Sruti in Ancient, Medieval and Modern Contexts


`Sruti' is one of the concepts in music theory which has lent itself to much discussion from ancient times to the present. We find Sruti at the centre of very serious disputes between Musicologists in our own times. In their zeal for establishing the validity of their own views we find scholars attributing to ancient writers like Bharata and SArngadEva theories and notions which they did not hold.The object of this paper is to clarify the concept of Sruti as propounded by the ancients, namely Bharata, Dattila and matanga and elaborated by later writers like Abhinavagupta and SArngadEva and to discuss the manner of the use of the term in the modern context.

`SrUyanta iti Srutayah' - This is how the word Sruti is etymologically derived,and therefore Sruti refers in general to all that is heard. The vEda-s are called Sruti since they are said to have been `heard', not composed by any mortal. Sruti also means the various ways of intoning svara-s in music (as used in nAradIyASikshA ). It is applied to the quality of the voice which changes due to the predominance of one or other bodily humours- vAta, pitta, etc. (Br.D p.4). As pointed out by some commentators, the word also denotes the different registers into which sounds fall (Br.D p.4). Thus we find the word Sruti being used in various senses, all having to do with sound, especially musical sound. We are however concerned with only one part of the connotation of the word Sruti, namely, Sruti as a measure of svara-s.

A technical term like Sruti cannot be understood if we remove it from the system or systems in which it functioned. In order to understand the concept, we have to acquaint ourselves with the entire musical system spoken of by Bharata and other ancients and the changed theoretical set-up in which medieval writers like rAmAmAtya used the term. And so any discussion of the concept will necessarily entail reference to other related elements of the system, namely grAma and jAti, mEla and rAga. The theoretical system which was formulated to codify the music of the ancient period can be called the "grAma system". jAti-s and mUrcchanA-s furnished the melodic basis of ancient music. These svara groupings were classified into two grAma-s. Sruti was the unit of measure of svara-s, and also the basis on which svara structures were classified into two grAma-s. Bharata states this clearly ---

jAtibhih SrutibhiScaiva svarA grAmatvamAgatAh
NS 18, 5-6.

This statement literally means "Through jAti-s and Sruti-s, svara-s attain the state of grAma. svara-s were defined as to their magnitude, i.e., as two-Sruti-ed, three-Sruti-ed or four- Sruti-ed and also their sequential order, i.e., rshabha that svara which is between shaDja and gAndhAra. A grAma is a theoretical construct, a purely theoretical arrangement of svara- s, which could fit several jAti-s, which are therefore said to belong to that svara-group or grAma. The order of arrangement of the svara-s is the same in both the grAma-s, the difference lying in the magnitude (expressed in terms of Sruti-s) of the pancama of one grAma from the pancama of the other. Thus the difference between the two grAma-s hinges on a measure called Sruti. Sruti, grAma, etc. are theoretical concepts which were postulated for classifying melodies. Melodic structures were classified on the basis of the magnitude of their component svara-s measured in terms of Sruti-s, into two grAma-s or groups. In fact it is for the purpose of classification that the concept of Sruti was formulated, it had no existence in performance. Abhinavagupta, the commentator on nATyaSAstra points this out again and again.

grAmavibhAgArthmEva SrutikIrtanam
AB on NS vol.IV, p.12, l.11

Evam grAmadvayam tadupayOgi ca Sruti sadbhAvE svarANAm SrutiniyamaprAmANyabhidhAya ....
ibid, p.24, l.4

As a unit of measure of svara-s, what does Sruti mean? In the ancient system, svara was conceived not merely as a sound of fixed pitch position, but also as comprehending the entire tonal range between itself and its previous svara. The interval which separated one svara from another and which was included in the definition, was measured in terms of Sruti-s. The interval between one svara and the next can be theoretically divided into infinitesimal parts, i.e., any number of sounds of very minute pitch difference can be produced between one svara and the next, but there are, according to the ancients, only one or two and at the most only three sounds which can be distinctly cognised by the ear in the interval between one svara and the next. These distinctly congnisable sounds or the intervals thereof, between one svara and the next, are Sruti-s. The term Sruti is appropriately given, for it is of a perceptible, audible interval sound that the ancients are speaking. Thus a svara is said to be of the measure of two Sruti-s (dviSrutika), when there are two Sruti-s separating it from the previous svara, triSrutika and catuhSrutika if there are three and four Sruti-s respectively between itself and the previous svara.

We saw above that svara was conceived both as an interval and as a pitch position. Similarly Sruti was thought of both as the least audible interval between two sounds, as well as the sounds themselves which were separated by such an interval. We find Bharata speaking of Sruti as the least perceptible difference in pitch between two sounds, while Dattila (Datt.9- 10), who is followed by matanga (BrD p.4) and SArngadEva(SR 1,2, 8&13), uses the term Sruti to denote the sounds themselves which are thus distinguished from one another. We might say that for bharata Sruti is basically an interval, while for dattila and SArngadEva it is firstly a pitch position.

Bharata's description of Sruti indicates his standpoint clearly. He first describes the two grAma-s into which jAti-s can be classified.

shaDjagrAma is the group of svara-s having the following intervallic sequence --

shaDja of four Sruti-s, rshabha of three Sruti-s, gAndhAra of two Sruti-s, madhyama of four Sruti-s, pancama of four Sruti-s, dhaivata of three Sruti-s and nishAda of two Sruti-s.

In the madhyama grAma pancama has to be diminished by one Sruti. The pancama is lower than that of shaDjagrama by the least possible difference in pitch. Such dimunition or augmentation is measured in terms of Sruti-s.

Evam SrutyutkarshAdapakarshAdvA yadantaram mArdavAdAyatadvAdvA tatpramANam SrutiH
NS 28, 26-27.

bharata adds that the lowering or raising should be brought about by loosening or tightening of the strings in the case of stringed instruments. The above description clearly indicates that bharata viewed Sruti mainly as an interval, the smallest possible one which distinguishes one sound from another as lower or higher in pitch. abhinavagupta makes it explicit.

Evam tIvramandatva hEtubhyAm yadantaram yO vishEshAvabOdhaH tatpramANam niScAyakam yasyAH sA SrutiH
AB on NS ibid. p. 21, l.25

We find an almost identical explanation in brhaddESI too. In the above passage from bharata, the term pramANa denotes in general the measure of Sruti. The pramANa or measure of Sruti is the intervallic difference between the pancama of shaDjagrAma and the pancama of madhyamagrAma . pramANa is not the value of any particular kind of Sruti such as `pramANa-Sruti'. This term has been coined by modern scholars and often wrongly attributed to bharata. Sruti is a unit like `inch' or `second', and there are no two or three kinds of Sruti. However it is not a precise mathematical measure. Although sound, which is physically measurable, is the medium of music, in music it is not measured as in physical science. Sruti is not a precise, acoustically standardised measure, but a perceptual measure like a spoonful of sugar which is of use in the field of art. It is meaningful in music because it is cognisable by the ear and does not require any instruments for recording it. This aspect of the ear being the sole judge of Sruti has been stressed by ancients. Abhinava states that augmenting or diminishing a sound should be perceivable by the ear.

SrutEH Sabdasya SrOtragrAhyasya utkarshastIvratA apakarshO mandatA
AB on NS vol. IV, p.21, l.24

He is aware that , theoretically, augmenting and diminishing can be negotiated through very minute or literally atomic intervals. But he explains that such infinitesimal intervals cannot be grasped by the ear.

yadyapi paramANutOpyutkarshaApakarshO vA bhavEd dhvanErviSEshastathApi nAsau grhItum pAryatE.
AB on NS vol.IV, p.21, l.28

It is on the basis of this audible unit of measurement that the intervals of the svara-s in the jAti-s were measured, resulting in two groupings of svara-s, namely, shaDjagrAma and madhyamagrAma. The interval between rshabha and shaDja was, as we saw above, one of three Sruti-s, that between gAndhAra and rshabha was of two Sruti-s. The grAma-jAti (and the grAma- mUrchanA ) system was such that within one grAma, scales were formed by taking different svara-s as commencing notes, while the interval between two svara-s remained the same. Hence rshabha was always three Sruti-s away from shaDja and it was often referred to as the three Sruti-ed svara.This means that two distinctly perceptible sounds can be produced between shaDja and rshabha, and the third distinct sound is rshabha. We saw above that Sruti was the term used for these sounds as well as the intervals which separated them. So "triSrutika" applied to rshabha would mean that it is the third distinct sound from shaDja as well as that it is separated from shaDja by three distinct intervals. With the help oftwo vINA-s bharata (NS,vol.IV,p.20) demonstrates the Sruti-s of each svara. When we take all the seven svara-s of grAma into account, the Sruti-s number 22. This number has no sanctity attached to it. What is important in the grAma system is the number of Sruti-s within a svara.

A question which is usually posed regarding the ancient Sruti system is, if Sruti was a standard unit, then were the octaves equally tempered? The answer is "no". Equal temperament means that the octave is equally divided and that all the svara-s are separated from each other by equal intervals. In the grAma system this was not so. The interval between shaDja and rshabha was different from that between rshabha and gAndhAra and so on. But the interval of svara was internally divided into a number of roughly equal Sruti-s. The exact magnitude of a Sruti is of no importance, what matters is that the four Sruti-s within, say, pancama, are equal. Similarly, the pancama of madhyamagrAma has an interval of three equal Sruti-s. It is not to be presumed that this pancama is sung exactly on the third Sruti of shaDjagrAmika pancama. On the other hand, Sruti is demonstrated by the two known entities, the shaDjagrAmika and madhyamagrAmika pancama-s. Sruti does not make svara-s perceivable, but is postulated as a unit to measure their intervallic differences. As such, it facilitates the classification of melodies which are constituted by the svara-s, into the two grAma-s. It is this application of Sruti for the purpose of classification that has been extended to the mEla system of the later period.

Medieval Period :-

The classification of svara groups in treatises of the medieval period is based on the concept of mEla . A mEla is a grouping of svara-s of specific intervals (measured in terms of Sruti-s) which can fit several melodies or rAga-s. We might adapt Bharata's statement to read "Through rAga-s and Sruti-s svara-s attain the status of mEla."

rAgaiSca SrutibhiSca svarA mElatvamAgatAh

Like grAma, the term mEla also means a group. Both terms mean a group of svara-s, but here their similarity ends. While, in the grAma-system, different scales are formed by taking differing commencing svara-s and keeping the intervals constant, in the mEla system, different scales are formed by taking the same commencing svara and by increasing or diminishing the intervals of the subsequent svara-s. So in the mEla system, we find the intervals of svara-s undergoing changes and therefore svara-s are not assigned fixed intervals in terms of Sruti-s. For instance, the intervals of kAkalI-nishAda and Suddha-madhyama would vary depending on the dhaivata and the gAndhAra which precede them respectively. The interval of kAkalI-nishAda would be of three different Sruti values depending on whether Suddha, pancaSruti or shaTSruti-dhaivata preceded it. Thus Sruti as a measure of intervallic magnitude is not fully utilised in the mEla system. The only svara-s which would have fixed intervallic values would be rshabha and dhaivata, which are preceded by shaDja and pancama which do not change their positions. vEnkaTmakhi has noted this point, for he states that there are two forms of sAdhAraNa-gAndhAra depending on whether the Suddha or pancaSruti-rshabha precedes it (CP3,42-43). This kind of thinking has not come down to the succeeding period.

vEnkaTamkhi also makes it clear that the svara-s are first located in pitch positions having known them through lakshya and then the intervals between them divided equally into the respective number of Sruti-s (CP 2,26-44). Thus Sruti-s within a svara were equal and not the entire sthAna divided into equal parts and svara-s located on them. Thus what was mentioned in the context of grAma system was spoken in the context of mEla too.

Modern Period :-

In the modern period the mEla system continues to hold sway. While the principle is the same as vEnkaTamakhi's time, we find the svara-s given different Sruti values, the Suddha-rshabha of three Sruti-s becoming dviSrutika, and the pancaSrutika-rshabha becoming catuhSrutika .The fixing of pitch positions for svara-s is to develop a simpler system for mEla classification. The abstraction of positions such as `Suddha-rshabha' and `catuhSrutika-rshabha ' is merely an attempt to choose positions around which the svara-s in the rAga-s which have them, are centred. Only by limiting the number of svara-s would it be possible to have a simple system of classification in karNATaka music where svara-s are sometimes conceived in a fluid state of existence, as for instance, the gAndhAra of tODI, Anandabhairavi etc., it is very difficult and also not proper to pinpoint a position for the svara.

The second svara of kalyANI is identified as catuhSruti- rshabha for the purposes of classification. This does not mean that it is sung exactly on that position, it is sharpened or flattened often in the contexts of rAGa phrasing. This rshabha cannot be identified as rshabha of so many Sruti-s, but it would only be right to call it the kalyANI-rshabha , since the rAga character would permeate it. A musician who has a grasp of the rAga would be singing the rshabha proper to it, not the catuhSuri-rshabha. SankarAbharaNa and shaNmukhapriya are also classified as having catuhSruti-rshabha, but a musician knows how different the rshabha-s in the three rAga-s sound . It is the classification system which has brought about an identity, a superficial one, between the rshabha-s of the three rAga-s.

Forgetting that the terms such as `catuhSruti' and `shaTSruti' have been invented for the purpose of classification, and taking these as their starting point, some scholars find that the rshaba of kalyANI is not `actually' catuhSruti but falls short of it, and promptly designate it is triSruti-rshabha. Similarly the rshabha of bhairavi is stated to be `actually' triSruti. The fallacy lies in taking catuhSruti-rshabha as a fixed svara, while it is only an approximation affected for rshabha of several rAGa-s. The actual differences between the renderings of the same svara in different rAga-s do not call for such theoretical explanations. As we said, catuhSruti-rshabha itself is a theoretical entity derived as the average position of several rshabha , and to complain that the actual rshabha of rAga-s do not conform to this and to attempt to create more theoretical entities such as triSruti-rshabha, would result in complicating rather than simplifying the classification, and would defeat the purpose of classification itself. If theory were to postulate as many rshabha-s as there are in actual practice, then there would have to be as many mEla-s as there are rAga-s. Then enumeration, rather duplication, and not classification would result and theory would lose sight of its purpose.

Another misconception which prevails regarding Sruti-s is the identification of the twenty-two Sruti-s in the various rAga- s. Some scholars have attempted to locate the twenty-two Sruti positions in the various rAGa-s, for instance, the EkaSruti- rshabha in gaula, dviSruti-rshabha in mAyAmAlavagaula, etc. It is clearly seen that in such an event these twenty-two do not remain Sruti-s but become svarasthana-s or practically svara-s. Sruti-s are not singable entities. If the rshabha is flattened and sung, ekaSruti is used as an adjective to indicate this flattened rshabha, we cannot claim that the first Sruti of rshabha is sung. Sruti-s are employed to indicate the magnitude of svara-s, they are not exactly twenty-two positions occupied by svara-s. The number of positions occupied by svara-s in the course of singing may be twenty-two or more, it may be even more than fifty-three.

Further, there seems to be no point in trying to interpret these positions in precise mathematical terms. Such measured values do not in anyway contribute to the understanding of music. I would refer the readers to the article "Sruti: The Scalic Foundation by R.Sathyanarayana, Sangeet Natak no.17. I quote here the last three paragraphs of the article. -------

"Some misconceptions of the sruti prevailing in modern musicological studies may be briefly mentioned. Thus there is an overwhelming bias for expressing sruti values through mathematical techniques; among the more prominent of these may be mentioned the system of ascribed equations. (2 srutis=16/15; 3 srutis=10/9; 4 srutis=9/8), Bosanquet's method of cyclic order and class, and of equal temperament. At the outset it must be pointed out that frequency ratios or string length ratios etc. are only empirical aids and do not contribute anything to theory. They also do not explain any of the svara or scalic facts or phenomena. The havoc wrought by this method is evident in that to about 250 frequency values already proposed for 22 srutis by different scholars, more are being added. The system of ascribed equations is a failure because these ratios are not ascribed to intervals required by textual authority but to others which obviously do not have these sruti values. Even the ascribed values differ from calculated values. This method repudiates theory, practice and logic. Bosanquet's method is also unacceptable because it involves (a) methodological error; (b) derivation by ascending cyclic fifths; (c) techniques and needs for remedying errors of equal temperament unknown in Indian music; (d) derivation of only the chromatic scale and not 22 srutis; (e) uses ascribed numerical equation; (f) the use of major third as an instrument of cyclic rotation. The system of equal temperament assumes a physical equality or equivalence of all the srutis which is nowhere implicitly or explicitly provided for in our musical theory at any time. Moreover, such equality may be repudiated unequivocally. Even if equivalence is a criterion, there are at least three other equivalences: psychic, aesthetic and musicological, which are more relevant and more important. The values obtained for different intervals here differ widely from their actual ones.

Two musicological theories of the srutis must be pointed out to be very untenable. One is that the 22 srutis can be derived by the quintan cycle. Firstly this requires a range of 13 registers which is beyond sensory perception. It also involves an intolerable amount of clash of intervals (vaditvabhagnata- parinama). It gives only svara intervals and not srutis. The fundamental attributes of a sruti and a svara are completely ignored by this system. The extrapolation of intervals in the higher pitch ranges to the experimental octave necessitated in this method involves loss in tone quality and brightness; it is also unfounded in theory.

Another theory seeks to identify sruti concept with shades of a svara employed in different melodic contexts. It tacitly assumes that the 22 srutis are actually 22 svarasthanas, that sadja is at the first sruti and that everyone of the srutis shall be found used in some raga or another; if not, it has no place in tha scale. Such mistaken ascription clearly arises out of an equation of the tonal inflexion due to gamaka or contextual usage with sruti. This theory cannot be supported because gamaka is applicable only to svaras; gamakas are of infinte subtle variation and this finds usage in raga, by no means in only 22 ways. A single note assumes slightly different and often intangible, indefinable shades due to gamaka in different contexts by different artists indifferent ragas. If these are to be called srutis each svarsthana will then have a large number of srutis, not a definite fixed one, which will be obsolete or will radically change with time and with aesthetic requirements or conventions. All these cannot be designated as srutis. Folk music employs several intevals never used in art music. Furthermore, these shades can find neither a fixed position, nor consonantal or other relations in the scale. Illusory notes, medial shades, augmented or diminished svaras - all these must be explained through the extensivity of svara and not through the sruti concept. The shades obtained by gamaka are vibrato, acent, swing, oscillation, glissando, portamento etc, which is always relative to two tonal limits in a range and are never definable with respect to position or extent. Often certain shades are obtained through illusory notes, association, signification, implication etc. Many of these have no independent existence. It thus violates many fundamental requirements or rules of theory." -----

Thus by interpreting catuhSruti-rshabha as 9/8 or kalyANI- rshabha as 10/9, we are only putting the known facts in a different language. 1, 16/15, 10/9, 9/8 etc. are relative frequency values of sound arrived at on the basis of a formula relating frequency to the string lengths. These are precise physical measures while Sruti is only a approximate measure. Thus they are two different kinds of measures. One should not try to find equivalences between the Sruti and mathematical relative frequencies.

One should either try to define svara on the basis of mathematical ratios or on the basis of Sruti-s. If we measure intervals in terms of Sruti then we should not equate it with mathematical ratio and vice-versa. It would be just like equating a "spoonful" of sugar to 5 "grams'. While a "gram" is a precise measure, a "spoon" is only a rough one where every spoon many not be identical in weight. For a cup of tea a unit like "spoon" is fine while "gram" would be necessary in contexts requiring precision.

It is not proper to say that 16/15 is dvi-Sruti and 10/9 is tri-Sruti. Similarly it is wrong to say that there are three values for Eka-Sruti. In fact, as pointed out by Dr B C Deva in his article "Continuity in Music and Sruti" (p.100) in the book "Music of India: A Scientific Study" (Munshiram Manohar Lal, Delhi, 1981) there are at least 7 values between 1 and 16/15. There is no textual sanction for terms like `pramANa', `nyUna', `pUrNa', `mahatI' and `upamahatI' for denoting "varieties" of Eka-Sruti, met with in the books by modern scholars.

It will be pertinent to quote Dr B C Deva (p.102, ibid) here. -----

"2. Are the sruti-s equal or unequal ?

Expressed as ratios they are unequal. We have seen this particularly in the case of one-sruti intervals. If they are unequal what are their mathematical values ? Much discussion and difference of opinion obtain on this problem. However, most of these opinions are theoretical with very little practical evidence.

But we have also seen that when ratios are expressed as additive numbers like sruti-s, they are logarithmic and equal.

In other words, sruti-s are both equal and unequal.

In so far as they indicate a position in an octave (as ordinal numbers), they are equal. In this capacity they do not measure but show only a position in a series of pitches. But each position or sruti may have many close ratios which we measure. In this capacity they are unequal. Much confusion can be avoided by follwoing this idea." ----

Thus if `Sruti' is the unit adopted for measuring svara intervals then it is assumed that intervals of the all the 22 Sruti-s are equal. As pointed out earlier Sruti is a rough measure like `steps'. If mathematical relative frequencies are used then it is quite likely that the intervals of the 22 pitch positions may not be equal. In fact once we use mathematical ratios then it is not proper to call the pitch positions as Sruti-s.

I have tried to present a general view of the ancient system and to indicate the role of Sruti in it. For more detailed information on the role of Sruti in the ancient system see

(i)The book 'Dattilam: A treatise on the Sacred music of ancient India' by Mukund Lath

(ii) The article 'TheConcept of Sruti as Related to Svara - A Textual and Critical Study' by R K Shringy in the Journal of the Music Academy, Madras, 1973.

I have then attempted to show the extension of its usage in the musical system of the middle ages. And lastly, I have tried to point out the misinterpretation and misuse of the term Sruti in the modern period and to show how the theories arising out of these notions go against the spirit of our music.

References :

NS AB  nATyaSAstra of Bharatamuni with Commentary of abhinavabhAratI of Abhinavagupta pub. in the Gaekwad's Oriental Series by Oriental Institute, Baroda.vol.IV- ed. by M.Ramakrishna Kavi and J.S.Pade, 1964.

Datt. Dattilam of Dattilamuni, ed. by K.Sambasiva Sastri, pub. in the Trivandrum Sanskrit Series, Trivandrum, 1930.

BrD   brhaddESI of matangamuni, K.Sambasiva Sastri, pub. in the anantaSayanam samskrta Series , no.94,Trivandrum, 1928.

SR   sangItaratnAkara of SArngadEva with the commentaries kalAnidhi of kallinAtha and sudhAkara of simhabhUpAla, vol.I, paNDita S.Subrahmanya Sastri and revised by S.Sarada, pub. in the Adyar Library Series by Adyar Library, Madras, 1992.