Kangra was painted somewhere else

In the standard account, Kangra painting is the lyrical last phase of Pahari painting: cool greens, figures in profile, a landscape that participates in the mood instead of framing it. The tell is the face. From about the 1790s the women are drawn with a straight nose running down from the forehead in one unbroken line — a convention so specific that it will date a folio to within a decade.

It dates it. It does not locate it. And the location is the strange part.

A workshop is a firm, not a place

The paintings we call Kangra were made by a family. Pandit Seu worked at Guler, in the hills above the Beas. His sons Manaku and Nainsukh took the workshop further: Nainsukh spent years in the service of Balwant Singh of Jasrota, painting his patron fishing, listening to music, inspecting a building site, sitting in a doorway in the rain.

That is the correct unit of analysis. A court painter followed a court. When the patron moved, the workshop moved; when the patron died or ran out of money, the workshop found another patron. The style travelled with the family, not with the soil — which is why a folio painted at Jasrota and one painted at Nadaun can be by the same hand and look like each other, and why the name of a state is a poor label for a picture.

The economics are visible in the materials. A folio is wasli — several sheets of paper pasted together and burnished — and the pigments are ground minerals and vegetable dyes; in the earlier Basohli phase the luminous green was made from the iridescent wing cases of beetles. A set of a hundred folios is not a picture, it is a capital project. Only a court could commission one, and only a court could keep a room of specialists fed while it was being made.

The address problem

Kangra town was not available. It stayed under Mughal control until 1786, and after 1809 it belonged to the Sikhs. The painting that carries its name was made elsewhere: at Alampur, at Nadaun, and at Tira Sujanpur — three places on the Beas, upriver and downriver of the town, none of them the town.

The reason is banal. A court painter needs a court that is present, and Sansar Chand's court was where he was. When the Gurkhas besieged Kangra fort around 1805 he moved to his hill palace at Tira Sujanpur, and the atelier went with him. So the greatest concentration of "Kangra" painting was produced at addresses that a modern postal service would not call Kangra.

What ended in 1809 was a payroll

In 1809 Ranjit Singh's armies drove the Gurkhas out of the western hills, and the fort at Kangra passed to him. Sansar Chand kept his title and a reduced state; the atelier continued, but the work after that date is not the work of 1785 to 1805, and the surveys say so plainly.

What collapsed was not a way of painting. It was the revenue that paid for a room full of specialists — a designer, a colourist, a face painter, a gilder, each doing one operation on the same folio. That division of labour is expensive, and it only exists where a court can afford it. When the money thinned, the painters left: to Lahore and Amritsar under Sikh patronage, to Chamba, Mandi and Jammu, and eventually to the bazaar workshops that painted for British buyers.

The style did not die. It stiffened — the same figures, the same greens, painted faster and for less. Which is what usually happens to a tradition whose problem is a payroll rather than an audience.

The name on the box

Now the part that makes this more than local history. Almost none of these paintings is signed or dated. Attribution runs on family style: a hand, a workshop, a generation, traced through the Seu family and their descendants. That method can tell you who painted something with reasonable confidence. It cannot tell you which courtyard they were sitting in.

So when the corpus had to be organised in the twentieth century, the label went to the biggest patron. Kangra was the largest state, Sansar Chand the most prolific buyer, and "Kangra" was the obvious word. The name then absorbed everything painted within a hundred and fifty kilometres over eighty years — Guler work, Basohli work, Nurpur work, Chamba work.

The man who did most of the absorbing was a civil servant. M. S. Randhawa was Deputy Commissioner of Ambala from 1948, and from that year he spent his leave going from hamlet to hamlet in the Kangra valley, buying paintings and locating collections. He published eight volumes on the subject between 1954 and 1967, then more, including Kangra Valley Painting in 1972. He later became the first Chief Commissioner of Chandigarh and planted the city's avenues. The books are still the shelf anyone starts from.

Why the label matters commercially

Because a label is a price. A folio catalogued as Kangra will outsell a folio from the same hand, in the same decade, catalogued as Guler — and the difference is not connoisseurship, it is filing. The market needs a school the way a library needs a spine label, and it pays for the spine label.

This is not a scandal, and nobody involved was dishonest. Randhawa's volumes are careful, and the scholars who followed him — Goswamy above all — spent decades putting the family workshops back together and taking the corpus apart again. The point is subtler: the correction happened in the library, and the price did not follow it out.

Which leaves the paintings themselves. They are extraordinary, and nothing here disputes it. But a tradition, in the form we usually meet it, is often a bibliography — a set of decisions about which box to put things in, made by particular people in a particular decade, and then inherited as if it were a fact about the world.

Kangra painting exists as a category because somebody needed a name for the box. The box has been the unit of value ever since.