Raga and tala: the frame that makes improvisation legible

Two listeners can hear the same hour of music and describe it in incompatible ways. One says the performer played a raga; the other says the performer played in a raga. The difference matters, because the first description implies a composition and the second implies a place. In practice it is the second. A raga is a territory with a boundary, an entrance, some places you may linger and some you may only pass through, and the performer's skill is measured by how much of it he can make you see without ever leaving it.

That is the useful way into this pillar's machinery. Everything the previous three entries described — the salaries, the stages, the colour codes, the audiences — sits on top of a grammar that is unusually explicit. Indian classical music does not ask the listener to feel something unnameable. It asks the listener to follow a set of rules in real time, and it publishes those rules well enough that a knowledgeable audience can hear a mistake.

A raga is not a scale

A scale is a set of pitches in order. A raga is that set plus a set of instructions about how each pitch may be used. Some notes are strong and can be held; some are weak and exist to pass between the strong ones; some appear in the ascent and are omitted in the descent, which is why the two are given separately as arohana and avarohana. A raga also has a signature phrase — the pakad — a short melodic figure that identifies it faster than any list of notes could.

Two ragas can contain exactly the same pitches and be unmistakably different, because the difference lives in the treatment rather than the inventory. The same note held steady in one raga is oscillated in another; the same note is approached from above in one and from below in the other. This is the part that notation handles badly, and it is why the tradition remained oral for as long as it did: what has to be transmitted is not the pitches but the way of touching them.

A tala is not a beat

A tala is a cycle that repeats, and it is marked by hand. The performer's right hand counts: a clap on the strong divisions, a wave on the empty one, finger counts between. In the sixteen-beat cycle above, the audience too is counting — many of them silently, some of them not silently at all, because arriving together on the first beat is the whole point of the exercise.

The empty division is the detail that makes the system work. In the ninth beat of teental the drummer plays the pattern but withholds the resonant bass stroke, so the loop develops a hole. The ear hears that hole as clearly as it hears the downbeat, which means the listener has two orientation points instead of one, and can tell not only where the cycle starts but where he is inside it. Cycle and emptiness are doing the same job as a cadence: they tell you where you are without stopping the music.

What this buys the performer

Inside those two frames an enormous amount is permitted. The melodic improvisation proceeds in stages — in Hindustani music from unmeasured alap through a pulse, then a composition, then increasing rhythmic play; in Carnatic music through the composed kriti, then neraval on a single line of text, then kalpana swaram, improvised solfa passages that must land on a chosen point in the cycle.

That last constraint is the one that impresses musicians most. To improvise a passage of arbitrary length and have it end exactly on sam, in a cycle you have been counting internally while also singing, is a computational task as much as an artistic one — and it is the standard party piece of both traditions. The audience follows it because the audience is counting the same loop. Nobody needs to be told what just happened.

The parts that are convention, not physics

Two features of this system are often presented as deeper than they are. The first is the association of ragas with times of day: Bhairav in the morning, Yaman in the evening, Malkauns late at night. This is a real and observed convention with a long history in the treatises, and performers still respect it at the start of a concert. But it is a convention of practice, not an acoustic property, and a raga played at the wrong hour sounds wrong mainly to listeners who have been trained to expect otherwise.

The second is the number system. Seventy-two melakarta, ten thats, thirty-five talas, three hundred talas named in the old texts: these are classification schemes, and classification is a twentieth-century activity as much as a medieval one. Bhatkhande's ten thats were proposed between 1910 and 1932, in the middle of a debate about whether Indian music should be notated, and they are best read as an argument about how to teach rather than a description of what musicians do.

Which is the pillar's recurring shape. The grammar is old and genuinely constraining. The organisation of the grammar into a syllabus, a set of numbered categories and a printed textbook is recent, and it happened because the patronage changed — from a court that employed musicians for life to a classroom that had to produce them on a schedule.

This is entry 04 of Arts & Performance in About India. The next entry takes the rhythm half of this frame and looks at the instrument that speaks it — a drum whose strokes have names, and whose names are spoken before they are played.