The Rags of North in D Ia N Music: Their Structure and Evolution

THE RAGS OF
NORTH INDIAN MUSIC
Their Structure and Evolution
by
N. A. J A I R A Z B H O Y
F A B E R A N D FA B E R
3 Queen Square
London
ProQuest Number: 10752720
All rights reserved
INFORMATION TO ALL USERS

The quality of this reproduction is d e p e n d e n t upon the quality of the copy subm itted.
In the unlikely e v e n t that the a u thor did not send a c o m p le te m anuscript

and there are missing pages, these will be noted. Also, if m aterial had to be rem oved,
a n o te will ind ica te the deletion.
uest
ProQuest 10752720
Published by ProQuest LLC(2018). C opyright of the Dissertation is held by the Author.
All rights reserved.

This work is protected against unauthorized copying under Title 17, United States C o d e
M icroform Edition © ProQuest LLC.
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 4 8 1 0 6 - 1346
First published in 1971
by Faber and Faber Limited
Printed in Great Britain by
William Clowes & Sons Limited
London, Colchester and Beccles
All rights reserved
IS B N 0 571 08315 3
© N. A, Jairazbhoy, 1971
Acknowledgements
I t gives me great pleasure to acknowledge my indebtedness to the many associates
and students who have helped in the preparation of this work. In particular I should
like to mention A. A. Dick, O. Wright, R. Clausen, J. M ontagu and K. W oodward
for their valuable advice and thought-provoking comments. I also wish to express my
thanks to G. Gibberd for his help in designing the model of the system of scales
discussed in the Appendix and to Miss M, Bennett for her assistance with the notation
of Vilayat K han’s record.
I feel greatly honoured that one of India’s leading musicians, TJstdd Vilayat Khan,
should have consented to play the musical examples for this work. Limited as he
was to under two minutes for each rag, he has nevertheless managed to illustrate the
melodic movement as well as to capture the essence of the rag, and the result in each
case is a work of art. I should like to express my extreme gratitude to Ustdd Vilayat
Khan for his contribution to this work.
In seeing this book through the press, I am deeply indebted to M. Kingsbury of
Faber Music for his patient and critical editorial assistance which included many
valuable suggestions for making this book more readable.
Much credit is due to the late Dr. A. A. Bake, my mentor for many years, and to
my wife for her forbearance and encouragement as well as technical assistance.
N.A.J.
University of Windsor,
Windsor, Ontario,
Canada, 1970
Contents
Preface 11
Note on Transliteration and Pronunciation 13
Introduction to the Historical Background 16
I An Outline of Present-Day N orth Indian Classical
Music 27
II Basic Elements of Theory 32
III That 46
IV The Effect of Drones 65
V Evolution of the Circle of Thats 90
VI Alternative Notes 102
VII Transilient Scales 122
VIII Symmetry, Movement and Intonation 151
Summary 179
Appendix A : The System of Thirty-two Thats 181
Appendix B : Description and N otation of Recorded
Music Examples 186
Bibliography 211
Index 213
Preface
T h e re is a remarkable uniformity in the performance of classical music in N orth
India, an area comprising various geographical regions, which, in this context,
includes Pakistan and extends southward into the Deccan, There are, of course,
differences in detail—in the interpretation of various mgs, in style of performance
and in the types and texts of compositions—but on the whole these are only minor
differences. The overall uniformity is especially remarkable in view of the fact that
these regions contain a heterogeneous population—both racially and culturally—who
speak a variety of languages and differ widely in their religious beliefs. N orth Indian
classical music cuts across the usual barriers imposed by differences of language and
religion, much as does classical music in the West. Nevertheless, many classical songs
have religious texts, both Hindu and Muslim. But religious content is not an essential
requisite of the music, for some songs are concerned with mundane subjects and some
are even composed of meaningless syllables. Just as in Western classical music where
great religious works written specifically for the Roman Catholic Church can be
appreciated as works of art by those of all religious beliefs, so too in Indian music
religious themes often serve as vehicles for artistic expression.
Classical music is not the music of the masses but is largely confined to the urban
areas of N orth India. It is performed either in concert halls or in private homes.
Its raison d ’itre lies in its purely musical content and it is basically on melody and
rhythm that its quality is assessed. While a study of the cultural background of the
people is essential for a social and historical perspective of this music, its appreciation
depends largely on comprehension of the musical idiom, and it is to this end that
the present work is dedicated. It had its origin in a series of lectures given at the
School of Oriental and African Studies, London, to university students who had no
previous knowledge of the subject. At an equivalent age level in India, students would
have had several years of musical study at High Schools in both theory and practice,
and this would have been supplemented by many hours of listening to both radio
broadcasts and recitals. Some of the Western students had not even heard N orth
Indian classical music until they attended the lectures at the School. Thus it was
necessary to adopt a completely different approach to the subject from that which is
usual in Indian universities. To the Western students Indian music was only incidental
to their main course of study and therefore the amount of time which they could
devote to it was severely limited. In view of this, it was necessary to concentrate on
broad principles and outlines rather than on the details which are the main concern
in Indian music colleges.
11
Preface
The critical attitude of the Western student provided a stimulus for the formulation
of many of the ideas expressed in this work. With his training in and experience of
Western music he has contributed new ideas and interpretations; and by his reluc
tance to accept traditional Indian explanations, frequently lacking coherence, he has
also provoked further enquiry into many topics. The question ‘why’ has been upper
most in his mind. ‘Why does Indian music have its present form ? Why are only
certain scales used in Indian music?’ To these and other similar questions the
traditional reply—‘because it was performed in this way by my teacher’—has been
unsatisfactory. To a large extent this work has been motivated by such questions
and attempts to provide some of the answers. In this respect, it is an exploration into
certain aspects of Indian music which have not hitherto received sufficient attention.
It is hoped that the reader will be stimulated to further enquiry.
12
Note on Transliteration and Pronunciation
Since this book is concerned primarily with present-day Indian music, terms are
generally given in their m odem Hindi forms in preference to the classical Sanskrit
forms. Exception is made in the following instances: (1) the Sanskrit form is used
when referring to Sanskrit treatises, their authors and the musical theory described
by them; (2) the common English spelling is used when referring to well-known
place names and personalities, for example, Delhi rather than the Hindi Dilli or the
Urdu Dehli—this follows the procedure adopted by Vincent Smith in the Oxford
History o f India (Oxford 1958); (3) Muslim names, other than those in common use in
English, are transliterated according to the system used in the Bulletin o f the School
o f Oriental and African Studies.
The DevnagrI (Devanagari) script is syllabic and all consonants carry the inherent
vowel a unless otherwise indicated. The principal difference between modern Hindi
and the classical Sanskrit forms is the omission in Hindi of this inherent a when in
final position (e.g. raga in Sanskrit and rag in Hindi) and frequently in medial
position (e.g. Mdravd in Sanskrit and Marvd in Hindi).
Approximate guide to pronunciation
(based on Received Standard English
Vowels Transliteration pronunciation)
short 3T a as in shut
i „ „ bit
u „ „put
n a Sanskrit vowel, in Hindi treated as a consonant r
+ vowel i and pronounced as in rip (with rolled r)
long a as in bath
i i ,, „ seed
u „ „ boot
*"V
e » Sê
ai in Hindi approximately as in bear (in Sanskrit as
in isle)
o as in boat
au Hindi as in saw (Sanskrit as in cow)
Consonants In English the difference between aspirate and
(without inherent d) non-aspirate forms is not generally recognised
whereas in Hindi and Sanskrit the majority of the
consonants have both forms.
13
unaspirated The English examples in this group are accompan
ied by a certain measure of aspiration which should
be eliminated for a more accurate representation of
the unaspirated Indian consonants.
k approximately as in baker
q derived from Arabic, it is a ‘k ’ sound produced as
far back as possible, i.e. uvular as against the velar k.
It has no aspirated form. In Hindi, often replaced
by k.
q- as in get
g
=5T c „ „ chat
j „ „jab
3 ti „ „ toe but with tongue curled back
z d ,, ,, do ,, ,, ,, ,, ,,
r not found in Sanskrit. An V sound produced by
drawing the tongue back and flapping it forward
t as in toe but with tongue against the teeth.
d ,, ,, do ,, ,, ,, ,, ,, ,,
T „ „pot
P
ST b „ „ bat
aspirated These can be approximated by exaggerating the

aspiration in the examples given above. They can
also be illustrated by the fusion of certain words
as below.
kh as in bajck hand
T gh „ „ sla/g heap
ch „ „ mu/ch hope
jh „ „ bri/dge hand
th „ „ carjt horse with tongue curled back
£ dh 55 roajd house ,, ,, ,, ,,
rh not found in Sanskrit. The aspirated form of r
th as in coajt hanger but with tongue against
teeth.
q* dh „ „ roajd house but with tongue against teeth.
ph » „ lea/p high
bh „ „ rujb hard
% h „ „ perhaps, a voiced h
1 f and d, their corresponding aspirates, th and dh, and the corresponding nasal are retroflex
or cerebrai sounds produced with the tongue curled back and pressed against the hard palate. The
English t and d are mid-way between these and the Indian dental t and d. The Indian th and th
should never be pronounced as in English thick, this', nor should the Indianph be pronounced as in
English physic.
14
Nasals
3? n as m smg
or n „ „ ni in onion
n „ „ running but with tongue curled back for the n
n „ „ now
*T m » » man
Semi-vowels Traditionally classified as a group, but in Hindi the

r and I are treated as consonants,
y as in yet
r the r is rolled as in the Scottish pronunciation of
road
1 as in light
v generally mid-way between the English v and w
and less emphatic than in never
Fricatives s as in show
s in Hindi generally pronounced as above (in Sans
krit with tongue curled back)
s as in sit
kh of Persian and Arabic origin, pronounced as in
the Scottish loch (approximately). In common
Hindi replaced by kh.
also of Persian and Arabic origin and is the
voiced equivalent of kh. In Hindi often replaced
by gh.
hT z as in zoo. Persian-Arabic origin. In common
Hindi often replaced by ].
f as in father. Persian-Arabic origin. In Hindi often
replaced by ph.
Others 5r: h voiceless /?, occurring in Sanskrit and Sanskrit
loan-words in Hindi
sr m a nasal, which may represent one of the nasal
consonants, in which case it is transliterated by the
appropriate consonant. Where it occurs before a
sibilant or a semi-vowel it is transliterated as
indicated (m).
a nasalisation of a vowel
For a fuller discussion of pronunciation see T. Grahame Bailey, Teach Yourself

Urdu, English Universities Press Ltd., London 1956.
15
Introduction to the Historical Background
JM odern N orth Indian classical music has its roots in ancient Indian music, but
appears to have acquired its present form after the 14th or 15th century a .d . Indian
musical theory is expounded in considerable detail in the Natyasdstra, probably the
earliest extant treatise on the dramatic arts, among which music is included. This
work, attributed to the sage Bharata, has been dated variously from the 3rd century
B.C. to the 5th century a .d . Some of the technical terms in present-day musical theory
and practice derive from this ancient source. Nevertheless, internal evidence shows
that the musical system of ancient India as described in the Natyasdstra differed
considerably from that of today.
The ancient melodic system was based on modes (jdti), each with its character
istic features, which were constructed on heptatonic series of notes (murcchana),
beginning on the successive degrees of two parent scales, Sadjagrama and Madhyama-
grdma. These scales were composed of intervals of three different sizes, comparable in
some respects to the major wholetone, minor wholetone and semitone of Just Inton
ation, which were expressed very approximately in terms of their highest common
factor—about a quartertone—called sruti.1 The musical intervals in the two parent
scales are described as being of four, three and two srutis, and since there were in
both parent scales three of the large intervals and two each of the medium and small
intervals, the octave comprised a total of twenty-two srutis. An interval of one sruti
was not considered musically satisfactory. The only difference between the two parent
scales was in the location of one single note which was one sruti flatter in the second
parent scale. In this period the sruti was a functional element since it was the only
distinguishing feature between the two parent scales.
Rag, which is the present basis of melody in Indian music, was not yet a technical
term in the Natyasdstra. It was apparently evolved during the centuries following for
it is first discussed in detail in M atanga’s Brihaddesi (c. 9th century a .d .) and later
expanded in Sarngadeva’s Sahgitaratnakara (first half of the 13th century a .d .).
This latter work is particularly interesting as it was written at the court of the
Yadava dynasty in the Deccan shortly before the Muslim conquest of this area and
is, to a large extent, free from Islamic influence. New conventions had evidently
already entered Indian music and rdgas had proliferated, for Sarngadeva mentions
264 of them.2 It is difficult to assess positively, however, whether the ancient music
1 Many scholars have given precise values for these srutis. Fox Strangways in Music o f Hindo-
stan, Oxford 1914, pp. 115-17, concludes that srutis are of three different sizes: 22, 70 and 90 cents.
2 Sahgitaratnakara, ‘Adyar Library Series’, II (2), 19. However, many of these are described as
‘ancient’ and were probably not current in Sarngadeva’s time.
16
based on the jdtis and the two parent scales was also in existence at this time, for the
Sangitaratnakara, like many other Indian musical treatises, does not always distin
guish clearly between current practice and antiquated theory.
The conquering Muslims encountered in India a musical system which was highly
developed and probably quite similar to their own. Their reaction to it was clearly
favourable.1 The poet Amir Khusraw, who was expert in both Indian and Persian
music at the court of ‘A la’ al-DIn Khilji, Sultan of Delhi (1296-1316), is unsparing
in his praise of Indian music,2 and his attitude is one which probably prevailed in the
Islamic world, for both al-Djahiz3 in the 9th century a .d . and al-Mas‘udi4 in the 10th
had commented favourably on it.
Music flourished in Islamic India in spite of the puritan faction, supported by the
Muslim legal schools, which believed that music was unlawful in Islam.5 However,
the gathering momentum of the Sufi movement with its unorthodox doctrines based
on the practices of ascetic and mystic groups, who found in music a means to the
realisation of God, more than compensated for the restrictions imposed by orthodox
Islam.
From Amir Khusraw’s time until well into the Mughal period, foreign music,
particularly from Iran, was commonly heard at the Indian courts together with
Indian music. Under these circumstances it is not surprising that Indian music was
subjected to new influences. Amir Khusraw, in spite of his dedication to traditional
Indian music, was a great innovator and is credited with the introduction of a number
of Persian and Arabic elements into Indian music: new vocal forms as well as new
mgs, tdls, and musical instruments including the sitar and tabla which are so pro
minent today. Of the vocal forms two are particularly important: Qaul, which is said
to be the origin o f Qawwali, a form of Muslim religious song, and Tardnd (or Tardna),
a song composed of meaningless syllables, both of which are still common today.6
During the reign of Sultan Muhammad b. Tughluq (1325-1351), music was appar
ently encouraged on a grand scale, although he was a ruler with strong religious
convictions. The traveller, Ibn Batuta, reports that the Sultan kept 1,200 musicians in
his service and had, in addition, 1,000 slave musicians.7 Similarly, Ibrahim Shah
1 Music apparently flourished in the Deccan under the Yadava kings to such an extent that, after
the Muslim conquest led by Malik Kafur (c. 1310), all the musicians and their Hindu preceptors
were taken with the royal armies and settled in the North. V. N. Bhatkhande, A Short Historical
Survey of the Music of Upper India, Bombay 1934, p. 11.
2 ‘Indian music, the fire that burns heart and soul, is superior to the music of any other country.
Foreigners, even after a stay of 30 or 40 years in India, cannot play a single Indian tune correctly.’
M. W. Mirza, Life and Works o f Amir Khusraw, Calcutta 1935, p. 184.
3 M. Z. Siddiqi, Studies in Arabic and Persian Medical Literature, Calcutta 1959, p. 32,
4 A. Sprenger, E l MasEidVs historical encyclopaedia, ‘Meadows o f Gold. . London 1841, p. 186.
5 H. G. Farmer, A History o f Arabian Music, London 1929, p. 20, discusses music in Islam. See
also M. L. Roychoudhury, ‘Music in Islam’, Journal o f Asiatic Society, Letters, Vol. XXIII, No. 2,
1957*
6 Kbyal, which is the most prominent type of song in classical music today, is also sometimes said
to have been invented by Amir Khusraw, but the evidence for this is inconclusive. Similarly, M. W.
Mirza, ibid., p. 239, draws attention to the fact that there is no mention of the sitar in Amir
Khusraw’s own writings, nor for that matter in any Indian treatises until much later.
7 Mahdi Hussain, Rehla o f Ibn Batuta, Baroda 1953, pp. 50-1.
2 17
Sharqi of Jaunpur (1401-1440) and Sultan Zain-ul-‘Abidm of Kashmir (1416-1467)
were both renowned for their patronage of the arts. A musical treatise (in Sanskrit),
Sahgitasiromani, was dedicated to Ibrahim Shah in 14281 and Zain-ul-4Abidin is
said to have been responsible for the composition of a treatise named M amak (?)
which is, unfortunately, not extant.2
In the following years music received further impetus from rulers, some of whom
were excellent musicians themselves. One of these was the Hindu Raja, M an Singh
Tomwar of Gwalior (1486-1516). His principal contribution was the rejuvenation of
the traditional form of song, Dhrupad (Sanskrit Dhruvapada), by his compositions in
Hindi,3 some of which are still said to exist today. M an Singh was also responsible
for the formulation of a progressive treatise in Hindi entitled M an Kautuhal, a work
which was compiled by the leading musicians of his court and incorporated many of
the innovations that had been introduced into Indian music since Amir Khusraw’s
time.
A contemporary of M an Singh, Husayn Shah Sharqi (1458-1528), initially Sultan
of Jaunpur, was also an excellent performer and an innovator, in importance perhaps
second only to Khusraw. He is credited with the introduction into N orth Indian music
of a new form of song, Khyal (khayal), which gave greater scope for improvisation and
technical virtuosity than did the traditional and austere Dhrupad. The rivalry between
the advocates of these two forms of song and their respective styles of performance
continued until the beginning of the 19th century when Khyal finally gained supre
macy.4
Sultan Sikandar Lodi of Delhi (1489-1517) was a bigot and in most respects a
strict follower of Quranic law. Yet he was himself a poet of considerable merit and
keenly interested in music. Under his patronage probably the first treatise on Indian
music in Persian, the Lahjat-i Sikandar Shdhi, was written. This was a traditional work
based on existing Sanskrit treatises.5
Before the efforts of Man Singh and Sikandar Lodi, musical treatises had always
been written in Sanskrit, a scholarly language which was beyond the comprehension
of most musicians, Hindu as well as Muslim. There was now a growing interest in
musical theory and especially in the systems of aesthetics with which it was associated
—the relationship of sound with sentiment or emotion (rasa), colour, the Hindu
deities, etc., as well as the visual representation of rags. This interest was particularly
notable during the reign of Ibrahim ‘Adil Shah II of Bijapur in the Deccan (1580-
1626), who by his patronage and enthusiasm for the arts attracted poets, musicians,
1 Abdul Halim, Essays on History o f Indo-Pak Music, Dacca 1962, p. 15.
2 Ibid., p. 18.
3 N. Augustus Willard, Music of India, Calcutta 1962, p. 67. Writing in the first half of the 19th
century, Captain Willard states that this kind of composition has its origin from the time of Raja
Man Singh, who is considered as the ‘father’ of Dhrupad singers.
4 Captain Willard says that in his time Dhrupad was not generally understood or relished and
its use seemed about to be superseded by ‘lighter compositions’ {ibid., p. 81).
5 The Lahjat-i Sikandar Shahi is discussed in some detail by Dr. Nazir Ahmad in Islamic Culture,
Vol. 28, 1954, pp. 410 ff. '
18
artists and architects to his court. He was himself a renowned poet and musician and
the Kitab-i Nauras (Sanskrit nava rasa—the nine emotions) contains a collection of his
poems intended to be sung in different rags.1 His reign is characterised by his liberal
views and his earnest attempts to integrate the opposing elements in Islamic and
Hindu philosophy.
The patronage of music reached its peak under the Mughal Emperors, Akbar
(1555-1605), Jahangir (1605-1627) and Shahjahan (1628-1658). M uch of Akbar’s
reign was devoted to the expansion and the consolidation of the M ughal Empire.
Nevertheless, he maintained a magnificent court at which literature, philosophy and
the arts occupied a prominent place. Music was presented on a lavish scale,2 and
Akbar himself is said to have been a prolific composer.3
The most famous musician of this period was undoubtedly Miya Tansen, around
whom so many legends have grown that it is now difficult to separate fact from
fiction. He was unquestionably a great musician as well as a composer. Several rags
still bear his name, M iya Malhar for example, and many of his songs are still sung
today. Another prominent musician at Akbar’s court was Baz Bahadur, the last
Muslim ruler of the state of Malwa, whose tragic affair with Rupmati, a singer and
dancing girl, has also become legendary.4 In the later part of his life, after he had lost
his empire, he became one of the leading musicians in Akbar’s retinue.
The Dhrupad style of singing was pre-eminent in Akbar’s time and the majority
of vocalists came from Gwalior, presumably following the tradition initiated by
Raja M an Singh, and it is in this city that Tansen is buried. Many of the instrument
alists, however, were foreigners who came from as far as M ashhad and Tabriz in
Iran and from H erat in modern Afghanistan.5
Jahangir’s court was perhaps even more opulent and ostentatious than A kbar’s
had been. As he too was a great patron of the arts (being himself skilled at painting),6
music continued to flourish. One of the principal musicians of his court was Bilas
Khan, the son of Tansen, whose compositions are occasionally heard today.
Shahjahan’s reign marks the culmination of the Mughal dynasty. The wealth of his
extensive empire, coupled with the conditions of comparative peace, permitted him
to maintain a magnificent court and to devote a great deal of attention to the arts.
While this does not appear to have been a period notable for innovation, the art of
1 This work has appeared in print with introduction and notes by Nazir Ahmad, published by
Bharatiya Kala Kendra, New Delhi 1956. It is interesting that the rags are referred to as maqdms, a
fact which suggests the similarity between the Indian and Arabic or Persian systems even at this
time.
2 According to Abu’l-Fazl ‘Allami, A ’in-i-Akbari, tr. H. Blochmann, Calcutta 1873, i, pp. 50-1,
for instance, the orchestra which played at the gateway of the Royal palace (naqqarakhand) had
more than sixty members.
3 Akbarnama, Beveridge, i, p. 50, quoted by O. C. Gangoly, Rcigas and Raginls, Bombay 1958,
p. 54.
4 This legend is the subject of a Persian manuscript by Afimad-ul-Umri, written in 1599, which has
been translated by L. M. Crump under the title of The Lady o f the Lotus, London 1926.
5 Abu’l-Fazl ‘Allami, op. cit., i, pp. 611-13.
6 Vincent A. Smith, The Oxford History o f India, Oxford 1958, p. 373.
19
music is said to have reached a polish and grace unprecedented in the past.1 The leading
musician of the period was Lai Khan, pupil and son-in-law of Bilas K han who,
presumably, continued the tradition of M iyI Tansen. He was a matchless Dhrupad
singer, and was frequently presented with large gifts by Shahjahan. Although foreign
musicians were still imported, their numbers had decreased since A kbar’s time.2
The 16th and 17th centuries are of great importance for the musical literature of
India. W ritten in 1550, the Svaramelakalanidhi of Ramamatya—a minister of Rama
Raja, prince of Vijayanagar3—focuses attention on the fact that the music of South
India, which had experienced relatively little Islamic influence, was evolving in its own
way and was beginning to acquire an independent character. This is corroborated
by the Ragavibodha of Somanatha (1609), although there is some evidence of his
contact with N orth India—for example, the occurrence of Muslim rag names in his
work. One of the most im portant treatises on the South Indian system was the
Catnrdandiprakdsikd of Venkatamakhi written in 1660, in which the classification of
ragas in terms of 72 basic scales (mela) was first advocated. This system still prevails
in South India.
Several important N orth Indian treatises were also composed during this period.
Of these the Ragatarahgini by Locana Kavl (of uncertain date),4 Sadrdgacandrodaya
and other works of Pundarika Vitthala (end of 16th century), Hridayakautaka and
Hridayaprakasa by Hridaya Narayana (c. 1660), and Sahgitapdrijata by Ahobala
(c. 1665) have considerable bearing on the history of present-day classical music.5
These works, with the exception of the Sahgitapdrijata, also follow the classifica
tion of ragas in terms of basic scales (mela), and for this reason can be more clearly
comprehended than the ancient system based on murcchand (which is followed in the
Sahgitaratndkara), where the tonic or ground-note of the mode is not explicitly
stated. Thus there is some confusion as to whether the base note of the murcchand,
the im portant note (amsa)6 or the final note of the mode (nydsa) should be considered
as the tonic.
1 Abdul Halim, op. cit., p. 38, quoting Faqirullah (Faqir Allah), Rag Darpan. Muslim University
Ms. f., 16a, dated 1661-1665.
2 Ibid., p. 43.
3 O. C. Gangoly, op. cit., p. 51. Ramamatya is said to be a descendant of Kallinatha who wrote a
commentary on the Sahgitaratndkara in the 15th century.
4 The date of this work is discussed by O. C. Gangoly, op. cit., p. 41>f.n. 3. The argument is as
follows: The colophon in the work itself gives the date of 1082 of the Saka era, i.e. 1160 a .d . The
occurrence in this work of Indo-Persian rags, some of which are said to have been invented by
Amir Khusraw, indicates that this date is too early. Further, there is a reference in the work to a
poet, Vidyapati, which could refer to the well-known poet who lived 1395-1440. The evidence
suggests that the earliest date of this work could be the second half of the 15th century. Bhatkhapcfo
A Comparative Study of some o f the Leading Music Systems of the 15th, 16th, 17th and 18th centuries,
p. 22, states that Hridaya Narayana has borrowed a whole section from the Ragatarahgini, and as
the date of Hridaya Narayana’s works is in little doubt the middle 17th century would appear to be
the latest possible date for the Ragatarahgini.
5 The Sahgitadarpana by Damodara Misra is another well-known work of this period written in
1625 a .d . Bhatkhande, A Short Historical Survey of the Music of Upper India, Bombay 1934, p. 26,
describes it as being as unintelligible and mysterious as the Sahgitaratndkara.
6 See p. 44.
20
By the second half of the 17th century we can be sure that the ancient musical
system as conveyed in the Natyasdstra was no longer in existence, and that the
prevailing system was very similar to that which pertains at the present time. The
treatises begin with the traditional description of the scale in terms of twenty-two
srutis. The srutis are, however, no longer functional as one of the two ancient parent
scales, the Madhyamagrama, is no longer in use.1 In spite of the mention of twenty-
two srutis, the octave seems to have been composed of twelve basic semitones.2 In the
Sadrdgacandrodaya, for instance, the octave is said to contain fourteen notes, but in
his description of the fretting of the vina (stick zither) Pundarika locates only twelve
frets, because, he says, the frets for the other two notes would be too close to their
adjacent frets on the fingerboard. He adds that if these two frets should be needed in
any rdga, the adjacent higher frets would be quite acceptable, as the difference of one
sruti will not make much of a difference in the general effect of the rdga?
A large number of musical treatises were concerned primarily with the iconography
o f ragas and were devoted to establishing familial relationships between ragas on
some extra-musical basis. In this very brief survey, we are necessarily obliged to
forego any mention of these.4
Shahjahan’s reign was followed by that of Aurangzeb (1658-1707). The latter was
fond of music and skilled in its theory, but he chose a life of asceticism in keeping
with the tenets of Islam, relinquished all pleasure and withdrew his patronage of the
arts. Musicians were obliged to leave the Mughal court and seek their livelihood at
the lesser provincial courts. It was only with the later Mughals, Bahadur Shah
(1707-1712) and M uhammad Shah (1719-1748), that music regained some of its
former glory. Although the reign of Muhammad Shah was beset with troubles and
the Mughal Empire was rapidly declining, he was keenly interested in music and was
an accomplished singer and composer. Largely as a result of M uhammad Shah’s
own endeavours and the compositions of his two leading musicians, Sadarang and
Adarang, Khyal finally came to the fore, and a large proportion of the modern
repertoire stems from this source.
This was not a fruitful period for musical literature. Bhava Bhatta wrote three
works at the end of the 17th century, but these are said to be largely in imitation of
Sahgitaratndkara? In the second half of the 17th century Faqir Allah wrote two
works in Persian, Rag Darpan and Man Kautuhal, the latter being to a large extent a
translation of the 16th-century Hindi treatise, Man Kautuhal, by Raja M an Singh,6
1 Bhatkhande, A Short Historical Survey, p. 25.
2 Locana uses only twelve notes in describing his ragas (see Bhatkhande, A Short Historical Survey,
p. 9). Similarly, Ahobala only uses twelve notes in describing his ragas although he gives the names
of nineteen altered (vikrit) notes {ibid., p. 27).
3 Bhatkhande? A Comparative Study, pp. 47-8. If Pundarika had indicated the lower adjacent frets
as a substitute, it could have been argued that the desired notes could have been achieved by the
technique of deflecting the playing string in order to raise its pitch, a technique which is commonly
used today.
4 Many of these are discussed in O. C. Gangoly, op. cit.
5 Bhatkhande, A Comparative Study, p. 69.
6 Halim, op. cit., p. 20.
21
In 1724 the Sahgitapdrijata was translated into Persian by Pandit D lnanath.1 These
translations were very necessary, for, while the Muslims took readily to Indian music,
the treatises and the words of the traditional songs were in Sanskrit and the Indian
vernacular languages, and were generally quite meaningless to Muslim musicians.
In addition, they were frequently based on Hindu religious and mythological subjects.
These must all have proved formidable barriers to the Muslims. While many songs
were composed in Persian, it is very likely that Muslim musicians were required to
sing traditional Indian songs, particularly at the courts of the more broad-minded
rulers such as Akbar and Ibrahim ‘Adil Shah II. It is equally probable too that Hindu
musicians were sometimes required to sing Muslim compositions in Persian, some
of which were based on religious Islamic themes. In either case the words can have
been of little significance to the musicians,2 and in practice the voice came to be
used more and more as a musical instrument, with words serving primarily to lend
colour and timbre to the music.
In the second half of the 18th century India was divided into several conflicting
factions, the most im portant of which were the Marathas, Mughals, Afghans and a
coalition headed by the Nizam of Hyderabad. It was just at this time too that the
British began to assert themselves in Indian politics. Musicians were dispersed to the
various courts and palaces of noblemen throughout the country, their fortunes, as
always, depending on the affluence of their patrons.
There was little sign of British interest in Indian music, except for a treatise written
by the Oriental scholar, Sir William Jones, entitled On the Musical Modes o f the
Hindus, which appeared in 1799. Two im portant treatises were written at the
beginning of the 19th century: the Hindi Sahgit-sar (c. 1800), compiled as a result
o f a conference of leading musicians in the court of the Jaipur Maharaja, Pratap
Simh Dev; and the Persian Naghmat-i-Asafi(1813), written by M uhammad Reza, a
nobleman of Patna. The latter has received considerable attention because it is said
to be the first ‘reliable’ authority in which Bilaval that is referred to as the natural
{suddh) scale.3 The fact remains that until about the 19th century the natural scale
described in N orth Indian texts was based on the ancient Sadjagrdma, comparable
to the D mode (the ecclesiastic Dorian). Today Bilaval that, comparable to the
Western major scale or the C mode (the ecclesiastic Ionian), is generally accepted as
the natural scale.4
1 Bhatkhande, A Comparative Study, p. 31,
2 Some of Ibrahim ‘Adil Shah’s songs in the Kitab-i-Nauras, composed in the Dakhani language,
are dedicated to the Hindu deities (mainly Sarasvatl and Ganes), others to Muslim saints (Sayyad
Muhammad Husayn-i-Gesu Daraz). Musicians of his court, whether Hindu or Muslim, would
presumably have been expected to sing all of them.
3 Bhatkhande, A Short Historical Survey, p. 35. However, G. H. Ranade, Hindustani Music,
Poona, 1951, p. 12, draws attention to the fact that, in his Hindustani Sangit Paddhati, III, p. 136,
Bhatkhande has written that Reza has nowhere referred to his notes as suddh.
4 It is tempting to think that this might be a result of Western influence, but this seems unlikely
in view of the widespread acceptance in India of Bilaval that as the natural scale. It should be noted
that very few traditional musicians have any familiarity with Western music, and most of them find
it completely alien.
22
In 1834 Captain N. Augustus Willard, an army officer attached to a small princely
state, wrote A Treatise on the Music o f Hindusthan, in which he drew attention to the
considerable gap that had grown between musical theory and practice over the
centuries.1 In the second half of the 19th century musical theory was rejuvenated in
Bengal. The publication of K. Goswami’s Sahgita Sara in 1868 was followed by
various publications by S. M. Tagore2 and a particularly im portant work by K.
Banaiji, Gita Sutra Sara (1855), written in Bengali. Banarjl made a serious attempt to
integrate musical theory and practice and his work is remarkable for its critical assess
ment of musical theory. In 1914 Fox Strangways wrote Music o f Hindostan, another
commendable attem pt to relate the numerous aspects of Indian music. The work shows
an extraordinary perception and grasp of the subject. Fox Strangways’s comments
on contemporary Indian music are particularly praiseworthy and his analogies with
Western music are often enlightening. But his discussions of ancient Indian music
must be viewed with caution as they contain some very basic misinterpretations.3
The beginning of the 20th century was, however, dominated by the works of Pandit
V. N. Bhatkhande. His first im portant work, Srimal-laksyasahgitam, was written in
Sanskrit and published in 1910 under the pseudonym of Catura Pandita. Although
Bhatkhande quotes from many prominent Sanskrit sources, it is quite clear that his
main intent is to reconcile musical theory with existing practice. This work was
followed shortly by the first of four volumes of a magnum opus in M arathi entitled
Hindusthdni Sangit Paddhati (hereafter abbreviated to H.S.P.) which was finally
completed in 1932 and later translated into Hindi.4 Bhatkhande here expands many
of the ideas expressed in Srimal-laksyasahgitam and introduces many new concepts
to explain the musical practice of his day. He traces the historical development of
rags through Sanskrit treatises and attempts to analyse and present a standard form
for each, while acknowledging divergent traditions. Bhatkhande’s second major
work, Kramik Pustak M alika, in six volumes (hereafter abbreviated to K .P .M .) was
published between 1920 and 1937 and was also later translated into Hindi.5 This
work is primarily devoted to the notation of more than two thousand traditional
songs in different rags and tals which Bhatkhande was able to collect from musicians
belonging to different gharanas (family traditions) throughout N orth India.6 K.P.M.
1 The works of both Augustus Willard and Sir William Jones have recently appeared in a second
edition as Music of India^ Calcutta 1962.
2 The article ‘Hindu Music’ has been reprinted from the Hindoo Patriot, 1874, in Hindu Music
from Various Authors, Varanasi 1965, compiled by S. M. Tagore.
3 For instance, the ancient Sadjagrdma is assumed to begin on the Ni note which he equates with
the Western note C: Music o f Hindostan, Oxford 1914 (reprinted 1966), p. 109. This would mean that
the Sadjagrdma was similar to the C mode or the Western major scale. In fact, it has since been
firmly established that it is equivalent to the D mode.
4 Published by Saiiglt Karyalay, Hathras 1956-7. All the references to H.S.P. in this work refer
to this translation.
5 Published by Sangit Karyalay, Hathras 1954-9. All the references to K.P.M. in this work refer
to this translation.
6 The following list of contributors and their provenance is given by L. N. Garg, the writer of the
preface of K.P.M., Vol. IV. The list is, however, incomplete and includes only those musicians who
gave permission for their names to appear in print. For instance, XJstdd Bundu Khan, a famous
23
also contains quotations from Sanskrit and Hindi sources on each of the rags
(numbering about 180) and brief verbal descriptions of them. In addition, Bhatkhande
gives his own interpretation of the musical characteristics of these rags: the ascending
and descending lines (aroh and avroh), a typical or ‘catch’ phrase (pakar) by which
each can be recognised (Vols. V and VI employ a slightly different method), and
at the end of each volume as many as twenty or twenty-five series of phrases
(svarvistar—extension of notes), compiled to illustrate the melodic contours of each
rag,
Muhammad Nawab ‘All Khan, a pupil of Bhatkhande, followed his preceptor in
that he too based his musical treatise, Ma'arif-ul naghmat, written in Urdu, on songs
which he had collected from practising musicians. In recent times there have been
many musical texts written in the Indian vernaculars which for the m ost part borrow
heavily from Bhatkhande’s works. A considerable number of publications on Indian
music have also appeared in English and other European languages, the standard of
scholarship often leaving much to be desired. A. A. Bake’s publications, although
they have not found expression in a m ajor work, are one of the noteworthy excep
tions.1 Herbert Popley’s The Music o f India (Calcutta 1950) is generally reliable
and is a useful guide to both N orth and South Indian music. A particularly
sdrahgi player of Indore who spent the last years of his life in Pakistan, is also said to have
contributed to Bhatkharnje’s collections (L. N. Garg, Hamare Sangit Ratna, Hathras 1957,
p. 479).
1. H. H. Hamid ‘All §ahib Bahadur Ruler of Rampur and follower of Tansen’s
descendants.
2. §ahibzada Sa‘adat ‘Ali Khan Sahib Rampur—follower of Tansen’s descendants.
Bahadur
3. Khan JjJahib Muhammad ‘All Basat Khan Rampur—descendant of Tansen.
4. Khan Sahib Muhammad Vazir Khan and Rampur—descendants of Tansen, and teachers
Amir Khan of His Highness.
5. Khan $ahib Muhammad ‘All Khan Jaipur—Manarang (son of Sadarang)
ghardnd.
6. Khan $ahib ‘Ashiq ‘All Khan Jaipur—Manarang ghardnd.
7. Khan $ahib Ahmad ‘All Khan Jaipur—Manarang ghardnd.
8. Khan §ahib Haidar Khan Dhar—pupil of Bahram Khan.
9. Khan §ahib Faiyaz Khan Baroda—Ranglle ghardnd.
10. Khan §ahib Amir Khan Gulab Sagar Baroda—instrumentalist.
11. Sri Raoji Buva Belbagkar Bombay—follower of ‘Abdullah Khan,
dhrupad singer.
12. Sri Eknath Pan<jit Gwalior—follower of Nathan PIrbakhsh’s
descendants and of khyal singer, Sankar
Paphit-
13. Sri Vi§nubuva Vaman De£pa$<je Gwalior—descendant of Vamanbuva, a
dhrupad singer.
14. Sri Rajabhaiya Puchvale Gwalior—pupil of Sahkar Pandit.
15. Sri Krisparao Gopal Date Gwalior
16. Sri Kri§nabuva Gokhle Miraj—follower of Amin Khan’s descendants.
17. Sri Kri§na Sastri Sukl Ujjain (Gwalior).
18. Sri Ganpatibuva Bhilvadikar Satara.
1 ‘The Music of India*, in Ancient and Oriental Music (New Oxford History o f Music, Vol 1),
London 1957, and ‘Indische Musik* in Die Musik in Geschichte und Gegenwart (Allgemeine
Enzyklopadie der Musik, Bd. 6), Kassel 1957.
24
valuable critical work, discussing the comments and theories put forward by Willard,
Bhatkhande and K. Banarjl, is H, L. Roy's Problems o f Hindustani Music (Calcutta
1937). Roy highlights the inadequacies of some of the present-day terminology and
gives suggestions for the reconstruction of musical theory. G. H, Ranade’s Hindu
stani. Music: An outline o f its Physics and Aesthetics (Poona 1951) is particularly
useful for its analysis of the acoustics of the drone. A number of other writers have
been concerned primarily with precise intonation in Indian music. Their work is
based on the acoustic properties of individual intervals as determined by mathemati
cal ratios, without reference to the varying musical context in which those intervals
occur. In addition they attem pt to explain modern Indian practice in terms of ancient
musical theory. This ‘school1 was initiated by K. B. Deval, The Hindu Musical Scale
and the Twenty-Two Shrutees (Sangli 1910), and E. Clements, Introduction to the Study
o f Indian Music (London 1913), and has recently been followed by A. Danielou,
Northern Indian Music (London 1949, 1954).
On the other hand, there have been several valuable works on the history of Indian
music. Foremost among these are, once again, the writings of Bhatkhande. His two
monographs in English—A Short Historical Survey o f the Music o f Upper India
(Bombay 1934), originally a speech delivered at the First All-India Music Conference
at Baroda in 1916, and A Comparative Study o f some o f the leading Music Systems o f
the 15th, 16^/?, 17//* and 18th Centuries (Bombay, n.d.)—present in concise form some
of the historical material which extends throughout his other writings. O. C. Gangoly’s
Ragas and Rdginis (Bombay 1935, reprinted 1948) is another scholarly work which
deserves to be mentioned. As a source of reference it is of considerable value, but it
does not go into details of musical theory. The works of Swami Prajnanananda,
Historical Development o f Indian Music (Calcutta 1960) and A History o f Indian
Music, Vol. 1 (Calcutta 1963), also contain valuable material. M ost of the historical
research has been based on Sanskrit sources. Perhaps the only useful work based
on Islamic sources is a collection, Essays on History o f Indo-Pak Music (Dacca 1962)
by Abdul Halim. M uch work still remains to be done in this field.
In this brief resume of the musical literature of the present century many works
have not been mentioned.1 Considering the body of material on Indian music, it is
surprising that so very little is concerned with the analysis of present-day Indian
music.
This century has seen fundamental changes in the preservation and presentation
of N orth Indian classical music. The traditional system of patronage has been
gradually disappearing and musicians now earn their livelihood mainly by public
recitals, radio broadcasts, gramophone records, and teaching in schools and music
colleges, and only incidentally by private recitals and individual tuition. New devices
have already been evolved to cope with the formal atmosphere in the concert hall
where rapport with the listener is not so easily achieved. Many musicians have been
1 Some of these have been discussed by Harold S. Powers in ‘Indian Music and the English
Language: A Review Essay’, Ethnomusicology, ix, January 1965.
25
experimenting with microphone techniques and, since, as a result, they are now less
concerned with producing a large volume of sound, there has been greater emphasis
on tone production. This is once again a period of exploration and change, and it will
certainly influence the form of Indian music in years to come. A t the present time,
however, there is no reason to believe that the basic fundamentals of Indian music
are in any danger of distortion in the foreseeable future.
26
An Outline o f Present-Day
North Indian Classical Music
P resent-day classical music is directly descended from the court tradition of earlier
centuries and some of the prominent musicians of today can still trace their ancestry
back to the court musicians of the Mughal period.1 Under the patronage system musi
cians were continually vying for the favours of the court and this gave rise to a highly
competitive atmosphere in which virtuosity, invention and showmanship played a
vital part. These characteristics still apply today. The musician aims to impress as
well as entertain, but above all to convey an aesthetic experience. He is not rendering
a traditional piece in a stereotyped manner, but refashioning his musical material
afresh in each performance. Although a traditional song or melody often serves him
as a basis, it is usually very short and in performance is elaborated and varied, and
repeated statements of it are interspersed with improvisations. Thus the length of the
performance is, to a large extent, determined by the inventiveness of the musician.
There are four m ain aspects of Indian music to be considered:
1. Main melody line. The Indian musical scheme is essentially monodic—it has a
single melody line with an accompaniment.2 The voice is usually thought to be the
most effective carrier of the melody line, not because it is also capable of conveying
verbal content, but because of its flexibility and expressive properties. However, any
instrument can be used for this purpose, some naturally being more suitable than
others. The following are the most prominent melody instruments: the sitar, a
long-necked plucked lute with frets; the surbahdr, a larger version of the sitar; the
sarod, another plucked lute with a shorter neck and without frets; the sarahgi, a
bowed lute; the basri, a side-blown bamboo flute; and the shahndf (shenai), a double
reed wind instrument similar to the oboe, but without keys. Many other instruments
are also used; some, like the violin and the clarinet, have been borrowed from the
West.
1 For instance, Ustdd Vilayat Khan whose background is mentioned in Appendix B on p. 186.
2 Duets (jugalbandi), in which there are two carriers of the melody line—two voices or two instru
ments, who generally perform alternately—are becoming increasingly popular.
27
An Outline o f Present-Day North Indian Classical Music
2. Drone. The melody line is generally played against a fixed, unchanging drone
which is based on the tonic, its octave and its fifth or fourth.1 This is usually played
on a tamburd (tanpitra), a long-necked lute with four or five strings which has no frets
and consequently sounds only the open-string notes. The drone may also be produced
on a hand-pumped harmonium (sur-peti). The shahncfi is often accompanied by other
drone shahna’is.
3. Accompanying melody line. A vocalist is accompanied by a secondary melody line,
usually played on a sarangi or a harmonium, which echoes the phrases produced by
the singer. The sarangi is usually played by an accompanist, while the harmonium
is often played by the singer himself. When the vocalist pauses, the accompanying
instrument assumes momentarily the role of the main melody carrier.
4. Percussive line. This is usually produced on the tabid, a pair of small kettledrums
struck with the hands. Occasionally, a two-ended barrel-shaped drum, pakhvaj
(pakhdvaj) or mridang, may be used instead. The shahna'i is generally accompanied
by another type of kettledrum, the khurclak or dukar, also played in pairs. The
percussive instrument serves primarily as a time-keeper, but is also used for rhythmic
variations and improvisations.
Many musical instruments fulfil more than one function. The sitar, for example,
not only carries the melody line, but also has special strings (cikari) for supplying its
own drone, and in addition has sympathetic strings (tarab) which provide an echo,
in some ways like the effect produced by an accompanying instrument.
Indian classical music has two fundamental elements: rag, the melodic framework,
and tdl, the time measure.
r a g
The term rag has no counterpart in Western musical theory. The concept of rag is
based on the idea that certain characteristic patterns of notes evoke a heightened
state of emotion.2 These patterns of notes are a fusion of scalar and melodic elements,
and each rag can be described in terms of its ascending and descending lines (which
may involve ‘turns’) as well as its characteristic melodic figures in which certain
intervals are emphasised and attention is focused on particular notes. M ore than two
hundred rags are extant and each is a melodic basis for composition and improvisa
tion. Most of the rags have been in existence for several centuries and have evolved
to their present form as a result of successive interpretations by generations of
musicians.
A performance of a rag usually begins with an dldp, a kind of improvised prelude
in free time in which the melodic characteristics of the rag being performed are
clearly established and developed. It is rendered on a melody instrument or by the
voice, and is usually accompanied by a drone. The vocal dldp may also be accom
panied by a secondary melody instrument. The instrumental dldp tradition is very
1 At the present time variant drone tunings are also used (see p. 187).
2 The word rag is derived from the Sanskrit root rahj or raj = to colour or tinge (with emotion).
28
prominent today and the dldp generally consists of a number of sections, some of
which, like for and jhald, are played against a pulse or beat but without fixed metre.
A t the conclusion of the dldp a composed piece set in a particular tdl is introduced.
TAL
The term tal, perhaps best translated as 'time measure’, is conceived as a cycle. It
has two principal aspects: (1) quantitative, meaning the duration of a cycle measured
in terms of time units or beats (mdtra), which are generally held to be in three tempi
(lay)—slow (vilambit), medium (madhy) and fast (drut); and (2) qualitative, meaning
the distribution of stresses or accents within the cycle. These stresses occur at different
levels of intensity: the principal stress at the beginning of the cycle (sarn); secondary
stresses within the cycle (tali); and then there is a negation of stress (khdli) which
always occurs at points where a secondary stress may be expected but is consciously
avoided.1 The following illustrations show the quantitative and qualitative patterns
of three prominent N orth Indian tdh. Following Bhatkhande’s system, X represents
the sarn, the numbers 2, 3 and 4 the talls and 0 the khdlis:
Ex. 1.
(a) Tintdl (Tritdl)
Time units 1 2 3 4 5 6 7 8 9 10 11 12
Stresses X 2 0
(b) Ektdl
Time units 1 2 !3 4 5 6 ; 7 8 9 10 11 12
1\
Stresses X \
0 2 0 3 4
(c) Jhaptdl
Time units 1 2 3 4 5 !6 7 8 9 10
Stresses X 2 !0 3
The metrical framework of each tdl is represented by a basic drum pattern, thekd,
which is a fixed sequence of drum-syllables produced on a pair of tabla.2 These sounds
are produced by striking different parts of the two skins on the drum heads and are
symbolised by mnemonic syllables such as dim, dhin, nd, tin, ke, ghe, etc.3 A common
thekd of Ektdl, for example, is:
Ex. 2. Ektdl
1 2 _ 3 4 5 6 7 8 9 10 11 12
dhin dhin dhage tirakita ttx na kat ta dhage tirakita dhin na
X 0 2 0 3 4
1 Most of the common North Indian tdh have an even number of time units, the prominent
exceptions being Rupak and Tivrd which have seven units. The khdli frequently occurs midway
between two ‘positive* stresses creating something of the effect of an ‘up’ beat against the ‘down’
beat of the sarn and the tails. When keeping time the khdli is usually indicated by a wave of the
hand, while the sam and the talls are marked by claps.
2 Certain tdh are played primarily on the pakhvdj, which has its own drum syllables. The basic
pattern of a tdl is then called thapiyd.
3 A description of drumming techniques is found in A. H. Fox Strangways, The Music o f
Hindostan, p. 225 ff., and in W. Kaufmann, Musical Notations of the Orient, Bloomington 1967,
pp. 218-63.
29
It is in the composed piece that rag and tdl meet on common ground. In instru
mental music this piece or tune is called gat and in vocal music, set to words, it is
called ciz (cij). The gat or clz is not only in a particular rag but also has a fixed
relationship with the metre of a particular tdl. In instrumental music, especially that
on plucked stringed instruments such as the sitar and sarod, the gat is constructed on
percussive patterns obtained when the instrument is plucked. The following pattern,
called Majid (Masit) Khanigat, is a very common example:1
Ex. 3. M ajid Khani gat
Time units 12 13 14 15 16 1 2 3 4 5 6 7 8 19 10 11
Percussive pattern
Stresses
JJVJ J J J J ii J £ i J J ! J J J
3 X 2 I0
While the melody of the gat is based on a rag, the percussive pattern is a feature
not of rag, but of tdl. It is the counterpart on a melody instrument of the thekd on
the drums. In practice these patterns are not rigidly maintained and are commonly
embellished, but the relationship with the basic pattern is nevertheless preserved.
In vocal music the song {ciz) serves the same purpose as the gat, providing a fixed
relationship between rag and tal. There are at present several different types of
songs—khyal {khyal), thumri, tarana, dhrupad, etc.—each with their individual forms
and styles of performance. They are sometimes also associated with particular rags
and tdh. The text of the ciz is generally traditional, particular themes tending to be
associated with specific types of songs. The words are not generally noted for their
poetic content and are relatively unim portant in classical music. The voice is used
rather like a special instrument, capable of varying its timbre through the enunciation
of different syllables.
The composed piece in both instrumental and vocal music generally has two
sections, sthdyi {astai) and antra. The former is the main part of the composition and
is said to be usually limited to the lower and middle registers, while the antra extends
from the middle to the upper registers.2 The vocal composition may sometimes have
additional sections {sancari, abhogmdbhog), and in its full form may have a number o f
verses and extend over ten or more cycles of the tal. However, since the main function
of the composition is to provide a frame of reference for the tdl, only the first verse
extending over one cycle is really essential,3 the others being used by the musician
as and when needed, either to introduce variety or to draw attention to a different
part or aspect of the rag.
1 It is composed of two equal parts, from 12-3 and 4-11. The melody, beginning on beat 12,
leads to a climax at 1 and tapers away to 3. The second part of the melody builds to a false climax at
9 and tapers away to 11, the first part resuming at 12.
2 H.S.P. I, p. 41.
3 In the bara khyal, sung in very slow tempo (e.g. J = 12), even this much is not usually used, for
the full cycle of the tal may last more than a minute and the song is too long to repeat in its tradi
tional form. Here the relationship with the tdl is established only one or two time units (matras)
before the sam, the remainder of the cycle being devoted to improvisation. In the short fragment
of the composition which leads to the sam (called mukhyd), only the last few syllables of the full
text may be used.
30
The composed piece is not the primary focus of attention in N orth Indian classical
music. This role is occupied by the melodic and rhythmic extemporisations which
the musicians introduce during the course of their performance. The composition
serves as a springboard for these and a frame of reference to which the musicians
periodically return. Thus the form is similar to that of the Rondo, the composition
alternating with the improvisations.1 It must be stressed that the melodic improvisa
tions are not variations on the composition itself, but elaborations of the different
features of the rag phrased against the metre of the tal. Similarly, the rhythmic
extemporisations are not variations of the basic pattern of the tal but are phrases
played against the metre of the tal and regularly go across its stress patterns. The
improvisations of the melody instrument (or voice) and those of the drums are gener
ally undertaken successively, the metre of the tal being maintained by the non
improvisator who plays his basic pattern softly in the background. Occasionally,
both may improvise at the same time, but this would probably occur late in the per
formance when the basic patterns have already been heard several times and the
listener is, to some extent, able to supply the underlying patterns for himself.
Traditionally it is the melody instrumentalist or the vocalist who is the leader of the
ensemble inN orth Indianclassicalmusic.2It is he who determines the extent to which the
drummer may improvise, if at all. In recent times the drummer has been getting an
increasing share of the improvisation, though there are some musicians who still prefer
their drummers to be merely accompanists and do not permit them much licence.
The improvised variations may begin at any point of the tal, they may continue for
any part of one cycle or several cycles and the return to the composition m ay be
accomplished at any point, provided that the original relationship between the
metre of the tal and the composition is maintained. In practice many variations begin
either on the sam 0 1 *shortly after, and are very frequently concluded either at the sam or
just before the beginning of the composition (e.g. in Ex. 3 just before time unit 12).
Whilst the form of N orth Indian classical music resembles that of the Rondo, the
successive cycles generally increase in intensity, thereby creating the effect of an
upward spiral. This is accomplished by the development of melodic ideas,3 the
increasing complexity of both melodic and rhythmic variations, and the accelerating
tempo which frequently culminates in a powerful climax.
1 The degree of creativity in these extempore passages is not easily assessed, for in playing the
same rag and tal again and again, musicians acquire musical habits and evolve favourite phrases
which may recur from time to time. It is, however, when the musician is performing beyond his
normal capacity that the music becomes ‘alive’.
2 The roles are reversed in what is referred to as a ltabla solo’ where a melody instrument and
drone may accompany the drums. Here the melody instrument serves as time-keeper, repeating a
short tune (lahra> similar to gat) against which the tabid player improvises. In the normal instru
mental performances of a rag it is quite usual for the melody instrument to assume the role of time
keeper at appropriate moments, to give an opportunity for the drummer to improvise.
3 As, for instance, in the gradual expansion of the range of the rag, A more detailed discussion of
the development of melodic ideas can be found in N. A. Jairazbhoy, ‘Svaraprastara in North Indian
Classical Music’, Bulletin of the School of Oriental and African Studies, Vol. XXTV, part 2, 1961,
pp. 307-25.
31
II
Basic Elements o f Theory
A rag does not exist in any precise form in the sense that a symphony can
be said to exist in score, but is a complex of latent melodic possibilities.
Although this seems to suggest an amorphous quality, each rag is an
independent musical entity with an ethos of its own, which becomes
manifest through recognisable melodic patterns. In the course of time a
corpus of technical terms has been evolved by theorists and musicians in
order to convey some idea of the nature of rags. Since these technical
terms are used primarily to supplement musical practice they are not
always precise enough for purposes of analytical study. Therefore, in the
following pages, as we consider the salient features of rags, it will be
necessary to discuss not only the pertinent technical terms but also to
extend the discussion to related musical principles.
N O TE S
SVAR In N orth Indian musical theory seven notes (svar) are recognised. In their
Hindi form, the names of these notes are Sadj (or Khadj), Risabh, Gdndhdr
(or Gandhar), Madhyam, Pancam, Dhaivat and Nisad (or Nikhad); or in the
commonly used abbreviated form, Sa, Re (or Ri), Ga, Ma, Pa, D ha and Ni. It
is these abbreviations that are used throughout this work, with the occasional
addition, for the convenience of the Western reader, of the note’s scale
degree in brackets. The Indian nomenclature is comparable to that of
Western tonic-solfa: there is no absolute or fixed pitch attached to the
notes, and the ground-note (the note which serves as the point of reference
of the scale) is called Sa, irrespective of its pitch. Once the pitch of the
ground-note has been established, however, it remains unchanged through
out the performance of a rag as there is no modulation in Indian music.
ACAL O f these seven notes, Sa and Pa (I and V) are ‘immovable notes’ {acal
svar)—they have no flat or sharp positions and Pa is always a perfect fifth
CAL above the Sa. The remaining five notes are ‘movable notes’ (cat svar).
These each have two possible positions, a semitone apart. One of these is
32
S U DD H called suddh (pure) which is comparable to the ‘natural* of the West. In
the suddh scale, Bilaval, composed of Sa, Pa and the five movable notes in
their suddh position, the distribution of tones and semitones corresponds
to that in the Western major scale.1
When the movable notes are not in the suddh position, they are called
VIKRIT vikrit—altered. In the case of Re, Ga, Dha and Ni (II, III, YI and VII)
they are a semitone lower than their suddh counterparts and are called
KOMAL komal—soft, tender. The altered M a (IV), however, is a semitone above
TIVR the suddh position, and is called twr—strong, intense.
The terms komal and twr are not exactly comparable to their Western
counterparts, fiat and sharp, as they apply only to specific vikrit notes,
whereas in the West every note has a flat and a sharp form. The Sa and Pa,
being immovable, cannot have either komal or tivr form s; nor can a
komal note be referred to as the tivr of the note below, which in the
Western use of flat and sharp is common practice. (The semitone above C
may be called either C# or Db, depending on the circumstances, but in
Indian music Re komal is not referred to as Sa tivr.) Notwithstanding this
difference, in this work we are using the symbol \? to indicate komal and #
to indicate tivr, and, where necessary to avoid confusion, t| to indicate suddh,
The full series of intervals in the gamut are set out below:
Suddh svar Sa Re Ga Ma Pa Dha Ni (Sa)
Vikrit svar Reb Gab Ma# Dhab Nib
These are represented in Western staff notation as follows, the Sa being
arbitrarily equated with the C but not implying its absolute pitch:
Ex. 4.
Suddh svar Vikrit svar
Sa Re Ga Ma Pa Dha Ni Sa Reb Gab Ma# Dhab Nib
This system of nomenclature has wide acceptance in India, and is

generally used by Bhatkhande (though he uses different symbols to repre
sent komal and tivr).2
1 In its present-day application the suddh concept does not entail the idea of parent scale from
which other scales are derived, but serves only as a standard for comparison.
2 Another system of nomenclature is also sometimes used in India, and is referred to by
Bhatkhande (K.P.M. II, p. 12) as being used primarily by vocalists. In this tradition, the higher
position of the movable notes is referred to as tivr and the lower position as komal. Here the term
tivr should be translated as the upper of two alternative notes, not as sharp, and komal as the lower
rather than as flat. A considerable amount of confusion is caused by the co-existence of these two
systems. Of the many examples which could be quoted, those from record sleeves are the most
obvious. For instance, on H.M.V. ALP 2312, the rag Jaijaivanti is described as having all seven
sharp notes in ascent. This is completely misleading and may even suggest to the Western reader
that the ground-note can be made sharp in certain rags. The writer has evidently equated the suddh
of Bhatkhapde’s system with tivr of the other. This is only justified in application to Re, Ga, Dha
and Ni. In fact, the ascending line of rag Jaijaivanti has ‘natural’ intervals.
q 33
R E G IS T E R S
N orth Indian classical music is not, of course, limited to one octave, and
the same names apply to the notes in the other octave registers above and
sthan below. There are three registers (sthdn—position; or saptak—aggregate of
saptak seven) generally recognised, each extending from Sa to the Ni above:
madhy middle (madhy); high (tar) which is here indicated by a dot above the note
tar name, e.g. Sa (I); and low (mandr) which is indicated by a dot below the
mandr note name, e.g. Ni (YU). Although musical theory usually acknowledges
only these three registers which are based on the natural limitations in the
range of the voice and most Indian instruments, the very low register
ati- (atimandr), indicated by two dots below the note name, is sometimes used
mandr by players of stringed instruments, especially the sitdr and the surbahar.
atitar The very high register (atitar) is rarely heard.
IN T O N A T IO N
While the present-day N orth Indian gamut is comparable to the twelve-
semitone octave of the West, some discussion on the subject of intonation is
necessary. In the classical music of N orth India there is no need for equal
temperament, since the factors which lead to this—changing harmonies
and the system of keys—do not apply. Moreover, the technique of temper
ing notes by the use of beats is generally unknown, and since it is un
common to find a number of melody instruments playing together, no
objective standard of tuning is in general use. The only Indian instrument
with fixed intonation is the harmonium which is often used for accom
panying singers, but even here the precise tuning varies with each instru
ment. In general, intonation is governed by the individual musician’s
feeling for intervals. Except for the simple consonances of the ground-note,
octave, fifth and fourth, these only approximate to a twelve semitone
standard. Electronic analysis has confirmed that there is variation in
intonation from one musician to another, as well as for a single musician
during the course of a performance.1
Apart from this unconscious variation in intonation, there are musical
traditions in N orth India which consciously recognise that in a few parti
cular rags one or two notes are flatter or sharper than that which they
conceive of as the standard in the rags as a whole. Bhatkhande refers to
these traditions on a number of occasions; for instance, when discussing
the rag Asavri he says, ‘Some say that the D ha (VI) of Asavri is flatter than
1 For further discussion on intonation see N. A. Jairazbhoy and A. W. Stone, ‘Intonation in
present-day North Indian classical music’, Bulletin of the School o f Oriental and African Studies,
Vol. XXVI, Part 1, 1963, pp. 119-32.
34
that of the rag BhairvV. However, he does not appear to give much
credence to this and prefers not to go further into the m atter.1
There is, however, one special case where subtle distinctions in intona-
ando - tion are particularly noticeable. This occurs when a note is subjected to a
LAN siow shake or an exaggerated vibrato (andolan or gamak), either as a
gamak decoration or as a functional feature in certain rags.2 It is in this context
that certain musicians use the term sruti to indicate the subtle intervals
Sr u t i produced as a result of this oscillation in pitch. They do, however, main
tain that these microtonal deviations from the ‘standard’ intonation may
only be used in oscillation and may not be sustained as a steady note.3
In the introductory chapter we have already suggested that the sruti,
which was the basis of distinction between the two parent scales in ancient
India, had certainly lost its original significance by the 17th century.
In m odern times certain musicologists and musicians still attem pt to apply
the old twenty-two sruti system to present-day music, while others go so
far as to assert that the present-day gamut can only be explained in terms
of forty-nine or even sixty-six different intervals. The fact remains that
srutis are no longer functional, that is they are not a primary basis of
distinction between rags.
Bhatkhande attempted in his early works to relate the twelve semitones
to the ancient srutis as follows: 4
Suddhsvar Sa Re Ga Ma Pa Dha Ni
Sruti 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Vikritsvar Ret? Gab Ma# Dhab Nib
The twelve-semitone system, however, is clearly at odds with the twenty-
two sruti system since some of the semitones are composed of one sruti
and others of two srutis.5 In his later writings Bhatkhande contradicts
this earlier opinion when he says, ‘To distinguish between two rags on the
basis of the difference of only one sruti would not be acceptable to any
present-day vocalist or instrumentalist’.6 If this statement is applied to the
1 H.S.P. IV, p. 428. He continues, ‘But I can see no reason why we should get involved in these
minute intervals. In current practice, the [following] rule always obtains: “svarasawgatyadhinani
svarasthanani nityasah” [The position of notes depends upon the notes they are combined with].*
Elsewhere, H.S.P. IV, p. 584, he is more explicit: ‘When a note is connected with lower notes, then
it is noticed to be lower [in pitch], and when with higher notes then it is seen to be raised. This
difference is noticed only by people with acute perception. That is why wise people do not like to
exert themselves unduly with the trouble of trying to ascertain the minute intervals.’ We shall be
discussing this question of intonation in Chapter VIII.
2 An example of this can be heard in the rag Darbari on the accompanying record.
3 This view has been stated by Bare Ghulam ‘All Khan. For further discussion see Chapter VIII.
4 K.P.M. II, pp. 10-11.
5 It is sometimes stated that the octave contains twenty-four srutis, presumably so that each
semitone can have two srutis.
6 K.P.M. VI, p. 21. This remark is reminiscent of that made by Pupdarika Vitthala more than
350 years ago which has been referred to earlier (see p. 21).
35
above scheme representing the semitones in terms of srutis, it would mean
that musicians could not distinguish between rags having a m inor third
(Gab) and a major third (Gan) or a minor seventh (Nib) and a major
seventh (Nin), for the difference between these is only one sruti.
Obviously this is not so. Bhatkhande goes on to say that there is no abso
lute measure of sruti available to him and that he recognises that the
position (intonation) of a note in any one rag fluctuates with the changing
context in which it occurs.1
The gamut is a conceptual standard and, though it is derived from
musical practice, it cannot take into account all the minute deviations from
the norm, many of which are quite unconscious. Thus we are obliged to
accept the twelve semitone standard, while making allowances for minor
variations, conscious as well as unconscious.
A L T E R N A T IV E N O TES
The suddh and vikrit varieties of each of the five movable notes are alterna
tives and do not normally occur as consecutive steps in a melodic sequence.
Thus, in principle, the complete musical series will consist of the two
immovable notes, Sa and Pa, and one of each pair of alternatives, Ret; or
Reb, Ga>i or Gab, Man or Ma#, Dhab or Dhab, and Nib or Nib. In general,
Indian music can be described as ‘diatonic’ in the sense that the successive
steps of a scale are different degrees, rather than as ‘chromatic’ where
the steps could include both alternatives of any note.2 But many rags are
quite complex and have both forms of one or more movable notes. These
usually occur each in their own particular melodic context from which
the other is excluded. It sometimes happens that a skilful musician will
merge the two contexts so that the two forms of a note may be heard in
succession. This generally requires some preparation of ground, as in the
1 K.P.M., ibid. This was written during the latter part of Bhatkliande’s life by which time he had
obviously modified his earlier views on srutis.
2 In this work the terms diatonic and chromatic are used in this rather specialised sense. Here
diatonic does not refer necessarily to scales whose steps are only wholetones and semitones. When
applied to a heptatonic scale, chromatic indicates the use of both alternatives of a note as scalar
steps and implies the corresponding omission of one of the other degrees, usually that just preceding
or just following the alternatives. Thus in the following illustration scale A would be diatonic,
in spite of its augmented-second intervals, while scale B would be chromatic because both alter
natives of Ga are used and Re, the second note, is omitted. The fact that scale B has an interval of a
minor third—virtually the same as augmented second—has no bearing on the subject.
following example illustrating the successive use of both forms of Ni
(VII):1
Ex. 5.
Ni Sa, Ni^ D ha Nitl Sa Ni^ Dha mk, S*~Nit Dha N ilf, Ni^ Dha
Ip-F If ‘'P
There is, however, a major exception to the scheme of alternative notes
as outlined above. This is provided particularly by the Lalit group of rags
in which both forms of the fourth, Mah and Mas, commonly occur as
consecutive steps. These will be discussed in greater detail in a later chapter.
We may note here that it is primarily the two M a’s which sometimes
provide exceptions to the rule that the suddh and vikrit positions of a note
are alternatives.
SC A LE SPEC IES
While many rags have both forms of one or more of the five movable
notes, there are some from which one or two notes are omitted entirely—
the Sa alone by definition cannot be omitted. Such rags are described as
transilient. In N orth Indian theory rags are sometimes classified according
to the number of notes they contain, the classes thus obtained being
ja ti known as jdtis (species): rags with all of the seven notes are called sampurn
sam puri^(com plete), those with six, sadav (or khadav) and those with five, audav.
sadav These terms are equivalent to the Western hepta-, hexa- and pentatonic.
audav It should be noted that alternatives do not count here as separate notes:
in a heptatonic rag any or all alternatives may be used as accidentals;
similarly, in a pentatonic rag any alternatives of the five notes of the rag
may be used as accidentals. The rag Vrinddvm (Brindabni) Sarang, for
instance, is classified as pentatonic although both alternatives of N i (VII)
are used :2
Ex. 6. rag Vrindavni Sarang
N i Sa, R e, M a Pa, Ni Sa Sa, Nit> Pa, Ma Re, Sa
J J i J
1 This is often an oversimplification of what actually occurs in practice. The circumstances are
complicated by the fact that musicians have been preparing the ground for this sort of movement in
certain rags perhaps for several generations. Consequently, there are instances when the preparation
of the ground is taken as read. Some musicians avoid this apparent chromaticismentirely, but
probably for the majority this is something which can be done in a few specific instances, and then
only with nicety.
2 K.P.M. III, p. 496. Bhatkhande does not explain the exact significance of commas in his nota
tions of rags. The commas are not used in his notations of clz where the duration values are regu
lated by the tal. In the atap-type of phrases of the svarvistar, the aroh-avroh and pakay, which are
37
The most im portant system of classifying rags is, however, in terms of
Th a t heptatonic scales, that (that), which are discussed in some detail in the
next chapter.
M E L O D IC M O V E M E N T
It is not enough to define a rag in terms of mode or scale alone, as a number
of rags have the same notes, yet each maintains its own musical identity.
When we examine different performances of the same rag we find that,
allowing for divergence of tradition and the possibility of experimenta
tion, not only are the same notes consistently used, but also particular
figurations or patterns of notes occur frequently. The most characteristic
pakar pattern of notes in a rag is described as pakar, a ‘catch’ phrase by which
the rag can be easily recognised. This is inevitably a subjective concept
as rags are not generally limited to just one pattern and the ‘catch’ phrase
of a rag varies, to some extent at least, with the interpretation of the
svar- musician, A more complete delineation of a rag is obtained in the svarvistar
vistar — a series of phrases devised to show the various note-patterns which are
permissible in, and characteristic of, the rag. These, too, are subject to
varying interpretations.
These patterns of notes can be described in terms of their melodic
varn movement, vani. Sanskrit treatises have recognised four types: sthayi—
sthayi steady, unchangeable; aroh (arohi)—ascending; avroh (avrohi)—descending;
aroh and sancari—wandering, i.e. a mixture of ascent and descent. Only the
avroh terms aroh and avroh are now commonly used in the description of rags
sancari and refer to the most characteristic ascending and descending lines of a
rag, whether step by step or including irregular movements (sancari varn)
if these are essential functional features of the rag. For instance, in the rag
Des (Des) the common aroh (ascent) is a step by step pentatonic movement
not regulated by tal, a comma could indicate either a pause or the lengthening of the preceding note.
The former seems highly improbable in view of the frequent occurrence of the comma which, if
interpreted as a pause, would disjoint the melodic line, as can be seen in the following typical
example {K.P.M. Ill, p. 23):
rag Bhupall
Thus it would seem more reasonable to interpret it as extending the time value of the preceding
note. There are no specific breathing indications except, by implication, at the end of variations in
the svarvistar which are marked by bar lines, and we presume that breath may be taken as required.
In this work the notes preceding the comma have been given double the time value of the other
notes; however, there is no evidence that Bhatkhande intended such precise values and our notation
system has been adopted for the convenience of the reader.
38
—which can be described as directional transilience—while the common
avroh (descent) is heptatonic and has two irregular turns at x and y :
Ex. 7. rag Des
Sa, Re, M a Pa, Ni Sa Sa Nik Dha p a> Qha M a Ga, R.e Ga Sa
M 1M f c r -
These turns, which are characteristic features of certain rags, are

vakr designated by the term vakr (crooked or oblique), and the note from
which this oblique movement begins, i.e. Pa and Re in the example above,
v a k r s v a r is called vakrsvar1 (oblique note).
On the basis of the given aroh and avroh, the rag Des could be described
as having a pentatonic ascent in which the Ga (III) and D ha (VI) are
omitted, and a heptatonic descent in which Pa (V) and Re (II) are vakrsvar
and Nib (Vllb) replaces Nit* (VIIb). The terms aroh and avroh do not always
refer to the typical ascending and descending lines in a rag, but are some
times used to indicate specific upward or downward movement. The dual
implications of these terms occasionally create confusion. For instance,
in describing the rag Kamod, Bhatkhande states that the Ma# (IV#) is
used only in the aroh, and yet when he gives the typical aroh and avroh
of the rag, the Ma# occurs in both the lines :2
Ex. 8. rag Kamod

Sa Re, Pa, Ma# Pa, Dha Pa, N i Dha Sa
■ in n
IP
Sa, N i Dha, Pa, Ma# Pa Dha Pa, Ga Ma# Fa, Ga M a Re Sa
ap
There is a further complication in the description of this rag, for although

the Ga (III) is omitted in the typical aroh line, it occurs in the ascending
phrase G a Mab Pa (III IVb V) which is in the typical avroh line,
and Bhatkhande describes this rag as being heptatonic in both aroh and
avroh. It thus becomes necessary to distinguish clearly between the use of
the terms to indicate the typical ascending and descending fines (which
1 According to Bhatkhande, only Re is vakr in the rag Des (.K.P.M. III, p. 521). However, in the
svarvistar of this rag (pp. 760-1) the Pa is frequently vakr, as in the example above.
2 K.P.M. IV, p. 92.
39
may involve oblique movement), and the use of the terms to indicate the func
tion of each individual note appearing in an ascending or descending context
within a rag, It is the latter which we must now discuss in greater detail.
There are two aspects to a note which belongs in a simple ascending move
ment: that it is approached from a lower note, and that the note following
is higher. When these two conditions are fulfilled, it can be said that the
note is clearly an ascending note. However, in certain rags it is permissible to
approach a note from below, but the following note may not be a higher one.
Here only one condition is fulfilled and it is a m atter of interpretation
whether this note should be considered as ascending or not. In fact, in
both Indian musical theory and practice, it would not be considered an
ascending note, as it leads downwards. This is commonly taken for granted
in the system, and a note prohibited in ascent may generally be approached
from below but must be followed by a lower note. The descending line in
rag Des provides a good illustration of this, where, although the D ha (VI)
and G a (III) are prohibited in ascent (except in certain phrases to be
discussed later), the descending line has turns leading upwards to these
notes (see * and y, Ex. 7). On the other hand, if a note may not be ap
proached from below, but the following note is a higher one, that note
is commonly thought to be in an ascending line; for instance, in the rag
Kamod the Mas (IV#) can only be approached from above and is always
followed by a higher note (see Ex. 8, w). These three possibilities are shown
in the following examples, where L stands for a lower note, H for a higher
note, and the note under consideration is represented by N :
Ex. 9.
(a) (b) (c)
Pa Dha Ni Pa Dha Pa Ni Dha Ni
L N H L N L H N H
(a) N is clearly a direct ascending note.

(b) N is not an ascending note.
(c) N is an incomplete ascending note,and since it can only be ap
proached by a turn from above (as in Ex. 8, w) it can also be referred
to as an oblique (vakr) ascending note.
These same three possibilities also occur in relation to descent:
Ex. 10.
(a) (b) (c)
Ni Dha Pa Ni Dha Ni Pa Dha Pa
-j ■ 1 "
I i N L H N H L N L
40
(a) N is a direct descending note.
(b) N is not a descending note (cf. Ex. 9c).
(c) N is an oblique descending note (cf. Ex. 9b).
There still remains one further distinction to be made. In some rags a
note which is generally omitted in the ascending line may nevertheless
occur as an ascending note in certain characteristic figures: for instance,
in the rag Kamod (Ex. 8), where the Ga (III) is normally omitted in the
ascending line but may be used as an ascending note in a melodic figure
usually found in the descending line:
Ex. 11. rag Kamod

(Paf) Ga Ma Pa, Ga M a Dha Pa, Ga Ma Re Sa
Here the use of the G a as an ascending note limits the possibilities which
may follow. In step by step movement the D ha (VI) may not be exceeded
and the phrase is only felt to be completed by the cadential fragment v.
A determining feature of this movement is that it does not extend into the
next octave but turns back on itself. Thus Ga in Kamod is an oblique
ascending note (as it can only be approached from above) which occurs
only in a discontinuous ascending figure, and can be described as a
discontinuous, oblique ascending note. The Ma^ (IV s) in this rag is not
usually used in ascent, but occurs as a discontinuous, direct ascending note
in the above example.
Similarly, in the rag Des, both Ga (III) and D ha (VI), while omitted in
the continuous, direct ascending line (see Ex. 4), may be used as discon
tinuous direct ascending notes, the former in melodic figures beginning
and ending on the Re (II), the latter on the Pa (V):
Ex. 12. rag Des

Re Ga Ma Pa Dha M a Ga Re Pa Dha Dha Pa
Bhatkhande describes Des as heptatonic in both ascent and descent, with the
qualification that the G a and D ha are generally omitted in ascent,1 but in
fact, the continuous ascent of Des is pentatonic, the G a and D ha being
used only occasionally as discontinuous direct ascending notes.
1 K.P.M. Ill, pp. 250-1. These discontinuous direct ascending notes can be heard in the rag Des
on the accompanying record.
41
IM P O R T A N T N O TE S
In every rag two notes, in theory, are given greater importance than the
vadi others. These notes are called vadi—sonant, and samvadi—consonant.
sam vad! According to Bhatkhande the prime character of a rag appears in the
vadi.1 The vadi is that note which is sounded clearly again and again, a
note which is superabundant in a rag.2 The samvadi is described as being
a note used less than the vadi but more than the other notes in the rag.
The samvadi should not be near the vadi as it will tend to detract from the
importance of the vadi. Ideally it should be a perfect fifth away or, if that
note is not present in the rag, it should be one of the adjacent notes, the
fourth or the sixth, preferably the former.3 These definitions of vadi and
samvadi appear to relate primarily to frequency of occurrence, but statistics
applied to Bhatkhande’s own notations reveal irreconcilable inconsist
encies.4 Obviously much depends on the interpretation of the key phrase
‘sounded clearly again and again5, which Bhatkhande does not clarify.
He seems aware of the inadequacy of his definition and quotes a divergent
view from the Gita Sutra Sara by K. Banarjl (Bannerjee) in which the
author questions the validity of these terms.5
Much of this difficulty seems to arise from the fact that rags have
different facets which are successively developed in the course of a
v i S r a n ti performance. In this connection Bhatkhande equates vadi with visranti
svar svar (or maqam sthan), terminal or resting notes, when he states that singers
choose different notes on which to end their melodic phrases, momentarily
presenting each of these notes as vadi, finally returning to the prescribed
vadi without detriment to the rag.6 Thus in a particular rag there are
several im portant notes which may be emphasised either by frequency of
occurrence or by their use as terminal notes. In theory the vadi is chosen
because it is the most important note in the characteristic phrase {pakar) of
that rag. There are, however, further qualifications. In all rags, the Sa (I)
is a vitally im portant note, both as a frame of reference and as a melodic
terminal. Yet the Sa is not a good candidate for the position of vadi
because it is a feature common to all rags and gives no indication of the
1 H.S.P. I, p. 20.
2 K.P.M. II, p. 14 and K.P.M. VI, p. 23.
2 H.S.P. I, p. 22.
4 In the svarvistar of rag Yaman, as set out in K.P.M. II, pp. 487-8, there are 62 Sa, 83 Re,
70 Ga, 54 Ma, 74 Pa, 47 Dha and 45 Ni. On a statistical basis, Re should be vadiand Pa samvadi.
Bhatkhande, however, gives Ga as vadi and Ni as samvadi. In the other rags examined there is also
a similar deviation between the most often used notes and Bhatkhande’s given vadi and samvadi.
This is discussed further by A. N. Sanyal, Ragas and Raginis, Calcutta 1959, p. 20.
5 H.S.P. I, pp. 79, 80. Banarjl gives an example of the rag Yaman in which some say Pa is vadi,
others Ga or even Re and Ni, suggesting that, in the hands of an expert, there may be even greater
latitude. The important notes of this rag are discussed in Appendix B on p. 205.
6 K.P.M. V, p. 49.
42
character of a particular one. The same applies, although to a lesser extent,
ti m e to the Pa (Y). Further, Bhatkhande’s choice of vadi is often influenced by his
theory time theory which is an attem pt to relate the musical characteristics of a rag
to its hour of performance.1 In this connection, he divides the octave into
p u r v a n g two parts, purvang, first portion, the lower tetrachord Sa to M a (I to IV)
UTTRANGor the pentachord Sa to Pa (I to V); and uttrdng, second portion, the
upper tetrachord Pa to Sa (V to 1), or the pentachord from M a to Sa (IV
to i). According to his theory, in the rags performed between noon and
midnight the purvang is emphasised, i.e. the vadi is in the lower tetrachord;
while in the rags performed between midnight and noon the uttrdng is
prominent, i.e. the vadi is in the upper tetrachord.
This theory tends to influence the choice of vadi in Bhatkhande’s system.
For instance, in the rag Tilak Kamod the Ni (VII) is very prominent and is
considered the vadi by a number of musicians. Bhatkhande fully recognises
the importance of this note in Tilak Kamod when he says that the quality
of the Ni in this rag is so spectacular that nearly everyone recognises it
from the (particular) way this note is used.2 Tilak Kamod is, however,
sung at night and according to Bhatkhande’s theory should have its vadi
in the lower tetrachord. In K.P.M. Bhatkhande gives the vadi as Re (II)
and the samvadi as Pa (V),3 but in the H.S.P. he says that, according to
experts, the Re is weak in descent4 and gives the vadi as Sa (I).5
From the foregoing discussion it is apparent that the concept of vadi
and samvadi is not quite consistent with present-day musical practice.
The terms have been used in the musical treatises since the Ndtyasastra
where vadi—sonant, samvadi—consonant, vivddi—dissonant and anuvadi
—assonant (i.e. neutral) represent a general theory of consonance which
is now either forgotten or has at least lost its earlier significance as Fox
Strangways has pointed out.6 The terms, however, have persisted to the
present time. The original concept appears to have been quite reasonable.
Only perfect fourths and fifths were recognised as consonant, while the
semitone and/or perhaps the major seventh was recognised as dissonant.
The other intervals were considered assonant. These terms were thus
1 Bhatkhande’s time theory has been described in Rdgas and Rdginls, by O. C. Gangoly, Bombay,
reprinted 1948, pp. 90-2. The time theory of rags is a controversial subject and there are several
different attitudes which may briefly be expressed here. There are some who will not tolerate a rag
at any but the prescribed time. Bhatkhande is not so dogmatic, but states that a particular rag
sounds especially beautiful at a particular time. Some musicians look at this matter in an entirely
different light; they feel that if a particular rag is performed well it will create an atmosphere of a
particular time of day or night. Finally, there are those who believe that the time theory has no
application to present-day practice and Banaijl, quoted in H.S.P. I, p. 75, says that the tradition of
performing rags at particular times of the day and night is ‘purely imaginary’.
2 H.S.P. I, p. 243.
3 K.P.M. III, p. 297.
4 H.S.P. I, p. 250.
5 H.S.P. I, p. 243.
6 Fox Strangways, The Music o f Hindostan, p. 114.
43
intended to express the phenomena of consonance and dissonance as
conceived in that period. Obviously consonance and dissonance were
particularly significant in relation to the important notes in a mode (jati).
Amsa These im portant notes were designated by the term amsa. Bharata, the
author of Natyasdstra, says, ‘That note which is the amsa, that note is vddV,1
indicating that the amsa is the sonant note whose consonance and
dissonance are particularly important, not that vadi is a synonym of amsa.
But later writers have equated the two terms, and so vadi has come to
mean im portant note and the term amsa has now become redundant.
This has led to some confusion. Whereas in Bharata’s time modes fre
quently had several im portant notes (arnsas), and indeed there was one,
Sadjamadhyamd, in which all the seven notes were considered important,
the present-day rags can have designated only one vadi and one secondary
im portant note, samvadi. The ancient samvadis comprised the consonant
fourth and fifth, while the present samvadi refers to the second most impor
tant note in a rag, which, to preserve the importance of the vadi, is removed
from it by generally a fifth or fourth, not necessarily perfect intervals,2 or
perhaps by a sixth.3
vivad! The terms vivadi (dissonant) and anuvddi (assonant) are also occasion-
a n u v a d I ally used at the present time, especially by theoreticians. Vivadi as ‘disputing’
is particularly meaningless in the present context in which the minor
second and the major seventh have a very prominent place in the system.
Bhatkhande explains vivadi as that note which, when used in a rag, would
varjit- damage it, and refers to it as varjitsvar—omitted note. He concedes that
svar the vivadi may, however, be used by expert singers and players without
detriment to the rag.4 Here again the precise meaning of the term re
mains unclear. Are all the omitted notes called vivadi, or just those notes
which may occasionally be used by experts, but are not essential to the
rag? In discussing the rag Kamod,5 he says that sometimes Nib (VHb) may
be used in descent as a vivadi note, indicating that it is the latter meaning
1 Ndtyasastra, ‘Kashi Sanskrit Series* (No. 60), prose following si. 20, chapter 28.
2 K.P.M. Ill, p. 612. In rag Pilu, for example, the vadi is given as Gab and the samvadi as Nib—an
augmented fifth. The same applies to the rag Mdrvd where vadi and sarpvddi are given as Reb and
Dhab. Some musicologists are disturbed by the fact that these two do not form a perfect interval
and give Dhab and Gab as its vadi and samvadi. V. N. Patvardhan in Rag Vijndn, Vol. II, p. 1,
discussing rag Mdrvd, says, ‘Reb is prominent in its lower tetrachord {purvang), Dha in its upper
tetrachord (uttrafig). . . . Sometimes one also pauses on Ga, because Ga makes a consonant
(samvadi) relationship with Dha. But if this is done often it gives the appearance of the rag Puriyd.
. . . It is customary to give Reb and Dha as vadi and samvadi of Mdrvd, but seen from the point of
view of the sastras (treatises) it is not possible for Reb and Dhab to be sapwddi (i.e. consonant) to
each other. For this reason, in our opinion it is proper to accept Dha as vadi and Ga as saipvddi.’
These comments reflect the confusion which prevails among musicologists regarding the interpreta
tion of these terms. A further discussion of the rag Mdrvd will be found in Appendix B on p. 202,
3 Some musicians also accept the third as sarpvddi.
< K.P.M. II, p. 14.
s K.P.M. IV, p. 92.
44
that he has in mind. When this vivadi or accidental is used with sensitivity,
it is considered particularly beautiful—a far cry from its original meaning
of dissonant. The term anuvddi still refers to the notes in a rag other than
vadi, samvadi and vivadi, though these may, in the present period, include
the perfect fourth or fifth of the vadi.
To summarise, Bhatkhande’s choice of vadi for a rag is influenced by
three factors:
1. It should be an im portant note in the characteristic phrase of the rag.
2. It should belong to the correct part of the octave in relation to his
time theory.
3. Sa (I), and to a lesser degree Pa (V), are less meaningful as vadi than
the other notes because they give little indication of the character of
the rag and so become vcidi only when there is no other reasonable
possibility to fit his time theory. It will be seen that m uch depends
on the validity of the time theory. This is difficult to assess, but the
fact that the theory is widely accepted in India suggests that it is
reconcilable, at least to some extent, with the time of day at which
rags are traditionally performed.1 We shall have more to say about
the time theory in the chapter following.
SU M M A R Y
This discussion of technical terms can be concluded with a summary of the principal
features by which a rag may be distinguished from others:
1. Basic notes used (that).
2. Transilience (sampurn, sadav, audav).
3. Emphasised notes (vadi, samvadi).
4. Ascending and descending lines (aroh, avroh):
(a) alternative notes used as accidentals (vivddil);
(b) directional transilience;
(c) oblique movement (vakr).
5. Register of emphasis (sthan-mandr, madhy, tar).
6. Shakes (andolan) and intonation (sruti).
These factors will be discussed in the following pages.
1 There are, of course, differing traditions regarding the time at which rags should be performed
and no time theory can satisfy all of these.
45
Ill
That
* •
I n N orth Indian classical music rags are generally classified into groups according to
scale. These scales are called that (that, thath—framework) or sometimes mel (mela).
In Bhatkhande’s system the term that applies only to those scales which fulfil certain
conditions:1
1. A that must have seven notes.
2. The notes must be in sequence Sa Re G a M a Pa D ha and Ni (whether suddh
or vikrit position—both versions of a single note being forbidden).
3. A that does not have separate ascending and descending lines (as do rag's).
4. A that does not have any emotional quality (in contrast with rags which, by
definition, have the power to convey emotions).
5. Thdts are named after prominent rags in order to make them easy to remember
and recognise, whether or not these rags are heptatonic.
In summary it can be said that, of all possible musical scales, only heptatonic,
diatonic2 scales are called that.
In accordance with these conditions, thirty-two scales are possible: one with all
natural notes, five with one altered note, ten with two altered notes, ten with three
altered notes, five with four altered notes and one with all five altered notes. With a
few exceptions to be discussed later, these thirty-two scales could provide the frame
work for the classification of the rags used at the present time.
Each of these thirty-two scales is related to five others which differ from it in only
one note. Thus the suddh or ‘natural’ scale {Bilaval that) is related to the five other
scales which have only one altered note. Each of these, in turn, is connected with five
scales, one of which is Bilaval, while the remaining four each have an additional
altered note. Similar relationships extend throughout the system, and for convenience
the thirty-two scales (as well as three other im portant scales to be considered later
in these pages) are set out in full on p. 47. Their sequence is discussed and their
relationships further defined in Appendix A, The System of 32 Thdts, on p. 181.
1 K.P.M. II, p. 12.
2 The terms diatonic and chromatic are defined on p. 36, f.n.2.
46
A1 Kalyan that B1 Cl
Sa Re Ga Mag Pa Dha N i Sa Sa Re Ga Mag Pa Dha Nil* Sa Sa Re Gab Mag Pa Dha N i Sa
A2 Bilaval that B2 C2
Sa Re Ga Ma Pa Dha N i Sa Sa' Re Gab Ma Pa Dha Ni Sa Sa Re Ga Ma Pa Dhab N i Sa
A3 Khamaj that B3 C3
Sa Re Ga Ma Pa Dha Nib Sa Sa Re Ga Ma PaDhab Nib Sa Sa Reb Ga Ma Pa Dha Nib Sa
A4 Kaf t that B4 C4
Sa Re Gab Ma Pa Dha Nib Sa Sa Reb Gab Ma Pa Dha Nib Sa Sa Re Gab-Mag Pa Dha Nib Sa
A5 AsavrT that B5 C5
Sa Re Gab Ma Pa Dhab Nib Sa Sa Re Gab Mag Pa Dhab Nib Sa Sa Re Gab Ma Pa Dhab Ni Sa
A6 BhairvT that B6 C6
Sa Reb Gab Ma Pa Dhab Nib Sa Sa Reb Gab Ma Pa Dhab Ni Sa Sa Reb Ga Ma Pa Dhab Nib Sa
A7 B7 C7
Sa Reb Gab Mag Pa Dhab Nib Sa Sa Reb Ga Mag PaDhab Nib Sa Sa Reb Gab Mag Pa Dha Nib Sa
AS Ton that B8 C8
Sa Reb Gab Mag Pa Dhab Ni Sa Sa Reb Gab Mag Pa Dha N i Sa Sa Re Gab Mag Pa Dhab N i Sa
A 9 PurvT that B9 C9 Bhairav that

Sa Reb Ga Mag Pa Dhab Ni Sa Sa Re Ga Pa Dhab Ni Sa Sa Reb Ga Ma Pa Dhab N i Sit
A10 Marva that BIO CIO

Sa Reb Ga Mag Pa Dha N i Sa Sa Reb Ga Ma Pa Dha N i Sa Sa Reb Ga Mag Pa Dha Nib Sa
D1 D2
Sa Reb Gab Ma Pa Dha N i Sa Sa Re Ga Mag Pa Dhab Nib Sa
33 34 35
Sa Reb Ga Ma Mag Dha Ni Sa Sa Reb Ga Ma Mag Dhab Ni Sa Sa Reb G a'M a MagDhaNib Sa
That
• *
These scales are implicit in Bhatkhande’s definition of that, yet they are n ot all
mentioned by him.1 Having defined that, he states that nearly all the rags used in
N orth Indian classical music can be distributed among the following ten thdts:
1. Kalyan (No. A l) 6. Marva, (No. A10)
2. Bilaval (No. A2) 7. K dfi (No. A4)
3. Khamdj (No. A3) 8. Asavri (No. A5)
4. Bhairav (No. C9) 9. Bhairvi (No. A6)
5. Purvi (No. A9) 10. Tori (No. A8)
Bhatkhande does acknowledge the existence of rags in scales other than the above,
but, as these are relatively few, he ascribes them to one or other o f the above ten
thats as follows:
rag Lalit (No. 33) ascribed to Mdrvd that (No. A10)
rag Basant ( Vasant) Mukhdri (No. C6) ascribed to Bhairvi that (No. A6)
rag Anand Bhairav (No. BIO) ascribed to Bhairav that (No. C9)
rag Ahir Bhairav (No. C3) ascribed to Bhairav that (No. C9)
At the present time there are several other rags not mentioned by Bhatkhande,
whose scales do not readily fit into the ten that system. Some of these are given below:
rag Madhukdnt (No. C4) rag Ahir Lalit (No. 35)

rag Madhuvanti (No. C l) rag Lalit (No. 34)
rag Nat Bhairav (No. C2) (common version today)
rag Carukesi (No. B3) rag Kirvdni (No. C5)
rag Simhendra Madhyama (No. C8) rag Patdip2 (No. B2)
About twenty of the thirty-two scales are actually in use today and the number
appears to be increasing. Some, like Carukesi, Simhendra Madhyama and Kirvdni,
have recently been borrowed from the South Indian classical system which has many
more scale possibilities than the N orth Indian system since it admits chromaticisms
foreign to the North. The South Indian system acknowledges seventy-two scales
(melakarta) which include the thirty-two we have been discussing. The remainder
involve the use of the augmented second and sixth (which in N orth Indian terms may
be expressed as Re# and Dha#) and the diminished minor third and seventh (Gabb
and Nibb). The South Indian system of seventy-two scales was first introduced, largely
as a theoretical exercise, by VenkatamakhI in the Caturdandiprakdsika (1660), only
nineteen of them being currently found in practice.3 Subsequently, songs4 have
1 H. L. Roy, in Problems o f Hindustani Music, p. 118, states that he wrote of these thirty-two
scales to Bhatkhande in 1932 who agreed that these scales would be adequate for the needs of
present-day North Indian music. However, Bhatkhande’s system of ten thats had been well estab
lished by this time and one can understand his reluctance to introduce a new system.
2 Bhatkhande does refer to this rag as PradlpkJ or Patdlpki and ascribes it to Kdfifhat (No. A4),
since in his version there is also a Nib (VHb). At the present time it is commonly heard with only
Nib (Vllb).
3 Caturdandiprakdsika, The Music Academy, Madras, 1934, p. 52.
4 The instrumental tradition in South India is based largely on vocal compositions.
48
That
actually been composed in all seventy-two, with the result that a greater variety of
scales is used in South Indian music than in the North. We may expect enterprising
and progressive N orth Indian musicians to continue to adopt some of the South Indian
scales, particularly those which are diatonic. It is thus valuable to extend the North
Indian system beyond Bhatkhande’s ten thats and to accept the thirty-two diatonic
scales as the basis for the classification of rags in N orth India.1
In addition to these thirty-two scales there are three rags in which there is appar
ently a form of chromaticism. In these, the two versions of Lalit and Ahir Lalit,2
both the natural and the sharpened fourth (Mab and Ma#) are used in succession.
This is a particularly N orth Indian development since they are not included in
the seventy-two South Indian scales. There are indications that this is not really
‘chromaticism’ and that the Ma# is, in effect, a diminished fifth which would perhaps
have been called Pab were it not that Pa is considered an immovable note. This view
can be justified to some extent by a consideration of alternative notes. If the Ma# is
indeed an augmented fourth we should be able to find some evidence among the
mgs to show that it temporarily replaces the natural fourth (Mat]). Similarly if it is an
alternative for the fifth (Pa) and temporarily replaces it, it would be quite legitimate to
consider it a diminished fifth (Pab). There are many rags in which the Ma# is in fact
an augmented fourth and temporarily replaces the natural fourth. The rag Ramkali
of Bhairav that (No. C9) is one such example. Bhatkhande gives its ascending and
descending lines as follows:3
Ex. 13. rag Ram kali

Sa Ga, M a Pa, Dhab, N i Sa Sa N i Dhab p a> pa .j^b Dhab pa Ga, Mat] Reb Sa
Initially the natural fourth is followed by the fifth (at x) but later (in the phrase j )
the fifth is preceded by the augmented fourth. The two fourths, however, do not
occur together in sequence. It is evident that the augmented fourth temporarily
replaces the scalar natural fourth of Bhairav that in the phrase y which is constructed
around the Pa, just as Nib temporarily replaces the scalar NiN in this same phrase.
In certain other rags, however, the Ma# is really a diminished fifth and is an alterna
tive for the Pa. In the rag Bhairvi, for instance, the first two phrases (a and b) in the
following example are permissible and are often played,4 but the third phrase (c) is
not generally acceptable:
1 This is particularly significant in view of the fact that certain well-known musicians, among
whom Pandit Ravi Shankar is one, find the ten that system quite inadequate and prefer to use the
South Indian seventy-two melakarta system as a theoretical basis, in spite of the fact that it includes
a number of ‘chromatic’ scales.
2 A rag said to have been invented by Pandit Ravi Shankar.
3 K .PM . IV, p. 312.
4 These phrases can be heard on the record of Bhairvi played by Ustad Ali Akbar Khan, H.M.V.
ALPC 2.
4 49
That
Ex. 14. rag Bhairvi
(a) (b) (C)
Dhab PaMa GabRe# Ga# Sa Re^Sa DhabMa#Ma#GabRe# Gab Sa Re** Sa Dhab Pa Ma$Gab Re# Gab Sa Reb Sa
--Q- ......
1— ----- rjrrr i-""
M mk - J jrJ X G
X
It is interesting that musicians do not appear to distinguish between the Ma# as an
augmented fourth and the Ma# as a diminished fifth, thereby reinforcing the argu
ment for a twelve semitone system.1 Because this distinction is not made, it is not
always possible to determine whether, in a particular instance, it is an alternative for
the natural fourth or the natural fifth, especially as there are rags in which the Ma#
appears to function in both ways at different moments. In the rag Lalit Pancam, for
example, which Bhatkhande gives in Bhairav that (No. C9), both the following two
phrases appear:2
Ex. 15. rag Lalit Pancam
(a) (b)
Sa Ni Dhab pa( Ma8 p a> Qa Ma# Ga R et Sa Ni Dhab Ma# Ma#
E g #■ mp o p
¥
In the former the Ma# (IV#) presumably replaces the natural fourth of the that, while
in the latter it replaces the fifth.
The fact that in the two versions of the rag Lalit and in Ahir Lalit as well the Ma#
and the Ma# may be used in succession (the natural fifth is absent in these rags)
suggests quite clearly that the Ma# is, in fact, a diminished fifth:3
Ex. 16. rag Lalit (Bhatkhande’s tradition)
Reb N i D ha, Ma#iDha Ma# Ma#[ Ga, Reb, Sa
m ¥ i
Bhatkhande classifies Lalit as a hexatonic {sadav) rag, in which the fifth is omitted,
implying that the Ma# replaces the Ma#, There is some justification for this as there is
1 In Just Intonation the augmented fourth and diminished fifth are, respectively, 590 cents and
610 cents. In Pythagorean Intonation they are 612 cents and 588 cents. It is in the Tempered twelve
semitone system that these two are the same at 600 cents (Willi Apel, The Harvard Dictionary o f
Music, London 1960, p. 362).
2 K.P.M. V, p. 362 and p. 509, var. 6. The function of the Ma# is extremely complex in this rag
and Bhatkhande also gives phrases in which a succession of semitones is used, e.g. (V, p. 509,
var. 10);
Sa Ni Dha'’ Pa MaS M a\ Ga Pa Ga Sa
3 K.P.M. IV, p. 490.

50
That
• •
a clearly noticeable tendency to omit the natural fourth in the flow of movement;
yet it is used as a resolution since the secondary drone is generally tuned to this n ote:
Ex. 17. rag Lalit

Ga Ma# D ha Ni D h a, M a# D h a, Mail, Ma#
This, however, does not necessarily indicate that the Mati is omitted to avoid chroma
ticism. In the same rag the ground-note Sa (which has no chromatic counterpart) is
also omitted in the flow of movement, while it, too, occurs at the end of a phrase as
a resolution:1
Ex. 18. rag Lalit

Ni Rc^ G a Ma# D ha Ni Re^ Ni D h a M ai D ha Ma# Mai) Ga Ma# G a Re Sa
It would seem reasonable to consider the Ma# in these rags as a diminished fifth
rather than an augmented fourth. The rags could then be described as heptatonic and
diatonic, in spite of the occurrence of the two M a’s in succession. The scales of these
three rags cannot be accommodated in the thirty-two that scheme, but would require
another complete system of thirty-two scales having a diminished fifth. In view of the
limited number of such scales in use this hardly appears to be justifiable at the present
time. Admittedly, chromatic passages, some of which may be inspired by Western
music, are increasingly heard in N orth Indian music. As yet they are used only for
special effects and have not altered the structure of the music, but there is a possibility
that continued experimentation in this direction may eventually lead to the intro
duction of chromatic scales.
Before we examine the validity of Bhatkhande’s system of ten thats in relation to
present-day N orth Indian classical music, we must discuss the inadequacies of this
kind of heptatonic that system as a basis for classifying rags.
A prime difficulty occurs in the classification of hexatonic and pentatonic rags
which could fit equally well in more than one heptatonic that. The pentatonic rag
Bhupall, for example, could be classified in any of four thdts: Kalyarjt (No. A l),
Ex. 19. rag Bhupall

Sa Re Ga Pa D ha Sa
1 K.P.M. IV, p. 821.

51
That
Bilaval (No. A2), Khamaj (No. A3) and the that with Ma«-(!Vif) and Nib (Vllb) (No.
Bl) which is not used in classical music at present. These thdts, differ from each other
only in the use of the alternatives of the M a and Ni, the very notes omitted in the rag
Bhupall. Bhatkhande classifies this rag in Kalyari that, implying that the omitted
notes are, in fact, the sharpened fourth and the natural seventh. This is, of course,
impossible to confirm on purely objective grounds. One reason he gives for this
classification is that Bhupali is also called Bhup Kalyap by some musicians.1 However,
there seems less justification for his classification of the pentatonic rag Malsri in
this same that, and a strong case can be made for its inclusion in Purvi that (No. A9)
as the rags Sri and Jetsri, which are connected at least in name, both belong to this
that. Musically too Malsri resembles the heptatonic Jetsri which is pentatonic in
ascent, omitting Reb (IIb) and Dhab (VIb), the very notes omitted in Malsri. In ascent
Jetsri and Malsri are identical in scale:
Ex. 20. rags M alsri and Jetsri (ascent)

Sa Ga Ma# Pa Mi Sa
It might be thought reasonable to consider that the resemblance both in name and in
melodic features is not entirely fortuitous.2 But this argument is not conclusive since
there are instances of rags which, though related in name, do not have the same
scale.
A second difficulty lies in determining the basic scale of a rag in which more than
one of the alternative notes are used. In the rag Des, for example, Nib is used in the
ascending line and Nib in the descending:3
Ex. 21. rag Des
This rag could be classified either in Bilaval that (No. A2) with the Nib as an acci
dental, or in Khamaj that (No. A3) with the Nib as an accidental. Bhatkhande chooses
the latter course, thereby giving importance to the descending line. This is a reason
able criterion of classification since cadential phrases, which are of considerable
importance in rags, are generally descending. However, he does not apply it con
sistently. Many of the rags which he ascribes to Kalyap that (No. A l), for instance the
1 H.S.P. I, p. 86.
2 In the chapter on Transilient Scales (p. 122) we suggest, however, that Malsri is the pentatonic
derivative of Marva that.
3 K.P.M. Ill, p. 251.
52
That
rags Kedar (Kedara) and Hamir, have MaN (IVn) in both ascending and descending
lines, while the Mas (IVs) occurs only as an oblique ascending note. The classification
of Kedar in Kalyan that becomes quite incomprehensible as Bhatkhande gives the
Ma4 as the vadi in this rag. It seems quite unreasonable to consider the most impor
tant and the ‘most often used note’ of a rag as an accidental, i.e. as not being part of
the basic scale of the rag.
There are genuine difficulties in the classification of certain rags in which several
alternative notes may be used. The rcig Pilu, for example, in which all the alternatives
are permitted, is classified by Bhatkhande in K dfi that (No. A4), while Grosset gave
its scale as the that with Gab and Dhab (No. C5).1
From the foregoing discussion it will be apparent that there is an element of
subjectivity in the classification of rags according to scale. The rags Kedar and Hamir
referred to above might easily be classified in Bilaval that (No. A2) with the Ma# as
an accidental. In this particular instance, as in many others, the choice lies between
two scales, both of which are among the ten thdts of Bhatkhande’s system. There
seems little doubt that his ten thdts are very prominent in N orth Indian classical
music today. In each of them there are a number of mgs, varying from seven in Tori
(No. A8) to as many as forty-two in K dfi (No. A4) according to Bhatkhande’s system
o f classification. In contrast there is at most only one rag in each of the other scales
listed 0 1 1 p. 48.
Where there are many rags in one that, each of these is, of necessity, clearly delin
eated by its characteristic melodic patterns. Were this not so, rags would merge one
into another and lose their identity. When there is only a single rag in a particular
that—such rags are described as ‘isolate’—the melodic possibilities are not so closely
defined since the rag can always be distinguished from others on the basis of scale.
Two musicians may evolve quite different melodic characteristics in these ‘isolate’
mgs. For instance, in the rag N at Bhairav (No. C2), Pandit Ravi Shankar tends to
use successive minor thirds (Dhab to Ni is really an augmented second) thus omitting
the Sa, Ga and Pa (I, III and V)2 (Ex. 22a), while Ustad Ali Akbar K han often omits
the Ga and Ni (III and Y l i y (Ex. 22b):
Ex. 22. rag Nat Bhairav

(a) (b)
Ma D hab Ni Re Ma Dhab Dhab Sa Re Ma Pa Dhab
This difference is found even though both musicians belong to the same tradition
1 J. Grosset in A. Lavignac, Encyclopedie de la Musique, Paris 1921, premiere partie, p. 326.

2 H.M.V. disc, 78 r.p.m. No. N 94756.
3 H.M.V. disc, 45 r.p.m. No. 7 EPE 1219.
53
(gharand) and have been taught by the same teacher, Ustad Allauddin Khan. Never
theless, both versions are recognisable as Nat Bhairav, principally because there is no
other rag with this scale. Here we may see the beginning of a process which could
eventually result in two differentiated melodic patterns crystallising from the one
scale. If both versions are propagated, they may some day be differentiated in name.
It is not unreasonable to suppose that it was a similar process which has led to the
differentiation of rags based on the same scale. From this we could say that a large
number of rags in any scale is an indication of the antiquity of that scale, since the
process of crystallisation of rags and their widespread acceptance must be reckoned
in terms of generations. The very large number of rags in K dfi that is a case in point,
for this scale is the modern counterpart of the ancient Indian Sadjagrdma. While
time is an essential factor, the rate of this process of crystallisation is also influenced
by the popularity of the scale, and this factor is particularly significant today in
view of the ease of communication. Bhairav that (No. C9) did not belong to the ancient
Indian modal system, and yet Bhatkhande gives as many as eighteen rags in this
that.
On the other hand, the ‘isolate’ rags clearly suggest their relatively recent introduc
tion into N orth Indian music, although there is the slight possibility that an old rag,
which has not been generally accepted, is preserved relatively unchanged in a family
of musicians. If, in the course of time, these rags are accepted and performed fre
quently, we may expect that individual melodic features will crystallise into several
separate rags and that they will no longer be ‘isolate’.
Bhatkhande’s ten that system does not allow for these ‘isolate’ rags, most of which
may be thought of as modern experiments. Experimentation has certainly been going
on in India for several hundred years and some of the rags which are mentioned in
texts over this period are no longer heard in N orth India. Bhatkhande’s ten thats give
an indication of the principal scales used in present-day N orth Indian classical music
—scales which have withstood the test of time. It is not unreasonable to presume
that these ten scales have greater musical justification within the Indian system than
do the others of the thirty-two.
Bhatkhande has organised the ten thats in a classificatory order, to some extent in
imitation of the South Indian melakarta system.1 The first three thats, Kalyan, Bilaval
and Khamaj, have Res and Gas (II s and Ills); the next three, Bhairav, Purvi and
Mdrvd, have Reb and Gas (IIb and Ills); the next two, K dfi and Asavri, have Res
and Gab (IIs and M b); while the last two, Bhairvi and Tori, have Reb and Gab (lib
and Mb). Although this kind of arrangement is convenient, it tends to obscure the
relationship between the thdts. For example, Khamaj that (No. A3) is followed by the
very dissimilar Bhairav that (No. C9).
Six of the ten thdts, Kalyan, Bilaval, Khamaj, Kdfi, Asavri and Bhairvi, are musically
related as they are serial progressions of each other; in other words, beginning with
1 The South Indian melakarta system is described in the introduction to the Mela-raga-malika
of Maha-vaidya-natha Sivan, Adyar 1937.
54
That
any one of these scales, the others can be derived by treating each successive degree as
the ground-note for the new scale. In Western musicology these progressions, known
as modes, are sometimes referred to by the names of their ground-note, taking the
‘natural’ scale beginning on the note C as the standard.1 Thus the D mode refers to
that succession of intervals obtained by playing, for instance, the white notes of a
piano from D to d. These serial modes with their modern Indian counterparts are
shown below.
‘C ’m ode Sa Re Ga Ma Pa D ha Ni Sa
Bilaval
that
‘D ’ m ode ' Sa Re Gat Ma Pa Dha

K afI that
J Sa R et G at Ma Pa D h a 1, N it Sa
E mode -Q ■ — ,
Bhairvi that , «, « '<* -----*7...
Sa Re Ga Matt Pa Dha Ni Sa
F m ode
K alyan that
Sa Re Ga Ma Pa D ha N it Sa
‘G ’ m ode
Khamaj that
Sa Re G at Ma Pa D hat N it Sa
‘A* mode
Asavri that
Sa R et G at Ma Pat. D hat N it Sa
A similar system of nomenclature was in use in ancient Indian musical theory where
the seven suddh jatis (pure or natural modes) were named after the notes of the scale:
Sadji after Sadja (Sa), Arsabhi after Risabha (Ri), Gandhdri after Gandhara (Ga),
Madhyama after Madhyama (Ma), Pahcami after Pahcama (Pa), Dhaivati after
1 In this context the Western notes have no implications of precise pitch.
55
That
Dhaivata (Dha) and Naisadi after Nisada (Ni). These seven serial modes1 were com
parable to six of the modern Indian thats as well as to the seven ancient Greek diatonic
modes and the mediaeval Ecclesiastical modes. These are ail shown, with C as the
ground-note, in the table below.
A ncient Ecclesiastical A ncient M odern

G round- Scale transposed Greek nam e nam e Indian nam e Indian
note to C (Ptolem y) (Glarean) (Natyasastra) nam e
Si Rs Gi Mi Fa. Dha Ni Sa
C
Sa Lydian Ionian Naisadi Bilaval
Ss Ro ■G i Ml Pa Dha N i. Sa
D
Re Phrygian D orian SadjI K afI
Sa ReV C i M s Pa D hi N it Sa
E
Ga D orian Phrygian Arsabhl Bhairvx
Si Re Ga MaJ Fa Dha NT Sa
F
Ma H ypolydian Lydian G andharl K alyan
Sa Re Ga Ma Fa Dha N it Sa
G
Pa H ypophrygian M ixolydian M adhyam a Khamaj
Si Re GaV M* Pa D hai Nfc Sa

A
Dha f~F—m— ^ = “^ 1 H ypodorian A eolian PancamI Asavrl
B
Ni M ixolydian Locrian D haivati
The similarity between the modern thats and six of the ancient Indian jatis clearly
suggests an unbroken continuity of tradition. However, the serial relationship of
these six thats is not commonly appreciated in India at the present time.
The relationship between these six thdts can also be expressed in a different way.
In each there is only one imperfect interval, a tritone, which as its name implies,
consists of three wholetones making an augmented fourth, with its complementary
interval, a diminished fifth. Elsewhere the fourths and fifths are perfect. The position
of the tritone differs in each of these serial thats. If we correct the tritone in any of
them, lowering the upper note by a semitone (or raising the bottom note by a semi
tone), this shifts the tritone a fourth higher (or lower) and produces another of these
1 Strictly speaking, these were not serial modes as they were derived from two parent scales,
Sadjagrcima and Madhyamagrama, which differed microtonally in their tuning of one note, the Pa.
56
That
thats. When this process is continued it results in the following succession (the
tritones are indicated by square brackets):
augmented 4th
‘F m ode
1 1
K alyan that
Sa Re Ga Pa Dha Ni Sa
/ augmented 4th
6C ’ m ode
Bilaval that j
Sa Re Ga Ma Pa Dha Ni Sa
augmented 4th augmented 4th

‘G ’ m ode ■ V ~
Khamaj that
Sa Re G a Ma Pa Dha Ni^ Sa
augmented 4th
*D’ m ode
K afI that
Sa Re Ga1, Ma Pa Dha N i1" Sa
augmented 4th Iaugmented 4th

‘A ’ m ode
Asavrf that
Sa Re Gah Ma Pa Dhafc Ni^ Sa
augmented 4th
‘E ’ m ode
-W-
BhairvI that r|j# k®-
Sa Re^’ Ga'-' Ma Pa Dhab N i^ Sa
If this series were continued, the next scale would be the seventh mode which, as
we have stated earlier, has a diminished fifth (Pah) in direct relationship with the
tonic (Sa). As this scale would violate a fundamental premise of modern N orth Indian
musical theory, the series cannot be continued any further in this direction. If we
retrace our steps, raising the bottom note by a semitone, the next step after Kalyan
would be a scale in which the tonic would have to be raised (Sa#). This too is prohi
bited in the system. Thus, strictly speaking, this series of modes, as applied to N orth
Indian music, should end at BhairvI and Kalyan.
Bearing in mind, however, that the enharmonic difference between the augmented
fourth, Ma#, and the diminished fifth, Pab, does not appear to be recognised, we could
57
That
* •
continue the series beyond Bhairvi, replacing the Pab by the enharmonic M at.
Similarly, we could continue the series beyond Kalyari, replacing the Sa# by its en
harmonic Reb. In Chapter V this enharmonic change is discussed further and we have
attempted to show how this could have occurred in musical practice.
As a result of this enharmonic change the tritone irregularity of the six serial
thats now becomes a diminished fourth (with its complement the augmented fifth).
This can now be corrected by raising the upper note a semitone (or lowering the
bottom note an equal amount). It will be seen that the two extensions beyond
Bhairvi and Kalyan meet:
augmented 4th
t------------------------------------- !
Bhairvi that
-----
Sa Reb Gab Ma -Pa .Dhab jsiib jja
diminished 4th
\rr---------------- 1
Hypothetical
b*
Sa Reb G at M a# Pa Dhab Nil^ Sa
diminished 4th diminished 4th
T ori that
\
¥
Sa Reb G at M a$Pa Dhab Ni Sa
diminished 4th
PurvI that £
S (;o---- ¥
Sa Reb Ga M a#Pa Dhab Ni Sa
diminished 4th diminished 4th

1
M arva that
''-Srh*---- • ■ -ff*
Sa Reb Ga Ma# Pa D ha Ni Sa
perfect 4th
K alyan that
¥
augmented 4th /
Sa Re Ga Ma# Pa D ha N i Sa
We can therefore present these ten thats as a circle.1 Nine of the ten are, in fact,
1 This Circle of Thats is identical with the outer circle of the thirty-two thats shown in Appendix
A on p. 181, which was especially designed to show the most consonant scales.
58
That
Bhatkhande’s thats; the other, No. A7, is not used in N orth Indian music at the
present time. A tentative explanation for this will be offered in a later chapter.
KALYAN (No.A1?
Sa ReG aM a^Pa DhaNi Sa
MARVA (No. A10) BiLAVAL (N o.A 2)

Re SaR eG aM aP a DhaNi Sa
S a RebGa Ma# Pa DhaNi Sa
Ma
PURVt (NQ.A9) /
Sa RebGa Ma#Pa DhabN iS a / r ,b KHAMAJ (No.A 3)
'^^/bhat'> r»-s. k a* D n n k f i M l^
GqV 7 Ga1
TORT (N o . A S)
KAFl (NO.A4) ,
S a RebGabMa*Pa DhafcN iSa
Sa Re Gab Ma Pa Dha NibSa
(N o .A 7 )
S a RebGabMa# Pa D h a W s a ASAVRT (N o.A 5)
Sa Re Gab Ma Pa DhabNibSa
BHA1RVT (N o.A 6)
Sa RebGa1’ Ma Pa DhabNib Sa
Circle of Thats
In this diagram we have indicated the note that changes from one scale to another.
The succession of changing notes in a clockwise direction form a circle of fourths (or
in an anticlockwise direction a circle of fifths) similar to that in the Western Circle of
Keys. The true succession of fourths is a spiral in which the twelfth successive fourth
is a Pythagorean comma (24 cents) flatter than the fifth octave of the initial note,
while the twelfth successive fifth is a comma sharper than its seventh octave.
In the Western Circle of Keys this spiral is broken at one of three possible points
and completed as a circle by an enharmonic compromise. This theoretically implies
an adjustment of a comma, but in the Western tempered twelve semitone system does
not involve any actual change. In the Indian Circle the enharmonic change takes
place at a particular point to avoid the Pab (Vb), or, in the reverse direction, to avoid
the Sa# (Iif). In the Indian system, however, as the Sa and Pa never change, there are
only ten changes, and there is a discontinuity in the Circle. This appears in the
Circle of Thats in which the sequence of fourths (or fifths) is broken, Rett (IN) being
59
That
followed by Mat! (IVn). The reason for this is simply that Pa (V) and Sa (I), being
immovable notes, cannot be included in the circle of changing notes. If they were
inserted in the Circle between Re (II) and M a (IV)—Re Pa Sa M a (II, V, I, IV)—this
discontinuity would be eliminated.
For ease of comparison, we give below the true spiral of fourths, the Western
Circle of Keys, and the Indian circle of fourths as they occur in the Circle of Thats,
in both Western and Indian notation.
Western notation
True spiral o f fourths: c
C F Bb Eb Ab Db Gb Cb Fb Bbb Eb'b Abb Dbb
Circle o f K ey s: c F Bb Eb Ab Db Gb Cb
C# F# B E A D G
Circle o f T h ats: (C) F Bb Eb Ab Db
F# B E A D (G)
Indian notation
Sa M a Nib Gab Dhab Reb Pab Sab Mab Nibb Gabb Dhabb
Circle o f K e y s: Sa M a Nib Gab Dhab Reb Pab Sab
Sa# Ma# N i G a D h a Re Pa Sa
Circle o f T h ats: (Sa) M a Nib Gab Dhab Reb
Ma# N i G a D h a R e (Pa)
The fact that Marva, Purvi and Tori thats (Nos. A10, A9 and A8) are prominent in
N orth Indian classical music appears to be clear proof that this circle is not mere
theory but has substantial basis in practice. From the theoretical standpoint this
Circle of Thats has several satisfying features. Above all, it shows that nine of the ten
thats in common use can be connected in a logical scheme consistent with the theory
and practice of N orth Indian Classical music. In the Circle, opposite poles always
represent scales in which the opposite alternative notes are used, i.e. opposite Bildval
(No. A2), the natural scale, lies scale No. A7 in which all the altered notes are used;
opposite K d fi(No. A4), in which Gab and Nib are used, lies Purvi (No. A9) in which
the other three altered notes, Reb Ma# and Dhab are used, while the G a and Ni are
natural. At the north pole of the diagram lies Kalydy (No. A l) in which each of the
alternative notes is in its higher position, while at the south pole lies Bhairvi (No. A6)
in which they are all in their lower positions. It is thus easy to see which of these thats
are related and the nature of their relationships.
The cyclic nature of the N orth Indian scales has been recognised by Fox Strang-
ways.1 This appears to have arisen from his intuition rather than from any logical
process, a fact which says much for his understanding of N orth Indian classical rags.
His circle differs from ours in two ways: it includes Bhairav that and it does not
include scale No. A7. Having established the succession of the six principal serial
thats (Nos. A l-6 ) just as we have done, he acknowledges that the B mode is not
applicable in present-day music because it requires a flattened fifth (Pab). He then
1 Fox Strangways, op. c i t pp. 169-70.
60
That
proceeds to Bhciirav (No. C9) through the intermediate scale (No. B6) which he calls
Bilaskhani Todi. From this he proceeds to Todi (No. A8) through Gauri (No. A9).
From Todi he returns to this same scale, calling it Basant, then going on to Marva
(No. A10) and back to the serial modes through Kalyan (No. A l) as follows:
Fox Strangway s’s mgs Bhatkhande’s thats

Sa Reb Gat Ma Pa Dhat N it Sa
Bhairvi Bhairvi (No. A6)
Sa Ret Gat Ma Pa Dhat Ni Sa

Bilaskhani Todi — (No. B6)
Sa Ret Ga Ma Pa Dhat Ni Sa
Bhairav Bhairav (No. C9)
Sa' Ret Ga Mail Pa Dhab Ni Sa
Gauri Purvi (No. A9)
Sa R et Gat Maif Pa Dhat Ni Sa

Todi Tori (No. A8)
Sa Ret Ga Map Pa Dhat Ni Sa

Basant Purvi (No. A9)
Sa Ret Maff Pa Dha

Puriya Kalian Marva (No. A10)
Sa Re Pa Dha Ni
Kalian Kalyan (No. Al)
It will be noticed that Fox Strangways has been obliged to repeat Purvi that in
order to complete his cycle. In the following chapter we shall be discussing Bhairav
that and its relationship with the circle of thats.
The only cyclic concept in Bhatkhande’s works is connected with his theory that
the time of performance of a rag is related to its musical characteristics, particularly
to its scale and its vddi. In this theory the cyclic concept is implicit as the succession
of rags is resumed afresh at the beginning of each new day and it is in this connection
61
That
that Bhatkhande refers to a circle of rags.1 According to his time theory the rags are
divided into three groups:2
1. Those which have Re*? (IIm), Dhab (VIb) and Gab (111*0; ie ., rags belonging to
Kalyan, Bilaval and Khamaj thats.
2. Those which have Reb (lib), Gab (Ills) and Nib (VII b); i.e., those belonging to
Bhairav, Purvi and Marva thats.
3. Those which have Gab (Illb) and Nib (Vllb); i.e. those belonging to K dfi,
Asdvri and Bhairvi thati.
Strictly speaking this scheme does not include Tori that which has Gab (Illb) and
Nib (VII b), but Bhatkhande includes it in the third group on the grounds that some
rags of the Tori family have Gab (Illb) and Nib (VIIb).3
For the purpose of his time theory Bhatkhande divides the day into two twelve-
hour periods, from midnight to midday and from midday to midnight. Each of these
is further divided into three sections: the period from four to seven which is called
sanclhiprakas and is the period of sunrise and sunset; the period immediately preceding,
from approximately ten to four; and the period immediately following, from seven
to approximately ten.4 He then associates the rag groups with the three sections o f
the twelve-hour periods. The rags of group 1, i.e. those with Reb, Dhab and Gab, are
ascribed to the period after sunrise and sunset, from seven to ten either morning or
evening. The rags of group 2, i.e. those with Reb, Gab and Nib, are the sandhiprakas
rags to be performed at either sunrise or sunset. The rags of group 3, i.e. those with
Gab and Nib but including Tori, are ascribed to the period before sunrise and sunset,
from approximately ten to four. The feature which finally determines whether a
particular rag is to be performed in the morning (from midnight to midday) or the
evening is the vddi of that rag, as we have mentioned earlier. If the vddi is in the lower
tetrachord (purvahg) it will be an evening rag, if in the upper (uttrahg) it will be a
morning rag. If, however, the vddi is Sa (I), M a (IV) or Pa (V) the rag may be
performed in either period.5
The cyclic succession of thats therefore repeats after twelve hours, while the cyclic
succession of rags only repeats after twenty-four hours or two that cycles. Bhatkhande
refers only to a circle of rags but does not carry the m atter any further. If, however,
we arrange the various thats in the circle implied by Bhatkhande according to the time
1 H.S.P. IV, p. 22.
2 Ibid., p. 7.
3 Ibid., Preface, p. 7.
4 Bhatkhande avoids mentioning the precise duration of these periods. The figures given above are
taken from H. A. Popley, The Music of India, Calcutta 1950, p. 63. O. C. Gangoly, op. cit., p. 90,
quoting a paper read by Bhatkhande at the Fourth All-India Music Conference at Lucknow (1925),
gives the period after sunrise and sunset as extending from seven to twelve and the period before,
from twelve to four. It seems reasonable to presume that Bhatkhande intended a certain latitude in
the precise duration of these periods.
5 A further modification to the time theory is provided by the Ma# which, according to
Bhatkhande, is also indicative of the time of performance and occurs generally in night rags
CK.P.M. V, p. 31). However, there are some day-time rags such as Multani and Tori which have Ma#.
62
That
of performance of the groups, the resulting circle is similar to that suggested by Fox
Strangways and differs from our Circle of Thats in that it omits scale No. A7 but
includes Bhairav that.
KALYAN
MARVA BILAVAL
7 -1 0
PURVI KHAMAJ
4 -7
BHAIRAV
KAFI
1 0 -4
TORI ASAVRT
BHAIRVI
Circle of Thais after Bhatkhande’s time theory
In this circle Tori and Purvi are separated by Bhairav that. Yet Bhatkhande clearly
mentions the relationship between Tori and Purvi thats in his discussion of the rag
Multani, a rag of Tori that>when he says that this rag is ‘parmelpravesalc* (introducing
a new group) and is followed by rags of Purvi that.1 This connection would by-pass
Bhairav and consequently bring it closer to our Circle of Thats. From the diagram
on the following page it will be seen that the fundamentals of Bhatkhande’s time
theory can be coherently related to our Circle of Thats.
The times specified for the performance of rags are only approximate and in
practice there is considerable latitude. It does appear, however, that this theory
conforms in a large number of instances to the traditional time of performance of
rags.2
The relationship between the scale of a rag and its time of performance has been
expressed as follows: ‘The principle that really emerges from Bhatkhande’s theory
is nothing but the tendency of rdgas to follow the line of least resistance in the easy
transition from scale to scale and it is observed to a certain extent by all musicians.’3
The concept of gradual change is apparent in many aspects of Indian music: in the
extension of the range of a rag in the dlap (prelude); in the acceleration of tempo
1 H.S.P. IV, p. 714.
2 There are a few ra^s whose traditional time of performance does not correspond to that which is
obtained in relation to the Circle of Thais. The rag Top, for instance, is said to be a morning rag, but
in the circle the time of its performance is approximately three a.m. or p.m. Multani, another rag
of Top fhcit, is in fact performed at about this time in the afternoon. In a subsequent chapter we shall
suggest an explanation for Top being one of the exceptions to the time theory.
3 H. L. Roy, Problems o f Hindustani Music, p. 82.
63
That
• *
7 ° /c
8 °/c
KALYAN
5 ° /c BILAVAL
MARVA
9 °/c
KHAMAJ
BHAIRAV
TORI KAfT
No.7 ASAVRI
BHAIRVI
1 1 ° /c
1°/C
Time theory and the Circle of Thats
during a performance; and in the development of melodic and rhythmic ideas

throughout a performance.
The significant feature of the Circle of Thats is that it shows an easy transition
from scale to scale, and it is not surprising that the rags are generally performed in this
sequence during the course of each day. There is thus some reason to suppose that
the scales may have evolved in this same sequence during the course of the centuries
and to say that the daily succession of rags is, in some respects, a reconstruction of
the course of evolution.
In the next chapter we propose to consider the thats from the point of view of
musical practice rather than as a m atter of musical theory and to attempt to justify
the evolution of the Circle of Thats from this standpoint.
64
IV
The Effect o f Drones
A prominent feature of Indian music is the use of a drone, which sounds at least
the ground-note, Sa, throughout the whole performance. The ground-note is the point
of reference for measuring the intervals used in any rag. It can be said of any modal
system, whether or not a drone is used, that the primary significance of the various
intervals is their relationship to the ground-note, so that the notes are perceived not
in terms of absolute pitch but in terms of this relationship. The particular relationship
of any note to the ground-note is responsible for the dynamic quality or function of
that note. This quality has been described as the ‘particular kind of unfulfilment
peculiar to each tone, its desire for completion’.1 Only the ground-note is at rest and
needs no completion. All other intervals manifest instability, each to its own parti
cular degree, and require fulfilment which can only be achieved by a return to the
ground-note. The degree of instability and the corresponding tension does not
increase in proportion to the distance from the ground-note but is governed to a
large extent by the smoothness or roughness (consonance or dissonance) experienced
in the relationship of that note with the ground-note. The degree of dissonance of
the successive intervals has been calculated by Helmholtz on the basis of two violin
tones, one static and the other varying in pitch, as shown in the first diagram on p. 66.2
In this diagram the vertical axis represents consonance and dissonance: the deeper
the valley, the more consonant the n ote; the higher the peak, the more dissonant the
note. W ith C as the ground-note (Sa) the F and the G (Ma and Pa) are clearly the
most consonant, the A and E (Dha and Ga) next in consonance. The Db, B, Ab and
F# (Reb, Ni, Dhab and Ma#) are the most dissonant in the series.
This scheme cannot be applied directly to Indian music primarily because a secon
dary drone is generally used. This is usually the fifth (Pa) but it is sometimes the
fourth (Ma), depending largely on the relative importance of these notes in a parti
cular rag. This secondary drone too has its own consonance-dissonance series which
must be superimposed on the original series to give a more realistic picture for Indian
1 V. Zuckerkandl, Sound and Symbol, Music in the External World, Bollingen Series, New York
1956, p. 94.
2 Helmholtz, Sensations o f Tone, trans. A. J. Ellis, Longmans, Green and Co., London 1875,
p. 520.
5 65
PURE INTERVALS (i.e. based on sim p le r a tio s )
TEMPERED INTERVALS
Consonance-Dissonance with C Drone (after Helmholtz)
music. In the diagrams following we have added the effect of the secondary drone,
the fifth (Pa) in the first and the fourth (Ma) in the second, working on the assumption
that the principal drone (Sa) is twice as prominent as the secondary drone. The dotted
lines in both diagrams show Helmholtz’s original graph.
Sa Reb Re Gab G a Ma M a* pa Dhab Dha Nib Ni Sa INDIAN

NOTATION
C Db D Eb E F G Ab A Bb B C PURE INTERVALS
TEMPERED INTERVALS
Consonance-Dissonance with Sa Drone twice as prominent as Pa Drone

66
It will be apparent that the addition of the secondary drone changes the pattern of
consonance and dissonance. In both diagrams the ground-note (Sa) is no longer a
perfect resolution. When Pa is introduced as a secondary drone the most prominent
changes are that the Ga is now nearly as consonant as the Ma, while the Dha is much
less consonant than it was with merely one drone, and that both the Reb and Reb are
much less dissonant, while the M ai is more dissonant. When Ma is introduced as a
secondary drone, both Ga and Pa become less consonant and the Dha, Nib and Nib
are now slightly more consonant, as are the Reb and Reb.
Sa Reb Re Gab Ga Ma Ma* Pa Dhab Dha Nib Ni Sa INDIAN

NOTATION
C Db D Eb E F G Ab A Bb B C PURE INTERVALS
TEMPERED INTERVALS
Consonance-Dissonance with Sa Drone twice as prominent as Ma Drone
These diagrams are at best a vague approximation and indicate only the directions
in which the consonance-dissonance pattern is altered by the introduction of a
secondary drone. There are a number of factors which might affect the real picture.
For instance, our calculations are based on Helmholtz’s work with two violin tones.
Obviously there is a considerable difference in the tone quality of the violin and the
usual Indian drone instrument, the tambura; the latter accentuates different over
tones which would alter the consonance-dissonance relationships of the various
tones. If the second overtone (sounding Pa) is stronger on the tambura than on the
violin, we may expect the Ma and the Dha to sound slightly rougher against the
tambura drone. The tambura tone itself is not constant but varies from instrument to
instrument, and one must also take into account the particular acoustic quality of the
hall or room in which the music is performed. Musicians sometimes add other drone
67
notes1 which would further alter the graphs. These variables make it impossible to
predict an entirely accurate picture of the consonance-dissonance relationships of the
tones in any actual performance.
The precise dynamic function of a note varies not only in different performances
but, to a lesser extent, even during one performance, as for instance when an instru
ment goes very slightly out of tune. It will be noticed that most of the 'pure’ intervals
are located in the bottoms of the troughs or valleys, while the tones of the tempered
system are frequently located on the rise or fall and are slightly more dissonant than
their ‘pure’ interval counterparts. This does not necessarily mean that the tempered
scale is musically less satisfactory than the ‘pure’ scale. A musical scale can aptly be
described as a scale of dynamic tensions, a gradation of degrees from the most con
sonant to the most dissonant. There is considerable evidence to indicate that in
Western music as well as in Indian music the tensions inherent in either the ‘pure’
scale or the tempered scale are, by themselves, insufficient for the needs of musical
expression. It is well known, for instance, that instruments without fixed intonation
as well as the voice tend to sharpen the leading note (Nib) even beyond the tempered
interval, thereby increasing its already considerable dynamic function and, by
contrast, enhancing the effect of resolution on the tonic (Sa). The development of
Western music in the past few centuries has been characterised by the increasing use
of discords, and, particularly in this century, by experiments with new scales, such
as the division of the octave into six equal parts (Debussy’s whole-tone scale) and
into twenty-four equal parts (Haba’s quarter-tone scale). These have served to intro
duce a greater variety of dynamic functions and to extend the dynamic range of
music.
From this evolutionary standpoint we may conclude that the tempered scale was
an improvement on the ‘pure’ scale since it permitted a greater dynamic range.
This is particularly evident in the tempered thirds where the slightly increased dis
sonance gives a vibrant quality to the common m ajor and minor chords. The fact
that the intonation in performance does not always follow the tempered intervals
indicates that these are by no means perfectly satisfactory. It would seem that the
ideal scale is one which permits a certain measure of latitude, so that intonation may
vary slightly, depending on the context in which the notes occur and the interpretation
of the musician. This, of course, is not generally possible in concerted music, but in
Indian music, since it is invariably performed by soloists, flexibility in intonation
is quite usual.2
In discussing the dynamic function of notes we have so far referred only to the
primary level on which this operates, where each note seeks resolution on the ground-
note. On a secondary level dissonant notes seek their more consonant neighbours.
Here the terms dissonance and consonance m ust be taken to mean the roughness and
smoothness of the notes as modified by the occurrence of a secondary drone. These
1 See Appendix B, p. 187.
2 N. A, Jairazbhoy and A. W. Stone, op. c i t p. 130.
68
two dynamic functions can be shown in schematic form, the solid lines indicating
the primary function and the dotted lines the secondary function:
Sa Re Pa Dha Sa
Sa + Pa Drones
Sa Re Ga Ma Pa Dha Sa
Sa + Ma Drones
In recent times the third (Ga) is often added as a supplementary drone to the
Sa + Pa drones1 resulting in a further modification of the consonance-dissonance
pattern of the notes. Here the dynamic function of the notes is similar to that in
Western classical harm ony where the major triad (Sa, Ga, Pa) is implied and may,
for short periods, even function as a drone. The following diagram shows the con
sonance-dissonance pattern with these three drones, the Sa drone being twice as
prominent as the others. The dotted lines once again show Helmholtz’s original
trace:
Sa Reb Re Gab Ga Ma Pa Dhab Dha .Nib Ni Sa INDIAN
NOTATION
C Db D Eb E F G Ab A Bb B C PURE
INTERVALS
TEMPERED INTERVALS
Consonance-Dissonance with three Drones, Sa, Pa and G a; Sa twice as prominent as Pa and Ga
1 The principal advocate of this drone accompaniment is Ustdd Vilayat Khan who also uses other
drone combinations to suit particular ra^s. Some of these are described in Appendix Bon pp. 187,188
and can be heard on the accompanying record.
69
The principal modification resulting from the introduction of theG a drone is that the
Ga is much more consonant than the M a which now has a greatly increased secondary
dynamic function in that it leads either to the G a or the Pa.
We have been discussing the inherent dynamic function of notes which is not
derived from their pitch but from their relationship to the ground-note and to the
other notes sounded in the drone. There are, however, other factors which also have
a very great influence on the function of a note. These arise from the context in which
the note is heard and are connected with melody, rhythm and metre. While a full
treatment of these factors is beyond the scope of this work, a brief examination of the
dynamic function induced by the melodic context is contained in Chapter VIII, pp. 17 Iff).
A fundamental question arises out of the foregoing discussion. How can we
recognise and appreciate a rag when the dynamic function of its notes is variable ?
The only explanation which appears to fit this condition is that the mind has con
siderable latitude in the comprehension of musical intervals. This is borne out by the
fact that in Indian music the precise intonation of notes also varies from performer to
performer, from recital to recital and even within the same recital,1 and yet the rag
being performed is clearly recognised by the audience. Perhaps the best way to
understand this is in terms of an analogy. Let us imagine that the consonance-
dissonance graphs represent the terrain on which we are walking. As we walk down
from a peak into a valley, at a certain point we suddenly recognise the valley and can
say this is D ha or this is Ga. The point of lowest potential energy of this valley is at
its bottom, but recognition dawns somewhere on the slopes. The analogy must now
be carried into three dimensions if we are to convey the dynamic function of the
notes, as the particular valley we are concerned with may be located in the mountains,
and a river in this valley will run into a lower valley and continue downwards until it
finally reaches the ocean. In two dimensions the bottom of the valley appears to be
a state of minimum potential energy; in three dimensions, however, it is seen that the
bottom of the valley is itself sloping towards a lower valley. The incline is less steep
in the valley than on the slopes, thus the kinetic energy, which can be correlated with
the dynamic function of the notes, is lower in the valley than on the slopes. Would
a musician necessarily choose the point of lowest kinetic energy when he wishes to
convey suspense, anticipation or tension ? It has been noted that the leading note
(Ni) is often sharper in ascent than in descent. Is not this sharpening of the Ni a
subconscious device to increase its dynamic value so that it more urgently demands
resolution on the tonic (Sa) ?
To summarise, music is concerned, from one viewpoint, with states of tension and
release, with contrasts of energy levels. Where the musician wishes to convey the
feeling'of relief from tension, he must seek the bottom of the valley, and particularly
those valleys which have a low potential energy level, in other words the more
consonant notes. When he aims to convey tension, however, he would not necessarily
1 Ibid., p. 130-1. There is reason to believe that the same occurs in Western music played on
non-keyboard instruments or sung, despite its basis of equal temperament.
70
seek the bottom of the valley of the less consonant notes. Yet he cannot stray too far
up the slope, else the note would sound disturbingly out of tune.
While the drone affects the dynamic function of notes, this is by no means its only
influence in Indian music. Generally the drone is taken so much for granted and is so
much part of the music that the exact nature of its influence is difficult to perceive.
Even on the infrequent informal occasions when music is sometimes performed with
out a drone, it is nevertheless implied, and it is likely that the memory of the drone
compensates, to some extent, for its physical absence. However, extraordinary
occurrences often enable one to have an insight into normal events. For instance, on
a recording by Ustad Bismillah K han1 playing on the shahna'i, one of the drone
shalmd’is suddenly introduces the third (Ga) as a subsidiary drone note. The result is
that the melodic improvisations gravitate to this point and one clearly hears the modal
series beginning on G a rather than the original scale (based on Sa). This is not a
transposition, merely a temporary shift of the point of reference and a corresponding
shift of tessitura. It is, of course, an unconventional practice, but for this very reason
we can clearly appreciate the result. After a short while the Ga drone ceases and the
melodic line returns to its original framework.
Another interesting occurrence can be heard on a record of Ustad Bundu K han
playing the sarangi2 Here he plays the pentatonic rag M dlkos (Mdlkaus) in which the
secondary drone is usually the fourth (Ma). On this occasion the primary Sa drone
is abandoned entirely and only the secondary M a drone is played; on analysis, the
proper scale of the rag (Ex. 23a) appears to be inverted so that the M a now becomes
its ground-note (Ex. 23b):
Ex. 23. rag Mdlkos

(a) (b)
Ma
drones ' drone
There is, nevertheless, no difficulty in recognising that it is rag Mdlkos which is

being performed. This can only be explained if we acknowledge that the series of
notes in Ex. 23b, the Ma-inversion provoked by the secondary drone (Ma), is always
implicit in this rag.
These two examples suggest the following conclusions: first, that the secondary
drone may become the temporary ground-note of the rag, particularly when it is
brought forward as it was in the first example by its sudden introduction; and
secondly, that the modal sequence starting on the secondary drone is also registered
in the mind of the listener, whether overtly realised or not, and is an essential aspect
of any rag.
1 H.M.V. N 94755 (78 r.p.m.), side entitled Kajn.
2 H.M.V. HT 83 (78 r.p.m.).
71
It may be argued, with some justification, that these conclusions need not apply
when Pa occurs as the secondary drone since there is a basic acoustic difference
between M a and Pa in this context. Let us go into this m atter in some detail.
Analysis of musical tones reveals that no tone produced on a musical instrument is
pure, but is composed of a fundamental and, in addition, a number of overtones
which are generally much softer than the fundamental. These overtones are explained
by the fact that a string on an instrument or a vibrating column of air in a pipe
vibrates not only in its full length but also in proportions of its length—half, third,
quarter, etc. In theory this overtone series1 is limitless, but, as each successive over
tone is softer (except where one of these is amplified by the shape of the resonating
chamber of a particular instrument), for all practical purposes only the first few over
tones are musically significant.
We should note that the perfect fourth, Ma, is not one of the notes which is signi
ficant in this series.
Ex. 24.
Sa Pa Ga
Fundamental
1st 2nd 3rd 4th 5th
Overtones
When the Sa drone is sounded, the overtone series is evoked and the first few
overtones are often clearly audible. The second overtone, Pa, is usually quite pro
minent, especially on the tambura, and is thus always present in Indian music as a
secondary drone whether or not it is actually sounded. To a lesser extent this also
holds for the fourth overtone, Ga. The addition of Ga, therefore, as an extra drone
note is an extension of a natural phenomenon and not a radical development to be
associated with Western influence. Ma, however, although consonant to Sa, is alien
to the overtone series and is not evoked in the sound of the Sa. On the other hand,
Sa is evoked in the sound of M a since Sa is a fifth above M a and is its second over
tone. For this reason it can be argued that the tendency to view M a as the ground-
note has a ‘natural’ basis. The same cannot be said for Pa as Sa is not part of its
overtone series. This thesis can be expressed in the following way: If two drones either
a fourth or a fifth apart are sounded, one of these will ‘naturally’ sound like the
primary drone. It is not always the lower of the two which will sound primary, but the
one which initiates the overtone series to which the other note (or one of its octaves)
belongs. By amplifying a prominent overtone the secondary drone lends support to
the primary and intensifies its ‘primary’ character.
1 A second system of nomenclature refers to these as the harmonic series, in which the funda
mental is the first. The first overtone, i.e. Sa, is the second harmonic; the second overtone, i.e. Pa,
is the third harmonic, etc. One advantage of this system is that it permits easy calculation of the
intervals between harmonics; for instance, the interval between the ninth harmonic (the eighth
overtone) and the eighth harmonic (the seventh overtone) is nine: eight.
72
Nevertheless, there are instances when Pa also becomes, in effect, the ground-note,
evoking its own modal series. Notable examples can be found in the rags Pancam se
Pilu (lit. Pilu from Pa) and Pancam se Gara (Gam from Pa) where the parent rags
Pilu and Gara are virtually transposed to the secondary drone, Pa.1
That the fifth should evoke its own modal series is not peculiar to Indian music, for
the Western Ecclesiastical inodes show, in their plagal forms, a similar tendency.
This is not exactly the same phenomenon as we have been speaking of in Indian
music, for in the plagal modes the finalis, which we might equate with the Sa, remained
the same as in its authentic mode, and only the ambitus, the range of the octave, was
shifted a fourth lower, extending from Pa to Pa.
It would therefore appear that rags have a certain measure of dual or even multiple
modality. When a secondary drone is brought to the fore, as, for instance, when it
serves as the pivot note in a series of melodic phrases, it serves temporarily as the
ground-note of the rag and evokes its own particular modal series, which may not,
however, be appreciated on a conscious level. In theory this applies to any terminal
note but is less significant unless the terminal note is also amplified in the drone.2
Naturally, the authentic series initiated by the Sa is predominant and the melodic line
is inevitably drawn back to this base.
If we accept this hypothesis, it is easy to see how the six primary thats, Kalya# to
Bhairvi, might have evolved without the conscious process of beginning each one
on the successive degrees of a primary scale, as was apparently the case in ancient
Indian music theory. From each that two modal series are brought to the fore by the
two commonly used secondary drones, Pa and Ma. Each of these series can become
a scale in its own right when the primary and secondary drones are interchanged,
as shown in the table on the following page.
By this process, too, we arrive at the B mode which is not one of the Indian thats,
but whose influence can be seen in the rag Bhairvi, where the diminished fifth is
sometimes used in descent (see Ex. 14b, p. 50).
It will be noticed that we cannot continue this process beyond the B mode because
Pa would be flat here and there could be no secondary Pa drone. In the same way,
retracing our steps, beginning each successive scale on the M a of the previous scale,
we finally arrive at Kalya#. Here again the process cannot be continued further
because Kalya# has a Mas, and thus a secondary Mati drone is not feasible.
We have been seeking musical justification for some of the theories expressed in the
previous chapter, and, in particular, a practical justification for the Circle of Thats.
We can now explain how the six primary thats might have arisen out of musical
practice, but we have also seen that we can go no further by this process and the
1 One of the characteristic features of the rags Pilu and Gara is the use of both forms of Ga (III)
and Ni (VII). In the ‘transposed* rags these are also a fourth lower (or a fifth higher) and appear
as both forms of Ni (VII) and Ma (IV). These rags can be heard on the two following records:
Pancam se Pilu, played by Vilayat Khan, H.M.V. 7 EPE 59; and Pancam se Gcira by Ravi
Shankar, H.M.V. N 94754 (78 r.p.m.).
2 For a further discussion of this, see Appendix B, pp. 188, 189.
73
remainder of the Circle of Thats is still unexplained. Seen in terms of consonant
fourths and fifths, the primary thats are nearly perfect, each having only one imperfect
relationship. The occurrence of the other thats in the Circle cannot be justified in
terms of consonance as each of these has several imperfect relationships; for instance,
Marva that has two augmented fourths, Sa-Ma# (I-IVs) and Reb-Pa (Ilb-V), and an
augmented fifth, Reb-Dha (llb-VI). We have also seen the primary thats as a
connected series beginning on the successive degrees of a parent heptatonic scale.
This is a complete, conscious system and excludes all other thats. Lastly, we have
seen these primary thats growing out of an imperceptible evolutionary process in
which the primary and secondary drones are inverted. Here too we have been unable
to go beyond the seven serial modes.
Sa Re Ga M at pa Dha Ni Sa
Sa Re Ga Ma Pa D h a 'N i Sa
I t
Sa Rc Ga Ma Pa DhaN ili Sa
I I
Sa Re G at M a Pa Dha N it Sa
Sa Re G at Ma Pa D h at N it Sa
Sa Ret G at M a P a D hat N it Sa
Sa R et G a t Ma P at D h at N it Sa
Let us look at our subject once again in a different light. In all our discussion of
scales we have taken the upper limit of these scales entirely for granted. The feeling
of identity which one experiences in the octave of the ground-note provides a ‘natural’
opportunity to terminate a scale. It has been said that the Octave has an unique status
in the series of notes, comparable to the role of the number One in the set of all
numbers.1 There can, of course, be no denying the exceptional status of the octave;
1 J. L. Dunk, The Structure of the Musical Scale, London 1940, p. 65.
74
we can, however, question whether the pre-eminence of the octave is a m atter of
degree or of kind. In other words, we can question whether the identity of the octave
with the ground-note is absolute or whether all notes can be identified to some extent
with the ground-note, while the octave has the highest degree of identity.
Perhaps we can gain an insight into this question by a consideration of what one
would actually hear if the ground-note and its octave were sounded loudly and
simultaneously. Under normal circumstances, the fundamental notes, Sa and Sa,
would be the most prominent. One would also hear the overtones of the fundamentals,
and, in addition to these, other series of tones produced by the interaction of the two
fundamentals and their overtones. These are known as summation and difference
tones.
The first summation tone is the sum of the frequencies of the two fundamentals,
while the first difference tone is the difference between their frequencies. The relative
loudness of the overtones and the summation and difference tones has been calculated
by Sir Janies Jeans.1 Based 0 1 1 his table, we give in order of diminishing intensity
various tones heard when Sa and Sa are sounded together:
1. Sa-Sa (I-I) The two fundamentals

2. Sa-Sa (I-i) The first overtones
Pa (V) The first summation tone
Sa (I) The first difference tone
3. Pa-Pa (V-V) The second overtones
If we extend this list we will have, in diminishing intensity, the whole of the over
tone series as well as the other summation and difference tones. For our purposes
the above is sufficient for we already see the very considerable prominence of Pa in
the sound we hear.
In comparison let us see what happens when Sa and Pa are sounded loudly and
simultaneously:
1. Sa-Pa (I-V) The two fundamentals

2. Sa-Fa (1-Y) The first overtones
Cla (III) The first summation tone
Sa (I) The first difference tone
3. i*a-Re (V-II) The second overtones
There is a marked similarity in the two lists. At the same time there are also some
differences. G a (III), which is the first summation tone of the Sa-Pa fundamentals,
is only the fourth overtone given by the Sa-Sa fundamentals, and consequently will
sound more prominent with the Sa-Pa fundamentals. The second overtone of Pa,
1 Science and Music, Cambridge 1937, p. 236.
75
i.e. Re (II), appears only as the eighth overtone of Sa, and will certainly be more pro
minent with Sa-Pa as fundamentals.1 A t the same time the Sa is further strengthened
by the first difference tone (Sa) of the Sa-Pa fundamentals, and this would tend to
enhance the feeling of identity of the Pa with the Sa. This evidence seems to suggest
that the difference between the relationship of the octave to the ground-note and
the fifth to the ground-note is a m atter of degree and not of kind. It is interesting to
note that against a Sa drone the octave Sa, when sounded, produces Pa as a summa
tion tone, while the Pa, when sounded against the same drone, produces Sa as a
difference tone.
In Indian music particularly, where drones are prominent, the Sa and the Pa tend
to acquire a certain ambivalence. The present writer has often noted the difficulty
students experience in differentiating between Sa and Pa even when there is a clearly
audible conventional drone in which the Sa fundamental is much more prominent
than the Pa.
In our discussion of scales we have taken no special account of the Pa, considering
it merely as one of the steps, albeit a consonant one, between the ground-note and
its octave. It is of importance to note also that we felt no need to go beyond the octave
on the presumption that the series of notes would repeat themselves from the octave
onwards.2 However, if the Pa can also be identified with the ground-note Sa, will
there not be also a tendency to consider the Pa as the end of a register and the begin
ning of the next one ? If this is so, do we not also expect the intervals to repeat them
selves beyond the Pa, just as we expect the intervals to repeat themselves beyond
the octave Sa?
This raises considerable difficulties, for the Pa does not divide the octave into two
musically equal parts, Sa-Pa being a fifth and Pa-Sa being a fourth. In spite of this,
there is a strong tendency to view the octave in two parallel parts. The half-way point
of the twelve semitones of the octave is Ma# (IV#), but the dissonance of this note to
the ground-note should preclude its use as the end and beginning of a register. On
either side of the Ma# are located the two most consonant notes (excluding the octave)
i.e. those notes which are most easily identified with the Sa, and it is with these notes
that the division of the octave is generally associated.
In ancient Greek musical theory the octave was divided into two tetrachords plus
a wholetone. The wholetone could appear between the two tetrachords, Sa-M a and
Pa-Sa (I-IV and V-I), in which case the tetrachords were said to be disjunct, or the
wholetone could appear at the end of the two tetrachords to complete the octave,
Sa-M a and Ma-Nib (I-IV and IV-VIIb), and the tetrachords were then said to be
conjunct. The octave can also be divided into two overlapping pentachords, or into
a pentachord and a tetrachord. These have all been tried at one period or another,
1 This may help to explain why the Re, which occurs as a terminal note in a number of ragst
does not convey quite the same element of suspense as does the second in Western classical music
based on harmony of the triad.
2 This is not always true in Indian music, notable exceptions being the rags Des and Sorafh of
Khamaj fha{ in which some musicians use Gab only in the upper register.
76
but no perfect division of the octave is possible so long as the consonance of M a and
Pa is recognised. This may explain why, in theoretical systems, the names of the notes
do not repeat beyond the M a or the Pa as they do beyond the octave.
Bhatkhande refers to the two parts of the octave as piirvang and uttrang, which he
sometimes defines as the disjunct tetrachords Sa-M a and Pa-Sa,1 and at other times
as two overlapping pentachords, Sa-Pa and Ma~Sa.2 In Bhatkhande’s theory the
only significance of this sort of division is to provide a basis for determining the time
of day at which a rag should be performed. We are suggesting, however, that each
successive note in one tetrachord has a certain measure of identity with its counter
part in the other (e.g. Sa with Pa, Re with Dha, etc.) and that this identity is of the
same kind, but of a lesser degree, as that which we experience between a note and
its octave.
As we have pointed out, there are several possible ways in which the octave can
be divided. W ith Pa as the secondary drone, it would be reasonable to think of the
octave as consisting of two disjunct tetrachords, Sa-M a and Pa-Sa3—a view which
seems particularly reasonable when we consider the scale as an ascending series,
with the Pa as the initial of the second tetrachord. In descent, however, this same
principle leads to a division of the octave into two conjunct tetrachords, Sa-Pa and
Pa-R e, with a wholetone, Re-Sa, appearing at the bottom of the scale. From this it
becomes immediately apparent that a scale may easily be perceived in the light of
more than one tetrachordal scheme, and that there is a certain measure of ambiguity
in the location of the wholetone disjunction.
If we now consider the same scale with a secondary M a drone, where M a is the
initial note of the second tetrachord, we have in ascent two conjunct tetrachords,
Sa-M a and M a-N fi, with the wholetone disjunction, Nit>-Sa, appearing at the top
o f the scale, whereas in descent we have two disjunct tetrachords, Sa-Pa and M a-Sa,
with the wholetone disjunction appearing between Pa and Ma. These four types are
shown in the following schema:
Ex. 25. Tetrachord Species
Pa Ma
Sa Sa Ma Pa Sa Sa Sa Pa Ma Sa
i _ r — ■ | -------- --------- —I—®---------------- :------------ - - ---------- ------------ 11
—c*-------------- ~i
-e -
drones droncs
(a ) Ascending disjunct (b ) r )escending disjunct
Pa Ma
Sa Pa Re Sa Sa Sa Ma N it Sa
A Sa
---- ---------- ----- [j ^ ---
——
1— | ® -C- "O- j _-------- 1
drones drones
(c) Descending conjunct (d) Ascending conjunct

1 H.S.P. I, p. 41.
2 K.P.M. V, p. 31.
3 If one permits the repetition of Pa, one could also consider the octave as a pentachord plus
tetrachord, Sa-Pa and Pa-Sa. We shall, however, discuss this matter in terms of tetrachords.
77
It will be noticed that the ascending and descending disjunct tetrachord species are
virtually the same, since the disjunction occurs between M a and Pa.
There is, however, a measure of ambiguity in conjunct and disjunct tetrachord
types; conjunct tetrachords become disjunct when extended above or below a single
octave register, and vice versa. In musical practice the ascending disjunct tetrachords
could be realised as ascending conjunct in the tessitura from Pa to Pa (or Pa to Pa)
as in Ex. 26a. Similarly, the other tetrachord types may be realised as in Ex. 26b, c,
and d, respectively:
Ex. 26.
(a) (b)
Pa Ma
Sa Pa Sa Ma Pa Sa Ma Sa Pa Ma
-o -
d rones
(c) (d)
Pa Ma
Sa Pa Re Sa, Pa Sa Ma Sa Ma
E
*
drones drones
We have expressed these tetrachord types in relation to the drones, but it must be
remembered that even when the secondary drone in a rag is Ma, the Pa does have a
considerable degree of consonance and may be an im portant note of that rag.
Similarly, when the secondary drone is Pa, the M a also has a considerable degree
of consonance and may also be important. Consequently, the tetrachord types are
not always mutually exclusive and there are rags in which different tetrachord
groupings may be emphasised as the rag is developed through its various stages.
Later in this chapter we will have occasion to discuss other tetrachordal divisions o f
the octave.
In analysing the six primary thats we notice that some have parallel ascending
disjunct tetrachords (i.e. the successive intervals in the two disjunct tetrachords are
identical), while others have parallel descending conjunct tetrachords. Obviously, no
scale may be parallel in both respects, for if it is parallel in one, the altered position
of the disjunct wholetone ensures that it will be unbalanced in the other.
Inseparable from the consideration of these tetrachord types is the concept of the
consonance of fourths and fifths. The successive intervals of two parallel conjunct
tetrachords will be a fourth apart, while the successive intervals o f two parallel
disjunct tetrachords will be a fifth apart. In the primary thats, as we have indicated
earlier, there is only one imperfect relationship—one pair of notes bearing the
78
relationship of augmented fourth/diminished fifth. It is this imperfect relationship
which destroys the parallelism in one of the tetrachord types. If within the octave
register the two notes stand as an augmented fourth, the conjunct tetrachord relation
ship will be unbalanced while the disjunct tetrachord relationship o f fifths will
remain perfect and balanced. If, however, the two notes stand as a diminished
fifth within the octave register, the disjunct relationship of fifths will be disturbed.
Viewed from the standpoint of musical practice, this discrepancy in the scale could
be noticed in two ways, neither of which need be on a conscious level. First, it could
be noticed through the absence of a consonant fourth or fifth. This would be particu
larly significant in a musical system where fourths or fifths might be sounded simul
taneously. A good example of this can be found in early Western Church music
where the practice of parallel Organum (two voices moving in perfect fourths or
fifths, note against note) drew attention to the tritone1 (augmented fourth) and led
to the introduction of accidentals. This is probably not so significant in Indian music
where only a single melody line is generally used, and jumps of fourths and fifths are
exceptional. Secondly, it could come to notice as the musician tries to repeat a
melodic phrase in the second tetrachord register. From our earlier discussion of the
identity associated with the successive notes of the tetrachords, it would seem that
the inability to repeat a phrase or even an interval in the second tetrachord would
be disturbing in the same way, but to a lesser degree, as the inability to repeat a
melodic figure or an interval in the next octave register. From the long range evolu
tionary point of view this disturbance, as we hope to show, provides the vital spur
for the evolution of new musical scales.
Let us now consider the application of these principles. In Bilaval that (No. A2)
the ascending disjunct tetrachords are parallel, while the descending conjunct
tetrachords are unbalanced:
Ex. 27. Bilaval that
Sa Re Ga Ma Pa Dha Ni Sa
Re Ga Ma / Pa
Ascending disjunct tetrachords Descending conjunct tetrachords
In this scale the lack of balance is created by the difference in the first descending
steps in the two conjunct tetrachords: the interval between Sa and N i is a semitone,
while the interval between Pa and Ma is a wholetone. This is a characteristic feature
1 In Western plainsong the tritone was forbidden and in early polyphonic music was referred to
as diabolus in musica (the devil in music).
79
of Bilaval that', nevertheless, the lack of balance here demands special treatment, and
it appears that certain melodic features are directly motivated by this irregularity in
the scale and are manifest in many rags of Bilaval that. Some of these melodic
features will be discussed in the chapters that follow.
In general it may be said that these melodic features tend to diminish the disturbing
effect of the imbalance, but the final solution is to replace one of the unbalanced
notes by a balanced one: in Bilaval that, to replace the Nil? by a Nib, or the Mab by a
Mas. The former leads to Khamaj that (No. A3), and the latter to Kalyan (No. A l).
Both of these have balanced descending conjunct tetrachords, but the balance in the
ascending disjunct tetrachords is now disturbed. In these scales too melodic features
tend to arise to compensate for the imbalance, but once again the final solution can
be achieved only through replacement of one of the unbalanced notes, thus leading to
new scales. Passing over Kalyari that for the time being, let us look more closely at
Khamaj that:
Ex. 28. Khamaj that
Sa Re Ga Ma
Pa Dha Ga Ma
In the ascending disjunct tetrachords of Khamaj, the thirds Sa-G a and Pa-Nib
are unbalanced. This can be corrected either by replacing the Nib by a Nib, or the Gab
by a Gab. The former returns to Bilaval that, while the latter leads to K afi that (No.
A4). Both these have balanced ascending disjunct tetrachords and the imbalance is
apparent in their descending conjunct tetrachords.
Ex. 29. Kafi that
I Sa Re Pa Dha Sa
D ha Nit* Re Ma
In K afi that Sa-D ha is a descending minor third, while Pa-Gab is a descending

major third. This can be corrected either by replacing the Gab by a Gab, or the Dhab
80
by a Dhab. The former returns to Khamaj that, while the latter introduces the new
scale Asavri that (No. A 5):
Ex. 30. Asavri that
Re Ma Dliab
In Asavri that the ascending disjunct tetrachords are unbalanced: Sa-R e is a

m ajor second, while Pa-Dhab is a minor second. This can be balanced either by
replacing the Dhab with a Dhab, or the Reb with a Reb. The former returns to Kafi
that, while the latter leads to Bhairvi that (No. A 6):
Ex. 31. Bhairvi that
Sa Reb Gab Ma
[& , -
fry^
^ Pa Dhab Nib Sa . Ma
a
In Bhairvi that the descending conjunct tetrachords are unbalanced; Sa-Pa is a

descending perfect fourth, while Pa-Reb is a descending augmented fourth. This
could be balanced either by replacing the Reb by a Reb, or the Pa by a Pab. The first
course leads us back to Asavri that. In the second instance we are once again con
fronted with the same difficulty we have faced so many times. The Pab is not permitted
in Indian music.
We have once again arrived at a dead end. It will have been noticed that the six
scales we have covered are once again the six primary thats. We now return to a
consideration of Kalyan that:
Ex. 32. Kalycin that
Sa Re Ga MaB Dha Ni Sa
Pa D ha Mart Pa
6 81
In Kalyan that the ascending disjunct tetrachords are unbalanced: Sa-Ma? is an
augmented fourth, while Pa-Sa is a perfect fourth. This could be corrected either
by replacing the Ma? by a Man, or by replacing the Sa by a Sas. The former leads back
to Bilaval that, while the latter is not permissible. Are we once again at a dead end ?
The musician, faced with this imbalance between Ma? and Sa in the ascending
disjunct tetrachords which may not be solved by replacing the Sa with a Sa?, will
naturally try to minimise its effect. One way this can be done is by the omission o f one
or both of the offending notes, and although the Sa may not be altered, there is no
restriction against the temporary omission of it in a particular melodic phrase.
Thus in the principal rag of Kalya$ that, Taman, Sa is often temporarily omitted in
both ascent and descent,1 and phrases in which the Sa is omitted (see Ex. 33a) are
much more characteristic than those including the Sa (Ex. 33b). Once the Sa is omitted,
however, the Pa has lost one of its supports, and there is an equal tendency to omit the
Pa and phrases parallel to those in which the Sa is omitted are frequently heard (Ex. 33c).
Ex. 33. rag Yaman
(a) (b) (C)
Ni Rc Ga Re Ni Re Sa Sa Rc Ga Re Sa Ma? D h a Ni D h a Ma? D ha Pa
4 . -------------------------------------------------------------------:----------------------- - r ^ r r r
. -m '^ _ ---------------- m ----------------------- ^ 0 — U
^ ^ ^ - 1 m m — • — ^ -------------------—
Phrases in which Sa and Pa are omitted are often greatly extended and may encompass
a full octave2 or m ore:
Ex. 34. rag Yaman

Ni Re Ga Ma? D ha N i Re Ni D ha Ma? G a Re Ni D h a, N> Re Sa
i
Let us consider for a moment the balances of the rag Yaman in the phrases where
Sa and Pa are om itted:
Ex. 35. rag Yaman, Kalyan that [Sa and Pa omitted]
> Ni (Sal Re Ga M a? (Pa) D ha Ni (Sa)
<»
-----0 —
(-«•)
M a? (Pa) Dha Ni (Sa) \ Re Ga Ma? (P a) <—
4 = ---------- 0----------= + ■ = = w = 4 1
it- <*> * 1 * 1
1 The Ma? is also temporarily omitted in rag Yaman and this omission is also characteristic of
several other rags in this that. Yaman is played on the accompanying record and is discussed further
in Appendix B, pp. 204,205.
2 In Ustad Vi iayat Khan’s rendering of Yaman on the accompanying record there is no extended
passage in which Sa and Pa are omitted. Such passages are quite normal in Yaman and are also
played by Vilayat Khan, for example on his commercial record, E.M.I. ASD 2425.
82
When the Sa and Pa are in the scale, the Re balances the Pa, and it is the Sa which
is out of balance with the Mas. When the Pa is removed from the scale, it could be
either Sa or Re which can be said to be out of balance with the Mas, but when the
Sa is also omitted there is no alternative but to see the Re in relationship with the Mas.
Now there is only one possibility of creating balance, for the Ma# cannot be raised
any higher. The Re must be replaced by Rei>. This gives us the scale of Marva that
(No. A 10), and we have finally progressed beyond the primary thats.
Ex. 36. Marva that
Ga MfiS (P a ) Dha Ni Sa
------- £2---
H* v«,T"
Uha Ga M a# (Pa)
3>t »— ...................
--------------- = # ■ =
It will be seen that we have put the Pa in brackets; although the Pa is necessary to
the definition of that in which each of the seven degrees is required, neither tetrachord
will be balanced unless the Pa is omitted. In view of this it is not surprising that many
of the rags in this that, including the rag Marva after which the that is named, are
hexatonic and omit the Pa. Just as in Kalyan, the omission of Pa means that the Sa
has lost one of its supports and there is a strong tendency to omit the Sa temporarily
in particular phrases. This phenomenon is so characteristic of rag Marva, that its
ascending and descending lines have been given as follows ‘3
Ex. 37. rag Marva
Ni Re^ Ga M ai D ha Ni Sa Ret- N i D ha Mhlt Ga Ret> Ni Dha Sa
¥ 1
Earlier we had indicated that the dissonant quality of the Ma# tended to prevent
the use of this note as the initial of a tetrachord. In view of the fact th at the Pa is
omitted completely and the Sa is frequently omitted in certain phrases we could
consider the parallelism occurring in the descending conjunct tetrachords, Ni-Ma#
and Ma#-Ret>, with a semitone disjunction at either end.2
Ex. 38.
1 This is sung in demonstration by Ram Narayan, the sarangi player, on the long-playing record
B.A.M. LD 094.
2 The conjunct parallelism could also be expressed as being in the two tetrachords Dha-Ga and
Ga-Ni (VI-III and III-VII).
83
If we now consider the scale of rag Marva, without the Sa and Pa, it will be seen
that the ascending disjunct tetrachords are unbalanced due to the augmented fifth
relationship of the Reb and the Dha. This can be corrected either by replacing the
Reb with a Ren or the Dhan with a Dhab. The former returns to Kalyan (see Ex. 32),
the latter to the new that, Purvi (No. A9).
Ex. 39. Purvi that

Ni Sa Reb (Mafy Pa Dhab Ni Sa
Ga
It will be seen that the ascending disjunct tetrachords of Purvi that are parallel if we
think of them from Ni to G a and from Ma# to Ni. If we go further, either the Sa must
be omitted temporarily, as is sometimes the case in rags of this that, or the Mas must
be replaced by a Man (in order to provide balance for the Sa), thus making two
parallel ascending disjunct tetrachords, Sa-M a and Pa-Sa. This leads to Bhairav
that (No. C9) which is not part of the Circle of Thats, but is, nevertheless, of great
importance in Indian music.
Before discussing Bhairav that in greater detail, we propose to complete our discus
sion of the Circle of Thats. Returning to a consideration of Purvi that (Ex. 39), it will
be seen that the descending conjunct tetrachords are unbalanced in two respects:
first, the Pa has no perfect fifth to support it, so there is a tendency to omit this note in
some rags of this that, just as in the rags of Marva that; and secondly, the Dhab
is not balanced by the Ga. This can be corrected either by replacing the Dhab with
Dhai), which leads us back to Marva that, or by replacing the Gat) with Gab, leading
us to the new that in the Circle, Tori.
Ex. 40. Tori that

Ni Ga Ma$ (Pa) Dhab
Ma* (Pa) Dha1 Maft (Pa)
Since the Pa has no support in the descending conjunct tetrachords, it is not

surprising that one of the very few rags in this that, Gujri Tofi, is hexatonic omitting
the Pa. The rag Tori (also called M iy t K i Tori), too, manifests this same tendency
84
to omit the Pa in particular phrases, and, correspondingly, to omit the Sa which is
unbalanced in the ascending disjunct tetrachords.1
The ascending disjunct tetrachords of Tori are unbalanced in two respects. The Sa
has no perfect fourth, and, as we have indicated, is frequently omitted in phrases in
the rag Tori as well as in Gujri Tori. There is not the same urgency as in Purvi that
to replace the Ma# with a Mab to support the Sa, as the ascending disjunct tetra-
chords will still remain unbalanced due to the augmented fifth relationship of the
Gab and Nib. This second imbalance can be corrected either by replacing the Gab by
a Ga^, which takes us back to Purvi that, or the Nit? by a Nib, which leads us to the
hypothetical that (No. A7) and completes our Circle of Thats. Before leaving this
Circle, we must consider how the hypothetical that could also have been arrived at
from Bhairvi. It will be seen from the tetrachord scheme of Bhairvi (Ex. 31 on p. 81)
that in the descending conjunct tetrachords the Pa and Reb are unbalanced. We have
already mentioned one possibility of creating balance, that of replacing the Reb by a
Re??, and have indicated that the second possibility of replacing the Pa by a Pab is
not acceptable in Indian music. In discussing Kalyan that, which has similar problems
to those of Bhairvi, we had suggested that the musician would naturally try to
minimise the effect of this imbalance, and would tend to omit one or both of the
offending notes. This is readily apparent in Bhairvi that also, where in the very pro
minent pentatonic rag Malkos just these two notes, the Reb and Pa, are omitted. In
the rag Bhairvi too, the same tendency is also manifest and the Reb and Pa are
frequently omitted in particular phrases (see p. 125). Once the Pa is omitted, the Reb
can only be balanced by changing the Mai? to Ma#, and thus one could arrive at
the hypothetical that.
Having argued that the hypothetical scale could have arisen in Indian music, we
must now attempt some explanation for its absence in the current repertoire of the
musician. In one respect the scale of the hypothetical that is musically unstable.
Ex. 41.
Sa Re'* G a1* Ma& Pa D h a6, Sa
From the consonance-dissonance charts (on p. 66) it will be apparent that the most
dynamic notes in the octave are the Reb (IIb), Ma# (IV#), Dhab (VIb) and Nib (VII).
The Reb and the N i demand resolution in the ground-note, Sa, while the Ma# and the
Dhab demand resolution in its fifth, Pa. When Pa is a secondary drone, the demand
for resolution on it is intensified. Three of these dissonant notes, Reb, Ma# and Dhab,
are in the hypothetical that; as there is no Nib, only the Reb resolves in the Sa, while
both Ma# and Dhab demand resolution on the Pa. As a consequence, it may very
easily be that the Pa has a strong tendency to usurp the place of the Sa and lead to a
1 This rag can be heard on the L.P. record ASD 498 (E.M.I.) played by Ustad Imrat Khan on
the surbahdr.
85
plagal inversion.1 There would also be a very strong tendency to introduce Ni*i both
as a leading note in ascent2 and to provide symmetry in the descending conjunct
tetrachords.
From the standpoint of balance too, this scale appears to be more unstable than
the others. As in Marva, Purvi and Tori thats there is a tendency to introduce Mat;
to balance Sa, but in this that Man is a double balance note since it also balances Nib.
Thus there must be a very strong tendency to introduce Man and move the scale back
to Bhairvi. If instead Nin were introduced to balance Mas, the resulting scale, Tori, is
not nearly so unstable as the Man would only balance one note, Sa, while putting Nin
out of balance.
Ex. 42. Hypothetical that
Sa R e G a ^ MaS (Ma8) Pa Dha*> Ni^ Sa ^
We now return to a further consideration of Bhairav that and the related pheno
mena which lead to the curious scales of the Lalit rags, which use both Man and Mas
and are thus outside the thirty-two that system. In the ascending disjunct tetrachords
of Purvi we have already explained how, if we progress beyond the tetrachords N i-G a
and M as-Ni, either the Sa must be omitted or the Ma# must be replaced by the Man
forming Bhairav that.
Ex. 43. Bhairav that
Sa Ret* Ga Ma Pa Dhak
Bhairav that has an extraordinary structure. The ascending disjunct tetrachords

are parallel while the descending conjunct tetrachords are completely unbalanced.
1 In Western terms the Sa lacks a leading note, while the Pa has one and thus tends to receive
greater importance than the Sa. The resulting scale:
Sa R et G at Ma M at D hat N it Sa
is itself unstable, having no perfect fifth and no semitone leading notes below either the Sa or the
Ma. There are at present two rags in which the Pa has a leading note, while the Sa does not. These
two rags, Madhukos (Madhukaus) and Madhukant, are both modern rags which appear to be evolving
rapidly. These are discussed in Chapter VII (pp. 136, 137).
2 See pp. 113, 114.
86
Perhaps the popularity of this scale can be explained, in part at least, by the fact that
there is both simple parallelism in the disjunct tetrachords as well as complete im
balance in the conjunct, thus making for extreme contrasts between symmetry and
asymmetry. The imbalance in the descending conjunct tetrachords can lead to the
replacement of any of the five movable notes by their chromatic counterparts. Some
of the resulting new scales still bear names which hint at their origin, Ahir Bhairav,
(No. C3), Nat Bhairav (No. C2) and Anand Bhairav (No. BIO). In fact, all thirty-two
thats could eventually be obtained through Bhairav.
There is no doubt, however, that Mas is the most prominent alternative note in
Bhairav that> and occurs as an accidental in a number of rags in this that (e.g.
Ramkali, see p. 49). It is probably this impulse which has also led to the introduction
of the rag Lalit (No. 34) which appears to be a compromise between Bhairav and
Purvi thats. It can be seen more clearly as a derivative of Purvi. The Reb (IIb) and
Pa (V) are unbalanced in this that and we have noted earlier that the Pa is some
times omitted. Since this leaves the Sa without either a perfect fourth or a perfect
fifth it seems very plausible that the Mail (IVii) is introduced to provide conjunct
balance for the Sa. In other circumstances the Ma? would probably have been
omitted, but here it is supported by both Reb and N i:
Ex. 44. rag Lalit
Ni Sa Reb Ga Mall
It will be seen that the ascending disjunct tetrachords hfi-G a and M as-N i are
balanced only when the Sa is omitted as is frequently the case in this rag.1 The version
of Lalit given by Bhatkhande (No. 33) can be derived in a similar manner from
Marva that, the Pa (V) being omitted as it is unbalanced with the Reb, the Mab (IVii)
being introduced to balance the Sa, and the Ma# (IV#) being too well supported by the
Reb (IIb) and the Ni (VII) to be omitted:
Ex. 45. rag Lalit
Ni Sa Reb Ga Man
Ga Mah Ma#
1 See p. 51.
87
The rag Ahir Lalit (No. 35) can be explained as a derivative of this rag, having a
further conjunct balance, the Nib being introduced as a support for the Ma^:
Ex. 46. rag Ahir Lalit
—»
Sa Reb Mall
S
Ga Ma^ Man
These three Lalit rags can also be explained in another way as plagal inversions
(i.e. the series beginning on Pa) of three thats: the inversion of Purvi gives Bhat-
khande’s version of Lalit (No. 33); the inversion of Tori gives the commonly heard
Lalit of today (No. 34) and the inversion of Bhairav gives Ahir Lalit (No. 35). This
explanation does not appear to be very plausible, particularly in connection with the
two Lalits since the Pa is generally weak in the rags of Tori and Purvi thats and there
seems no reason for the plagal inversion. However, Ahir Lalit may well have been
derived in this way.
In spite of the strange scale of these Lalit rags, their musical justification in terms
of balanced tetrachords will be evident (in Ex. 45, D ha-G a and G a-N h descending
conjunct, are parallel; in Ex. 46, Sa-M a and Ma-Nib, ascending conjunct). However,
balanced tetrachords are difficult to achieve in some of the ‘isolate’ rags which have
recently been introduced into N orth Indian music. This may indicate that the
importance of balanced tetrachords is diminishing. The rag Carukesi (No. B3), for
example, has unbalanced conjunct as well as disjunct tetrachords:
Ex. 47. rag Carukesi

Ga Ma Pa Dhab
Pa Dha^
Can it be that one of the fundamental impulses in N orth Indian music is under
going change or modification, or will time ensure that balance is somehow created
in one of its tetrachords ? The pentatonic rag Hamsdhvani which was also taken from
South Indian music, probably more than forty years ago,1 now shows signs of artificial
balance being introduced to compensate for an inherently unbalanced scale:
1 Bhatkhancje described this rag as one which had been introduced from Karnatic music and was
quite rare in his time. K.P.M. V, p. 255.
88
Ex. 48. rag Hamsdhvam
— ->
This is accomplished by means of turns in which Sa and Re are temporarily omitted,

for example in the following phrases, where in (a) Ga Pa is parallelled by Ni Re,
and in (b) Ni Pa is parallelled by G a Sa.1
Ex. 49. rag Hamsdhvam

(a) (b)
Ga Pa
In the rag Carukesi, too, there are indications of similar to n s which produce a
measure of symmetry within the rag, but these are not yet clearly defined. It is too
early to tell if the urge to balance tetrachords has now weakened and will finally be
lost, or whether it will reassert itself once the novelty of these new scales has worn off.
1 This can be heard on the recording of the rag Hamsdhvam played by Ravi Shankar, E.M.I.
ALP 1893.
89
V
Evolution o f the Circle o f Thats
W hile the ten thats of Bhatkhande’s system are musically coherent through their
relationship to each other, the course of their evolution was neither coherent nor
systematic. It was frequently diverted by mutations which were either retrogressive
or led to blind alleys or were too far in advance of their period to be accepted. So it is
not entirely surprising that the thats have not evolved in an orderly manner around
the Circle.
Indian music appears to have undergone a fairly radical change probably some
time before the 10th century a.d., a change which may have been associated with
the replacement of the ancient bow-harp vina (which is now obsolete) by stick zithers
and long-necked lutes, which are now commonly found in India. In the introductory
chapter we outlined the musical system described in the Natyasastra, perhaps the
earliest work on Indian secular music. Here the music was based on seven serial
modes (suddhajdti) derived from two very similar parent scales, the Sadjagrama and
the Madhyamagrama.1
A system based on serial modes of this nature implies a difference in the pitch of
the starting note of each mode, particularly in application to musical instruments.
The practice of changing ground-notes for the successive modes can be accomplished
easily on bow-harps where, presumably, all the strings are melody strings and are
tuned to the successive degrees of the scale. This could, however, cause considerable
inconvenience when applied to stick zithers and long-necked lutes where, as today,
a number of strings could be used to supply the drone and would have to be retuned
for each change of mode. Thus it seems probable that the prominence of the drone2 is
associated with the widespread acceptance of stick zithers and long-necked lutes and
that the inconvenience as well as the practical difficulties entailed in retuning the
instruments led to the acceptance of a standard ground-note, Sa, to which the modes
1 A third parent scale, the Gandharagrama, mentioned in other early works does not appear to
have been significant in Bharata’s time, although it is referred to under the section on Marjana
(tuning of drums). Nafyasastra, Chapter XXXII, si. 37, Bib. Ind. no. 272. The system of ‘pure’
jatis had already been extended, in the Natyasastra, to include eleven other ‘altered’ modes (vikrita
jati).
2 Unfortunately, the early treatises do not mention the drone and the later treatises follow their
pattern. As a consequence, there is even now no commonly accepted term for drone in North India.
90
were then transposed. As we have indicated earlier, this does not imply a fixed and
invariable pitch for the Sa, which merely provides a common frame of reference for
the different modes, just as does the D oh in the Western sol-fa system.
When serial modes are transposed to a common tonic, they lose their cyclic
significance, and it is neither so obvious nor of particular importance that they are
derivatives of a parent scale or scales. The acceptance of a standard tonic, influenced
by the requirements of stick zithers, was probably responsible for the loss of distinc
tion between the two ancient parent scales, although it is not until much later that this
is clearly admitted.1 The subtle difference between the four sruti m ajor whole-tone,
the three sruti m inor whole-tone and the two sruti semitone in the modes of the two
parent scales would require about eighteen frets (or stops) in the octave and these
would have to be positioned so close to each other as to be impractical.2
Six of the seven serial modes of the early period have continued in Indian music to
the present time, reflecting the continuity of tradition in India. It must be remembered,
however, that these six, the primary thats, are the most consonant possible as they
have only one tritone (augmented fourth/diminished fifth relationship), the other
fourths and fifths being perfect. Consequently, they might arise in any musical
culture in which the consonance of perfect fourths and fifths is recognised. The
fact that the seventh mode, the ja ti Dhaivati (B mode), prominent in ancient India, is
not now used in Indian music is an indication that the seven ancient modes were
not held in any great sanctity after the changes had entered the musical system.
Many of the N orth Indian treatises dating from about the 11th century a .d .
describe rdgas in terms of the hierarchy, rdga (masculine) and rdgini (feminine), a
fanciful classificatory scheme which was extended to include putra (son) and bhdryd
(wife of the son) as the number of rdgas increased.3 These rdga-rdgini classification
schemes do not appear to have been based primarily on musical principles. In the
most commonly used schemes there are six masculine rdgas, each having either five
or six rdginis. There are, however, divergent opinions as to which are the six rdgas
and which are their rdginis. According to Muhammad Reza, c. 1813, four principal
traditions appear to have been current,4 but Gangoly presents a more realistic view
when he draws attention to the ‘bewildering variety of catalogues, groups and classi
fications5.5
Concurrent with the rdga-rdgini schemes occur the much less fanciful classification
systems based 0 1 1 scale, mela (present-day that), which presuppose that all scales have
been transposed to a common tonic. The majority of the treatises adopting this
method appear to have been describing the South Indian musical system. One of the
1 For instance in the Rasakaumudl of Sri Kaotha (probably 18th century) and Saftgltasaramrita
of Raja Tulaja (18th century) in both of which only the Sadjagrama is said to be in existence.
2 See p. 21.
3 Sai)gitamakaranda of Narada which is probably the first text in which the melodic system is
classified into three groups, masculine, feminine and neuter. Gangoly, op. c/7., p. 23.
4 Quoted in Gangoly, op. cit., p. 220.
5 Ibid., p. 92.
91
earliest N orth Indian texts of this nature is the Ragatarangini written by Locana
Kavi (probably 16th or 17th century),1 a work of considerable importance for evidence
of the evolution of modern N orth Indian rags.
Locana classifies ragas in terms of the following twelve scales.2
Lo cana’s
me las Scale Modern that
Sa Re G at Ma Pa Dha Ni^ Sa
Bhairav! 4m KafI (No. A4)
Sa Rc^ Gat jâ pa Dhat Ni^ Sa
Tor! Bhairvi (No. A6)
Gaurl Bhairav (No. C9)
Karnata Khamaj (No. A3)
Sa Re Ga Ma Pa D ha Ni Sa
Kedara Bilaval (No. A2)

m
Sa Re Ga Ma# Pa D ha Ni Sa
Imana Kalyan (No. Al)

¥
Sa Re Ga# Ma# Pa Dha# Ni Sa
Saranga
* »■ »■ • t|- * '!
Sa Re Ga Ma Pa Dha# Ni Sa
Megha
Sa R et Ga Ma# Pa Dha^ Ni Sa
DhanaSrf Purvi (No. A9)
1 The date of this work is discussed on p. 20, f.n.4

2 These twelve scales are also mentioned in Hridayaprakasa of the late 17th century.
92
Dha1* Ni
Purava
Sa Re Ga*- Ma Pa D ha1, N i1, Sa
Mukhari Asavri (No. A5)

Dlpaka no description
It will be apparent that eight of the modern ten thats were in existence at the time
o f Locana, probably more than three hundred years ago. The melas Sarahga and
Megha involve the use of Gas and Dhas which is not permissible today, and they
would therefore appear to be chromatic scales from the modern point of view. But
if we interpret the Gas and Dhas as their chromatic counterparts Mat; and Nib
respectively,1 the description would then be consistent with the modern rag Suddh
Sdrang which is largely pentatonic but uses both alternatives of M a and Ni. Similarly,
Megha may be interpreted as a hexatonic scale having both alternatives of Ni.
Puravd is, however, a different matter, for according to the Rdgatarahgini the D ha
in this scale is raised by one sruti.2 Since Puravd appears to have no counterpart at the
present time, it may be described as an unsuccessful m utation.3
The modern Marva (No. A10) and Tori (No. A8) thats are not among Locana’s
melas and appear to have evolved since his time. This is a clear indication that the
modern Bhairav (No. C9) and Purvi (No. A9) thats did not originate as parts of the
Circle of Thdts as there are discontinuities in both clockwise and anti-clockwise
directions. The following diagram shows the discontinuity of Locana’s melas in
terms of the Circle of Thats:
1MANA (A1)
(A 10)
DHANASRI KARNATA ( A 3 )
(A 9)
GAURl
<C9)
(AS) BHAIRAVT (A 4 )
( A7) MUKHARl ( A 5 )
TOR! (A6)
1 This interpretation finds support in the 18th-century treatise Saiigitasardmrita, where the
author, Tulaja, discussing Sdraftga mela, equates Gait with Mali (Music Academy Series, No. 5,
1942, p. 111).
2 Hridayaprakasa (second half of the 17th century) which follows the same scheme as the Raga-
tarafigitfi, states that the Dha of Purava is very sharp (tivratara) as are also the Ga and Ma. Quoted
by V. N. Bhatkhancje, A Comparative Study, p. 30.
3 Purava does not seem to have been a prominent mela even in Locana’s time, for only one rdga
is given in it, namely, Puravd.
93
It was suggested by the late Dr. A. A. Bake that the scale of Bhairav was first
introduced into South India, probably from Persian or Arabic music, under the name
of Hejuji (Hejciz, Hajiz, Hijej), from whence it spread into N orth India appropriating
the name Bhairava from a rag which was then losing popularity.1 This may be a
possible explanation in view of the fact that the Durrat al-Taj, an encyclopaedia in
Persian dated c. 1300, describes the tetrachord Hejaz as having an augmented second
interval (i.e. Reb-Ga) which is characteristic of Bhairav that. This argument is by
no means conclusive, for this scale has the most rdgas in Locana’s system, a fact
which would suggest that it had been in existence in India for a considerable period
before the 16th century.
Since Locana gives only two rdgas in what is modern Purvi that, we may deduce
that this scale was of more recent origin than Bhairav and it is very probable that
Purvi evolved out of Bhairav. Locana mentions two versions of the raga Dhanasri,
one in Dhanasri mela (modern Purvi that) and the other in Gauri mela (modern
Bhairav that). This establishes a connection between the two fhdts2 but does not help
us to determine which of these originated earlier. However, a number of rdgas given
by Locana in Gauri mela are now classified in Purvi that, for example Malvi, Triveni,
(Puriya) Dhanasri, Vasant, Reva and Jetsri. In addition, the rag Gauri now occurs in
two main variants, one in Purvi, the other in Bhairav, while the rag Vibhas occurs in
three, one each in Purvi and Bhairav and one in Marva that as well. There seems,
therefore, to be sufficient evidence to substantiate the theory that the rags of Bhairav
that have a tendency to merge into Purvi that,3 and this, we suggest, may have been
responsible for the origin of Purvi that.
Some rdgas of Locana’s Gauri mela are now in modern Tori that, the present-day
Multdni and Gurjri (Gujri Tori), for example; others are in modern Marva that:
Bhatiyar, one version of Vibhas, {Mali) Gaurd, and perhaps Locana’s Mdlava which
may be the origin of modern Marva. N ot all the rdgas in Gauri mela have changed,
however, and the rdgas Gauri, Bhairava, Vibhdsa, Ramakali, Gunakari and one version
of Khata still belong to Bhairav that.
The relationship between Locana’s Gauri mela ragas4 and their present-day
equivalents can be shown more clearly in tabulated form :
1 That Hijej (cf. Hejuji) was an importation was first suggested by William Jones in 1799. Music o f
India, Calcutta 1962, p. 94. A further connection between Bhairav and Hejuji is suggested by_the
divergent readings in the various manuscripts of the Kitab-i-Nauras composed by Ibrahim ‘Adil
Shah II, c. 1600. Both Bhairav and Hdjtz (cf. Hejuji) are mentioned in some of the manuscripts.
In others, however, only one or the other is mentioned; the songs which are said to be sung in
Bhairav in one manuscript are ascribed to Hajiz in others. Kitab-i-Nauras translated by Nazir
Ahmed, Delhi 1956, p. 62 f.n.
2 The occurrence of two variants of a rag may be purely coincidental, i.e. they may have com
pletely different origins. On the other hand, if the two recensions are closely related, it is very pro
bable that one has evolved from the other, the original being preserved perhaps in a different part
of the countiy.
3 There are, however, also indications that rags of Purvi that merge into Bhairav that.
4 A full list of Locana’s rdgas is given in Gangoly, op. cit., p. 196, and in Bhatkhanê, A Compara
tive Study, p. 21 (in Devnagri script). The deviant readings of Locana’s rdgas in the latter are given
in brackets.
94
Locana’s Gauri mela Modern
rdgas (modern Equivalents
Bhairav that) (conjectural) M odern thats
’ M arva Marva that (No. A lt
M alava
Malvi Purvi that (No. A9)
Trivana Triveni 5? J? 5?
Dhanasri Puriya Dhanasri 55 55 55
Vasanta Vasant 55 55 55
Reva Reva 55 55 55
Jayantasri (JayatasrI) Jetsri 55 55 55
"Vibhas (1) 55 55 51
Vibhasa <Vibhas (2) Bhairav (No. C9)
_Vibhas (3) Marva (No. A 10)
‘ Gauri (1) Purvi (No. A9)
Gaura Gauri (2) Bhairav (No. C9)
„Maligaura Marva (No. A 10)
Bhairava Bhairav Bhairav (No. C9)
Ramakali Ramkali J5
Gunakari (Gunakari) Gunkri (Gunkali) JJ
'K h a t( l) 55 55
Sadraga (Khata)
K hat (2) Asavri (No. A5)
Saveri Bhairav (No. C9)
Asavari
Asavri Asavri (No. A5)
Bhatiyara Bhatiyar M arva (No. A 10)
M ulatani (Mulatani) Multani Tori (No. A8)
Gurjarl Gurjri (Gujri) Tori
Desakara (Desakari) Deskar Bilaval (No. A2)
Devagandhara Devgandhar Asavri (No. A5)
"Desi Tori (1) 55 5, (?)
Desi Tori <
. „ „ (2) K afi (No. A4) (?)
Six other rdgas mentioned by Locana in this mela do not exist today. It will be seen
that the ragas of Locana’s Gauri mela are distributed principally among the modern
Bhairav, Purvi', Marva, Tori and Asavri thats. The first four of these are obviously
related, while Asavri is three steps removed from Bhairav (see diagram on p. 184).
There are indications, however, that the Asavri that rags above may have evolved
from a different tradition from that which is represented in the Ragatarahgini.
Pundarika Vitthala (late 16th century) gives several versions of the rdga Devagdn-
dhara in his works. In his Sadragacandrodaya its scale appears to be that of our
modern Bhairav that, while in his Ragamahjari its scale is that of our modern K dfi
thatd It is probably from this second version that the existing Devgdndhdr has
descended. Similarly, in Ragalaksana2 two variants of Asavari (Saveri) are men-
i H.S.P. IV, p. 476. ^ Ibid., p. 438.
95
tioned, one in modern Bhairav and the other in modern Bhairvi. The former is very
commonly heard in South Indian music as the rag Saveri and occasionally in the
N orth, while the latter is probably the ancestor of the modern N orth Indian Asavri.
To recapitulate, it would appear that the early period of the seven serial modes
was followed by a period of expansion and exploration in which the modes were
altered and mixed and basic changes were accepted within the system. As a result
either of this exploration or of foreign influence Bhairav that was introduced and
because of its melodic qualities was rapidly absorbed into the system. We have
suggested that Purvi that may well have evolved out of Bhairav which, as we have
indicated in the previous chapter, is extremely prone to change. It seems probable
too that the scales of Tori and Marva also grew out of Bhairav. Both of these are two
notes removed from Bhairav, i.e., for Bhairav to evolve into Tori, two notes must be
changed: the Gan must become Gab and the Man must become Ma#. Similarly, for
Bhairav to evolve into Marva, the Dhab must become Dhan and the Man become
Mas. If the course of evolution is gradual, these changes will take place one at a
time, and form in transition one of two possible intermediate scales depending on
the note which changes first. In both cases if M a changes first, the intermediate scale
will be Purvi. If in the process of evolution from Bhairav to Ton the G a changes
first, the intermediate scale will be No. B6 (see p. 184). If in the process of evolution
from Bhairav to Marva the Dha changes first, the intermediate scale will be No. BIO.
Let us consider the Ton rags, Gurjri and Multani, which are classified by Locana in
his Gauri mela (modern Bhairav that) but are now in Tori that. Ahobala in the
Sahgitapdrijata (second half of the 17th century) describes two forms of Gurjri,
Daksina (Southern) and Uttara (Northern).1 The former has the scale of our Bhairav,
while the latter has Reb, Gab and Dhab and belongs to scale No. B6. The rag M ultani
too seems to have evolved through this same intermediate scale, for Ksetramohan
Goswarm, in the second half of the 19th century, says that some use the Ma*i instead
of the Ma# in Gurjri Ton,2 evidently a reflection of an earlier tradition. Thus the
evolution of Gurjri and Multani appear to have proceeded as follows:
Sa Reb Ga Ma Pa Dhab Ni Sa
L ocana’s Gauri mela ragas— Gurjri,
—\r» - +— - 9■ —- .......... ^ M ultani
s'a Rel> G ab Ma Pa D h ab Sa A h ob ala’s raga, Uttara Gurjri (17th

century)
b«~~ * * * ■ =ft and G osw am l’s raga M ultani (variant)
(19th century)
Sa R eb G ab Ma§ Pa Dhab Ni Sa
M odern Gurjri and M ultani

^ if* *— ■ -----
1 H.S.P. IV. pp. 699-700. 2 Ibid., p. 716.
96
The evidence for the origin o f Marva that is not nearly so conclusive, primarily
because the many traditions of the rags in this that are somewhat confused. For
instance, the rag Marva could have its origin in Locana’s raga Malava of Gauri mela,
in his raga Maru of Karyata mela (modern Khamaj that), or in his raga Maru of
Keddra mela (modern Bildval that). The raga Vibhasa given in Locana’s Gauri mela
now occurs in three versions, in Bhairav, Purvi and Marva thats. This suggests that,
in this case at least, Marva that may have evolved through the intermediate Purvi
scale. The evidence for the intermediate scale of the present rag Bhatiyar, given by
Locana in Gauri mela but now occurring in Marva that, is also inconclusive. It seems
to have maintained its Bhairav scale into the second half of the 19th century.1
Perhaps the only evidence which can be brought to bear on this m atter is connected
with the present use of the Man as an accidental in this rag which may be an indication
that the process of evolution of the rag Bhatiyar from scale No. BIO is not yet com
plete. The general principles relating accidentals to the evolution of mgs will be
discussed in the following chapter.
So far we have been discussing the evidence for the origin of Tori and Marva thats
from Bhairav that. This evidence tends to diminish the significance of our Circle of
Thats. There is, however, im portant evidence to prove that both Tori and Marva. thats
have also been derived around the Circle, the former from Bhairvi that»the latter from
Kalyan.
Let us first consider the case of Tori, which from the evolutionary standpoint is
one of the most interesting mgs. From the 15th to the end of the 18th century it is
said to have had the scale of modern Bhairvi, as indeed it still does in South Indian
music. This is stated in a number of texts some of which apparently describe South
Indian music, e.g. Rdgavibodha of Somanatha c. 1610, while others such as Rdga-
tarahgiiii and Sahgitapdrijata describe the N orth Indian system. The present-day
N orth Indian Tori has, of course, a very different scale, and it is usually thought
that it represents a different tradition from the earlier Tori. This, however, does
not appear to be so.
In the Sahgitsar compiled in Hindi by Maharaj Sawai Pratap Simh Dev of Jaipur
after a conference of the leading musicians in his court (c. 1800 A.D.), several varieties
o f Tori are described. Among them we find a rag called Marg Tori which has the
scale of our hypothetical that (No. A7), and has, in addition, Msm as an oblique
descending accidental. The rag is given as hexatonic, the Pa being said to be omitted,
but in the example given, Pa occurs in the ascending line.2 The common Tori of
this period is called M iya k i Tori and still has the scale of the original Tori, that of our
1 H.S.P. Ill, p. 330, quoting Goswaml.
2 Quoted in H.S.P. IV, p. 691. In the Satigitsar the notations are given in the form o f ‘magic
squares’ (jantr), in this case consisting o f three vertical columns of notes. In the following example,
they are transcribed in sequence with the bar lines separating the columns:
Rag Marg Tori o f Satigitsar
Dhat N it D hat MaS G at R e t MaB Pa D hat M a# G at R et Mali G at R e t Sa
7 97
modern Bhairvi that. In Marg Tori we can see, however, a departure from this scale
towards the present-day N orth Indian Tori.
This is not the only available evidence to corroborate this thesis. In a Bengali
work by Ksetramohan Goswami (late 19th century) Tori appears to lie between
scale No. A7 and our present Tori. From the notations given in this work,1 its
ascending and descending lines may be given as follows:
Ex. 50. rag Tori of Goswami
Sa RcVG at M a# Pa N it D h a t Ni# Sa Sa N it- D h a t p a M a# G a t Re t Sa
£ £
In addition, Nib occurs occasionally as an oblique ascending note—perhaps a vesti

gial reminder of the original scale of the rag.2
The connection between Tori and Bhairvi thats can also be established in a different
way. Two modern rags, Bhiipal Tori and Bilaskhani Tori, are even now classified in
Bhairvi that. The former is pentatonic, while the latter is identical in ascent but is
heptatonic in descent:
Ex. 51.
(a) Bhiipal Tori and
Bilaskhani Tori (ascent) (b) Bilaskhani Ton (descent)
Sa R et G at Pa D hat sa Sa N it D hat pa Ma G at R et Sa
I
The omitted notes M a and Ni are, in fact, the very notes which distinguish Bhairvi that
from Tori, and the rag Bhiipal Tori could equally well be classified in modern Tori that.
In view of this evidence it seems probable that the rag Tori has evolved in the fol
lowing manner within the past two hundred years:
Sa R et G at Ma Pa D hat N it Sa
Locana’s Tori, 17th century
(modern Bhairvi that)
Sa R et G at Ma# Pa D hat N it Sa M a#
d e s c e n d in g = = = i Pratap Simh’s Marg Tori,
t accidental c. 1 8 0 0
Sa R et G at M a# Pa D hat N it Sa N i#
ascending;
Goswaml’s Tori, c. 1870
accidental:
1 Quoted in H.S.P. IV, p. 695. Bhatkhande does not specify which o f Goswaml’s works he is
quoting from, his Saiigitsar which appeared in 1868 or his Safigitsdrsaiigrah which appeared in
1875.
2 This statement will be clarified in the next chapter.
98
Sa Rel> G at M ali Pa D hat Ni Sa
Modern Tori
4 — !>»
We mentioned earlier (p. 63, f.n. 2) that the rag Tori was one of the principal excep
tions to the time theory of rags, as its traditional time of performance was the late
morning, whereas according to the time theory it should be performed either at 3 a.m.
or 3 p.m. Texts such as the Rdgavibodha (c. 1610) and Pundarika’s Rdgamdla (late 16th
century) which state that the raga Tori should be performed in the morning are
undoubtedly referring to the Tori of their period which, as we have shown, had the
scale of modern Bhairvi, The traditional time of performance has remained associated
with the name of the rag while its scale has evolved to its present state.1 This is an
excellent example of the imperceptible working of the evolutionary process.
We have now considered the evidence for the evolution of Ton from Bhairvi
through the hypothetical that, No. A7, and incidentally provided some justification
for the inclusion of the hypothetical that in our Circle of Thats. We now present the
evidence which suggests that Marva that has evolved from Kalyan that around the
Circle in an anti-clockwise direction.
Among the ragas placed by Locana in Imana mela (modern Kalyan that) we will
first consider the raga Puriya which is now in Marva that. There can be no doubt that
the modern tradition is directly connected with Locana’s P u riy a for in the Hridaya
prakasa (middle 17th century), a work which closely follows the Ragatarahgini,
there is mention of the raga Puriya Kalydna (in Imana mela),2 which also exists in
the present period in Marva that. The occurrence of the rag Puriya, Kalyan in Marva
that obviously suggests its connection with Kalyas that. In the previous chapter we
had suggested that Marva that might have arisen from Kalyan by a comparison of
the tetrachords of Kalyan that in which the Sa and Pa had been omitted.3 This would
certainly seem to apply to the rag Puriya (the origin of rag Marva is uncertain as we
have suggested earlier) which, like the rag Marva, is hexatonic, omitting the Pa.
The connection between Kalyan and Marva thats is further suggested by the raga
Jayata Kalydna (modern Jet Kalyan). Both Locana and Hridaya give this raga in
Imana mela and it does not appear to have changed since it is still classified in Kalyan
1 Some of the other rags deviating from the time theory may perhaps also be explained in this
way. In the diagram showing the time theory in relation to the Circle of Thats (p. 64) Bhairvi that is
indicated at approximately 12 o ’clock. In Locana’s system Bhairvi had the scale o f modern Kafi
that which is indicated at about 10 o’clock. This still does not compare with tradition which ascribes
Bhairvi to the early morning. It is quite evident that, in the case of some rags at least, there are other
factors involved.
2 Hridaya Narayapa classifies his rdgas in two ways: in the Hridayakautuka he follows Locana’s
melas, but in the Hridayaprakasa he uses melas based on the number o f altered notes. Thus the
term Imana mela is taken from the scheme used in the Kautuka. It is interesting that, in this mela, he
mentions the raga Puriya in the Kautuka and Puriya Kalyan in the Prakasa, the other three rdgas
in this mela being the same in both works. The full lists of rdgas are given in Gangoly, op. cit.,
pp. 207-8.
3 See p. 83.
99
that. It does appear, however, to be a combination of the modern rags Jet and
(Suddh) Kalyan• The former is a rag of Marva that in which some musicians, accord
ing to Bhatkhande, use both alternatives of Re.1 The Re is, of course, the distinguish
ing note of the two thats Marva and Kalyan, and the use of the Reb as well as the Refc
in the rag Jet shows a merging of the two thats.
Many rags have undergone change in the past four or five centuries, generally
moving from one of these ten thats to another. M ost o f our discussions have been
limited to thats on the left side of the Circle, from Kalyan to Bhairvi and including
Bhairav. It is not only the rags of these thats which have been subject to the process of
evolution and before we close this chapter we will briefly consider a few examples
from among the six primary thats.
The first of these is the very commonly heard rag Bhairvi. Early writers representing
both the N orth Indian school (Locana, for example) and the South Indian school
(for example, Ramamatya in the Svaramelakalanidhi, middle 16th century) give its
scale as modern K afi that (No. A4). In the 17th century both schools—Ahobala in
the Sahgitapdrijata representing the N orth and Venkatamakhi in the Caturdandi-
prakasika representing the South—give Bhairavi as modern Asavri that (No. A5).
This remains the scale of the rag Bhairavi in South India. In N orth India, however, the
evolution of this rag has progressed one step further and Pratap Simh (c. 1800) is
probably the first writer to give the scale of modern Bhairvi. Although the scale o f
Bhairvi has not changed since this period, the occasional use of the Ma# (IV#) as an
unconscious diminished fifth indicates that the next evolutionary step, the B mode,
has already been considered, a step which may never be completely achieved as the
diminished fifth would be in direct relationship with the ground-note.
The raga Mdlakausika (modern Malkos) is described by Locana as belonging to his
Karnata mela (modern Khamaj that, No. A3). This same rag belongs to modern K afi
that (No. A4) according to the description in Ragamahjari of Pundarika Vitthala
(late 16th century). Malkos is now a rag in Bhairvi that (No. A6), and there appears
to be no textual evidence to indicate that it has ever been a rag in Asavri that (No. A5)
which lies between K afi and Bhairvi in the Circle. This is not entirely surprising in
view of the fact that Malkos is pentatonic, omitting the Pa and the Re—the latter
being the distinguishing note of the thats Asavri and Bhairvi. Thus modern M alkos
could equally well have been classified in Asavri that if judged purely from the point
of view of scale.
Of the many other examples which could be discussed, we shall refer only briefly
to three other rdgas in Locana’s Karnata mela (modern Khamaj that, No. A3),
Vagisvari (modern Bagesvari), Sughrai and A4ana. The first two rags now have
Gab and are classified in K afi that (No. A4), while A<jdna has evolved even further
and is now classified in Asavri that (No. A5).
While the majority of the rdgas of Locana’s period can be explained in evolutionary
terms in relation to the circle of thats, a few prominent rdgas appear to have under
i H.S.P. HI, p. 250.
100
gone sudden and extreme change which is not accounted for in this gradual evolution
theory. It will be useful to consider two such instances.
The raga Hindola is given in Karnata mela (modern Khamaj that) by Locana.
About two centuries later, in the Sahgitapdrijata, it is given as a pentatonic raga
(Ex. 52a). This is still the South Indian version of this raga. In N orth India, however,
Hindol, while still remaining pentatonic, now uses all the opposite alternative notes
(Ex 52b):
Ex. 52. rag Hindol
(a) (b)
Sa G at Ma D h a I* N i t Sa Sa Ga M a# D h a Ni Sa
y ----- ■ ------;------ 1
----- „— u* — K , ---- — :— --- ----------------------- "■■■..... U
---------------------------
— «------------------------------
There is no certainty that the modern N orth Indian tradition is connected with that
described in the Sahgitapdrijata. The treatises give no evidence of gradual evolution,
and yet the fact that both the ancient raga and the m odem rag are pentatonic and
omit the Re and the Pa suggests that they are connected. Perhaps, in this instance, we
are dealing with a different phenomenon, something in the nature of a semitonal shift
o f the ground note, which could, in one move, create such a vastly different scale.
Similarly, the raga Sri is generally described in N orth Indian treatises (Locana,
Ahobala and Pundarlka in Rdgamanjari) as in present-day Khamaj that, while the
South Indian writers (Ramamatya and Somanatha) describe it as in m odern N orth
Indian K afi that. The scale of the raga $ri in South India remains unchanged (Ex.
53a); in N orth India, however, the rag now belongs to Purvi that, and once again, all
the alternative notes are reversed (Ex. 53b).
Ex. 53. rag £ri

(a) (b)
Sa Re G at Ma Pa Dim N it Sa Sa R et Ga Matt Pa D hat Ni Sa
-0---------- —--------------------------------------------— t ——— m--------
-y
ffm
— -------o — -------—--------------
- f r " ------------------------------------------------ u
' e f wr- *+----*---------------------------- ...... *
Bhatkhande suggests, however, that the modern N orth Indian rag £ri may have
been descended from the raga Srigauri of Locana’s Gauri mela (modern Bhairav
that).1 If this is acceptable, Sri can readily be explained in terms of gradual evolution
as Bhairav and Purvi thats are adjacent in our scheme.
In the previous pages we have been presenting the evidence to show that rags have
tended to evolve around the Circle of Thdts. The introduction of Bhairav that was,
from the standpoint of the Circle, premature, but has not detracted from the Circle;
it appears to have provided a new point of departure which may, in fact, have
hastened the completion of the Circle. This analysis should be considered as a tenta
tive beginning, rather than accomplished fact, for there still remain many unknowns.
1 H.S.P. III, p. 62.
101
VI
Alternative Notes
I n the previous chapters we have been discussing Indian musical scales and attempt-
mg to connect them in an evolutionary sequence. A discussion of rags purely on the
basis of scales alone cannot obviously be comprehensive; and a case has been made
out to minimise the importance of scale as a basis for the classification of rags.1
However, in this work we have tried to show that the notes used are much more than
a means for classification, and are an im portant functional element in the evolution
of rags. While it is not surprising that fashions change and new rags replace old ones,
it is remarkable that the scales of many rags should have been modified over the years
apparently without conscious appreciation of this fact. In this chapter we propose to
consider why these changes should have come about and how it is that they could
pass unnoticed.
There can be no denying that tradition is a powerful conservative factor in Indian
music. The inadequacy of the traditional methods of notation, coupled with the
intangible structure of rags, has tended to accentuate the teacher-pupil relationship.
This conservative element is, however, more than offset by the creative aspect of
improvisation. The pupil can only become a master when he can go beyond what he
has been taught and create his own music. The rag is the traditional restraining
element within which the creative musician must find the freedom to express himself.
In his search for freedom he frequently stretches the bounds of the rag and, as the
limits of mgs are never exactly prescribed, we can say they are being re-defined, to
some extent at least, in every performance. Under these circumstances change
becomes inevitable. There is, of course, a point of endurance in any particular
context which the musician may not exceed without violating the traditional basis
of the rag; but this does vary, depending on the sensitivity and skill of the performer
as well as on the discrimination and receptivity of the audience.
A second conservative influence, which might be expected to retard the rate of
evolution, the textual definition (which can be consulted in moments of doubt), has
1 It has been argued, for instance, that classification according to melodic features—omissions,
turns, etc.—may provide a more meaningful basis; but these features can, in any case, often be
related to scale.
102
Alternative Notes
not been of great importance until this century in N orth India, for many of the
great musicians were illiterate, while others had only been able to consult where
and when available the few treatises which had been written in the m odern Indian
languages.
A characteristic feature of the change we are discussing is that the process is
imperceptible. We can see the change in the scale of a rag only by looking back to the
description of that rag as it was hundreds of years ago. We may presume, however,
that change is not generally noticeable during a musician’s lifetime, for if this were
so, tradition would assert itself, ensuring that the original rag is preserved and the
changed version either discarded or given an independent status. Thus the process
by which the scale evolves is gradual, not a sudden replacement of one note by its
alternative as it appears on paper. This is accomplished by an intermediate phase in
which both the note and its alternative occur in the rag, not as chromatic steps,
but each in its own melodic context, the one temporarily replacing the other.
W ith both alternatives of a note in a rag, the shift of emphasis from one to the
other can be both gradual and imperceptible. Conversely we can say that the
presence of both alternatives of a note in a rag is a manifestation of the process of
evolution.
The prime essential in this gradual course of evolution is that the whole process
must flow smoothly and without any discontinuity. This being so, we m ust first
attem pt to explain how the initial introduction of the accidental can be accomplished
with such subtlety that no change is apparent. In Chapter IV we discussed certain
concepts of dual tonality initiated by the Sa and Pa drones and the related perception
o f intervals in terms of tetrachords. We indicated that the unbalanced tetrachord
types led to the replacement of one or other of the unbalanced notes, thus producing
new scales. This same principle provides the functional raison d ’etre for the introduc
tion of an accidental to produce temporary balance in the tetrachords. This accidental
provides an interval which is by no means foreign to the rag for it already occurs in
the tonality initiated by one of the two drones. For instance, in Khamaj, where the
lower of two ascending disjunct tetrachords has a major third, Sa-G a (I-III), and the
upper a minor third, Pa-Nib (V-VIIb), the experience of both the m ajor third and
m inor third is familiar. The introduction of the minor third, Gab (IHb), into the lower
tetrachord or the m ajor third, NU (VII b), into the upper merely extends the experience
already inherent in Khamaj. Initially, this impulse to introduce the accidental may
manifest itself as a wavering or an exaggerated vibrato on the nearest diatonic note.
This would probably be interpreted as a form of expression on the part of the musi
cian rather than as a separate musical interval and it would only be after a period of
acclimatisation that the accidental would gradually emerge as a grace note {lean
svar). This ‘wavering’ can even now be heard in a number of rags. In the Kanhrd
rags, Durban, for example, it has become a characteristic and obvious feature of the
rag (see p. 162). On a more subtle level, it often occurs in certain rags, for example
the rag Kedar, where some musicians consciously use the Nib (VIIb) as an accidental,
103
Alternative Notes
while others merely put a slight inflexion on the D ha (VI).1 From such a modest
beginning the accidental may gradually become m ore prominent at the expense of its
chromatic2 counterpart (the scalar note), until finally the accidental becomes the
more im portant and can be called the scalar note. This process of absorption of the
accidental is clearly associated with either the ascending or descending line; if the
accidental is lower than its chromatic counterpart it will be, initially, an ornament of
the preceding (lower) note of the scale, which is its nearest diatonic neighbour, and
will be introduced first into the descending line of the rag. We may here remind the
reader that the criterion of both descent and ascent is the position of the succeeding
note: if it is higher, then the note is ascending; if lower, the note is descending. In the
instance we are considering the accidental will occur as an ornament attached to the
lower note, i.e. the ornament will begin on the lower note, rise to the accidental and
return to the lower note. It is this descending return which is the determining factor.
Similarly, if the accidental is higher than its chromatic counterpart, it will be asso
ciated with the following (higher) note of the scale, and will thus be introduced into
the ascending fine. Thus, if the accidental is Gab (IIIb) it will appear as an ornament
attached to the Re (II) and if Gan (Illn) will be attached to the Ma.
Ex. 54.
* = T , M
tS W - " -jj. * i 'I
We shall now trace, in hypothesis, the course of the Gab (IIIb) from its first intro
duction as a grace note to its final replacement of the Gan (IIIt?) as a scalar note. It is
convenient to discuss the course of evolution in terms of stages, but it should be
understood that there is no precise line of demarcation between these stages, and
that they flow smoothly one into another. The first stage can be expressed concisely
as follows:
(1) Ascent Gan Descent Gan (+ Gab)
The Gab in brackets here indicates that it forms part of an oblique descending line,
while the Gan continues to be the direct descending note.3
The next evolutionary stage can be said to be reached when the accidental Gab
forms a descending line of its own. This may, for instance, be accomplished gradually
through an intermediate stage in the following way:
1 A particularly good example of this can be found on the record of Kedar (Raag Kidara), CLP
1514, sung by Roshanara Begam. About 1 min. 20 secs, from the beginning, the Dha (VI) is
sustained, first as a steady note, then there are two slight inflexions suggesting the Nib (VHb). This
can also be heard in rag Kedar on the accompanying record.
2 Chromatic is not used here in the usual Western sense—see p. 48.
3 The terms ‘oblique’ and ‘direct’ are defined on pp. 40, 41.
104
Alternative Notes
Ex. 55.
(a)
Ma Ga Re Gab Re Sa
-* -0
(b)
Ma Re G ab Re Sa
• •
(c)
Ma Gab Re Sa
It frequently happens, however, that the original descending line containing the
Gan will continue to exist alongside the new descending line, and there will be two
alternative forms of descent. At first, the new descending line may be used only occa
sionally and the traditional method of descent will be dominant. In the course of
time, however, the new descending line with the Gab may supersede it. This second
stage can be expressed as follows:
(2) Ascent Gan Descent Gan + Gab
W ith the predominance of the descending line containing the Gab, the importance
of the original descending line diminishes until, finally, the Gan may no longer appear
in a full descending line but, as a vestigial reminder of the original descending line,
remain as a discontinuous descending note, as in the example below :
Ex. 56.
Ma Ga Re Ga Ma Gab R e Sa
' • m * m —o ^ ^
This third stage can be expressed as follows:

(3) Ascent Gan Descent Gab (+ Gan)
Next, this vestigial figure in the descending line may become progressively shorter
until it is completely dropped and onlythe Gab occurs in the descending line. A t this
point the rag is exactly half-way between two thats, for the ascending line contains
the Gan and the descending line the Gab:
(4) Ascent Gan Descent Gab
The Gab would now begin to cast its influence on the ascending line. In the next
stage it may occur, at first, as a small discontinuous ascending figure which would
gradually become longer and more im portant without actually forming a full ascend
ing line:
105
Alternative Notes
Ex. 57.
Re G at M a GaJj Ma . . •
4 ^ . ~m m
This stage is exactly parallel to stage 3 and if we were considering the evolutionary
steps of the Gan accidental instead of the G at accidental, this would indeed have been
stage 3 in which the G at is only a discontinuous ascending note and a vestigial
reminder of the original ascending line. To return to our original scheme, this stage
can be expressed a s :
(5) Ascent Gan (+ Gab) Descent Gab
The Gab would next form an ascending line which appears as an alternative to the
original ascending line with the Gan. This new ascending line may gradually gain
prominence at the expense of the original ascending line. This stage can be expressed
as:
(6) Ascent Gan -f- Gab Descent Gab
The original ascending line which contained the Gan may then, as it gradually falls
into disuse, become oblique and the Gan remain merely as an ornament around the
M a:
(7) Ascent Gab (-f Gan) Descent Gab
The change of scale would be complete when, finally, the Gan is no longer used in
the rag.
These would be the evolutionary steps of a rag as a note of its scale changes from
Gan to Gab. If the change were from Gab to Gan, the steps would have been exactly
reversed. The changes would first occur in the ascending line until the half-way point,
stage 4, where the ascending line contains the Gan and the descending line the Gab;
thereafter, the changes would take place in the descending line until finally there
remained no sign of the Gab.
We must underline that this is only a hypothetical scheme and need not apply in any
particular instance. We have been concerned with the kind of process by which the
scale of a rag may, without intent, change over a period of time. In the course of its
evolution a rag may retrogress or find a considerable degree of stability and remain
in a particular stage, perhaps for several hundred years. On the other hand, it may go
through a period of relatively rapid change within a few generations, and may possibly
omit some of these intermediate stages. The uneven tempo of this change is reflected
in the diverse renderings of the same rag in the different geographical regions of
N orth India, and explains the differences in interpretation in the various musical
traditions (,gharana), some of which have been more conservative than others. There
has, however, been a considerable am ount of interchange between these traditions
and consequently their interpretations of the details are, in m ost cases, not so widely
divergent as one might expect in view of the size of the country and the difficulties of
communication which have existed until the modern period.
106
Alternative Notes
There are many rags which do not appear to have changed in scale, at least since
Locana’s Rdgatarahgini. Does this mean that certain rags are less subject to evolu
tionary forces than others? This is not necessarily so. It is probable that they have
also evolved, but their new forms have at some stage received recognition. The new
forms have here not replaced the old, but have been granted independent status and
been given their own names. The names of such rags often give an indication of
their origin. Thus, Alhaiya Bildval, Suld Bildval and Devgiri Bildval all suggest that
they have evolved out of Bildval. Sometimes they are differentiated from the parent
rag in the notes of emphasis, or in the ascending and descending lines in which the
accidentals may have reached a different stage of evolution. The evolutionary process
has, in some instances, progressed so far that the derivatives can no longer be classi
fied in the same scale.
It seems probable that the pattern of evolution may be very much the same whether
a rag evolves from one scale to another, or other rags evolve from it. In the former
instance there remains no memory o f the original state of the rag and no record is
left of the various evolutionary stages except those which we can glean from the
treatises of the different periods. These texts, however, are generally not detailed
enough to permit us to establish positively the various subtle differences between one
evolutionary stage and the next. In the latter instance the original rag continues to
exist alongside the new rags which are left as milestones in the path of evolution.
Consequently we should be able to find evidence for the various stages of evolution
among the rags which exist at the present time.
We shall now consider the rags of Bildval (No. A2) and Khamaj thats (No. A3) in
order to show that there are, in fact, rags which corroborate most of the seven inter
mediate evolutionary stages between these two thats. The following examples are
taken from Bhatkhande’s notations.
1. Beginning from Bildval that, the first stage is the introduction of the Nib (VIIb),
initially as an ornament around the Dha, later used more freely but always remaining
an oblique descending note. The rag Alhaiya Bildval is an excellent illustration of
this stage of development. In Bhatkhande’s tradition the Nih is found in both the
direct ascending and descending lines, while the Nib occurs only in an oblique
descending line marked x : 1
Ex. 58. rag Alhaiya Bildval
Sa, R e, Ga R e, Ga P a, D ha, N i Dha, N i Sa
i-i- j -J-j' Jj J 1 rJ
Sa N i D ha, P a , Dha Nib D ha, P a , M a G a, Ma R e, Sa
- i r- , — M = l — — .---- -------- j—
[ _ a ... J .
L .V— t
1 K.P.M . II, p. 75.
107
Alternative Notes
2. In the second stage the descending line, with the Nib, is completed and there are
now two concurrent descending lines as seen, for instance, in the rag Sukl Bilaval:1
Ex. 59. rag Sukl Bilaval

(a)
Sa, Rc Sa Ni D ha, Ni Dha Pa, Ma, Pa Ma Ga, Ma, Re Re Sa
(b)
Sa, Nib D ha, Nib Dha Dha Ma Ga, Ma R e, Sa
•
4-r -F--- ^ J J J J|J i
3. In the third stage the descending line with the Nib has become dominant and the
original descending line containing the Nib is now discontinuous, remaining as a
vestigial reminder of the original. This is apparent in the rag Bihagrd where Nib is
freely used in the descending line. The Nib is occasionally used in a discontinuous
descent, but is followed inevitably by a descent in which the Nib is used.
This will be evident from the following examples :2
Ex. 60. rag Bihdgfa
(a)
Pa Pa NL Sa Re Sa Ni Dha N ib D h a P a ,. . .
4 - r - r r r L_E__C_r
(b)
. . . . Pa Ni Sa Re Sa Ni D ha P a, N ib D ha Pa..............
: [_ !_ £ -/ -j-= j
4. In the fourth stage the Nib no longer appears in the descending line. The rag Des,
as it is usually performed, illustrates this stage. The Nib occurs in the ascent, while the
Nib is used in the descent:3
Ex. 61. rag Des

Sa, R e, Ma Pa, Ni Sa Sa N ib D ha pa, Ma Ga, Re G a, Sa
1 (a) K.P.M. V, p. 480, variation 3. (b) Ibid., p. 153, penultimate line.

2 Ibid., p. 486, (a) variation 9, (b) variation 7.
3 K.P.M. III, p. 251.
108
Alternative Notes
5. In the next stage the Nib is used as a discontinuous ascending note. This is some
times done in the rag Des where the phrase (a) may sometimes be followed by (b):1
Ex. 62. rag Des
(a) (b)
R e M a Pa D h a N ib D h a Pa D lia M a G a R e. R e M a Pa D ha N ib Sa N ib D ha Pa D ha M a G a Re
=---------
A : ;r~ — --------
■ ~ * • ---#---* --- -------* --- J --- * ....
J” *
6. In the sixth stage the Nib forms a separate alternative ascending line. The rag
Khamaj is an example of this stage of development. The most commonly heard
ascent has Nib (a), but occasionally the Nib is used (b):
Ex. 63. rag Khamaj
(a) (b)
Sa, Ga M a Pa Dha N i Sa Sa, Ga M a Dha N ib Sa
Bhatkhande gives the latter2 but prefers to use the Nib in his svarvistar of the rag,3
There is also corroboration of the ascending use of the Nib in Bhatkhande’s notations
of songs, where this occasionally occurs.4 The Nib as an ascending note is, however,
rarely used, but this has an independent explanation which we shall consider later
in this chapter.
7. In the seventh stage the Nib is used only as an incomplete ascending note, as in
the following phrase:
Ex, 64,
Dha N ib Sa N ib Sa
I ................
In fact, there is no rag in either Bildval or Khamaj that in which this occurs. Once
again the explanation for this will appear later. This stage can, however, be illustrated
1 K.P.M. III, p. 761, variation 10. It will be noticed in the phrase above that we have called the
Nib a discontinuous ascending note in spite of the fact that the phrase proceeds to the upper Sa.
An essential requirement of a direct ascending note, in our interpretation, is that the movement
above should not be restrictive and normally, a note which permits access to the Sa would not be
so. In this exceptional instance, however, the ascent to the Sa following a Nib does not permit free
access into the upper register but is invariably followed by a descent. The explanation probably
lies in the fact that the leading note, Nib, in Des paves the way for the Sa, but the Nib does not
convey the same sense of anticipation for the ground note and the Sa is then treated as an appen
dage of the Nib. This rag can be heard on the accompanying record and is discussed further in
Appendix B.
2 K P .M . II, p. 122.
3 K.P.M, II, p. 490.
4 For instance in the song, K P .M . II, p. 124.
109
Alternative Notes
from a corresponding position between two other thats, for instance Bildval (No. A2)
and Kalyan (No. A l), in which the two alternatives of M a are involved. Here we
would need a rag in which the Mas (IV?) occurs as an incomplete ascending note,
while the Mab (IVb) forms the complete ascending line. This stage is illustrated by the
rag Hamir, in which the Ma# appears only as an incomplete ascending note, as in the
phrase x below, while the ascending line has the M ab:1
Ex. 65. rag Hamir

Sa ( R e S a ), Ga M a D h a, Ni D ha, Sa Sa N i D ha Pa, MaftPa D h a P a , G a Mab R e Sa
(• J-')
These are then the seven principal stages between two adjacent thats. They can be
seen more clearly in the following diagrammatic form:
BILAVAL KHAMAJ
THAT (No.A2) THAT (No.A3)
1
ASCENT Ni*1 Ni11 Ni*1 Nir Ni*1 Nii V N i b) Ni*1+ Nib Ni^+Ni*1) Nib
DESCENT Ni11 NiHt+Nib> NiH+Nib Nb(+Ni^) Nib Ni* NT Ni* NT
< K w ac i c_ ic-
J d
> i< 74 ifla ®
K o 2< > -o
k x
^ <D ± j -*~
x %£%-g
OP c a
z A .2 o
< c~ >
X
'■ZD
(/) t [/) ID
O o •■=
m c l co
Although these seven stages should apply between each that and its neighbour, it is
not always possible to find rags to corroborate all these intermediate stages. In Tori
that, for example, there are only three commonly known rags,2 and none of these
has any accidentals, while in Bhairvi that, in which once again there are very few
rags, the rag Bhairvi permits a great deal of freedom, not only in the way an accidental
is used (see p. 120) but also in the number of accidentals permitted.
In the previous chapter we showed that Locana’s Bhairavi was similar to our
modern K dfi that (No. A4), whereas it now includes, in addition to the Gab (Illb)
and Nib (Vllb) of Kdfi, Reb (lib) and Dhab (VIb). The two most prominent accidentals
1 K P.M . m , p . 6 8 .
2 Bhatkhande gives seven rags, but four of these, Lacarl Tori, Laksmi Tori, Bahaduri Tori and
Anjni Tori, are seldom heard, and with the exception of Bahaduri Tori do not really belong in the
modern Tori that since they have Mab as a scalar note.
110
Alternative Notes
in the rag Bhairvi are, however, the Ren (II b) and the Dhab (VIb), remaining as if they
were a memory of its original form.
This hypothetical scheme may be of value in the clarification o f the evolutionary
process; it does not, however, help us to determine whether a particular rag in one
of these evolutionary stages has evolved from its neighbouring that in a clockwise
or an anti-clockwise direction round the Circle of thats, as both are equally possible.
In most instances there does not appear to be any musical criterion for determining
the that of origin and the only clear evidence sometimes occurs in the name of the
rag; thus we can be fairly certain, as we have suggested earlier, that Sukl Bildval has
originated from the rag Bildval. N or does this scheme automatically resolve the
problems in classifying rags in terms of thats. The obvious difficulty lies in determining
the parent that of a rag which is exactly half-way between two thats and would be
classified in stage 4 of our evolutionary scheme. This may perhaps be resolved on the
grounds that the descending fine, on which the perfect cadence is generally con
structed, is the more important. A much more serious difficulty is caused by the fact
that many rags have more than one accidental.
We have, so far, been speaking of only one kind of accidental, that which grows
out of the unbalanced tetrachords of a scale. This type of accidental can be called a
first order balance note, as it creates a temporary balance in the scale. In Bildval that
for instance, where the descending conjunct tetrachords are unbalanced, the intro
duction of either the Nib (VIIb) or the Ma# (IV#) will temporarily achieve this end.
As the Nib occurs initially in the descending line and the Ma# in the ascending line,
these are not mutually exclusive and there are several rags in which both of these
accidentals are used. This complicates our simple evolutionary pattern for it can be
said that these rags are evolving in both directions at the same time: the Nib
leading towards Khamaj that, the Ma# towards Kalyan.
The introduction of the first order balance note may create temporary balance in
the descending conjunct tetrachords of Bildval, but at the same time, it disturbs the
balance in the ascending disjunct tetrachords. This can manifest itself in the intro
duction of a second order balance note which would temporarily restore the balance
in the disjunct tetrachords which has been disturbed by the first accidental. For
instance in the rag Bhairvi (No. A6) the Reb (lib) may occur as a first order balance
note which restores the balance in the descending conjunct tetrachords, §a-P a
(1-V) and Pa-R e (V-II). However, it destroys the balance in the ascending disjunct
tetrachord, Sa-M a (I-IV) and Pa-Sa (V-I). This might then lead to the introduction
of the second order balance note, Dhab (VIb), which will then temporarily create
ascending disjunct balance. This process could, of course, be extended further to
include third, fourth and fifth order balance notes. There are, however, at least two
first order balance notes in every scale and from each stems a second order, so that
four of the five possible accidentals can be obtained without proceeding beyond the
second order. The remaining accidental which belongs to the third order does not
appear to be significant in N orth Indian music. In Bhairvi that, for instance, the first
111
Alternative Notes
order accidentals would be Ren (IIn) and Mas or rather Pab (IVs or Vb); the second
order accidentals, Dhan (VI b) and Nib or rather Sab (VII b or lb); the third order
accidental, Gan (Illb). In the rag Bhairvi it is commonly understood that all the five
possible accidentals may be used, but in fact the third order accidental, the Gan, is
very rarely used.
Second order accidentals too are relatively infrequently used. This is not an
easy m atter to establish for it is dependent on the determination of the that to
which a particular rag belongs. The rag Keddr, for example, is ascribed by Bhatkhande
to Kalyan that. In this rag the two alternatives of M a (IV) and N i (VII) are used. If
we accept Bhatkhande’s classification of this rag in Kalyas that, then the Man and
Nib are accidentals for these are chromatic alterations of the parent scale Kalyan
(No. A l). The Man will be a first order accidental, the Nib a second order. Earlier1
we had suggested that there was little justification for classifying the rag Keddr in
Kalyan that, in view of the fact that Bhatkhande himself gives Man as the vddi
(important note) of this rag. If this rag is classified in Bildval that„ the accidentals
would be Ma# and Nib, both of which would then be first order balance notes.
The number of first order accidentals depends on the number of unbalanced
intervals in the two tetrachords, in other words, the number o f imperfect fourths
and fifths in the scale concerned. In Bhairav that all the five possible alternatives can
be achieved as first order accidentals since the descending conjunct tetrachords are
completely unbalanced. In Marvd, Purvi and Tori thats there can be four possible
first order accidentals (three in Marvd since Ren occurs twice). These first order acci
dentals are shown in square brackets below:
Ex. 66.
(a) Marva that
Ni Sa R et Rd} Ga Mab M ai Pa Dha Ni Sa
I i
Mail Pa Dhab Dha Ni Sa Maft Pa
$
(b) Purvi fhat
Ni Sa R eb Mail Pa Dhab Dhaij ‘ Ni Sa
. . - l i
■■ - §
I 1
Mafi Pa Dhab Ni Sa Reb Re[] Gab Ga Mail Pa
4
—• —
Lb* J
1 See p. 53.
112
Alternative Notes
(c) Tori that
Ni Sa .Reb- G at Go!} Mat] MaB Pa Dhab Ni Sa
MaB Pa D h at
i t t Ma# Pa
(d) Bhairav that

Sa R et Ga Ma Mat Pa Dhat Dhal; N it Ni Sa
V V
tt
Pa D hat Ni Sa Ret Ret] Gat Ga Ma Ma)} Pa
fr* j" *
Two of the accidentals in the three thats, Marva, Purvi and Tori, are constant, the
Reh and the Man. In many of the rags in these thats, as we have previously indicated,
the Pa is either very weak or completely omitted. Consequently, the Ren, which
would, in the that, balance the Pa, is, except in one instance, not actually found as an
accidental in the rags of these thats. On the other hand the Man which balances the
Sa is a very prominent accidental in the rags of these thats, for although there is a
tendency to omit the Sa in the phrases of many of these rags, the importance of the
fundamental note of the system cannot be entirely suppressed. For all practical
purposes, then, these thats have three first order balance notes. In Marva that the
Ren occurs twice as a first order balance note. There is, however, only one instance of
its occurrence as an accidental, as will be noticed in the table which is given later in
this chapter.
N ot all the accidentals used in the rags of N orth Indian classical music can be
accounted for in terms of balance. The other accidentals, with perhaps one or two
exceptions, can all be explained as leading notes. This concept, familiar to Western
musical theory, is based on the principle that the extreme dissonance of the major
seventh (Niti) demands resolution in the ground-note and conveys, in the resolution,
a considerable measure of finality. The Ni*i appears to be used in a number of rags
in just this manner, as an accidental in the ascending line (even when it is not an
immediate balance note) showing the way to the Sa. Of the ten principal thats only
four have Nib, and it is obviously in these four, Khamaj (No. A3), K dfi (No. A4),
Asavri (No. A5) and Bhairvi (No. A6) thats, that the leading note, Ni*, occurs as
an accidental. In Khamaj the Nib is also a first order balance note and it is not sur
prising that this note is found in all the heptatonic rags of this that. This explains why,
8 113
Alternative Notes
in our illustration of the seven intermediate stages between Bildval and Khamaj thats,
we were unable to give an example of stage 7 in which Nib would have been a direct
ascending note, with the Nit? only an oblique ascending note. In K dfi that the Nib
is a second order balance note which can only be significant if the first order acci
dental, Ga^ (Hit)), is used in the rag. In fact the Nib occurs much more often than the
Gab in the rags of this that, & fact which clearly indicates that the importance of the
Nib is not primarily as a balance note. From Bhatkhande’s notations it is apparent
that the use of this leading note in some of the rags in this that is n° t accepted in all
the musical traditions.1
The prominence of the leading note in Khamaj and K dfi thats may possibly suggest
a survival of the early musical system of India, in which the two parent scales, the
Sadjagrama (similar to modern K dfi that) and the Madhyamagrdma (similar to
Khamaj that), were each permitted an accidental. These accidentals, the kakali Ni
in the former and the antara Ga in the latter, were in fact leading notes in the parent
scales and could only be used in ascent.2 In application to Khamaj and K dfi thats, at
least, the use of the leading note is a long venerated tradition in India.
In Asavri and Bhairvi thats the leading-note accidental is much less common.
Sometimes, as in the rag Adana of Asavri that, the occurrence of Nib is a m atter of
controversy, and although Bhatkhande does not draw attention to this in his works
there are at present a number of musicians who do not use the Nib in this rag, but
show, nevertheless, a tendency to sharpen the Nib slightly in ascent.3 In Bhairvi that
the Nib is even less frequently used, though it occurs occasionally in the little-known
rag, M otki,4 and of course in the rag Bhairvi, where accidentals are used quite
freely.
While there would appear to be no doubt that the Nib occurs in a number of rags
as an accidental leading to the Sa, it should not be presumed that the Nib, where it
occurs as a scalar note, is invariably used in this way. In the thats on the left of the
Circle, Marva, Purvi, and Tori, which have Nib as a scalar note, the Reb (IIb) very
frequently serves as a leading note to the Sa. There are probably several explanations
for this. The Reb is perhaps even more dissonant than the Nib (depending, of course,
on the accompanying drones) and consequently we may expect the resolution in the
Sa to be even more satisfying. In addition, the final cadence of m ost rags is generally
descending, a fact which would tend to give prominence to Reb as a leading note.
Finally, in a number of rags in these thats, there is a tendency to omit the Sa, often in
both ascent and descent, as in the rag Marvd, where the Ni, in some traditions, does
not lead directly to the Sa, but either to the Reb or the D ha and thence to the S a:
1 For instance in the rag Bhimpldsi, some songs have Nib in ascent, e.g. K.P.M. Ill, p. 564, others
Nib, e.g. ibid., p. 573-4.
2 Ndtyasastra, ‘Kashi Sanskrit Series’ (No. 60), prose, p. 321.
3 Some musicians refer to the slightly raised ascending Nib as carhi or sakdri. This can be heard
on the accompanying record in the rags Darbarl and Suha and is discussed in Appendix B. It may
be this note which Bhatkhande has notated as Nib.
4 KP.M. VI, p. 438.
114
Alternative Notes
Ex. 67. rag M arva
D ha Ni Reb Sa R et N i D ha Sa
f t H *- V “
The balance-note accidentals are a manifestation of the same inherent instability

of musical scales which, as we have suggested earlier, is a fundamental factor in the
evolution of the musical system. On the other hand, leading-note accidentals do not
appear to have influenced the course of evolution to any great extent. There are,
however, two or three rags which may be exceptions. The rag Pilu, in which acciden
tals are freely permitted, is given in K dfi that (No. A4) by Bhatkhande. This rag has
a very prom inent leading note (Nib) which often serves as a terminal note. Perhaps
because of its importance in the rcig the Nib is sometimes also used in descent.1
Ex. 68. rag Pilu

Sa Ni Dha Pa or Sa Ni D hat Pa
Similarly, in one version of the rag Patdip (this tradition is not mentioned by Bhat
khande, but is frequently heard at the present time) the scale could be given as
follows:
Ex. 69. rag Patdip

Sa Re Gab Ma Dha
In both these instances the importance of the Nib, probably initially introduced
as a leading-note accidental in either K dfi or Asavri that, may finally have led to its
acceptance in the descending line as well.2
There are indications that the accidental Gab (III may be used as leading note
to the M a (IV) in one or two rags. The rag Devgdndhar is a case in point. Bhatkhande
refers to two versions of this rag, both in Asavri that (No. A5). In one of these the
Gab occurs as the only accidental, while the other has no accidentals. The scalar
Gab is, however, perfectly balanced, having both a perfect fourth, Dhab, and a
perfect fifth, Nib. In the normal ascending line of the former the G a is generally
omitted (Ex. 70a); the Gab only occurs in the phrase leading to the Ma, which is then
generally sustained (Ex. 70b)3:
1 The scale of Pilu, given by Grosset, is in fact our No. C5 with Gab (Illb) and Dhab (VIb), but
with Nib, as we have indicated.
2 The rag Candrkos, discussed on p. 136, may be a third instance.
3 K P.M . IV, p. 383.
115
Alternative Notes
Ex. 70. rag Devgandhar
(a) (b)
Sa Re Ma Pa . . . . Sa Re Gaij, Ma
Similarly, the Gan as a leading note also occurs in one version of the rag Patdipki
(Pradipki) as given by Bhatkhande. This rag is in K dfi that (No. A4) and the use o f the
Gan could be explained as a balance for the Dhab. It is used also, as in the rag
Devgandhar, as a leading note to the Ma. Here the Gan occurs as an incomplete
ascending note, while the Gab occurs in the complete descending line as in the
following phrases:1
Ex. 71. rag Devgandhar

NiU Sa, Ma, Gall Ma, Pa M a, D ha Pa, Ma, Pa Gafi, M a, Nib D ha Pa, Sa, Nib Dha Pa,
Dha Pa, Gafj, M a, Pa Gab, Re, S a,
■t|J .,-■1
The use of the Gan as a leading note occurs in a very limited number of rags, and is
only possible in those rags in which the M a may be used as a terminal note. In the
same way the Ma# (IV#) may also be used as a leading note to the Pa (V) in several
rags. There are, however, no instances of the use of the Ma# in a rag where it is not
explicable as a balance note.
It is quite clear that the accidentals used in the rags of N orth Indian classical music
are not fortuitous, but can be explained on musical grounds, either as balance notes
or as leading notes. Further, the consistent occurrence of the same accidental or
accidentals in many rags of the same that indicates that they are primarily influenced
by the scale o f the rag. The occurrence o f these accidentals can be conveniently
summarised in statistical form. This summary is based on Bhatkhande5s notations as
found in both his works, K P .M . and H.S.P. For the purpose of these statistics we
have only taken into consideration the heptatonic rags, except in Tori, Purvi and
Marvd thats, in which hexatonic rags are basic. In the case of divergent traditions of
the same rag, we have chosen that tradition in which one or more accidentals are
used. These are not necessarily traditions which are preferred by Bhatkhande, nor
are they necessarily those which are commonly heard today. It was our concern,
1 K P .M . VI, p. 479, variation 3.
116
Alternative Notes
however, to show that even among the lesser-known traditions the use of accidentals
appears to be governed by the musical principles indicated.
heptatonic rags
Second-order
anticlockwise
Second-order
anticlockwise
Number of
First-order
First-order
clockwise
clockwise
Observations
Other
Kalya# Mab Reb Nib Dhab Gab The rags with two accidentals, Mat)
and Nib, first and second order
N o accidentals 2 clockwise respectively, would be more
reasonably considered in Bildval fhaf,
One accidental 3 X — — — — where the accidentals would then be
Nib and Ma#, first-order clockwise
Two accidentals 6 X X and anticlockwise respectively—see
discussion of rag Keddr, p. 53. The
Total No. of rags 11 only example of a rag with Reb is
given in Marva that, where the
accidental occurs as Ret).
Bildval Nib Ma# Gab Reb Dhab
N o accidentals 2
One accidental 14 X — — — —
One accidental 3 — X — — —
Two accidentals 3 X X —■ --- —

Total No. of ra^s 22
Khamaj Gab Nib Dhab Ma# Reb The Nib occurs as a leading note in
all the rags of this that. Consequently,
No accidentals 0 there are neither rags with no acci
dentals, nor rags with just Gab
One accidental 5 — X — — — accidental.
Two accidentals 4 X X ■
-- --- ---
Total No. of rags 9

Kdfi Dhab Gab Reb Nib Ma# In these rags too the prominence of
the leading note, Nib, is very marked.
No accidentals 0 ~~ — — — — Some of the rags of Kdfi fhaf are
commonly heard without the Nib,
One accidental 16 — — — X ■ but in all of these there are diver
—
gent traditions found among Bhat-

Two accidentals 8 — X — X — khapde’s notations of songs in which
the Nib is used as a leading note.
Three accidentals 3 X X — X — The rag with all accidentals permitted
is the rag Pilu,
Five accidentals 1 X X X X X
Total No. of rags 28

117
Alternative Notes
Observations
Asavri Reb Dhat* Ma# Gab Nib The leading note is not quite so
prominent in this that. The rag with
No accidentals 3 — Gab accidental is the rag Dev-
gdndhar in which it occurs as a
One accidental 1 leading note to the Ma (see text,
p. 116). The importance of the first
One accidental 1 order balance notes in Asavri and
Khamaj thats can be appreciated if
Two accidentals 1 X — X the leading-note accidentals are
ignored.
Two accidentals 1 X — X
Two accidentals 1 X X —
Four accidentals 3 X X — X X
Bhairvi Ma# Reb Nib Dhab Gab The rag with four accidentals is the
rag Bhairvi in which all accidentals
No accidentals — are said to be permitted. In Bhat-
khanfle’s notations, however, there is
Two accidentals X X — no indication of Gab.
Four accidentals X X X X
Total No. of rags
All First-Order Balance
Bhairav Reb Gab Ma# Dhab Nib N o clear pattern emerges from the
use of accidentals in the rags of
No accidentals Bhairav thdf. The Nib and the Ma#
occur most frequently. The latter
One accidental — — X — — indicates the connection with Purvi
that. The absence of the Reb as an
One accidental 2 — X — accidental may seem surprising in
view of the fact that it is a balance
One accidental _ _ _ _ _ _ X for the Pa. However, this could be
appreciated only by a comparison
Two accidentals „ _ x — X of conjunct tetrachords (since Re-Pa
is a fourth) which are in Bhairav
Two accidentals X X completely unbalanced. Evidence for
the influence of Reb may be seen in
Two accidentals — X X the rag Nat Bhairav (No. C2).
118
Alternative Notes
c/1
■8t O
« irt «/1>
o O *8 T<D c
»-3<*5
>
CD o S3
& fl ■oa o £ ■g^ o o TaZ) ■S* *d
tH
Observations
J3 o Rj+->
ft 0
g ta t/3 O o«•§
.y <i>
.2 ~
9< .So ■3 fl 8'^ O 4-J
55 aRj
r i +toi
i j CO
M h O P-l Rj co T3 CO o<S o£
To/7 Gait Nib Dhab(Mab) Reb Mab In Top, Purvi and Marva
thats, hexatonic rags have
No accidentals 3 — — — — — — been included in the tables.
Several other varieties of Top
are mentioned by Bhatkhande-
He does not, however, des
cribe them in any detail, as
they are very rarely heard and,
Total No. of rags 3 from the scalar point of view,
do not belong in Top that.
Purvi Dhab Gab (Reb) Nib Reb Mab The first - order clockwise
(Dhab) and anticlockwise
No accidentals 7 — — — — — — (Gab) do not appear to be
used as accidentals. This is,
One accidental 5 — — — — — X indeed, the case with respect
to the Gab. There are, how-
Total No. of rags 12 ever, several rags which could
be indicated in Purvi that with
Dhab as accidental, but these
are given in Marva that, with
Dhab as accidental, by
Bhatkhande.
Marva Reb Dhab (Mab) Gab (Reb) Mab The Reb is the connecting link
with Kalya/.i, and thus with
N o accidentals 4 — — ■
— — — — the serial fhafs. This link is
undoubtedly weak and the
One accidental 3 — X — — — — Reb occurs in only one version
of the rag Jet. Nib, not given
One accidental 4 — — — —- — X i n this table, is neither first-
nor second-order, and is not
Two accidentals I X X — — — — used as an accidental in any
rag of this that.
Two accidentals 3 — X — — — X
These tables show that the first order balance notes are of prime importance in the
system, and that leading-note accidentals have considerable effect in some of the
thats. Second-order balance notes can, to some extent, be eliminated by the re-classi-
fication of certain rags. There are, however, a few rags in which these would still be
significant. In these, for instance Bhairvi and Pilu, the accidentals are used more as
a temporary change of scale rather than as ascending or descending melodic features.
This temporary change of scale is characteristic of the thumri, a form of song in which
119
Alternative Notes
the use of accidentals is considered desirable provided the transition from one scale
to the next is accomplished smoothly. This is another instance of the re-creation of
the evolutionary pattern which also depends on the smooth transition from one scale
to the next. Thus, in the rag Bhairvi, the accidental Reti is not introduced merely
into the ascending line, but may completely replace the R et until the final cadence
at the end of a variation as in the following example:
Ex. 72. rag Bhairvi
Pa N it sa Re!) G at G at, Rdj Sa ‘ D hat, N it Sa Re!] G at Re1!) l<Iit p a
¥ a G at, pa N it Sa Relj G at Rctf Gat Sa R et N it Sa
It is primarily in those rags in which the first order balance accidental may be used
so extensively, that the second order accidental becomes significant. In Bhairvi, the
Dhat) can be used in a similar manner:
Ex. 73. rag Bhairvi
Pa Dhalj N it Dhat) N i t D hal|Pa ^ G a t , Ma P a Dhat) N it D h a ilN it Pa N i t D h at Pa
i£ _JS m m #=i
The following diagram is designed to provide a convenient reference for the order
of accidentals in the thats. The ten radial lines here represent the ten movable notes.
The thats are represented as segments of a circle, each crossing five of these lines
which are the constituent movable notes of that that. Thus this diagram shows the
scale of each that (not including Bhairav) when the immovable Sa and Pa are added.
The first radial beyond the segment representing the that is the first order clockwise
balance note. The following radial in the same direction is the second order clock
wise. Similarly, the two radials beyond the segment in the anticlockwise direction,
are the first and second order anticlockwise balance notes. In the diagram we have
also added two more radials in dotted lines to indicate that the thats crossed by these
lines are those affected by the leading note, N h ; and the first order balance note, Mat).
Although our prime purpose in this work is to further the understanding of Indian
120
Alternative Notes
music rather than to suggest the reform of musical theory, certain principles for the
classification of heptatonic and the ‘functional’ hexatonic rags of Marva, Purvi and
Ga
jAHiyHgJ
Tori thats become evident from our analysis. These can be expressed briefly as
follows:
1. For the purpose of classification leading notes may be ignored.
2. A rag in which one pair of alternatives is used should be classified according to
the way in which the alternatives are used with reference to the seven stages
between two thats. If the usage is such as to be exactly half-way between two
thats (stage 4), then the rag could be classified, perhaps somewhat arbitrarily,
according to its descending line.
3. If two pairs of alternatives (excluding the leading note) are used, the rag should
be classified in that that in which the alternatives occur as first-order balance
notes, irrespective of the way in which the alternatives are used.
4. If more than two alternatives are used, i.e. second-order balance notes are clearly
significant, then the rag should be classified according to the cadence or cadences.
121
VII
Transilient Scales
1 he Indian term for a heptatonic series, sampurri (lit. complete), suggests that hexa-
tonic and pentatonic series are incomplete, lacking either one or two notes. This
provides justification for the use of the term transilient (or gapped) scales for the
hexatonic and pentatonic series in Indian classical music. The heptatonic series has
been the basis of the musical system from a very early period.1 The grama-jati system
of the Natyasastra is, of course, also based on the heptatonic series and there are
instructions given for the creation of transilient forms by the omission of specific
notes from the heptatonic jatis.
In the Muslim period, however, the pentatonic series appears to have been given a
measure of importance in the raga-ragini schemes of classification which were then
in vogue. In these schemes the masculine rdgas were frequently pentatonic, with the
implication that the system was based on the pentatonic series. This was a period
when poetic imagination had free rein and rdgas were associated with the Hindu
deities, colours, stars and other natural and supernatural phenomena, culminating
in the raga-mala paintings in which the ragas and raginis are represented in their
visual symbolic form. There is no reason to suppose, however, that the association
of pentatonality with the masculine raga at this time reflects the real origins of art
music and it is much more probable that the connection between the two was based
on aesthetics.2
In the modern period the heptatonic series is thought of as the parent of the hexa
tonic and the pentatonic. Bhatkhande says that pentatonic, hexatonic and heptatonic
rags are all produced from thats, and consequently each that must necessarily have
seven notes.3 From one heptatonic series of notes Bhatkhande derives 484 possible
combinations of penta-, hexa- and heptatonic forms, his calculations being based on
the independence of the ascending and descending lines as follows:
1 Elk Pratisakhya, a Vedic text of about the 4th century B .C . dealing with phonetics, mentions
seven notes (yama) and three registers (sthana).
2 That this association of raga with pentatonality is largely subjective is emphasised by
Bhatkhande when he quotes an opinion that raga, being male, will have a serious character, be
sung in a leisurely and grand manner and be heptatonic. H.S.P. Ill, p. 36.
3 K.P.M. II, p. 14.
122
Transilient Scales
Num ber of
Ascent Descent possible forms
heptatonic heptatonic 1
heptatonic hexatonic 6
heptatonic pentatonic 15
hexatonic heptatonic 6
hexatonic hexatonic 36
hexatonic pentatonic 90
pentatonic heptatonic 15
pentatonic hexatonic 90
pentatonic pentatonic 225
Total 484
These combinations, each of which could provide the basis for at least one rag, are
all derived from just one heptatonic that. To impress the reader still further, Bhat
khande multiplies this number by the seventy-two me/s (melas) of the South Indian
system, arriving at a grand total of 34,848. But even this figure does not exhaust
all the possible rags, for it does not take into account the differentiation of one rag
from another on the basis of altering im portant notes, i.e. vddi and samvddi.1
Bhatkhande does not, however, place too much importance on these theoretical
possibilities for he recognises that in current practice there do not appear to be many
more than 200 rags in all.
Transilient rags could obviously have been derived from heptatonic scales by a
similar intellectual process, but in fact there is little reason to believe that this has
actually occurred, except perhaps in a few isolated instances. Whereas the intellect
can conceive hundreds of transilient scales, only a very small percentage is actually
in use; for instance, among Bhatkhande’s notations there are only twenty-seven rags
pentatonic in both ascent and descent. This enormous discrepancy between theory
and practice is explained by Bhatkhande on the grounds that rags must have the
capacity to give pleasure and, by implication, only a few of these combinations have
this capacity. Our experience with heptatonic scales indicates that the prominent
scales have evolved within the system in a logical subconscious process, and only a
few m odem rags have been borrowed or created intellectually. It would not be un
reasonable to expect the transilient scales to follow this same pattern and to be closely
related to the ten heptatonic thats.
The concept of balance has been seen to be a fundamental aspect of the heptatonic
thats and appears to be of equal importance in transilient scales. The inherent im
perfection o f heptatonic scales, where at least one pair of notes stands in an augmented
i K P.M . m , pp. 13 and 14.
123
Transilient Scales
fourth/diminished fifth relationship, precludes the possibility of producing a perfectly
balanced scale. This imperfection, we suggested in the previous chapter, was the prime
instigation for the introduction of accidentals which then supply temporary balance
in the scales. A second obvious solution is to omit one or both of the notes which
cause this imperfection. This has already been mentioned in our discussion of the rag
Marva (see pp. 82,83), where we suggested that the omission of the Pa was a necessary
device in order to create balance in the rag. The tendency to omit the unbalanced
notes in a heptatonic scale is readily demonstrated in those rags for which there are
few prescribed rules of movement, where musicians are relatively free to use virtually
any combinations of the notes and, in the process, to omit notes according to their
inclinations. Even in these rags certain conventional phrases have evolved and are
very commonly heard, while other theoretical possibilities are used much less fre
quently or not at all.
In the rag Bhairvi, for instance, the correlation between the first-order accidentals
and the commonly omitted notes is clearly apparent. The unbalanced intervals are
the Ret- (IIb) and the Pa (V), and the first order balance notes are the Re 13 (IIm) and the
Pab (Vb), called Mas (IVs). Both the scalar note Reb and the accidental Ren, may be
used in ascent, but a third ascending line in which the Re is omitted is equally as
prominent. All three of these ascending lines are found among Bhatkhande’s
notations:
Ex. 74. rag Bhairvi

(a)1 (b)2 (C )3
Sa R eb Gab Ma Sa Rclf G ab Ma Sa G ab Ma
Similarly, the Pa may either be included in the ascending line or omitted. The Pab,
or rather Ma# as it is called, occurs however only in the descending line:
Ex. 75. rag Bhairvi

(a)4 (b)5 (C)6
Gab M a Pa Dhab N ib Sa Ga Ma D hab N ib Sa N ib Dhab M a | Mail R elr G ab
In rag Bhairvi two characteristic symmetrical, gapped phrases are apparent. In the
first (Ex. 76a), Re is omitted, in the second (b), Pa. Some musicians also introduce a
1 K.P.M. II, p. 498, variation 10.
2 K.P.M. II, song, on p. 400.
3 K.P.M. H, p. 498, variation 3.
s K.P.M. II, p. 498, variation 6.
6 K.P.M. II, p. 427, song.
124
Transilient Scales
third gapped phrase (c) which provides disjunct symmetry to (a) in spite of the fact
that this tends to place emphasis on Ren which is really an accidental in Bhairvi that:
Ex. 76. rag Bhairvi (a)
(a) (b) (C)
N it Sa G at Ma Pa G a t M a D h at N it Sa Ma Pa N it Sa Relj
tj**
Neither of these omissions is essential in Bhairvi, and theoretically there would be

no objection to the omission of other notes, for instance, the G at, but in fact this
occurs very seldom.
Our second example is the rcig Pilu, where in several songs given by Bhatkhande
our argument is well illustrated, as in the following:1
Ex. 77. rag Pilu— trital

Sa N i S a R c G a t R c Rc N i N i' Sa Ni D hat P a^ P a D hat Ni Sa Sa
* r r ' - - w-
G at G at G at G at G at — G at Re G at Ma Pa Pa M a G a t G atR e TSIi Sa
Ma
Gal} Gal} G all Gall Ma Ma M a M a Re Ma R e M a P a D h a t M a P a Ma G a t G a(^° 3î Sa
Bhatkhande ascribes this rag to K dfi that, but in this song the Nin is used in both
ascent and descent, the Nib appearing only once in oblique descent in the second part
of the song (antra) in the following fragm ent:
Ex. 78. rag Pilu

Pa Gal} M a Ma"~^N it Pa Ma~ G a t G at R e Ni Sa
In Ex. 77 we are concerned with the three possibilities: the occurrence of the scalar
Gab, the occurrence of its alternative Gan (probably introduced to balance the Nin
1K.P.M. Ill, pp. 615-16, sthdyi and antra.
125
Transilient Scales
and used as a leading note to the Ma) and the omission of the G a altogether. In the
second line of the sthdyi (main verse) the Gat> is clearly established as an ascending
note in the figure marked x. In the first bar of the next linethe accidental Gab is used
in ascent as a leading note to the Ma, marked y. In the following bar,however, the
Ga is omitted in the ascending phrase, marked z.
In the rags Bhairvi and Pilu the omitted note is correlated with the first order
accidentals. The tendency to omit the unbalanced notes does not, however, depend
on the use of accidentals. In the rag Yaman {Kalyari that, No. A l) the notes of the
ascending and descending lines are given as consecutive sequences but, in fact, any
combinations of the notes of Kalydii that are permissible.1 The absence of precise
rules of ascent and descent permits us to note trends which have entered in the
performance of this rag. Of these, the four most prominent are:
(1) The tendency to omit the Sa:2

Ex. 79. rag Yaman
Dha N i, R c, Ga Re, N i Re G a, R e, Ni R e, Sa
^ j} j- ^ ^ J ^ 4’
(2) The tendency to omit the Mas and the N i:3
Ex. 80.
Pa G a, Pa Dha P a, Sa
4Ji ^ Li ■
(3) The tendency to omit the Mas and the Ga in a descending cadence :4
Ex. 81.
Ni Re Ga Mail P a , R e, Sa
(4) The tendency to omit the P a:5

Ex. 82.
Mail D ha, Pa
1 H.S.P. I, p. 51.
3 Ibid., p. 488, variation11.
4 Ibid., p. 488, variation11.
5 Ibid., p. 488, variation 10. The omission of Pa in Yaman is not so prominent in Bhatkhanqle’s
notations but is very much in evidence at the present time.
126
Transilient Scales
In Kalyan that the Sa (I) and Ma# (IV#) are unbalanced (see p. 81) and here we see
evidence of the omission of both these notes. These omissions can be called ‘first order’
to be consistent with the terms used in the previous chapter on alternative notes.
However, once the Sa or the Ma# are omitted there is a tendency to make a second
order omission of Pa (V) and Ni (VII) respectively. Thus the Ma# and the Ni are
omitted in the same phrase (as in Ex. 80 above). The Sa and Pa may also be omitted
in one phrase as follows, although this does not occur in Bhatkhande’s notations:
Ex. 83.
D h a N i Re G a M atfDha N i fe e, 5a
The notes Sa and Pa are, however, omitted in exactly parallel phrases (see phrase in
Ex. 79 marked ^ and Ex. 82). Kalya$ that is in some ways an exceptional case, for
the second order omissions are of greater importance than is usually the case. This
can perhaps be explained by the fact that the first order omissions, Sa and Ma#, which
have such an extreme contrast of dynamic functions, are not usually thought of as a
pair. Although a comparison of the ascending conjunct tetrachords is not feasible
owing to the absence of the Mat* (IVti), the descending conjunct tetrachords, Sa-Pa
(1-V) and Pa-R e (V-II), are not only comparable but parallel in Kalyan that. This
tetrachord scheme is emphasised in the two cadences, Pa-Sa in Ex. 80 above and
Pa-R e-S a in Ex. 81, and this may explain the omission of the Ma# (IV#) and Ga (III)
in the descending cadence.1
In these examples from the rags Bhairvi, Pilu and Yaman, we have attempted to
demonstrate that it is, in fact, the unbalanced notes that tend to be omitted. As the
heptatonic thats, excluding Marva, Purvi and Tori, are balanced in one tetrachord
scheme, either conjunct or disjunct, the omission of one note is bound to disrupt the
balance. In Marvd and Tori, however, the omission of one note, the Pa, is essential
to produce balance, so that these two thats are, in this respect, exceptional. On the
other hand, balance can be preserved in pentatonic scales if the omitted notes are
in corresponding positions in the two tetrachords. Thus for a scale in which notes I,
II, III, and IV are balanced by V, VI, VII, and 1 respectively, the omission of II and
VI will still leave I, III, and IV balanced by V, VII, and i. From the standpoint of
balance, therefore, pentatonic scales are generally much more satisfactory than
hexatonic scales.
In Bildval that the omission of one of the two unbalanced notes, either M a (IV)
or Ni (VII), will disturb the ascending disjunct balance which is characteristic of this
that. This does not mean that there can be no hexatonic mgs in Bildval that, but that
balance is not inherent in these hexatonic scales and may have to be achieved in some
1 This rag can be heard on the accompanying record and is discussed further in Appendix B,
p. 204.
127
Transilient Scales
manner in the rags themselves. Therefore, we propose to discuss pentatonic scales
first and to consider the omissions in pairs.
In Khamaj that the corresponding pairs of notes can be seen vertically:
Ex. 84. Khamaj that
Dha Nib
Dha N it G a Ma
Here the Ga and N it are the first order omissions. If these are omitted, the ascend
ing disjunct tetrachord will become balanced, while the balance of the original
descending conjunct tetrachord scheme will be disrupted, as the G a will no longer
be there to balance the D ha and, similarly, the N it will not be there to balance the
Ma. The resulting scale will be quite satisfactory as one tetrachord scheme will be
balanced. There is also a certain justification for omitting the pairs G a-D ha or
N it-M a, first and second order omissions clockwise and anticlockwise respectively
round the Circle (see p. 59). These would leave the descending conjunct tetrachords
balanced and cause no great disturbance in the scale. There is, however, no functional
need to omit these notes, for in the descending conjunct scheme of Khamaj they are
already balanced. These two pairs are the first order omission notes of the adjacent
thafs, Ni and Ma in Bilaval, Gab and D ha in Kdfi, where the need to omit them is
functional.
Ex. 85. Bilaval that
Ga Ma Ni Sa
Ga Ma
Ex. 86. Kdfi that

Transilient Scales
The resulting pentatonic scale whether derived from Khamaj or Bildval will nat
urally be the same, for the note which distinguishes Khamaj from Bildval, the Ni,
is one of the two notes omitted. Consequently, the pentatonic scales derived from
omitting the respective first and second order notes from any that are identical with
the scales derived from their adjacent thats by the omission of the two first order
unbalanced notes.
There still remain two other pairs in Khamaj which could be omitted, R e-D ha
(II-VI) and R e-Pa (II-V). (The pairs in which the Sa is one of the notes may not be
omitted.) The omission of the former destroys the parallelism in both tetrachord
types, but the omission of Re and Pa still leaves the possibility of parallelism in
ascending conjunct tetrachords, Sa-G a-M a (I-III-IV ) and M a-D ha-N ib (IV -V I-
VII b). Here again, there is no reason why these notes should not be omitted but there
is no functional need for doing so.
It must be stressed that none of the pentatonic scales are perfectly balanced in both
conjunct and disjunct tetrachords, and in this respect they are no better than hepta
tonic scales. In Bildval that, for instance, if the first order unbalanced notes, M a (IV)
and Ni (VII), are omitted, the original unbalanced descending conjunct tetrachords
of Bildval will become balanced, but the ascending disjunct tetrachords will now be
unbalanced:
Ex. 87. Bildval Pentatonic
Sa Re Ga — Pa Dha — Sa
. -------------- ■ = &
'S * .... ....
Pa Dha — Sa Re Ga — Pa
A A
i k ------------------ ------------ = ^ T - r - — = ----------------— &
L , ----------. ---------- J - i g i
The G a and the Sa are unbalanced in the ascending disjunct tetrachords. If the
process of omitting the unbalanced notes is carried further and the G a omitted, this
will remove the balance for the D ha in the descending conjunct. The omission of this
D ha would remove the support for the Re, and if the Re were omitted this would in
turn have an effect on the Pa and through this on the Sa. That this tendency is not
without substantiation in N orth Indian music will be shown later in this chapter
when the rag M dlsri is discussed. The reverse process, that of adding the balancing
notes (i.e. perfect fourths and fifths), has equal justification. The scales of the six serial
thats, which are the foundations of the Indian system, can be constructed by succes
sions of fourths and fifths.1 These are, in fact, the intervals which separate conjunct
1 For instance, Kdfi that, by three consecutive upper perfect fifths from Sa, giving Pa (V), Re (II)
and Dha (VI), and three upper perfect fourths from Sa, giving Ma (TV), Nib (Vllb) and Gab (Mb).
Similarly, Kalydij that could be constructed by a succession of six upper perfect fifths, and Bhairvi
by a succession of five upper perfect fourths and one upper perfect fifth.
9 129
Transilient Scales
and disjunct tetrachords. The heptatonic scale is readily obtained from its related
pentatonic scale merely by adding the balance notes—in the above example, N i (VII)
to balance the Ga (III) and Ma (IV) to balance the Sa (I). This being so, it does not
make much difference whether the Indian system originated from a heptatonic or a
pentatonic base, for they are complementary. To be consistent with the modern Indian
tradition, however, we shall discuss transilient scales in terms of their heptatonic
‘parents’.1
From each of the ten thats in the Circle of That?,, a pentatonic scale can be derived
by the omission of the two unbalanced notes. The derivatives of five of the six serial
thats are quite regular and are given below, where the unbalanced notes of the that
to be omitted in the pentatonic derivative are shown in brackets:
Ex. 88.
That Representative rag
. Sa Re Ga " M a ' Pa Dha ~ Ni ’ Sa
1 .. ■ ■ -J
Bilaval (No. A2) Bhupali
^ Sa' Re "G a * Ma . Pa‘ .Dha * N i t ' Sa

Khamaj (No. A3) Durgd
— m—
4 ^ — —
a Sa Re * G at“ Ma Pa ~Dha~ N it £a
Kdfi (No. A4) Jtj------—-------- “1-------1 Sarahg
- t # —1
. a Sa * Re ‘ G at Ma Pa DhafT N it Sa
Asavri (No. A5) ^ --------- Dhani
k * ----------- * ------------- *
a Sa " R e t' G at Ma " Pa " D h at N it §a

Bhairvi (No. — ^ — — — • ----------
Mdlkos
These are the five most common pentatonic scales whose occurrence is not limited
to India alone. A characteristic feature of these scales is that they have no semitones
and, for this reason, are sometimes called anhemitonic.
The pentatonic derivative of Kalyan that is not so easily resolved, for Sa is one of
the unbalanced notes and may not be omitted. We have, however, already referred to
1 This terminology follows Bhatkhande who says, for instance, that the pentatonic rag Bhupali is
bom of Kalya$ fhdf (bhupali rag kalyan that se utpann hotd hai). K.P.M. Ill, p. 23.
130
Transilient Scales
the enharmonic compromise growing out of the temporary omission of Sa, as a re
sult of which Re (II) is seen in relation to Ma# (IVs)- This, we have suggested, was
the basis of Marva that. In the same way, we can derive a pentatonic form related to
Kalya# that by the omission of Re and Ma# instead of Sa and M as:
Ex. 89.
“ R e~ Ga *M aC Pa Dha Ni Sa
Kalya# (No. Al) Sahkra1
1' • — •—
:!^ = H w
In our Circle of Thats we had accepted a second enharmonic compromise, that

between Bhairvi and the hypothetical that where the Pab (Vb) is interpreted as a Ma#
(IVs). For the derivation of a pentatonic scale from Bhairvi that, this compromise is
unnecessary since the omission of Pa is not forbidden. Thus the pentatonic derivative
of Bhairvi', Malkos, is formed by the omission of Reb (IIb) and Pa (V), the unbalanced
notes.
The thats on the left side of the Circle, Marva, Purvi and Tori, could each have more
than one derivative as more than one pair of notes are unbalanced. The derivatives
around the Circle are as follows:
Ex. 90. Representative rag

"Reb" Ga Mail Pa "Dha" Ni Sa
0
M arva (No. A 10) M alh'i
^ Sa R eb "G a ' Mail Pa 'D h ab Ni Sa
Purvi (No. A9)

h i —
ft Sa Reb "Gab" M a# Pa D hab "* N i “1 Sa

Tori (No. A8) ~
a Sa Reb G ab "M af Pa D h ab " N ib " Sa

(No. A7) Bhupal Toj-i
■J L —
It will be seen that the derivatives of Purvi and Topi are not in common use at the
present time. However, the rag Sri shows, in its pentatonic ascent, a close connection
with the pentatonic Purvi derivative, while the derivative of Tori does apparently
1 There are two main versions of rag Sankrd. One is pentatonic as above, while the other is hexa-
tonic, Ma being the only omitted note (K.P.M. IV, p. 208). This second version can be heard on
the accompanying record and is discussed in Appendix B, p. 206.
131
Transilient Scales
occur in a very seldom heard rag called Yasd RanjnL1 Neither of these two derivatives
has either conjunct or disjunct balance, a fact which may help to explain why these
pentatonic forms are so rare.2 The pentatonic derivatives of the Circle of Thats are
shown in the following diagram, where dotted lines indicate the notes which are
unbalanced and therefore omitted:
SANKRA
S a G a P a DhaNi Sa
MALSRl BHUPAU
S a Ga Ma# Pa Ni Sa Sa Re Ga Pa Dha Sa
Ma1
A1_
Dha KALYAN
A10 A2^
M ARVA BI LAVAL
DURGA
S a RebMa# P a Ni S a — Dha1 S a Re Ma Pa Dha Sa
A3 G a1
G a p Dr v T
Ga’ ,8 TOt^
'" - - 2 ^ SARANG
DhaC 5 ^ Sa ReMa Pa Nib £ a
A7
b h a ir v T
A6
Re'
i dhanT
S a Re^Ga^Pa Dha^Sa S a G ab M aP a Nib S a
BHOPAL TOtjiT
S a Gap Ma Dhap N r Sa
m a Lk o s
Marva that has a second possible pentatonic derivative if Reb and Pa are seen as
the unbalanced notes. The omission of these results in the rag Hindol:
Ex. 91.
'R cT Ga Mail “ P a " D h a Ni Sa
0 Sa
M arva (No. A10) (ift — Hindol
«3 V*
There are three possible pentatonic derivatives of Bhairav that since the three pairs
of notes, Reb-Pa (Ilb-V), Ga-Dhab (III-VIb), and N i-M a (VII-IV), are unbalanced
1 Vasant, Rag Kos, Hathras 1962, p. 21.

2 Several commonly heard pentatonic rags are, however, also basically unbalanced, and yet
an element of symmetry is produced in them. Some examples of these are discussed later in this
chapter.
132
Transilient Scales
in the descending conjunct tetrachords. Of these, only one pair appears to he signifi
cant. In the previous chapter we had noted that Ma# (IV#) and Nib (VII b) were the
most commonly used accidentals in Bhairav that (see table p. 119). The omission of
the two unbalanced notes, M a and Ni, results in the commonly heard pentatonic rag
Vibhas:
Ex. 92.
Sa Re !> Ga "Ma " Pa Dhab " N i " Sa
Bhairav (No. C9) ~ t — — Vibhas
Another pentatonic rag, Gunkri, ascribed by Bhatkhande to Bhairav that, is better

discussed as a derivative of the rag Vasant M ukhari (No. C6). According to Bhat
khande the latter has been introduced from the South Indian system,1 but it would
appear that the impulse to produce this scale must be indigenous to N orth Indian
music since the introduction of Nib (VIIb) as an accidental into Bhairav that suggests
it. It will be noticed below that neither the ascending disjunct nor the descending
conjunct tetrachords in this scale are balanced, a fact which may explain why it has
no lengthy tradition in N orth Indian music:
Ex. 93. Vasant Mukhari

Sa R et Ga Ma Pa Dhab
1 ^ u -------------- m—-
__ >
Pa Dhab N ib Sa
\l = L -Z — U
The pentatonic rag Gunkri can be derived from this by the omission of the two
unbalanced notes, Ga and Nib, in the ascending disjunct tetrachords of Vasant
M ukhari, leaving two parallel disjunct, gapped tetrachords:
Ex. 94.
Sa Reb Ga Ma Pa Dhab "N ib" Sa
Vasant Mukhari (No. C6) Gunkri
— *—
The scale of Vasant M ukhari provides a link between Bhairav (No. C9) and
Bhairvi (No. A6) in the Circle of Thats. There are indications of similar connections
between Bhairav and other thats of the Circle—the connection with Marva (No.
1 K.P.M. VI, p. 446.
133
Transilient Scales
A10) through the intermediate Anand Bhairav (No. BIO), with Bilaval (No, A2)
through Nat Bhairav (No. C2), and with Khamdj (No. A3) through the two inter
mediate scales Anand Bhairav (No. BIO) and Ahir Bhairav (No. C3). These connec
tions cannot, however, be established conclusively.
O f the twenty-seven pentatonic rags discussed by Bhatkhande, sixteen appear to be
derivatives of the Circle of Thats and two others of Bhairav that. This still leaves a
significant percentage of the pentatonic rags unexplained. Three others clearly have
their origin in South Indian music, a fact which is acknowledged by Bhatkhande:
Hamsdhvani (Ex. 95a), Abhogi (b) and Devrahjni (c). In addition, Meghranjni (Ex.
95d) appears either to have been imported from South India or to be the modal
series from the fifth of Devrahjni with N h instead of the Nib,1 and these four rags do
not appear to represent evolution within the system.
Ex. 95.
(a) rag Hatpsdhvani (b) rag Abhogi
Sa Re Ga Pa Ni Sa Sa Re Gab Ma D ha Sa
h i
(c) rag Devrahjni (d) rag Meghranjni

Sa Ma Pa D hab Ni(b) Sa Sa Reb G a Ma Ni Sa
^ U « -
There still remain five pentatonic rags which have apparently originated within the
N orth Indian system but do not seem to be connected with the derived pentatonic
scales we have discussed. On the assumption that the omitted notes are first order
unbalanced notes, we can attempt to reconstruct the heptatonic scales of origin o f
these rags. In the following table the unbalanced notes to be added to the pentatonic
rag are shown in brackets:
Ex. 96.
Pentatonic Conjectural
rag Heptatonic Parent
Sa G a ’ Ma pa D hat N it’ Sa
Vasant Mukhari
<5 L• J - (No. C6)
Gunkri
. Sa R et 'G a t ' Ma Pa D hat 'N i Sa
(No. B6)
v vm L J _
1 Meghranjni and Devrahjni are the only two rags in which consecutive notes are omitted and it
seems probable that they are connected.
134
Transilient Scales
Pentatonic Conjectural
rag Heptatonic Parent
'R e t Ga Ma Pa 'D h a ' N il Sa
rt Sa Ahir Bhairav
—~ —
■p»— (No. C3)
Tilahg
Sa "R e ” Ga Ma Pa D liaf N il Sa
Cdrukesi (No. B3)
*) *
Sa ' R e t ' Ga Ma
Durga ‘ Pa " D ha N il Sa
A fur Bhairav
(2nd tradition) -*■
. .
(No. C3)
Sa R et Qa
Sa Anand Bhairav
(No. BIO)
Jet (Marva that)
. Sti R e t Ga ‘ M o#' Pa D ha r Sa .
A I=::=: •• - • * - 1 (No. CIO)

«) m
Candrkos Sa 'R e t ' G at Ma 'P a " .D h a N it Sa

m (No. B4)
(Bhatkhande’s 3
tradition) - _ __
The first four of these rags could have been derived from off-shoots of Bhairav
that, Vasant Mukhari, Ahir Bhairav or Anand Bhairav. There is, however, no con
clusive evidence which can be brought to bear on this matter. The connection be
tween Ahir Bhairav and Khamdj that is apparent in certain traditions where the Re 4
is sometimes used in ascent in Ahir Bhairav ;x there is, however, no indication of a
tendency to omit the Reb in the notations of this rag. There appears to be no justifi
cation for the assumption that the above Candrkoti, which is now extremely rare,
may have originated from a heptatonic scale. The explanation that it is a combina
tion of the rags Malkos and Bagesri2 may very well apply in this case.
It will be seen, however, that the majority of Bhatkhande’s pentatonic rags appear
to have been derived from the heptatonic thats or from the heptatonic derivatives of
Bhairav, while only a small proportion are either borrowed from outside the N orth
1K.P.M. V, p. 346, and in songs on pp. 348, 349 and 351.

2 This is a rag of Kafi that whose ascending and descending lines are given by Bliatkhapdc
(K.P.M. HI, p. 444) as follows:
rag Bagesri
Sa, N il Dha N il Sa, Ma' O a t, Ma Dha N il Sa §a. N il Dha, Ma Gal, Ma Gal Re Sa
135
Transilient Scales
Indian system or intellectually created. In present-day practice there are several other
pentatonic rags which have not been mentioned in Bhatkhande’s w orks: Madhukos
(Madhukdus), Kaldvti\l Sivranjni, and another version of Candrkos (Candrkdus):
Ex. 97.
(a) rag Madhukos (b) rag Kaldvti
Sa Gat Ma# Pa N it Sa Sa Ga Pa Dha N it Sa
(c) rag Sivranjni (d) rag Candrkos

Sa Re G at Pa Dha Sa Sa G at Ma D hat Ni Sa
Since these rags are not mentioned by Bhatkhande they would appear to be recent
innovations. In spite of this, their origins are not clearly known. The second version
of Candrkos is very likely derived from Malkos with the leading note Nib replacing
the N it in both ascent and descent. This would not be the only instance of the
ascending leading note becoming a scalar note, for this also appears to occur in the
rag Pilu where Nib is frequently used in descent (the scalar Nib occurs in the more
usual descending line) and in the rag Patdip in which this is carried further to the
complete exclusion of the Nib.2 The rag Madhukos is the modal series beginning on
the fourth of this version of Candrkos. This series would naturally be prominent in
Candrkos as the secondary drone would be tuned to M a (IV), there being no fifth in
the rag. The rag Madhukos appears to be unstable, however, as the Mas leads to the
Pa while there is no comparable pull from the N h (Vllb) to the Sa (I). Thus there
will be a tendency to shift back to Candrkos which is the series from its Pa. Madhukos
is, however, clearly connected with the heptatonic Madhukant (No. C4), and if
indeed Madhukos has its origin in Candrkos, the heptatonic Madhukant has been
derived from it by the addition of the Re (II) and D ha (VI), first and second order
balance notes respectively. Madhukant, too, appears to be unstable as it also has a
Mas and Nib. This may account for the heptatonic rag Madhuvanti(No. C l) in which
the Nib has replaced the Nib of Madhukant. From Madhukant too the pentatonic
rag Sivranjni may have been derived by the omission of both Ma# and Nib. This
possible chain is schematised below:
1 This rag may also be of South Indian origin.

2 In Bhatkhande’s time this rag, called Patdlpkl, generally had Nib in ascent and Nib in descent.
Since then it appears that the Nib has gradually been replaced by Nib in descent.
136
Transilient Scales
Ex. 98.
Malkos
Sa Gal» Ma Dhab ‘N i Sa
Candrkos
Sa G at Ma§ Pa N it Sa
Madhukos
— if 1
I
Sa Re G at Matt Pa Dha N it Sa
Madhukant d): —
4^ —It3
if \
Sa Re G at 'Mail Pa Dha N i Sa Sa R e G a t Pa D ha Sa
Madhuvanti Sivranjni
t o
This is, perhaps, an oversimplification of the evolutionary process, and it may never
be possible to establish conclusively the validity of this sequence. T hat this whole
process has taken a relatively short time, perhaps less than fifty years, can be explained
by the fact that all these derivatives of Malkos are essentially unbalanced scales. This
does not necessarily mean that they are musically unsatisfactory, although balance
appears to be an extremely im portant aspect of the scales which have withstood the
test of time. There are, however, pentatonic scales even among the derivatives of the
Circle of Thats which are not, in themselves, balanced.
The five anhemitonic scales have simple balanced tetrachords either conjunct or
disjunct. These five are related to each other serially in exactly the same way as are
the seven primary heptatonic scales in that they too can be derived by starting on the
successive degrees of any one of the five scales. They can also be derived in succession
by redressing the balance in each of the scales. In the example overleaf the arrows
show the notes introduced to create each new balance.
It will be apparent that the series is discontinuous at both ends as the next step
would require in each case the omission of the Sa. If this series is to be continued
beyond Bhupdli the Ni replaces not the Sa, but its neighbour, the Re; and beyond
M alkos the Reb must replace the Nib. The changing notes in this series of pentatonic
scales are a succession of perfect fourths (or fifths) as in the Circle of That?,. If this
pentatonic Circle is to be completed, a similar enharmonic compromise is necessary,
occurring between Bhairvi and the hypothetical that where the Pab (Vb) is interpreted
137
Transilient Scales
as Ma# (IV#), and between Kalyan and Marva where Sa# (I#) is interpreted as Reb (lib).
Ex. 99.
N i (replacing Sa)
Sa Re
t
Ga Pa Dha Sa
Bhupall
Pa Dha Re Ga
Sa Re Ma Pa Dha
Durga Dha Re Ma Pa
t
Sa Re Ma Pa N ib Sa
Sarang N ib Re Ma Pa
Sa Gab Ma Pa N ib
Dhani Pa
Gab Ma
Sa Gab Ma Dhab Nib Sa
W
Malkos hab Ni b Sa Gab Ma Dhab
R eb (replacing the Sa)
138
Transilient Scales
The scales of Sankra (pentatonic version)1 and Hindol are unbalanced both as
conjunct as well as disjunct tetrachords:
Ex. 100.
(a) Sankra
Pa Dha Ni Sa
Pa D ha Ni Sa
(b) Hindol
MaS D ha Ni Sa
Dha Ni Sa
Although these scales are clearly unbalanced, a measure of symmetry is neverthe

less produced within the rags themselves. In Sankra this is accomplished, to some
extent, by frequently omitting the D ha and making the Ni oblique (vakr) in ascent.
Bhatkhande gives its ascending line as follows:2
Ex. 101. rag Sankra
Sa Ga, P a, Ni D ha, Sa
t— .v -J1— y
This twist in the ascending movement is an unconscious but functional element

for this enables the m ajor third * and the minor third y to be balanced in the upper
tetrachord, the former disjunct, the latter conjunct. In descent the D ha is generally
omitted, accentuating the m ajor third, N i-Pa, which is finally balanced by the other
major third, G a-Sa.3
1 Bhatkhande also mentions the fairly common occurrence of a second version of rag Saiikra
which is hexatonic. It is this hexatonic version which is played by Ustad Vilayat Khan on the
accompanying record and is discussed in Appendix B, p. 206.
2 K.P.M. IV, p. 208.
3 These turns, slightly modified, are well illustrated on the accompanying record.
139
Transilient Scales
Similarly, in the rag Hindol an element of symmetry is also created by an oblique
movement. Bhatkhande gives the ascending line as follows i1
Ex. 102. rag Hindol
i x
— — i i —y ■— i
Sa Ga, Matt D ha Ni D h a, Sa
•J - ' ~ 11
Here the wholetone x and the minor third y are repeated in the upper tetrachord.
Although M a#-Dha and D ha-Sa are not parallel in the tetrachordal sense, they do,
nevertheless, convey an illusion of symmetry. The significance of this oblique move
ment is not, however, appreciated on a conscious level. Bhatkhande states that the
Ni in Hindol is oblique in ascent and adds that the less the Ni is used the more clearly
is the rag delineated. His explanation for this is that excessive use of the N i will
evoke the rag Sohni.2 This is not a veryconvincing argument in view of the fact that
Sohni’ a rag of Marva that, is hexatonic and has a very prominent Reb (lib), and is thus
easily distinguished from Hindol. The frequent omission of the Ni in the ascent of
Hindol (in spite of its function as a leading note) and its descent can be reasonably
explained as an unconscious attempt to produce symmetry in a rag whose scale is
clearly unbalanced. Non-tetrachordal symmetry may be more readily acceptable in
this scale where both the Mas (IVs) and the Pa (V), the initial degrees of the second
tetrachord conjunct and disjunct, are absent.
The rag Mdlsri is perhaps the most extraordinary of the pentatonic rags. From the
standpoint of symmetry the scale of Mdlsri is not nearly so awkward as Hindol and
$ankrd} N i-S a-G a being balanced by M as-Pa-N i (ascending disjunct in the octave
from Ni to N i):
Ex. 103. rag M dlsri
tl^ i) Sa Ga Ma#
(M a#) Pa Ni Sa
However, the expression of this symmetry is by no means characteristic of the rag,

although the parallelism between the m ajor thirds Sa-G a and Pa-N i is sometimes
brought into focus. Apparently, Sa and Pa are so strong here that they do not permit
the octave to be viewed as a range from Ni to Ni as is characteristic of Marva, Purvi
and Toyi thats. The absence of the Re in Mdlsri focuses attention on the fact that the
Ma? and the Sa are unbalanced, with the consequence that the Ma? is also frequently
1 K.P.M. IV, p. 176.
2 Ibid., p. 176.
140
Transilient Scales
omitted. This omission is then countered by the temporary omission of the Ni.
According to Bhatkhande, some musicians say that M atin should have only three
notes, Sa, Ga and Pa,1 and this is certainly evident in some of the songs in this rag.2
In the following example the Mas and Ni are used merely as grace notes:3
Ex. 104. rag Malsri—Jhaptal4
Ni
Pa Ga
x 2 0 3
Ni Ga
Ga Ga Ga
Xj
In Bhatkhande’s system, however, no rags may have less than five notes, thus a small
measure of symmetry is still maintained.
In the rag Bhupal Tori once again the simple tetrachord schemes beginning on
Sa and Pa (I and Y) or Sa and M a (I and IV) must be disturbed if symmetry is to be
produced, for the parallelism in the scale of this rag lies between Dhab-Sa-Reb and
Gab-Pa-Dhab (ascending disjunct in the octave from Dhab to Dhab), in both instances
a semitone following a m ajor th ird :
Ex. 105. rag Bhupal Toyi
( Dhab) Sa Reb Gab Pa Dhab Sa

= i
i0. 1?#' "■
® < l«)
--- >
* (Gab) Pa Dhab Sa Reb Gab Pa
---------- --------
*
h----------------
!)■ 1:------------- r• ...— ..... ^ 6vl
----- ---- !>•-------
1 Op. cit. V, p. 78.
2 K.P.M. V, pp. 93-4.
3 Ibid., p. 91—sthdyu
4 Jhaptal is a time measure of ten units, subdivided into four groups, 2 + 3 + 2 + 3.
141
Transilient Scales
This scheme is, in fact, followed in the rag and phrases frequently begin on Dhab and
Gab,1 as is well illustrated in the following song.2 In the second line of this song the
parallelism is also apparent.
Ex. 106. rag Bhupal Tori— Dadra
Gab D hab Pa D hab Pa Pa G ab R eb Sa R eb Sa Sa.
r r J - ...r i j j j —i T -■
» 1" | “I
$
4f=tal = “ = • ! = “ — J— J------ 1** 1>«I J. 1 J -J. J. -
= \
\>
x 0 X 0
D hab Sa R eb G ab R eb G ab Pa D h ab Pa G ab R eb Sa
|_j. J 1>J bJ "lr m

The vadi and samvddi of Bhupal Tori, Dhab and Gab respectively in Bhatkhande’s
system, also give an indication of their parallel roles in the rag, for these are the two
notes which initiate the parallel movements. In spite of the fact that Bhatkhande’s
choice of vadi and samvddi is not always based on objective musical principles, they
often provide a clue to the symmetry within a rag. In the rag Sankra (see p. 206),
Bhatkhande gives G a and Ni as the vadi and samvddi. These are the upper notes of
the parallel thirds Sa-G a and Pa-N i (Dha being an oblique ascending note is ap
proached from the Ni), or the initiators of the parallel descending thirds, N i-P a (Dha
is generally omitted in descent) and Ga-Sa. He also mentions a second tradition in
which Sa and Pa are recognised as vadi and samvddi in this rag;3 these are, in fact,
the lower notes of the parallel thirds. In the rag Hindol D ha and G a are given as vadi
and samvddi—the two notes at which we have suggested the parallelism is initiated. In
the rag Mdlsri too the vadi and samvddi, Pa and Sa, are the initiators of the parallel
ascending thirds. Unfortunately, this kind of relationship does not always hold.
It will be observed that even the pentatonic scales derived from the Circle of Thats
are not always balanced. The rag Vibhas, a derivative o f Bhairav, also has unbalanced
conjunct and disjunct tetrachords:
Ex. 107. Vibhas
Sa Reb Pa D hab
Pa D h a*
1 Op. cit. VI, p. 441. Bhatkhande begins on Dhab when illustrating the usual movement (sadharan
calan) in this rag,
2 Bhairav that afik, Special publication of ‘S a iig itHathras, January 1962, p. 114, sthayl of song
by Ye^vant D. Bhatt. Dadra is a type of song based on Dadra tdl, a time measure of six units,
subdivided into two equal parts.
3 K.P.M. IV, p. 207.
142
Transilient Scales
No clear solution to this problem is yet apparent in this rag. There is definitely a
tendency to introduce the Ni (VII) to balance Ga (III), particularly in descent1 where
it generally occurs as a grace note. This produces a measure of disjunct balance
N i-D hab-Pa and Ga-Reb-Sa. There is also a tendency to produce a non-tetrachordal
balance by omitting the Reb in ascent2 leaving two major thirds, Sa-G aandD hab-Sa.
Perhaps the justification for this rag may lie in the inverted symmetry occurring in
the two segments Sa-Reb-Ga, semitone and augmented second, and G a-Pa-D hab,
augmented second (minor third in a heptatonic series) and semitone.3 In Vibhas this
inverted symmetry appears to be a feature of considerable importance and can be
heard frequently in the characteristic phrase (pakaf) :
Ex. 108. rag Vibhas
Ga Pa Dha^Pa, Ga R et Sa
J—• — ——4-------u
In the rag Gurtkri the disjunct parallelism is quite clear:

Ex. 109. Gunkri
Sa R e t Pa Dhab
Pa Dha b Sa Ma Pa
Ex. 110. Tilahg

Sa Ga Ma Nit Sa
Nib Sa Ga Ma
In Tilahg, however, the scale is unbalanced, but a considerable measure of symmetry

is produced within the rag. In ascent, the leading note, Nib, is commonly used, thus
1 K.P.M. V, songs on pp. 392-9 and 402.
2 K.P.M. V, songs on pp. 392 and 397-9.
3 There is a second rag, Reva, which has this same scale, but is classified by Bhatkhande in Purvi
that because its vadi is in the lower tetrachord while the vadi of Vibhas is in the upper tetrachord.
This, according to his time theory, determines that Reva is sung in the second half o f the day, a
common characteristic of Purvi that rags, while Vibhas is sung in the first half of the day, a feature
o f Bhairav that rags. Reva is, however, a rare rag and Bhatkhande gives only one song from which
no conclusions can be drawn.
143
Transilient Scales
producing parallel ascending disjunct tetrachords S a-G a-M a and Pa-Niti-Sa. In
descent, however, the Nib destroys this parallelism and both conjunct and disjunct
tetrachords are unbalanced. It is not surprising, therefore, that a descent of continuous
steps is seldom heard although it is not prohibited. Instead, the usual descent is as
follows i1
Ex. 111. rag Tilahg
Sa N ib Pa, Ga M a G a, Pa Ga M a G a, Sa
Here the minor thirds Nib-Pa and P a-G a are balanced to some extent, while the flow
of the descending line is disrupted at Ga, thus preventing an easy comparison of
the complete tetrachords.
From the foregoing discussions it will be apparent that most pentatonic rags are
based on balanced scales, but an unbalanced scale does not necessarily preclude the
possibility of a pentatonic rag, for symmetry may still be produced within it by means
of temporary omissions, accidentals and oblique movements. Some of the recently
introduced pentatonic rags do not reveal the same propensity for tetrachordal
symmetry as the traditional rags, a situation comparable to that which also occurs
in the recently introduced heptatonic rags.2 The rags Devrahjni and Meghranjni are
obviously exceptional, containing gaps of a fourth and an augmented fourth respec
tively—virtually a denial of the two tetrachord scheme. The rags Abhogi and Kaldvti
(the modal series from the M a of Abhogi) seem to indicate a new trend for they are
largely unbalanced:3
Ex. 112. Abhogi
Sa Re Gab Ma Dha Sa
Dha Sa Re Gab Ma
Ex. 113. Kaldvti

Sa Ga Pa Dha N it Sa
Pa Dha Nib
1 K.P.M. V, p. 283.
2 See p. 88.
3 This also applies to the rag Harpsdhvani which has been discussed on p. 89.
144
Transilient Scales
The successive intervals of Kaldvti are interesting: major third, minor third, whole-
tone, semitone and wholetone, the intervals gradually decreasing by semitones to the
N ik The same of course applies to Abhogi where the successive intervals decrease by
semitones from its Ma. These two rags seem to be constructed on triads, Abhogi on
M a-D ha-S a and Kaldvti on Sa-G a-Pa. Nevertheless, there appears to be some sign
of balance even within these mgs. In Abhogi a measure of symmetry is possible by
making Gab vakr in descent, a characteristic feature of the Kdnhra family of rags to
which Abhogi belongs, and by occasionally omitting Gab in ascent:
Ex. 114. rag Abhogi

Ma
Sa Sa Dha Ma Gab M a Re Sa Sa Re Ma Dha Sa
In Kalavti the following two ascending phrases are quite characteristic:
Ex. 115. rag Kaldvti

Ga Pa Dha N ib Dha Sa Ga Pa N ib Dha Sa
In the first the minor third G a-Pa is balanced by Dha-Sa. In the second an extra
minor third, Pa-Nib, is added. In both these rags the symmetry does not extend over
a full tetrachord. The growing popularity of these and other similar mgs suggests
the possibility that full tetrachordal symmetry may no longer be an essential govern
ing factor in the natural selection of rags.
In contrast to pentatonic rags, which naturally lend themselves to balance,
hexatonic forms generally tend to disrupt the balance in a scale. For instance, the
omission of the N i in Bildval that will not only leave the M a unbalanced (descending
conjunct) but will also remove a support from the Ga (ascending disjunct). Thus it is
inevitable that most hexatonic scales are inherently unbalanced and symmetry can
occur only in the rags.
The most obvious examples of hexatonality occur in Marva and Tori thats in which
the Pa is omitted in a number of rags. This, as we have mentioned earlier, is the
usual device whereby balance is created. There is some justification, however, to think
of some of these as pentatonic, for the Sa (second order balance note) is also omitted
in many of the phrases in these mgs, but, being the ground-note of the system, cannot
be omitted entirely. In Purvi that) where the Pa is also unbalanced, the omission of the
Pa is not entirely necessary for balanced disjunct tetrachords already exist in the
octave from Ni to Ni. The omission of the Pa may be considered functional when
viewed in terms of conjunct tetrachords where the Pa is seen in relation to the Reb,
but this omission will disrupt the balance in the ascending disjunct tetrachords of
10 145
Transilient Scales
Purvi. In Marva and Tori thats, however, the Pa m ust be omitted in order to produce
any tetrachordal balance:
Ex. 116.
(a) Marva that (b ) Purvi that (c) Tori that
Mag (Pa) D ha NL (Sa) N i Sa Reb Mail (Pa) Dhab N i (Sa)
‘ ... I
Reb Ga Mag (Pa) Mail Pa Dhab N i (Sa) Reb Gab M ag (P a )
There are four hexatonic rags of Marva that in which the Pa is omitted: Marva,
Sohni, Puriya and Piirbya. O f the three principal rags in Tori that, one, Gujri Tori,
is also hexatonic, Pa being omitted. In addition, Pa and Sa are frequently omitted
in the rag Tori itself. In Purvi that, on the other hand, there are no hexatonic rags
omitting Pa, while there is one, Triveni, in which the Mag is omitted, still preserving a
measure of balance: Sa-Reb-Ga, and Pa-Dhab-Ni.
A part from the hexatonic rags of these thats, the majority of hexatonic rags given
in Bhatkhande’s works are closely related to pentatonic forms. These occur princi
pally in Bilaval, K afi and Bhairav thats, which all have the common feature of
balanced disjunct tetrachords. Nevertheless, the parallelism in their hexatonic rags
may be either conjunct or disjunct.
In Bilaval that Bhatkhande gives only one hexatonic rag, a version of the rag Hem
Kalyari. Here the G a is treated as an oblique note, thus maintaining a disjunct balance
in the descending line:1
Ex. 117. rag Hem Kalyari

Sa Dha Pa Ga Ma Re Sa
In K afi that there are two main groups of rags which are hexatonic. Four of these
belong to the Sarahg group of rags and three to the Kanhra group. In the pentatonic
Sarahg rags, Vrindavni (or Brindabni) and Madhmad, the symmetry lies in the des
cending conjunct tetrachords:
Ex. 118. rag Vrindavni Sarahg
Sa N ib Pa Ma Re
1 K.P.M. V, pp. 99 and 100. The descent given is abstracted from the calan on p. 100.
146
Transilient Scales
In the hexatonic rag Suddh Sarahg this symmetry is maintained while the added
note, Dha, is used around the pivot Pa, together with the accidental Ma#, as in the
following example:1
Ex. 119. rag Suddh Sarahg

Sa Nit* Pa, Matt Pa D h a P a , Mali R e , Sa
In M iya k i Sarahg the D ha is attached to the Nib, more or less as an ornam ent:2
Ex. 120. rag Mlyd k i Sarahg

£a Nib D ha Nib P a, M[a E.e, Sa
J •i = M
In Samant Sarahg the D ha is used in an alternative descending line while the

normal balanced pentatonic descent of Sarahg remains. In Barhams Sarahg the
D ha is only occasionally used as a passing note but only in ascent.
In all three hexatonic rags of the Kanhrd group, Nayki, Suhd3 and Devsakh, the
descent maintains conjunct balance by the oblique use of Gab:
Ex 121. rag Nayki Kanhra

Sa N ib P a, Ma Pa, " • ^ G a V M a Ro , Sa
I \ ------- 1-------
J ■j — m----- J : ------
Bhatkhande gives one hexatonic rag in Asavri that, Gopikd Vasant, in which the
Re is omitted so that it resembles Malkos but has a Pa. This rag follows the Malkos
parallel scheme of conjunct tetrachords by oblique use of the Pa as in the following
descending line:4
Ex. 122. rag Gopika Vasant

N ib D h a b M a , Pa Gab, M a Gab Sa
1 Abstracted from K.P.M. VI, pp. 482-3.

2 Abstracted from K.P.M. VI, pp. 189 and 489. Dha is often omitted leaving the pentatonic
descent.
3 See accompanying record and notation in Appendix B, pp. 200,201.
4 Abstracted from K.P.M. VI, p. 421. Here it is said to be a South Indian rag.
147
Transilient Scales
In Bhairav that, Bahgal Bhairav still maintains parallel descending disjunct tetra
chords by the oblique use of G a:1
Ex. 123. rag Bahgal Bhairav

Sa Dhab P a, Ma Pa Ga M a, R et Sa
I i -{n
The rag Jogiya occurs in several different variants. In Bhatkhande’s system it is
pentatonic in ascent and hexatonic in descent, the ascent having disjunct balance:
Ex. 124. rag Jogiya

. Sa R et Ma Pa Dhab Sa
The Ni, introduced in the descent, spoils this parallel movement. However, a
measure of non-tetrachordal symmetry is maintained by making Pa oblique and the
Ni-Dhab is balanced by D hab-M a:2
Ex. 125. rag Jogiya

Sa Ni Dhab Pa, Dhab M a, R et Sa
I - L f ! ^ ur J ^
In two of the six songs in this rag given by Bhatkhande, however, the Ni does not
occur at all and the rag is completely pentatonic and has disjunct balance.3 In another
song the Ni occurs only in descent as a grace note and is balanced by the G a which
is used in the same way.4 There is a very prominent tradition5 in which the ascent
remains pentatonic, but the descent is heptatonic and has disjunct balance:
Ex. 126. rag Jogiya

Sa RebM a Pa D h atS a Sa N i Dhab P a , M a Pa D habpa, M a G a R e t Sa
To summarise, the hexatonic rags of Marva and Tori thats appear to have been
derived from heptatonic scales. However, the other hexatonic rags seem to have been
1 K.P.M. V, p. 334.
2 Ibid., p. 378.
3 K.P.M. V, songs on pp. 382 and 384.
4 Ibid., song on p. 380.
5 The tradition followed by Ustad Bundu Khan and his son, Ustad Umrao Khan.
148
Transilient Scales
derived from pentatonic scales and the additional note generally appears in between
the two parallel descending segments of the pentatonic series, often in an oblique
melodic figure, as in Exx. 119, 121 and 123.
From the foregoing discussion it will he apparent that the rags of a balanced
heptatonic that do not always show this simple heptatonic balance. If the straight
ascent is characteristic of one of these rags, then the ascent may be oblique or transi
lient in others. In fact, the straight ascent and descent of the that can generally be
considered too simple a scheme for an individual rag. The rags of one that are often
distinguished by the different ways in which the symmetry of that that is worked out
in each. In many rags the that symmetry is altered by the omission of two balanced
notes in either ascent or descent. This can be called directional transilience. Direc
tional transilience is not motivated by imbalance, as we have suggested complete
transilience to be, and indeed the notes omitted in this way are usually perfectly
balanced. This could be brought about by experimentation (probably at an uncon
scious level)—notes being omitted more or less at random, until a pleasing combina
tion emerges. Here again, balance plays an im portant part in the result, for the
omitted notes are generally a fourth or a fifth apart. The rag Jogiya (Ex. 126) is a
typical example o f directional transilience, being pentatonic in ascent and heptatonic
in descent. The notes G a (III) and Ni (VII) which are omitted in ascent are perfectly
balanced.
The rag Khamdj is particularly interesting since it is described as being hexatonic
in ascent and heptatonic in descent.1 In spite of the hexatonic ascent, a large measure
of balance is maintained, as the hexatonic ascent is really an amalgamation o f two
alternative pentatonic ascents, one in which the Re (II) and Pa (V) are omitted, the
other in which the Re (II) and D ha (VI) are omitted. Since the Re is omitted in both,
it is described as hexatonic. The former leaves parallel conjunct tetrachords (Ex. 127a),
the latter parallel disjunct (b), while the third alternative ascent is really a combina
tion of the two (c):
Ex. 127. rag Khamdj

(a) (b) (c)
Sa G a Ma D ha N ib Sa G a Ma Pa Nil) Sa Sa Ga Ma Pa D ha Ni!} Sa
This last is unbalanced, but since phrases in the rag usually begin fro m G a , the
ascending imbalance is not easily apparent.2
The notes omitted in directional transilience are often ‘weak’ (<durbal) notes in the
rags? The term ‘weak’ here does not refer to the inherent dynamic function of a note
1 K.P.M. II, p. 122.
2 A further discussion of this rag is found in the next chapter,
3 This is not always the case. In the rags Suddh Asdvri and Jaunpuri, for example, Gab which is
given as the samvddi in these two rags by Bhatkhande is omitted in ascent. Similarly, in rag Jhinjhoti
the vadi Ga is one of two notes generally omitted in ascent.
149
Transilient Scales
due to consonance and dissonance, but rather to the dynamic function induced by its
melodic context in the scheme of the rag. Ga, for example, is a note with a high
degree of consonance, but is nevertheless a weak note in many rags (as in Jogiya,
where it is usually omitted in ascent). The notes omitted in the ascent of a rag are
often weak in descent and may sometimes be omitted altogether, suggesting th at
directional transilience may tend to become complete transilience. In general, how
ever, complete transilience is associated with scalar imbalance, while directional
transilience is an unique characteristic of the individual rag.
150
VIII
Symmetry, Movement and Intonation
I n the preceding chapters we have examined certain facets of scale and tetrachord
species which, we have suggested, lead naturally to the introduction o f accidentals
as well as to transilience in rags. A basic principle has emerged out of our discussions,
that rags show a tendency to align themselves in symmetrical units, frequently, but
not invariably, ranging over a tetrachord. These units are either conjunct, in which
the parallel notes are a perfect fourth apart, or disjunct, in which they are a perfect
fifth apart. Sometimes they are neither conjunct nor disjunct but express a ‘false’
symmetry. In view o f the fact that rags are still evolving it is impossible to be certain
whether this ‘false’ symmetry is just a temporary phase or a more perm anent state on
a par with conjunct and disjunct symmetry. While symmetry in a that is a relatively
simple and straightforward matter, the adaptation of this symmetry may become
quite complex in different rags. This complexity is conditioned by melodic features
which differentiate rag from that. One o f these is oblique (vakr) movement.
O B L IQ U E M O V E M E N T
In the previous chapter we referred to oblique movement in connection with hexa-
tonality, and suggested that this was a device whereby pentatonic symmetry could
still be preserved in a hexatonic rag. This is only one manifestation of oblique move
ment. There are many instances where it appears to be associated with the intro
duction of an accidental and may indeed have been motivated by it.
In Chapter V we suggested that an accidental is introduced initially as an inflexion
of the nearest diatonic note which gradually gains recognition and develops into an
oblique ascending or descending line. The Mas, for instance, occurring as an acci
dental would first be merely an inflexion of the Pa and gradually form an oblique
ascending line:
Ex. 128.
Ga Pa MaB Pa D ha Ni Sa
x
151
Symmetry 9 Movement and Intonation
Its chromatic counterpart, Ma#, will, at first, continue in its own ascending line as
a direct ascending note, and there may be two ascending lines, the oblique incorpor
ating Ma#, the direct incorporating Ma#. If the former gains in prominence, the
latter may fall into disuse and eventually become discontinuous—a vestige of the
original. The discontinuous line may become progressively shortened until the Man
is no longer an ascending note, as will be seen in the following series:
Ex. 129.
(a) (b) (o)
Ga Ma Pa Dha Pa Ma Ga Ga Ma Pa Ma Ga Ga M a Ga
Ex. 129c provides an oblique descending parallel for the oblique ascent incorporating
Ma# (Ex. 128, x). This kind of inverted non-tetrachordal parallel movement associated
with the two alternatives of M a is seen in a number of rags, either as in Bihag (Ex.
130a) or as in Hamir (b):
Ex. 130.
(a) rag Bihag (b ) rag Hamir
Pa Ma8 Pa Ga Ma^ Ga Pa Ma# Dha Pa Ga M a# Re
¥ ¥ ¥ ¥
The use of the two alternatives also makes possible directional non-tetrachordal
parallelism, as in the following examples:
Ex. 131.
(a) rag Gauf Sarahg (b ) rag Keddr
Sa Ga Re Ma Ga Pa Ma# D h a . Sa Ma Ga Pa Ma# D h a . . .
$
(c) rag Kamod
Pa Ma# Pa Ga Ma# Re Sa
* — IL_¥
In general terms then, associated with each that are characteristic oblique move
ments as well as typical accidentals and transilience. The oblique movements, too,
often appear to be connected with imbalance in scale and tetrachord species. As with
accidentals and transilience, we can speak of first and second order oblique move
ments. However, once the oblique motion is initiated there seems to be a tendency
to continue it, the classic example being the rag Gaur Sarahg, whose ascent and
descent are given as follows :l
i K.P.M . IV, p. 142.
152
Sym m etry , Movement and Intonation
Ex. 132. rag Gau}' Sarahg
Sa, Ga Re Ma G a, Pa MaltDha P a, N i Dha Sa SaD haN i Pa, Dha Ma# Pa GajMa^Re, Pa, Re Sa
Before continuing this discussion of oblique movement, we should remind the reader
of the different interpretations in the rendering of many rags. In most cases the
differences are matters of detail, which do, nevertheless, preclude a definitive
analysis. Bhatkhande’s own interpretations of many rags, as found in his svarvistdr,
are influenced to some extent by a theoretician’s desire to differentiate clearly between
one rag and another. As a result his own interpretations do not always compare
with the bulk of songs (ciz) notated in his works, most of which have been composed
by traditional musicians. It is clear that certain melodic features are common to
several rags in a that. This is not always apparent in Bhatkhande’s interpretations;
for instance, similar descending lines are found in many of the songs in the rags
Hamir, Keddr, Kdmod, Chdyanat, Sydm Kalydn and Gaur Sarahg:
Ex. 133. rags Hamir, Keddr, Kdmod, etc.

Sa D ha Nik P a, Mall Pa D h a P a, Ga Mail Re Sa
In these songs the Nib is quite unmistakable, yet in Bhatkhande’s svarvistdr the Nib
occurs in only two of these rags, Chdyanat and Keddr A It would appear that where
there are two or three alternative descents in these rags Bhatkhande has chosen,
wherever possible, the line which is unique, as representative of that rag. It is,
however, the similarities in rags which concern us in our study of the relationship
between rag and scale. Let us now consider the characteristic oblique movements
which occur frequently in the rags of a few thats.
In the rags of Kalya$ that, as given by Bhatkhaiide, the typical oblique movement
is associated with the two M a’s (Ex. 134a) which commonly leads to the second order
oblique movement involving the two N i’s (Ex. 134b):
Ex. 134. conjunct symmetry in Kalydn that rags

(a) (b)
Pa Mail Pa Ga Maf] Rc Sa Ni Sa Dha Nik pa
,<| « .. II ~ * « II
The second of these is not always accepted by Bhatkhande, but occurs quite
frequently in the songs. It is an im portant device for preserving symmetry in the rag,
1 Bhatkhande acknowledges, however, that the Nib may be used a little in those rags which have
both the Mab and Mail. H.S.P. I, p. 110.
153
Symmetry , Movement and Intonation
a symmetry which is not always apparent in Bhatkhande’s svarvistdr. In the rag
Hamir, for instance, Bhatkhande gives the descending line as follows:1
Ex. 135. rag Hamir

Sa N i Dha P a, Ma8 Pa Dha Pa, Ga Ma^ Re Sa
Here the two disjunct tetrachords are unbalanced. However, in nineteen of the
thirty-two songs (most of them composed by traditional musicians) notated by
Bhatkhande in the rag, the upper tetrachord has Sa-D ha-N ib-Pa, which is symmetri
cal with the lower tetrachord. In addition, there is a common tendency to omit the Ni
in descent, producing pentatonic balance as follows:
Ex. 136. rag Hamir

Sa D ha Pa Ga Ma Re Sa
This tendency is also apparent in Bhatkhande’s svarvistdr. Disjunct balance is, how
ever, characteristic of Bilaval rather than Kalydn that, another argument in favour of
the classification of rags such as Keddr and Hamir in Bilaval that.2
In the rags of Bilaval that both Ma# and Nib are first order accidentals, thus the
oblique movements involving both these notes are also of the first order. Once the
Nib has been introduced and the associated oblique movement (Ex. 137a) established,
this may lead to the second order conjunct parallel (b). Both of these are typical of
Bilaval that rags and also occur in Hamir, Keddr, etc.
Ex. 137. conjunct symmetry in Bilaval that rags

(a) (b)
Sa D ha N it Pa Pa Ga Ma Re
It will be apparent that oblique movement is not only associated with accidentals
but also with directional transilience since it requires at least temporary omission of
a note in ascent or descent. In Bilaval that rags the omission of the Niu in descent
can be said to be a preparation of the ground for the introduction of the Nib. The Nib
can then easily be introduced as an ornament attached to the D ha and may gradually
gain prominence. Similarly, the associated omission of the Gah in the descent of
1 K.P.M. IB, p. 68.
2 See p. 53.
154
Bilaval that rags paves the way for the introduction of the Gab as in the following
example, a phrase which is characteristic of the rags Jaijaivanti and Gdrd (Gard) of
Khamdj that:
Ex, 138.
Pa Ga Ma Re Gab Re Sa
A . _ «-------- [—------ —-----------

------ - *— 3 * P* m
This is not, however, the most typical oblique movement found in the rags of
Khamdj that. In a number of rags, Khamdj, Des, Jhihjhoti, Tilak Kdmod and
Khambdvti, there is a strong tendency to treat the Pa and the Re as oblique notes. In
view of the fact that Khamdj that has conjunct balance, Pa and Re should have
considerable importance, being the bottom notes of descending conjunct tetrachords,
Sa-Pa and Pa-R e. This oblique movement does not necessarily preclude the. possi
bility of emphasising the Pa and the Re, and indeed, these are the two m ost im portant
notes in the rag Des (in addition to the Sa). Yet a characteristic descending line1
prevents the direct comparison of the two parallel tetrachords Sa-Pa and Pa-R e
since the Pa is omitted in the lower tetrachord x :
Ex. 139. rag Des

£a N ibDha Pa, D]3a hla G a, R e Ga SI
4=5 U - a n i--] 1— 1-------

«! L.
---J
- X —____
L
However, the balance in Khamdj that is also ascending conjunct, Sa-M a and Ma-Nib
where the Re and Pa are the second degrees in the two tetrachords respectively. The
parallel movements can thus be expressed as follows:
Ex. 140. rag Des

Sa N ib D ha Pa Dha Ma Ga Re Ga Sa
4 ~ 7 ;li - r ...
i-----------------ii------------- ----- i
1 K.P.M. Ill, p. 251. Bhatkhan4e gives the descending line of Des as follows:
rag Des
Sa N itD ha F a, Ma G a, Ro Ga Sa
In his svarvistdr, however, he frequently treats the Pa as oblique (ibid., p. 760, variation 9):
rag Des
. . . . Sa Sa N it N it Dha Pa Dha M a Ga R e . . .
155
While the justification for this oblique movement seems apparent, there does not
appear to be any necessity for it as the tetrachords are initially balanced. It would
seem, nevertheless, to be a functional element in view of the fact that this tendency
is apparent in a number of rags. Perhaps the explanation may be as follows: at one
stage in the evolution of these rags the straight parallel conjunct descent (Sa-Pa and
Pa-Re), which is still used in Des and Khamaj, may have been characteristic. The
movement G a-Sa may have been introduced to emphasise the position of Re
as the base of the tetrachords by providing a discontinuity in the movement and
making Re oblique:
Ex. 141. rag Des

'Sa N it Dha Pa M a Ga Re., Ga Sa
This use of the oblique Re could then have been transferred to the Pa which is its
conjunct parallel in the upper tetrachord:
Ex. 142. rag Des

Re Ga Sa Pa Dha Ma
■ i
In Des these turns are often carried one step further, to the conjunct tetrachord
above Pa, and the following descent is quite com m on:
Ex. 143. rag Des

Sa Re N it Dha Pa Dha Ma Ga Re Ga Sa
These descending major thirds, G a-Sa and D ha-M a, can be particularly satisfying
in view of the ascending m ajor thirds which occur in both Khamaj (Sa-G a and P a -
Nin) and Des (Pa-Nin). The leading note, which is so prominent in this that, however,
prevents the completion of the octave in the ascending conjunct scheme. Thus, in the
rag Khamaj, the disjunct parallel ascending line (Ex. 144a) is more usual than the
conjunct (b):
Ex. 144. rag Khamaj

(a) (b)
Sa Ga Ma Pa Ni Sa Sa G a Ma- D ha N it
In Jhinjhoti and Khambavti the symmetry is once again disjunct:1
Ex. 145, rags Jhinjhoti—Khambavti
Sa Re Ma Pa D ha Sa
The rag Des is, however, unbalanced in ascent because of the leading note (R e-M a is
a minor third, while P a-N i is a m ajor third):
Ex. 146. rag Des
Sa Re Ma Pa NiS} Sa
This unbalanced ascent suggests the possibility that the Nib may initially have been
an ascending note in Des, but that it has gradually been drawn up to enhance the
effect o f the resolution in the Sa. Earlier we had mentioned that the Nib is still used in
discontinuous ascent in certain traditions, including Bhatkhande’s.2 This may be
a vestige of the original ascending line. The rag Tilak Kamod is very similar to Des
but its upper tetrachord often tends to be completely transilient, as though the leading-
note quality of the Nib could not adequately compensate for the resulting imbalance
and was then often omitted. This is an over-simplification, for in other melodic
contexts, Nib is a very im portant note in Tilak Kamod. In descent the conjunct
parallel scheme of this rag requires the Ni at the end of the lower tetrachord, as in the
following phrase:
Ex. 147. rag Tilak Kamod
Pa D ha Ma Ga Re Ga Sa Ni
The Ni is of course a very dissonant note and will finally resolve in Sa, so that the
Ni does serve as a leading note in certain phrases.3 Bhatkhande gives the ascending
and descending lines of this rag as follows :4
1 Op. cit. V, p. 261. Bhatkhapxje gives the ascending line of Jhinjhoti as follows:
Sa Re Ga Ma Pa Dlia N it Sa
However, he does not use the Nib in ascent in his svarvistar. We are following the tradition repre
sented by the songs on pp. 265, 267, 270, 271 and 272, where the ascent is pentatonic, with the
occasional use of the Ga as a discontinuous ascending note.
2 See p. 109.
3 Ustad Vilayat Khan uses the Ni quite frequently as a leading note in his recording of rag Tilak
Kamod on the accompanying record, and there is only one instance in which he makes the upper
tetrachord transilient in ascent.
*K P .M . Ill, p. 298,
157
Sa Re Ga Sa, Re M a Pa Dha Ma Pa, Sa Sa Pa Dha Ma Ga, Sa R e Ga, Sa Ni
J i-y
P
The ascending line can be seen as two symmetrical conjunct segments ending on
Pa (w), the Sa being added above just to complete the octave. Bhatkhande does not
refer to symmetry in his works and it seems very likely from the phrasing (i.e. the placing
of the commas in his notations) that he is quite unaware of this melodic symmetry. In
descent the Pa initiates the symmetry (x) and the upper tetrachord is once again
transilient. Bhatkhande gives the characteristic phrase (pakaf) of this rag as follows i1
Ex. 149. rag Tilak Kamod, characteristic phrase

Pa lî Sa Rc Ga, Sa, Re Pa Ma Ga, Sa Ni
Here the Ni of the lower register is clearly used in ascent and the phrase begins
from Pa thus stressing the octave register from Pa to Pa. The ascending N i in this
phrase can be justified in terms of inverted symmetry since the figure Pa-Iî-Sa, a
rising major third followed by a semitone with which the rag is usually introduced,
is balanced by the cadence G a-Sa-N i, a falling m ajor third followed by a semitone
(z). A further indication of symmetry is the tendency to make the lower tetrachord
transilient in keeping with the upper tetrachord 0>).
One of the characteristic oblique movements in several rags of this that is the
oblique use of Pa in descent as we have mentioned earlier in connection with the rag
Des. In Tilak Kamod, too, this is a central feature and is also found in other rags, for
instance, rag Jhinjhoti. Consequently there is a certain measure of ambiguity among
these rags. The characteristic symmetry of rag Des (x), Tilak Kamod (y) and Jhinjhoti
(z) can be seen in the following example:
Ex. 150. rags Des, Tilak Kamod, Jhinjhoti
Sa Rc N it Dha Pa D ha Ma Ga Re Ga Sa Ni
in
It will be seen that the central figure involving the oblique use of Pa is, in Des,
balanced in the tetrachord above, while in Tilak Kamod and Jhinjhoti it is balanced in
the tetrachord below. Des and Jhinjhoti require the Nib to produce symmetry, while
1 K.P.M. III, p. 298.
158
Tilak Kamod needs Nib. In many respects Des and Tilak Kamod complement each
other and there is a tendency to merge the two. Thus it is not surprising that some
musicians use Nib in Tilak Kamod1 producing a symmetry in the upper tetrachord
very much as in Des.1 In spite of the prominence of the Nib and the absence of Nib in
his version, Bhatkhande has no hesitation in ascribing Tilak Kamod to Khamaj
that, a fact which lends support to the argument that it has evolved from a rag in
Khamaj that.
The rags in K afi that have their own characteristic oblique movement. Let us
consider a simple rag in this that, rag Bhimplasi. Bhatkhande gives its ascent and
descent as follows:3
Ex, 151. rag Bhimplasi

Nib Sa Gab M a, Pa, N ib Sa Sa Nib D ha Pa Ma, Gab Re Sa
This is one of the rags in K afi that in which some musicians use the leading note
Nib, while others, like Bhatkhapde, do not. However, Bhatkhande’s simple ascent
and descent does not really do justice to the rag for musicians frequently tend to
increase the tension of the Nib in two ways: by sharpening the note slightly in ascent,
the raised note being referred to as carhi or sakari, and by introducing a very subtle
oblique movement involving a slide (portamento) from Sa to Nib. In view of this,
the upper tetrachord in ascent could be given more accurately:
Ex. 152. rag Bhimplasi

Sa X
Pa Nib Sa
.. J- |
Ail extreme example of the way Nib is used to create tension is found in the follow
ing fragment of a song from Bhatkhande’s works:4
Ex. 153. rag Bhimplasi—trital
x 2 0 3
1 According to Bhatkhande the musicians of Maharashtra use both Ni’s in this rag, and this
does occur in a few of the songs notated by him.
2 Both the rags Des and Tilak Kamod are played by Ustad Vilayat Khan in the accompanying
record and are discussed further in Appendix B, pp. 193, 195. In Vilayat Khan’s rendering of Tilak
Kamod Nib occurs as a grace note, an ornament of Dha.
3 K.P.M. Ill, p. 562.
4 K.P.M. III, p. 565, Antra.
159
By means of this slide the Nib is used as a substitute for the leading note Ni*. The
slide from Sa to Nib passes through the Ni*i and must create a momentary impression
of it. This slide is characteristic not only of the rags in this that but, in general, of those
rags in which the Nib is used in ascent. In Bhimplasi, as in many other rags of K afi
that, a similar movement also occurs in the lower of the two parallel disjunct tetra
chords in the slide from M a to Gab. In Bhatkhande’s system these are not essential
features of the rag Bhimplasi; nevertheless, they occur very frequently in the songs
found in his works, as will be seen in the following example:1
Ex. 154. rag Bhimplasi—trital

R e"^ M a ''\ Ma"N Nib'S
N it Sa Gab Ma Pa Gab Ma Pa Sa Nib D ha Pa Pa Dha Pa M a Gab Ma
..... r
I ~ L— J * r'*'
3 % 2
Ma"s Sa'">. Sa"\ Re- ^

Pa M aG abM a Gab Re Sa Nib ‘ Nib Sa Sa Nib Sa M a Gab Re Sa
j. i i 3 . 1 Q ...................... i i
Sa “s SaN . Ma*N Ma ~s
Nib Nib Sa Pa — Pa Ma Gab Ma Pa Pa Gab Re Sa
--1-- ■■J i —1.... —

r * - iLp^LJUJ..■ '-p. b — .J
Pa Ma Pa Sa Nib Nib Sa Nib Sa Sa Gab Re Sa Nib D ha Pa Pa
JL=- 1
*2
— ,‘u.-
*-----i:— --
. .-^=b r
-iiirr r
M - -
“ J->
The Gati is a first order accidental in this that and, presumably, there must be a
measure of expectation associated with this note in K afi that rags. The slide from
M a to Gab exploits this feeling, suggesting the Gati and supplying it for a brief
instant in passing. It parallels the slide from Sa to Nib in the upper disjunct tetrachord.
Thus the complete ascending line of Bhimplasi could be given as follows:
i K.P.M . III, p. 582, sthayi.
160
M a~\ , Sa
Sa G at Ma Pa N it Sa
In descent, as may be expected, the Nin is not particularly significant since its
importance lies in its function as a leading note in ascent. Consequently, the
descending movement from Sa to Nib is not usually accomplished in a slide. However,
the G at, being a balance note rather than a leading note, remains a force and often
leads to an oblique movement in the descending line as well, the disjunct symmetry
being, nevertheless, maintained:

Sa Nil* D ha Pa, Ma Pa Ma~'(jab, Ma Gab Re Sa
Although this slide from M a to Gab is not considered by Bhatkhande to be an

essential feature of Bhimplasi, it is generally thought to be an essential characteristic
in a number of other rags in both K afi and Asavri thats. These include all the rags in
the Kanhra (Kama) group and many in the Malhar (.Mallar) group. As we have
already indicated, these groups contain rags which share certain melodic features,
but do not necessarily have the same scale. For instance, the Kanhra rags generally
have the following descending cadence:
Ex. 157. Kanhra cadence
Ma 'bab, Ma Re, Sa
They differ from each other either in the ascending line, or in the upper tetrachord
of the descending line, or both. The upper tetrachord is generally treated in one of
three ways: first by omitting the Dha, thus preserving a pentatonic symmetry, as in
the rags Suha1 and Nay Id Kanhra (Ex. 158a); secondly by making the D ha oblique,
as in Sughral (Ex. 158b) or as in Sahana (Ex. 158c), both of which show essentially
the same pentatonic symmetry; and thirdly by using Dhab as an oblique note in a
slide from Nib, thus producing two perfectly symmetrical conjunct segments as in
1 This rag can be heard on the accompanying record and is discussed in Appendix B, p. 200.
11 161
Darbari {Darbari Kanhra) (Ex. 158d). It is not surprising that this last rag, belong
ing to Asavri that, is the most popular of the Kanhra rags.1
Ex. J58.2
(a) rags Suha and NaykX Kanhra.
Sa Nib Pa ■Ma~G ab' Ma Re Sa
4-----------------— — *
(b) rag Sughrai

rN iC ; I Ma Re Sa
Sa Dha Nib Pa
a —
*
--------- 9 -t : — m —
J Lf *•*
(c) rag Sahana

Sa " Nib Dha “ Nib pa ~ Ma G'ak' Ma Re Sa
$ — -----------
--------- m
<F‘
« • ■- 1
1
i____________________ 11_______________ i
(d) rag Darbari

Sa N^~"bhab Nib Pa Ma~Gab Ma Re Sa
Jl_______________ 1
In Darbari particularly, the method of increasing tension by a slide between two

notes is carried much further, resulting in an oscillation around Dhab, Gab and, fre
quently in ascent, Nib. The D h a 4 and the Gati are, in Asavri that, the first order and
second order accidentals respectively, while the Niti is the leading note. Although
these are not used as such, the oscillations clearly seem to imply the accidentals, as
will be seen in the following graphic notation of a fragment from a record of the
1 According to Bhatkhande (H.S.P. IV, p. 555), Kanhra is a corruption of Karpafa which was
in Locana’s Rdgatarafigim, and still is in Kamatic music, the equivalent of modem Khamaj that.
Thus it appears that the Kanhra rags have evolved around the Circle from Khamaj to Kafi and some
to Asavri. The rag Darbari can be heard on the accompanying record and is discussed further in
Appendix B, pp. 197, 198,
2 In these examples the lower of the two notes placed in brackets is a kind of appoggiatura, an
ornamental note connected with its upper neighbour. It appears that, from the standpoint of
symmetry, such notes may be ignored.
162
Dagar Brothers singing the rag Darbari.1 The illustration begins approximately
forty seconds from the beginning of the record.2
TIME IN SECONDS
14 16 18
fc Sa
Here we see the voice rising from the lilib to the Sa and Re, then dropping to the
Dhab and oscillating approximately between the Dhab and the Dhan. After a short
break for breath, the voice lesumes this oscillation, then rises to the Sa and drops to a
slightly sharp Nib. After another short break, the voice oscillates between the Nib
and the TsTin before rising to the Sa which is then sustained.
From this recording it seems quite clear that the range of the oscillations is approx
imately a semitone and does not extend to the next diatonic note, but to the vicinity
of its own chromatic counterpart. Thus the tension created by this oscillation is
through the alternation of the scalar note, either Dhab or Nib, with its accidental, Dhati
ot N h. In the lower tetrachord of rag Darbari there is a similar suggestion o f the
accidental Ga* (second order balance). As in the rag Bhimplasi, there is seldom a
slide between Sa and Nib in descent, while the other two slides between Nib and Dhab
and between M a and Gab remain, and the notes Dhab and Gab are made oblique
(Ex. 158d). In Dai bari there is also a tendency to omit these oblique notes producing
a pentatonic descending conjunct balance, but the Kanhra cadence (Ex. 157) invar
iably concludes these phrases.
1 H.M.Y. EALP 1291. See also Appendix B, p. 198.
. 2 SraPh was constructed without the aid of electronic instruments, and as such is bound to be
inaccurate. The only instruments used were a stop watch and a dodecachord tuned in semitones.
To estimate the extent of the oscillation, the record was slowed to half speed and the extremes of
each oscillation were compared against the semitones on the dodecachord. From this graph it will
be seen that the voice is a fluid instrument and that the notes of a scale should not always be thought
of as exact points in the spectrum of sound. Notation systems are inevitably deceptive as they
reduce the fluidity of sound into precise units, as will be seen from the following example in which
we have attempted to translate the graph into approximate musical notation:
It is interesting that in Darbari there is no suggestion of Reb which is one of the two
first order balance accidentals in Asavri that—Reu being held as steady as Sa and Pa.
Pa and Re are the bases of the conjunct parallel tetrachords o f Asavri that, a scheme
which is very important in the rag Darbari, and Ret! is here treated virtually as an
immovable note in parallel with Pa. The very satisfactory conjunct parallel scheme
of this rag, to some extent, diverts attention from the diminished fifth Reb-Dhab.
From the preceding discussion it will be apparent that oblique movement is, in most
instances, a functional melodic feature of rags, and as such is intimately connected
with accidentals, transilience and the internal symmetry of a rag.
I N T O N A T IO N
In Chapter II we mentioned the opinion of a leading musician, according to whom
srutis can only be heard in oscillations of the kind which occur in the rag Darbari.1
We have also referred to other traditions in which certain notes in a few rags are
consciously recognised to be slightly flatter or sharper than is considered to be normal.
A coherent discussion of intonation is extremely difficult, for there is no common
accepted performing standard against which these divergent intervals can be com
pared. In theory, of course, there are a number of standards available, but these are
absolutes, which do not take into consideration the context in which the sound
occurs. It is certain that unconscious variations in intonation—as between ascent and
descent, one performer and another, the beginning of a performance and the end, slow
passages and fast, to mention but a few—may be as sizeable as the conscious devia
tion, and until such time as we can separate one from the other, no conclusive word
can be brought to bear on the subject. Therefore, we do not intend to discuss this
matter in terms of absolute values, but to attempt to explore on the basis of a few
examples the motivating factors which might lead to intentional changes in intonation.
In all of these traditions, so far as it is known, the M a and Pa have a constant
relationship with the Sa. N ot only are these notes the most perfect consonances, but
they are commonly the base notes of the second tetrachord and are in this respect
secondary ground-notes. In the conjunct tetrachord scheme, however, the intonation
of the Nib (upper note of the ascending conjunct scheme) and the Re (lowest note of
the descending conjunct scheme) will be equally positive. This last is clearly apparent
in the rag Darbari as we have shown. The remaining notes are not so positively
positioned, although they all tend to be self-regulating, i.e. the intonation of one note
will tend to influence the intonation of its counterpart in the second tetrachord, both
conjunct and disjunct. This, in turn, will influence its second counterpart in the first
tetrachord. For example, the intonation of the Gab (IIIb) will tend to influence the
intonation of the Dhab (VIb) (conjunct) and the Nib (Vllb) (disjunct). The Dhab will, in
turn, tend to influence the Reb (IIb) (disjunct), while the Nib will tend to be heard in
relationship with the Mab (IVb) (conjunct) and be regulated by it since its position
1 See p. 35.
164
is fixed (see tetrachord scheme of Bhciirvi that>P* 81). This self-regulating scheme1 can
be applied perfectly to scales which are abstract entities. In particular rags, however,
the attention is focused on specific relationships and symmetries, and the imperfect
relationship between two notes, for instance the Reb and the Dhab in Darbari, m ay
not in fact be realised.
There are occasions in certain rags when unbalanced intervals are sometimes
brought into focus, and it is on these occasions that we can expect divergent inton
ation. In the rag Suddh Asavri of Bhatkhande’s system, for instance, the Gab and Nib
are omitted in ascent, leaving two unbalanced conjunct or disjunct tetrachords:
Ex. 159. rag Suddh Asavri
Sa Re Ma Pa Dhab Sa
The imbalance between Reb and Dhab is particularly noticeable in the ascending line
of this rag, the lower tetrachord comprising a wholetone and minor third, and the
upper a semitone and major third. It would not be unreasonable to expect some
feature which will ease the effect of this imbalance. In most of the songs notated
by Bhatkhande this is accomplished by means of a slide (portamento) from Nib to
Dhab. In ascent the suggestion of Nib used as an oblique grace note hints at the
completion of a parallel conjunct tetrachord scheme:
Nib' . DhalT'>
Sa Re Ma Pa Dhab Sa
In descent this slide is even more prominent, as will be seen in the following
example;2
Ex. 161. rag Suddh Asavri—cautal3
NilT"^. . NiiT>.. , Nit";:. . Nitres
(Sa 'Nl^~Dhab Dhab — Dhab Dhab Dhab — Dhab — Pa
. Nib*-^ . . Nib'x, . NiP>

Dhab Dhab Dhab Sa _ NllT w _
1 This is, in fact, the cyclic application of perfect fourths and fifths, which is sometimes also
called the up and down principle, i.e. the alternation of upper fifths with lower fourths, or vice
versa.
2 K.P.M. II, p. 357, Sancarl.
3 Cautal is a time measure of twelve units (mdtrd) similar to Ektdl.
165
Since Re is the bottom note of the parallel conjunct tetrachord scheme (Sa-Pa and
Pa-Re) its position is fixed and the slide from Nib to Dhab is parallelled in the lower
conjunct tetrachord by a slide from M a to Gab, thus producing two symmetrical
conjunct segments:1

Nil> 'n Ma' 'n
Sa Dhab Pa Ma Pa Gab Rc Sa
According to Bhatkhande there is a tradition in which the Dhab of Asavri is made

consciously flatter than the same note in the rag Bhairvi.2 In this latter rag, however,
the Dhab is balanced in the conjunct tetrachords by the Gab and in disjunct tetra
chords by the Reb. Here the intonation of the Dhab is much more stable than in the
rag Asavri, where in the essentially unbalanced ascending line Dhab is often seen in
relation to Reb. As a result Dhab is not held as a steady note but oscillates, suggesting
Dhab, somewhat as in the rag Darbari. This oscillation is then transferred to its con
junct counterpart, Gab. However, if we interpret Bhatkhande’s statement correctly, in
the tradition referred to, D ha is considerably flattened not merely in the process of
oscillation, but is sustained in this position as a steady note. This can be interpreted in
two ways: on the one hand it may be an attem pt to accentuate or draw attention to
the imbalance (which is unstable from the evolutionary point of view) so that it becomes
a prominent and characteristic feature in the rag; on the other hand it may be seen
as an attempt to camouflage the discrepancy by flattening the note D ha beyond
the chromatic limit so that it is not easily compared with its counterpart Reb. (Raising
the Dhab slightly would serve to accentuate the imbalance by making its relation
ship with Re even more dissonant.) In any case, this obvious imbalance appears to
1 K.P.M. II, p.[355. Bhatkhande gives the straight descent:
Sa N it Dhai, Pa, Ma Ga?, Re, Sa
But, in fact, he only uses this once in his svarvistar, where he generally has an oblique descent, e.g.,
ibid., p. 490, variation 2:
Sa, N it, Dhai, Pa, Ma Pa Dha!> Ma Pa, Gat, Re, Sa
Unfortunately, he does not indicate slides in his svarvistar. He does, however, show this slide quite
clearly in H.S.P. IV, p. 430:
S* ^"chat, NirDliat, P a___
2 H.S.P. IV, p. 428.

166
have been a spur to evolution, for a second tradition of the rag Asavri—often
called Komal Asavri because it has Reb in place of the Ren of Suddh Asavri—has
ascending disjunct balance:
Ex. 163. rag Komal Asavri
Sa Reb Ma Pa Dhab Sa
4 ^ u - n - 1
i------------- 1
Komal Asavrif which has probably evolved from the other, still retains the slide from
Nib to Dhab which is now no longer a functional element. This slide is balanced in the
disjunct tetrachord by a slide between Gab and Reb (whereas in Suddh Asavri it was
balanced by a conjunct parallel slide from M a to Gab), as will be seen in the following
example:1
Ex. 164. rag Kotnal Asavri—riipak tal (7 units)
Pa Dhab Ma Pa Dhab Sa - Sa ^ " k e b Gal Reb Sa Sa Nit~Dhab - Pa
# = H .J 1 1R ] — U-*
—m------ - J ------ 1 fdUr
< T n
1 i-
T
r 1-L---|;--------------------------
f c p t a * - : — ^*1— -
2 3 x 2 3 X
Pa-N . L ^ Ma"'' , G ab~\ , GalT'' ,

Dhab Ma Pa Nib Dhab — Pa Gab — Reb — Reb — Sa
J N 7 fr 44 - M y f a l : l- v p j ■ Ifc^J ...
A second frequently quoted example of divergent intonation is the Gab in rag Tori,
which is generally said to be very flat.2 Tori has parallel conjunct tetrachords, Reb to
Ma# and Ma# to Ni, provided that Pa is omitted as it frequently is.3 This omission is
usually accomplished by a slide from Dhab to Ma# in most of the songs given by
Bhatkhande, as in the following example :4
1 K.P.M. II, p. 366, sthayi.
2 The various traditions do not always agree; for instance, in the January issue of the Indian
musical journal, Saftgit (1960), G. N. Goswaml, on p. 11, gives the intervals of the rag Tori in which
the minor third, Sa-Gab, is 315 cents (slightly larger than the tempered minor third). In the same
issue, Urmila Surl, on p. 19, says that the Gab in Toyi is ati-komal (very flat), which should indicate
something flatter than the tempered minor third. This view is also expressed by Danielou in
Northern Indian Music, London, 1954, II, p. 49, when he gives a value equivalent to 275 cents for
this same interval.
3 The parallel movement of rag Toy! can also be expressed as Dhab-Ni-Reb and Gab-Ma#-Dhab.
4 K.P.M. II, p. 436, sthayi.
167
Ex, 165. rag Tori—trital
Pa Dhab Pa Dhab Ma# Ma# Pa Dhab Ma# G at—Gab Ma# Ma# Pa Pa
Ma^*Gab Ma# Pa D hat Dhat"Ma# - Gab Gab Reb - R et Reb Gat,~ R e b Reb Sa Sa
Reb" Dhs!>^^
Gab Gab Dhflt™MaJ| Mafi Dhab Dhab Ma# Dhab N i Dhab Sa Ni Dhab Pa
However, the Pa is not completely omitted in rag Tori and its presence disturbs the
parallel movement. As Reb and Ma# are the initiators of the parallel tetrachords,
Gab and Dhab are the balanced second degrees in the two tetrachords where the Pa
is omitted. When the Pa is used, however, it becomes the second degree in the upper
tetrachord and is unbalanced with the Gab. This could well influence the intonation
of the Gab, especially as it has no disjunct balance in Tori (Gab to Nin being an
augmented fifth) which would help to stabilise it. It is not surprising that the inton
ation of Gab is affected, but here, as in Asavri, it would be meaningless to give any
exact values to the deviations.
As a final illustration, we consider the rag Sri, where the Reb and Dhab are said by
some to be very flat.1 In this rag these notes are closely tied to each other, being in
perfect disjunct relationship, but neither is in a perfect relationship with any other
notes in the rag. The Dhab has no conjunct balance, G a to Dhab being a diminished
fourth, while the Reb has only conjunct enharmonic balance, Reb to Ma# being an
augmented third rather than a perfect fourth, and is not significant in this rag where
disjunct balance is characteristic. This in itself would not necessarily affect the
intonation of these two notes. According to Bhatkhande, however, the rag S ri has a
pentatonic ascent and a heptatonic descent :2
Ex. 166. rag Sri

Sa Reb Reb, Sa, Reb, Ma# Pa, Ni Sa Sa, Ni Dhab, Pa, Ma# Ga Reb,G a Reb, Reb, Sa
1 H.S.P. III, p. 46.

2 K.P.M. Ill, p. 363.
168
The ascent is unbalanced, but in descent symmetry is achieved by the temporary
omission of Sa, as in the following phrase:1
Ex. 167. rag S ri
. . . Sa, keb N i Dhab, N i Dhab Pa, M ail P a D tlab M a#G a Reb, Ia#G a Reb, G i Reb, Reb,. S a . . t
-j, . V [--I-
—*
J—------- ------ ------- — — — m-H—
-W ■ — J —;— J —J ... J j
The ascent draws attention to the imbalance between Reb-Ma# (augmented third)
and Pa-N i (major third) and it is probably this which is responsible for the deviant
intonation of the Reb.2 Once this deviant intonation is established it is transferred to
Reb’s disjunct counterpart, Dhab.
It appears that it is a combination of two factors which leads to the relatively rare
instances where divergent intonation is consciously used. These a re : (1) that under
certain circumstances the scalar imbalance in a rag becomes a focal point of that
rag, and (2) that at least one of the two notes concerned in the imbalance is not
precisely and firmly fixed in the rag. There is no indication, however, that this
divergent intonation itself has influenced, or will in the future influence the course
of evolution. It is a counter to the imbalance in a rag) but not a complete solution,
and as such may, at best, delay the evolutionary process a little.
F U N C T IO N O F N O T E S
In Chapter II we discussed the terms vadi and samvadi and indicated that the choice
and application of these in Bhatkhande’s system was not entirely based on objective
principles. The usual definition of vadi as the most important note in a rag is mean
ingless unless it is preceded by a clarification of the various functions of notes and the
establishing of a principle whereby the function of vadi can be determined. To take a
simple example, the leading note, Nit?, demands resolution in the Sa. This is an expres
sion of inherent function, but which of these is the more important, the tension
created by the leading note, or the relaxation of this tension in the ground-note?
Surely one is meaningless without the other.
In the rag Jaunpuri, for instance, Bhatkhande gives Dhab and Gab as the vadi and
1 K.P.M. III, p. 768, variation 15.
2 There are indications that this unbalanced ascent is not very satisfactory as most of the songs
given by Bhatkhande begin in the upper tetrachord, proceed to the upper Sa then gradually descend
to the middle Sa. The infrequent ascents are generally as follows:
rag Sri
Sa R et Ga R et Sa, Ma* Pa Sl N i Sa
The straight pentatonic ascent only occurs in one or two of the songs given in this rag.
169
samvadi. These notes are treated as parallel in the descending conjunct tetrachords,
Sa-Pa and Pa-Re, as follows:
Ex. 168. rag Jaunpuri
Sa ^ '^ D h a b Pa Ma Pa Dhab Pa Ma Pa Gab Re Sa
4-r—^
Dhab is an extremely dissonant note and has, in addition, a very high dynamic
function since it is the penultimate note in the descending tetrachord Sa-Pa. It there
fore acts as a powerful ‘leading note’ to the Pa. The longer Dhab is sustained, the
more urgently it needs to resolve on Pa and, correspondingly, the more satisfying is the
resolution. To call Dhab vadi is to say that in rag Jaunpuri suspense is more im portant
than resolution. In our opinion this is an oversimplification. In the following two songs,1
for example, it will be seen that the emphasis can be placed on either Dhab (Ex. 169a)
or Pa (b) without violating the rag:
Ex. 169. rag Jaunpuri—trital

(a)
Pa ' • Nib''
• i i----- 1 Ma
Pa DhabMa Pa Sa Dhab— Pa Pa Dhab Ma Pa Pa Pa Dhab Ma Pa Gab -
¥
m
(b)
Fa"'?, ■ • Nib"' , _ Ma"'V, , Sa~^ .. w
Ma — Pa Sa Dhab — Pa Gab — Re M a Ma Pa — Pa — Pa — Ga ■ Re
I f r f-
m
■ ■ Set_^ # #
— Sa Sa Sa N ib — Sa Sa Nib Sa Re Dhab Pa Dhab Ma Pa Sa
f-r r|nf r r m m
1 K.P.M. Ill, songs on pp. 649 and 665. The simplest method for determining the note which is
emphasised in these songs is to see which note is on the main beat (sani) just after the double bar
and marked with X in our notations. However, this does not necessarily apply in all instances.
170
In the rag Darbari we have a similar situation since the descending line, using
the same notes, also consists of the same two parallel conjunct tetrachords (see Ex.
158d, p. 162). Here it will be seen that Dhab is denied the role of leading note to Pa
and the symmetry is basically pentatonic, Sa-Nib-Pa and Pa-M a-R e. Thus the high
inherent dynamic function of Dhab is, to some extent, counterbalanced by the memory
of the parallel movement characteristic in this rag, where Dhab rises to Nib before
resolving on Pa. In this rag too, Dhab and Gab are extremely im portant notes and
can be emphasised (in spite of the fact that the balance is basically pentatonic and
both these notes are often omitted) as in the following example:1
Ex. 170. rag Darbari—trital
Re Re
Sa'
, 1 ~i r
Sa Sa
Sa - - Tfib Sa Re - Re Re Gab Re Gab Re Re
Gab
However, Bhatkhande gives the Re and Pa as the vadi and samvadi in rag Darbari,
thereby stressing their significance as the base notes of the parallel tetrachords.
The above examples show how arbitrary is the attempt to determine the most
im portant note in a particular rag. There is obviously a certain measure of latitude in
the treatment of the notes, so that a musician might, on one occasion, place greater
emphasis on the aspect of suspense, withholding the resolution for quite long periods,
while on another he may resolve his phrases relatively easily and sustain the conson
ant notes. Thus it will be seen that the terms vadi and samvadi, if they are to be
meaningful, must be purely descriptive and relate to the dynamic structure of the notes
in a rag.
The inherent dynamic function of notes based on consonance and dissonance, which
falls into two main patterns founded on the most frequently heard drone combin
ations, Sa-Pa and Sa-M a, describes the dynamic function of notes in the abstract,
1 K.P.M. IV, p. 663.
171
without reference to particular rags and context of occurrence. Superimposed on this
inherent structure of the scalar notes are the dynamic functions which are induced
by the characteristic melodic movements in particular rags. In the rag Tilak Kamod,
for instance, Niu is the base note of the two parallel descending conjunct tetrachords,
D ha-G a and Ga-hJin, as seen in the following characteristic phrase:

Pa Dha Ma Ga, Sa Re Ga, Sa N i,
^ J” J J — J - - J) j,-' -I
1-------------------1 i____________ i
As such, the N is is the base of the system and has something of the quality of perman
ence and repose associated with the ground-note. It has, in this instance, acquired
a much lower dynamic level and is commonly used as a terminal note of phrases in
this rag. At the same time, the dynamic level of the Sa has been raised considerably,
for there is now a sense of incompleteness upon arriving at it, the tension caused by
the need for symmetry overriding the very low inherent dynamic level of the ground-
note. Here we have a complete reversal of the inherent dynamic functions of the two
notes. Sa is now the ‘leading note’ to the IsTi.
In the rag Marva the natural function of the Sa is affected even further by the
melodic context of the rag. Here the parallelism in the two tetrachords can only be
maintained if the Sa is omitted, N i-Reb-G a and G a-M as-D ha, and the inclusion
of the Sa in either ascent or descent is superfluous. As a consequence, the Sa is
omitted for a great deal of the time and when it is used it is generally in an oblique
movement, as is exemplified by the ascending and descending lines of Marva given by
Ram N arayan:1
Ex. 172. rag Marva

Ni R et Ga M ai Dha Ni R et, Sa R et N i Dha Ma» Ga R et N'1 E h a , Sa
Here we can say that the Sa has so high an induced dynamic intensity when used in
direct ascent or descent that it is altogether omitted in these contexts.
In these two examples we have considered the dynamic functions of the two notes
Sa and Ni because they are dynamic extremes as far as their inherent function is
concerned. The melodic context may also impose its own dynamic considerations on
the other notes of the scale. Re, for instance, can become an important terminal note
as in the rags Des (see pp. 155, 156) and Darbari (see p. 171) because of their parallel
descending conjunct tetrachords where Re is the base note. While there are many
1 On the record Inde du Nord, B.A.M. LD 094.
172
instances where the dynamic function induced by the melodic context reinforces the
inherent dynamic function of the notes, there are indications that it is the interplay
between the inherent and the induced functions, especially when they differ widely,
which may be an im portant subconscious factor in determining the melodic
features of rags. Thus the natural scalar symmetry which tends to emphasise Sa, Ma
and Pa, the notes with low inherent dynamic function, is often replaced by a more
complex symmetry which places the emphasis on more dissonant notes. This may
explain, for example, why nearly all the rags in Bilaval that (which has disjunct
balance and would naturally tend to emphasise Sa and Pa) have either the accidental
Nib or Mas (or both), notes which make conjunct balance and so alter the inherent
dynamics of the notes in the rags of Bilaval that.
In any melodic context the induced dynamic function is caused by a sequence of
notes1 and the dynamic function of each note depends on the memory of the previous
notes and, to some extent, on the anticipation of the notes to follow. It cannot be
measured on an objective level. Its existence is transitory, while the inherent dynamic
function is a more permanent state which can be measured, at least in theory (see Ch.
IV). The interaction of these two, the inherent and the induced, results in a constantly
changing dynamic pattern of the notes in a melody line. If, for instance, a note, whose
low dynamic function is induced by its occurrence at the base of two parallel tetra
chords, is sustained, its inherent dynamic function will gradually reassert itself as the
memory of the melodic context in which it has occurred slowly fades. Thus the in
the rag Tilak Kamod which is the base of the tetrachord scheme will gradually appear
to become more dissonant when sustained, until finally the full inherent dynamic
function of the note will return.
We have mentioned earlier that the vadi and samvadi of Bhatkhande’s system
frequently give an indication of the parallel movements in a rag, but that there are
also instances where they do not appear to do so. This would seem to be, at least in
part, caused by Bhatkhande’s attempt to reconcile the position of the vadi with his
time theory of rags, as for instance in the rag Tilak Kamod, in which he acknowledges
the importance of the Ni, but gives the vadi as either Re or Sa, which would suit his
theory better.2 In other instances, however, it may be that the obvious parallel
tetrachord scheme of the scale is not so significant in the rag as a complex non-
tetrachordal scheme, and that the given vadi may relate to the latter. For instance,
in the rag Khamaj there are two obviously balanced alternative ascending lines, the
1 The induced dynamic function of notes is undoubtedly influenced by factors other than the
symmetry manifest in a rag. The associations often expressed (for example in The Harvard Diction
ary o f Music under Melody) of predominantly ‘rising’ melodies with tension and energy and of
‘falling’ melodies with the relaxation of this tension; of melodies moving in consecutive steps
(‘conjunct motion’) with emotion and expressiveness and melodies with wide steps (‘disjunct
motion’) with stability and reservation, etc., suggest that ascent, descent, transilience, oblique
movement and use of accidentals may all influence the dynamic functions of the notes. Similarly, time
measure (tdl) and the rhythm of each individual phrase will also be significant factors. The precise
influence of these factors has not yet been investigated.
2 See Chapter II, p. 43.
173
disjunct (Ex. 173a) and the conjunct (b). Nevertheless, perhaps the most common
ascent in the rag has no tetrachordal balance (c):
Ex. 173. rag Khamaj

(a) (b) (c)
This last ascent has a complex inverted symmetry, G a-M a-P a (semitone andâ
wholetone) and D ha-N i-Sa (wholetone and a semitone). The main tessitura of
Khamaj is from Ga to Sa and these two notes are often treated as parallel, for example,
in the characteristic terminal figures, G a-M a-G a and Sa-N i-Sa. This is well illus
trated in the following song in the figures marked x :1
Ex. 174. rag Khamaj—Dtpcandi2

Ga Ni Sa Ga - - Ga Ma Ma - Ga M a Pa Dha
j ^=~ H J- J = IJ J J ^ —1
% 2 0 3
Sa'
Ni
Sa" **
Ni _ Sa Sa
Sa"’''N i Sa Nib - Dha Pa~D ha
rr
Sa' '4---- n i--- n
Ni - Sa — Sa — Ni Sa Ni Sa - Nib — — Dha
) h a PPa
a
u r ■| T
m
Pa"
Ga Ma Sa - Nib Dha Pa Ga Ma Ga - - -
=P
m
This suggests that complex symmetry may sometimes be preferred to the more
obvious balanced movement. Bhatkhande gives the vadi of Khamaj as G a which is
the base of the non-tetrachordal inverted symmetry rather than the obvious tetra
chordal symmetry. For the samvadi he gives Ni but does not specify whether he is
1 K.P.M. n , p. 159.
2 Dipcandi is a time measure (taf) of fourteen beats Qnatrd). The rhythmic cycle is subdivided into
four groups, 3 + 4 + 3 + 4.
174
referring to Nib, the leading-note accidental, or to Nib, the scalar note.1 In fact, in the
ascending inverted symmetry (Ex. 173c) neither of these is treated as the parallel of
G a and it is, instead, the upper Sa whose induced function compares with that of Ga.
If the terms vadi and samvadi are to express parallel function, as they generally do
in most rags where tetrachordal symmetry exists, Ga and Sa should be designated as
vadi and samvadi in rag Khamaj where this inverted symmetry is characteristic in a
particular performance. However, in the descending line of Khamaj other symmetries
are sometimes brought to the fore. Of these, the two most prominent are the scalar
conjunct symmetry, Nib-M a and M a-Sa2 (Ex. 175a), and a pentatonic conjunct
symmetry which is found in the ‘catch’ phrase (pakar) of this rag3 (b):
Ex. 175. rag Khamaj
(a) (b)
Sa N it Dha Pa Ma Ga Re Sa Sa, Nib Dha, Ma Pa, Dha, Ma G a, Sa
■! j J J J J ^^1
The second of these is particularly interesting: Re, being omitted in descent, in
fluencing its conjunct counterpart, Pa, which then occurs in a turn between the two
parallel segments. In both of these the bases of the two parallel tetrachords, M a and
Sa, are also entitled to be referred to as vadi and samvadi. In the disjunct pentatonic
ascent Sa and Pa are emphasised: in the non-tetrachordal ascent, G a and Sa; in the
conjunct descent, M a and Sa. So, in the course of a normal performance of Khamaj,
the (vadr will change from one m oment to the next as the musician explores the
different facets of the rag.4 While the rag Khamaj is particularly noted for its variety,
it is by no means an exception in this respect. In many other rags, too, different notes
may be emphasised during the course of a performance and the indication of a single
vadi is often quite inadequate.
In a rag which has tetrachordal symmetry the vadi will generally be a terminal
note of melodic phrases as it is usually at the base of one of the two symmetrical
tetrachords. When the terminal note is sustained its inherent dynamic function
1 Danielou, Northern Indian Music, II, p. 168, gives Nib but many other writers, following
Bhatkhancje, do not specify which N i is samvadi.
2 In Khamaj the parallel tetrachords are necessarily ascending conjunct as Re, the base o f the
descending conjunct tetrachords, is a weak note in the rag and is often omitted.
3 K.P.M. II, p. 122. The given pakar actually begins on Nib and ends on Ga. We have added the
Sa on each end to complete the octave. There are indications that the Ga is treated as the base, and
we may once again have an instance o f inverted non-tetrachordal symmetry, N ib-D lia-M a (a
semitone followed by a major third) being balanced by D ha-M a-G a (a major third followed by
a semitone).
4 Bhatkhande is aware o f the importance o f other notes, for he says that the individuality of
Khamaj lies in the notes Ga, Ma, Pa and N i and that these are commonly used as terminal notes of
tans (melodic figures comparable to the 17th and 18th-century method o f divisions) (ibid.). Accord
ing to Sanyal, Raga and Ragini, p. 58, only 8 % of songs in Khamaj considered by him show Ga as
the dominant note. He concludes that Sa, Ga, Ma and Dha may each be emphasised. While
agreeing with him in principle, we question the inclusion of Dha in this category and can quote
several instances where Pa is the dominant note.
175
gradually returns, so that it is largely with these terminal notes that consonance and
dissonance are significant factors. In the flow of melody, however, it is primarily the
induced functions which are significant. This explains why consonant and dissonant
notes can both be used in passing without any distinction between the two. It also
explains why, in the course of melodic movement, the tonic and its fifth—the most
consonant notes—may both be omitted without creating a sense of loss and the fifth
may be omitted entirely in certain rags even though the drone will be sounding it as
a prominent harmonic of the tonic.1
On the other hand, with terminal notes, consonance and dissonance are of some
importance—not in the choice of the terminal note, but in the psychological effect of
the note when its inherent dynamic function returns. Here we can distinguish two
basic categories of terminal notes. The first is exemplified by the Sa, whose natural state
is one of repose and occurs as a terminal note in all rags, even if its induced function
is highly dynamic. This applies, in a diminishing degree, to Pa and G a when the
drones are Sa and Pa, and to M a and D ha when the drones are Sa and Ma. The
second category is exemplified by the Nib, Dhab, Ma# or Reb whose natural function
is so highly dynamic that they only occur as temporary terminal notes when in the
melodic context they have a very low dynamic function. The other notes fall some
where in between these two extremes.
There is a difference in principle between these two categories. The natural
terminals, Sa, Pa, etc., make no demands on the future; a phrase terminating on the
Sa is complete and the following phrase may either be an elaboration of the same
phrase, or a completely new melodic idea. The dissonant induced terminals do, how
ever, influence the future, for although the melodic idea may be complete on the Nib,
the feeling of completeness rapidly vanishes as it is sustained. Consequently, the
following phrase either extends this suspense, or resolves finally on the Sa. In prac
tice, there are exceptions, as for instance when certain musicians, following a conven
tion, habitually leave phrases incomplete, the mind of the listener, guided by the
drone and the memory of the cadence of the rag, supplying the final note or notes.
We have attempted to show that the concept of vadi must be related to the induced
dynamic structure in a rag and that the vadi need not be consonant to the ground-note.
This induced dynamic structure is composed of two symmetrical sections and it is
generally the base of one of these two sections which is given as the vadi of the rag by
Bhatkhande. The samvadi will then naturally be the base of the second section and
will be either a perfect fourth (conjunct) or a perfect fifth (disjunct) from the vadi,
except when non-tetrachordal symmetry is characteristic in a rag.
Ascribing the vadi to one of the two symmetrical segments naturally places emphasis
on a particular tetrachord. Bhatkhande often decides which of the two tetrachords
is more important in a rag on the basis of his time theory.2 The relative importance of
1 Musicians often perform rags such as M arva, where Pa is omitted, with the conventional
Sa-Pa drone tuning.
2 We have suggested earlier (p. 43) that, in some instances, Bhatkhande’s choice is quite arbi
trary,
176
one tetrachord over the other has an influence on the ambitus of a rag which, in a few
instances, is an im portant distinguishing feature between two otherwise similar rags.
The pentatonic rags Bhupali and Deskar, for instance, are distinguished by their
ambitus and their vadi and samvadi, which in Bhupali are Ga and D ha and in Deskar
Dha and Ga. While their scalar symmetry is descending conjunct where Pa and Re
are the base notes, the predominant symmetry in these two rags is in the two parallel
segments R e-Sa-D ha (or R e-Sa-D ha) and D ha-Pa-G a, with D ha and G a as the
base notes. This does not completely exclude the scalar symmetry and both Pa and Re
are occasionally used as terminal notes.1 In Bhupali with Ga as the vadi, Bhatkhande
states that the lower tetrachord is more prominent, and in Deskar with D ha vadi, the
upper. This is apparent in many phrases of the two rcigs, where in Bhupali the tessitura
is largely from D ha to D ha with G a as the focal point, while the emphasis in Deskar
is on D ha with a predominantly higher tessitura extending to Re in the upper octave.
Thus in Bhupali the parallel phrases tend to be disjunct, and in Deskar to be
conjunct. This can be seen in schematic form as follows:
Ex. 176.
{a) rag Bhupali
i----------—ii_________ !
(b) rag Deskar
In both instances the vadi is the base note of the upper tetrachord and the distinc
tion between the two rags manifests itself on the conscious level by a regulation of the
ambitus in the two rags. In Bhupali it is not only the lower tetrachord which is
prominent, but also most of the upper tetrachord in the lower register, while in
Deskar it is the upper tetrachord of the middle register and part of the lower tetra
chord of the upper register.2
Similarly, ambitus is an im portant distinguishing feature between the rags Darbari
and Adana of Asavri that. In the former the emphasis is on the lower tetrachord
where the vadi, Re, is located, while in the latter, where the vadi is Sa, the upper
tetrachord is more prominent. Thus the characteristic ambitus of the two rags can
be shown in the following schematic form:
Ex. 177.
(a) rag Darbari
i i i--------------------- r
Pa Dhat Iît Sa Re G at Ma Pa Dhat N it Sa
# —
* T? ^ ^ ^ 7 —‘
(b) rag Adana
1 For instance, in rag Bhupali the Pa occurs as a terminal note in K.P.M . Ill, p. 750, variations
5 and 10, the Re in variation 12.
2 N ot all musicians observe the indications o f ambitus and there are many instances among
Bhatkhanê’s songs where the distinction between the two rags is not clearly maintained.
12 177
The differentiation of two rags primarily on the basis of ambitus and the related vadi
and samvadi occurs in only one or two instances in N orth Indian classical music. It
does, however, draw attention to the relationship between the tetrachord species
and the octave registers, where conjunct parallelism becomes disjunct when the
symmetrical segments are extended above or below the middle register, and vice
versa. Thus even the simplest rag will have an element of both conjunct and disjunct
symmetry.
178
Summary
I n this work our primary purpose has been to discuss some of the underlying
principles which have helped to shape the rags of the modern period. It has been our
contention that rags have been conditioned largely by musical factors, the main
exceptions being those which have been introduced recently into N orth Indian music
and have not yet had sufficient time to develop a clearly defined shape.
Contrary to commonly accepted opinion, we feel that rags are unstable and that
change is one of their most prominent characteristics. Yet the rate of change is
controlled to a large extent by the force of tradition as well as the need to keep rags
distinct from each other.
The primary motivating force in the evolutionary process of rags is the inherent
imperfection of all diatonic scales, based as they are on the principles of consonance.
A t least one imperfect consonance m ust exist in all these scales. In Indian music,
however, this imperfection is not appreciated directly except perhaps when it occurs
in immediate relationship to the ground-note, and even in this context it is not
necessarily undesirable, as is evident from the prominence of the Mas in Indian
music. The drones tend to divide the octave into two tetrachords plus a wholetone
which may occur between the two tetrachords (disjunct species), at either end (con
junct species), or be divided and placed at both ends. The significance of a tetrachord
as a unit is comparable to that which underlies the octave as a unit where any note
can be identified with its octave. In the same way a note can be identified, to a lesser
degree, with its consonant fourth and fifth. Consequently, there is an urge to repeat
the intervals of one tetrachord in the other, just as the intervals are repeated in
different octave registers. While this repetition is possible in the two tetrachords of
one species, either conjunct or disjunct, it is not possible in both the species of any
one scale. It is the unbalanced tetrachord species of a scale which provides the
stimulus for evolution.
We have applied this principle at different levels. In connection with scales, we
have indicated that the successive correction of this imbalance led from one serial (hat
to another. The progress beyond these could only be accomplished by enharmonic
compromise leading finally to a circle of ten thats} of which one is not used at the
present time. Although Bhairav that does not belong to this circle we attempted to
show its connection with the circle and suggested that its unstable nature could easily
lead to the introduction of many new scales. The historical evidence suggests, how
ever, that Bhairav that was introduced before the Circle of Thats was completed.
179
Summary
In connection with rags, the use of accidentals and alternative notes also appears
to be associated with scalar imbalance, except in the case of those accidentals which
appear as leading notes. Alternative notes have the effect of producing temporary
balance in the rags. They can be seen as evidence for the evolution of scales, and we
have attempted to show the gradual process by which the scale of a rag might change
over a period of time without the conscious realisation of the process.
Similarly, we have explained transilience as an attempt to correct the scalar imbalance
by the omission of the unbalanced notes, and were able to derive a number of penta
tonic scales from the Circle of Thats and from Bhairav, most of which are prominent
in North Indian classical music today. Since symmetry is an im portant feature,
pentatonic scales fit quite naturally into the system, whereas hexatonic scales, which are
unbalanced, have to be modified by an oblique movement if balance is to be created.
Thus oblique movement is associated with transilience, but, we have suggested, it
may also be induced by the introduction of an accidental. It is often parallelled in the
second tetrachord, either conjunct or disjunct, and sometimes in both. Once initiated,
oblique movement tends to become extended.
The accidental need not occur as a steady note, but may instead be suggested by a
slide (portamento) between the adjacent scalar notes. This is an essential characteristic
of certain rags. An extension of this idea in some rags results in a slow oscillation
about the scalar counterpart of the accidental which extends approximately a semi
tone towards the accidental. The oscillation is a response to the imbalance in a scale,
which once again tends to transfer itself to the second parallel tetrachord. In some
rags the imbalance still remains very prominent and it is in these that divergent
intonation appears to be used. This, we have suggested, is probably an attempt to
camouflage the imbalance, and is a temporary stage in the evolutionary process.
It will be seen that the scalar imbalance may be resolved in the rags, to some extent at
least, by the introduction of accidentals, omission of notes, oblique movement, slides,
etc., each rag providing its own temporary solution in which symmetry appears to be
a vital factor. Memory and anticipation, which are involved in the perception of
symmetry, impose new sets of dynamic values on the notes during the flow of melody
giving emphasis to the notes either at the base or at the top of the symmetrical
segments. In some rags more than one symmetrical scheme may be characteristic and
several notes may be emphasised. The induced dynamic function is often in striking
contrast to the inherent dynamic functions of the notes, and there are indications
that this contrast may be an important factor in determining the melodic features
of rags.
180
APPENDIX A
The System o f Thirty-two Thats
I n presenting this system our primary concern is to show relationships between the
thirty-two thats, rather than to list them in groups based, for instance, on the number
of altered notes (vilerit svar), or on tetrachord species (ahg). The underlying principle
in our system is that scales which differ in only one note are directly related to each
other, and it is this relationship which the system shows.
For this purpose the gamut of twelve semitones can be represented by an icosahe
dron, a regular three-dimensional form with twenty sides meeting at twelve points,
each of which would then correspond to a semitone of the gamut. Any two opposite
points may be selected as the Sa and Pa which then become the immovable axis of
each scale. The direction of the axes of all the thirty-two scales are made consistent
within the system. The icosahedron can now be thought of as being in two p a rts: one
with the Sa and the five points surrounding it, the other with the Pa and the five
points surrounding it. The two sets of five alternative notes can then be distributed,
one around the Sa and the other around the Pa. These alternative notes may be
arranged in a number of different ways. In the following scheme we have placed the
Re G a Ma# Dhab and Nib around the Sa, and the Reb Gab M a D ha and Ni around
the Pa in such a manner that each note is directly opposite its alternative (see Fig. 2,
p. 182). Although, at first sight, this might seem arbitrary, there is justification for
this arrangement. Each of these series is made up of six wholetones, and thus the two
groups balance each other. Further, there are 1 1 0 perfect fourths or fifths in either of
these groups, and, as the two series of notes interlock to form the icosahedron, the
interlocking notes are largely perfect fifths.1 This can be seen in Fig. 1, p. 182.
As a result of this arrangement the more consonant musical scales can be shown by
the interlocking notes from the two groups, the Sa and Pa being implicit in the system.
On paper this icosahedron can be represented in a view looking down the Sa-Pa axis.
In this view, eleven of the twelve points can be seen. The only point not visible is
the other pole representing the Pa, which lies directly below the Sa. In Fig. 2
the notes connected to the Sa by a solid line belong to the Sa series, while those
1 There are two exceptions: Ma# to Reb which is really a diminished minor sixth, and Mato Re (the
last of the Pa series and the first of the Sa series) which is a major sixth. Both these are discussed
in Chapter III.
181
Fig. 1
(Sa) Re Ga Ma? Dhat* N it
- > - - n- « - - ‘' - - H I
Re Ga Mail Dhab N it
Dha Ni Rat G at Ma
(Pa) Dha Ni R et G at Ma
^ * [>» ■■■ {?» ■ ■ * ■
Pa
connected to the centre by dotted lines belong to the Pa series and are connected
to the Pa at the other pole:
Fig. 2
Gab
D ha
Ma Re'
Pa 'Sa
Re Ma
Dha
Ga
With this icosahedron as a basis we can now show each of the thirty-two scales.
Of the ten alternative notes shown on the perimeter of the diagram above, only five
are used in any of these scales. These can be represented by lines extending outwards
from the five appropriate points. A scale with all the alternative notes in their higher
position, Sa Re G a Ma? Pa D ha Ni Sa, is shown by lines extending from the five
points in the lower half of Fig. 2, while a scale with the alternative notes in their
lower position, Sa Reb Gab M a Pa Dhab Nib Sa, is shown by lines extending from the
five points in the upper half. In this way, the thirty-two scales are represented, each
with its own combination of arms extending from the icosahedron.
The icosahedron can readily be represented on paper as a decagon (a ten-sided
figure), with the understanding that the Sa and Pa are taken for granted in these
scales. Now we can connect the scales which differ from each other in only one note
182
by linking the appropriate arms. This will become apparent from the following
illustration in which five scales are shown in their relationship to Kalyan, the scale
with the alternatives in their higher positions:
Fig. 3
Re Mcf
Re Re Ma
.Ga
.Dha Ni
Each of these five scales is connected to four others, the process continuing until
all the thirty-two scales are tied together forming a complete closed system. This
system, as we have indicated earlier, is three-dimensional, so that the scales are not all
on the same plane, although we are obliged to show them thus on paper. It therefore
appears that the connecting arms of two scales frequently go through a third, whereas,
in fact, the third scale is in a different plane, either above or below the line of the
other two. To obviate this difficulty, we have drawn these connecting arms to circum
vent the intervening scale in Fig. 4 on the next page which shows the complete system.
In this figure all the decagons have a constant alignment and each pair of alter
native notes is placed opposite each other, as in Fig. 2, retaining the same positions
for all of the scales; thus the Gab and Gati will always be found at the top and
bottom respectively of each scale, even when they are not specifically indicated.
In the centre of the diagram there are two scales, one at the top of the system (No. D l)
Sa Re Ga Mas Pa Dhab Nib Sa, the other at the bottom (No. D2) Sa Reb Gab M a Pa
D ha N i Sa. The arms of the latter are shown in dotted lines. These two scales being
furthest removed from the outer ring of consonant scales are, of course, the most
dissonant in the system.
The photograph facing p. 184 shows a model of this system. F or practical reasons
scales are represented here not on icosahedrons as discussed above, but on dodecahe
drons which are closely related to the former. While the icosahedron has twenty sides
meeting at twelve points, the dodecahedron has twelve sides meeting at twenty points.
Thus in the model the semitones are represented not by the points, but by the sides,
each of which is painted a different colour, and the notes of each scale are shown by
arms extending from the centre of the appropriate sides. In all essentials, however,
the model corresponds to the diagram.
The scales in this system have been given numbers for easy reference. The method
183
Fig. 4
A2
BIO”
B9n
mivn.
rvir
A9 H
11
^mV!i
JA3
JILbl
C3
TV 31 llb
A8
C6
B6
finally adopted is designed not only to facilitate the location of these scales, but to
show certain relationships, particularly those associated with the outer circle (here
designated as the ‘A ’ group) which contain the most consonant scales and are ex
tremely important in N orth Indian music. The five scales related to an ‘A ’ group
184
Model o f System o f Thirty-two 7'hats
The System o f Thirty-two Thdts
scale are the two adjacent ‘A ’ group scales, the corresponding ‘B’ and ‘C’ group
scales, and a second ‘B’ group scale two integers below. These relationships can be
expressed in symbols as follows, where n stands for the number of the scale under
consideration i1
An is related t o : An + 1
An — 1
Bn
Cn
B/i —2.
Similarly the relationships of the ‘B5 group and the ‘C’ group scales can also be
expressed in symbols as follows:
B72 is related t o : A n
An + 2
C n- 1
Cn + 3
if n is odd, with D1
if n is even, with D2
Cn is related t o : A n
B« + 1
Bn- 3
Cn + 3
1 If the resultant number in the following examples exceeds ten, the related scale is obtained by
subtracting ten. If, however, the resultant number is less than one, the related scale is obtained by
adding ten.
185
APPENDIX B
Description and Notation o f Recorded Music

Examples
T h e re is always a considerable gap between musical theory and musical practice

which can perhaps never be completely bridged. It is with a view to reducing this gap
that we felt it imperative to include a sound recording in this book. The record
contains short illustrations showing the characteristic melodic features of eight of the
rags which have been discussed in the main body of the text. These are not repro
ductions of the melodic phrases quoted in the text, but are in the form of aldp as one
might hear in a real performance. The reader will immediately see the enormous gulf
between the bare skeleton notations in the text and the highly elaborated melody line
as it occurs in practice.
The performer has not seen the text of this book, nor was he required to conform
to any theories therein. As a result it was inevitable that there would be some diver
gence from the notations of rags in the text which were based largely on Bhatkhande’s
works. These divergences will give the reader some idea of the deviations in the
precise interpretation of rags to be found among musicians. The recordings have
provided new source material to which some of the theories expressed in the book
could be applied and provide an independent check on their validity.
Three factors influenced the choice of musician for these illustrations: (1) that he
should have had a traditional musical training in the gharana system; (2) that he
should not be influenced by Bhatkhande’s theories; (3) that he should be one of the
leading musicians of India.
Ustad Yilayat Khan fulfilled these conditions perfectly. Recognised as one of the
foremost musicians of India, he is descended from a family of musicians who trace
their ancestry back about three hundred years to the fabulous Tansen, the leading
court musician of the Moghul Emperor Akbar. Vilayat K han’s grandfather, Imdad
Khan, and his father, Enayet Khan, were both brilliant musicians and pre-eminent
in their generations. Vilayat K han was brought up to be a professional musician and
186
Description and Notation o f Recorded Music Examples
began his concert career before he had reached his teens. He has had a traditional
musical education and has been trained in singing as well as in sitdr technique. His
repertoire includes songs which are said to have been composed by Amir Khusraw
at the beginning of the 14th century. Nevertheless, his creativity refuses to be circum
scribed by conservative orthodoxy. This is in the best Indian musical tradition where
pride of place has always been given to those great musicians who have left their
mark on Indian music through their innovations. In this respect Vilayat K han is very
much like his father and grandfather who also played a very prominent part in the
reshaping of instrumental music.
Vilayat K han’s most outstanding innovation is the introduction of the gayki (vocal)
style on the sitdr in which the phrasing and ornamentation are based on voice
technique. This has only been possible because Vilayat Khan is also an accomplished
singer. The voice is regarded in India as the most versatile of instruments and virtually
without limitations, while all other instruments are restrictive. W ith plucked stringed
instruments, such as the sitdr, the sharp sound of attack and the subsequent decay
of the sound are unavoidable and legato passages are extremely difficult to produce.
In this Vilayat K han has succeeded to a remarkable extent. This has necessitated a
modification o f the sitdr as well as an adjustment in playing technique. The gayki style
depends to a great extent on the increased use of portamento effects which are achieved
by the sideways deflection of the melody strings, as will be apparent from the
accompanying record in which there is extensive use of ornaments and grace notes
produced in this manner. This style has now gained wide acceptance and has in
fluenced many Indian musicians.
Among Vilayat K han’s other innovations one is of particular importance in con
nection with this record. He has evolved a new method of tuning the playing strings
of the sitdr, which necessitates the removal of the bass, third and fourth strings,
usually of copper or bronze, and the replacement of these by a single steel string.
Given overleaf are Vilayat K han’s new tuning of his six-stringed sitdr1 and two
traditional sitdr tunings.
In Vilayat K han’s method the im portant notes of the rag being played often
determine the tuning of strings 3 and 4, The traditional method of tuning is based on
the Sa-Pa or Sa-M a drones. Vilayat Khan generally keeps this basis, but adds to it.
If the rag has a prominent Ga, he will often tune string 3 to this note. This, as we have
suggested earlier (p. 72), is merely an extension of a natural occurrence as G a is a
prominent part of the sound spectrum produced by the Sa-Pa drones. Vilayat Khan
uses this tuning in the rags Yaman and Sankra on the accompanying record. It should
be mentioned that the third string is only occasionally sounded in the drone and
the listener may not hear it every time the drone strings are plucked. There are also
some occasions when Vilayat Khan tunes this third string to the Gab, for instance
in the rag M iyd k i Tori where Dhab and Gab are the two im portant notes. In the rag
1 N ot including the sympathetic strings, usually 11 or 13, which remain more or less as in the
traditional method of tuning.
187
String Number
Ma Pa or Ma Pa or Ma
Traditional
Tuning 1
or
Pa orMa Pa or Ma gva
Traditional
Tuning 2
or
A ltern a tiv e n o tes A lternative notes

Ma Ga JGabtMa Pa Re Pa Ma D M D h at Sa
Vilayat Khan’s
Tuning
Gujri Tori there are great problems with conventional sitdr tunings since the rag has
neither Pa nor Ma*. Here Vilayat K han tunes his third string to Gab and the fourth
string to Dhab, the two im portant notes of the rag.1 In the traditional tunings, one is
obliged to tune the secondary drone to Pa, a note which is omitted from the rag, and
whenever the Dhab, which is the vddi in this rag, is sustained, its extreme dissonance
against the Pa drone demands resolution on it, and of course this is forbidden in the
rag Gujri Tori. The same applies to the hexatonic rags in Marva that, Marva, Puriya
and Sohni, which have Mas and no Pa. In rag Marva, which can be heard on the
accompanying record, Vilayat K han tunes his third string to Ga and his fourth to
Dha, this in spite of the fact that Ga is not an im portant note in this rag (Dha is the
vddi of this rag). The justification for this is that G a is not only a prominent harmonic
of Sa but also of Dha, and would be heard in the drone in any case. This does not,
however, influence Vilayat K han’s rendering of the rag and he gives due importance
to Reb which is the other important note in rag Marva (yadiin Bhatkhande’s system).
The result of these additional drones is that greater emphasis is placed on the
important notes of a rag in Vilayat K han’s rendering than is usual. This is exemplified
in the musical examples on the record where Vilayat K han generally ends his illustra
tions, not on the ground-note, but on one of the im portant notes of the rag. This is pri
marily for purposes of demonstration and in a normal concert, which is usually
accompanied by a tambura drone, the drone notes, in which Sa would be predominant,
would continue to sound after the melody line had been concluded.
Perhaps the greatest virtue of this kind of tuning is that the added drones provide
a richer texture against which the musician can improvise. It also enhances the
multiple modality of the rag and gives greater emphasis to the im portant notes
1 This tuning can be heard on the record of Gujri Tori performed in duet by Vilayat Khan and
Bismillah Khan, E.M.I. ALP 2295.
188
which may then tend to sound even more like the ground-note. A case in point is the
rag Mdrvd on the accompanying record where the Sa-G a-D ha drone may suggest
to the Western listener the first inversion of the minor chord with Dha as the root and
thus the tonic of rag Mdrvd. But to one who is familiar with Mdrvd, the D ha drone
seems to intensify the feeling of the rag and does not in fact suggest that Dha has
replaced Sa as the tonic. This tends to corroborate our earlier suggestion that all the
notes evoke, to a greater or lesser degree in relation to their importance, their own
modal series (p. 73). D ha being a prominent note in Marva evokes its modal series
even when it is not supported in the drone, and when it is, this merely accentuates
this series which is already an inherent feature of rag Mdrvd.
These experiments with added drones are, by and large, to be expected and are
developments within the musical system. Vilayat K han’s tunings have been adopted
by a number of other sitdr players and it is now even possible to purchase six-stringed
sitdrs designed specially for this tuning. While the extra drone notes on sitdrs is a
relatively modern innovation, this type of tuning of tamburds appears to be somewhat
older. In recent times five-stringed tamburds which allow greater variety in tuning have
become increasingly popular. It remains to be seen whether this will have a lasting
effect and will change or modify some of the fundamental elements of the musical
system.
N O TA TIO N
The notation of the musical extracts on the record is intended to help the listener to
follow the music, but is not intended as a score for a performer. The extreme subtlety
of the ornamentation virtually defies accurate notation, even when the recording is
studied at half its original speed. This is inevitable since a characteristic feature of the
music is the manipulation of the dying sound, often suggesting notes which in a cold
analytical light are not actually present. In notating this kind of music, one is faced
with a dilemma; whether to attempt to notate it absolutely objectively as a machine
would (assuming a sufficiently precise machine were available) or to take into
account human factors, such as the occasional discrepancy between the musician’s
conception and its realisation. We are concerned, in this work, primarily with the
musician’s conception of rags, but since this conception is not on a verbal level and
can only be conveyed through the medium of musical sound, we have endeavoured
to notate the music as accurately as possible. Nevertheless, the careful listener will
find one or two instances where we have made allowance for the human factor.
Following the practice in the text, the musical examples are notated in both Western
staff and in Indian sargam which, for convenience, has been abbreviated further to
the initial consonant of the Indian note-names. It must be remembered that the use of
C for Sa does not indicate its actual pitch in the recording, which is somewhere
between C# and D, but is again a matter of convenience. Since the shape of the melodic
189
line is our main concern, only very approximate duration values are indicated.
Alap is of course in free time and there is no fixed pulse against which durations can
be measured. Thus we distinguish only three duration levels: quick notes, which are
indicated by quaver tails (whether as grace notes or as full quavers), medium notes,
which are indicated as blackened circles without tails, and sustained notes, which
are shown as white circles, i.e. as semibreves.
One of the greatest difficulties we have faced with this notation concerns the
concept of grace note as a non-essential embellishment of a melody. From the stand
point of a work of art one could argue that there is no such thing as a non-essential
embellishment; each nuance is essential in the production of such a work. From the
standpoint of a particular mg, however, it would be reasonable to say that some of
the nuances were obligatory in a particular context of a rag, while others could be
thought of as ornaments because they were non-essential in their context of occur
rence, as for instance, in the following phrase in the rag Kedar, where the main notes
seem to suggest the straight descent Ga Re S a:
But in the descending line of Kedar G a is an oblique note and the descent G a M a Re is
a characteristic feature. In the phrase above the notes shown as grace notes may be of
the same duration; nevertheless, one of the M a’s between G a and Re is obligatory,
the others we could classify as ornaments. Of course, it has sometimes to be an
arbitrary decision as to which one of these is obligatory, but without such an indica
tion the reader could well be misled into the notion that in Kedar Ga may be used as
a direct descending note. Thus it would seem more reasonable to show one of the
M a’s (perhaps the last one) as a full note, while keeping the quaver tail to show that
it has the same duration as the grace notes:
A particular advantage of this kind of notation is that it provides a bridge between

the notations that are found in theory books and in the main body of this work,
which are concerned primarily with the main (obligatory) notes, and the actual music
with all its elaboration.
In the Indian notation the quick notes are written in slightly smaller letters and
placed slightly above the main notes. Sustained notes are indicated by commas after
190
the notes, following traditional Indian practice. The duration of the quick obligatory
notes is not shown in the Indian sargam, but the Indian reader can easily recognise
these in the Western staff by their tails.
An important aspect of sitar technique is the sideways deflection of the melody
string producing portamento effects or slides {mind, mir). It is possible for a musician
to play, in this manner, a number of notes over a range of about a fifth with just one
plectrum stroke. All the notes produced in this manner with one stroke are linked by
a slur sign:
4 i
All notes produced by deflection are naturally slides, although some are less
noticeable as such than others. The more obvious slides are shown by a line joining
the two notes:
Various other devices are used on the sitar, partly to prolong the duration of a plucked
note, partly to extend the apparent range of the slide, or just as a device to produce
ornaments. These are produced by manipulations of the left hand—sliding from one
fret to another, plucking with the second finger while stopping with the first, or
hammering with the second finger to the next higher fret. These are also linked by a
slur, but have in addition accents below the appropriate notes to show that the attack
on these notes is more acute than that produced by the deflection of the string:
Ma Pa Dha Ma Ga Ma
The periodically struck drone strings are shown;
J J
A sustained note slightly shaken in pitch, but not rising or falling as far as its neigh
bouring note (either chromatic or diatonic) is shown:
Ma
191
Indeterminate rise or fall, before or after a note, is notated:
M U SICA L EXAM PLES
The musical examples on the accompanying record were recorded in London in 1968.
The original recordings also included the ascending and descending lines (aroh-avroh)
of the rags. These have, however, been omitted from the record, but can be seen in
the notations.
The alaps played by Vilayat Khan on the record each have two sections, sthayi
and antra. These sections are usually separated by a melodic figure called mohra,
which signifies the conclusion of any section of the dlap.1 The two parts, sthayi and
antra, are generally associated with tessitura and vary from rag to rag, sthayi usually
being limited to the low (mandr) and middle (madhy) octave registers, while the antra
begins about half-way in the middle register and extends into the upper {tar). It will
be seen from the notation of rag Des that the sthayi extends to the Re in the upper
register, presumably because the characteristic descending symmetry of the rag begins
at this point (see p. 158). In rag Tilak Kamod the sthayi extends only to Dha, the
uppermost note of its characteristic symmetry (see p. 158). In rag Darbari the range
of the sthayi is from Pa to Ma. This includes the characteristic Kdnhra cadence (see
p. 161) and its parallel below. In rag Sitha the characteristic symmetry (p. 162)
extends to the upper Sa, and this is, in fact, the extent of the sthayi in the recording
(see p. 200). This also applies to rag Sahkra, where the characteristic overlapping
symmetry also extends to the upper Sa (see p. 139). In rag Yaman the range is from
f a to Pa which provides for the two symmetrical disjunct tetrachords Pa M a Ga Re
and Sa hft Dha Pa. In the other two rags, Mdrvd and Kedar, the correlation is not
nearly so clear;2 nevertheless, the evidence seems to suggest that there may be a
connection between the sthayi of a rag and the range of its characteristic symmetry.
If this is so, then the antra would, in general, begin at the second half of the sthayi
and carry the symmetry a tetrachord higher as, for instance, in the following
example:
1 Two of the other well-known sections are jor and jhdla. In the notations the mohrd is followed
by a single bar and is specifically shown in the notation of rag Darbari.
2 In rag Mdrvd the range of the sthayi is from Ma# to Reb, only an enharmonic fifth (actually a
diminished minor sixth). This naturally precludes any symmetry. In rag Kedar the tessitura of the
sthdyl is mainly from Sa to Pa, whereas one would have expected a range from Pa to Ma to include
the symmetry Ma Re Sa and Sa £>ha Pa. In the beginning of the antra section Vilayat Khan quite
unexpectedly starts with this descent into the lower register. This suggests that he may have felt the
need to complete the characteristic symmetry.
192
sthayi
a----------------— ^
"1 l---------------- 1
Ma Pa §a Ma
t---------------- 1 i _______ i
antra
Thus the sthayi and antra show both the disjunct and conjunct aspects of the
characteristic symmetry of a rag. While this might be the basis underlying the concept
of sthayi and antra, musicians, in fact, do not limit themselves to just these three
tetrachords. It remains, however, that anything beyond these three tetrachords is a
duplication since both aspects of symmetry are contained in just three tetrachords.
We will now briefly consider each of the eight rags as they have been played by
Ustcicl Vilayat Khan.
RA G D ES1
In the aroh-avroh of this rag Vilayat K han gave one ascending line and two, alter
native, descending lines. His ascending line is unbalanced due to the occurrence of
Niti as leading note (see p. 157). His first descending line shows the simple descending
conjunct symmetry of the that which concludes on Re (a). This descent is not usually
heard in alap (perhaps because it is too simple), but is quite common in faster passages.
The second descending line shows the characteristic turn, Pa D ha Ma, but does not
show its upper conjunct parallel, Sa Re Nib. Nevertheless, this descending line also
has two symmetrical conjunct tetrachords Nib-M a and M a-Sa (b). In the alap,
however, one can clearly see the upper conjunct parallel, with its turn around Sa, for
instance on line 2 (6) which is then balanced on line 3 (&), with the turn around Pa, and
is finally echoed on line 4 (b) with the t o n around Re—the parallelism being particu
larly striking in the last two. But behind this one can also see the larger symmetry,
Sa Re Nib D ha Pa and Pa D ha M a Ga Re, by extending the first two of these (adding
b'). Here Pa and Re are seen as the base notes of the symmetrical segments and these
are the two m ost im portant notes of the rag.
Another characteristic symmetry used by Vilayat K han is the conjunct parallel
Pa-Nib for the ascending minor third Re-M a. This is clearly seen on lines 9 and 10
(c). The figure Re Nin Sa, a descending minor third followed by a rising semitone (on
lines 5 and 7), may be interpreted as an inverted echo of this minor third in the figure
Re M a G a . . . (a rising minor third followed by a descending semitone). In the record
ing one can also find the use of D ha and G a as discontinuous direct ascending notes,2
the former on fine 3, the latter on line 7 (d).
1 Rag Des is discussed on pp. 38ff, 155ff.

2 Discontinuous direct ascending notes are discussed on p. 41.
13 193
Aroh-Avroh
) I I- I i L _ JL, i i i --
w ♦ •
M 5 GR M P W S ^ i R Sv ^ _ s Nb ]> P H G -R
Alternative Avroh
X j i
t> » •
S Nb X> P
Aldp
i i i i
if -ft*1 T P c t'1 v T
li he
rm c & m*'R m p n n I n $
5 USJ4S
1 1 1 j; i
r.
s R.
In s n# nr ? tR -r M 6 MS - P
T:
$=>=£
^ P D%p D p NVP S N b DNb
KJb,,NV p I p ! ? p i> V :sm
i i
-X -
r- ^ i' f } i .v 1- V * 1"
i
Â rV - _GC X ? l ----- ^
X r g
". b ■
V - s ■ s^ M
v£> W ? < p - j?
i LA i
52b • m ±$=F
>>> > rr *71 J’ <(• -e-
V *
Nb D P D v M S MG R G R s s s M—S S S 5 N fi
1 J L • e-
£T-]> (j- >“5------

5 M- 5 * ** T? S-p M P N N N S H U R
■9 u
1 I i J 1 _
"£ *:rfX
(S - ci>
^■R R ^ S R P N s ‘i ! ? G * tR ? r<5 M-5
i T T . I
i, * tr ^ t g t T (TPf.j tT*".-1> • mLtA' r » [ H « ty » a.
H *P t M ^ S G-R iV ? I r- P ? H 6G \t« d”
V S NS 63 15 W1* D PvTI s M r m g G R
1*. J kv -+ • ^ ^ kW-*■ *
X)-p N ^ S R M -p N S vJR M P N 5 R-M G
F MG N6 N6 5 NPDNTj, D P M G
P D
y = - . . . f j '; ir. . n . . : r ; ^ i 'y * > «— #-
R S MSR Mj 6 3 - 1 r —r 7 M_S G*_R -> G MS R y

_ J ____________ 1 1 1 1 1 I " '
SsE
ii o' 1 ,0 •O-
N , s *
SMS
' - N S , S , yJV ^ 5 N * *R M G R
r a g t il a k k Am o d 1
In the aroh-avroh of this rag Vilayat Khan gave one ascending line and three, alterna
tive, descending lines. In the ascent Vilayat Khan clearly used Nin as a leading note,
with the result that the ascent is asymmetrical as in rag Des. Nevertheless, the tendency
to omit the Ni in ascent and to make the upper tetrachord transilient is quite
apparent in the alap, for instance on line 6 (a). Of the three descending lines, the
second is perfectly symmetrical (b) and is virtually the same as that discussed on p. 158.
It would seem that Vilayat K han is conscious of this symmetry, for in the first descent
he stops on Sa, a note short of the complete tetrachordal symmetry. In the second
he completes the symmetry which takes him to Ni. In the thud he adds an extra note,
resolving the phrase on the Sa. A particularly interesting feature of these three
descending lines is that the upper segment is identical and, although the lower seg
ments vary, the number of plectrum strokes remains constant at four, providing a
kind of rhythmic parallel to the upper segment. This suggests the possibility that
rhythmic symmetry may, on occasion, be a temporary substitute for melodic sym
metry. In the aldp, line 2 (c), he shows yet another melodic variant of this segment
which also maintains the rhythmic unity of four plectrum strokes.
All these descending lines draw attention to the transilient upper tetrachord and
1 Rag Tilak Kamod is discussed on pp. 157, 158, 172.
195
to the characteristic symmetry (b), which places Ga and Ni, the two most im portant
notes of the rag, at the base of the two segments. In Vilayat K han’s tradition, as in
Bhatkhande’s, Nib is not a scalar note in Tilak Kamod, presumably because extensive
use of this note will eventually lead to the characteristic symmetry of rag Des. Yet
Vilayat Khan does use Nib as a grace note, for instance, on line 4 (d) which is more or
less parallel to (e) on the following line, suggesting that there is an inclination towards
ascending conjunct symmetry. The slight hint of Nib obviously adds something to
the rag as it satisfies the needs of symmetry. This could be given as an example of the
modern concept of vivddi,1 that is, a note which is generally not permitted in a rag,
except when played or sung by a great musician where it seems to enhance the melodic
features of the rag, rather than to add a new feature.
Aroh-Avroh
L_J L
r+ r
5y S "R, M P N S K S N-S P I> ,v_M G S, H G ^_S
' ---------------- i------------ b — --------------- 1
Avroh Alternatives
_i L i ___ L
35 r ... f?
$
P 2> . \ . M Cf P G .v S N N S P P X h G v.S N S
A lap
(i ‘ y'f :.....
^ •* ■
. %
- j f- - o„ rr H
Ar ‘'
- y
-O; i » ♦ • ^ >* > > -e-
S P N S P ^ G ^ S P PMG Sv^_G 6 , S , s j j g3
r i _________ i______i__________ i _________ i ___________________ i
. « r , fT .
_L i -I
r P FI > ~ 7 = ^ * -•—— t—*-----------
P M 6 „„ 6 M ^ ^ G _ ^
^ M M 6 Tt ^ ^ 6 - R R M P N ,
1_
c~- *-• eft
■ e-
* - U _ NSM
8vet bat*A
S -" 0 S -"* p—s —p , M P N S gmgp - g' r ~ m
.1 Jt ,> 1 a 1 it
fit D * ■*
> > >
G
* R G
/
G -
VR
*
&
R S M -
v S N N S
G•
G-' CR V^P
•. * RS
M G ^
NS*
_jL_ I I L
$ • • *
a •:
1 1.
--■■ ■; J j j > ....J r . , j7 7
N DN
K G **
P M G
SR MPB MP NS U Q4-S VS P l> M G S R M V'
• • r r
PMG
R—M G R
r r» f " 7^ = ^ = r v
R 6 N S
RAG DARBARI1
From the standpoint of symmetry, Darbari is perhaps the most obvious example.
Yet the ascending line given by Yilayat Khan does not show the ascending conjunct
symmetry, Sa Re *^~G ab M a and M a Pa Dhab Nib, which is sometimes quite

explicit. His ascending line is composed of two asymmetrical parts. However, descend-
1 This rag is discussed on pp. 162ff, 171,177.
197
ing conjunct symmetry is clearly apparent in his descending line (a). W hat is perhaps
not so obvious is that elements of this symmetry appear to be carried a stage further.
The basic symmetrical segments are Sa Dhab Nib Pa and Pa Gab M a Re, where Pa
and Re, the important notes of the rag, are at the base of the two tetrachords. The
first three notes of these segments appear to have a further conjunct balance Re X^Tib
Sa (fine 1, b), and there appears to be a tendency to repeat the descending minor
third, Nib Pa, a further conjunct tetrachord above, Gab Sa (line 10, c).
The most remarkable feature of this rag is the extended oscillation (andolan) of
Dhab and Gab. These, we have suggested earlier, extend not to the adjacent diatonic
notes, as is shown, for instance, in Bhatkhande’s notations, but to their chromatic
counterparts. Vilayat K han’s rendering of the rag substantiates this thesis. The
oscillation of the Nib is basically melodic since it occurs in ascent and is associated
with a slightly sharpened (sakari) Nib, which increases the dissonance of the note so
that the effect of resolution in Sa is enhanced. This occurs on lines 7 and 8 and is
marked with a + above the note. While the oscillation of the Nib is basically melodic,
it must be remembered that there will be some tendency to transfer the oscillation
from the Gab to its disjunct counterpart, Nib, which may account for the extensive
oscillation of this note. But the Nib does not always lead directly to Sa, ISTib Re Sa
(as in the aroh-avroh) being equally characteristic. This could be interpreted in melodic
terms as a withholding of the Sa. It could also be interpreted as the inverted parallel
of Re Nit' Sa.
There is a fairly strong pentatonic tendency in Darbari. It may seem curious that
the Dhab and Gab, the oscillation of which is such a characteristic feature of the rag>
are the very notes which tend to be omitted, for instance on line 6 (d) and line 10 (e).
In Darbari Dhab is out of balance with Re, one of the im portant notes of the rag. The
oscillation of Dhab, we have suggested earlier, results from this imbalance and is then
transferred to Gab to produce conjunct symmetry. We have also suggested earlier
that the unbalanced notes are often omitted (pp. 125,126), thus the Dhab tends to be
omitted. This feature is also transferred to Gab which is then a second order omission.
Aroh-Avroh
J JL J L ji L „.l
fW f r r r .
* j
1 1 Ji
S Nb 5 HV DV ?!j>v V M P M M 4bM 5
_____*— —1 a ___ ^__ ii_______________ .a — __^ ’
198
Alap
X J L X X A L
fr-tr
-e-
Qvft bass*.
£*“ ^r s r ' n1’ »^ v - . s , s__, ^ ’>*
b 1 ' 1 T 1
p , p ^ * V
b _ 6 ^ R cf1R s _ s b >e L0bsl> GÛ gb gbgljgb

*-Gp v "
L I I I i
t ~ il-l °
S N>-S S N b- S P =bR° k/ S
— ^ — ■ ... ■ , TTFnoh
1 I 1 ~T
—pv-7,------ffflpp h---£-*—£—£-a ... »- -5 - r t r K .---- >-- « pPv tA* ■
6b „ p P P , MP _, D1* Bbubj^pb^Db
yM , R sjl P v , M! T ’ p P __/ P ^Jr
,, ..J„ „1.- ^ - ^ - _____ __________
i L
. ,.----- .. 1_____ l . J U , - „ L i.
/ b ^ b A ' - f i n M h t . ' tS:' c ...... — P-«-^ > 0 ""' - 1"
^b , b »*>T>b&»l N'DNSP
fc,brA,fc P V ^ V p , p n ,bPNo M- g“,
p 'C p m
Lîr. ^-------——J — —=-I . "
J 1____ 4 i i 1 _____ 1 V * ------ = - J ^ e
_j L _ J 1 1 1_________ i
T -■yr "c r
RS N1 6N L
V -M R b S ' J , R
R A G SU H A 1
This is a hexatonic rag (Dha being omitted) which is ascribed to K afi that in Bhat
khande’s system. It is one of the Kanhfd group of rags, as is Darbari, whose character
istic feature is the oblique Gab in the lower descending tetrachord, ^^~Gab M a Re
Sa. In ascent th&rdg is pentatonic and has disjunct symmetry (aroh-avroh, a). Since
Gab is an oblique note, the descending line is also pentatonic. Here the conjunct
symmetry can only be appreciated if the turn around Gab is ignored (aroh-avroh, b).
Thus this rag has elements of both conjunct and disjunct symmetry. There is an
indication that there is a tendency towards disjunct symmetry, which is characteristic
o f K afi that, in the descending line as well, since Dhaii occurs occasionally as a grace
note attached to its nearest diatonic neighbour, Nib. This results in a kind of turn,
for instance on line 1 of the aldp (c), which is paralleled in the lower disjunct tetra
chord, as on line 4 (c). The emphasis, in this rag, is in the upper tetrachord, and there
fore on the ascending line,2 so that M a and Sa, the final notes of the ascending
disjunct symmetrical tetrachords, are the two im portant notes in the rag. As in rag
Darbari, the Nib is slightly sharpened when leading to the Sa, and is shown by a +
sign above the Nib, for example, at the beginning of line 4.
Aroh-Avroh
(L_^l J-J. J L 1 i.
•O'
t w b S M-61 N P M S
t ^ -2rs?- b—IL.
Aldp
I 1 L
E m
i' I *
5 -o-
5^5, P
_ J L _____
P M P MP 1? P M
■p _N bPS P P,
1 This rag is discussed on pp. 161, 162.

2 Vilayat Khan concludes his aldp with an ascent to the upper Sa.
200
RAG M ARVA1
We have suggested earlier that the scale of this rag is asymmetrical unless Sa is
omitted. This tendency is clearly evident in Vilayat K han’s aroh-avroh of Marva {a).
This leaves two conjunct symmetrical segments which can be seen either as TsTi Reb G a
and Ga Ma# D ha or as Reb Ga Ma# and Ma# D ha Ni. The former is the basis of the
rag Puriya which is, in scale, identical to Marva. In Puriya G a and Ni, which are the
base notes of the two symmetrical segments, are its two im portant notes. The latter
scheme appears to be the basis of Marva. We should then expect Reb and Ma#, the
base notes of the two symmetrical segments, to be the im portant notes of the rag.
In fact, Reb is an important note, but Ma# is not. This is not surprising in view of the
fact that if Ma# were sustained as a terminal note, its extreme dissonance would
require resolution in Pa, the note which is omitted in the rag. In Vilayat K han’s
rendering of the rag, the octave register Reb-Reb is clearly emphasised and Reb
sustained as a base note, for instance on line 10 (b), but the Ma#, which is often the
initial note of a phrase, for instance on line 4 (c), is never used as a terminal note.
Going across this scheme is a second scheme based on the tessitura D ha-D ha, for
instance on lines 3 and 4 (d). Here D ha is not the base note of any symmetry within
the rag and appears to function as an independent ground-note, on a par with Sa,
which of course need not be justified in terms of symmetry. The successive intervals
of Mdrvd (without Sa) from D ha give the common pentatonic scale of rag Bhupali
(see p. 130). Similarly, if we consider the successive intervals from Reb (again omitting
Sa), we get the pentatonic scale of the rag Mdlkos (see p. 130). The evocation of these
two well-known pentatonic rags is a basic feature of Mdrva>but this is not appre
ciated on a conscious level.
It will be clear from the recording that Reb and D ha are the two m ost im portant
notes in Mdrvd and these are recognised as such by Bhatkhande, in spite of being an
augmented fifth apart. Some musicologists, confusing the modern and ancient
concepts of vddi and samvddi, are disturbed by the fact that the im portant notes o f a
rag (modern concept of these terms) are not a perfect fourth or a perfect fifth (ancient
concept), and give D ha and G a as the vddi and samvddi of this rag.2
1 See pp. 83, 114, 115, 188 for a discussion of this rag.
2 See f.n. 2, p. 44.
202
Aroh-Avroh
1 1 1 1 1 i t i l 1 L 1
qFi ______
*±fc
'* ' ».....
* J ) T h ”Rb 6 » * J I I ft* S !n ! D-> M# S t&, S ‘ N* D V * ; R 'j. S
Alap
DJ 1 1 1 1 J. 1. JL i 1_
m ■0VPOSto.
SnS S N -Rb ^ D, p *#d -î* D yV
-D
1 JL i ____ Ik! 1 1 I____ J,
is s:
*++*s-'* :*-----~ ■" •—Tji -©■ »** *
s s S. , _ l. „b -S M 5 N S
s 5 . ,5 S .. e
N vN S
_L 1- i 1
J p \ —Jr
'■»— O'".....B~
a D E Z ZZ 3E ^ « -r t E
• ''■I
J ) v— }
I _________ L _ T
3=3;
$» r:.J. • T '' ST ‘S'-"' Nj. ^
v G^ N'^M) Vt '■M# P MSD N'
X A_ T - JL
»• O -i1 i * t *t~ —^ «~E=^=fc
P n ^KSNNm# * u^s n s n 5- jj.4^4
Si Si D P ,
b
J 1 1 J 1 1
=fc£ r.t-Lt-, ;.•.■■ t * — i
N 5 N-R S ^ ■R*----,
5 M . NiNj S N_# N.
1 i 1
-■ . ...q : ^ ..r . r r > b t
_ D ,, S N 5 N _
D ^-D t>, PVM^ PNT>*f "•'b'* NS NP N—i&A} NS P P
1
u !■ - f
— *------■------ • ---- • -- | ---- & ~z
J> N M* MfrI) G G M# T** T?4, ■

5 »n.J"
M". 11h ^ ^ SvRr<,
>bM# W
*#;^G<i>
, £J'M-P s> t
^G-M*
203
if f £ Qr ; i ; j i - i jTp J, 17 p ~ ^ ~ J V ^ ^
n^ h n * d n1) d I m#dm* m*? gm*s Rb- M

*MV ,D p m^ n
RAG YAM AN1
The scalar symmetry of rag Yaman is descending conjunct, Sa-Pa and Pa-R e. When
Sa and Pa are omitted, which is often the case particularly in ascent, the symmetry
becomes disjunct, Ni Re Ga and Ma# D ha Ni. This disjunct symmetry is quite obvious
in Vilayat K han’s aroh-avroh of this rag (a). In descent, however, one would expect to
see the conjunct symmetry emphasised. This is not so readily apparent; in fact, in the
descending line played by Vilayat Khan, there is still evidence of disjunct symmetry,
although it only extends over a major third (b). In the course of his alap, however,
there is a perfect example of this descending conjunct symmetry (line 8, c) where the
Re functions as the base note of the lower of two symmetrical segments. This sym
metry would appear to be quite unconscious, since Vilayat K han gives no indication
of it in his aroh-avroh.
In Vilayat K han’s present rendering of Yaman, Sa and Pa are not omitted to any
great extent,2 although the tendency is clearly there. Equal emphasis appears to be
given to the omission of Ma# and Ni. We have pointed out earlier that there is a ten
dency to omit the unbalanced notes, in Yaman Sa and Ma#. Each of these may lead to
a second order omission, Sa to the Pa, and Ma# to the Ni. This leaves four pairs of
notes which are all related in terms of symmetry, as follows:
Ni-Re^s— ►Mas-Dha (disjunct parallel)

G a-Pa (conjunct parallel)
^ ^ D h a - S a (conjunct parallel)
By combining these in pairs, Vilayat Khan also achieves an oblique symmetry,
D ha Sa Ni Re and Ga Pa Ma# Dha, for instance, on line 2 (d).
1 This rag is discussed on pp. 82, 126, 187.
2 On Vilayat Khan’s record of the rag Yaman, E.M.I. ASD 2425, the omission of Sa and Pa is
more evident.
204
It is also interesting to note the slight suggestion of Matt which occurs on several
occasions as an inflexion of the Ga. This is, of course, the conjunct parallel of Sa, and
can be regarded as another instance of vivddi.1
In a rag such as Yaman the vddi and samvddi would naturally fluctuate, depending
on whether the ascending disjunct segments were being emphasised or the descending
conjunct. In the former, G a and N i would qualify as the two m ost im portant notes
(Ivfi Re Ga and M ai D ha N i); in the latter, Pa and Re (Sa N i D ha Pa and Pa M a
G a Re). It is not surprising that there are a number of divergent opinions, and
Banarji mentions just these four notes which qualify as vddi in the different tradi
tions he has encountered.2
Aroh-Avroh
1 1 L_J I 1 L_J l I 1 1 1 I 1
§* ^ o -J: <! " - *■ J'»l . .
-*•
VN \R 6, vj) N S N vj) P G S
l—
------ J 1 f i----- 1 1----F~-... <
Alap
1 i 1___1_________ 1_______ I______ 1 1. L_
■j j j j »— =— — titit
0Y* s ... Sl1w 9N. D S SMSM NSIV1SN
s "vD J R ’S H * r>~ S
1 1 —I i ___L
fu u r f f*— I 1
N, S M 5 N S ND 1r t g VR ^ , V-G P 1{ S y G R ^
1 I 1 L
fh M T * ■j C j ‘« k f'j'f J’ f r ; .r . j ^
SSN? V s N —1> P
.C -- =^1
P M*
M^-G JR
X 1 X 1 X X J-
i r* ii - f f ^ J
ma
P R~G
......
4 i L i ___L
h » . J t J'J'x r . " f t - t-Cxl. i .'X f i--------* -•—■-
>pM*t> P M*P M#. ^ " ' ' * 3 N DP M* 5 S .J X r. S T> R 6 —C
1 1 1 1 1 _L L
l r£:
S N ? S.N S* R. s
RAG Sa n k r A1
We have discussed earlier (p. 131) the pentatonic version of this rag which has the
notes Sa Ga Pa D ha Ni Sa. In the hexatonic version Re is added and provides the
disjunct balance to Dha. Thus the ascending line has a measure of disjunct symmetry
(a). We had suggested, largely on the basis of Bhatkhande’s notations, that the char
acteristic symmetry involved a turn whereby S a-G a-P a was balanced by Pa-N i
(disjunct) and D ha-Sa (conjunct). This symmetry is also apparent in Vilayat K han’s
rendering of Sankra, for instance on line 3 of the alap {b). Equally prominent seems
to be a tendency to produce a mixed conjunct-disjunct symmetry in the descending
line, where D ha Sa Ni D ha Pa are balanced by Ga Pa (conjunct) G a Re Sa (disjunct),
as at (c). The invariable use of Re between these conjunct and disjunct elements can be
interpreted as a device to disguise the discontinuity and perhaps to provide a rhythmic
substitute for the Mas which would have been necessary to complete the conjunct
symmetry.
In Vilayat K han’s rendering of rag Sankra, Ni is a very prominent leading note, so
that when Sa is approached from Dha, as in the characteristic turn, Pa N i D ha Sa, it
is invariably followed by a descent (for example on line 3, following b), as it is not felt
to be resolved. The extensive use of Ni in turns (mordents) around Sa is paralleled
1 This rag is discussed on pp. 131, 139.
206
by a similar treatment of its disjunct Mas (which remains basically a grace note) in
turns around Pa, as will be seen on lines 1 and 2 (cl). The hint of Mati, largely as an
inflexion of Ga (line 3, e), suggests that Sa also has some influence on its conjunct
counterpart.
Aroh-Avroh
G • » •
G P R G R SD
RAG KEDAR1
The ascending line of rag Kedar is quite extraordinary and consists basically of two
transilient tetrachords bridged by a disjunction, Sa M a, M a Pa, Pa Sa. This is, of
course, an over-simplification as all three notes, Sa, M a and Pa, act as tonal centres
around which turns (mordents) are based. These are perfectly symmetrical, as in the
ascending line given by Vilayat Khan (a). Thus Kedar has elements of both conjunct
and disjunct symmetry, so that the melodic figure Sa N i Re is echoed a fourth above,
M a G a Pa, as well as a fifth above, Pa Mas Dha, the last necessitating the use of the
accidental, Mas, for reasons of symmetry.2 In descent, however, the symmetry is
basically disjunct and pentatonic, Sa D ha Pa and M a Re Sa, as in Vilayat K han’s
descending line (6), although an element of conjunct symmetry is found in the occur
rence of the Nib—sometimes just as an inflexion of D ha—suggesting the figure D ha
Nib Pa, as on line 6 (c), which provides symmetry for G a M a Re in the lower tetra-
chord (e.g. on line 5, c).2
W ith the exception of this conjunct symmetry, which is not yet well defined, all
the other symmetries in the rag place emphasis on the three most consonant notes,
Sa, M a and Pa, which are im portant terminal notes. However, only two can be given
as vddi and samvadi, and it is usual for M a and Sa to be recognised as such. This gives
importance to the ascending line in which M a and Sa are the final notes of the
1 This rag is discussed on pp. 103, 104, 152ff.

2 Bhatkhancje’s own version of this rag shows virtually no symmetry, K.P.M. Ill, p. 118:
Sa Ma, Ma Pa, Dha Pa, N i Dha Sa Sa NI Dha, Pa, MaJ Pa Dha Pa, Mai;, Ga Ma Ro Sa
However, the majority of songs he has notated show more or less the same elements of symmetry
as are found in Vilayat Khan’s rendering.
208
disjunct tetrachords and a performance of this rag usually opens with the character
istic jum p Sa-M a, as in Vilayat K han’s recording.
Aroh-Avroh
1 L_L A l J t l I
3 :
$3C
j? vm
*pm* n p M,
A L _ _ _ 1 i L
f f. j . --r
M s^ G 6 ^ GM, M, M * P
h M , M & G M G- P
1 1____L
f I 'm j 'J j i fj j / . J1J ’ j f
V f PM* V* P"* "aM GMR \ ‘ *S I-'l HS"“ 6PN*Ili

u. 1 L_ 1
V » .. r r . - 'g y p J =v
PM* M*l6 SH^ M
M
_L 1 i_
_L i _____L
s i-v - is * -fi ♦
R6 r4£ -D ? ^*0 M*r s *5 R S*s
“s se s ;
l x i 1 1 i_
3pc -G-
M G .J
MG?
~
P
M *P „ * _
M D -i
P M*L
^ P
M* D
vJP — M:L— ^ M
. MG s MG P M
^
^
M G,
M
r
MR
'
s ,
4 1 1___ _ i
r ^ F - ; J T p - t? r s-- r iM >* >-p. HM x
"* M • H SM S V*P p "pm* b f *‘ b "V ’‘"'’ “ ’’p
i 1 I i 1
tj ■' t.
M* M *FM *P P * _ # P P M* P M* P . . f r „ PM * M *F M * * b f iM
_________ ^ v M P V M*D M y
14 209
Description and Notation of Recorded Music Examples
_ 1 ..I 1
It * t ' f
N s
M s m &mg D ™*p f pc s , S " su p
i - 1 i L
U
" tr--* r, ' rJr'—u" cr~i
N6^D h^ ns^dn ^ nSN "R V*s n 5- d N
J p N> D-P M1
i ____ 1
? ) «" , K f » iCi'. j * £ ? ; n jT l 4 f7 ^ i' ■ «
l 1
P 77 7
R S H S PM"
■R ^ 5 j1 M*P 3 I^ D v*J» "vSt SM*S
1 1 1 i i I L
f\ l . n<T R i....f .
J J V - J f * i1
M6M6 P M f D
GM- M—R V*S l ? s j
i
•R s * s * S 0 S~ ^ M
210
Select Bibliography
1 European Languages
Ahmad, Nazir, ‘Lahjat-i Sikandar Shahi’ in Islamic Culture, Vol. 28, 1954,
Ahmad-ul-Umri, The Lady o f the Lotus, tr. L. M. Crump, London, 1926.
‘Allami, Abu’l-Fazl, Aln-i-Akbarl, tr. H. Blochmann, Calcutta, 1873.
Bake, A. A., ‘The Music of India’ in The New Oxford History o f Music, London, 1957.
Bake, A. A., ‘Indische Musik’ in Die Musik in Geschichte und Gegenwart, Allgemeine Enzyclopadie
der Musik, Bd. 6, Kassel, 1957.
Bhatkhapqle, V. N., A short historical survey of the music of Upper Lidia, Bombay, 1934.
Bhatkhande, V. N., A comparative study o f some of the leading music systems of the 15th, 16th, 17th,
and 18/A centuries,—A series of articles published in Safigita, Lucknow, 1930-1.
Danielou, A., Northern Indian Music, Vol. II, London, 1954.
Dunk, J. L., The Structure o f the Musical Scale, London, 1940.
Farmer, H. G., A History of Arabian Music, London, 1929.
Fox Strangways, A. H., Music o f Hindostan, Oxford, 1914.
Gangoly, O, C,, Ragas and Raginis, Bombay, 1958.
Grosset, J., ‘Inde: Histoire de la musique . . . ’ in A. Lavignac, Encyclopedie de la Musique, /, Paris,
1921.
Halim, Abdul, Essays of History of Indo-Pak Music, Dacca, 1962.
Helmholtz, H. von, Sensations of tone, trans. A. J. Ellis, London, 1875.
Jairazbhoy, N, A., ‘Svaraprastara in North Indian Classical Music’, Bulletin o f the School o f
Oriental and African Studies, Vol. XXIV part, 2, 1961, pp. 307-25.
Jairazbhoy, N. A., ‘Bharata’s concept of Sadharana’, Bulletin of the School o f Oriental and African
Studies, Vol. XXI, part 1, 1958, pp. 54-60.
Jairazbhoy, N. A., with Stone, A. W. ‘Intonation in present-day North Indian classical music’,
Bulletin of the School of Oriental and African Studies, Vol. XXVI, Part 1, 1963, pp. 119-32.
Jeans, James, Science and Music, Cambridge, 1937.
Kaufmann, W., Musical notations of the Orient, Bloomington, 1967.
Mangahas, R., The Development of Ragalaksat.ta, A thesis submitted to the University of London
for the degree of Doctor of Philosophy (Music), June, 1967.
Mirza, M. W., Life and Works o f Amir Khusraw, Calcutta, 1935.
Popley, H. A., The Music of India, Calcutta, 1950.
Powers, H. S., ‘Indian music and the English language; A Review Essay’, Ethnomusicology, ix,
January, 1965.
Ranade, G. H., Hindustani Music, Poona, 1951.
Roy, H. L., Problems o f Hindustani Music, Calcutta, 1937.
Roychoudhury, M. L., ‘Music in Islam’, Journal of Asiatic Society, Letters, Vol. XXIII, No. 2,
1957.
Sanyal, A. N., Ragas and Raginis, Calcutta, 1959.
Smith, V., Oxford History of India, Oxford, 1958.
Sprenger, A., El Mas'udi’s historical encyclopaedia, *Meadows of Gold. London, 1841.
Willard, N. A., Music o f India, Calcutta, 1962.
Zuckerkandl, V., Sound and Symbol, Music in the External World, New York, 1956.
211
Select Bibliography
2 Sansk rit a n d other I n d ia n L an gu ag es
Ahobala, Sangitapdrijata, Hathras, 1956.
Banarji, K., Gita Sutra Sara, Calcutta, 1934 (Bengali).
Bharata, Natyasastra, Kashi Sanskrit Series, No. 60.
Bhatkhande, V. N., Hindustharii Safiglt Paddhati, Vols. I-IV, Hathras, 1951-57 (Hindi edition).
Bhatkhande, V. N., Kramik Pustak Malika, Vols. I-VT, Hathras, 1954-9 (Hindi edition).
Bhatkhande, V. N., (under pseudonym, Catura Pandit a Vi§nu Sarma), Srimal-lak$yasangitam,
Poona, 1934.
Damodara, Sangitadarpana, Hathras, 1962.
Kantha, Rasakaumudi, Gaekwad’s Oriental Series, 1963.
Locana, Rdgataradgitii, Bihar, 1934.
Matanga, Bnhaddesl, Trivandrum, 1928.
Narada, Sangltamakaranda, Baroda, 1920.
Narayana, Hridaya, Hridayakautaka, Bombay, 1920.
Narayana, Hridaya, Hridayaprakasa, Bombay, 1920.
Patvardhan, V. N., Rag Vijnan, Vols. I-VTI, Poona, 1962 (Hindi).
Ramamatya, Svaramelakalanidhi, Annamalai, 1932.
Sarngadeva, Satigitaratndkara, Adyar Library Series, 1943.
Shah, Ibrahim ‘Adil, Kitab-i-Nauras, Poona, 1956 (Dakhani).
Simh, Pratap, Saiigitsar, unpublished manuscript, quoted by Bhatkhande (Hindi).
Sivan, Maha-vaidya-natha, Mela-raga-mdlikd, Adyar, 1937.
Somanatha, Ragavibodha, ridyar Library Series, 1945.
Tomwar, Man Singh, Man Kautuhal, unpublished manuscript, quoted by Bhatkhande (Hindi).
Tulaja, Saiigitasdrdmrita, Madras, 1942.
Vasant, Rag Kos, Hathras, 1962 (Hindi).
Venkatamakhi, Caturdaydiprakdsika, Madras, 1934.
Vitthala, Pundarika, Rdgamdla, Bombay, 1914.
Vitthala, Pundarika, Rdgamanjarl, Bombay, n.d.
Vitthala, Pundarika, Sadragacandrodaya, Bombay, 1912.
212
Index
Abdul Halim, 25 Amwadi, 43, 44, 45

Abhog, 30 Appoggiatura, 162 n.
‘Abidin, Zain-ul-, 18 Aroh-avroh, 24, 37 n., 38-9,45,192
Acal svar, 32 dual implication of, 39-40
Accidental^) Asdvrl That, 47, 48, 95, 113, 54-64 pass.
as balance notes, 111-13,115-21 accidentals in, 114,118,121
as exaggerated vibrato, 103 balance and imbalance in, 81
as grace note, ornament, 103,104 hexatonic rag of, 147
as leading notes, 113-21 pentatonic derivative of, 130,132
as temporary change of scale, 119-20 Ascending, descending movement (line), 40,104-
as vivddl, 45, 196, 205 10pass.
in Ancient Indian music, 114 alternative, 105
in evolution, 97, 104-6 discontinuous, 41,105
in rag classification, 52-3 see also Aroh-avroh, Transilience (directional)
in Western Church music, 79 Ascending, descending note
suggested in slides, 165-6,167 criterion of, 40,104
summary of, 116-19 direct, incomplete or oblique, 40, 41, 97-8,
that, diagram of, 121 104—10 pass.
Adarahg, 21 discontinuous, 41,105, 109 n.
Ahobala, 21 n. see also Aroh-avroh
see also Safigltapdrijdta Astai, see Sthayi
Akbar, 19-20, 186 Atikomal, 167 n.
Aldp, 28-9, 37 n., 186,190, 192, 193-210 Atimandr, 34
All Akbar Khan, 49 n., 53 Atitar, 34
Allah, Faqlr, see Rag Darpan, Man Kautuhal Augmented fourth, 49, 50, 51, 56
Alternative notes, 36-7,49-50 see also Tritone
in classification of rags, 52,120,121 Augmented fifth, 58
see also Accidentals Aurangzeb, 21
Ambitus Avroh, see Aroh-avroh
of plagal modes, 73 Audav, 37, 45
of rags, 177 see also Transilience, Pentatonic series
see also Tessitura
Amir JChusraw, 17,17 n,, 20 n,, 187 Bahadur Shah, 21
Anisa, 20,44 Balance and imbalance, 82, 85, 111-3, 115-21,
Ancient melodic system, 16, 20-2, 23 n., 44, 56, 123-4
90,114 see also Symmetry
Andolan, 35,45,198 Banarji, K. M., 42 n.
Afig, 181 see also Gita Sutra Sara
see also Purvaiig, attrang, Tetrachord Bake, A. A., 24, 94
Anhemitonic, 130,137 Bard Ehydl, 30 n.
Antara Ga, 114 Basrl, 27
Antra, 30,125, 125 n., 192-3 Baz Bahadur, 19
Index
Bhairav That, 47,48, 54, 63, 84 of primary thats, 91
accidentals in, 112,119,121 see also Satrivddi, Vivddl
balance and imbalance in, 86-7
hexatonic rags of, 148 Dadra tdl, 142, 142 n.
links with Circle of Thats, 93-7, 101, 133, 134 Dagar Brothers, 163
origin of, 94, 96 Damodara Misra, see Sahgitadarpatia
pentatonic derivatives of, 132,133 Danielou, A., 25
plagal inversion of, 88 Descending movement, see Ascending, descend
Bhairvi That* 47, 48,113, 54-64pass. ing movement
accidentals in, 113-14, 118,121 Descending note, see Ascending, descending
balance and imbalance in, 81 note
pentatonic derivative of, 130-2 Deval, K. B., 25
transilience in, 85 Devnagri, 13
Bharata, see Nafyasdstra Dhaivat, 32
Bhatkhande, V.N. Dhanasri mela
approach, 24, 35, 35 n. of Locana, 92, 94
method of notation, 37 n.-38 n. ragas of, 94
time theory, 43, 45, 61-4, 99 Dhrupad, 18,19, 20, 30
works, 23, 25 Diatonic, 36 n., 46,104
Bhava Bhatta, 21 Difference tones, 75, 76
Bhog, 30 Diminished fourth, 58
Bilas Khan, 19, 20 Diminished fifth, 56-7
Bilaval group (rags), 107, 108 in Lalit rags, 50-1
Bilaval That* 22, 33, 46-8, 54-64 pass. in rag Bhairvi, 49-50,73
accidentals in, 117, 121 see also Tritone
balance and imbalance in, 79, 80 Dipcanditdl, 174, 174 n.
evolutionary stages to Khamaj that, 107-9 Disjunct tone, 77, 78
hexatonic rag of, 146 divided, 83
oblique movements in, 152-4 Dissonance, see Consonance-dissonance, Vivddl
omissions in, 128-9 Drone, 28, 65, 90 n.,
pentatonic derivative of, 129-30, 132 instruments, 28
Bismillah Khan, 71, 188 n. secondary Ma, 51, 6 6 -7 ,11-4-pass., 136
Bow-harp vina, 90 secondary Pa, 65-7, 72-6 pass.
Bundu Khan, 23 n., 71,148 n. subsidiary, 69-72 pass.
tunings, 187-9
Cadence, 41,114,120,127,161 Drut, 29
in classification, 52,121 Dukar, 28
Cal svar, 32 Durbal, 149
Cathi, H4 n., 159 Durrat al-Taj, 94
Caturdai.idiprakasika, 20,48,100 Dynamic function of notes, see Notes
Catura Papdita, see Bhatkhande, V. N.
Characteristic phrase, see Pakat
Chromatic, 36 n., 48,49, 51,103,104, 163 Ektal, 29
Enharmonic compromise, 57-9,137-8
Cikari, 28
Circle of fourths, fifths, 59, 60
Circle of Thats, see That Finalis, 73
Ciz, 30, 37, i53 Fixed pitch, 32, 91
Classification of rags, 51-3,121 Flat, 33
see also Mela, That Fox Strangways, A. H,, 23,29 n.
Clements, E., 25 Fundamental, 72,75
Composed piece, 30, 31
Consonance-dissonance, 34, 78,164 Gamak, 35
ancient Indian concept of, 43-4 Gamut, North Indian, 34, 36
influence on function of notes, 65-9,171-3 Gdtidhar, 32
214
Index
Gdndhdragrdma, 90 n. Important notes, see Amsa, Sairivddi, Vddi
Gangoly, O. C., 25 Improvisation, 31, 31 n.
Get, 30, 31 Imrat Khan, 85 n.
Garni mela Intervals
of Locana, 92, 94 ‘pure’, 68
ragas of, 94-7, 101 tempered, 59, 68
Gayki style, 187 see also Intonation, Notes
Ghardnd, 23,106, 114,186 Intonation, 25, 34—5, 70
Ghulam Ali Khan, Bare, 35 n. divergent, 166,167 n., 168
Gita Sutra Sara, 23, 42 oscillating notes, 35,163-4
Goswami, K., 98 ‘pure’ intervals, 68
see also Sahgita Sara self-regulating, 164,165
Grace note, 104,190 tempered intervals, 59, 68
Ground-note, 32, 55, 65, 90-1,172, 188 variable, 34, 70
as tonal centre, 208 see also Just intonation, Sruti
identity of octave with, 74-5 Islam, and music, 17
primary and secondary, 72-6, 164
semitonal shift of, 101
shift of, 71-3 Jahangir, 19
Jairazbhoy, N. A., 31 n., 34 n.
Jdti{s)
Harmonium, 27, 34 Ancient, 16, 44, 55-6, 90-1
Harmonic series, 72 n. Dhaivati, 91
Helmholtz, 65-6 Modem, 37
Heptatonic series, 37,122-3 Jhdld, 29
as ‘parent’, 122,130 Jhaptdl, 29, 141,141 n,
balance in, 127 Jones, Sir William, 22, 94 n.
reconstruction of from the pentatonic, 134—5 Jor, 29
Hexatonic series, 37,122-4 Jugalbandi, 27 n.
balance in, 127,145 Just intonation, 16, 50 n.
directional, 149 as ‘pure’ intervals, 66-9
in Mama, Tori and Purvi thats, 83-5, 113, ‘pure’ scale, 68
145-6
of rags, 145-8
Hindusthani Sangit Paddhati, 23 Kdfi fhat, 47-8, 54-64pass.
Hridaya Narayana, 99 n. accidentals in, 113-14, 117, 121
Hridayalcautaka, 20, 99 n. balance and imbalance in, 80
Hridayaprakdsa, 20,92 n., 93 n„ 99, 99 n. hexatonic rags of, 146,147
Husayn Shah Sharql, 18 oblique movements in, 159, 160,161
omissions in, 128
pentatonic derivative of, 130,132
Ibrahim ‘Adil §hah, II, 18, 22, 22 n. Kajri, 71 n.
Ibrahim Shah Sharqi, 17-18 KakaliNi, 114
Imana mela, of Locana, 99 Kalydn That, 47, 48, 54-64pass.
Imbalance accidentals in, 117,121
correction of, 82, 85, 87 balance and imbalance in, 81-2
leading to accidentals, 103, 111 connection with Mama fhat, 99-100
leading to divergent intonation, 166-9 oblique movements in, 153-4
leading to omissions, 82-3,124 pentatonic derivative of, 131-2
leading to oscillations, 162-4 Kan svar, 103
leading to slides, 165 as grace note, 104,190
spur to evolution, 167 Kdnhrd rags, 146-7,161, 192
Immovable note, 49, 60,164 Karnata mela, of Locana, 92, 93,100
see also Acal svar Khali, 29, 29 n.
215
Index
Khctmaj Thai, 47-8, 54-64 pass. Matrd, 29
accidentals in, 103, 113-4, 117, 121 Mela{s), 91
balance and imbalance in, 80 interpretation of Puravd, Sarafiga and Megha,
evolutionary stages to Bilaval {haf, 107-10 93
oblique movements in, 155-9 Locana’s, 92-3
omissions in, 128, 129 South Indian, 20, 46, 48, 123
pentatonic derivative of, 130,132 Melakarta, 48, 54
Khiljl, ‘Ala’ al-DIn, 17 Melodic movement, as Yarn, 38
Kitab-i Nauras, 19, 22 n., 94 n. see also Ascending, descending movement,
Komal, 33 Oblique movement, Symmetry, Transilience
Kramik Pustak Malika, 23 Melody line, 27,28
Khurdak, 28 Metre (Meter), see Time measure
Rhydl, 17 n., 18, 21, 30 Mind, 191
baya, 30 n. Mnemonic drum syllables, 29
Modality, 73, 188-9
Lahjat~i Sikandar Shdhi, 18 see also Tonality
Lahra, 31 n. Mode(s)
Lai Khan, 20 ‘B’, 57, 73, 91
Lay, 29 Dorian, 22
Leading note, 86, 109 n., 156-7, 169 Ecclesiastic, 55-7
as accidental, 113-9 Greek diatonic, 56
descending, 170-1 plagal, authentic, 73
influence on scale, 115, 136 plagal inversion, 86, 88,134
in Western music, 68 serial, 54, 56, 74, 90-1, 96
to terminal note Ma, 115-16 Mohra, 192, 199
Locana, see Ragatarangini Mordent, 206, 208
Long-necked lute, 90 Movable note, see Cal svar
see also Sitar Mridang, 28
Mughal period, 19, 27
Ma!1drif-ul naghmdt, 24 Muhammad Nawab ‘Ali Khan, see Ma'arif-ul
Madhy naghmdt
register, 34,45,192 Muhammad Reza, 22,22 n.
tempo, 29 Muhammad Shah, 21
Madhyam, 32 Muhammad b. Tughluq, 17
Madhyamagrama, 16, 21, 56 n., 90, 114 Mukhra, 30 n.
Majid CMasit) Khdrii Gat, 30 Murcchana, 20
Major scale, 33 Musical instruments, 27-8
‘natural’ scale, 22 Musicians referred to :
triad, 69 Ali Akbar Khan, 49 n., 53
Malhdr rags, 161 Amir Khusraw, 17, 17 n,, 20 n., 187
>'Mamak\ 18 Baz Bahadur, 19
Man Rautuhal, 18 Bilas Khan, 19, 20
Persian translation, 21 Bismillah Khan, 71,188 n.
Mandr, 34,45,192 Bundu Khan, 23 n., 71,148 n.
Man Singh Tomwar, 18,19, 21 Ghulam Ali Khan, Bare, 35 n.
Maqam, 19 n. Imrat Khan, 85 n.
Maqam Sthan, 42 Lai Khan, 20
Marjand, 90 n. Ram Narayan, 83 n., 172
Mdrvd That, 47-8, 58-64 pass. Ravi Shankar, 49 n., 53, 73 n., 89 n.
accidentals in, 112, 113,119,121 Roshanara Begam, 104 n.
balance and imbalance in, 83 Tansen, Miya, 19, 20,186
hexatonality in, 83,113, 145-6 Umrao Khan, 148 n.
origin of, 94-7, 98-100 Vilayat Khan, 27 n., 69 n., 73 n., 82 n., 139 n„
pentatonic derivative of, 131,132 159 n., 188 n., 186-9, 193-210
216
Index
Naghmdt-i-Asafi, 22 Percussive line, 28
Naqqdrakhdna, 19 n. Portamento, 187, 191
Narada, 91 n. Pratap Simh Dev, 22
‘Natural’, 22, 33, 55 Pulse, 29
see also Suddh Pundarika Vitthala, see Sadragcandrodaya
Ndtyasdstra Purvaiig, uttrahg, 43, 44, 77
concept of amsa, vddi, samvddi, 43-4 see also Tetrachord
melodic system, 16,21, 90 n., 114 Purvi 7''hat, 47-8, 58-64 pass.
Nava rasa, 19 accidentals in, 112-13, 118, 121
Nisdd, 32 balance and imbalance in, 84
Notation, 189-92 connection with Toy! that, 63
Bhatkhanqle’s, 37 n. hexatonality in, 84, l i 3 , 145
of accompanying record, 193-210 origin of, 94-6
Notes pentatonic derivative of, 131-2
first and second order balance, 111-13,115— plagal inversion of, 88
21 Pythagorean intonation, 50 n.
first and second order omissions, 127-9 Pythagorean comma, 59
important, see Samvddi, Vddi
Indian, 32 Qaul, 17
induced dynamic function of, 70, 150,172-6 Qamvali, 17
inherent dynamic function of, 65-70, 149-50,
169-75 Rag{s), 16, 28, 32
oscillation of, 162-3,166,198-200 evolution of, 54, 94-101, 104-6
scalar and accidental, 104 Indo-Persian, 20 n.
suspense and resolution of, 175-6 maqam, connection with, 19 n.
terminal, 76 n,, 175-6, 208 new ‘isolate’, 48, 53-4, 88,136-7,144
twelve semitones, 21, 21 n., 50 principal features of, 45
see also Accidentals, Alternative notes, transilient tendency in, 124-7
Consonance-dissonance, Ground-note see also Hexatonic, Melakarta, Pentatonic,
Nyasa, 20 That
RAGS referred to :
Oblique movement, 39, 41, 45,151-61 Abhogi, 134 (Ex.), 144-5 (Ex.)
first and second order, 152-3 Addnd, 110, 114, 177 (Ex.)
in relation to accidentals, 104-6, 151-3 Afiir Bhairav, 48, 87, 134-5 (Ex.)
in relation to hexatonality, 146,149 Ahir Lalit, 49, 88 (Ex.)
in relation to symmetry, 139-40 Alhaiyd Bilaval, 107 (Ex.), 110
to provide discontinuity, 156 Anand Bhairav, 48, 87, 134-5 (Ex.)
Omitted notes, see Hexatonic series, Pentatonic Ahjni Toy1,110 n.
series, Transilience Asdvri (kornal), 166-7 (Ex.)
Organum, 79 Asdvri {suddh), 149 n.
Overtones, 72,75-6 evolution of, 95-6,165-6 (Ex.)
intonation of, 34,166
Pakay, 24, 37 n., 38, 42, 143 Bdgesvari (Vdgisvari), 100
Pakhvaj (Pakhdvaj), 28, 29 n. as Bagesri ,135, 135 n. (Ex.)
Pancam, 32 Bahaduri Toyi, 110 n.
Pause, 37 n. -38 n. Baiigdl Bhairav, 148 (Ex.)
Pentachord, 43, 76, 77 Barhams Saraiig, 147
Pentatonic series, 37, 122-3 Basant, 61
anhemitonic, 130, 137-8 see also Vasant
balance in, 127,129,148, 149 Basant Mukhari, see Vasant Mukhdri
Circle, 132, 137 Bhairav, 61, 94-5
derived from heptatonic, 130-7 Bhairvi, 35,166
problems in classification of, 51-2 accidentals in, 110, 111, 114, 118-20 (Ex.),
unbalanced, 137-9,142,145 124-6
217
Index
764 GSJreferred to—cont. Gujri Tori, 188
Bhairvi—cont. evolution of, 94-6
after Fox Strangways, 61 hexatonality of, 84,146
diminished fifth in, 49-50 (Ex.), 73 Gurtkali, see Gunkri
evolution of, 100 Guijkri
omitted notes in, 85,124-6 (Ex.) evolution of, 94-5
temporary change of scale in, 119-20 parent of, 133-4 (Ex.)
Bhatiyar, 94-5, 97 symmetry in, 143 (Ex.)
Bhimplasi, 114 n., 159-61 (Ex.) Gurjri
Bhupal Tori, 98 (Ex.), 131-2 (Ex.), 141-2 XJttara and dak$ina, 96 (Ex.)
(Ex.) see also Gujri Tori
Bhupdli, 38 (Ex.), 130 (Ex.), 132, 137-8 Hamir, 53,110 (Ex.)
(Ex.), 202 oblique movements in, 152-4 (Ex.)
problems in classification of, 51-2 (Ex.) Hamsdhvani, 88-9 (Ex.), 134 (Ex.)
range of, 177 (Ex.) Hejujiy HejaZy 94
Bihdg, 152 (Ex.) Hem Kalydn, 146 (Ex.)
Bihagra, 108 (Ex.), 110 Hindol, 132 (Ex.)
Bilaskhdni Tori, 98 (Ex.) dynamic functions in, 142
after Fox Strangways (Bilaskhdni Todi), 61 evolution of, 101
Brindabni Sdrang, see Vrindavni Sdrafig imbalance in, 139 (Ex.)
Candrkos (Candrkaus) oblique symmetry in, 139-40 (Ex.)
Bhatkhapê’s version, 135 (Ex.) South Indian and North Indian, 101 (Ex.)
modern version, 115 n., 136-7 (Ex.) Jaijaivanti, 33 n., 155 (Ex.)
Cdrukesi, 48, 88 (Ex.), 135 Jaunpuri, 149 n., 169-70 (Ex.)
Chdydnaf, 153 Jet, 100,119,135 (Ex.)
Des (Des) Jet Kalydn, 99
accidentals in, 52 (Ex.), 76 n. Jetsri, 52 (Ex.), 94-5
hypothetical reconstruction of, 156 Jhinjhoti, 149 n., 155, 157-8 (Ex.)
melodic movement in, 38-41 pass. (Ex.), Jogiydy 148-9 (Ex.)
108-9 (Ex.), 157 (Ex.) Kaldvti, 136 (Ex.), 144-5 (Ex.)
range of, 192 Kalydn {Kalian), 61
recorded example, discussion of, 193-5 Kdmod, 44
symmetry in, 155-6 (Ex.), 158 (Ex.) melodic movement in, 39-41 (Ex.)
Desi Tori, 95 oblique movements in, 152-3 (Ex.)
Darbdri Kedar {Keddra)
dynamic functions in, 164,171 (Ex.) accidentals in, 112,117
intonation of, 114 n. classification of, 53, 154
oscillation in, 35, 103, 162-4 inflexion in, 103-4, 104 n.
range of, 177 (Ex.), 192 range of, 192
recorded example, discussion of, 197-200 recorded example, discussion of, 208-210
symmetry in, 162 (Ex.) symmetry in, 152-3 (Ex.)
Deskar, 95, 177 (Ex.) Khamaj
Devgandhar, 95, 115-6 (Ex.), 118 alternative ascending lines in, 109-10 (Ex.),
Devgiri Bilaval, 107 149 (Ex.), 174
Devrahjni, 134 (Ex.), 144 dynamic functions in, 173-5 (Ex.)
Devsakh, 147 oblique movement in, 155,156
Dhandsri, 94—5 Khambavti, 155-6 (Ex.)
Dhdni, 130 (Ex.), 132, 138 (Ex.) Khat, 94, 95
Durgdy 130 (Ex.), 132, 138 (Ex.) Kirvdni, 48
second version, 135 (Ex.) Komal Asdvri, see Asdvri (komal)
Gardy 73 n., 155 (Ex.) LdcariTori, 110 n,
Gauri, 61, 94, 95 Laksmi Tori, 110 n.
Gaur Sdraiig, 152-3 (Ex.) Lalit (Bhatkhancje’s version), 37, 48-51 pass.
Gopikd Vasant, 147 (Ex.) (Ex.) 87-8 (Ex.)
218
Index
7L4GS referred to—cont. Sdmant Sdrafig, 147
Lalit (Modem version), 37, 48-9, 87-8 (Ex.) Safikrd, 131 (Ex.), 132,187, 192
Lalit Pancam, 50 (Ex.) balances in, 139 (Ex.)
Madhmdd Sdrafig, 146 drone tuning in, 187
Madhukdnt, 86 n., 136-7 (Ex.) pentatonic and hexatonic version of, 131 n.,
Madhukos, (Madhukaus) 86 n., 136-7 (Ex.) 139 n.,
Madhuvanti, 136-7 (Ex.) range of, 192
Maligaurd, 94, 95 recorded example, discussion of, 206-8
Malkos (Malkaus) 85,130 (Ex.) 132,135-6,147 Vddi and samvddi of, 142
balance in, 138 (Ex.) Sdrafig, 130 (Ex.), 132, 138 (Ex.)
derivatives of, 137 (Ex.) Saveri, 95-6
drones in, 71 (Ex.) iSimhendra Madhyamd, 48
evoked in rag Mdrvd, 202 Sivranjni, 136 (Ex.)
evolution of, 100 Sohni, 140,146,188
Mdlsri 52 (Ex.), 129, 131-2 (Ex.), 140-2 Sorath, 76 n.
(Ex.) Sri, 52, 131
Mdlvi, 94-5 evolution of, 101
Mcirg Tori, 97 (Ex.) intonation in, 168-9 (Ex.)
Mdrvd, 44 n. 114-15 (Ex.), 146 South Indian and North Indian, 101
dynamic functions in, 44 n., 172 (Ex.) , (Ex.)
drone tunings in, 188-9 Srigauri, of Locana, 101
evolution of, 94—5, 97 Suddh Asdvri, see Asdvri (suddh)
omissions in, 83 (Ex.) Suddh Kalydn, 100
range of, 192 Suddh Sdrafig, 93,147 (Ex.)
recorded example, discussion of, 202-4 Sughrdi, 100, 161-2 (Ex.)
Meghranjni, 134 (Ex.), 144 Siihd
Miya ki Sdrafig, 147 (Ex.) intonation of, 114 n.
Miya ki Tori, 84, 97, 187 range of, 192
see also Top recorded example, discussion of, 200-202
Miya Malhdr, 19 (Ex.)
Motki, 114 ^ symmetry in, 147,161-2 (Ex.)
Multani, 94-6 (Ex.) Sukl Bilaval, 107-8 (Ex.), 110
Nat Bhairav, 53 (Ex.), 87, 119, 134 Sydm Kalydn, 153
NdykiKdnhrd, 147 (Ex.), 161-2 (Ex.) Tilak Kamod
Pancam se Gdrd, 73 dynamic functions in, 172 (Ex.)
Pancam se Pihi, 73 oblique movements in, 155
Patdip, 115 (Ex.), 136 range of, 192
Patdipki, 48 n., 116 recorded example, discussion of, 195-7
Piiii symmetry in, 157-8 (Ex.)
accidentals in, 73 n., 117,119 vddi, samvddi of, 43
classification of, 53 Tilafig, 135 (Ex.), 143-4 (Ex.)
influence of accidental on scale of, 115 (Ex.), Tori, 84,146
136 after Fox Strangways (Todi), 61
omissions in, 125-6 (Ex.) change of scale, 97
vddi, sairivadi in, 44 n. divergent intonation in, 167-8 (Ex.), 167 n.
Piiravd, 93, 93 n. evolution of, 97-8 (Ex.)
Piirbya, 146 North and South Indian, 97
Puriyd, 44 n., 99,146,188, 202 see also Miya ki Tori
Puriyd Dhandsri, 94, 95 Triueni, 94-5
Puriyd Kalyditi, 100 Vagiivari, see Bdgesvari
after Fox Strangways (Puriyd Kalian), 61 Vasant, 94-5
Rdmkali, 49 (Ex.), 94—5 after Fox Strangways (Basant), 61
Reva, 94—5,143 n. Vasant Mukhari, 48, 133-4 (Ex.)
Sahdnd, 161-2 (Ex.) ViMas, 94-5, 133 (Ex.), 142-3 (Ex.)
219
Index
RAGS referred to—cont. Samvddi, 42,45,123,169
Vrindavnl (Brinddbni) Sdrafig, 37 (Ex.), 146 ancient concept of, 43-4
(Ex.) in relation to symmetry, 142,169-78
Yaman modern concept of, 42 n., 44
balance in, 82-3 (Ex.) see also Notes, dynamic function of,
omissions in, 82, 126-7 (Ex.) Saiicdrl
range of, 192 as section of song, 30
recorded example, discussion of, 204-6 (Ex.) as varn, 38
transilience in, 82 (Ex.), 83 Sangitadarpana, 20 n.
tuning in, 187 Sahgitamakaranda, 91 n.
vddi, samvddi, 42 n. Safigitapdrijdta, 20, 96, 97, 100, 101
Yarn Rahjni, 132 Safigitaratnakara, 16, 20
Raga-rdgitR classification, 91,122 Safiglt-Sdr, 22, 97, 98 n,, 100
Ragalak$ana, 95 Sahglta Sara, 23
Ragamdla, 99 Sangltasdramrita, 91 n., 93 n.
Rdga-mala paintings, 122 Sangitasiromarii, 18
Ragamahjarl, 95, 100, 101 Sanyal, A. N., 42 n., 175 n.
Rdgataranginl, 20, 21 n. Saptak, 34
melas and ragas of, 92-7, 99, 100, 101 Sarangi, 27, 28, 71
Ragavibodha, 20, 97, 99 Sdrafig rags, 146-7
Rag Darpan, 21 Sargam, 189, 191
Rag Vijhdn, 48 n. see also Tonic-solfa
Ram Narayan, 83 n., 172 Sarngadeva, see Safigitaratnakara
Ramamatya, see Svaramelakalanidhi Sarod, 27, 30
Ranade, G. H., 22 n., 25 Scale(s)
Rasa, 18 anhemitonic, 130,137
Rasakaumudl, 91 n. diatonic, 36 n., 51
Ravi Shankar, 49 n., 52, 73 n., 89 n. instability of, 79
Registers leading note influence on, 115, 136
octave, 34 most consonant, 91,182-4
relationship of tetrachord species to octave, most dissonant, 183
178 quarter-tone, 68
tetrachord, 76-8, 177 represented diagrammatically, 181-4
varying range of, 158, 202 tempered, 59, 68
see also Tessitura, Tetrachord wholetone, 68
Rik Pratisakhya, 122 n. see also Hexatonic series, Madhyamagrdma,
Ri?abh, 32 Mela, Modes, Pentatonic series, Sadjag-
Rondo form, 31 rdma, That
Roshanara Begam, 104 n. Scale species, 37,122-3
Roy, H. L., 25, 48 n., 63 n. Semitone(s)
Rupak tdl, 29 n., 167 Just intonation, 16
system of twelve, 33, 50
Western tempered system, 59
Sadarang, 21 Shahjahan, 19
$adav (Khddav), 37, 45 Shahnd% 27, 28, 71
see also Hexatonic series, Transilience Sharp, 33
$adjagrama, 16, 22, 23 n., 56 n,, 90,114 Sikandar Lodi, 18
$adj (Khadj), 32 Sitdr, 17, 28, 30, 34
§adjamadhyama, 44 gayki style, 187
Sadragcandrodaya, 20, 95 tunings 187-9
Sdkdrl, 114 n., 159, 198 Somanatha, see Ragavibodha
Sam, 29, 30 n., 31 South Indian music
Sampun.t, 37, 45,122 classification of ragas in, 20, 48,123
see also Heptatonic series Islamic influence in, 20
220
Index
South Indian Music—cont. Tansen, Miya, 19, 20, 186
ragas of, 96-7, 100-1, 133-4 Tdr, 34, 45,192
rags imported from, 48, 49, 88-9, 133-4 Tarab, 28
treatises, 91, 97, 100 as sympathetic strings, 187 n.
Spiral of fourths/fifths, 60 Tardnd, 17, 30
&rimal-lak$yasa?igitam, 23 Tempo, 29
M s ) , 35, 36,45 Terminal notes, see Notes
ancient concept of, 16, 91 Tessitura
cents value of, 16 n. of rags, 177
in oscillation, 35, 162-3, 166 of sthayi, antra, 192-3
non-functional, 21, 35 n., 91 varying range of, 158, 202
see also Intonation see also Ambitus
Sthan, 34 Tetrachord(s), 43,103,151,181
Sthdyi ascending, descending, 77-9
as main verse of song or composition, 30, conjunct, disjunct, 76-9,177-8
125 n. parallel, 78-81
as varn, 38 unbalanced, 88
range of, 192-3 see also Balance and imbalance, Symmetry
Stick zither, 90 Thapiya, 29 n.
Suddh, 22, 33, 36, 37, 46 Thdt(s), 38,45, 46 ff.
Suddh svar, 33 (Ex.), 35 Bhatkhancje’s ten, 48, 63
Sufi, 17 Circle of, 59-60, 73-4, 97, 111
Summation tones, 75-6 Circle after Fox Strangways, 60-1
Surbahar, 27, 34, 85 n. development of, 93-6,104-6
Sur-peti, 28 hypothetical, 58-9, 97-9
Svar, 32 inadequacy of, 51-3,102
see also Notes pentatonic derivatives of, 130-2
Svaramelakalanidhi, 20,100 serial, 56, 60,129-30
Svarqprastdra, 31 n. six primary, 73-4, 78, 91, 100
Svarvistar, 24, 37, 39,42 n., 109, 153,154,155 n. system of thirty-two, 46-7,181-5
Symmetry see also Mela, Time Theory
conjunct and disjunct, 77-8, 139, 154, 157, Thekd, 29, 30
177-8 Thumri, 30,119
incomplete, 145 Time measure, see Tdl
inverted, 143,174 Time Theory, 43, 45,61-4, 99
non-tetrachordal, 88-9, 140,143,152,173 Tivr, 33
pentatonic, 145-6,148-9,161-3,171 Tintdl (Tritdl), 29,125,159,160,168,170,171
rhythmic, 195 Tivrd tdl, 29 n.
tetrachordal, 78-87 pass, Tonality, dual, 103
vddi clue to, 142 see also Modality
Tones, see Notes
Tabid, 17, 28, 29, 31 Tonic, see Ground-note
Tdl, 17, 28, 29-31 (Ex.), 37 n. - 38 n. Tonic-solfa, 32, 91
Tals referred to : Toj'i That, 47-8, 58-64 pass.
Dadra, 142,142 n. accidentals in, 112-13
Dipcandi, 174,174 n, balance and imbalance in, 84-6
Ektdl, 29 connection with Purvi that, 63
Jhaptdl, 29, 141, 141 n. hexatonality in, 84-5, 113,145-6
Rupak, 29 n., 167 origin of, 94-6
Tintdl (Trital), 29, 125,159,160,168,170,171 pentatonic derivative of, 131-2
Tivrd, 29 n. plagal inversion of, 88
Tali, 29 Transilience
Tamburd (Tanpura), 28, 67, 72,188-9 directional, 39,149-50
Tan, 175 n. first and second order omissions, 127
221
Index
Transilience—cont. Vakr, 39
motivated by imbalance, 82, 85,123-4 see also Oblique movement
omissions of pairs of notes, 128-9 Vakrsvar, 39
temporary, 82-3,124,126,140, 141 Varjitsvar, 44
tetrachordal, 144,157 Varn, 38
to create symmetry, 139-40 Venkatamakhl, see Caturdaijidiprakdsikd
Tritdl, see Tintdl Vikrit, 21 n., 33, 36-7, 46
Tritone, 56, 79, 79 n„ 91 Vikrit svar, 33 (Ex.), 35,181
Vilayat Khan, 27 n,, 69 n., 73 n., 82 n., 139 n.,
159 n., 186-9
Umrao Khan, 148 n. musical examples performed by, 193-210
Uttrang, see Purvang, uttrang Vilambit, 29
V m , 21, 90
Violin, 65, 67
Vddi, 123 Visranti svar, 42
ancient theory, 43-4 Vivddl
defined, 42,169 ancient concept of, 43
in relation to ambitus, 177-8 modem application of, 44-5,196, 205
in relation to suspense and resolution, 169-71 Vocal music, 30
in relation to symmetry, 142,173-7
in relation to time theory, 43,45, 62 Wholetone
tuning of strings to, 187-8 major, minor, 16,91
variableness of, 42 n., 175,175 n. see also Just intonation
see also Notes, dynamic function of, Willard, N. Augustus, 18 n., 23
222

The Rags of North in D Ia N Music: Their Structure and Evolution

Uploaded by

Copyright:

Available Formats

The Rags of North in D Ia N Music: Their Structure and Evolution

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Rags of North in D Ia N Music: Their Structure and Evolution

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What are some of the acknowledgements mentioned in the book?

What are some of the acknowledgements mentioned in the book?

THE RAGS OF

NORTH INDIAN MUSIC

Their Structure and Evolution

All rights reserved

INFORMATION TO ALL USERS

In the unlikely e v e n t that the a u thor did not send a c o m p le te m anuscript

Published by ProQuest LLC(2018). C opyright of the Dissertation is held by the Author.

All rights reserved.

aspirated These can be approximated by exaggerating the

Semi-vowels Traditionally classified as a group, but in Hindi the

For a fuller discussion of pronunciation see T. Grahame Bailey, Teach Yourself

Basic Elements o f Theory

This system of nomenclature has wide acceptance in India, and is

Ex. 7. rag Des

Sa, Re, M a Pa, Ni Sa Sa Nik Dha p a> Qha M a Ga, R.e Ga Sa

These turns, which are characteristic features of certain rags, are

Ex. 8. rag Kamod

Sa, N i Dha, Pa, Ma# Pa Dha Pa, Ga Ma# Fa, Ga M a Re Sa

There is a further complication in the description of this rag, for although

(a) N is clearly a direct ascending note.

Ex. 11. rag Kamod

Ex. 12. rag Des

A 9 PurvT that B9 C9 Bhairav that

A10 Marva that BIO CIO

rag Madhukdnt (No. C4) rag Ahir Lalit (No. 35)

Ex. 13. rag Ram kali

3 K.P.M. IV, p. 490.

Ex. 17. rag Lalit

Ex. 18. rag Lalit

Ex. 19. rag Bhupall

1 K.P.M. IV, p. 821.

Ex. 20. rags M alsri and Jetsri (ascent)

Ex. 21. rag Des

Ex. 22. rag Nat Bhairav

1 J. Grosset in A. Lavignac, Encyclopedie de la Musique, Paris 1921, premiere partie, p. 326.

‘D ’ m ode ' Sa Re Gat Ma Pa Dha

A ncient Ecclesiastical A ncient M odern

Si Re GaV M* Pa D hai Nfc Sa

augmented 4th augmented 4th

augmented 4th Iaugmented 4th

diminished 4th diminished 4th

diminished 4th diminished 4th

MARVA (No. A10) BiLAVAL (N o.A 2)

Fox Strangway s’s mgs Bhatkhande’s thats

Bhairvi Bhairvi (No. A6)

Sa Ret Gat Ma Pa Dhat Ni Sa

Sa' Ret Ga Mail Pa Dhab Ni Sa

Gauri Purvi (No. A9)

Sa R et Gat Maif Pa Dhat Ni Sa

Sa Ret Ga Map Pa Dhat Ni Sa

Sa Ret Maff Pa Dha

Kalian Kalyan (No. Al)

Time theory and the Circle of Thats

during a performance; and in the development of melodic and rhythmic ideas

The Effect o f Drones

In addition, Nib occurs occasionally as an oblique ascending note—perhaps a vesti