commissioned performance, subject to the self-selection of ethnographer and informant and influenced by the need of the performer to please the patron. Furthermore, the quantity of material that can be collected through recitation and transcription is minuscule compared to that available from printed texts. These texts enable many comparisons and contrasts that would otherwise be impossible.
Pritchett makes ingenious use of the publication records of the South Asia Microform Project (SAMP) to outline a history of mass printing in Hindi and Urdu beginning in the 19th century. This data reveals both the extent and continuity of the qissa phenomenon, and gives us a feel for the changing conditions that shape a once-oral folk tradition as it becomes commercial, popular, and written. Pritchett is the first person I know to have talked extensively with popular publishers in India. However, she is somewhat naive about the role of publishers in the production process. She echoes their statements that they publish only on demand, overlooking the possibility that they manipulate reader interest (through advertising and packaging of their texts), or that they intervene in the construction of texts (by prescribing length, plagiarizing other authors, cutting and pasting). The statistics of text production may not accurately gauge popularity, given the additional interference of middlemen in the distribution process, the effect of local markets and competition with rival publishers, and the widespread "remaindering" of excess copies (which are sold off as waste paper).
For these and other reasons, I am somewhat dubious of Pritchetfs use of SAMP statistics to identify a "core" of "best-seller" qissas, and an "outer layer" of marginal interest. (27) The construction of a "classic qissa" prototype (29), besides freezing a fluid folk tradition into an exclusive canon, serves to exclude important Indo-Islamic romances such as Laila Majnun, Shirin Farhad, and Hir Ranjha. These stories have exerted a great impact on the folk and elite literatures of South Asia and beyond, and they are also popular texts in the regional musical drama (sangit). Omitting these masterpieces of tragic love, Pritchett asserts that the qissa "always reaches a 'happy ending'" (163) The ambiguity of different narrative structures has no place in her definition of qissa.
There is also a rather misguided attempt to reduce the diversity of the narratives she chooses to a single line of diffusion from the Persian dastan. Pritchett privileges the Persian tradition, positing it as the elite parent that produced the popular offspring. Her Hindi tales constitute "changes" in which "some of the stock characters of Islamic story tradition are replaced by their Hindu counterparts, " a faqir becoming a yogi, or a pan an apsara. (29) While there is a certain poetic justice in the primacy Pritchett assigns to the Islamic traditions, considering how neglected they have been in Indian folklore studies, she is simply wrong in asserting that "there were no other readily available generic categories into which such works [Raja Vikram, Saranga Sadabrij, Tota Maina] might fall." (26) In Sanskrit, narrative genres existed such as itihas, pur ana, and katha', in medieval Hindi there were akhyan, gatha, and git, as well as specific types identified by subject, e.g. alha, bharthari, dhola, soratha, What is needed instead is some sort of model acknowledging the porous boundaries across which tales travel from one genre to another, one community to another, one region to another. The question of indigenous labels also needs to be further thought out, for the printed attribution of a text to a genre (e.g. qissa) may reflect publishers' concerns rather than readers' responses, to say nothing of the intent of the author or
Annual of Urdu Studies, #7 ^4