Note: A post after a long time. Perhaps I am rusty as I had a real hard time penning this one. I had broach this subject (a deeper look into the rhythmic maze woven in kalpanaswaras) for a while, but I could not get my thoughts ordered correctly, and thus  it seemed quite hard to prevent it from getting out of hand and too expansive. Also, this is certainly an area where I am shaky at best, and so in spite of the effort I may have gone off the beam. Advance apologies if that turns out to be the case!

The term kaNakku in Tamil means calculation, and is used to refer to many things, the most common of which is a colloquial reference to the subject of math in general.

In Carnatic Music also, it finds use albeit mainly by the tamil speaking cross section of the Carnatic Music populace. Here, like most unofficial, colloquial terms, it does not have a precise, specific meaning. I think it refers to “complexity”, in particular “rhythmic complexity” – an obviously nebulous reference. When used to describe the music of an artist,  it is used to describe the artistic expression of the musician’s kalpanaswaras. When someone says His/Her music has a lot of kaNakku, it means that during kalpanaswaras, the musician typically weaves seemingly complicated rhythmic patterns presumably involving elaborate calculations i.e. kaNakku. This may involve swaras/pauses of varied lengths, and also employing complex techniques like switches in gati/gait. This to most of us would seem complicated, and that is perhaps so because it is “non intuitive” in the sense that a steady, smooth flowing rhythm is not easily apparent in the constructed melodic patterns.

The other end of the spectrum to this type of kalpanaswaras is one which is referred by another nebulous term – sarvalaghu. I do not know what exactly the word means, or again, whether it has a precise meaning that every one agrees upon. At best, I can think we can describe it as weaving kalpanaswara patterns where a somewhat steady, smooth flowing rhythm natural rhythm is obvious to the listener. Of course, I realize that the phrase steady, smooth flowing rhythm is perhaps as nebulous as the term it is trying to shed light upon on 🙂 !

Hence, perhaps the difference would be better demonstrated via examples:

Here is an example of the sarvalaghu type of kalpanaswaras:

The musician here is Semmangudi Srinivasa Iyer (SSI) who almost always did kalpanaswaras in this fashion, and he always does a superb job. In fact, this is the more common type of kalpanaswaras which almost all musicians follow. The great Madurai Mani Iyer who was famous for his kalpanaswara renditions excelled in sarvalaghu style. Today, for example, T.M. Krishna amongst the current day vocalists employs a style very similar to SSI above. In fact, even those who employ “kanakku” usually start (or end) with this type of kalpanaswaras.

Contrast the above with this by T.N. Seshagopalan (TNS), particularly the iteration starting with him:

The piece starts with the violinist’s response which is more or less a straightforward sarvalaghu type style – you can clearly make out swaras in groups of 4. However, compare that with TNS’s next iteration? Can you see that starting with it, the rhythm of the melody seems somewhat complicated? Of course the pattern is certainly catchy but to me, it seems it isn’t as “easy flowing” as the SSI sample.

But why is this so? .

Sarvalaghu style – a deeper look:
Tala, gati, and mAtra: The song that SSI is singing swaras to is mArubalka (Sriranjani, Thyagaraja), which is set to Adi Tala, in  catusra gati. What does this mean? This implies that the “beat” of the song, which has the same duration as an akshara of the tala (but need not align with it and hence is not necessarily synonymous with it) can be further divided into 2 or 4 or even 8 smaller units of time. This can be in theory be 16, 32 etc. in theory, but let us consider simpler cases.

If we assume that it is 4 divisions (most common),  then within the duration of a whole cycle of Adi tala (which has 8 aksharas)  you have 32 such time units as follows:

1 2 3 4 5 6 7 8
1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4

This smaller time division is called a mAtra or mAttirai (tamil).  So if a musician is singing swaras for one tala cycle  as above, and each swara‘s duration is equal to this smaller time unit, then he/she would be able to fit 32 swaras in a cycle of Adi tala. This is the typical way things are done in most Adi tala songs (in catusra gati).

Now in sarvalaghu singing,  for catusra gati, I believe that the swaras (and pauses) are almost always 1, 2, and 4 mAtras, (i.e. 1/4th 1/2, or 1-whole akshara durations), and you will also find them grouped such thatas per the tala “beats” (either on the “beat” or half-way in between). Not all swaras need be of same duration  (e.g. like srgm rgmd gmdn mdns), and thus it can be a mixture of these.  If you mostly have 1, 2, or 4 mAtras for swaras (and pauses), then you can see in general you will find that they align “to the beat”.

Here is the notations for the swaras the vocalist (SSI) sings in that piece:

Note that the swaras shown above are grouped in 4s (not necessarily aligned to tala), and a set of 4 have the same duration as the akshara, which is the span of the natural “beat of the song”. As you can see, the overwhelming # of swaras are 1/4th akshara in duration, and there are almost no pauses. This is typical of sarvalaghu type of swaras singing and it helps establish a riveting, easy to grasp rhythm.

“Kanakku” style: In the case of that T.N. Seshagopalan piece, as we will see, the swaras do not follow this style. In fact, it took me quite a bit of analysis and very very careful listening to arrive at the correct rhythmic pattern being used here (a reflection of my weakness w.r.t laya 🙂 ). Here is the notation for the first iteration that TNS sings:

Again swaras are grouped in 4 which span an akshara. The blue [ ] markers are used to group the swaras as per the musical pattern. Just from the look of the above, it should be clear that the piece is more complex in its composition that the previous case. If you look at it carefully, you can see that TNS is weaving a 5-count (i.e. khaNDa) based melodic pattern within the catusra gati framework (i.e. where the mAtras are spaced apart in groups of 4).  The base melodic pattern is 10 mAtras i.e. 2 and 1/2 aksharas long, and is mostly made up of 5 swaras divided as 1, 1 (pause), 2, 1, 5  (e.g m , d , m g , , , ,  and m , d ,  m d , , , , ). There is one case in the middle where I think he intentionally or unintentionally leaves a pattern as 9 mAtras, but stretches the following one to 11 mAtras, and thus maintain the 20 mAtra count for two successive patterns. But in general it is a 10-mAtra pattern. Although, I have not analyzed all iterations it seems the violinist’s response to this iteration, and the entire next iteration is based on such a pattern.

As I mention above, it took quite a bit of effort for me to arrive at the pattern, and be reasonably sure it is indeed the case! Let us say I am still not 100% sure! I actually I had to slow down the song a lot for me to be able to count the mAtras for the swaras and thus determine that it is a khaNDa pattern.

What was interesting is that I got the rhythmic sense of the pattern quite easily registered in my brain, could relate to it well, and even able to reproduce it from memory almost immediately. However to find out how many mAtras it took etc. was still a difficult challenge. Even now, I have to sing it extremely slowly to correctly count the 5-akshara swara as well as the 1+1(pause)+2+1 pattern right :-).

So given that, some of you may also be wondering if this is really correct. So I present another way to look at this – a “video” that demonstrates that the notation above is correct. Note however that the tempo slowed down by about 30% so that each mAtra is stretched out to be long enough to be more discernible to the eye:

In that video, I have placed markers at regular intervals spanning the duration of a mAtra. Then I placed labels for the swara pattern for the first iteration. As you can see the swara pattern matches the audio (mostly – there is some jitter  as humans are not robots). Of course it might seem silly to do this but I also think that this is a nice instructional method for illustrating laya aspects. What do you think?

In any case, why is this piece “more complex” and “less intuitive”  (at least to me)? Based on above, I am thinking that perhaps the odd numbered 5-mAtra long pause-laden swara is something “atypical” that my mind does not grasp easily as compared to a 2-mAtra or 4-mAtra swara (which is ubiquitous).  This is especially so when done within a catusra gati song which till that point had laid out the more ubiquitous 4-based i.e. ta-ka-dhi-mi rhythm clearly (this is why I chose to start the sample with the violinist’s portion).  On top of that we have 10-base pattern for an entire iteration, done within a catusra gati which  seems atypical.

So I wonder if this is perhaps the reason why the rhythm while attractive has a complicated feel to it. When musicians employ “kANakku” it is typical for them to employ notes and pauses like these of “atypical” length. When we listen, we do know something complicated is going on, but perhaps most of us are not able to tell why.

Only when you peer inside you see the sophistication involved, and you begin to understand why many feel that Carnatic Music has one of the most sophisticated rhythmic components amongst all music systems of the world!

But you ain’t seen nothing yet !!!
Carnatic Musicians go even further in terms of rhythmic sophistication. I will post about this with one mind-boggling kalpanaswara performance from a genius of a musician in the next part