Rahaim, Matthew. 2008. "Gesture and Melody in Indian Vocal Music." Gesture 8(3): 325–347

The gestures that accompany improvisation in Indian vocal music, like the gestures that accompany speech, are closely co-ordinated with vocalization. Though linked to what is being sung, these movements are not determined by vocal action; nor are they taught explicitly, deliberately rehearsed, or linked to specific meanings. Students tend to gesture recognizably like their teachers, producing lineage-based gesture dialects, but the gestural repertoire of every vocalist is nonetheless idiosyncratic. This paper aims to trace a brief history of song gesture in India, and to show some of the links between gesture and vocalization. It also adapts Katharine Young’s theory of the “family body” to the transmission of gesture dialects through teaching lineages. Gesture and sound are taken to be parallel channels for the expression of melody.

The body and the voice work together in music as well as in speech. Flamenco singers, for example, heighten the rhetorical impact of their performance with dramatic movements of the hands, arms, and eyes. Singers of Peking Opera as-sume stylized gestural dispositions according to specific role types. Systems of chironomy are an integral part of Yemenite Torah recitation (Katsman, 2007) and are found in other liturgical chant traditions as well.

Indian classical vocalists use gesture space together with sonic space as a medium for the depiction of improvised 1 melody. Vocal forms such as dhrupad, khyal, and thumri in the North, and ragam-tanam-pallavi in the South feature highly cultivated vocal improvisation practices that give rise to ever-novel melodic utterances. These improvisation techniques are guided by flexible frameworks that operate like grammatical structures. In this sense, the melodic improvisation that occurs in the course of performance is akin to extemporaneous speech.The gestures that accompany this vocal improvisation (hereafter song gesture), like those of improvised speech (hereafter speech gesture), are closely co-ordinated with vocalization. Though linked to what is being sung, these movements are not fixed or strictly determined by vocal action. Unlike the formalized systems of hand shapes, or mudras, of classical Indian dance (Coomaraswamy, 1936), the gestures that accompany singing are not taught explicitly, deliberately rehearsed, or linked to specific meanings. Nor, in most cases, do the gestures relate to the meanings of song texts. Students tend to gesture recognizably like their teachers, producing lineage-based gesture dialects, but the gestural repertoire of every vocalist is nonetheless different.Of course, not every movement of the body that occurs during a performance is a depiction of melody. There are specialized, conventional signs to mark the progression of metric cycles, to guide accompanists, and to indicate tempo. AsClayton (2007) has shown, gesture serves as an important mediator between performers and audience members. Audience members — particularly in intimate musical gatherings — gesture in conventionalized ways to encourage the performers, and performers greet eminent musicians in the audience. Musicians also occasionally direct certain segments of the performance at particularly appreciative connoisseurs, as though music were an object offered by the performer. Singers also, of course, adjust their sitting position, drink water, and carry out other physical actions that are unrelated to sound production in the course of performance.The gestures that I will focus on here are those most closely linked to improvised melodic action: tracing curves in space, stretching virtual materials, sculpting virtual objects. These gestures serve as three-dimensional, kinetic representations of melody. I will first give some examples of how these gestures are linked to phrasing and melodic shape, and how topographies of gestural action may beisomorphic with raga spaces. Then I will explore some ways that various gripping gestures may be used to manipulate virtual melodic objects. Finally, I will suggest some ways of understanding how gesture dialects, and their attendant disciplines of melodic improvisation, may be transmitted through teaching lineages.The research for this paper is based on approximately a hundred hours of video recordings collected in 2006–2008 in India. These recordings consist of performances, lessons, and formal interviews with twenty vocalists, with a primary focus on eight vocalists in two teaching lineages. It is also based on my own amateur study of North Indian classical vocal performance over the last ten years,primarily with Vikas Kashalkar in Pune, India. For this reason, nearly all of the material reviewed here is focused on North Indian music, and the genre khyal in particular.

Song gesture in India

Little can be said with much certainty about the history of song gesture in India before film. Indian music-theoretic literature has concerned itself primarily with the grammars of note combinations that provide the frameworks of various ragas. Just as gesture has until recently been regarded as something outside of language (Kendon, 2004, p. 356; Sweetser, 2004, p. 197), Indian music theory, in its focus on note combinations, has tended to regard gesture as something outside of music. Music criticism, on the other hand, has been primarily concerned with the performance styles of individual musicians in relation to musical lineages. This body of critical literature, developed over the course of the last hundred and twenty years or so, in newspapers and in popular books, as an adjunct to the growing public sphere for the performance of classical music. The extent to which music critics contributed to the creation of a new climate for the reception of song gesture — instead of merely carrying on an older tradition — is unclear. Both music-theoretic and critical music literatures, however, had a place in a national musical reform project that adapted diverse court and temple performance traditions to the standards of a national public stage. 2 In this literature, body movement is generally regarded as an incidental accessory to music, or at worst, a distraction from music. Generally, when a music critic mentions gesture, it is to admonish singers against making “violent and spasmodic physical movements,” (Bailur, 1955, p. 40) or “futile gestures of the hand” (Deshpande, 1976). Some music writers even directly prescribe the suppression of gesture: “the activities wasted away by bodily actions and movements, should be concentrated on the voice…” (Pingle, 1962 [1894], p. 103–104).When a music critic does praise a singer for their gestural habits, it is usually in praise of a singer’s restraint. In the following passage, the relative lack of gesture in the performance of Amir Khan, a prominent Hindustani vocalist of the 20th century, is discursively linked to spirituality. This is achieved by associating his stillness with both yogic practice and Sharangadeva, a thirteenth century music scholar who posited spiritual liberation as music’s primary goal. [Amir Khan’s] dignified bearing and his upright, yogi-like posture, when performing on the stage, was perfectly in tune with the serene grandeur of his music. Acrobatics and contortions of any kind were conspicuous by their absence in his case — and, that too, at a time when “making faces” (as also a bewildering variety of physical gestures) had already become a specialized art by itself. Here, indeed, was an executant who appeared to live up to the dictum of Sharngadeva, that distortion of the face and indulgence in physical gestures was a gross disqualification (Nadkarni, 1999, p. 178).

The opposition of song gesture to spirituality appears implicitly in Indian cinema, as well. An example of this can be found in the two opening scenes of Sant Tukaram (1936), perhaps the most well-known and most influential of Indian saint-biography films. Here, gesture operates alongside elocution, dress, and posture as a sign of the main opposition represented in the film: between hierarchical, elite Brahminical religion and the populist, devotional spirituality of the Marathi singer-saint Tukaram. These extremes are enacted in the performance styles of Tukaram and the vain, hypocritical priest Salomalo. Salomalo gestures wildly as he peppers his devotional songs with florid, distracting melodic runs — at one point he nearly hits someone in the face with his flailing arms. Tukaram himself, the humble, impoverished saint who is the movie’s protagonist, hardly moves at all as he sings his simple, unadorned songs — transported, as it were, beyond his body. In many Indian films that feature classical music, musicians associated with Muslim courts 3 are likewise depicted as gesturing unnecessarily as a token of meaningless, showy virtuosity in contrast to devotional (and usually Hindu) piety (Booth, 2005).Although there is no clear link between religious identity and gestural dis-position as performed on the modern classical music stage, gender does seem to matter. Female singers, on the whole, gesture less dramatically in concerts, and in a smaller gesture space, than male singers. The reasons for this are unclear. But it may be partially explained by some of the processes involved in the music reform movements, mentioned above, that began in the latter half of the 19th century. In the years leading up to independence, the image of the prototypical female singer shifted from a courtesan who sang and danced to a respectable middle-class woman who sang but, crucially, did not dance (Quinn, 1982, pp. 91–92). The new breed of female singer sat on the ground, relatively still, and sang with a tanpura in her lap like a male singer. 4 Overt gestures that pantomimed the verbal content of songs were eliminated from these performances (Pradhan, 2004, pp. 341–343). There may also have been pressure on these female singers to make these gestures smaller as a means of downplaying the role of the female body in performance. The separation of music and dance as a means of distinguishing respectable female singers from courtesans also occurred in the South during roughly the same time. Anthropologist Amanda Weidman (2007) describes this shift:

for female musicians in particular, a convention of music performance developed in which the body was effaced; too much physical movement or ‘show’ on the stage was seen not only as extraneous to the music but as unseemly … the ideal became a kind of performance of nonperformance: nothing visible was supposed to happen on the music stage (pp. 130–131).

Male and female singers alike tend to downplay the role of their own song gestures. Nonetheless, no singer does without song gesture. Even when a singer’s hands are occupied with an instrument, they manage to make melodically relevant movement with their head and eyes.As this paper aims to demonstrate, this movement is not merely a persistent bad habit. Song gesture is coordinated with vocalization, and appears to be deeply connected with the conception and performance of melody. 5 This relationship be-tween vocal phrasing and gesture phrasing seems to be akin to the relationship between speech phrasing and gesture phrasing that has been described in the work of Kendon (2004) and McNeill (1992, 2005) and which has led to the view that gesture and speech are co-produced and coordinate because they are two aspects of a single linguistic process. Song gesture appears likewise to be an integrated with vocalization, suggesting that vocalization and gesture are two integral aspects of a single musical process.

Khyal in particular

This article will focus on khyal, a genre of North Indian (hereafter “Hindustani”) vocal music. Khyal, as used in North Indian vernaculars, usually means “thought” or “imagination.” Accordingly, the art of Khyal performance is primarily the art of large-scale, imaginative melodic and rhythmic elaboration of song material ac-cording to given melodic and metrical frameworks. The compositions that serve as the seeds for this improvisatory elaboration are quite short, usually consisting of six- or eight-line poems set to simple tunes. The phonetic vehicles for vocal improvisation in khyal largely consist of vocables (“aa” “re” “eh,” etc.) and short phrases from the composition at hand. Significantly for studies of gesture in khyal, melodic gesture is generally found during sections of improvisation rather than during sections of pre-fixed melodic material.Improvisation in khyal unfolds within metrical and melodic frameworks (talaand raga, respectively), neither of which lend themselves easily to verbal defini-tion. Talas are metrical cycles marked by strokes on a tabla or other drum, and which provide a measured, periodic place of return for melodic elaboration. Raga-s are complex melodic entities that are less specific than individual songs, but more specific than scales. While allowing for infinite possibilities for improvisation, they indicate sets of notes, guidelines for their arrangement in phrases, and, most importantly, characteristic ways of moving between them. 6 These melodic features are often associated with specific emotional colorings, and in some cases, to times of day or seasons. As we shall see later, a raga is also often spoken of metaphorically as a space in which melodic action takes place. Khyal improvisation, then — like extemporaneous speech — takes place within structures even as it requires the creation of novel material in the moment of performance.For readers unfamiliar with khyal, I have given below an extremely rough oultline of the form of a typical performance, to aid in understanding the analyses that follow. 7

Table 1. Outline of khyal performance

Phrasing

As with speech gestures, song gestures consist of excursions from a resting position, through one or more significant movements, and back to rest. Furthermore, as with speech gesture, hierarchies of kinesic units, marked by various degrees of rest, correspond roughly to hierarchies of vocal phrasing. For example, a brief re-traction of the hands in front of the chest typically marks a less significant phrase marker than allowing the hands to come completely to rest in the lap for several seconds.Singers also must breathe in the course of singing. The necessity of stopping phonation while inhaling every fifteen seconds or so constitutes an important phrase marker in solo vocal music. Hindustani vocalists time their breaths to avoid “breaking” a musical phrase or a sung word with a breath in the middle. But there are also perceptible macrophrases in music longer than a single breath — sometimes on the scale of minutes — as themes are developed and refer back to themselves. Singers often will continue the course of a gesture while taking a breath, linking two vocal phrases into a longer melodic phrase. Hierarchies of me-lodic phrasing that are longer than the breath are thus perceptible in gestures.There are times when an initial motion from rest to gestural action reveals that certain features of a coming gestural action was conceived before vocalization. For example, a handshape may already be fixed before a sound is uttered, or the arms may “wind up” for a dramatic downward jerk. Idiomatic melodic phrases, some-times several seconds long, often serve as building blocks as well as individual notes. Gesture usually operates at the level of these phrases, rather than at the level of individual notes, and offers a way of apprehending melody as continuous shapes rather than as discrete points.

Melodic shapes

Hindustani musicians have at least two complementary, but quite independent, models of melody. The dominant model sees melody as a sequence of discrete notes. This model is represented in a precise, explicit, centuries-old analytic sys-tem, similar to {do re mi fa sol la ti} as used in the West. This system likewise names the seven notes (svara-s) in turn: {sa re ga ma pa dha ni} and analyzes melodic motion in terms of the discrete stopping points marked by these pitches. This model has the virtue of being both precise and easily notated, and is used widely by both musicians and theorists to describe melody.The second model of melody used by Indian musicians envisions melody as shapes rather than sequences of discrete notes. 8 Many Hindustani vocalists speak explicitly about these shapes. Vocalist Veena Sahasrabuddhe says that while improvising, she is not generally thinking in svaras (individual notes) but in shapes: “I make all kinds of designs, and the designs are in my mind. But … I’m not putting any efforts … instead of svaras [i.e. notes], I’m just thinking about the curves and the lines.” 9 Arun Kashalkar, another Hindustani vocalist, emphasizes that these shapes are abstract, in contrast to the representational gestures of dancers, but also unfold over time, in contrast to a drawing: “They are amoeba-like, the shapes … they would be abstract forms, like abstract drawings … but the entire thing has some shape that is musically explained, you can’t explain it like a drawing.” 10 Below is a simple example of how gesture delineates moving melodic shapes, in which the pitches of four notes in a phrase (sa-ma-ga-pa) correspond to the relative position of the left hand in the y-axis. The singer is Mukul Kulkarni. The phrase is sung on the pure vowel “e” as in “Mexico”, and is approximately one second long.

Figure 1. Sa-ma-ga-pa phrase: gesture and pitch vs. time

Pitch and hand position are clearly linked in this case, but the relationship between a sequence of sounded notes and the shape of a phrase is not rigidly fixed. Sometimes, the left side of the gesture space is associated with lower notes, while higher pitched notes are located to the right. For some singers this is reversed: lower notes correspond with a hand position to the right, and higher notes cor-respond with the left hand side. Pitch may be mapped onto radial dimensions as well, or onto the relative position of two hands. Sometimes singers conjure an invisible cord that they stretch between their hands, and pitch corresponds to its virtual tension. Any of these particular correspondences between space and pitch are temporary. Singers will use one for several phrases, and switch to another when expedient. Indeed, many singers employ a melodic-rhetorical technique in which they repeat a sequence of notes several times in a row with different gestural content each time. This is parallel to cases of verbal repetition in which gesture is deployed differently each time to create a novel utterance, and highlights that gestural-melodic coherence is achieved, rather than merely given (Kendon, 2004, pp. 151–156).Furthermore, pitches do not have absolute “homes” along spatial axes the way that every key has a fixed place on a piano keyboard. That is, “sa” is not always in front of the belly, “ma” is not always three inches to the right, etc. The movement from left to right is context-dependent — each left-right gesture is calibrated to a single phrase. Macrophrases tend orient themselves around the “goal” note of each phrase, which tends to be located near the center of the gesture space.

The four notes of this phrase are not merely independent data points with separate pitch values. Nor is the gestural articulation of this phrase merely a sequence of snapshots. (Recall Arun Kashalkar’s caveat above about interpreting a melodic shape as a static form: it is “musically explained — you cannot explain it like a drawing”.) The melodic fragment is performed both gesturally and vocally as a single curvaceous melodic action. This phrase traces an ascending, serpentine curve — low to high to low to high. The phrase has both periodic and linear components — it changes direction periodically, but maintains a net motion in a single direction. The technical term for this kind of motion is vakra (lit. crooked) a term also used for the meandering progress of a river and for planetary retrogression.Mukul’s hand is not only moving up and down to map pitch levels onto the y-axis. The hand is rotating as well. When the melody is descending (i.e., when the time derivative of pitch is less than zero), the hand is inflected downward. When the melody is ascending (i.e., when the time derivative of pitch is greater than zero), the hand is inflected upward. In the course of the gesture, the hand rolls upward and forward, describing an ellipse in the sagittal plane, about the wrist joint. This is the periodic component of the vakra ascent. Meanwhile, the arm articulates the linear component, gradually moving the hand upward and away from the trunk. The resultant hand motion in the sagittal plane, then, includes a loop.This low-high-low-high pattern is both the smallest vakra motion and the smallest possible looping motion. Often, it is more extensive. An ascending phrase may include eight or more oscillations on its way up so that the hand describes an extended looping form.

These kinds of curvaceous melodic patterns are often described in words by singers. Arun Kashalkar, for example, describes certain kinds of melodic motion as “twists”, and Sheila Dhar describes how a master vocalist melodically “traces delicate three-dimensional arcs, and draws from silence deeply searching spirals” (Dhar, 2001, p. 22). But it is important to note that these loops are not given or obvious merely from looking at sequences of pitches. The shape-based descriptions of melody above refer to melody in the course of performance, not graphs of pitch mapped against time. A mechanical mapping of the pitch content of the vakra phrase sung above in merely sonic terms would not produce a loop or a spiral, but an undulating line as in Figure 1. Indeed, a loop on such a machine-drawn pitch vs. time graph would be absurd. It would imply progress backwards in time, which requires attention, memory, and imagination — all three of which are necessary to apprehend melody as simultaneously sequence and shape, sound and gesture.

Gesture and raga

A raga is one of the foundational melodic constructs of Indian music, but it is eas-ier to hear than to describe. Musicians and non-musicians alike can typically perceive the difference between contrasting ragas as a kind of gestalt, often described in terms of color or emotional impact. Music theorists, on the other hand, have tended to describe the differences between ragas in terms of grammar. Among other things, these grammars prescribe which notes are to be used in ascent and which are to be used in descent, and which key phrases, omissions, and emphases will generate acceptable melodies characteristic of that raga.But grammar is only one of the tools that musicians have at hand for speak-ing about raga. Most significant for the current discussion is that musicians often speak about a raga as if it were a space: as if one were “in” a certain raga while performing. Melodic action, then, is motion within the space. For example, one might say that in Raga Bihag, one ascends from pa to ni by leaping over dha; in descent, dha can be touched in passing, but should not be rested upon. These patterns of melodic action are described as the chalan (“movement”) of a raga. The establishment of an inside of the raga through motion implies an outside, and a common way of describing an inappropriate melodic act is that it has gone outside of the raga. This may refer to a grammatical violation, and be analyzed in grammatical terms, but more often it is subtler than this. Even the strictest melodic grammarians agree that great musicians can sing non-grammatical phrases whose motion remains inside the raga. Indeed, this is often cited as the sign of musical greatness. Singers from musical families who learned by ear but never had systematic theoretical training often can move easily within ragas without explicit knowledge of the grammar of the raga or even the exact notes they are singing.
11 Musicians often speak of various places (jagah) within ragas.

12 As Sheila Dhar puts it: “each jagahor place in the scale of a raga is not a point but a musical area that must be explored anew each time and brought to life in the living moment” (Dhar, 2001, p. 22).Through performance of a raga, a one-dimensional continuum of pitch is developed into a flexible but stable landscape. The space around the body is likewise transformed into a landscape with consistent features. These landscapes are not given or fixed, but simultaneously created and explored through consistent melod-ic action, as paths are created in a forest through repeated use. Singers maneuver in this space while improvising, moving through particular regions via particular melodic paths articulated by both sound and gesture.

Many Hindustani ragas are nuanced variants of other ragas. 13 The performances of these ragas often seem to include subspaces. 14 Rag Ramkali, for ex-ample, is a variant of Rag Bhairav. Nearly all of the melodic action of Ramkali resembles Bhairav. 15 Ramkali is distinguished from Bhairav by an occasional, distinctive introduction of use of two notes “outside” of Bhairav (tivra ma and komal ni) to elaborate on pa, the fifth note of the scale. For brevity’s sake, I will call uses of these notes — motion within this particular jagah — “Ramkali shifts”, and the space of Bhairav-like melodic action in Ramkali “quasi-Bhairav”.In the particular performance 16 of Ramkali that I analyze here, the melodic action of Ramkali shifts was dramatically and consistently contrasted with quasi-

Bhairav action by the spatial and movement features of the hand gestures accompanying them. These differences are summarized below.17 While moving within the Ramkali shift subspace, the hands curl around a small empty space about the size of an apple without touching. This seems to be parallel to the sonic action of the Ramkali shift, in which the notes center around, but never rest on, the stable fifth scale degree. Unlike the melodic action in quasibhairav, the Ramkali shifts in this performance yield no net change in pitch. Correspondingly, the action of the hands results in no change in physical space. Some gestural features are found across several ragas with melodic similarities. For example, there is a gentle swinging, oscillating manner in which the lowered third scale degree of several ragas18 is sung. Depending on the singer and the context, this swinging may be accompanied by various gestures: a churning motion with the fist, a waving motion of the open palm, or a slow tracing of a line back and forth in the air. Although these gestures are different from one another, they share an underlying periodicity that is coordinated with periodic sonic action. Gina Fatone’s term for this is “meta-gesture” — a recurring formal entity that is articulated via gesture variously in different situations.19 The link between melodic subspace and gestural subspace is quite consistent in the context of a single performance (of 34 Ramkali shifts in the above performance, for example, 30 are accompanied by the curling gesture). But just as there are various possible isomorphisms of pitch and space, there are various ways that the structure of a raga can be mapped onto space — some of which may only be maintained for a section of a single performance. Recurring gestural patterns within this landscape might be compared to what McNeill has called “catchments” (McNeill, 2005, p. 116), which are recurring features of gesture performance (such as hand shape or movement pattern) that link discourse features together within the context of a single performance.

This close linkage between raga space and physical space seems to be mostly limited to the slow introductory segments of performance. As a performance progresses and phrases are strung together more densely, rhythmic action becomes more important than the fine details of raga exposition. This is reflected in the course of lessons, as well. During alap — slow, unmetered melodic exploration — a teacher will often stop a student who misses a crucial note, but during later, faster elaboration, the overall flow of a phrase is more important than the exact sequence of notes. In these later segments of performance, gestures often become more vigorous, and the hands twist, flap, and shake in ways that seem to foreground rhythmic flow instead of precise melodic shapes.

Hand shape and grip

So far, we have focused on pitch as the primary sonic feature of melody and have discussed how pitch and gesture are related. However, gesture seems to be linked to several other sonic dimensions as well. Perhaps the widest gestural-sonic link held in common among Hindustani vocalists is connected to timbre. Across Hindustani vocalists, closed vowels (such as the “u” in “shut” or the “ee” in “feed”) tend to be associated with a fist or pursed fingers,20 while open vowels (particularly the “a” in father) are associated with an open or loose palm.21 Singers also use various gripping hand shapes as if to form and manipulate virtual objects. In the course of improvisation, singers will sometimes appear to hold an established, repeated melodic pattern, as though it were a small object, in one hand. Upon suddenly changing the melodic pattern, they might abruptly invert the hand holding the object, as if to flip it over in space, exposing an aspect of the virtual object that was previously hidden by position of the grip. Some performers also use their fist to grab and stretch a virtual elastic material in front of the body. Performers may also use a hand shape much like the grappolo hand shape that
Kendon (2004, p. 229) has described as a form of precision grip. This hand shape is used in the form of a gripping action at moments when the precise intonation of notes is crucial, particularly when holding a single note for several seconds. Figure 5 shows Vikas Kashalkar holding his hand in a grappolo and slowly pulling it to his left while sustaining a note. As Kashalkar himself explains it, he uses this hand shape so that the note “should not go out of my control, out of

my grip”. After gripping the note, he sustains it for several seconds while moving his hand horizontally. He describes this action as depicting a “hairline”. This thin horizontal line is isomorphic with a graph of his pitch against time — the pitch remains constant. These three functions associated with the grappolo seem quite different on the surface: closing (as opposed to opening), gripping (as opposed to
releasing), and drawing a thin horizontal line (as opposed to movement up and down). But all three are simultaneously articulated by the handshape, the motion, and the vocalization. The closed vowel (“ee”) is linked to the firm grip of the note: most singers agree that pitch is easier to control on such vowels. Likewise, the line is “hairline” thin both by virtue of the bunched handshape and the articulation of the vowel: “ee” requires the tongue to rise high in the mouth, narrowing the vocal passage. Melodic functions in several dimensions are brought together in this single vocal/gestural action.

Transmission: The paramparic body

Singers tend to gesture like their teachers, and there are recognizable resemblances within teaching lineages. Although no two singers have exactly the same manner of gesticulating, these gestural family resemblances are quite striking in many lineages of khyal vocalists. The development of gestural styles is similar to the development of recognizable regional speech gesture dialects (Kendon, 2004, p. 349). The difference is that these gestural patterns are specific to musical lineages, regardless of the language or region of the singer, are lived specifically in the moment of musicking, and are learned through hours of sitting in front of their teacher.

The transmission of musical knowledge through gesture, however, is nearly always implicit. (Remember that my comments are specifically focusing on khyal pedagogy, though some may apply to other genres of Indian music.) Although most teachers give copious explicit advice about vocalization, they almost never tell students how to move. 22 The student instead learns by repeating what the teacher does. Teachers move their hands vividly in lessons to indicate phrasing and the shapes of melodies. Teachers sometimes even “guide” a student’s melody by gesturing along with them as they sing. Sometimes in the course of a lesson, a student will be unable to reproduce a phrase sung by the teacher; the teacher then will repeat the sonic content of the phrase for the student using different gestures to model an aspect of melody that the student doesn’t yet see. The teacher, may, for example, accentuate similarities and differences between different subphrases by placing them in contrasting locations in space. The teacher repeatedly provides the same sonic content, but uses space to suggest various alternate ways of conceiving the music. These are among the most intense moments in the course of a lesson because they require a student to unlearn a habitual way of conceiving of a melodic fragment, and to reconceive it according to the teacher’s demonstration. Even if a student has the sequence of notes right, the particular way of shaping these notes into a phrase may be mistaken and is often corrected through gestural demonstration. These moments are perhaps reminiscent of moments of gestural “mismatch” in math students that immediately precede a cognitive breakthrough (Goldin-Meadow, 2003).

The transmission of musical knowledge via gesture is made possible by an intense, long-term relationship23 between teacher and student. As written notation is seldom used in lessons, teacher and student spend hundreds of hours sitting face to face over the course of many years. This extensive time spent together fosters a close bond, often involving a great deal of time spent together outside of formal lessons. When the student is not actually a blood relative of the student, the relationship is often modeled on the relationship between a guru and his disciple: guru-shishya parampara.24 In the strictest form of this relationship, the disciple and the guru are ritually bound to one another as virtual parent and child. When musicians are introduced at concerts, among the first pieces of information offered is the name of their teacher. Even after the period of formal, regular tutelage has ended, musicians identify with their teacher, often keeping pictures or other tokens of them in their music rooms where they in turn teach their own students.

In addition to a particular teacher,25 nearly every singer identifies with other musical “ancestors” — even, in some cases, legendary musicians from hundreds of years ago. These links between great foundational musicians and present-day performance are sustained by teaching lineages, some of which in the last few centuries or so have crystallized into widely-recognized schools of musical thought. These lineages are important identifying features of a singer’s style, each of which carries certain aesthetic ideals for how a performance ought to be developed. For a singer, a musical lineage serves as a disciplined way of musicking. With so much possibility for improvisation within a khyal performance, musical lineages provide
direction by focusing their musical scope. Young singers are generally expected to commit to such a discipline for many years before developing an individual style. Vocalist Arun Kashalkar26 describes this process by analogy with handwriting: everybody’s handwriting is different, but in learning to write, a child must commit to a single method of writing each letter, and do it a thousand times. Only after a student has mastered a certain basic discipline (of musicking or of forming letters) can the student’s own style emerge. After this point, students may freely borrow features from other lineages, as well. This process is neither a matter of pure choice nor unmediated inheritance.

This dialectic between inheritance and choice is captured well by what Katharine Young calls the “family body”:
… bodies are passed down in families, not as assemblages of biological traits enjoined on the bodies of children by parents but as intentional fabrications devised by children out of the bodies of parents … Within families, memory is passed down, not only as oral lore or material artifacts but also as something that is neither mentifact nor artifact: corporeal dispositions … Family bodies, like family stories, provide their heirs positions, situated perspectives, on parents’ ways of being in the world, out of which children can devise their own ‘presentations of self ’ … (Young, 2002, pp. 25–26).

To adapt this term to the specific case of musical transmission, I will use the term paramparic body, borrowing the Sanskrit word parampara (as in guru-shishya parampara) which usually used in North Indian vernaculars to refer to a traditional lineage. There is great gestural variety both among teaching lineages and among the paramparic bodies of individual singers. I do not, however, see these only as “styles” or arbitrary patterns of movement. As David McNeill says, “Gesture and speech are not only ‘messages’ or communications, but are a way of cognitively existing, of cognitively being, at the moment of speaking” (McNeill, 2005, p. 99). Each paramparic body is a way of musically being,27 formed out of corporeal dispositions passed down tacitly through teaching lineages.

For example, certain musical lineages put great value on singing with an open voice (khula avaz). Open voice is linked to two features of bodily disposition: an open mouth and, as we saw earlier, a hand shape in which the hand is held open, with the fingers extended and the palm exposed. Among the most dedicated adherents to this disposition is Pavan Naik, a young singer from Maharashtra. For Pavan (as for other singers in his lineage), an open voice is connected to a certain open musical disposition as well: musical elaboration should be straightforward, simple, and clear. The voice should not be hidden behind the tongue or the jaw; likewise there is nothing concealed in the hand. On one occasion, I was showing Pavan a video of his teacher’s teacher, Gajananrao Joshi, performing. Gajananrao Joshi had died before Pavan had a chance to meet him, and this was the first time he’d seen a video of him. Very soon after the video began, Pavan remarked, “He looks like me here” (pointing to his open jaw), “and here” (pointing to his flat palm). A few minutes later, Pavan looked at me in amazement and said, “I am the next birth of Gajananrao Joshi”.28

It would miss Pavan’s point to interpret this as a claim about the literal transmigration of a living soul into a separate material body.29 The sense in which Gajananrao Joshi has been reborn in Pavan bridges body and soul: the incarnation occurs continuously in the moving body, the volitional body, the musicking body. Pavan’s recognition of his own gestural disposition in his grand-guru’s musicking was akin to that of a person recognizing his smile in a photograph of a distant ancestor. But to describe this resemblance merely in terms of passive inheritance obscures the role of volition in this process. Pavan has not simply absorbed this body through his teacher30 — he has also chosen it. Indeed, he has consciously taken it further than his own teacher does. Even in the early stages of performance, he avoids closed vowels and their attendant gripping gestures. He also has deliberately chosen not to adopt other features of his teacher’s paramparic body. For example, inspired by his drama training, he has chosen an unusually upright orientation of his torso while performing, and makes a point to make eye contact with the audience from the beginning. Pavan’s gestural disposition, then, is neither merely inherited nor merely chosen. Through discipline and deliberate construction, through vocalization and gesture, he incarnates a lineage of embodied musicking through his paramparic body.

Conclusion

I have suggested here several ways of understanding musical processes in terms of space and motion. These processes all have parallel sonic and gestural expressions: phrasing is delineated both by sound and by excursionary movements of the hands. Melody, it seems, is depicted both as sung notes and as moving shapes, and raga is enacted both through note grammars and spatial topographies. I have also described ways in which musicians act as if they grip, hold, and move melody while singing. Finally, I have tried to account for the transmission of gestural dispositions through the construction of paramparic bodies. Spatializations of melody, motion in raga space, and the inhabitation of paramparic bodies are all both locally consistent and flexible, both coherent and constructed. Just as a singer may, over the course of several phrases, substitute tension for height as the analog of melody, he may, over the course of several performances, alter the spatial features of a given raga, and, over the course of years, transform his paramparic body.

The most obvious interpretation of the linkages between gesture and vocalization is that space and sound are metaphorically related. This is akin to what McNeill calls “gestural metaphoricity”, as when for example an English speaker “holds” an idea in his outstretched hand (McNeill, 2005, p. 45.) Clearly, gesture and vocalization share an underlying logic in many cases, as in Figure 1. But does this shared logic imply a unidirectional mapping? In other words, does physical space serve as a grounded source domain for the depiction of an abstract sonic target domain? In a few special cases, singers are thinking primarily in notes, and seem to use gesture to metaphorically represent those notes. For example, a teacher may deliberately “sketch” notes in the air along a clearly delineated linear axis for the sake of a student who repeatedly fails to reproduce the pitch sequence of a phrase. There are even rare moments in performance when singers seem to be using such representations to aid themselves in conceptualizing the relationships between a particularly complex series of pitches.

Apart from these special cases, however, there is no reason to assume in general that the gestures these vocalists use are directed metaphorical representations of vocalization. Many common features of song gesture performance are not accounted for by this model:

1. Gesture phrases often continue smoothly through silences during breaths (see “Phrasing”, above)
2. Singers report that they often think in shapes rather than notes when singing. (see “Melodic shapes”, above)
3. Singers can move freely (both sonically and gesturally) in raga spaces without being able to name the exact sequences of notes that they are singing (see “Raga and gesture”, above)
4. Preparation phases of coming gestures often reveal that a hand shape or kinesic arrangement is fixed before the sonic content begins (see “Phrasing” and “Hand shape”, above — also cf. Kendon, 2004, p.125)
5. Teachers sometimes guide a student by gesturing while a student is singing, but without singing themselves (see “Transmission”, above.)

None of these features are accounted for by a model that takes gesture to be a unidirectional metaphoric representation of sound. Instead, in most cases of vocal performance by trained singers, vocalization and gesture seem to act as complementary, parallel channels for the expression of melody. The kind of inaudible,
always-yet-unarticulated melody that underlies song gesture seems to be in the same family as the entities that David McNeill describes as growth points (2005, p. 105). Further, the construction and maneuvering of virtual objects and spaces in Hindustani song gesture suggests a parallel with Kendon’s view (2004, p. 360)
that language is linked to the fabrication and manipulation of objects. It is by now fairly common to speak about the “language of music” (in reference to, say, the grammatical structures underlying note combinations) or the “music of language” (in reference to prosody). But the similarities between speech gesture and song gesture suggest that the processes of both musical and spoken improvisation may have more in common than is revealed in the structural and prosodic features of notes and words.

Acknowledgments

The research upon which this paper is based has been supported by the American Institute for Indian Studies, the Qayum Family Foundation, the UC Berkeley Center for South Asian Studies, the UC Berkeley department of Music, and the Berkeley University Fellowship. I am grateful to the UC Berkeley Gesture Studies Group, Bonnie Wade, Ben Brinner, Eve Sweetser, Vikas Kashalkar, Mukul Kulkarni, Pavan Naik, David Wessel, Marié Abe, Martin Clayton, Adam Kendon, Irfan Zuberi, and Jon Barlow for their ongoing inspiration, support, and feedback.

Notes
1. “Improvisation” can mean many things. Here I use it to loosely refer to practices in which composition and performance occur together in a single moment. See below for more detail about what it means in khyal, the particular genre at issue in this paper.
2. See Bakhle, 2005; Quinn, 1982; Weidman, 2007; Subrahmanian, 2006, and Neuman, 2004, for detailed investigations of the changes wrought by this music reform movement. Another important connection is provided by Kendon (2005, p. 357), who suggests ways that the rise of the printed word as the paradigm for language may have contributed to the sense of gesture as something outside of language. The printed word was integral to Indian musical nationalism as well, as music notation was a crucial feature of nearly every attempt at music reform.
3. Although gesture tends to be mapped onto the prevailing Muslim-Hindu binary as “Muslim”, not every Muslim is always depicted as a voracious gesturer. Amir Khan, for example, the “yogilike” singer from above, was a Muslim.
4. This remains the default singing position for both males and females.
5. In the last five years, there have been several fine studies of various aspects of movement in music Indian classical music. Nikki Moran’s PhD (2007) thesis deals with physical action in Hindustani music. Martin Clayton (2007) has studied various aspects of motion in Hindustani music performance, including rhythmic entrainment and performer interaction. Laura Leante has studied the gestures of Hindustani vocalists as anaphonic signification of sonic qualities (Fatone et al., forthcoming).
6. For those wishing to learn more about raga, Nad by Sandeep Bagchee is a good introduction, as is Rajan Parrikar’s website parrikar.org.
7. There is great variety in the details of how a khyal performance is structured, and techniques of raga development are linked closely to musical lineages (see “Transmission”, below). There are many fine samples of khyal recordings available online: Sarangi.info, musicindiaonline.com, and Parrikar.org have wide selections. Note that many of these recordings are truncated for the sake of recordings. Videos of segments of some performances are available on YouTube.
8. This second model corresponds roughly with what Charles Seeger called the “stream” model of melody; the first corresponds roughly with what he called the “chain” conception (Seeger, 1958). Most scholars of Indian music have focused exclusively on the first model — but see Roy (1934) for a striking exception in which an attempt is made to rigorously analyze melody in terms of melodic shapes.
9. Interview with Martin Clayton in Bombay, April 12, 2003. Many thanks to Martin for the kind permission to reproduce this here.
10. Interview in Bombay, November 25, 2007.
11. Of course, such a singer’s implicit knowledge could be considered grammatical. After all, the fact of linguistic competence in one’s mother tongue without explicit grammar study doesn’t deny the existence of implicit grammatical structures. But as we will see later, the fact that singers don’t always think in terms of notes in performance has implications for the processes that underlie musicking.
12. For an extensive discussion of jagah in raga performance, see Neuman, 2004, Chapter 4.
13. Many Hindustani (North Indian) ragas are mixed or “joint” (jod) ragas, or distinguished from close relatives by virtue of subtle differences. This is less frequently the case in Carnatic (South Indian) music.
14. The terms “space” and “subspace” may call to mind, for some readers, the algebra of vector spaces. Raga spaces are like vector spaces in two key respects: they are defined largely by relations between its elements rather than by pre-given boundaries, and they are closed over certain melodic operations as a vector space is closed over certain arithmetic operations. Like the temporary relations between pitch and space enacted by singers, however, this is a limited analogy. If raga spaces were actually as strictly delineated as vector spaces are, a singer would be unable to stop singing one raga and start singing another.
15. There is some controversy over the exact difference between these two ragas. Some musicians say, that Ramkali has its own essence that is distinct from Bhairav. For example, a greater weight is given to pancham (the fifth scale degree.) Others say that these differences are merely incidental consequences of the Ramkali shifts themselves, which nearly always circle around the fifth scale degree and thus weight it more heavily. In the particular performance analyzed here, however, the performer performs Ramkali like Bhairav except for these shifts.
16. This performance was of Girja Devi at the Savai Gandharva Music Festival in Pune, India, December 12, 1991. The recording is available on VCD from Fountain: FMRVCD-13.
17. A more detailed analysis of this case can be found in Fatone, Clayton, Leante, and Rahaim (forthcoming).
18. E.g., Mian ki Malhar, Jayant Malhar, Darbari Kanada.
19. Fatone et al. (forthcoming).
20. See John Ohala’s (1994) proposed iconic linkages between size and vowel and Gentilucci and Dalla Volta’s (2007) proposed hand/mouth motor command system.
21. See Leante in Fatone et al. (forthcoming).
22. Although I have heard of cases of a teacher telling her students to gesture less in general.
23. Although these generalizations apply broadly to Indian music pedagogy, there is a wide variety of pedagogic practice even among teachers of khyal. Teachers may or may not teach in groups, may or may not have students live with them, may or may not accept payment, may teach for one hour at a sitting or up to five hours at one sitting, may require the student to practice only within earshot or not, may prohibit their students from hearing singers from other lineages or not, may expect to be called guru or not, etc.
24. A shishya is a disciple; parampara can roughly be translated as “tradition”.
25. Most students, of course, have had more than one teacher at one time or another, and this too has gestural consequences. But nearly everyone chooses one teacher with whom they identify.
26. Personal communication, November 28, 2007.
27. Young’s and McNeill’s formulations, as they each point out, owe much to Heidegger and Merleau-Ponty.
28. Interview, December 2, 2007.
29. As Pavan was well aware, he was born years before Gajananrao Joshi died.
30. His own direct teacher in the Gajananrao Joshi line is Vikas Kashalkar, shown above in the illustration of grappolo. Despite the unusually “closed” disposition reflected in the grappolo, excerpted from an alap, Kashalkar also stresses the importance of open voice in general during other parts of performance.

References

Bagchee, Sandeep (1996). Nad: understanding raga music. New Delhi: BPI.
Bailur, Chitra (1955). Music and personality. Lakshya Sangeet, 3 (2), 40–42 (quoted in Neuman,
2004).
Bakhle, Janaki (2005). Two men and music. Oxford: Oxford University Press.
Booth, Gregory (2005). Pandits in the movies: contesting the identity of Hindustani classical
music and musicians in the Hindi popular cinema. Asian Music, 36 (1), 60–86.
Clayton, Martin (2007). Time, gesture and attention in a khyal performance. Asian Music, 38
(2), 71–96.
Coomaraswamy, Ananda K. (1936). The mirror of gesture, being the Abhinaya darpana of
Nandikesvara. New York: E. Weyhe.
Deshpande, Vamanrao (1976). The National Center for the Arts presents Mogubai Kurdikar
with Kishori Amonkar. Bombay, September 16, 1976 (cited in Quinn, 1982).
Dhar, Sheila (2001). The cooking of music and other essays. New Delhi: Permanent Black.
Fatone, Gina, Martin Clayton, Laura Leante, & Matt Rahaim (forthcoming). Imagery, melody,
and gesture in cross-cultural perspective. In New perspectives on music and gesture. Aldershot:
Ashgate.
Gentilucci, Maurizio & Riccardo Dalla Volta (2007). The motor system and the relationships
between speech and gesture. Gesture, 7 (2), 159–177.
Goldin-Meadow, Susan (2003). Hearing gesture: how our hands help us think. Cambridge, MA:
Harvard University Press.
Katsman, Roman (2007). Gestures accompanying Torah learning/recital among Yemenite Jews.
Gesture, 7 (1), 1–19.
Kendon, Adam (2004). Gesture: visible action as utterance. Cambridge: Cambridge University
Press.
McNeill, David (2005). Gesture and thought. Chicago: University of Chicago Press.
Moran, Nicola Stephanie (2007). Measuring musical interaction: analysing communication in
embodied musical behaviour. Unpublished PhD diss., Open University, Milton Keynes.
Nadkarni, Mohan (1999). The great masters: profiles in Hindustani classical vocal music. New
Delhi: Somaiya.
Neuman, Dard Abraham (2004). A House of Music: the Hindustani musician and the crafting
of traditions. PhD diss., Columbia University.
Ohala, John J. (1994). The frequency codes underlies the sound symbolic use of voice pitch. In
Leanne Hinton, Johanna Nichols, & John J. Ohala (Eds.), Sound symbolism (pp. 325–347).
Cambridge: Cambridge University Press.
Pingle, Bhavanrav (1962 [1894]). History of Indian music with particular reference to theory and
practice. Calcutta: Susil Gupta.
Pradhan, Aneesh (2004). Perspectives on performance practice: Hindustani music in nineteenth
and twentieth century Bombay (Mumbai.) South Asia: Journal of South Asia Studies, 27 (3),
339–358.
Quinn, Jennifer Post (1982). Marathi and Konkani speaking women in Hindustani music,
1880–1940. PhD diss., University of Minnesota.
Robinson, Marilynne (2004). Housekeeping. New York: Picador.
Roy, Rabindra Lal (1934). Hindustani ragas. The Musical Quarterly, 20 (3), 320–333.
Seeger, Charles (1958). Prescriptive and descriptive music-writing. The Musical Quarterly, 44
(2), 184–195.
Subramanian, Lakshmi (2006). From the Tanjore Court to the Madras Music Academy: a social
history of music in South India. New Delhi: Oxford University Press.
Sweetser, Eve (2004). What we mean by meaning: conceptual integration in gesture analysis and
transcription. Gesture, 4 (2), 197–219.
Tagg, Philip (2004). Gestural interconversion and connotative precision. Université de Montréal.
2004. Accessed at http://mmmsp.com/tagg/articles/filminternat0412.html on December
27, 2007.
Weidman, Amanda (2007). Singing the classical, voicing the modern : the postcolonial politics of
music in South India. Calcutta: Seagull.
Young, Katharine (2002). The memory of the flesh: the family body in somatic psychology. Body
& Society, 8 (3), 25–47.

Author’s address

Matt Rahaim
1196 Keith Avenue
Berkeley, CA 94708
USA
mrahaim[at]gmail.com

About the author

Matthew Rahaim is a student of Hindustani vocal music and a member of the Berkeley Gesture Studies Group. He has taught in the department of music at the University of California, Berkeley and in the department of religious studies at Stanford University. Apart from gesture, his research interests include devotional performance, the history of tuning in India, and speech melody. He has contributed articles to New perspectives in music and gesture (Aldershot: Ashgate, forthcoming), World of Music, and the Yearbook for Traditional Music. He is currently writing his doctoral thesis at the University of California, Berkeley on the transmission of musical
knowledge through gesture.