Stylistic Difference and Narrative Choices pieces in museums, one comes face to face with the unexpected richness and the sometimes frustrating complexity of the Bengal pata. Here we have not one but multiple traditions, each with a distinguishable style. The usual width of the scrolls and the module for register size; standardized compositional formats, settings and 92 backgrounds; linear qualities, preferred colours and their application; the approach to the problem of the third dimension; the whole structural approach to the human form, including iconometric formulae, chosen perspectives, articulation of limbs; the vocabulary available for characterization of figures, the repertoire of gestures, postures and expressions; and even certain details of the performance tradition, all seem to differ from region to region. Any attempt, however, to map the extent of a style is immediately confronted with a number of problems, some of which are material and some methodological. The problem of scarce material, in a poor state of preservation, with little or no documentation, is familiar to all art historians but much more so to those who work with folk art. We are fortunate that collectors such as J.C. French and Gurusaday Dutt acquired and preserved a number of patas and even noted their provenances, themes and in some cases the names of the painters themselves. However, while working with these sources one is always aware of the smallness of the sample and the operation of that, chance moment in history when the collection happened to be made; and one feels as a constant presence the collector's individual taste.2 It is difficult in these conditions to make any authoritative generalizations about the various stylistic idioms, or at any rate, to make them with a good conscience. Possibly the most unattractive aspect of setting down the notion of a style is the way it establishes itself as a disembodied template: once something is defined as the style of such-and-such it begins 'to perform a normative function, whereby other objects discussed in the future are obliged to measure up to or seen as differing from such norms. During fieldwork one comes across many instances that contradict the notion of a monolithic, localized style: not only does one witness a number of idioms operative in one village but also at times within the one individual who playfully displays his ability to paint in 'this style" and then in 'that style'.3 The most attractive solution for these problems is naturally to adopt the perceptions or terminology of the patuas themselves. Unfortunately, questioning the patnas about their perceptions of each other's styles is often a puzzling experience: one is often faced with what seems like a dogged refusal to see, or a disinterest in, any idiom not their own.4 Patuas are quicker to acknowledge distinctions within their community than in their work. The community has vertical divisions of an internal heirarchy, as well as regional divisions, with few kinship linkages between patuas of different regions. While writing for the 1951 census, S.K. Ray had put forward the now controversial concept of patua samajik schools. Ray said these were 'probably circles of families in a particular area (linked) in the bonds of kinship and (which) protected the monopolized Gharwana qualities — something akin to a trade secret or patent. As a general rule a particular Samaj Bandhani roughly coincided with the bounds of its particular traditional and regional school of painting. . . /5 Ray identified three important sa- Joiirnal of Arts 6' Ideas