WHEN IS ORAL NARRATIVE POETRY? GENERATIVE FORM AND ITS PRAGMATIC CONDITIONS Dell Hymes I want to persuade you of three things: (1) that spoken narrative has a level of patterning that is likely to be found everywhere, or nearly so, but that has been missed in most research (in the common kinds of entextualization (Blommaert 1997: 15-24)); (2) that we are only beginning to grasp the complexity of such patterning as a part of narrative competence; (3) that such patterning may be present to different degrees, or even absent, in ways dependent upon personal and community circumstances and concerns. A fair amount has come to be known in regard to the first point (see 1. below). Something has begun to be known in regard to the second. What it would be like to know something about the third can be suggested. The first point has to do with transcending two conventional practices. One is the long-standing assumption that the narratives consist of paragraphs. Of course one knows that epics like the Iliad and Odyssey, and many other long narratives, are organized in terms of lines. We readily recognize their lines because they are organized internally. There are patterns in terms of which a line has a certain number of stressed syllables, alliterating initial consonants, feet of types defined by a tradition. We have a name for such lines, metrical. When metricality is absent, internal organization has been taken to be absent. Even those devoted to avoiding imposition of alien frameworks on languages, to describing the organization of languages inductively in terms of relations within the languages themselves, scholars such as Boas and Sapir, have left that devotion behind at the sentence's edge.Much of the traditional oral narrative of the world is published only in terms of paragraphs. Today there is a great deal of work that transcribes oral narratives in terms of lines. The working assumption is to attend to pauses, phrases, tone groups and the like. Such features are noted to divide a narrative. Largely ignored are features that enter into relationships to organize a narrative (see discussion of John L's story below). In recent years it has become clear that in many languages, perhaps all, there is an organization of lines in terms of which the shape of a narrative can be discovered inductively and shown on the page. The relations are not internal to lines, not metrical, but among lines, 'measured'. There are regularities in the relations among measured lines, just as there are regularities in metrical lines. Especially clear accounts of the identification of lines in spoken narratives are given by Virginia Hymes (1987) and Joel Sherzer (1987, ch.4; 1990, pp.17). On the vocal articulation of lines in ritual wailing in three South American cultures, see Urban (1991), pp. 110-1, 152-9. These regularities have to do with cultural patterns, but also with the explorations and skill of narrators. In terms of cultural patterns, communities appear to build upon one of two alternatives: Relations in terms of two and four, or relations in terms of three and five. English narratives in the United States appear to work in terms of three and five, but in parts of Ireland in terms of two and four. Among Native American groups, Navajo, Zuni and others build in terms of two and four, while Chinookan and Sahaptian build in terms of three and five. In a flourishing narrative tradition one may find a variety of ways of making use of such relations, and of varying their aesthetic effect. A set of three or of five may be a set of three or five pairs. A sequence of ten lines may be a single rhetorical unit. A set of lines may be balanced internally, or a run from start to end. Uses of quoted speech and of catalogs are variables also. (See Hymes 1994, 1999) For many narrative traditions, discovering lines and relations is pretty much all that we can do. To be sure, it can lead to understanding and interpretation otherwise not possible. We can recognize artistry and subtleties of meaning otherwise invisible. For a true account of the human capacity for verbal art, this is crucial. But as with the Iliad and Odyssey, there are no people to observe and question, only texts. The second and third points have to do with when and where and to what degree narrators are conscious of such relations, or sensitive to them, vary in their deployment of them, as in regard to gender, are perhaps unconsciously affected by their presence. Pragmatic research should not be ignorant of the first point, when dealing with narrative. It can make major contributions to the second and third.