Skip to content

Chapter 6: The Treatises of Music: From the Dattilam to Bhatkhande

For two thousand years Indian musicians and their patrons wrote about music. The result is one of the largest technical literatures on music in any pre-modern civilisation: several hundred works in Sanskrit, with others in Persian, Telugu, Kannada, Tamil, Hindi, Marathi and Bengali, ranging from short manuals to encyclopaedias of many thousand verses. Chapter 4 traced the central idea of this literature, the rāga, from the Nāṭyaśāstra to the modern classifications, and Chapter 5 its theory of rhythm. This chapter looks at the texts themselves: who wrote them, for whom, where and why; what each added to the tradition it inherited; and how they should be read. It follows the literature from the music chapters of the Nāṭyaśāstra and the Dattilam, through Mataṅga, the royal compendia of the eleventh and twelfth centuries and Śārṅgadeva’s Saṅgītaratnākara, to the regional treatises of the fifteenth to eighteenth centuries and the reformers of the colonial period. The Persian literature on Indian music is treated in Chapter 7 and the dance treatises in Chapter 12; here they appear only where they bear on the history of musical theory.

6.1 The Shape of a Tradition

The Sanskrit literature on music (saṅgītaśāstra) has some features that a modern reader needs to keep in mind. Its works are mostly in verse, usually the anuṣṭubh metre of the epics and the śāstras, with occasional prose; they were meant to be memorised and recited as well as read. Their authors worked by compilation as much as by original statement. A medieval treatise typically begins with the received doctrine, often quoting or paraphrasing earlier authorities at length, and then adds what its author regarded as new: the regional (deśī) rāgas, compositions, instruments or dances of his own time and place. Many works survive only because later compilers quoted them, and many of the authorities named in the texts — Kohala, Viśākhila, Tumburu, Kāśyapa, Yāṣṭika, Durgāśakti and others — are known only from such quotations (Nijenhuis 1977; Rowell 1992).

The literature is also, for most of its history, a literature of courts. Several of its authors were kings or claimed to be: Someśvara III of the Cāḷukyas, Nānyadeva of Mithilā, Kumbhakarṇa of Mewar, the Recharla king Siṃhabhūpāla, the Maratha king Tulajā of Thanjavur. Others were ministers, court poets or Brahmin scholars attached to a court, and the works name their patrons with care. To write on music was a mark of royal accomplishment, and a treatise was a gift to a dynasty’s reputation as well as a manual. This explains both the strengths of the literature — its learning, its system and its preservation in court libraries — and its limits: it records the music that was valued at courts, describes performers only as the patron saw them, and has almost nothing to say about the music of the village, the women’s quarters or the low-caste professional, except as the raw material of new deśī categories.

Three ideas organise the whole literature. The first is the distinction between mārga and deśī: the “path” of the ancient, canonical music believed to have been revealed by the gods and recorded by Bharata, and the “regional” music of particular places and people. The second is the distinction between lakṣaṇa, the theoretical definition of a thing, and lakṣya, the thing as found in practice, which theorists from Mataṅga to Bhatkhande invoked when they had to admit that the music of their time did not fit the books. The third is the authority of the Nāṭyaśāstra itself: almost every later author begins by placing his work in the line of Bharata, even when, as often happened, the music he describes would have been unrecognisable to the earlier text. The history of the literature is in large part the history of how successive writers reconciled the authority of the past with the music of their own day.

6.2 The Gāndharva Texts: The Nāṭyaśāstra and the Dattilam

The music chapters of the Nāṭyaśāstra (chapters 28–33 in Ghosh’s numbering) were described in Chapter 2. Their companion is the Dattilam, a compact treatise of some 240 verses ascribed to Dattila, whom later tradition made one of the sons or pupils of Bharata. It is probably somewhat later than the musical chapters of the Nāṭyaśāstra, which it closely resembles, and is usually placed in the early centuries CE. Mukund Lath’s study and translation, and Emmie te Nijenhuis’s edition, made it the best-understood of the early texts (Lath 1978; Nijenhuis 1970).

Both texts describe gāndharva, a music that they define as consisting of three components: notes (svara), time-measure (tāla) and words (pada). Gāndharva was not ordinary song. It was a sacred, fixed repertoire, performed in the preliminaries of the theatre and in ritual, whose correct performance was believed to please the gods and to produce unseen merit (adṛṣṭa); its rules were correspondingly strict. The songs of the play itself, the dhruvās, belonged to a different category, gāna, which was valued for its visible effect on the audience (Chapter 19). Lath has argued that this distinction explains much about the later tradition: the theory that the treatises preserved for centuries was the theory of a ritual music, not of the music people heard, and when later authors repeated the grāmas, mūrchanās and jātis, they were preserving the theory of a sacred art rather than describing the music around them (Lath 1978; Rowell 1992).

The content of the Dattilam is technical and systematic. It defines the twenty-two śrutis and their distribution among the notes; the two grāmas and their fourteen mūrchanās; the forms of a mūrchanā with six or five notes; the jātis with their characteristic notes; and the “path” (mārga) tālas with their gestures (Chapter 5). It also describes the gītakas, the large ritual song-forms of gāndharva, built from fixed sections and performed in strict order. What makes these early texts so valuable is their coherence: the jāti system is a carefully worked-out modal theory, in which each melodic type is defined by a set of functional notes — initial, predominant, final, penultimate, limits of range, notes to be emphasised or avoided — and this functional way of thinking about a mode survived the disappearance of the jātis themselves and passed into the theory of rāga (Widdess 1995).

6.3 Mataṅga and the Discovery of Deśī

Between the Dattilam and the next major surviving text lies a gap of several centuries, partly filled by the authorities quoted in later works. The text that ends the gap is Mataṅga’s Bṛhaddeśī (“The Great Treatise on Regional Music”), composed in the second half of the first millennium CE; its date cannot be fixed more closely, and proposals range from about the sixth to the ninth century. The text survives in a damaged state: its surviving portion breaks off in the middle of its account of composed songs. Prem Lata Sharma’s edition and translation is the standard one (Sharma 1992).

The title announces the novelty. Mataṅga defines deśī music, in contrast to the ancient mārga, as the music current in the different regions, sung for their own pleasure by women, children, herdsmen and kings. He does not reject the gāndharva theory; much of his work restates and discusses it, and he quotes earlier authorities, some of them otherwise lost, on disputed points. But he extends the theory to new material. His treatment of the notes includes a discussion of nāda, the sound from which music arises, and of the nature of śruti; his account of melodic types introduces the grāma-rāgas, derived from the jātis, and a wider class of regional melodies called bhāṣā and vibhāṣā, which he takes partly from the lost work of an authority named Yāṣṭika. These melodies are grouped under styles of singing called gīti. Above all, he gives the first known definition of rāga (Chapter 4), which treats it as a melody characterised by its notes and movements and defined by its effect on the listener, and he applies the term to melodic types of the deśī world (Sharma 1992; Widdess 1995).

The surviving part of the text also begins the systematic description of composed songs (prabandha), the vernacular and Sanskrit song-forms of the courts and temples, which later treatises were to describe at much greater length. With Mataṅga the centre of gravity of the theory moves from the ritual music of the past to the living music of the regions, and every later Sanskrit treatise follows his example in combining the two. The Kudumiyamalai inscription in Tamil Nadu (Chapter 4), which sets out melodies under the names of grāma-rāgas of the kind Mataṅga describes, shows that this theory was used by practising musicians in his period (Widdess 1979).

Abhinavagupta’s commentary on the Nāṭyaśāstra (c. 1000), the Abhinavabhāratī, belongs to the same world. Its discussion of the music chapters, though damaged in the manuscripts, shows a scholar trying to explain an ancient system he no longer heard by reference to the rāgas and practices of his own time, and it preserves many quotations from lost writers on music. Its main influence lies in the theory of rasa (Chapter 3), but for the historian of music it is a reminder that the ancient theory was already being read, by the eleventh century, as a classical text in need of interpretation.

6.4 The Royal Compendia of the Eleventh to Thirteenth Centuries

The eleventh and twelfth centuries produced a cluster of works on music written at the courts of the Deccan, of eastern India and of the west. They share a pattern: a restatement of the mārga theory, followed by a long account of the deśī rāgas, compositions, instruments and dances of the court.

Nānyadeva’s Bharatabhāṣya.
Nānyadeva (r. c. 1097–1147), the Karṇāṭa king of Mithilā in northern Bihar, wrote a long commentary on the music of the Nāṭyaśāstra, also known as the Sarasvatīhṛdayālaṅkāra. Besides discussing the ancient theory it describes the rāgas of its time and gives notated examples, and Richard Widdess has used its notations, with those of Mataṅga and Śārṅgadeva, to reconstruct the outlines of early rāgas (Widdess 1995).
Someśvara’s Mānasollāsa.
The encyclopaedia of royal life compiled at the court of the Cāḷukya king Someśvara III of Kalyāṇa (1131) treats music, instruments and dance among the “pleasures” (vinoda) of a king (Chapter 12). Its section on song describes rāgas and many kinds of prabandha, among them songs in Kannada and other regional languages, and it is valuable precisely because it records the courtly music of the Deccan without much concern for the mārga theory.
Jagadekamalla’s Saṅgītacūḍāmaṇi.
A treatise attributed to Someśvara’s successor, Jagadekamalla II (r. 1138–1151), continues the same Cāḷukya tradition.
Pārśvadeva’s Saṅgītasamayasāra.
The work of a Jain author of the twelfth or thirteenth century, it gives an unusually full account of deśī music: rāgas, phrases, compositions, regional tālas and dance. Its author’s Jainism is a reminder that the music treatises were not the preserve of any one religious community.

These works show music as one of the arts of a cultivated court, discussed in the same terms across a large part of the subcontinent. They also show that the repertoire of rāgas was growing and changing rapidly. Names that become prominent in later music — Bhairavī, Toḍī and Gurjarī among them — appear in this period among the bhāṣā and deśī rāgas, and the older grāma-rāgas retreat into the theory. The Gītagovinda of Jayadeva, composed in Bengal or Odisha in the twelfth century, belongs to the same moment: its songs, as transmitted in the manuscripts, carry the names of the rāgas and tālas in which they were sung, and the work became one of the most widely performed texts in Indian music and dance (Chapter 17).

6.5 The Saṅgītaratnākara

6.5.1 The author and his court

The Saṅgītaratnākara (“Ocean of Music”) of Śārṅgadeva summed up this literature and became its standard authority. Śārṅgadeva tells us that his family came from Kashmir: his grandfather Bhāskara had migrated to the Deccan, and his father Soḍhala served the Yādava kings of Devagiri. Śārṅgadeva himself was an official at the court of Siṃhaṇa (r. c. 1210–1247), and the work belongs to the first half of the thirteenth century. Its importance owes much to its timing. Within a century of its composition the Yādava kingdom fell to the Delhi Sultanate, and the music of the north was transformed by the courts of the sultans; the Saṅgītaratnākara thus became the last great statement of the theory before the northern and southern traditions diverged, and both looked back to it as their common source (Shringy and Sharma 1978; Rowell 1992).

6.5.2 The seven chapters

The work is divided into seven chapters, whose contents are summarised below; the chapters on tāla, instruments and dance are discussed in Chapters 5, 10 and 12.

  1. Svaragata (on the notes).
    The chapter opens with the origin of sound in the body, drawing on yogic physiology and the doctrine of nāda: the word itself is explained as a union of breath (na) and fire (da). It then treats śruti, svara, grāma, mūrchanā, the patterns of notes (tāna), the varṇas (types of melodic movement) and their ornaments (alaṅkāra), and the jātis.
  2. Rāgaviveka (the discrimination of rāgas).
    The chapter classifies 264 rāgas in the categories inherited from Mataṅga and his successors — grāma-rāgas, uparāgas, bhāṣās and others — and then turns to the rāgas “famous in former times” (pūrvaprasiddha) and “famous now” (adhunāprasiddha), the latter including categories called rāgāṅga, bhāṣāṅga, kriyāṅga and upāṅga. The distinction between ancient and current rāgas is Śārṅgadeva’s own acknowledgement that the music of his day had moved beyond the older theory.
  3. Prakīrṇaka (miscellany).
    Here are the qualities of the composer (vāggeyakāra) and the singer, the qualities and faults of voice, the fifteen gamakas (Chapter 4), the elaboration of rāga without text (ālapti) and the phrase-types (sthāya) from which a performance was built.
  4. Prabandha (composition).
    The theory of composed song (below).
  5. Tāla.
    The mārga tālas and 120 deśī tālas.
  6. Vādya (instruments).
    The four classes of instrument, with detailed accounts of the vīṇās, flutes and drums and their techniques.
  7. Nartana (dance).
    Dance and drama, mārga and deśī.

6.5.3 The composer and the composition

Two parts of the work have been particularly influential for the history of musical form. The first is Śārṅgadeva’s portrait of the ideal composer, the vāggeyakāra, “maker of words and song”. Such a person, he says in effect, should be master of grammar, lexicon and prosody, of the rāgas and tālas, of the regional languages and their idioms, of the rasas and the arts of poetry, and should have a fine voice and a sense of propriety. The ideal of the composer who writes both text and music and sings his own compositions is one that later traditions — the Haridāsas, the dhrupad singers of Gwalior, the Karnatak Trinity — were to embody.

The second is the theory of prabandha. Śārṅgadeva distinguishes “bound” (nibaddha) music, composed and set to tāla, from “unbound” (anibaddha) music, such as the ālapti that explores a rāga without text or metre — a distinction that corresponds to the alternation of improvised ālāp or ālāpana and composed song in modern concerts. He analyses a prabandha in two ways. Its “sections” (dhātu) are the udgrāha (opening), the melāpaka (link), the dhruva (the fixed, refrain-like section that must not be omitted) and the ābhoga (conclusion), sometimes with an additional antara. Its “limbs” (aṅga) are six kinds of material that may be combined: svara (sol-fa syllables), biruda (words of praise), pada (meaningful text), tena (auspicious syllables such as tena), pāṭa (drum syllables) and tāla. The number of limbs a song contained determined its class. The prabandhas themselves are grouped into large families, among them the sālagasūḍa, a group of seven song types (dhruva, maṇṭha, pratimaṇṭha, nissāru, aḍḍatāla, rāsaka and ekatālī) from which the names of several later southern sūḷādi tālas (Dhruva, Maṭhya, Aṭa, Eka), and the sūḷādi song itself, appear to derive (Shringy and Sharma 1978; Rowell 1992).

The vocabulary of this theory survived the songs it described. The names dhruva and ābhoga reappear in the dhrupad (dhruvapada) and its fourth section, the ābhog (Chapter 7); the combination of sol-fa, meaningful text and drum syllables reappears in the tarānā of Hindustani music, the tillāna of the south and the rhythmic compositions of the dancers. Continuity of names does not prove continuity of music, but the analytical habits of the prabandha theory — the composition as a set of sections and materials — persisted in the way Indian musicians thought about form.

6.5.4 The commentators

Two great commentaries extended the authority of the Saṅgītaratnākara into the following centuries. The Sudhākara of Siṃhabhūpāla, a fourteenth-century king of the Recharla dynasty in Telangana who also wrote on drama, and the Kalānidhi of Kallinātha, written at Vijayanagara under Devarāya II (r. 1424–1446), explain the text line by line and discuss its difficulties. Kallinātha in particular was aware that the rāgas of his own day often differed from those described by Śārṅgadeva, and his comments are among the earliest explicit discussions of the gap between lakṣaṇa and lakṣya. Both commentaries are printed with the text in the Adyar edition, the standard Sanskrit text, and the edition and translation by R. K. Shringy and Prem Lata Sharma makes the work accessible in English (Subrahmanya Sastri 1943–1953; Shringy and Sharma 1978). The commentaries also show how quickly the text became a classic: within two centuries of its composition, a work that had recorded the music of the Yādava court was being read as a description of timeless principles.

6.6 After Śārṅgadeva: Regional Treatises, c. 1300–1800

6.6.1 North and south

After the thirteenth century the literature divided along the lines of the music itself (Chapter 4). In the north, Sanskrit treatises continued to be written at Rajput, regional sultanate and Mughal courts, but they increasingly shared the field with works in Persian and in Hindi; in the south, the courts of Vijayanagara and its successors produced the treatises that led to the mela system. Table 6.1 lists the main works of the period from the twelfth century to the early nineteenth; the dates of several are uncertain.

Table 6.1. Some treatises on music after Mataṅga (dates approximate)
Work Author or patron Place and date
Bharatabhāṣya Nānyadeva Mithilā, c. 1100
Mānasollāsa Someśvara III Kalyāṇa (Deccan), 1131
Saṅgītasamayasāra Pārśvadeva 12th–13th c.
Saṅgītaratnākara Śārṅgadeva Devagiri, 1st half of 13th c.
Saṅgītopaniṣatsāroddhāra Sudhākalaśa western India, 1350
Sudhākara (commentary on the Saṅgītaratnākara) Siṃhabhūpāla Telangana, 14th c.
Saṅgītaśiromaṇi compiled by scholars for a Muslim ruler north India, 1428
Kalānidhi (commentary on the Saṅgītaratnākara) Kallinātha Vijayanagara, c. 1430–1446
Saṅgītarāja Kumbhakarṇa Mewar, mid-15th c.
Rāgataraṅgiṇī Locana Mithilā, date disputed
Svaramelakalānidhi Rāmāmātya Vijayanagara, 1550
Ṣaḍrāgacandrodaya, Rāgamañjarī Puṇḍarīka Viṭṭhala Khandesh and Amber, late 16th c.
Rāgavibodha Somanātha Andhra, 1609
Saṅgītadarpaṇa Dāmodara Miśra north India, early 17th c.
Caturdaṇḍīprakāśikā Veṅkaṭamakhin Thanjavur, mid-17th c.
Saṅgītapārijāta Ahobala north India, 17th c.
Anūpasaṅgītaratnākara and other works Bhāvabhaṭṭa Bikaner, late 17th c.
Saṅgītasārāmṛta Tulajā Thanjavur, c. 1730
Naghmāt-i Āṣafī (Persian) Muḥammad Riẓā Patna, 1813

6.6.2 The northern compendia

The fourteenth and fifteenth centuries produced several large compilations in the tradition of the Saṅgītaratnākara. The Jain author Sudhākalaśa wrote his Saṅgītopaniṣatsāroddhāra in 1350 as an abridgement of his own larger work, with accounts of rāgas, prabandhas and dances. The Saṅgītaśiromaṇi, compiled in 1428 by a group of scholars for a Muslim ruler in north India, arranges the received doctrine in a clear and systematic way — a Sanskrit compendium in the line of Śārṅgadeva made for a sultan’s court, and one more sign that Hindu, Jain and Muslim patrons shared a learned musical culture — and Emmie te Nijenhuis’s edition and translation of its musical portions has made it one of the most accessible of the medieval handbooks (Nijenhuis 1992). The Saṅgītarāja, ascribed to Rāṇā Kumbhakarṇa (Kumbha) of Mewar (r. 1433–1468), a king who also wrote a commentary on the Gītagovinda, is the most ambitious of them: a vast encyclopaedia of recitation, song, instruments, dance and rasa that draws heavily on Śārṅgadeva and on many other sources (Nijenhuis 1977).

These works were monuments of court learning, and they say less about the music actually played at the courts of their time than their size suggests. The music that was transforming north India — the songs of the Sufi assemblies, the new genres and instruments of the sultanate courts, and the dhrupad of Gwalior — appears in them only indirectly. For these one has to turn to works in other languages. The Hindi treatise associated with Mān Siṃh Tomar of Gwalior, the Mānakutūhala, is lost in its original form and known chiefly through Faqīrullāh’s Persian adaptation, the Rāg Darpan of 1666, which is discussed with the other Persian writings in Chapter 7.

A northern work of a different kind is Locana’s Rāgataraṅgiṇī, from Mithilā. It classifies the rāgas of its time under twelve parent scales (saṃsthāna), in a way that anticipates the later thāṭs, and quotes songs of the Maithili poet Vidyāpati (fourteenth–fifteenth century) as examples. Its date is disputed: earlier writers placed it in the twelfth century, but the quotations from Vidyāpati show that it cannot be earlier than the fifteenth, and some scholars place it in the seventeenth. Whatever its date, it shows that a classification of rāgas by scale was being attempted in the north as well as the south.

6.6.3 The mela in the south

In the south the decisive step was taken at Vijayanagara. Rāmāmātya, a minister at the court of Rāma Rāya, wrote his Svaramelakalānidhi in 1550. He complained that the theorists of his time described a music that no one played, and set out instead to describe the rāgas of current practice. His method was to group them by their parent scale (mela), which he explained in terms of the frets of the vīṇā: a set of rāgas that could be played on the same arrangement of frets belonged to the same mela. He recognised twenty melas. The classification grew out of the practice of the fretted vīṇā, then the principal art instrument of the south, and it subordinated the ancient categories of grāma and jāti to the organisation of real scales (Nijenhuis 1977; Widdess 1995). Later southern writers cite a lost Saṅgītasāra ascribed to Vidyāraṇya, the fourteenth-century scholar associated with the founding of Vijayanagara, as an earlier authority for the mela idea; the work does not survive and its contents cannot be checked.

The idea spread quickly. The Rāgavibodha (1609) of Somanātha, a scholar from Andhra, classified rāgas under a larger number of melas and gave its examples in a notation for the vīṇā with signs for ornaments. At Thanjavur, the Saṅgītasudhā, attributed to the Nāyaka king Raghunātha and probably written by his minister Govinda Dīkṣita, and then the Caturdaṇḍīprakāśikā of Govinda’s son Veṅkaṭamakhin carried the theory further. Veṅkaṭamakhin’s demonstration that seventy-two melas are mathematically possible, and the completion of the named scheme by Govindācārya in the eighteenth century, are described in Chapter 8. The title of Veṅkaṭamakhin’s work refers to the “four pillars” (caturdaṇḍī) of the music of his time — ālāpa, ṭhāya, gīta and prabandha — a reminder that even this most theoretical of works set out to describe the forms of a living practice.

The court of the Maratha kings of Thanjavur (1676–1855) continued the tradition. The Saṅgītasārāmṛta attributed to King Tulajā (r. c. 1729–1735) describes rāgas with their characteristic phrases and ornaments as well as their scales, and gives an account of the dance of the court (Chapter 12); it is one of the main sources for southern music in the generation before the Trinity (Subramanian 2006).

6.6.4 Bridges between north and south

Several authors of the late sixteenth and seventeenth centuries carried southern ideas north. The most important was Puṇḍarīka Viṭṭhala, a scholar from Karnataka who worked first at the court of the Fārūqī rulers of Khandesh, for whom he wrote the Ṣaḍrāgacandrodaya, and then for the Kachhwaha rulers of Amber, for whom he wrote the Rāgamālā, the Rāgamañjarī and the dance treatise Nartananirṇaya. His works apply the southern method of classification by mela to the rāgas of the north, and record northern names and forms of the period of Akbar; the Muslim and Rajput courts for which he wrote show how freely learning moved between Muslim and Hindu patrons (Nijenhuis 1977).

Ahobala’s Saṅgītapārijāta, of the seventeenth century, is notable for defining the positions of the twelve notes by the length of the string of a vīṇā at which each is produced, the first attempt in the Sanskrit literature to fix the pitches of the notes by measurement. It was widely copied in the north, and its method was taken up by the Gond king Hṛdayanārāyaṇa of Garha-Mandla in his Hṛdayakautuka and Hṛdayaprakāśa. Dāmodara Miśra’s Saṅgītadarpaṇa (early seventeenth century), by contrast, is largely an abridgement of the Saṅgītaratnākara, but it became one of the most widely copied of all Sanskrit music texts, partly because of its verses describing the rāgas and rāgiṇīs as persons, which fed the tradition of rāgamālā painting (Chapter 4).

At Bikaner, Bhāvabhaṭṭa, the son of a musician from the Mughal court, wrote several works for Mahārāja Anūp Siṃh (r. 1669–1698), among them the Anūpasaṅgītaratnākara, the Anūpasaṅgītavilāsa and the Anūpāṅkuśa. They are compilations, but they show the Rajput courts of the Mughal period continuing the Sanskrit tradition alongside the Persian and Hindi music culture of the imperial court, and the library that Anūp Siṃh assembled at Bikaner preserved many manuscripts on music that would otherwise have been lost.

6.6.5 The fretted vīṇā and the theory of scales

Many of the theoretical innovations of this period were made by theorists who were also players of the fretted vīṇā, and their theory bears the marks of the instrument. Already the Nāṭyaśāstra had demonstrated the śrutis with a pair of vīṇās (Chapter 4), but the harp it used had open strings, one for each note. The stick zithers and lutes of the medieval period had frets, and on a fretted instrument a scale is a physical arrangement: a set of positions on the neck, fixed for as long as the frets are not moved. Rāmāmātya’s melas were, in the first place, arrangements of the frets of the vīṇā, and his classification grouped together the rāgas that could be played without moving them. Somanātha’s notation in the Rāgavibodha records the strokes and finger movements of the vīṇā player, defining ornaments in terms of what the hand does on the string, and Ahobala’s method of locating the notes by measuring the length of the string is a description of where to place the frets. In the south, the Thanjavur tradition credits the Nāyaka king Raghunātha (early seventeenth century), the patron of Govinda Dīkṣita, with a vīṇā whose frets were fixed for all twelve positions of the octave, so that any rāga could be played without re-fretting; the modern Sarasvatī vīṇā, with its twenty-four fixed frets over two octaves (Chapter 10), is the result of this development, and Veṅkaṭamakhin’s demonstration of the seventy-two possible melas presupposes exactly such a fretboard of twelve positions.

This connection helps to explain both the power and the limits of the scale-based classifications. A fretted instrument makes the selection of notes visible and countable, and so encourages a theory of rāga as a choice among scales. But the rāgas themselves, as singers and players knew, were defined as much by phrases, ornaments and intonations that no fret fixes: the oscillating Ga of Tōḍi or the slow āndolan of Darbārī are made by bending the string or the voice between the frets. The mela and thāṭ systems are thus, in part, theories of an instrument applied to a music that was, above all, vocal — a point that modern musicians, who use the classifications for orientation but learn rāgas from compositions and teachers, have always understood (Chapter 4).

6.6.6 Persian and vernacular theory

From the fourteenth century the Sanskrit treatises are only part of the story. The Persian literature on Indian music (Chapter 7) began as translation and explanation of the Sanskrit theory for Persian-reading patrons, but it became a tradition in its own right, describing musicians, genres and repertoires that the Sanskrit texts ignored. In the vernaculars, works in Hindi, Bengali, Telugu, Marathi and other languages explained rāga and tāla to new readers, often in the form of song collections with theoretical introductions. One Persian work marks a turning point in the theory itself. In the Naghmāt-i Āṣafī (1813), Muḥammad Riẓā, a nobleman of Patna, examined the traditional families of six rāgas and their rāgiṇīs and concluded that they had no basis in the music of his time; he proposed instead to classify rāgas by their scales, taking the scale of Bilāval, with all its notes natural, as the standard. His proposal anticipates by a century the thāṭ system of Bhatkhande, who knew his work (Bor et al. 2010; Schofield 2010).

6.7 From Prabandha to Dhrupad and Kṛti

The treatises describe theory; they also allow the historian to follow, at least in outline, how the forms of composed song changed between the age of the prabandha and the modern repertoire. The story is not one of simple descent, since the medieval prabandhas themselves are lost and the names of forms often outlived the forms, but certain principles persist.

The first is the refrain. Śārṅgadeva’s dhruva, the section of a prabandha that must never be omitted, became in the following centuries the organising principle of vernacular devotional song. The songs of Jayadeva’s Gītagovinda (twelfth century) each have a refrain, marked in the manuscripts as dhruva, to which the singer returns after every stanza; the pads of the north Indian saint-poets, the abhaṅgs of Maharashtra and the kīrtanas of the Haridāsas and of Annamācārya are all built on a refrain and a series of stanzas. In the north, the dhrupad (dhruvapada, “refrain-verse”) that emerged at the courts of the fifteenth century took its name from this principle and kept a four-part form whose last section, the ābhog, recalls the terminology of the prabandha (Chapter 7).

The second is the movement from text to music. The devotional song of the fifteenth and sixteenth centuries was text-centred: its music served the words, and a singer might set the same words to different melodies. In the south, the songs of the Tāḷḷapāka composers, Annamācārya and his descendants at Tirupati, already show the division into a refrain (pallavi) and stanzas (caraṇa), and a Telugu work on the composition of such songs, the Saṅkīrtanalakṣaṇamu, is attributed to the family. The Haridāsas of Karnataka composed, besides their kīrtanas, the sūḷādi, a long devotional composition passing through a series of sections in different tālas, whose name links it to the sālagasūḍa of the treatises. Over the following centuries, and decisively in the generation of the Karnatak Trinity, the kīrtana became the kṛti, in which the music is as important as the text, the melody is fixed by the composer, and the composition provides the material for elaboration (Chapter 8).

The third is the place of the solfa and the drum syllable. The prabandha theory’s svara and pāṭa limbs anticipate the passages of solfa in the varṇam, the svarajati and the ciṭṭasvara of the kṛti, and the combination of drum syllables and meaningless vocables in the tillāna and tarānā; the dance traditions, which kept the recitation of drum syllables as part of their technique, preserved this element most fully (Chapter 13). The fourth, finally, is the ideal of the vāggeyakāra, the composer of both words and music, which the treatises formulated and which the great composers of both traditions — the Haridāsas, the dhrupad composers of Gwalior, the Trinity, Svāti Tirunāḷ — realised.

These continuities should not be overstated. The modern kṛti and khayāl are not prabandhas, and the treatises after Śārṅgadeva often describe the old forms long after they had disappeared, while saying little about the genres actually sung in their authors’ time. But the analytical vocabulary of the treatises shaped the way musicians named and thought about the parts of a composition, and it allowed the new forms of each age to be understood, by their makers and patrons, as continuations of an ancient art.

6.8 Colonial Encounters and the Making of Modern Theory

6.8.1 Orientalists and their informants

European interest in the theory of Indian music began with Sir William Jones’s essay “On the Musical Modes of the Hindus”, written in 1784 and published in Asiatic Researches in 1792. Jones read the Sanskrit and Persian treatises with the help of Indian scholars and tried to relate their rāgas to the modes of ancient Greek music; like many later Orientalists, he looked for the “true” Indian music in the ancient texts rather than in the music of contemporary performers. N. Augustus Willard, an officer in the service of the Nawab of Banda, took the opposite view in his Treatise on the Music of Hindoostan (1834), insisting that the texts must be judged against the practice of living musicians. The tension between the two approaches, textual and practical, ran through all later writing (Farrell 1997).

In Calcutta in the later nineteenth century, Indian scholars took up the subject on their own terms. Sourindro Mohun Tagore (1840–1914) founded the Bengal Music School (1871), published collections of Sanskrit and other texts on music and books on Indian instruments, and presented Indian music to European audiences as a learned art comparable to that of Europe. In the early twentieth century the question of the twenty-two śrutis became a subject of intense debate. K. B. Deval’s The Hindu Musical Scale and the Twenty-two Shrutees (1910), E. Clements’s Introduction to the Study of Indian Music (1913) and A. H. Fox Strangways’s The Music of Hindostan (1914) tried, by measurement and by reasoning from the texts, to establish the exact sizes of the śrutis; the debate was never resolved, and the modern understanding of śruti as a flexible framework rather than a fixed set of pitches is described in Chapter 4.

6.8.2 Bhatkhande and the new śāstra

Vishnu Narayan Bhatkhande (1860–1936) brought the history of the treatises to its modern conclusion. Between about 1904 and 1909 he travelled through southern, western, central and eastern India, collecting manuscripts, interviewing musicians and meeting scholars, among them Subbarāma Dīkṣitar at Ettayapuram. His study of the Sanskrit texts led him to a conclusion that earlier writers had avoided stating plainly: the theory of the ancient and medieval treatises could not be applied to the music of his day. The grāmas and jātis had disappeared; the śuddha (natural) scale of the texts, built on the ṣaḍja-grāma, was not the śuddha scale of modern Hindustani musicians, which is the scale of Bilāval; and the rāgas described under familiar names in the medieval works often had little in common with the rāgas that bore those names in practice (Bhatkhande 1934; Bakhle 2005).

His response was to write a new śāstra. In the Lakṣyasaṅgītam (1910), a Sanskrit treatise in verse published under the pen-name “Catura Paṇḍita”, he described the music of current practice in the traditional form and language of the texts, as if to give it a classical authority of its own; the Abhinavarāgamañjarī, also in Sanskrit, set out the rāgas. His Marathi Hindustānī Saṅgīt Paddhati (four volumes, 1910–1932) explained the theory in the form of dialogues, and the six volumes of the Kramik Pustak Mālikā gave the compositions he had collected in notation (Chapter 4). At the heart of the system were the ten thāṭs — Bilāval, Kalyāṇ, Khamāj, Bhairav, Pūrvī, Mārvā, Kāfī, Āsāvarī, Bhairavī and Toḍī — under which he classified the rāgas of the north, together with assignments of vādī and saṃvādī and a theory of the time of performance. His lecture to the first All-India Music Conference at Baroda in 1916, later published as A Short Historical Survey of the Music of Upper India, set out his reading of the history of the treatises (Bhatkhande 1910; Bhatkhande 1934).

Bhatkhande’s achievement was immense, and his thāṭs remain the framework in which Hindustani rāgas are taught. Its costs, discussed in Chapter 7, lay in the way it subordinated the knowledge of hereditary musicians to a textual system, and in its claim to finality. Musicians have always pointed out that many rāgas fit uneasily into the ten thāṭs, that the theory of vādī and saṃvādī misdescribes some of them, and that a rāga is defined by its phrases rather than its scale. In this respect Bhatkhande stands where Rāmāmātya and Muḥammad Riẓā stood before him: a theorist who, faced with the gap between the books and the music, chose to describe the music.

The south had no single figure comparable to Bhatkhande, in part because its mela system already offered a classification that connected theory and practice. Its modern theory was shaped instead by publications and institutions. Subbarāma Dīkṣitar’s Saṅgīta Sampradāya Pradarśini (1904) set out the rāgas in the tradition of Veṅkaṭamakhin together with the compositions that exemplified them (Chapter 4); Abraham Pandither’s Karuṇāmirta Sāgaram (1917) argued for a Tamil origin of the southern system (Chapter 8); and from 1929 the Music Academy of Madras held annual conferences at which experts debated the characteristics (lakṣaṇa) of particular rāgas, the number and size of the śrutis and the readings of the old texts, publishing their papers and conclusions in the Academy’s journal. These debates turned the old relation between lakṣaṇa and lakṣya into a modern institutional practice, in which a committee of musicians and scholars, rather than a royal author, pronounced on the correct form of a rāga (Weidman 2006; Subramanian 2006).

6.9 Reading the Treatises Today

The modern scholarly study of the treatises began with the editions of the early twentieth century, many of them published in the series of the princely and institutional libraries of Trivandrum, Baroda and Adyar, and was transformed in the second half of the century by the work of scholars such as V. Raghavan, Prem Lata Sharma, R. K. Shringy, Mukund Lath, Emmie te Nijenhuis, Lewis Rowell and Richard Widdess (Nijenhuis 1977; Rowell 1992; Widdess 1995; Powers and Widdess 2001). Their work suggests several principles of interpretation.

First, the texts must be read historically. A statement in a seventeenth-century treatise may be a copy of a thirteenth-century source describing music that had long since changed; the most valuable passages are often those in which an author departs from his source to describe something new, as Śārṅgadeva did with his “current” rāgas, Rāmāmātya with his melas and Muḥammad Riẓā with his critique of the rāga families.

Secondly, the notations and the technical descriptions can be used as evidence for actual music, but only with care. Widdess’s analysis of the notated examples in Mataṅga, Nānyadeva and Śārṅgadeva showed that the early rāgas can be partly recovered, but the manuscripts are often corrupt, and the reconstructions depend on assumptions about intonation, ornament and rhythm that the texts do not fully specify (Widdess 1995).

Thirdly, the treatises are evidence for the social history of music as well as its theory. They show which arts were valued at which courts, how authors claimed the authority of the past, how Jain, Hindu and Muslim patrons and scholars shared a common learned culture, and how the theory of music was used to give prestige to new repertoires. Their silences are evidence too: the absence of women and low-caste musicians from the authorship of the texts, and the near-absence of the music of ordinary people except as deśī material to be classified, tell us whose knowledge the literature recorded.

Finally, the treatises are not the music. Indian musicians have always learned from teachers, not from books, and many great musicians of the past three centuries could not have read a Sanskrit treatise. The literature records how a learned elite understood and organised the music it patronised; the music itself lived in the practice of performers, whose knowledge reached the written record only partially, and often only through the books of reformers such as Bhatkhande. To read the treatises well is to read them alongside that practice, as the best of their authors did.

Further Reading

Nijenhuis (1977) is the standard survey of the Sanskrit literature on music, and Rowell (1992) the best introduction to its thought. Widdess (1995) analyses the early rāgas and notations, and Powers and Widdess (2001) surveys theory and practice. For the texts see Lath (1978) and Nijenhuis (1970) (Dattilam), Sharma (1992) (Bṛhaddeśī), Shringy and Sharma (1978) (Saṅgītaratnākara) and Nijenhuis (1992) (Saṅgītaśiromaṇi). On Bhatkhande and the modern reorganisation of theory see Bhatkhande (1934) and Bakhle (2005), and on the Persian literature Schofield (2010).