Applying Bhatkhande's rules to enlist possible raga's
Bhushit Joshipura
The domain of discussion is in Hindustani (North Indian Classical)
music. However, please draw parallel from this public message to
Carnatic music if you find appropriate.
I want to share my efforts with and receive feedback on it from you
all. I am not a musician. My assumptions may be factually wrong.
Please be a strict reviewer.
My aim was to come up with a list of possible raagas under
Bhatkhande's rules.
[ I know these rules state neither necessity nor sufficiency - they
state common wisdom. So please, no flames on that matter. ]
The idea is something that can provide a crude grid for fitting raga's
more systematically.
[ Again, some of us may point out that raga's can not be bound to
rigid rules of grammar. Artists have been excercising and will
continue to excercise freedom of expression with the grammatical
purity of a raga. However, from my standpoint, we can handle such
situations later.]
I wrote a C program codifying (and at place modifying) the rules. I am
planning to share the whole code in this newsgroup and on my website.
I am still trying to simplify the code - so that it is in just one .c
file. Also I am planning to post all the possible ragas according to
the rules in this groups once you ratify my assumptions.
Please criticise my approach below:
-------------------Start ---------------------
1. Here are swara's: [ Taken from Pankaj Joshi's works ]
S r R g G m M P d D n N
2. Word "frame" is used to describe non-chronological set of swara's
[ Word aroha and avaroha may mean chronological renderence of swara's
in some books and in some minds ]
3 In a raga, "varjya" decision is based on a swara and not on saptaka
(mandra, madhya, tara) it belongs to, we do not need S' kind of
symbols to denote saptaka - nor we care about it in our calculations
at this stage
The rules I apply are as follows:
A frame (aroha or avaroha)
1. has to have S (no known exceptions)
2. has to have total 5, 6, or 7 swara's (a few exceptions known)
3. must not have both pure and vikrta form of a swara (a few
exceptions known) [ We can see a terminological muddle in term swara ]
4. must not be jumpy ( no two consecutive swara's both pure and vikrta
forms are varjya - no jumping from R to P is allowed ) (no known
exceptions)
4.1 r,R,g,G must not all be varjya
4.2 g,G,m,M must not all be varjya
4.3 m,M,P must not all be varjya
4.4 P,d,D must not all be varjya
4.5 d,D,n,N must not all be varjya
5. must have at least one valid vadi-samvadi pair (no known exception)
(I got one rule stating shruti distances between them and approximated
the pairs. One musician pointed out that M is not used as either of
them. However, I think it would be inappropriate to miss it out on one
single person's opinion.) The pairs are:
5.1 S,m
5.2 S,P
5.3 r,M
5.4 r,d
5.5 R,P
5.6 R,D
5.7 g,P
5.8 g,N
5.9 G,P
5.10 G,N
5.11 m,N
5.12 M,N
This gave so many frames:
97 oudava
110 shadava
32 sampoorna
--------
239 total
Pairing of frames:
6. Only pairs of frames that have at least one valid vadi-samvadi pair
(defined in 5 above) are substantial candidates for a raga
This gave so many possible raga's: [symmetric table]
avaroha
aroha \ oudava shadava sampoorna
oudava 5007 6701 2406
shadava 6701 8800 3060
sampoorna 2406 3060 1024
------------------------------
39165 total
-------------------End ---------------------
My question to you all:
1. I was told theoritically there are 9 jati's of raga's. However,
practically, no raga ever has more swara's in aroha than in avaroha.
In other words, there are no shadava-odava or sampoorna-odava or
sampoorna-shadava raga existing.
How true is this? If it is so, we are looking for a reduction of
~12000 possible raga's from above
2. To take care for darbari/ashawari kind of conflict wherein they
differ only in their vadi-samvadi pair, we have to further increase
the count. Given vadi-samvadi are the two most used swara's, it is
reasonable to assume that each valid vadi-samvadi pair inside a frame
pair represents possibility of finding *one* raga and not two.
Application of this "rule" explodes number of possible raga's to
~74000. This number is arrived at not considering question 1 above.
I am planning to finish approach of question 2 above tomorrow. Based
on your feedback about the rules and contents of rule, I shall modify
the numbers and post.
As a byproduct, we have got a numbering scheme to raga's. However,
that is worth a different thread.
Thanks for your attention and thanks in advance for your criticism.
-Bhushit Joshipura
Praful Kelkar
1. Raag as a musical concept precedes it description in these
reductionistic terms named swars, A/A, & V/S - which are used later in
order to communicate some simple rules about the basic structure that
the Raag uses for its evolution. These terms are useful for
classifying raags and studying them, as Pt. Bhatkhande did very
cleverly, but then to work backwards from mathematical calculations
like this is of limited value, since this is not how this music has
evolved.
2. Raags are not defined by the named swars, aroh/avaroh, and
vadi/samvadi alone -- e.g. bhairavi's v/s can be m/S, P/S, or d/g.
The aroh can be taken as SgmdnS or SrgmPdnS.
2. Many raags have vakra chalan - thus defying a linear definition of
A/A.
3. I don't think any raags have more swars in aroh than in avaroh.
4. I don't know if there are any raags that have teevra M as vadi or
samvadi.
5. I don't what you will do with the final number when you arrive at
one. You will probably come up with a larger number, but it may be
irrelevent since some will not have asthetic appeal, to be called a
"Raag".
PK
Sarvjit Goraya
Some years ago, I tried to learn Indian Classical music with
'physics-like' approach. I focused on frquency ratios and such like.
It is sad to say that I never succeeded; but I did learn that this is
an art form and not a C-programming assignement. Although
technologists keep on trying, the fine arts tend to defy the rigour
of structure. An arithmatic excercise such as this will be able to
count thousands of trees without eating a single mango. This is just a
personal opinion, no offence is meant or implied.
Regards
Sarvjit
naniwadekar
"Praful Kelkar" <praful...@yahoo.com> wrote -
>
> 1. Raag as a musical concept precedes it description in these
later in
> order to communicate some simple rules about the basic structure that
> the Raag uses for its evolution.
>
It is absurd to claim that raag as a concept precedes swara.
Raagas have evolved through experimenting with swara.
Aaroh-avaroh may be a (useful) 'reductionistic device' but
swara is a living thing.
>
> 3. I don't think any raags have more swars in aroh than in avaroh.
>
I tried a search on rmic archives for 'sampoorna-shadav'.
A post dated 18/12/1998 shows this entry :
"Kanada Bahar Thaat Jati sampoorna-shadav Aroha Sn SRSRgMPmDNS * Avaroha
S>nPmPgmRS Vadi m Samvadi S Pakad
Rn<-SR(m)g-(m)g-mP-m-n-DN-S>-nP-mP-g-m-R-R-g-mRS Time N Ras
Chalan "
- dn
Praful Kelkar
> It is absurd to claim that raag as a concept precedes swara.
> Raagas have evolved through experimenting with swara.
> Aaroh-avaroh may be a (useful) 'reductionistic device' but
> swara is a living thing.
No you misunderstood or I perhaps didn't state it clearly enough -
the "named swar" are reductionist simplifications...... Each swar is
indeed alive with the spirit of the raag - hece kR in Bhairav is
unique and different from kR in Todi, Shree, Multani, etc. (relating
to complicated differences which is a separate long discussion) .. but
the naming of both the swars as kR is reductionistic and misleading.
There is no just one kR.
It is like chicken and the egg.... the trick is Bhairav's kR has no
existance by itself without there being the concept of Bhairav. As
Jhasaheb says eloquently "No Swar sings alone. You cannot talk about
shruti without reference to swar, and you can't talk about swar
without reference to the Raag. Ref: Jha lec-dem in RP's posting on
Adana. (Just by the way, if WS is listening - the kN in Adana is
higher than the 'usual' kN - not sometimes higher but consistently,
persistently, and intentionally.)
PK
Warren Senders
>higher than the 'usual' kN -
I agree entirely.
>not sometimes higher but consistently,
>persistently, and intentionally.)
Except when it isn't.
WS
Ashok
You mean you agree entirely, except when you don't?
Ashok
Praful Kelkar
All people, small and great agree .. that it needs to be higher ..
everyone agrees in principle to the shastra ... but when it comes to
performance people can have slip ups, smaller one slipping up more
often than the greats, and not intonate swar correctly ... but that's
a separate issue.
I like Jhasahebs insights into this old debate we had .. he seems to
indicate that kG moves up in Multani being pulled by tM .. so he seems
to agree with some of the higher/lower stratification of shrutis. But
then he says too much of that is goofiness - phalane raagme phalane
swarki phalani shruti hai yeh sab bakwas. Truth may be somewhere in
the middle!
PK
Warren Senders
>
>All people, small and great agree .. that it needs to be higher ..
>everyone agrees in principle to the shastra ... but when it comes to
>performance people can have slip ups, smaller one slipping up more
>often than the greats, and not intonate swar correctly ... but that's
>a separate issue.
In the case of Adana, then, "slipping up" would result in the
correct interval. Perhaps you mean "slipping down."
In any case there is a distinction between the Platono-Shastric
ideal of a raag, and the actual intervallic realpolitik. The
implication of Praful's original assertion is that if the alteration
of Ni is not adhered to, the raag has been incorrectly performed.
Since we are able to recognize raags even when a note is out
of tune, it follows that the actual core identity of a raag lies not
in interval relationships as such, but in the relationships of notes
and phrases...in other words, in the chalan. Thus we can recognize
Adana as played by Shivji, for example...even if he doesn't tweak
his Ni string before starting. If Praful wants to criticize the santooriya
for incorrect intonation, that's his prerogative, but he should
recognize that it's not a very productive way to occupy all those
long lonely hours....I think I'll go practice.
Ta ta,
WS
naniwadekar
> indicate that kG moves up in Multani being pulled by tM .. so he seems
> to agree with some of the higher/lower stratification of shrutis. But
> then he says too much of that is goofiness - phalane raagme phalane
> swarki phalani shruti hai yeh sab bakwas. Truth may be somewhere in
> the middle!
>
In which clip has Ramrang said : 'phalaane raag me.n ...' etc ?
- dn
Praful Kelkar
You seem to be simply arguing, perhaps only because you like to or
something. There is absolutely no ground in what you are saying. We
have talked about this before - brain is a master instrument in
pattern recognition and can even read physician's or a child's
scriggly handwriting even if the letters are not drwan
'shastra-shuddha'. So, even if you sing besoora, the 'you' is generic
you, not implying your ability to sing soorila or not, brain can
recognize the raag. So, your argument doesn't disprove the shastra.
And yes, playing a fltter kN would be wrong for Adana, even if one is
able to name the raag as such. That would be child's doodling,
decipherable, but not master calligraphy!
Whether raag 'happens' due to intervals or not is a long separate
discussion - the interval in Adana may not be what creates Adana - the
'out of interval' kN is intentional and perhaps critical for creating
proper Adana rasa. Playing the flatter kN would display lack of depth
in performers understanding... it may be overlooked and one can ejoy
other aspects of raag-daari, but masters and also good listeners and
good students would know and realize that the person lacks pakka samaz
of the soor.
And BTW, Shivji is a master at tuning his santoor and meticuloulsy
tunes the shrutis for the raag before he plays. He doesn't use
generic tuning. His instrument obvioulsy cannot play two kG in
darabari, nor can it create the andolita and meends - but given the
limitations of it he is a master, altho' personally I don't care for
santoor. Certainly he is not besoora! I have not heard his Adana
personally to say anything more specific.
As to Nani's question the lovely lec-dem is in Adana section - Kanada
part 1. Listen also to section on Darbari where he beautifully shows
the two kGs, one moving down due to R and one moving up due to M -
what intervals are we talking about? I think ICM purposefully uses
"stange intervals" in order to create certain peculiar feeling -- so I
would be careful not getting too caught up applying the laws of
intervals to analyze ICM... that would be another reductionist
simplification.
Incedently, my respect for Jhasaheb has grown exponentially thanks to
his and RP's willingness to share the knowledge to freely and openly!
I like this spirit that they bring to ICM, whereas secrecy and small
mindedness has prevailed thru generations of gharanas; and more often
than not pandits and ustads tend to act secretive and create a
'mysterious' air about this very rational shastra.
Ramram,
PK
Rajan P. Parrikar
>Incedently, my respect for Jhasaheb has grown exponentially thanks to
>his and RP's willingness to share the knowledge to freely and openly!
Btw, Jha-sahab just returned from an event at the
SRA in Calcutta where he commended the organizers for
their kindness at finally looking his way at a stage
when he has one foot in the grave. Other participants
in this gathering included Dinkar Kaikini, L.K. Pandit,
Babanrao Haldankar, Shanno Khurana etc. I understand
that all the sessions have been videotaped.
Ramrang first sang Hem, then Nat, then Hem-Nat and
then Raga Sazgiri, presenting his own creations
without announcing them as such, deliberately
choosing those that did not carry his mudra. He
mischievously reports that at this point, Dipali
Nag innocently wondered if he would be now so kind
as to present some of his own compositions.
Warm regards,
r
Bhushit Joshipura
to right track.
"Reductionistic" does not mean "useless".
1. Description of position on the earth in terms of latitude and
longitude is reductionist but very useful.
2. Written words are reductionistic representations of spoken words.
[And we never codified Vedas in written languages on same lame
arguments. Remember what Malek Kafoor and Kaalaa PahaaD could do to
our culture then?]
3. Even current codification of ICM is reductionistic presentation.
Nothing (other than discrete mathematics) is inherently discrete.
Path of evolution and path of analysis and presentation need not be
the same.
1. Indian consonants might have originally been in maalini order.
However, (perhaps later) we adopted the best phonetic order in the
world.
Until the periodic table, chemistry (or alchemy or che-mystery) was
all mystery and saturated to a level. Once proper indexing of elements
was in place, chemistry exploded beyond comparison.
These points reply to Mr. Kelkar's point #1.
However, I think I have to start from a different angle this time.
Please distinguish between terms "possible raga" and raga.
1. The approach has given us an indexing scheme. This scheme
(reductionistically though) describes a possible raga structurally.
You can consider this as a next step to chakra-meLaadi or ThaaTh
system. It can even describe a "not possible raga" (something not
falling in our rules).
2. The indexing represents a possible raga as a quadruplet of four
numbers:
a/a/v/s
3. A raga may represent one or more of these quadruplets ( one-to-many
as Mr. Kelkar points out in case of Bhairavi or one-to-one in many
cases - notably aashaawaree and darbaaree kaanhDaa).
These points reply to Mr. Kelkar's point #2.
My posting read:
>2. Word "frame" is used to describe non-chronological set of swara's
>[ Word aroha and avaroha may mean chronological renderence of swara's
>in some books and in some minds ]
so Mr. Kelkar's point #3 (misnumbered as #2 again) is half-replied.
The other half will be replied later in the posting.
Mr. Naniwadekar gave a counter example to Mr. Kelkar's point #4
(misnumbered as #3).
Absence of M as v/s is well taken in point #5 (misnumbered as #4). I
shall remove reference to M as v or s in my program.
The most serious jump in conclusion by Mr. Kelkar is point #6
(misnumbered as #5).
(I introduce sub-numbering for better argumentation.)
>5.
>6.1 I don't what you will do with the final number when you arrive
at
one.
>6.2 You will probably come up with a larger number,
>6.3 but it may be irrelevent since some will not have asthetic
appeal, to be called a
"Raag".
#6.1( alongwith #3) will be dealt in later paragraph.
#6.2 - I have already come up with a large number - and I am here to
discuss rational way to reduce that large number.
#6.3 - How solid is the argument : "Some of human beings are thieves
so human life is irrelavent"? The counter argument is : "Some of human
beings are thieves, some of them are Buddha too! So human life is
worth analysis."
Sarvjit's argument:
>An arithmatic excercise such as this will be able to
>count thousands of trees without eating a single mango.
also falls apart in the fact that surveying mango fields can not yield
a mango fruit but can give an approximation of how much crop to
expect.
Now the most interesting part: what do we want to achieve with all
this discussion?
The idea is to
1. come up with as mathematical indexing as possible.
2. gather the database of existing ragas indexed by such a scheme - of
its finer characteristics like shrutis or minDs or chronological
renderences (replying other half of Mr.Kelkar's #3)
3. feed the whole thing to various neural networks (with index as
input and other characteristics as outputs) to learn
4. gather approximations about possible ragas from the trained neural
networks
But sheer number of possible ragas need to be addressed first. 70000
candidates are not possible to examine within a single human lifetime.
Given that this is going to be a "try and tune" exercise, it is
going to be a repeatatious task.
We need to intelligently cutdown number of possible ragas. That is
where the whole discussion is worth. We should still be able to teach
neural networks with exceptions like kaanaDaa-bahaar. However, we will
limit our query to more possible set of unknown candidates.
Again, what we will get will be approximations. But those will be
scientifically sound approximations. The chances of finding a raga
around such an approximation will be higher. (Remember the alchemy vs.
chemistry parallel?) This is just the first step towards a possibility
of explosive growth.
So, if you all still agree, I modify the rules once again:
5. must have at least one valid vadi-samvadi pair (no known exception)
The pairs are:
5.1 S,m
5.2 S,P
5.3 r,d
5.4 R,P
5.5 R,D
5.6 g,P
5.7 g,N
5.8 G,P
5.9 G,N
5.10 m,N
6. Only pairs of frames that have
6.1 number of swaras in aroha frame less than or equal to those in
avaroha frame AND
6.2 at least one valid vadi-samvadi pair (defined in 5 above)
are substantial candidates for a raga
Please continue discussing...
Thank you for all your time and attention.
-Bhushit
praful...@yahoo.com (Praful Kelkar) wrote in message news:<fb5417fe.02120...@posting.google.com>...
Surajit A. Bose
bhu...@hotmail.com (Bhushit Joshipura) wrote:
> 2. Written words are reductionistic representations of spoken words.
A long line of poststructuralists from Derrida on down would beg to
differ.
-s