Sullivan, Francis
Temple University

Francis Sullivan
Temple University
College of Education
Department of Curriculum, Instruction, and Technology
3rd Floor, Ritter Hall
Phila., PA 19122

E-mail address:

Strand: discourse English

Title: Accommodation, Resistance and Elaboration: Given-New and Strategic Conduct Toward Authority

This paper examines styles of elaboration as marking strategies whereby writers simultaneously project representations of information and social relations. It is argued that in situations in which social relations are institutionally hierarchical and asymmetrical, such styles function strategically. Writers use these styles to construct situationally specific representations of information and self that may accommodate or resist (Giroux, 1983) institutionalized asymmetries.

One of the more salient means by which writers construct relations with readers is through structuring patterns of information that distinguish "Given" (or old) from "New" information. Though most analyses of the Given-New distinction have focused on its consequences for sequencing information, others have examined the distinction in terms of how it affects the elaboration of information in a text. That is, how does a writer's assumptions about her audience's knowledge and beliefs affect the form in which she represents information as Given or New?

To answer that question, this paper adapts for use in a critical systemic framework a taxonomy of "Assumed Familiarity" outlined in a seminal article by Prince (1981). This taxonomy classifies representations of information through nominal groups on the basis of writers' assumptions about what readers know or believe and how they know it. "New" information, for instance, has two subcategories. Information may be new to the ongoing discourse but, presumably, familiar to readers from memory. Such information termed Unused, receives relatively minimal elaboration. Information may also be new to the discourse and presumed unfamiliar to readers. While such information may be introduced "as is," (e.g., a guy, a book), more often it is "linked" syntactically (Brand New, linked) to information assumed more familiar to readers (e.g., a guy I know, a book by Somerset Maugham).

The taxonomy also identifies two categories of representation--crucial to writers--that are not often addressed in the literature that treats Given New as a binary distinction. Information that is neither completely new nor completely old may be considered Inferable from other information in the discourse. Inferable information may be represented as a complex noun phrase (e.g., the burden of proof in this argument ); or information may be considered inferable from something mentioned earlier in the text (e.g., The next time that chair was used, the leg broke). Further, in a discourse situation characterized by a heterogeneous group of readers, information that is Unused for some will be Inferable for others. Such Blurred information is also represented through complex noun phrases, as in "Scholars, such as Kenneth Burke and Ernst Cassirer, have found the question of function. . . indispensable."

The importance of this taxonomy is twofold. First, the taxonomy is connected to Grice's Cooperative Principle, specifically the Maxim of Quantity. To be cooperative, writers will represent information only as explicitly as they believe necessary. All things being equal, the more familiar writers believe that readers are with the information, the less explicitly the information should be represented. Systematic deviations from this style of representation suggest that the maxim is being flouted for some purpose or that there may be a clash between fulfilling this maxim and another. In either case it suggest that this situation is not simply a cooperative exchange of information between equals.

Second, specific categories of elaboration--Unused, Inferable, and Blurred--make explicit the ideology, or world-view, that characterizes the register considered appropriate in this situation. Aspects of such world views range from stereotypical beliefs about common objects (e.g., that chairs have legs) to much more sophisticated, culturally-based beliefs, such as the basis of the connection between "argument" and "burden of proof" in the example cited above. The notion that arguments are the kind of thing to which a burden of proof is attached presumes shared membership in a highly circumscribed community. Thus, information represented in any of these categories identifies the kinds of knowledge, beliefs, and values considered shared and relevant in the discourse situation.

After locating this taxonomy within a critical systemic framework, I apply it to the results of research on university placement testing (Sullivan, 1990; Sullivan, 1995) as a situation in which institutionalized asymmetries between writers and readers should be played out in the features of texts. Analyses of the elaboration of information in 99 placement-test essays show specific patterns of elaboration that are substantively and significantly rewarded, as evidenced by the holistic scores received. Further, these patterns deviate systematically from what would be expected in a situation characterized by cooperation.