[I wrote most of this in 2012. This spring I decided to publish it (mostly) as-is; I may revise or add to it later.]
Govind Chandra Pande (1923-2011), an Indian historian of the Vedic and Buddhist periods, provided a set of criteria for separating earlier and later strata in the Pali suttas. This review will focus on those criteria as presented in his Studies in the Origins of Buddhism (1957; 4th revised edition Delhi: Motilal Barnasidas, 1995).
Pande (pp. 27-50) identified six basic criteria for stratifying the Pali suttas: doctrinal evolution, literary evolution, interpolation, vocabulary and linguistic features, geography, and political and social data. Before his basic list, he gives an overview of other scholars’ stratification criteria (or, indeed, of their lack of interest therein); after the basic list, the first half of the book applies those criteria, somewhat opaquely, to the contents of the Sutta Pitaka.
Here is a quick overview of Pande’s criteria:
|DOCTRINAL EVOLUTION||• collecting, explicating, harmonizing the Master’s sayings
• increase in the number, extent, subtlety, and frequency of “enumerated groups”
|• interpretations of interpretations, in sectarian context
• growth of systematic Buddhist “theology”
• specific shifts in ideas
|LITERARY EVOLUTION||• dialogues (reminiscent of Upanishads, or Socratic)
• style of “simplicity, spontaneity, earnestness”
• “sun-lit by apt similes and parables”
• concerned with practice or biography
| • scholastic controversies
• “dry-as-dust catechetical style”
• exaggeration in numbers
• similes worked out in extreme detail
• concerned with philosophy
|INTERPOLATION|| • looser statements
• statements of an “unorthodox” character, at odds with later editors
| • more compact formulae
• conventional beginnings or endings added
|VOCABULARY & LINGUISTIC FEATURES||• metaphoric use||• words rare in Pitakas, frequent in later works
• analytical use
• archaic forms continue to be used in verse
|GEOGRAPHY||• area referenced is limited to central Ganges area||• extension of area known, esp. to south & west|
|POLITICAL & SOCIAL DATA|
Pande dismissed the last two right in his initial discussion, however. The extent of geographical knowledge in the Buddha’s time is difficult to reconstruct, though presumably both information and travel by monks and others went primarily along trade routes. (He did mention, tantalizingly, the appearance in MN 93 of “yonas” who have only two castes—slaves and freemen—and wonders if this could be an pre-Alexandrian colony of Ionian traders!)
Pande complained that reconstructions of Buddhist social history have drawn freely from sources of different ages (and purposes) to create a “composite” he implies to be invalid. He noted that the social history of pre-Asokan India can only be traced in longer periods of time than are involved in the formation of the Nikayas; some future comparison with Brahmanic and Jinist sources may be of limited help.
Pande also mentioned a criterion he did not formally list: prose and verse. As of the time he wrote this book (1957), poetic form had been already been claimed by several scholars to indicate either earliness and lateness. He discredited the latter on grounds that prohibitions against verse seem to have had in mind a specific sort of “Sanskritization” and specific ways of chanting. He pointed out that verse could help preserve a text under conditions of oral transmission. He concluded that the prose-poetry distinction is of no stratigraphic significance.
This leaves, essentially, four criteria: doctrinal evolution, literary evolution, signs of interpolation, and vocabulary and other linguistic features.
In the absence of independent sources, the evolutions of three of these—doctrine, literary form, and language—are inevitably circular. Doctrines known to be early would, of course, mark the suttas in which they occur as early (at least in part)—but how do we know which doctrines are early? The problem is the same for literary evolution and for vocabulary and other linguistic markers. Most of what we know about the ideas and forms of expression from this period comes from the suttas themselves. Unless we have independent evidence, we risk making our preferences, hunches, or biases about these matters into criteria.
Signs of interpolation offer more promise. Some purported signs of interpolation are, of course, items the historian believes (on other grounds) to be “too late” to have come from the Buddha’s time. These are simply examples of the circular arguments above. But there are certain signs that a text has been stitched together from elements of different origins, signs that go beyond the circular reasoning cited above. These include grammatical awkwardnesses, puzzling repetitions, mismatched facts, mismatches between supports and conclusions, and shifts in vocabulary. Such redactional (i.e., editorial) seams point credibly to diverse sources; they do not indicate the age, or the relative ages, of those sources.
What Pande has offered is not so much a set of reliable criteria for separating early, middle, and late in the suttas as a conjectural snapshot. He imagines the Buddha’s original circle to have been a certain way—and everything in contrast to that, he imagines as later developments of the tradition. It is a reasonably judicious set of imaginings, however. And such conjectures may be all we have to work with, at least until/unless extensive textual-critical work is done on the suttas.