more or less dark to me'.

Geldner's struggle to make inherited mistranslations fit necessitates a considerable body of commentary. He notes, for example, to the third line of I, 162, 3, in which the word puroḷā́ś, mentioned above, appears to refer to a goat, that the line is "elliptical. puroḷā́ś (the appetizer consisting of a flat cake of rice in the ritual, see Atharvaveda 9, 6, 12) is used here metaphorically to describe the first-offered goat." His unshakeable conviction that the word has the later specialisation of sense in the context may seem strange, but the translation 'sacrificial rice cake' is hallowed by centuries of later use. To a scholar at home in the later literature the word can have no other meaning.

Geldner's complete translation, and, more particularly, the passages where 'the translation appears dark', forms the basis for much of the selection into English for Penguin Classics by the religious historian Wendy Doniger O'Flaherty, continuously reprinted since its first appearance in 1981. The Penguin selection has been the only version generally available in English for the past quarter of a century, and has introduced a generation of readers to the Rigveda. It perpetuates the belief that these ancient poems are full of arcane references to sacrificial practice, and that they are deliberately obscure.

The distance of O'Flaherty's interpretations from the text itself can be simply illustrated by her version of part of the opening verse of V, 85, "[the god spread the earth beneath the sun] as the priest who performs the slaughter spreads out the victim's skin" (op. cit. p. 211). These twelve words, "as the priest who performs the slaughter spreads out the victim's skin," translate śamitéva cárma 'like a worker a skin'. The word 'victim', together with others, is supplied to give the passage a 'sacrificial' interpretation (the text of V, 85, 1 is example 277 in Lesson 9 of this course). Despite the fact that there is no word for "victim" in the text, her index entry "victim, sacrificial (paśu)" cross-refers to this passage (she omits the accent throughout in conformity with the later language). The word paśú is not present; and what is more, the interpretation that she gives for paśú, "sacrificial victim," is the later, ritual sense used by the texts of the Veda. The word paśú is cognate with Latin pecus (Umbrian pequo, Gothic faihu 'money, moveable goods', Old High German fihu 'cattle', 'Vieh'). See the third verse of the Lesson 5 text, and examples 318 and 357, for passages where the word paśú 'beast, cattle' does appear in the poems.

Tradition colours translations in a number of ways that can be misleading for scholars in other fields. The archaeologist Colin Renfrew, in his stimulating and controversial book Archaeology and Language, chooses Rigveda I, 130, which he quotes in its entirety in Ralph Griffith's nineteenth-century translation, as typical of the whole "in its reference to Soma juice, and in its association of horses and chariots with the heroic practice of war." Leaving 'Soma juice' aside for the moment, is the second part of this conclusion valid? The only reference to human strife in the poem has svàr 'sunlight' as its prize (verse 8); 'chariots' only appear in similes describing streams running down to the sea (verse 5), and wise men fashioning a speech (verse 6); and the Sanskrit word áśva, related by linguists to other words for horse in the Indo-European language family, is absent from the poem. The three adjectives interpreted as 'horse' by the English translator could all have an entirely different meaning. The problem does not lie in the choice of a nineteenth-century translation; Geldner's version of I, 130 is similar, and Louis Renou, working in the 1960s, supplies a word for 'horse' to his French translation of this poem in two additional places.

What of Renfrew's other conclusion, about the typical reference to 'Soma juice'? Four pages on he quotes the first verse of Rigveda I, 102, again using Griffith's translation:

The picture conjured up is pleasing, calling to mind Greek gods supping nectar on Mount Olympus, or Anglo-Saxon heroes feasting in the mead-hall. But "when the Soma flowed" translates a single word only, the abstract noun prasavé (for which see the Lesson 3 text). This same locative form, prasavé, is repeated eight verses later in the poem, where Griffith interprets it entirely differently, as 'in attack': may Indra make us prasavé puráḥ (purás 'in front' again) "foremost in attack." So is the Rigveda typically about the drinking of an intoxicating juice whose identity remains unidentified, or about warfare? Or is it about neither?

3. The decipherable Rigveda.

As this course is designed as an introduction to Ancient Sanskrit I have tried to avoid controversy in my translations, but the misinterpretations permeate the text, and it has not always been possible. In listing the nouns in -van I have included the word grā́van, as it is used by Arthur Macdonell in his Vedic Grammar for Students to illustrate the declension. But I do not believe, as Vedic scholars do, that it means 'ritual stone for pressing out the Soma juice', but that it describes a man who sings (see section 22 in Lesson 5). The traditional interpretation 'ritual pressing stone' produces translations throughout the Rigveda that are puzzling in the extreme. The translation in the first verse of the Lesson 8 text, V, 42, 13, of the feminine plural noun vakṣáṇā also differs significantly from that of Indology. My suggestion 'fertile places' is based on a survey of the contexts in which the word vakṣáṇā occurs. Antiquity understood the word differently, and as referring to part of the body, perhaps as a result of V, 42, 13 where it is traditionally translated 'womb'. But 'womb' fails to fit the other contexts in which vakṣáṇā occurs in the Rigveda, leading to a broad range of interpretations, and ingenious attempts by modern translators to explain them. The most recent dictionary by Manfred Mayrhofer suggests "belly, hollow, entrails; probably also 'bend of a river' and similar." Translators add 'udders' (Geldner and Renou, explaining that the rivers in one passage (my example 76) and the goddess of dawn in another (III, 30, 14) are pictured as cows), 'breasts' (Stephanie Jamison at X, 27, 16) and 'wagon-interiors' (Geldner at X, 28, 8, again citing the authority of a later text). Wendy O'Flaherty offers 'boxes' at X, 28, 8: "[the gods] laid the good wood in the boxes," but her note shows that she is following Geldner: "they take [it] home in boxes on wagons." For another occurrence of vakṣáṇā see example 151 in Lesson 6; and see also section 45.1 for the misreading by the Atharvaveda, in perplexity at a context that is clearly terrestrial, of the noun here as a participle.

My translation 'fertile places', however, is at variance with a strong tradition that explains the first verse of the Lesson 8 text as a description of primeval incest. This is an idea that Wendy O'Flaherty enthusiastically embraces elsewhere: she offers, for example, as an explanation of her perplexing translation of III, 31, 1 the note, "the priest pours butter into the spoon, and the father pours seed into his daughter" (p. 155). Not only is there no word for 'seed' in the passage glossed here, there is none for 'priest', 'spoon', or 'butter' either.

The Rigveda remains open to imaginative exegesis because Indologists continue to believe that its poems are deliberately obscure. "As the Brāhmaṇas tell us so often, 'the gods love the obscure'... and in investigating Vedic matters, we must learn to cultivate at least that divine taste" (Jamison The Ravenous Hyenas and the Wounded Sun. Myth and Ritual in Ancient India, 1991, p. 41). But the Brāhmaṇas came into existence because the meaning of the poems had become lost. The ancient commentators didn't understand the Rigveda, and they were trying to work out what the poems were about. The American linguist William Dwight Whitney, writing over a century ago, had little time for "their misapprehensions and deliberate perversions of their text, their ready invention of tasteless and absurd legends to explain the allusions, real or fancied, which it contains, their often atrocious etymologies" (Oriental and Linguistic Studies, 1873, p. 110), but to be fair to the authors of the Brāhmaṇas, they lacked modern resources: a written text and a concordance, for example. Without the ability to compare contexts decipherment is extremely difficult, and "ready invention" is a tempting alternative. Indology today, which has these resources, nonetheless adheres to the ancient methods of investigation. In her paper quoted at the beginning of this introduction, Stephanie Jamison propounds the thesis that "many of the most obscure images and turns of phrase in the Rig Veda make sense as poetic realisations of specific ritual activities, and whole hymns and hymn complexes can poetically encode the sequences and procedures of a particular ritual," citing as an example "Joel Brereton's recent brilliant explanation of the fiendishly opaque mythology of the divine figures, the R̥bhus, as reflecting in remarkable detail the Third Pressing of the Soma Sacrifice" (p. 7). This is the approach that first buried the Rigveda from view in ancient times, and in continuing to apply it modern Indology is simply throwing earth onto the mound.

As an editorial postscript to an article published in 1965 on the word vidátha, the Iranian scholar H.W. Bailey commented, "It should not pass unnoticed that the most recent translation of the Rigveda by L. Renou... knows nothing of vidátha- as 'congregation'... Each translator tends to read into the obscure texts his own theories." Only attention to the text itself, which has been out of print for much of the past century, will lift the mists that have always enveloped the Rigveda. Study of the earliest Indo-European poetry may have languished in recent times, but the parallel discipline of Old English studies has notably flourished as a result of the application of rigorous scholarship, deriving from the 'new philology' introduced into England from Germany in the 1830s. "The greatest strength of Anglo-Saxon and Medieval Studies in general, I believe, is that by and large we have never lost our devotion to the text and to interpreting texts. We have not let theory estrange us from the life's blood of our enterprise, the texts and artifacts at the center of our study." (Fred C. Robinson, in the introduction to The Preservation and Transmission of Anglo-Saxon Culture, 1997). The Rigveda stands alone; unlike Old English it has not come down to us together with any artifacts that we know to be dateable to the same remote period in time. But it constitutes a considerable body of material, and remarkably, given its antiquity and importance, it remains largely undeciphered. This course has been written primarily to give access to the text to scholars from other disciplines, and to provide the means for a fresh approach to the decipherment of the earliest Indo-European poems.

4. The text, and the editorial tradition.

Until very recently the original poetic form of the Rigveda was also hidden. Luckily for modern students, this is no longer the case (see below). The artificial ordering of the poems by their ancient editors however remains in use today.

Books II to VII (of ten books) are arranged on a uniform pattern. Hymns addressed to Agni 'Fire' (Latin ignis) always come first: a frequent epithet of Agni in the Rigveda is puró-hita 'placed in front'. The Agni hymns are followed by hymns to Indra. Within these two groups the poems are arranged in order of diminishing length. Poems addressed to other gods form the third group of each book. Book VIII follows a more natural arrangement, and contains many poems of early date. The songs of Book IX are a special case, having been put together because of the similarity of their vocabulary, notably the obscure verb √pū, pávate and its derivatives. They contain many refrains (see section 40 in Lesson 8) that help to identify the groups to which they originally belonged. Books I and X appear to have been added later to the core collection. A different numbering system which is popular in India preserves this order but divides the material equally into eighths; still another, followed by Grassmann in his concordance (see the reading list in section 9), simply numbers the poems consecutively.

For much of its history this body of poetry was passed down orally. Even following the general introduction of writing, some time before the 3rd century BC, there was a strong reluctance to write down this sacred and cabalistic text, which was the exclusive and secret property of an elite. The date of the earliest written text that has come down to us, from which all others derive, is characteristically unknown. It is a 'continuous' text -- in Sanskrit, saṃ-hitā 'placed-together' -- in which adjacent sounds combine with each other across word boundaries according to strictly applied phonetic rules. This combining of sounds is known as sandhi, from the Sanskrit saṃ-dhi 'placing-together' (see section 7). A second ancient text, the pada or 'word' text, which gives all the words separately in their original form, appears to have been compiled at around the same time. The surviving manuscripts of these two texts in the Devanāgarī script were edited and published in a definitive edition by Max Müller in the second half of the nineteenth century. It was clear to Max Müller that the 'continuous' text obscured the original form of these poems. In 1869 he wrote, "if we try to restore the original form of the Vedic hymns, we shall certainly arrive at some kind of Pada text rather than at a Sanhitā text; nay, even in their present form, the original metre and rhythm of the ancient hymns are far more perceptible when the words are divided, than when we join them together throughout according to the rules of Sandhi." But it was not until 1994 that the metrically restored text, in a modern transliterated form, was published by the American scholars Barend van Nooten and Gary Holland. For the first time in its history, the Rigveda was clearly revealed, on the printed page, as poetry.

Van Nooten and Holland's edition has unfortunately been out of print for some years. In order to make the metrically restored text universally available, we have produced an edited online version, The Rigveda: Metrically Restored Text.

The system of modern transliteration used by van Nooten and Holland is also used in the full Unicode 3 versions of these lessons.

5. A note on methodology.

My aim throughout the grammar sections has been to provide a description of the language that is as straightforward as possible. Many factors have traditionally combined to make the Rigveda inaccessible to scholars in other fields, one of which is grammatical complexity. I have opted for the clearest presentation that I could find. As Arthur Macdonell's Vedic Grammar for Students is an excellent summary and remains in general use, I have tended to follow him in the attribution of verbal forms, but I have, for example, categorised the types of the aorist following Whitney, as his description seems more straightforward. Others may disagree with the choices that I have made, and I welcome comments. In addition, as Macdonell wrote in the Preface to his 1917 Vedic Reader (the immediate predecessor of this course), "freedom from serious misprints is a matter of great importance in a work like this." The Vedic Reader never reached a corrected edition, but one of the advantages of online publishing is the relative ease with which mistakes can be put right. I particularly welcome corrections.

Indologists have so far found no common ground for debate with my approach. I am very grateful therefore to Ramesh Krishnamurthy for constructive discussion and advice, and to Alexander Lubotsky for proof-reading the first four lessons and making some necessary corrections. Where however Professor Lubotsky urged the traditional interpretations over my revisions I have stuck to my guns: for example, in my translation of the feminine noun usríyā in the second verse of the Lesson 4 text (surely not 'cow'), and of páyas in a number of the examples (not, in my view, 'milk'; see section 50.2). Where my translation of words occurring in the lesson texts differs from the current consensus, the translation appears in italics in the glossaries. (Occasionally translations are in italics because there is no existing consensus.) Some retranslations are minor refinements of sense; others, like usríyā, and vakṣáṇā discussed in section 3 above, are more radical. Wherever possible, however, I have chosen passages that are free of problem words, and italicised translations of this kind are relatively few in number.

My greatest due of thanks is to Professor Winfred Lehmann and the Salus Mundi Foundation, for making it possible to put the course online.

6. The sounds of Sanskrit and the Sanskrit alphabet.

The 'dictionary' order of Sanskrit follows phonetic rules. The vowels come first.

The short vowel a is pronounced approximately as the a of English about, and i and u as in bit and put (in Classical Sanskrit the short a sound became even shorter, and is transliterated as a u sound). These vowels each have a long equivalent, ā, ī, ū, pronounced as in English bar, beat and boot. In addition Sanskrit has a vocalic r sound, r̥, which occurs frequently and is pronounced like the r in British English interesting with accent on the first syllable, 'íntr̥sting'. The word Rigveda itself in Sanskrit begins with this vocalic r, which is why it is sometimes transliterated without the i, Rgveda. (In this course r̥ is transliterated both as ri and as ar.) There is also a longer r̥ sound, r̥̄, and a vocalic l sound, l̥, which is very rare and is pronounced something like the l (with silent e) in table.

Four long vocalic sounds classified as diphthongs follow:

The equivalent English sounds are e (bait), ai (bite), o (boat), and au (bout).

The consonants are also arranged phonetically.

These are ordered according to their physical production in speech. The sounds produced at the back of the mouth, k, kh, g, gh are listed first, and are described as 'velar' because they are made with the tongue touching the soft palate (velum in Latin). 'Palatal' consonants, c, ch, j, jh, are made slightly farther forward in the mouth, with the tongue touching the hard palate; 'dentals', t, th, d, dh, with the tongue touching the teeth; and 'labials', p, ph, b, bh, with the lips. This is given in tabular form below. Each sequence or class comprises a 'voiceless' sound, pronounced without the vibration of the vocal cords, like k; the same sound aspirated, kh, pronounced with a following h sound; a 'voiced' sound, g; the same sound aspirated, gh; and a nasal.

Between the palatal and dental classes appears another sequence. The dental t sound is in fact like a French t (tout), made with the tongue touching the teeth. The Indian retroflex sounds are made with the tip of the tongue curved backwards (hence the name) behind the upper teeth, and then flicked forward. To Indian ears the t of try is more like a retroflex than a dental sound.

The nasals belonging to each class simply represent the sounds produced in each part of the mouth. English also has a range of nasal sounds, but they are not generally reflected in writing. Compare, for example, the sound of the nasal in these five words, which changes because of the different adjacent consonants: hunger (velar), punch (palatal), unreal (retroflex), hunter (dental), and, with a written change, lumber (labial).

Note: ḍ becomes ḷ (and ḍh ḷh) between vowels, as in the word puroḷā́ś mentioned in the first section of the introduction.

At the end of the alphabet come semivowels and sibilants, and h:

The semivowels and sibilants are again in phonetic order:

The semivowels are closely related to vowels: y corresponds to i/ī, r to r̥/r̥̄, l to l̥, and v (pronounced like English w when preceded by a consonant) to u/ū. The same close vowel/semivowel relationship is reflected in the eighteenth-century spelling of persuade, 'perswade'. In the earliest 'continuous' text the written semivowel often represents an original vowel. Palatal ś and retroflex ṣ are both pronounced something like English sh, the second again with the tongue slightly curved backwards.

In addition there are two sounds that occur very frequently, ṃ and ḥ, which are not original but represent other sounds under the influence of sandhi (see below). In most dictionaries, that by Monier-Williams for example, ṃ is positioned alphabetically according to the original nasal that it represents, which can be confusing. In the course glossaries these two sounds have been arranged to follow the diphthongs and precede the consonants.***Since the 2016 software change this is no longer the case; see note at the beginning of Lesson 1.***  ṃ (sometimes written m̐) is a pure nasal: táṃ is pronounced something like French teint. ḥ is an unvoiced breathing sound.

7. Sound changes, combination of sounds, or sandhi.

The word sandhi is used to describe the way in which sounds change as a result of adjacent sounds, both within words and across word boundaries, and it is a natural phenomenon in speech. Consider the English nasal sounds described in the previous section, for example. Because the extensiveness of its occurrence in Sanskrit is unparalleled in any other language, the Sanskrit name saṃ-dhi 'putting-together' has come to be used to describe this phenomenon in other languages.

The evidence of the Rigveda with respect to vowel sandhi (see section 45.1 of Lesson 9) suggests that many of the sandhi changes made by the later editors were in fact artificial, and the result of the imposition of fixed rules onto a language that was more naturally flexible. In English most sandhi changes are not written, but in Sanskrit they are extensively reproduced in writing. This, as Michael Coulson mildly expresses it in his guide to the Classical language, Teach Yourself Sanskrit, is "not necessarily a good thing." The complexity of the written sandhi system is potentially alienating for a beginner. This section therefore provides only a brief sketch of the principles involved to prepare the reader for the kinds of change that he will encounter in the lessons. Appendix 1 at the end of this course presents, in tabular form, the changes that occur.

7.1. Sandhi of vowels.

As mentioned in the previous paragraph, the form in which these poems were first written applied later rules of vowel sandhi which the metre indicates were inappropriate. Final i/ī and u/ū, for example, when followed by another vowel were systematically turned into the related semivowels y and v in order to avoid hiatus, that is, to give a smooth, continuous sound. But the syllabic loss that this change entails destroys the rhythm of the poems and the vowels must nearly always be restored. A language of a different character emerges. "The text of the Rigveda, when metrically restored, shows us a dialect in which the vowels are relatively more frequent, and the syllables therefore lighter and more musical, than is the case in classical Sanskrit. The Homeric dialect differs just in the same way from classical Greek" (Arnold, Vedic Metre p. 106).

Certain vowels when juxtaposed nonetheless do change in the Rigveda. Two short vowels that are the same, for example, -i at the end of a word followed by i- at the beginning of a word, usually combine to give the long vowel, here -ī-. Long vowels, or a mixture of long and short vowels, combine in the same way. In example 225, aśvinā́ in fact represents two words, aśvinā ā́, and in example 234 ā́gāt represents ā́ ágāt. This does not always happen: in example 334, for instance, the two adjacent a sounds in evá agníḥ have not combined, nor in example 136, stotā́ amatīvā́. Sometimes at the end of a line ā́ is written ā́m̐ to make clear that it does not combine with the initial vowel at the beginning of the next line. There are examples of this in the Lesson 4, 5, and 10 texts.

Some combinations of dissimilar vowels also regularly occur, particularly with final a/ā, which may combine with initial i/ī to give e, and with initial u/ū to give o. In example 26, for example, aśvinoṣásam = aśvinā uṣásam, and in example 277 śamitéva = śamitā́ iva. Again these rules are not invariably applied: see aśvinā ūháthuḥ in example 224.

7.2. Sandhi of consonants.

In the written system consonants are also regularly subject to change. s and m are frequently found at the end of words: nominative singular devás 'god', accusative singular devám. Final m, the labial nasal, under the rules of sandhi becomes the pure nasal ṃ if followed by anything other than a vowel, or another labial sound. Final s is regularly given as the unvoiced breathing sound ḥ by the editors -- this is the form it always takes at the end of a phrase or line. It is changed to r before a 'soft' sound like a vowel or a voiced consonant. With an immediately preceding a, however, it is treated differently: -as becomes -o before soft sounds. Examples of these changes in simple compound words have already been given: the word saṃ-dhi itself, saṃ-hitā 'placed-together', and Rigvedic puró-hita 'placed in front'.

Final t also occurs frequently, as in tát 'that, it'. When followed by a soft sound it becomes d, but before n or m it becomes n. This sounds complicated, but such changes soon become familiar. They occur naturally when a language is spoken at speed, and are a good source of the punning jokes beloved of children (as in "say iced ink very quickly").

The first line quoted in the introduction to the first lesson, agníṃ dūtám puró dadhe, shows sandhi effects at the end of the first and the third word. A word for word version would read agním dūtám purás dadhe (the m of dūtám was unchanged as it was followed by a labial consonant, p). The last two lines of the first lesson text,

with sandhi removed and final s restored, read

All the lesson texts are glossed word for word with the sandhi changes removed, and sandhi changes are also regularly explained in square brackets when they occur in the examples.

7.3. Retroflexion.

Included within the scope of sandhi are changes known as retroflexion. The sounds r̥, r̥̄, r and ṣ under certain circumstances make n retroflex, ṇ, even across word boundaries: see example 325, prá ṇaḥ for prá nas. Similarly, vowels other than a or ā, and k, r and ṣ, can change s to ṣ. See example 81, abhí ṣyāma [abhí syāma], and example 296, nū́ ṣṭutáḥ [nū́ stutáḥ], where the ṣ in turn has made the following t retroflex. This occurs very frequently within words: arká 'song', arkéṇa 'with song', arkéṣu 'in songs'.

8. Vowel gradation, or ablaut.

A characteristic feature of Indo-European languages is the variation of vowels in derivatives from a root. Found regularly in the verbal system, it also occurs in nouns, as in sing, sang, sung, and also song. This vowel variation is known as ablaut. Its occurrence in Sanskrit was recognised by the ancient grammarians, who described it as 'strengthening' of the vowel. The table shows how the simple vowel is strengthened.

Vowel strengthening is found in nominal derivatives, like the element vaiśvā- in the first word of the first lesson text, which is a derivative of víśva 'all', and pā́rthiva 'earthly' in the third verse of the Lesson 3 text, which is a derivative of pr̥thivī́ 'earth'. It is a feature of many parts of the verb, like the causative, viśáti 'he enters', veśáyati 'he causes to enter' (see section 33.1), and the aorist passive: ámoci 'it has been released' from √muc 'release' (see section 48.1).

9. Reading List.

Text.

• Rig Veda. A Metrically Restored Text with an Introduction and Notes. Edited by Barend A. van Nooten and Gary B. Holland. Harvard University Press, 1994.
• The Rigveda: Metrically Restored Text. Online version edited by Karen Thomson and Jonathan Slocum. University of Texas, 2006.
• Arnold, E.V. Vedic Metre. Cambridge University Press, 1905.

The most important resource for studying the Rigveda is the text itself, and the metrically restored text is the first to show its original poetic form. Previous editions are misleading in masking both form and meaning, as explained in section 45 of Lesson 9.

Arnold's 1905 study goes well beyond its modest title, not only in disentangling the original metrical form but also in using the metre, together with vocabulary and grammatical forms, to attempt a chronological arrangement of the poems.

Concordances.

• Grassmann, Hermann. Wörterbuch zum Rig-Veda. Leipzig, Brockhaus 1873.
• Lubotsky, Alexander. A R̥gvedic Word Concordance. New Haven, CT, American Oriental Society 1997.

Grassmann's dictionary and analytical concordance remains invaluable; the recent concordance by Lubotsky is useful in listing all the word forms, without translation, in the context of the line in which they occur. Though deriving from van Nooten and Holland's metrical edition, the text in Lubotsky's concordance is quoted in unrestored form.

Grammars.

• Macdonell, Arthur. Vedic Grammar for Students. Oxford, Clarendon Press 1916.
• Macdonell, Arthur. Vedic Grammar. Strassburg, Trübner 1910.
• Whitney, William D. Sanskrit Grammar. Second Edition. Harvard University Press 1889.
• Whitney, William D. The Roots, Verb-forms and Primary Derivatives of the Sanskrit Language. Leipzig, Breitkopf und Härtel 1885.
• Arnold, E.V. Historical Vedic Grammar. [in] Journal of the American Oriental Society, vol. 18, 1897, pp. 203-353.

As a compendium of Rigvedic grammar, Macdonell's Vedic Grammar for Students remains extremely useful. The same author's earlier and fuller Vedic Grammar is an outstanding work of scholarship, and is currently available from India as a reprint (Munshiram Manoharal, 2000; the reprint however lacks the last gathering and therefore much of the index).

In addition to the works by Macdonell, Whitney's nineteenth-century Sanskrit Grammar, which includes the early language, is useful in regularly clarifying what may seem unduly complex. His supplementary volume, The Roots, Verb-forms and Primary Derivatives of the Sanskrit Language, arranges nominal forms under the verbal roots to which they belong, and is a guide to the regularly transparent word formation of Sanskrit (see section 49 in Lesson 10).

Arnold's Historical Vedic Grammar, while not for the beginner, is a rich statistical resource for the historical study of pre-Classical Sanskrit.

Dictionaries and semantic studies.

• Monier-Williams, Monier. Sanskrit-English Dictionary. New Edition, Oxford University Press 1899.
• Mayrhofer, Manfred. Etymologisches Wörterbuch des Altindoarischen. [Part I. Ältere Sprache]. Heidelberg, Carl Winter 1992-1996.

All dictionaries contain translations that are misleading for the Rigveda. With this caveat, the Sanskrit-English Dictionary by Monier Monier-Williams, based on the seven-volume Sanskrit-Wörterbuch by Otto Böhtlingk and Rudolph Roth, is a work of great erudition. The most recent dictionary of early Sanskrit, by the eminent linguist Manfred Mayrhofer, is useful for presenting the Rigveda in its Indo-European context, and is distinguished by the regular unwillingness of its author to accept traditional interpretations without question.

Those interested in the reconsideration of inherited interpretations may wish to look at my studies of some of the words mentioned in this introduction. Thomson, Karen, "The meaning and language of the Rigveda: Rigvedic grā́van as a test case," Journal of Indo-European Studies 29, 3 & 4, 2001, 295-349; "The decipherable Rigveda: a reconsideration of vakṣáṇā," Indogermanische Forschungen 109, 2004, 112-139; "Why the Rigveda remains undeciphered: the example of puroḷā́ś," General Linguistics 43, 2005 [2003], 39-59; and, a sister paper to the last, "The decipherable Rigveda: tiróahnyam as an example," Journal of the Royal Asiatic Society, 15, 1, 2005, 63-70 (the words puroḷā́ś and the temporal adverb tiróahnyam, misunderstood by the authors of later Vedic texts as an adjective, frequently occur together).

Related Language Courses at UT

Most but not all language courses taught at The University of Texas concern modern languages; sometimes courses are offered in ancient languages, though more often at the graduate level. Indic language courses, including Sanskrit, are taught in the Department of Asian Studies (link opens in a new browser window). Other online language courses for college credit are offered through University Extension (new window).

Indic Resources Elsewhere

Our Links page includes pointers to Indic resources elsewhere.