Ambarish Sridharanarayanan

Personal website + blog of Ambarish Sridharanarayanan.

The mēḷakartā scheme considered harmful


The mēḷakartā scheme

Carnatic music is defined by its usage of rāgas. Each rāga can invoke specific emotions in rasikas, and each rasika will have their set of favourite (and disliked) rāgas. Both the kalpitasaṅgīta and manōdharmasaṅgīta aspects of music adhere to the underlying rāgalakṣaṇa, or the grammar of the rāga. With few exceptions, each piece is composed in one prescribed rāga, and that rāga constrains the music in that piece. Transgressing the boundaries of a rāga by mistake is an egregious error, and such an apasvara would be spotted immediately by many rasikas.

Much music theory is devoted to classifying these rāgas. The predominant scheme used to today is the mēḷakartā scheme, involving 72 mēḷakartā, or janakarāgas, and all other janyarāgas classified “under” them. Ramamatya (circa 1550 CE), in his svaramēḷakalānidhi, proposed a mēḷa‐based scheme, which Venkatamakhin (17th c. CE) extended into a mēḷakartā scheme in his chaturdaṇḍiprakāśikā. The scheme followed today is a standardised variant proposed by Govindacharya (17th–18th c. CE), in his saṅgrahacūḍāmaṇi.

How it works

Twelve svarasthānas are enumerated, from the base ṣadja to the ṣadja of the next octave (left‐inclusive), and are given the following names, picked from an earlier nomenclature.

  • Sa
  • Ri1
  • Ri2 or Ga1
  • Ri3 or Ga2
  • Ga3
  • Ma1
  • Ma2
  • Pa
  • Dha1
  • Dha2 or Ni1
  • Dha3 or Ni2
  • Ni3

The mēḷakartās, then, are all the possible rāgas that include the Sa & the Pa, and have one each Ri, Ga, Ma, Dha & Ni, with the natural restriction that the Ri chosen should be lower than the Ga, and likewise the Dha and the Ni. There are 6 possible Ri‐Ga combinations (Ri1‐Ga1, Ri1‐Ga2, Ri1‐Ga3, Ri2‐Ga2, Ri2‐Ga3 & Ri3‐Ga3), likewise 6 combinations of Dha‐Ni, and 2 options for Ma. There are thus 72 mēḷakartās, also termed sampūrnarāgas as each such rāga has all 7 notes.

Every other rāga, which does not adhere to the above lakṣaṇa, is then notated as a pair of scales of the above 12 notes – an ārōhaṇa and an avarōhaṇa – that define them. The ārōhaṇa is a scale that starts at a lower Sa and ends at the next Sa and is meant to be the set of all possible ascending sequences of notes valid in the rāga. Likewise, the avarōhaṇa, starting at a higher Sa and ending at the previous Sa is meant to include all possible valid descending note sequences. The janyarāga is then classified as a derivative the mēḷakartā rāga.

For instance,

  • Arabhi is considered to be the scale Sa‐Ri2‐Ma1‐Pa‐Dha2‐Sa; Sa‐Ni3‐Dha2‐Pa‐Ma1‐Ga3‐Ri2‐Sa, and is thus classified under mēḷakartā Sa‐Ri2‐Ga3‐Ma1‐Pa‐Dha2‐Ni3‐Sa; Sa‐Ni3‐Dha2‐Pa‐Ma1‐Ga3‐Ri2‐Sa, which happens to be Sankarabharanam.
  • Purvikalyani, which is assigned the scale Sa‐Ri1‐Ga3‐Ma2‐Pa‐Dha2‐Pa‐Sa; Sa‐Ni3‐Dha2‐Pa‐Ma2‐Ga3‐Ri1‐Sa, is considered a derivative of mēḷakartā Sa‐Ri1‐Ga3‐Ma2Pa‐Dha2‐Ni3‐Sa; Sa‐Ni3‐Dha2‐Pa‐Ma2‐Ga3‐Ri1‐Sa or Gamanasramam.
  • Bhairavi, nominally the scale Sa‐Ga2‐Ri2‐Ga2‐Ma1‐Pa‐Dha2‐Ni3‐Sa; Sa‐Ni3‐Dha1‐Pa‐Ma1‐Ga2‐Ri2‐Sa (note the 2 different Ga‐s), is classified under Sa‐Ri2‐Ga2‐Ma1‐Pa‐Dha1‐Ni3‐Sa; Sa‐Ni3‐Dha1‐Pa‐Ma1‐Ga2‐Ri2‐Sa, Natabhairavi, with the Dha2 considered an anyasvara.


In pedagogy and in musicology, a rāga is introduced as a child of a mēḷakartā and then in terms of its ārōhaṇa and avarōhaṇa. An excerpt from a songʼs lyric page follows:

raagam: bEgaDa
29 dheera shankaraabharaNam janya
Aa: S G3 R2 G3 M1 P D2 N2 D2 P S
Av: S N3 D2 P M1 G3 R2 S

Rasikas in the know seek to learn which mēḷakartā a rāga belongs to, and attempt to understand the rāga based on this scheme.

Unfortunately, this entire exercise is pointless.

Disconnect between definition and notation

Inherently, what defines a rāga is not the set of notes but the phrases. The notes used to describe the phrases, let alone the ārōhaṇa/avarōhaṇa, are but notational conveniences. Notice for instance, that Atana is notated with a Ga, but the Ga sung in Atana is an oscillation between Ri2 and Ma1. Likewise in Darbar, where the Ga a different oscillation, also between Ri2 and Ma1. A Pa‐Dha2‐Sa phrase from Suddhasaveri and one from Yadukulakambhoji are identical to a Pa‐Ni2‐Sa phrase from Madhyamavati and one from Dhanyasi. And a classic phrase from Anandabhairavi can be annotated as both Pa‐Dha2‐Pa‐Sa or Pa‐Ni2‐Pa‐Sa!

The result is a disconnect between rāgalakṣaṇa and their notation, leading to a huge disconnect between many classic janyarāgas and their nominal parents. None of Sahana, Khamas, Kambhoji, Natakurinji, Suratti and Mohanam have anything to do with any other, but for the fact that they are classified under Harikambhoji. Likewise, Reetigaula, Bhairavi, Darbarikanada, Hindolam, Suddhadhanyasi are all classified under Natabhairavi, albeit none of them sound like any other. So what use is such a system that emphasises the names of the notes?

To quote musicologist Dr. N Ramanathan,

People exclaim (the song) Gajavadana, set in Todi, has so many different kinds of Ga‐s; thatʼs not the case. How remarkable that various different phrases are all termed Ga! We use a single symbol Ga to identify many of the phrases. That, I feel, is a very sophisticated scheme which in the 1800s or 1700s we (Carnatic music practitioners) have developed. The word svara itself has been redefined. Svara is no longer associated with a pitch – svara is a phrase. We say that Sankarabharanam has 4 kinds of Ma‐s; Begada has 3 kinds of Ma‐s. No, we term all those phrases Ma.

Any system that emphasises notational conveniences at the expense of the lakṣaṇa of the rāga is missing the point.

Inapplicability to many rāgas

Rāgas inherently are not a taxonomy, and trying to fit them into a hierarchical structure is futile. There are undoubtedly relationships among some rāgas: Mukhari, Bhairavi, Huseni and Manji have some common phrases. But while as above Bhairavi is classified under Natabhairavi, people think Huseni sounds somewhat like Anandabhairavi, which is classified under Kharaharapriya. Our taxon is very unwieldy.

The janyarāga examples above – Arabhi, Purvikalyani, Bhairavi – happen to contain all the saptasvaras. What if the rāga does not? There would be multiple mēḷakartā super‐sets of this rāga, so which one do you pick as parent? There are 2 systems in usage:

  • Pick the earlier one: if the janyarāga does not contain say, Ri, then pick the mēḷakartā with Ri1 rather than Ri2 or Ri3. This arbitrary choice only helps derail the already tenuous janakajanya relationship.
  • Pick the mēḷakartā that sounds “closest”. This choice is also arbitrary, as many janyarāgas donʼt sound like any of the candidate parent mēḷakartā rāgas. Madhyamavati can be assigned to Natabhairavi, Kharaharapriya, Charukesi or Harikambhoji on this basis, but sounds like none of them.

Many rāgas, not amenable to being described as a linear progession of notes, are strait‐jacketed as they are fit into an ārōhaṇa/avarōhaṇa scheme. Atana, Suratti, Sindhubhairavi, Purvikalyani, Bhairavi, Kambhoji all include phrases that cannot be fit into a linear scale. Atana and Suratti are especially problematic: no ārōhaṇa/avarōhaṇa written for them cover even a portion of the phrases possible. Finding a mēḷakartā parent is futile.

Lastly, bhāṣāṅgarāgas exhibit different problems of classification, as its sometimes unclear which svaras belong in the scale and which are anyasvaras. Kapi, Sindhubhairavi and Brindavani exemplify this problem.


Generations of students has grown up focussing on learning the wrong bases of rāgas, the foundation of Carnatic music. Generations of rasikas has been cheated out of appreciating the grain of rāgabhāva, by concentrating on the chaff of mēḷakartās. A chance of a deeper understanding of Carnatic classical music has been lost on account of trying to marry Indian music ideas to European music concepts of pitches, notes and scales.

The path forward

Pedagogy must focus on learning rāgabhāva through saṅgatis and lakṣaṇa without focussing on ārōhaṇa/avarōhaṇa. And rasikas should be steered clear of trying to enjoy a rāga through learning scales.


Term Meaning
anyasvara Exceptional svara, absent in the parent mēḷakartā
apasvara Grave svara error
ārōhaṇa Ascending scale
avarōhaṇa Descending scale
bhāṣāṅgarāga Rāgas with one ore more anyasvaras
janaka Parent
janya Child
kalpitasaṅgīta Precomposed music
lakṣaṇa Grammar
manōdharmasaṅgīta Improvised music
mēḷam Musical scale
mēḷakartā Mēḷam‐creator
rāga Melodic mode
rāgabhāva The emotionss associated with a rāga
rāgalakṣaṇa The lakṣaṇa associated with a rāga
sampūrna Complete
saṅgati A musical phrase associated with a song
saptasvara The 7 svaras in an octave
ṣadja The base svara of an octave
svara Ornamented note
svarasthāna Note