Subject: Alaapana - a gentle introduction
Date: 3 May 1998 07:16:38 GMT
From: "Shivkumar Kalyanaraman" <>

Alaapana, a gentle introduction

By Shivkumar Kalyanaraman

I have been meaning to write a series of articles on manodharma sangita dubbed gentle introductions", taking after the lead of my well-known friend, Ramesh Mahadevan who wrote a few "gentle introduction" articles on carnatic music. Interested folks point your browser to
to read all his must-read articles. I would recommend you read those articles before this one if you are not familiar with some of the common terms in carnatic music.

Ok, here is a first shot at alaapana.

Alaapana is perhaps the most celebrated and well known components of manodharma (or "improvisation") in Carnatic music. My goal is to give the reader a concrete sense of what an alaapana is, how one should listen to alaapana, and how one can attempt alaapana ... Of course I will try to make the reading a little more interesting by giving examples of masters in various aspects of alapana.

What is alaapana ?

An alaapana is an abstract essay of a raaga. The word 'alap' means "spread" [1], which indicates that the purpose of the alaapana is to elaborate or "spread out" the raaga. Which begs the question, what is a raaga, and how does one go about "spreading it out" ?


The first step towards defining a raaga is laying out its arohana and avarohana. Arohana is the basic sequence of ascending notes and avarohana is the basic sequence of descending notes. An example of arohana/avarohana of Mohanam:

Aro: s r g p d s
Ava: s d p g r s

The avarohana need not have the same notes as the arohana, and both can criss/cross a little though they have to be ultimately ascending/descending by definition. The position/frequency of each note ("swarasthana") is one of 12 absolute possibilities. In fact the choice is more restricted than that. There is no choice for "sa" and "pa". For every other note there is two options to choose from. So, lets for the moment assume that this straightforward choice has been made...


The next aspect of a raga is the set of "permissable gamakas" associated with the arohana and avarohana. Gamaka is the collective term given to the various shakes, graces, ornaments and embellishments used in Indian music. In more simple terms, for every note in the arohana and avarohana, there is a set of shakes/graces which are common. What shakes/graces are allowed and what are not is in general defined by aesthetics. But some rules are clear - no gamakas for "sa" and "pa".

For example, the default gamaka for "ri" in Mohana is as follows. Start from "sa". Utter the syllable "ri", but slide from "sa" to "ga" and then back to "ri". That is, you have actually sang "sa-ga-ri" but uttered only one syllable "ri". Another key aspect is that you don't jump discretely from "sa" to "ga" to "ri" - you slide through all the

frequencies in between. This whole thing fits into the time-period for one note i.e. "ri". Similarly the default gamaka for "ga" is just a flat "ga"; for "dha" is "pa-sa-dha". Recall that "sa" and "pa" are also flat.

These are the default gamakas in the ascent (arohana). The corresponding ones for the descent are: "dha" = "sa-dha"; "ga" = "pa-ga"; "ri" = "ga-ri". Now we can replace the flat notes in the arohana/avarohana with these default gamakas. You can almost experience a change in the mood as you transition from a "flat" aro/avaro to a "gamaka-filled" aro/avaro. Try it !

Note that these are the "default gamakas" for the raga, which one normally learns before learning any song. However, depending upon context (and aesthetics), there are other gamakas which are possible for each note. For example, you could hold "ri" flat (esp. used when you are whizzing through "ri"), or use "sa-ga-ri-ga-ri" (esp. used when you are going very slow and you have ample time for saying "ri"). Technically, the janta swaram "ri-ri", and flashing "ga-ri-ri" are also gamakas for "ri"; but aesthetics says that you cannot arbitrarily use them. Let me stop here about gamakas and let you discover other possibilities which can appear in a alaapana.

Alphabet, Phrases, Sentences in a raga

The arohana/avarohana and the "permissable" gamakas define the "ABCD" or alphabet of a "raga", using which you construct the essay of alaapana. In other words the gamaka-filled aro/avaro to "raga" is what the alphabet is to a language.

Now, you don't directly learn essay writing in English after you learn ABCD right ? You need to first learn how to construct "phrases", "sentences" and groups of sentences called "paragraphs". The notion corresponding to a "phrase" in carnatic music is called the "sangati". Example phrases (sung with the default gamakas for each note): "sa-ri-ga-pa", "ga-pa-ga-ri-sa", "ga-ga-ri-ri" (the beginning of ninnukori varnam) etc

We have been assuming that each note is of equal length in a phrase. Not necessarily! You get another dimension by varying the length of notes in a phrase. For example (capitals imply notes spanning two time units): "sa-ri-ga-PA", "ga-pa-ga-ri-SA", "GA-GA-RI-RI".

You could have a set of sangathis in slow speed only - MD Ramanathan (MDR) was a singer who used to sing a majority of his sangathis in slow speed only. Typically, singing in slow speed allows one to clearly articulate the gamaka {which is called a "weighty" exposition of the gamaka/sangathi. Note that "weight" has more to do with a clear articulation of a gamaka (and punctuation between sangathis - described elsewhere) and less to do with having a deep voice, though the latter is also a definite asset}.

You could also have notes spanning fractional units of time (1/2, 1/4, 1/3 etc) - such sangathis are called "briga" sangathis. GN Balasubramanium (GNB) was a master of briga. Control of the length of notes and balancing them to form a meaningful sangati is challenging. More than that, stringing together briga sangathis with poise is a sign of a master. Shri Kalyanaraman, a disciple of GNB was known to be able to fit different kinds of briga sangathis within a unit time measure.


Back to the story of "sangathis", not all sangathis are equal (or even meaningful)... So, how does one learn to sift out "good" sangathis from the bad ones ? How does one form attractive musical phrases ? How does one string them to form a garland of phrases ?

Firstly let us agree that sangathis are every where - you yourself come up with them unknowingly. Any bathroom singer can attest to this. Perhaps the hum of the shower and the confines of room leads to special acoustics and the absence of worries creates a special ambience... I wish I could do a PhD studying this. Alas, the pressures of econmics against this are too much ! Forget bathroom singers. Even if you are a occasional concert goer and have a liking for the sabha canteen, haven't you experienced musical ideas popping up in your head ?

However, these are transient emotions and unless we come up with a method of capturing that information for posterity, we have to methodically search for good sangathis... There are several techniques which can be used for this.

One simple path is to learn a varnam, a krithi or even a geetham in the raga (but with the gamakas - not those flat geethas!). Learn it well, sing it in public gatherings and win praise... But then bring it back to the drawing board if you want to explore alaapana in that ragam. A simple technique is to then mark off the set of musical phrases corresponding to a word in the sahityam (text). For example, the word "Ninnukori" in the Mohanam varnam has the phrases: "GA-GA-RI-RI-sa-sa-ri-ri-ga-ga-ri-ri". Then the next exercise is to throw out the sahityam ("ninnukori") and replace it by "aa", and sing this phrase within a single breath (or a single bow for a violinist). The sangathi thus sung is called an "akaara". Practise this for every word in the krithi/varnam - it is a treasure of good sangathis.

An akaara is the way a sangathi appears in a alaapana. We only use the syllables "aa", "tha", "dha", "ri", "na" etc in an alaapana. We now know one secret - phrases can be simply picked off the krithi just as flowers are picked off a flower plant. Thrikaala varnam (three speed) practise gives a grasp of the dashavidha (technical lingo for a classification of gamakas under 10 types) gamakas and provide a starting point to develop "akaara" (i.e. singing of pharases, long or short, fast or slow, with the use of the abstract note "aa" ). Once a student has a basic assemblage of phrases in his/her vocabulary and a control of akaara, the structural aspects can be explored.

Other places to search for sangathis is to listen when anyone sings. No wonder the first (and sometimes only) bit of advice that any learned musician gives is "listen" ! I cannot overstate the importance of listening to grow your knowledge of sangathis ...


Ok, now we know what is a sangathi and how/where to find them. The next step is to stringing sangathis together to form a sentence. The key thing we learn in a language when we move from a "phrase" to a "sentence" is punctuation. Punctuation helps organize the ideas conveyed in phrases to enable effective communication of the meaning of sentence. The analogy is particularly good for constructing musical sentences.

First, note that music is a form of communication. If you are sitting on the stage, you are communicating with the audience. If you are alone in your basement practising, you are communicating with yourself. If you lose this communication, then whatever you sing can be safely termed as rambling. Havent you experienced rambling when you are practising and suddenly lose concentration ? Similarly if a performing artiste loses the communication of ideas with the audience OR him/herself, the result is a mechanical rambling output. Watch out for this in a performance. Anything mechanical is NOT improvisation and is NOT manodharma. An artiste should ideally strive to break free of anything that makes his output mechanical. For example, I dont like these so-called "chain-sangathis" ("ga-ma-ga- ga-ma-ga- ga-ma-ga- || -ri-ga-ri- ri-ga-ri- .. so on till infinity) which ironically a manodharma expert Madurai Mani helped popularize. The problem with this is that a) there is little punctuation b) there is little improvisation c) it is very mechanical... Sometimes, one gets into the fixation of chain-sangathis because they start learning alapana as an extension of alankaras (which have patternized forms such as "sa-ri-ga", "ri-ga-ma" etc) and not as an abstraction of krithis/varnas ...

Back to the definition of "punctuation" as a tool which helps organize phrases... Without belaboring the point, I'd like to assert that there is a need to have a punctuation between sangatis which allow the effective communication of the aesthetic form to be conveyed to the listener. The simplest form of punctuation is silence (or a pause). Especially care should be taken not to overrun the listener (possibly oneself) with junk sangatis. Silence as a form of punctuation is probably one of the ignored aspects of alapana on the concert platform - a sure sign is when the main artiste does not even allow enough silence for the accompanist to catch up or play the terminal part of the last phrase....

Another subtle form of punctuation is "correlation". The string of sangathis should have a strong aesthetic correlation. Patterning (as in alankaras) is a form of correlation, but is not the only form of correlation. In general I would conjecture that the moods determine aesthetics which in turn determine the kind of correlation between sangathis. If you observe any of the trinity's major composition, you will not see a pattern repeat thrice (or such an incident would be very rare). Every line in a krithi is an example of how to build correlation among sangathis.

Recall that we formed our "akaara" sangati out of a *bunch* of phrases of a krithi which corresponded to a word in the sahitya. Hence, we can safely say that the word in sahitya is a unit of correlation in a composition. Now words have meaning and the composers have then gone on and correlated the music between words to reflect the mood behind the meaning of the words. So this gives us two simple techniques: a) Study the correlation between the musical phrases forming a word of a krithi b) study the correlation between the sets of musical phrases underlying a set of words in a krithi. If nothing else, this exercise will help you sing the krithi better -- you will now know at least the rough meaning of words, and will be able to emphasize the music with a knowledge that it is meant for the words and not in a context-free manner. {this is invaluable for singing good neraval - a topic I will write about later}

I also wanted to quickly terminate this discussion of punctuation by saying that you can achieve punctuation by also varying the speed of delivering the notes (intra- or inter-sangathis). For example, a long note is sometimes equivalent to a silence period.

Zones, nyasa swaras, jiva swaras:

The next simple technique in alaapana is to choose a "zone" ie a range of swaras and confine your phrases within that range. This practise helps you dig deep within that area for phrases. Avoid patterning. Phrases can be repeated, but in different contexts.

The two technical terms one comes across in such exposition is the concept of a "jiva" swara/sangathi and that of a "nyasa" swara/sangathi. A jiva swara is a note which has the potential to reveal the "life" or the form of the raga. Quickly we can say that zones ought to formed around jiva swaras, or there should be at least one jiva swara in a zone -- otherwise there might be no scope for elaboration in the zone... A "nyasa" swara/sangathi is a swara/sangathi appearing at the end of a phrase or set of phrases (typically before a punctuation like silence). Jiva and nyasa swara/sangathis can be thought of as a basis around which one develops a set of correlated sangathis. So, when you hear a musician singing an alapana, can you identify the current jiva/nyasa swaras/sangathis ? Doing this is a step towards active listening ... You need not explicitly pinpoint it -- you should be able to lock onto it (and hence can communicate (aka listen) more effectively) ...

Master vs amateur:

Now, this singing in a zone is very demanding. Sometimes one runs out of ideas. More often we see musicians who have no concept of sticking to a zone at any point through an alaapana. This reflects lack of discipline and a lack of structure. Anyway, the first step is to find a zone to elaborate. The next step is to elaborate. What to do if you run out of steam ?? I have found that the best way is to just hold the "nyasa" swara or highlight the "jiva" swara and concentrate - the next sangati will present itself (after the confusion clears)! This belief is very central to the development of ragaalaapana. If it were not true, then how is the "mechanical" raga alaapana of an amateur any less worthy than a scintillating one by the master ?

The goal is to break all "mechanical" leanings and let the improvisation (or "manodharma") flow ! Getting this flow is one of the biggest hurdles for an amateur. One way to do this is to set a rather large zone and practice madhyama kaala akaara continuously -- when you are constrained not to stop and set your mind on the sruti and raga bhava, the new ideas will automatically emerge. Singing the whole varnam in akaara is one simple way to do this...

Reg mechanical vs true improvisation, lets go back to the example I gave earlier. It is a commonly observed phenomenon that just after we hear a good song or alapana by an artiste (and we take a tea break at the sabha canteen), ideas hitherito unknown and of amazing grandeur seem to pop up in our heads - one would wish those ideas could have been captured and archived for posterity ! This is a proof positive that there exists depths in a raga which can be explored given the right mood, technique and creative energy. My hypothesis is that zoning is a disciplined way to channelize the creative energy and dig for the depths.

My guru used to say that one should master all the techniques possible and then when it comes to manodharma, just forget the techniqes and concentrate on the abstract or bhava, laya etc. The techniques will employ themselves when necessary. This is an argument for "bhava-directed" alaapana rather than "swara-based" or "noty" alaapana. Great masters could do that, amateurs cant.

Great masters went in several directions.

Trying to classify every musician into a box like this is infeasible - I just wanted to paint broad strokes to help understand styles of various artistes.


The theory books also give an academic description and structure of raga alaapana. They say that just as a technical essay has an "abstract/introduction", the main body, and then the conclusion, the alapana has several divisions called the akshipita (abstract - where you quickly give a picture of the entire range of the raga), ragavardhini (the main body - split into lower octave, mid-octave and higher-octave expositions, including fast-speed sangathis in the ending part), stayi and vardhini (which include some aspects of "taanam" and finally wind to a conclusion of the alaapana).

A note on singing alapanas of "similar" or "close" ragas: suffices to S Ramanathan used to say that finding the difference between Nayaki and Darbar is similar to how the mother of two twins finds the difference between them. Once you "live" or experience a ragam deeply and abstract out concepts which are characteristic of it, differentiating two close ragas becomes a trivial affair.

The laya component of a raga is usually highly misunderstood and underestimated. Suffices to say that an alaapana built around the phrases taken from the swarajati "Kamakshi" in bhairavi is very different from those taken from "upacharamu" or "koluvaiyunnade" in bhairavi.

Another subtle aspect is one of "modulation" (or how much volume to give to different sangathis, in different octaves or in different moods). Ariyakudi used to say that the voice should be as broad as a mountain base in the region of the "adhaara shadjam" or the base "sa" and in the mandra stayi ( the octave below the base "sa"), and it should gradually become pointed like a mountain peak at the higher sa (tara stayi). You would appreciate this fact because Ariyakudi had a wavering voice and had to take these special measures to keep himself aligned to sruti. {On a tangent you could also learn how to punctuate a sangati from Madurai Mani because he couldn't utter more than a few syllables at a time ...} The absence of modulation can give the impression that an artiste is "shouting" at the higher reaches and is not "weighty" in the lower reaches. On a more subtle level, many moods cannot be created if the modulation aspect is not mastered - listen to Veena Balachander or Lalgudi to see modulation exploited to its fullest. For a violinist modulation means a strong control of bowing technique...

Finally ...
after much rambling ...

So how do we see if the alaapana sung is of high quality or not ? As we have seen , the creative energy, structure, sruti alignment, effecitiveness of expansion within a zone (correlation between sangatis), amount of mechanical output, punctuation, and finally overall aesthetics are key ingredients. The listener has to learn to identify and appreciate each of these aspects and judge how well the artiste treats or uses these aspects in his/her alaapana. Hopefully this article has helped identify some of those aspects more clearly.

[1] B.R.C Iyengaar, "Listening to Carnatic music", 1998

Go back to Mohan's Carnatic Reference Library