Early Indian Musical Speculation and The Theory of Melody

Yale University Department of Music
Early Indian Musical Speculation and the Theory of Melody

Author(s): Lewis Rowell
Source: Journal of Music Theory, Vol. 25, No. 2 (Autumn, 1981), pp. 217-244
Published by: Duke University Press on behalf of the Yale University Department of Music
Stable URL: http://www.jstor.org/stable/843650 .
Accessed: 30/10/2014 10:43
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Duke University Press and Yale University Department of Music are collaborating with JSTOR to digitize,
preserve and extend access to Journal of Music Theory.
http://www.jstor.org
This content downloaded from 151.100.161.184 on Thu, 30 Oct 2014 10:43:33 AM

All use subject to JSTOR Terms and Conditions
EARLY INDIAN MUSICAL SPECULATION

AND THE THEORY OF MELODY
Lewis Rowell
Our musicallegacy from Westernantiquity consists of many words

and few tunes. While the survivingdocuments are, in many cases,
extremely informativewith respect to the construction of the musical
systems, they tell us next to nothing about how actual melodies came
to be written and what sort of guidelinesconstrainedthe choices of the
composer. The present paper addresses some basic issues in ancient
melodic theory, by means of an exposition of the early Indian theory
of melody, with comparisons to what remains of Classical Greek
melodic doctrines. The comparativeapproachworks both ways: our
appreciation of the achievements of early Western music theory is
enhanced by consideringalternativesolutions, and we can gain insight
into a foreign musical system by discovering analogues-as well as
startlingcontrasts-to our own. We pose these questions:
1. What does a theory of melody seek to accomplish-lay out a
gamut, outline specific melodic pathways thereon, prescribea set of
musicalmodels, describean existing repertoire,establishcertainmelodic
functions?
2. By what principlesare the two primarydimensionsof pitch and
time structuredin sets of analytictopics?
3. To what extent can early melodic theory be read as a record of
evolution-of a musicalsystem and of a musicalrepertoire?
217

The approachin this paper will be in three stages,becomingincreasingly specific: first, an introductionto the genres,concerns,and distinctive features of early Indian musical thought and the intellectual and
cultural frameworkwithin which it flourished;second, an examination
of the Indian and Greek systems in relativelygross terms,by meansof
the principaltopics used to dissect the domains of pitch and time; and
third, a brief study of the Indian theory of melody, with particular
attention to the relationshipbetween the establishedpitch gamut and
the authorizedselections to be made therefrom,that is, melodic composition.
MUSICIN EARLYINDIANTHOUGHT
The Indiantradition of musical speculationis by far the richestand

most ancient treasuryof Asian musical thought. WhileChina'smusical
heritageis of comparableantiquity, the survivingbody of texts is considerablysmaller;1further,it is evident that musical speculationper se
held little interest for the practical scholars of early China. The few
texts that survivedthe Great Book-Burningand other similarcultural
upheavals focus upon the role of music in ceremony and are highly
prescriptive; the components of the musical system are treated as
givensand not subjectedto any kind of penetratinganalysis.The Indian
mind, on the other hand, delightedin mentalgymnastics,verbalpuzzles,
and the most profoundinquiryinto the natureof music and its context
in the Indianworldview.Andjust as the Sanskritlanguageis the easternmost branch of the Indo-Europeantree of languages,so is the early
Indian musical system more amenableto analysisin categoriesfamiliar
to the West than the musics that evolved in the regions of East Asia
dominatedby Sinitic culture.
Thus the music of India has a formidabletheoretical base of great
antiquity, and contemporaryIndianmusic (apartfrom the music of the
Bombay films and the tribal musics of peripheralIndia)is informedby
a rigorous body of melodic, rhythmic, and aesthetic theory whose
roots are clearly discerniblein the earliest survivingtexts from more
than 2,000 years ago. This extensive corpus of musica speculativais
still virtually inaccessible to Westernmusicians,due to the poor state
of the texts and the stylistic intricaciesof the Sanskritlanguage.And
the few translationsinto Westernlanguageshave done little to close this
gap, suffering either from an overly-freestyle of translationor weighed
down by the voluminous explanatory notes that must inevitably
accompany a translationthat attempts to remain accurate. In the end
translationsare quite uselessuntil one learnsthe system.
The scholars of ancient India faced the same set of problemsthat
confronted Westernthinkers:how to objectify the experienceof music
218

so it could be taught, learned, remembered,and repeatedaccurately;

how to separate the phenomena of music from such other cultural
activities as speech, chant, poetry, and ceremonialmovement;how to
identify and arrangein systematic order the components of the major
dimensions of music: pitch and time; how to define the units, groupings, and higher levels of the musicalhierarchyand prescribeguidelines
for selecting, combining, and recognizingvariousparts of the musical
system.
The Sanskritword for music is sarhgita("concerted song"), a term
that includes not only vocal and instrumentalmusic but also dance.
the study of treatises on music, is usually translated
Sarhzgitasdastra,
today as "musicology." Early Indian literature is full of puns on the
words s'Ostra
and sastra;the formermeansa venerabletreatise,the latter
a sword. Because of their phonetic similarity, both words have often
(and on dubious grounds) been attributedto the Sanskritverbalroot
meaning "to correct, chasten." The semantic developmentof the word
sffstrais quite clear: from the basic meaning "to correct,"it evolvedas
follows: (1) to restrain,control, rule, govern;(2) to command,enjoin,
decree; (3) to teach, inform; and a series of even further attenuated
meanings that include both "to praise" and "to blame!" Nowhere is
there any explicit connection with the root meaning "to cut" (as in
sastra), but there are clear indications of an aggressive,often punitive,
brand of pedagogy. The word for student (sisya) derivesfrom the same
root as s'astraand means literally "he who is to be punished." It is
evident that sarhgitasa'stra
is not only a prescriptivedisciplinebut one
designedto uphold tradition-not overthrowit.
There is bound to be lively debate over the dating of treatisesin a
country where there are no survivingmanuscriptsover 400 years old
and where worms and heat have taken their toll of the palm-leafpages,
but most scholars agree that the three oldest and most important
musical texts may reasonably be dated between the first and fifth
centuries A.D., although many portions of these texts may date from
as much as a thousand years earlier. Since later texts are invariably
designedas glossesor commentarieson earliertexts andincludeabundant
quotations, the lines of descent can usuallybe tracedand become quite
clearby the end of the first millennium.
The three texts provide not only the main evidence for this study
but also an excellent introduction to the genresof early musicalspeculation: the NAradTyasikksa
(the
["phonetictreatise"]attributedto
the legendarysage Nirada) is a manual
fiks.i that teaches the properchanting
of the hymns of the Sffmaveda;2Bharata'sNatyaJ'strais a wide-ranging
treatise on the theater, with chapterson such diversetopics as acting,
playwriting,makeup, singing, dancing, instrumentalmusic, and drumming;3 the Dattilam (an otherwise-untitled treatise by the author
219

Dattila) is apparentlya concise performancemanual for the style of

ritual theater music known as gffndharva.Dattila addressesonly the
topics of svara(pitch) and t/la (time), but-since his exposition closely
parallelsBharata'schapterson the same topics-furnishes independent
verificationof many difficult passagesand issues.4 By far the most important and authoritativeof the later commentariesis the exhaustive
commentarythat the eleventh-centuryKashmirischolarAbhinavagupta
despite all its opaquepassagesand textual
appendedto the
N.tyass'stra;
uncertainties, his
commentary Abhinavabhdratiremains the most
invaluablebody of exegesis for the earliertexts and the standardagainst
which all later commentariesmust be measured.s
Authority is an interesting issue in Indian scholarship.The Indian
philosophicaltradition attachesas greatvalue to testimony as it does to
perception and inference, and both the written text (no matter how
dubious) and the utteranceof text citations are held in a reverencethat
never fails to astonishWesternacademics.Indianscholarshipis designed
to reconcile,to harmonizeconflictingopinions,to explainawayapparent
contradictions,and above all to promote the status quo-not to reveal
erroror misinterpretation.Thus commentatorsgo to extremelengths to
rationalizewhat is most probably scribalerror, thereby accountingfor
some of the intellectual gymnasticswith which Indian music theory is
riddled. Modem scholars, notably at BanarasHindu University, are
making some progressin cleaningup the majortexts, but somethingof
the old venerationstill persists. The Indiantraditionof musicaspeculativa is often describedby a familiarbotanical metaphor: a tree that
accepts new grafts onto the central trunk, rejectingothers and sloughing off old growths. The central core of ideas is somehow beyond
body of amplificacriticism, surroundedby a gradually-accumulating
tion and interpretation,not unlike the tradition of scholarshipin the
WesternMiddleAges.
The authorityattachedto the texts is extremely heavy in the absence
of precise musical notations. Indian musical notations are even more
skeletal than early Westernnotations, although both sol-fa and drum
syllables are found in the early texts. But even if one takes the dubious
step of assumingthe modern equivalentsof these symbols, only the
barest bones of the music remain. Even in modem Indian practice, a
performer does not learn a piece from notation alone-it must be
learned from a teacher. So the most one can say for the most informed
modem reconstructionis that it can be shown not to be in conflict with
the sistras.
The discipline of music flourished amidst colorful and exotic surroundingsin the Indianintellectual tradition.It was held in the highest
esteem and invested with an aura of religious mysticism and cosmic
symbolism that is firmly embeddedin Indianconsciousness.Musicwas
220

often referredto as parichamaVeda(the "Fifth Veda") and regardedas

a means of revelation. It was also linked with the philosophyand practice of yoga (literally "attachment")and servedas a model for yogic
practice in the rigorousdiscipline and concentrationof the performer
and his attachmentto the ideal model he is following. Music,the least
materialof the arts,had obvious attractionsfor the Indianphilosophical
tradition which taught that matter is illusion. Music afforded both the
performerand auditor a glimpse of eternalreality:by their attachment
to what is real (often attained by means of a trance-likestate), they
obtain detachmentfrom what is illusion. Thusthe Indianphilosophyof
music has important ontological and epistemologicaldimensions,and
the performanceof Indianmusic has acquiredmetaphysicalovertones
which have profoundly affected all the contexts in which the subject
of music appears.
The origins of most ancient art musics can probably be traced to
ritual, but these roots lie more exposed in Indian music than in the
West. Music, insofaras we know, first appearsin Indianliteraturein the
genreknown as siksd. The siksdswerepracticalhandbooksfor the correct
utterance of the sacredVedic texts and the properperformanceof the
accompanying ritual gestures, supplemented by lists of maxims for
everydaylife-how to brush one's teeth with twigs, avoid the company
of gamblers,women, and actors, and the like. In the fiksaisthere is some
evidence that the tfla gestures, which control the rhythm of all later
Indian music, grew out of the hand motions made by the priest as he
chanted the Sfmaveda hymns. In fact I believe very stronglythat some
traces of the old Vedic notion of music as sacrificialritualstill inherein
the performanceof Indianmusic: the theory of niida,usuallytranslated
as "causalsound," is the foundation for the Indianconcept of music.6
It details the emergenceof articulatesound from the continuous substratumof vital sound that inhabits the innermostregion of the body,
as symbolized by the constant drone that has been a featureof Indian
music since the MiddleAges. Everyemissionof breath,every utterance
of a musical sound or of a sacred formula (mantra),puts the singerin
control of this inner reservoirof divine sound, and the idea of music as
pure emergencefrom within is one of the most basic concepts in Indian
thought. The accompanimentof music by gesture (as in Vedic chant)
provides a ritual measurement of the bodily channels and regions
through which the emergingsound travels. The philosophicaltradition
has addressedthese questions in detail too great for the scope of this
paper, but with this backgroundit is easy to see why vocal music is the
model for all Indianinstrumentalmusics and why music and phonetics
are so closely intertwined in the siksEliterature. Music was a sacred
utterance, and every aspect of its production had to be meticulously
prescribedfor the novice. The accompanyinggestureswere, to borrow
221

a phrase, "outward and visible signs" of an inward and spiritual

process.
The siksd genre formed one of the six Vedifngas,"limbs of the
Vedas."Hereare their domains:
1. Siksj, apart from music, was concerned primarilywith articulatory phonetics and was subdividedinto such topics as individualsounds,
accents, quantities, the organsof speech, delivery, and euphonic laws;
its overallpurposewas correctrecitationof the texts. 7
2. Kalpadescribedthe properrituals.
3. Vyjkaranawas a meansfor the grammaticalanalysisof the texts.
4. Niruktawas a method for understandingword etymologies.
5. Chandasmeant the study of poetic metres.
6. Jyotisa was the subject of astronomy, a science that was highly
developed in early India but here applied mainly to determine the
correcttimes for the Vedic sacrifices.
The disciplines fall into pairs: siksa and chandas were concerned
with performance,kalpa and jyotisa with ceremonies, vydkaranaand
nirukta with the texts themselves,which had to be preservedin exact
detail-so exact (so the story goes) that when the redactors of the
Vedas assembledtheir evidencefor the final fixing of the texts, not one
variantreadingturnedup! The six Vedifngasform the nearestequivalent
to the MediaevalEuropeantrivium(grammar,rhetoric,logic) and quadrivium (geometry, arithmetic,astronomy,music). Mentaldisciplineand
communicationskills are primarygoalsof both the Indianand European
curricula.The major difference appearsto be this: the Westernquadriviumwas designedto trainthe mind by means of the study of abstract
numbers and quantities, a pedagogy obviously in harmony with the
Greek belief that numbers were things; the ultimate concerns of the
Vedffhgaswere for more concrete matters-performanceand the spoken
word.
Ngrada'ssiksi is about 50% music and 50%phonetics (if we use a
modem definition of "music"), and the musical passagesare a naive
mixture of fact and myth. It containsinformationof tremendousvalue,
includinga passagewhich relatesthe degreesof the later secularscale to
the degrees of the Vedic scale.8 In another passagewe learn that the
peacock shriekson the first scale degree, while the bull bellows on the
second and the cuckoo whistles the fifth scale degrees!9But scattered
among such fragmentsof musicalfolklore are severalessentialpieces of
the ancient Indian pitch system, perhapsthe earliestsuch discussionto
survive.Some passagesof this most important text may be later accretions, as some scholarshave contended, but the centralcore must have
been set down in Vedic times.
A second important strandof musicalspeculationis concernedwith
the theater of Gupta India (fourth through seventh centuriesA.D.), of
222

which the plays of Bhisa and Kilidisa are representative.A substantial

corpus of dramatic theory evolved during these centuries, and two
musical genreswere prescribedin very greatdetail for theatricalspectacles: gandharvawas a strict, ritual music for the opening ceremony of
music and dance that dedicatedthe play to the appropriatepatronsand
deities, while gina was a freer style of music that could be tailored to
the various dramaticrequirementsof the performance.The treatisesof
Bharataand Dattila stressedthe musical components and organization
of gffndharva,although the former contains additionalinformationon
the less formal style of music. The word gindharvais cognatewith the
Greek kentauros(centaur) and in later Vedic times was used to denote
the class of handsome demi-gods,living in the sky and notorious for
their seduction of women, who were regardedas the patrons of such
diverse enterprises as oratory, music, and gambling. Their name is
carried not only by this musical genre but by the third scale degree
(GA) and the early scale built thereon that is practiced (as Nirada
tells us) "only in heaven."10 Both Bharataand Dattila set forth the
theoretical system of gffndharvain quite remarkabledetail, and all subsequent Indian music theory can be traced-in one way or another-to
their model.
By the end of the first millenniummusic had become detachedfrom
its earlier ritual and theatrical contexts and was considered as an
independent discipline, even if its venerableassociationslingered.There
are numerousclassificationsof the arts in Indianliterature,frequently
overlapping, seldom agreeing, and developed on obscure principles.
Despite music's privilegedstatus, it has no one identifiable niche, but
more often than not was listed alongside"arts"that must be considered
trivial by any standard. We find classificationssuch as the eighteen
professionalarts(silpas) and the sixty-fourkalds(which Coomaraswamy
labels as "avocational arts"') that Vitsyayana enumerates in his
famous Kffmasiitra:the latter range from major arts (such as singing,
dancing, writing poetry, and architecture)to such pastimes as tattooing, cooking, reciting tongue-twisters, applying perfumes, teaching
parrots to talk, speaking "Pig-Latin,"gambling, and gymnastics.12
Architecture,poetry, and music are usuallycited as "major"arts, since
they can be practicedindependently,that is, sculptureand paintingare
dependent upon architecture for their existence, as dance is upon
music.
Life, according to early Hindu tradition, has four ends: dharma
(righteousness),artha (wealth), kima (pleasure,especially through the
senses), and moksa (liberation from the cycle of birth and rebirthand
the consequences of karma). The arts in general have kima as their
immediategoal, and sensorygratificationis as explicitly recognizedand
valued as an effect of art as is the perception of the spiritualreality
223

signifiedby the sensory illusion. And indeed all of the artswere valued
for their assignedrole in heighteningsexual pleasure.But the ultimate
goal of the musicalexperiencewas moksa, approachedby meansof two
intermediate and more attainable stages of absorption in the musical
process:bhoga(enjoyment) and 7nanda(rapture).
Early Indian music theory, as will soon be evident, is distinguished
by the same rigorous intellectual precision and keen observationthat
marks P~inini'sgreat grammarof the Sanskritlanguage. At the same
time it flourishedamidst exotic surroundingswith strong implications
of cosmic symbolism and mysticism. The texts display some of the
same naiveblend of musicallore and technicalinformationthat we find
in the MediaevalEuropeanliteratureof musicaspeculativaand the same
tendency to dwell on music's origins and effects. The boundariesof
music were gradually defined, distinguishingit from speech, chant,
poetry, ritual, and drama.Musicalspeculationwas practicednot for its
own sake but with a clear, practicalgoal: to insure accurate,effective
performancethat could be repeatedin exact detail-and to this end we
owe many of the complexities of the musical system, especially the
mnemonic aids. Since the treatiseshad to be transmittedorally, authors
resortedto the aphoristicsfltrastyle in orderto presentmaximalinformation in minimal space. The literatureis characterizedby linguistic
puzzles and cryptic codes, perhaps to protect trade secrets from the
uninitiated. But hidden in the dense thicket of ambiguoustechnical
terms and obscure syntax is an abundanceof imaginativeideas and an
incisive, systematic grasp of the musical principleswith which early
Indian authors graduallyorganizedthe mysterious realm of articulate
sound.
TOPICA
The musical systems of ancient Indiaand Greeceare set out in ways
that invite comparisons. Each system is ordered by a set of clearly
hierarchicaltopics that serve to divide the two major dimensions of
pitch and time, as shown in Table 1. In this section we focus upon the
grossanatomyof the two systems, by pointing out importantsimilarities
and differencesin the stated topics and some of their consequencesfor
the musical repertoires.It is useful to bear in mind that there can be
unstated topics, properties of the music that are taken so much for
grantedthat they are assumed,not defined.
The Indian and Greek musical repertoiresalso have some common
features: both were basically ensemble musics designedfor the theater,
using physical gesture to regulatethe music. Both trace their originsin
part to ritual and liturgy and were influenced by the ceremonies,conventions, and styles of sacred chant. Both included some purely vocal
224

and some purely instrumentalmusic, were used to accompanydancing,

and both used the temporal patterns of speech and poetry as a model
for musicalrhythm.
Table 1 shows a numberof similarstrategiesof thought in the Greek
and Indian dissections of the pitch and temporal domains.13In each
case there are very clear indications of hierarchicalthinking,with each
dimension built up from the minimal unit to basic patterns, larger
structuralgroupings,and appropriateGestalts.Eachof the two systems
attempts(with varyingsuccess)to conceptualizefour extremely difficult
issues for the early theory of music: tempo, the role of silence, change,
and style. Each system, in its own way, stressesthe relativefunction of
the major components instead of defining their measurementor position in absolute terms, for example, the Greek intervalratios, genera,
and the dynamic nomenclaturefor the tonoi, the various Indianbeat
gesturesand the criteriafor recognizingvarioustypes of phrases.India
and Greecelackedthe official standardsof ancient China,set by bamboo
tubes of prescribedheight and scale and stone chimes of specific size;
their systems had to be describedin relativisticterms to remainoperative in a musical world in which the only major constants were the
rangeof the humanvoice, the length of the breath,the familiarrhythms
of speech, simple motions of the body, the mutual relationshipsof
numbers,and the facts of auralperception.Indeedit can be arguedthat
music theory flourishesin the absence of precise standardsand a common practiceof composition.Finally, and most germanefor the present
paper, each of the two systems lays the foundation for an elaborate
theory of melody.
Some of the differences are even more revealing: Greek authors
obviously sought to apply the same set of principlesto the organization
of the domains of pitch and time, an ideal that has managedto elude
theorists ever since. While there can be no such thing as a rhythmic
gamut (in terms of scale), the remaining categories are remarkably
parallel.On the Indian side of the Table,we find virtuallyno similarity
between the domains of svaraand tila other than the generalprogression from small to large categoriesand from the specific to the general.
Most obviously the Indian authors preferreda largernumber of categories, with the consequent potential for overlappingand ambiguityso
characteristicof Indianthought.
While both systems are hierarchical,I would describe the Greek
system as a fairly steep hierarchy:multi-leveled,one or few units on
each level, and integrated by the incidence of similarnumber proportions on the variousstructurallevels. In contrast,the Indianhierarchyis
ratherflat: fewer levels, many units within each level with much overlapping of function, and little relationship between adjacent levels.
These differences are more apparentin the temporal organizationof
225

TABLE1A: THETOPICSOF PITCH

GREEK:Harmonics
INDIAN:Svara
1. phthongos, note
2. diastema,interval
1. svara,note, tone
2. grdma,the basic scales
3. genos, genus,tunings
4. systema, system, scale
3. mfircchani,their modes
4. tdna,deficient mi7rcchands
5. sthdna,register
5. tonos, key
6. metabole, modulation
6. vrtti,style, mode of performance

7. Muska,
purely instrumentalmusic
7. melopoeia,melodic
composition 8.
sffdhfrana,intermediatenotes,
alterednotes
9. varna,the 4 basic melodic
progressions
10. alarhkffra,graces,ornaments
11. dhatu, the variousinstrumental

strokes
12. gruti,intonation in general,the
22 octave divisions,
microtones
13. jiti, the 18 proto-melodytypes:
7 pure, 11 modified
226

TABLE1B: THETOPICSOF TIME

GREEK:Rhythmics
INDIAN:Tla
1. chronosprotos, the basic

duration
1. kalJ, a silent and visiblebeat
2. pous, the rhythmicfoot
2. pita, a soundingbeat
3. laya, space between beats,
"tempo"
4. yati, the law of regulation
(3.) genos, genus,proportion
(4.) kenoi chronoi, rests

5. agoge rhythmike,
5. pinzi,the law of phasing,when
to enter
rhythmicprogression
6. metabole rhythmike,
6. pddabh~ga,one phraseof a
groupof four
rhythmicmodulation
7. rhythmopoeia,rhythmic
7. mdtrJ,anothertype of phrase,
a time duration
composition
8. parivarta,repetition
9. vid&ri,a phrasemarkedby a
pause or cadence
10. ahiga,a section of several
phrases
11. vastu, a longertype of section
12. prakdrana,the seven fixed
forms
13. avayava,expansionsor contractionsof these
forms by the interpolation or deletion
of beats
14. giti, treatmentof text, style
15. mirga, relativetime density,
standardpatternlength
227

the two musics, and I would like to suggest that such preferencesfor
certain types of hierarchiesexpress well-establishedculturalmodels. I
suspect that a culture tends to organize its music in the same way it
organizesits social structures;"flow charts"of modem Asianand Western businesses reveal similarsets of differencesin the basic pattern of
organization.
Cultural theories of form and substance underlie other major differences between the two systems: to the Greeksnumberswere things,
and number proportionsheld the universetogether. The forms signified
by these proportions were persistent;rhythmic patterns, for example,
retainedtheir indigenousaccents when appearingin differentcontexts.
In the Indianworldviewsubstanceis an illusion, and all forms are fluid;
melodic and rhythmic patterns easily took on new tonal functions and
accents, thus accountingfor the abundanceof superimposedpatternsin
Indianmusic. Whileboth culturesfound it necessaryto objectify musical
phenomenain the form of hand gestures,the Greekarsisand thesis, and
regionsof voice, I suggestthat both the musicalexperienceandthe musical principlesembodied therein enjoyed greater ontic status in Indian
thoughtthanin the shadowyGreekworldof abstractnumbersand forms.
Culturalpreferencesmust be regardedas causalfactorsin the organization of a musical system-we think in categoriesthat our culturehas
trained us to like, just as we think that which our languageallows us to
express. For the musicianas for the anthropologist"culture"is a grid
that shapes both our perception of and preferences for experience.
Westernersshare preferencesfor definite beginnings,straightlines, increasingtension, definite goals,a certainamountof conflict, a pleasantlytextured surface, sonorous tone, parts that relate proportionatelyto
wholes, clear functions, and the like. Many of these preferenceshave
evolved graduallythrough the more dynamic developmentof Western
society, but they are completely in harmony with the principles of
ancient Greek art. Indianslike indefinite beginnings,circularand spiral
designs,ambiguousfunction, abundantornamentation,organicdevelopment, an intensely-saturatedfield, reticulatedpatterns,the concurrence
and confluence of divergentstrands,and an entirely differentbrandof
teleology. This set of cultural preferences was firmly establishedin
Indianart by the time their musicalsystem was set down and is clearly
revealedin the basic organizationof the system and the specific principles for constructinga melody.
I will mention five other specific differencesthat revealmajor gaps
between early Indianand Greekmusicalthought:
1. There are no Indian equivalentsto the concepts of modulation
and key, no "changeof place." The contrastis between a stable system
(Indian) and a mobile one (Greek), "fixed do" as opposed to "movable
do."
228

2. Similarly,there is no preciseIndianequivalentto the concepts of

intervaland genus. Withinthe Indianpitch system, the relativeposition
of one component to anotherwas assumed,and greaterstresswas placed
on the relativeposition of that componentwithin the system as a whole.
Lannoy describesthis as a "unifiedfield awareness"that he finds typical
of Indian culture.14Svarascould be consonant,dissonant,or neutralto
one another, but the intervalthus formed could not be defined other
than by the number of srutis it included. It had no separateidentity.
The Greek concept of intervalwas one of their most brilliantachievements and served a variety of functional roles, for example, its size,
consonanceor dissonance,composite or incomposite.
3. The principal topics for Harmonics and Rhythmics furnish no
clues as to the resultingmusicalstyle; style as such aroseonly in the subtopics of melopoeia and rhythmopoeiaand in the separatediscussions
of ethos. But in the Indian system style is explicitly defined by means
of the correlationof variousmajortopics. 15From these correlationswe
can infer (in addition to the inevitable"moderate"style) the existence
in early India of two styles familiarin Indian music today: citra, an
exciting style modeled upon instrumentaltechnique, featuringword
repetitions, a persistent fast tempo, short phrases, frequent syncopations, and irregularaccents; and daksina, a slow and expressivevocal
style, appropriatefor an openingdlapanaimprovisation,with prolonged
syllables, a slow pulse, extended phrases,and expressivedeviationfrom
the beat.
4. The Indian theorists set down a more fully-developedtheory of
musical structure than their Greek counterparts,who relied on the
structuresof formalpoetry. The seven IndiangTtakaforms are probably
the oldest musical structures to be substantiallyindependent of the
principlesof prosody. These were large,fixed, modularforms specified
down to the individualbeat, showing few traces of the later cyclical
patterns that served (and still serve today) as frameworksfor improvisation.16
5. The word does not play an important role in Greek musical
speculation. Although rhythmics and metrics were consideredtwo of
the three main divisions of music and sharedthe same patterns,their
domains were kept quite separate. And none of the stated topics of
harmonicsor rhythmicswas in any way dependentupon text considerations. In early Indianmusicalthoughtpada (word) was similarlyone of
the three main divisionsof music, alongwith svaraand tdla,but several
of the main topics in the latter domainscould only be defined by means
of the text. Indianmelodic theory is obviouslymore word-conscious.
Thereare othercharacteristicdifferences:the Greeksstressedperception, the Indiansdid not. Whileboth groupsof authorsseem obsessedat
times with such technical matters as assigningnames, taking up topics
229

in their proper order, and enumeratingall possibilities, their general

approachto the subdivisionof a topic is somewhatdifferent:the Greeks
analyzed a topic by means of its differences,that is, how an intervalor
a rhythmic foot might differ from another-in magnitude,in composition, in structure,in genus, whether rationalor irrational,and the like.
Indian authors outlined a topic in terms of its laksanas(characteristics),
its gunas (qualities), its ahgas("limbs,"components),and variousother
sets of attributes.The Greek strategy resultedin a kind of intellectual
dissectionthat alignedall possiblevariantswithinthe system in mutuallyexclusive groupings;the Indian strategy was to list a large number of
positive (sometimes overlapping)attributesfor each topic, then to subdivide each of these in the same manner,but generallyavoidinga clear
"either/or."
THE THEORYOF MELODY
Earlymelodic theory survivesin the form of little more than a framework of ideas, although these ideas point clearly in certain directions
and give some indication of the elaborate nature of this branch of
ancient musical thought. It will be useful here to make a basic distinction between "precompositional"and "compositional"melodic theory:
by these terms we mean (a) melody as availablegamut and (b) melody
as pathway, the particularchoices one is allowed to make from the
established gamut. In Greek and Indian treatisesboth of these aspects
of the theory of melody are highly prescriptive,and we may take both
"composition" and "performance"(whether improvised or not) to
mean the arrangementof authorized patterns in accordancewith the
generalguidelinesof the system.
In the Greek scheme melopoeia (usually translated as "melodic
composition") is the final topic in the domain of harmonics. The
available gamut is laid down in the previous harmonic topics, so it
remains for melopoeia to outline the variouscompositional options.17
Similarly,rhythmopoeia is placed as the final topic of rhythmics, and
its subtopics are the same as those of melopoeia. It is unfortunatefor
our purpose that these sections are treated only superficiallyor are
among the missing portions of the survivingtexts. Figure 1 summarizes
the divisionsof melopoeia:18
We should note in passingthat, due to the essentiallymonophonic
nature of Indianmusic, the Sanskritlanguagecontainsno specific word
for "melody," as opposed to "song" or even "music."But, in contrast
to the sketchy set of options for melopoeia, much more of the early
Indiantheory of melody remainsintact. The relevanttopics are found
in the domain of svara, although, as we have pointed out, the final
delineation of style was accomplished by correlating the melodic
230

Melopoeia,melodic compositi
lepsis, selection
chresis,usage
ploke, network
agoge,progression
1
pettei
eutheia, anakampousa, peripheres,

bent
roundabout
straight
(ascend)
(descend)
(returning)
the tropoi ("styles") of Melopo
nomikos,
elevated
dithyrambikos,
moderate
tragikos
depressed
systaltiken
contracting
Figure1. The Theory of Melod

to)

product with various characteristicsof t/la.19 In the formal lists of

svaratopics as givenby Bharataand Dattila,we find no clearseparation
of the compositionaland precompositional;Table 2 makes such a separation and reordersthe appropriatetopics in what to us seems a more
logical sequence.20
The topics listed in the left column provide for a three-octavescale
with specific degree names, yet flexible enough to be divided and
inflected in a variety of ways and devoid of any implicationsfor such
things as tonal emphasis, boundaries,registers, or idiomatic patterns.
Like the Indian concept of matter, the melodic gamut was a neutral,
undifferentiated field; it could be marked off for convenience with
the sol-fa svara names [SA, RI, GA, MA, PA, DHA, NI] and in selected intonations, but these could be exchangedas easily as one erases
chalk-linesfrom a blackboardor moves the frets on a sitar.Therewere
two main grnimas,diatonic scales from D and G, plus a hypothetical
were formedby mechanicalrotathird scale from E.21The muircchands
tion through the scale degreesand seem to have carriedno functional
implications; if rendered in hexatonic or pentatonic versions, these
became the tdnas.SadhIrainaprovidedfor two intermediate"chromatic"
tones, each producedby raisingthe appropriatesvaraby two srutis.22
This Indian equivalentof musicaficta was limited to the svarasGA and
NI:
SvaraName
Equivalentto
antaraGA
F?, treatedas a "leadingtone" of MA (G)
kdkaliNI
CM,treatedas a "leadingtone" of SA (D)
of Table 2, as amplified by their
column
in
the
The topics
right
divisions and subdivisions,set out a comprehensivetheory of melodic
composition that (a) builds on the establishedmelodic gamut, (b) describesa repertoireof eighteen authorizedmelody-types, (c) states the
options for melodic motion within phraseunits, (d) establishescertain
melodic functions, (e) provides an elaborate assortment of melodic
ornaments,and (f) shows how all of these ingredientsmay be combined
so as to resultin certainapprovedmusicalstyles.
The most vital choices are those which direct the course of a melody
into one of the prescribedjifftis,the ancestorsof the modem concept of
rffga.The word ijfti is one of those bland words so useful in musical
terminology;it is a past passiveparticipleof the verbalroot jan [cognate
with the Greek word genesis] meaning "to be born, arise,"and thus its
developedmeaning: "kind, type, species." A glanceat the standardten
characteristicsof jifti discloses most of the familiarstandardsby which
a mode is recognized in MediaevalWesterntheory-incipit, final, confinal, and ambitus(high and low):23
232

TABLE2
PrecompositionalTopics
Composi
12 sruti, intonation,microtones
13 jiti, p
1 svara,note, tone
- vYdi,
2 grfma, scale
9 varna
3 mifrcchani,modes
4 tifna,deficientmi7rcchanis
5 sthina, register
8
alteredtones
sidhatrana,

10 alarh
6 vrtti,
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
graha,initial
arhs'a,prominent,usually called "sonant"by Indianauthors
tffra,high
mandra,low
hexatonic
sa~dava,
auduvita,pentatonic
alpatva,scarce,weak
bahutva,copious
nyffsa,final
apanyisa, confinal, an internalcadencetone
There were obviously numerous possibilities for producingvarying

degrees of tonal emphasis and distinguishingone jdatifrom another.
Indian authors were fond of enumeratingaggregatesfor pedagogical
purposes, and their figures are revealing:for the eighteen jiftis, there
were twenty-one possible finals, fifty-six confinals, sixty-threeinitials,
and sixty-threesonants!2 No such enumerationswere given for copious
or scarce svaras.It is not surprisingthat initial, sonant, and final coincided at times, but their functions were clearly separate.Nydsa (literally, "to sit down on") signifies final, consummation, and rest but
with no implications for profusion or any kind of control over the
remainingtones. Bahutva ("muchness")implied no more than copious
use. Arhia (sonant) comes the closest to the Westernconcept of "tonic"
and seems to have more in common with the initial than the final;25
althougharhzsaand graha are defined as separatefunctions, they are the
identical tone in all but one of the jftis.26 Apparentlytonal ambiguity
was less to be tolerated at the beginningthan at the end!
Arthsa,accordingto the Natyas-fstra,had its own list of ten laksanas:
it is the generatingtone, it determinesnot only the low tone but the
interval between low and high tones, it is the tone most frequently
heard, it determinesthe initial, the final, the three types of confinals,
and is the tone which all the others follow."2Aitha, one gathers,is no
trivialconcept!
Apanyfsa signifies a strong degree that could be used for cadences
on interior phrases. Bharata mentions two other types of confinals:
sanya-sa,which could not be dissonantto the arhizaand which was used
as the cadential tone in the first phrase,and vinydsa,a tone placed at
the end of a singleword within a phraseand which could be either consonant or neutralto the arzsa.28
The otherjiti-laksanaswere not as compJex:certainof the jiftiswere
restricted in range, but the generalprinciplelimited movement in the
low octave to one of the strong degrees or the tone below the final;
movement in the high octave was not allowed beyond the fifth tone
above the sonant. Certainof the jfftis could only be performedin their
234

full, heptatonic form, but others were permissiblein hexatonic and

pentatonic versions (when the note or notes to be dropped were not
those vital notes consonant with the arhisa).Bahutva and alpatvacould
be effected in a numberof ways, apartfrom mere statisticalpresencein
a melody: emphasisresulted not only from prolongationor repetition
but from using a tone as a turningpoint or the extremity of a melodic
figuration;deemphasis,similarly,resulted from a lack of repetitionand
progressingdirectly over the tone to be slighted.
Later authors added an intriguingconcept, antaramirga("internal
path").29 Many hints are dropped in the various texts, but nowhere
do we find very specific information on this most important subject.
It obviously involved the correct applicationof the laksanasprescribed
for an individualj/ti; it is also said by variousauthorsto involvecertain
tonal pairings which were emphasized in a back-and-forthmotion,
saficira. The result must have been similarto what today is called the
calan of a riga, a deeply-etchedindividualpath and a set of idiomatic
melodic patterns that distinguishedone jfti from all others. We will
returnto the idea of antaramirga.
Vffdior "sonance"is treated formallyas a subtopic of grfma (scale),
but the concept first becomes operationalwhen appliedto the structure
of an individualjiti. 0 Hereare the four possibilities:
1. vidi, sonant, "rulingnote" (arhisa)
2. sarhvidi,consonant,harmonicaffinity
3. vivddi,dissonant,distorted
4. anuvidi, neutral
The commentatorAbhinavaquotes an old analogy: "Valdiis the king,
sarhvidi is the ministerwho follows him, vividi is like the enemy and
should be sparinglyemployed, anuvadidenotes the retinue of followers."31Like most analogies,this one is dangerousto push too far! While
the sonant is clearlyidentifiedwith the athisaand servesas a fixed point
of tonal reference for a jiiti, accordingto the principal authors, the
remaining tonal relationships are defined as reciprocal relationships
between note-pairs.Sarhvffdi(consonant)affinity existed between pairs
either nine or thirteen srutis apart, intervalsequivalentto the perfect
fourth and perfect fifth. This relationshipentailed important consequences for range, structure,and "fullness"of a jiti, in that the note
which is consonant with the sonant could neither be dropped nor
deemphasized,thus setting up a kind of dominant/tonicaxis; it further
served as a natural limit of motion in the high and low octaves. An
interesting complication is that consonant affinities existed between
other note-pairsand presumablyinfluencedthe melodic motion in some
other, unspecified ways. Dissonant note-pairswere two Srutisapart, a
"half-step" that varied somewhat in size in the Indian nontempered
235

system of tuning; today the term signifies a foreign note in a riga.

Anuvodis were neutral in function and thus were not subject to any
stated constraints.
Thus far we have been discussinghighly functional aspects of the
theory of melody. The topics varna(basic progressions)and alarhklra
(ornaments) were more neutral in that they were not prescribedfor
any of the individualjftis and could be widely employed in different
melodic contexts. Varnais to alarhkirawhat structureis to decoration,
and both were reckoned on the basis of text units. The list of four
varnasis reminiscentof the divisionsof the Greektopic agoge:32
1. sthjyT,stable
back-and-forth
2.
saflciri, wandering,
3. JrohT,ascending
4. avaroh', descending
There is probably no precise distinction to be made between varnza
and alarhkfra, although any Indian musician can improvise typical
examples;the terms are as elusiveto define as "motive"and "gesture."
Varna(appearance,color, word, syllable) is clearly the largerunit and
were
indicates the generaldirection of the melodic line. The alarhkdras
melodic flourishesor gracesthat were, in certaincases, associatedwith
certain varnas or which could occur in the middle of an established
varna.The list of thirteen alarhkdraswhich follows is Dattila'sabbreviand
ated list;33 the Natyasrstra enumeratesthirty-threealarhkJras,M4
later lists go as high as eighty-eight.Indianscholarstoday speculatethat
certain combinations of progressionsand ornamentswere specific to
certainjftis, but the texts yield no supportingevidence for this claim.
One can only assumethat such matterswere reservedfor the guru/lisya
tradition of oral instruction on which the transmissionof the Indian
musicalsystem still largelydepends.Dattila'slist is as follows:
1. prasannidi,beginswith low note
2. prasanninta,ends with low note
3. prasannamadhya,low note in the middle
4. prasannidyanta,begins and ends with low note
5. bindu, highernote touched like lightning
lower note touched quickly
6. nivrtta-pravrtta,
7. preikholita, even swingbetween two notes
8. tira-mandra-prasanna,
gradualrise followed by suddendrop
suddenrise followed by gradualdescent
9. mandra-tira-prasanna,
10. sama, even ascent and/or descent
11. kampita,quiverin low register
12. harita,quiverin middle register
13. recita, quiverin high register
236

Of all the Indian cultural preferences expressed in the musical

system (some in the domain of pitch, others in the temporal dimension), none is more evident than this addiction to ornamentation.
Bharata,in a famous verse, states that "A song without any alarhkira
will be like a night without moon, a riverwithout water, a vine without
a flower, and a woman without any ornament."3sIt is true that an
Indian woman feels undressedwithout some ornament, and one can
point to other relevantculturalideas such as the heavy role of spicesin
Indian cuisine and the absence of a concept of "naturalbeauty." The
word alarhkfframeans literally "to make sufficient," clearly implying
that ornamentationis a vital aspect of the artistic process-not mere
frosting on the cake. The theory of music, just as eloquently as architecture and sculpture, tells us that structureand decorationjoined to
form a concept as basic to Indianart as it was secondaryin Greekart.
The lists of ornaments reveal typically Indian decorativepatterns:
arabesques,circles and spirals,and reticulatedlines. Whilemany of the
listed ornamentsare similarto some of the gamakasin modem Indian
music, others form largermelodic units that Nijenhuiscomparesto the
melismaticBlumen or melodic "garlands"of MediaevalGermansong.36
The common factor was their conception as ornament,not the length
of the unit. Bindu and nivrtta-pravrttaare commonly employed in
Indian singingtoday and are practicedin systematicexercisesby every
beginning student; when performed at "lightning"speed, they are
responsible for the highly articulated, "textured" style of melismatic
singingused for virtuosopassaggi.Wemay note also (in kampita,harita,
and recita) the survivalof vibrato as an ornamentfrom older traditions
of Vedic chant.
To conclude this discussionin as concrete a way as possible, I wish
to present the gamutand minimalconditionsfor two of the majorjatis.
one (perhapsthe most importantone) of the seven purejitis; its
name
indicates its derivationfrom the sadja-grdma,and it is equivalent
.SjffdTis
to the ecclesiasticalDorian mode in more than superficialaspectsof its
structure.Its gamutis as follows:
SA
RI
GA
MA
PA
DHA
NI
The basic conditions are these: (1) all seven degrees must be used,
(2) the final is on SA in the low octave, (3) SA is also the initial, confinal, and sonant, (4)
may appearin hexatonic form by dropping
appearin pentatonic form. The majorvariables
s.dj7
NI,37but (5) it cannot
are (6) all degreesexcept NI and RI are potential sonants,38(7)GA and
237

PA are acceptable confinals, (8) typical of this jiti is sahgati(association) between SA and GA, also SA and DHA-apparently part of the
"internalpath."39Bharataadds that (9) NI and RI are scarce, while
(10) GA is copious." The later author Matafigastates further that
(11) only the lowest five notes of the high octave may be used, and
(12) that this jfti's emotional affect is appropriatefor three of the
rasas:vbra,the heroic; raudra,the furious;and adbhuta,the wondrous.
And the thirteenth-centurySarhgTtaratnikara
adds that (13) kffkaliNI
used
in
if the sonant is GA, NI
be
the
form,
may
heptatonic
(14)
[cg]
cannot be omitted, and (15) this jti is traditionallyused for the exit
song in the first act of a play.
There can be no conceivable reason why this set of melodic prescriptionsshould have anything in common with the MediaevalDorian
-unless we admit the possibility of certain instinctive (and perhaps
"universal")tendencies, such as the stabilizing influence of perfect
intervals and the desire to cadence by small melodic intervals. So
with the appropriateamount of suspicion for cross-culturalcoincidence, it is still strikingto note these structuralfeatures:a cleartonic/
dominant axis, oscillating movement around the final, deemphasis
of degrees 2 and 7, copiousness of the third degree, the limitation
of range to the fifth degree in the high octave, the optional musica
ficta leadingtone, and the rathersober ethos prescribed.
By way of contrast, Nandayant7("bringingjoy") is a mixed jti. 41
It was held in very high esteem, and Abhinavaquotes an old saying
that "the j/ti NandayantT,performedeven once in accordancewith the
injunctions governingits forms, can expiate the supremesin of having
murdereda Brahman."42Its gamutis as follows:
GA
MA
PA
DHA
NI
SA
RI
Nandayant'iwas the only jiti in which the sonant and initial were not
the same svara. Its conditions are (1) GA is the initial, (2) GA is the
final, (3) PA is the sonant, (4) MA and PA are the confinals,(5) it may
become hexatonic by droppingSA, but (6) a pentatonic versionis not
possible, (7) low RI is used copiously (!), (8) the upperoctave may not
proceed beyond SA, (9) it is related in structureto four other gjtis43
but does not share their idiomatic patterns, and (10) the appropriate
rasasare srhgara,the erotic;hisya, the comic; and karuna,the pathetic.
Nandayanti has little in common with the WesternLydian, more
perhaps in common with the sixth of the reciting tones for Roman
238

psalmody.Semitones(4/5 and 7/8) are obviouslyvital to the expression

of this j'ti, since RI is used copiously and DHA may not be dropped;
perhapsthis helps to account for the more exotic ethos of NandayantT,
an ancestor of the more emotionally expressiveeveningragas.The significanceof separatingthe initial and sonantis not clear,especiallysince
NandayantTis unique in this respect. Like some of the other mixed
jotis, NandayantThas but one possible sonant; its conditions are thus
more specific than those governing
So even among this set of
distinction between (a) jatis that
early melody-types we can observe a .sajd/1
allow for variabletonal emphasis,and (b) jatis with a singleprescribed
tonal focus.
Most musicians realize instinctively that the simplest phenomena

require the most profound explanations, and this is perhapswhy the
theory of music reveals a persistent tendency to approachthe most
basic questions in extremely complex ways. A great teacher once said
that most questions in musical analysis reduce to one of these two:
"Whatis it?" and "What'sit doing here?" The essential tasks for early
music theory were the differentiationof varioussound phenomenaand
the assignmentof meaningin terms of identity andfunction. Precisely
these were the immediate goals of ancient Indian music theory and
hence the emphasison giving names and drawingdistinctionsbetween
one component and another, or between one melodic pattern and
another.
The separation and naming of components is one of the more
primitive tasks of a musical system, but assigningmeaningis a higher
achievement. We have seen that Indian authors cultivated a certain
ambiguity in musical function, just as the Greek authors strove to
minimize ambiguity. But apart from such basic preferences,musical
meaningarisesat differentstagesin the two systems:most of the Greek
melodic theory is precompositional,and melodic function (in termsof
the thetic and dynamicroles of the scale degrees)is built into the gamut
itself, not reservedfor the individualmelody; melopoeia has no functional terminology. In Indianmusical speculation,function arisesonly
when melody becomes specific; the topics that outline the melodic
gamutare virtuallydevoid of functionalimplications.Thus,
Greek
Indian
functionalgamut
nonfunctional,neutralgamut
neutralcomposition
function arisesthroughcomposition
One can only speculate why this is so. One of the primarycauses
must have been the differingrequirementsof a mobile system (Greek)
239

with a sliding scale of degree functions and a concept of modulationin

contrast to a stable system (Indian) in which the scale degree was a
more neutral concept. The idea of music was obviously freerin Indiaperhaps even then largely improvised-and, as Deva argues, function
may have been expressedmore in terms of a local relationshipamong
membersof a small pitch set, or of an individualmelody, than in terms
of a global relationshipamongthe membersof a whole system or which
governedan entire repertoireof melodies.44Function in ancient Greek
music, no matter how it evolved, is expressedin such globalterms and
is clearly an intellectual tour de force; its relationshipto the repertoire
is not known. In the prescriptionsfor the jftis we can see the development of certain basic specializedtonal functions: what Szabolcsi calls
the "groundtone" and the "pivotaltone," as well as an arrayof other
specific melodic functions.45
Ancient music theory may be viewed as the record of two parallel
evolutionaryprocesses-the evolution of a system and the evolution of
a repertoire. In any single treatise we may find musical concepts and
descriptionsof actual pieces embedded like fossils in a layer of rock.
The early Indiantheory of melody may be readfor suchtraces,although
many blanksremainto be filledin. Themelodicgamutof earlyIndiatells
us next to nothing about the repertoire,but it is a highly informative
record of the progressof the system. It had evolvedinto a true diatonic
system of great sophisticationwith many permutationaltechniques;it
shows only faint tracesof its prehistoricevolution-no specific evidence
of upwardor downwardgeneration,formationfrom melodic kernelsor
nuclei, chains of intervals,downward"cascades,"and similargenerative
processesthat areclearlyvisiblein the structureof the olderVedic scale.
The theory of melodic composition, as prescribed for the jfatis,
reveals traces of several stages in the development of the musical
repertoire:(1) a fairly primitivestage in which the concept of melody
is scarcelydifferentiatedfrom the gamutitself, (2) a set of proto-melody
types (such as sjdcli)that appearto have controlled a repertoirelarger
than the individual tune, (3) more specific formulae (such as NandayantT)with sharplyindividualcharacteristics,parentsof the more recent
rigas, and (4) a small collection of musical styles in which various
melodies could be sung.
Before the theory of melody could demonstrate any convincing
relationshipto this evolvingrepertoire,theoristshad to find a means of
objectifying the referential set of "grooves" into which acceptable
melodies fell and which underlie such concepts as mode, maqam, and
riga. Perhapsthe most admirableachievementof early Indianmelodic
theory was the definition of melodic path againstthe backgroundof
the establishedmusical gamut, explicitly and brilliantlyrealizedin the
theory of jfti and such ideas as antaramdrga.
240

If these concepts have any analoguein the music of ancient Greece,

it is the nomoi (literally, "laws"). Although the nomoi are frequently
mentioned in Greek literature, their musical characteristicsremain a
mystery, and they take no part in the theory of melopoeia." The
surviving fragments of the Greek theory of melody serve more to
outline melodic options than to record the repertoire;nowhere do we
find such concepts as melodic path or the description of specific
melody types. Perhapsthis helps to explain the enormous confusion
that has accompaniedthe applicationof the conceptual frameworkof
Greek theory to variousforeignmelodic repertoiresin the MiddleAges.
If, as some have claimed, the Greek musical system is the result of a
systematic codifying of the propertiesof a repertoire,the processis by
now quite impossible to trace. This void enhances the value of the
Indiantheory of melody, in which successivestageslie open to our view
and detail some of the vital early steps in the formationof a music.
I
\\
~---~
// Y
/%;~P\V~;J/
/?,
1
?--?'\"1,
'c.
I
'k ,CI?/??I
--
;e
241

NOTES
1. See Walter Kaufman, Musical References in the Chinese Classics (Detroit:
Information Coordinators, 1976) and two works by Babette Minnie Becker,
"Music in the Life of Ancient China as Reflected in the Ceremonial Books:
The I Li, the Chou Li, the Li Chi" (M.A. thesis, University of Chicago, 1954)
and "Music in the Life of Ancient China: From 1400 B.C. to 300 B.C."
(Ph.D. dissertation, University of Chicago, 1957).
2. Hereinafter N. For a study of this treatise see Lewis Rowell, "A Siksj for the
Twiceborn," Asian Music 9/1 (1977): 72-94. The only available edition is
with the commentary Siksivivaranopeta by BhattaNarada, Naradyda
sobhikara (Datia: Sri
Pitambarapitha Sanskrit Society, 1934).
Siks.,
3. Hereinafter B. The most reliable of the many editions is Bharata, Natyaaistra,
with the commentary AbhinavabharatT by Abhinavagupta, ed. M. Ramakrishna Kavi and J.S. Pade, 4 vols., Gaekwad's Oriental Series No. 36, 68,
124, 145 (Baroda: Oriental Institute, 1926, 1934, 1954, 1964). For an
English translation see The Natyasastra, trans. Manomohan Ghosh, 2 vols.
(Calcutta: The Asiatic Society, 1951, 1961).
4. Hereinafter D. For the text and an excellent English translation see Mukund
Lath, A Study of Dattilam (New Delhi: Impex India, 1978).
5. See Note 3.
6. Although the roots of this doctrine are extremely ancient, it was first articulated in Mataiga's Brhaddes' (8th century?).
7. For more details see Rowell, "A
8. N 1, 5, 1-2.
SiksJ.."
9. N 1,5,3-4.
10. N 1, 2, 7.
11. Ananda K. Coomaraswamy, "The Theory of Art in Asia," in Aesthetics
Today, ed. Morris Philipson (New York: World, 1961), p. 36.
12. Vatsyayana, The Kama Sutra, trans. Sir Richard Burton and F. F. Arbuthnot
(St Albans: Panther, 1963), pp. 20-23.
13. One finds some variation in the order of pitch topics from text to text. This
list is taken from Cleonides, "Harmonic Introduction," in Source Readings in
Music History, ed. Oliver Strunk (New York: W. W. Norton, 1950), p. 35,
and B 28, 13-14. D 6-7 begins with druti, omits dhatu, and inserts ja ti after
sadharana. The list of rhythmic topics is that found in Aristides Quintilianus,
De Musica Libri Tres, ed. R. P. Winnington-Ingram (Leipzig: B. G. Teubner,
1963), pp. 31-40. The lists of tala topics given in B 28, 20 and D 110-112
vary wildly, and B treats each of the eight tala beats as a separate topic.
14. RichardLannoy, The SpeakingTree:A Study of IndianCultureand Society

(London: Oxford University Press, 1971), p. 278.
15. The most complete version of such correlations appears in the 13th-century
Book 6, verses 165-170:
SahgTtaratnakaraof SMrrigadeva,
242

Citra
Vrtti,style:
instrumental dominates
fast
Laya, tempo:
Yati, regulated
flow of tempo:
Pdni, phasing:
Mode of tala:
Giti, treatment
of texts:
Marga,standard
even tempo
accenting after the
beat
one-beat units
frequent word
repetitions
2 matras
Vrtti
equal blend
moderate
Daksina
song dominates
slow
decelerating
accelerating
metrical accents anticipating the
beat
two-beat units four-beat units
many long
many short
syllables
syllables
4 matras
8 matras
pattern length
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
The gTtakasare described in B 31 and D 155-222.

See Table 1A, topics 1-5.
See Aristides Quintilianus, De Musica, pp. 28-31.
See Note 15.
See Note 13.
Although Indians routinely translate grama as "scale," the term should be
understood in its broadest sense as a pitch "collection" indexed on or around
a particular tonal referent. For a particularly lucid explanation of the problem see Harold S. Powers, "An historical and comparative approach to the
classification of ragas (with an appendix on ancient Indian tunings)," in
Selected Reports, vol. I, no. 3 (U.C.L.A. Institute of Ethnomusicology, 1970),
pp. 1-78, especially pp. 59-63.
Efforts to establish the exact size of the frutis have so far failed to convince
most scholars.
B 28, 66 and D 55-62.
B 28, 79-91 and D 90-91.
The etymology of arhva illustrates a typical Sanskrit semantic twist in the
development of a specialized technical meaning: the basic meaning is "a share
or portion of anything, the denominator of a fraction, a degree of latitude or
longitude," but in its application to the svaras, ariha has evolved to mean
"that which divides" or "the principle by which a thing is divided."
Nandayanti, see below.
B 28, 68-69.
B 28, 73.
Both B and D refer to antaramarga but do not define the term.
B 28, 23-25 and D 18-19.
Commentary on B 28, 24.
B 29, 15-16 and D 97-99.
D 99-108.
B 29, 20-45.
B 29, 45.
36. E. Wiersma-TeNijenhuis,Dattilam:A Compendiumof AncientIndianMusic

(Leiden: E. J. Brill, 1970), p. 314.
37. One assumes that this variant appears by license since it contradicts the first
prescription.
38. The contradiction of (3) has not been adequately explained.
243

39.
40.
41.
42.
D 63-64.
B 28, 97.
D 85-86.
Commentary on B 28, 9-10.
43. pahcamT,arsabh7,gandhafi,andandhr7.
44. B. Chaitanya Deva, Psychoacoustics of Music and Speech (Madras: The Music
Academy, 1967), p. 67.
45. Bence Szabolcsi, A History of Melody (London: Barrie and Rockliff, 1965),
p. 37.
46. See Solon Michaelides,The Music of Ancient Greece: An Encyclopedia

(London: Faber and Faber, 1978). pp. 222-223.
coo
CeC
0`
~jrlb
244


Early Indian Musical Speculation and The Theory of Melody - Rowell, L.

Uploaded by

Copyright:

Available Formats

Early Indian Musical Speculation and The Theory of Melody - Rowell, L.

Uploaded by

Document Information

Original Description:

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

Early Indian Musical Speculation and The Theory of Melody - Rowell, L.

Uploaded by

Copyright:

Available Formats

What are the main topics covered in the theory of melody according to the text?

What are the main topics covered in the theory of melody according to the text?

What are some of the principles used to structure pitch and time in early Indian musical theory?

What are some of the principles used to structure pitch and time in early Indian musical theory?

Yale University Department of Music

This content downloaded from 151.100.161.184 on Thu, 30 Oct 2014 10:43:33 AM