Note: Opinions expressed here are mine and mine alone.  Some of the ideas expressed here are controversial and question some established notions. I advise you to think this through and draw your own conclusions.

Certain songs by muttusvAmi dIkhsitar are labeled as tiSra tripuTa (3+2+2) or miSra Eka (7) in  authentic texts like the sangIta sampradAya pradarshini, but are sung in miSra cApu (3.5*2=7) nowadays. Similarly many songs of Tyagaraja are labeled in older texts as jhampa, implying miSra jhampA (7+2+1), but are sung nowadays in khaNDa cApu (2+3). There are also some songs in Adi tiSra gati (8*3=24) which are sung in rUpaka cApu (3*4*2), or tiSra Eka (3*4*2).

Now some questions.

  1. Is it really not right to sing a tiSra triputA song in miSra cApu as many rasikas may believe by default? Is it a corruption as some claim? Or is miSra cApu really the same as tiSra tripuTa?
  2. From a carnatic music classicism and tradition point of view, it would be pointless to argue about the “corruption, not right” sentiment, as it can be personal and subjective. But let us ask this from a more scientific standpoint: Does the tala itself impose a mandatory, unavoidable, influence on the melody? Thus, does changing the tala as above affect the melody itself?

I have not always answered as such, but my answer now is No.

I used to think that talas like tiSra tripuTa and miSra cApu are really two different animals in terms of their effect of melody, and while they can be equated only in some cases, any equivalence beyond that is contrived.

However,  now I think that from a strict rhythmic standpoint, they are fully equivalent, and thus cannot affect the rhythm and the melody of the rendition itself.  While I do not deny that the concept of tala may have helped people to come up with sophisticated rhythmic concepts, I do not think it is the only means to sophisticated rhythms, even the precise ones we find in carnatic music.  I think that may shock many 🙂 ! I will at least try to explain why I think so by first elaborating on why I think tiSra tripuTa and miSra cApu are fully equivalent and hence equally effective in keep time for the same song.

Read on (warning: gets technical)

Consider tiSra tripuTa. It has 7 aksharas divided as a 3 count laghu (beat + two finger counts), plus a 4 count dhrutam (beat + wave + beat + wave). Let us assume catuSra gati. Now, what does this mean? It means that if one were to divide one cycle of the tala into equal parts, you will come up a number that is an evenpower-of-2 multiple of 7. In other words, if one were to fill a cycle of talas with swaras/sollus which all have the same duration in time, you end with with “n” swaras/sollus where “n” is a evenpower-of-2 multiple of 7 (7,14,28,56 etc.).

Note: In western terminology, if the cycle is filled up as all whole notes, or all half-notes, or all quarter notes, you will use “n” notes, where n in an evenpower-of-2 multiple of 7. Of course in reality, carnatic melodies are not “monotonous” like this in terms of how they fill the cycle. But we take the simplest case here to illustrate our point.

Let us consider 3 such divisions:

(i) 7 swaras/sollus. Each swara/sollu spans the duration of 1 akshara. We will use the sollu pattern ta-ki-Ta   ta-ka-dhi-mi here.

ta ki Ta | ta ka |  dhi mi ||
s  r  g  | s  r  |  g   m  ||
B  L  L    B  W     B   W

(Note: B is beat, L is finger-count part of a laghu, W is wave)

(2) 14 swaras/sollus. Each swara/sollu spans the duration of 1/2 akshara. We will use the sollu pattern ta-ka-ta-ri-ki-Ta    ta-ka-dhi-mi    ta-ka-ju-Nu (6+4+4) for the 14 sollus.

ta ka ta ri ki Ta | ta ka dhi mi | ta ka ju Nu ||
s  r  g  s  r  g  | s  r  g   m  | s  r  g  m  ||
L     L     L       B     W        B     W

(3) 28 swaras/sollus. Each swara/sollu spans the duration of 1/4 akshara. We will use the sollu pattern ta-ka-ta-ri-ki-Ta-ta-ka-ta-ri-ki-Ta      ta-ka-dhi-mi-ta-ka-ju-Nu    ta-ka-dhi-mi-ta-ka-ju-Nu (12+8+8) for representing the 28 sollus.

ta ka ta ri ki Ta ta ka ta ri ki Ta |
s  r  g  m  g  r  s  r  g  m  g  r
L           L           L            

ta ka dhi mi ta ka ju Nu | ta ka dhi mi ta ka ju Nu ||
s  r  g   m  p  d  n  s    s  r  g   m  p  d  n  s
B            W             B            W

If e.g. we assume that each akshara in a particular rendition takes 1 second, then in all three cases, the cycle takes 7 seconds. In the first case, each swara/sollu spans 1 second, in the second, 0.5 seconds and in the third, 0.25 seconds.

Mapping to miSra cApu
Now if to we were map these to miSra cApu in a equivalent way, we can do it as such


ta ki Ta |  ta ka | dh mi ||
B  B        B       B     ||


ta ka ta ri ki Ta | ta ka dhi mi | ta ka ju Nu ||
s  r  g  s  r  g  | s  r  g   m  | s  r  g  m  ||
|-----1st-cycle--------|-----2nd-cycle---------|    (cycles)
B        B     B       B           B     B          (kriyas or tala signs)
1  2  3  4  5  6    7  1  2   3    4  5  6  7       (divisions of the tala)


ta ka ta ri ki Ta ta ka ta ri ki Ta | ta ka
s  r  g  m  g  r  s  r  g  m  g  r    s  r
B                 B           B
1     2     3     4     5     6       7
ta ka ta ri ki Ta ta ka ta ri ki Ta   ta ka

dhi mi ta ka ju Nu | ta ka dhi mi ta ka ju Nu ||
g   m  p  d  n  s    s  r  g   m  p  d  n  s
B                    B            B
1      2     3       4     5      6     7
dhi mi ta ka ju Nu   ta ka dhi mi ta ka ju Nu

Divying up the swaras/sollus
In case (2), the tiSra tripuTa tala divvied the syllables among the tala signals (i.e. beats, laghus etc.) as follows:

ta ka    B-Beat   (laghu-1)
ta ri    L-Count  (laghu-2)
ki Ta    L-Count  (laghu-3)
ta ka    B-Beat
dhi mi   W-Wave
ta ka    B-Beat
ju Nu    W-Wave

miSra cApu divied up as

ta ka ta    B  (1st cycle start)
ri ki       B
Ta ta       B
ka dhi mi   B  (2nd cycle start)
ta ka       B
ju Nu       B

In case (3), the tiSra tripuTa tala divvied the syllables among the tala signs (i.e. beats, laghus etc.) as follows:

ta ka ta ri   B-Beat   (laghu-1)
ki Ta ta ka   L-Count  (laghu-2)
ta ri ki Ta   L-Count  (laghu-3)
ta ka dhi mi  B-Beat
ta ka ju  Nu  W-wave
ta ka dhi mi  B-Beat
ta ka ju  Nu  W-wave

misra cApu divvied up as

ta ka ta ri ki Ta    B  (1st cycle start)
ta ka ta ri          B
ki Ta ta ka          B
dhi mi ta ka ju Nu   B  (2nd cycle start)
ta ka dhi mi         B
ta ka ju  Nu         B

Basically in each case, the swaras/sollus of one cycle of tiSra tripuTa can be represented exactly by 1 or 2 cycles of miSra cApu. In the first case, there is a exact one-one correlation, but this case is pretty rare in krithis, and really appears only in geethams. In the next two cases, 1 cycle of tiSra tripuTa maps to 2 cycles of miSra cApu, this is found in krithis.

We should highlight that in the end, for each case, the same number of swaras/sollus are executed in the same duration with both talas. The only difference is that the “time keeper” i.e the tala counted up to that duration differently. It is like to add up to 70 cents/paisas, one added up 50+10+10 (and further the 50 as 10+10+10+10+10), but other added as 15+10+10+15+10+10. In the end the value is the same – the internal divisions are different.

However, there seems basically no reason to believe that the internal divisions of these talas should necessarily impose any mandatory influence on the melody itself i.e. in particular how the melody itself is “sub-divided” into swaras of various duration. To bolster this, take the case of certain songs that are sung in both Adi tiSra gati as well as rupaka “pretty much similarly” – as in unlike other songs whose laya itself has a tiSra (triplet) gait, these songs have a more of a a standard catuSra gait. Here again this is like one adding up to 24 cents/paisas  as 3+3+3+3+3+3+3+3 vs. 4+4+4+4+4.

In fact, going further, one could have effectively mapped the above for miSra Eka (1+1+1+1+1+1+1 = 7), khaNDa rUpaka ( 5+2= 7), catuSra jhampa (4+2+1 = 7). One can render the same song as effectively in all these talas. It is not just a question of miSra cApu vs. tiSra tripuTa for me. All these 7 unit case talas are equally effective. Sure keeping tala with one vs. other may pose challenges for us, but it is a problem with us humans, it really has no effect on the inherent melody of the song itself which may or may not align with the tala. For example, many Adi songs no necessarily align with the angas (internal parts) of Adi.

This is the reason why I believe, tala is not integral to determining the rhythm of the melody itself. While the concept of tala has no doubt proven to be very useful in rhythm of carnatic music (in helping composers set rhythm etc.), it is a key to the treasure i.e. rhythm/laya, and not the treasure itself. It is a useful tool to chisel a product, but still it is only the tool, and not the product.

I may be wrong, but in Carnatic Music, one aspect where we have paid too much importance to the role of the tala is w.r.t its internal structure. This is what that leads us to notions like why a song in say Adi shouldn’t really be kept in khaNDa jhampa since it would somehow runs against the core of the rhythm of that song, and thus unnatural. This, in spite of similar evidences to the contrary (e.g. rUpaka and tiSra Eka effectively used for Adi tiSra gati). Like I said, I think there is no reason to not treat all 8 beat cycles (+ the 4-beat catuSra Eka) as equivalent w.r.t to how equally effective they can be in helping maintain rhythm of a song.

In fact, if we really look at it, keeping tala itself (for the song portion), is an aid, just like tampura is an aid to establish and maintain sruthi. Of course both have evolved in their roles (e.g. tala being very important in pallavis) but their core role is as an aid to more important aspects of music – sruthi and laya. And I think those aspects of music are fully defined on their own, and not in terms of these aids.

Update: There is a companion post to this one: A bit about Tala – himAdri sutE. It is based on the (very technical) discussion that went on in the comment section below.