Inevitably, any emerging field must find ways of defining itself apart from others or showing how it shares similarities and convergences with others; this is particularly true of an interdisciplinary or cross-disciplinary field such as composition studies. Another task is to outline its epistemological assumptions. As members of the composition research community have tried to define the field and its assumptions, we have discovered that English compositionis composed of a fairly loose confederation of members with differing intellectual histories, and, consequently differing perspectives on our teaching and research.

James Berlin provides us with one "map" of the conflicting positions in the field, outlining four kinds of pedagogical theorists: the NeoAristotelians, the Positivists or Current Traditionalists, the Neo-Platonists, and the New Rhetoricians. Even though a number of rhetoricians have disputed the rigidity of these categories, the differences he describes are useful because they offer one explanation for the various paradigms that exist within composition research. Because research in composition has been intimately linked to questions of pedagogy, differing pedagogical paradigms signal methodological debates within our research. Let us examine the underlying assumptions of these four theories.

For the Neo-Aristotelians and the Positivists, objective reality can be known through the senses, with the addition of either deductive or inductive reasoning. Language is a relatively unproblematic medium with clear connections between words and meaning. On the other hand, the Neo-Platonists and the New Rhetoricians, as Berlin describes them, believe that reality or truth must be somehow interpreted or constructed by human beings through language. Neo-Platonists believe that reality or truth is always mediated by the individual, while the New or Epistemic Rhetoricians believe that truth is constructed of the interactions among writer, audience, external phenomena, language, and culture. According to Berlin, "The New Rhetoric denies that truth is discoverable in sense impression since this data must always be interpreted--structured and organized--in order to have meaning" (774). The connections between language and meaning are far more complex for Neo-Platonists and New Rhetoricians.

Within composition research, we see similarities between NeoAristotelians and Positivists in the assumptions made by "quantitative" researchers. Neo-Platonists and New Rhetoricians are more similar to those we call "qualitative" researchers. Sandra Stotsky, in a very useful overview of research methodologies, provides a number of labels for the two kinds of research within our field: for "quantitative" research, she lists "positivistic," "scientific," and "hypothesis-testing" as rough synonyms; for "qualitative," she includes "holistic," "phenomenological," "hypothesis-generating," "participant-observational," "ethnographic," "humanistic," "naturalistic," "interpretivistic," and "hermeneutical," among others.

The American Educational Research Association, an organization to which some composition researchers belong, makes an additional distinction between "research" and "conceptual inquiry." The first depends on empirical data, information that can be sensed or experienced and collected, analyzed, and interpreted. The latter depends on philosophical or conceptual speculation and not on empirical data. Other writers ( Myers; Bridwell and Beach) have described positivist, rationalist, and contextualist paradigms for research, leaving open the possibility of some combination of conceptualizing and data-gathering. Systems of classification that separate conceptual theorizing from "research" are characteristic of social scientific or positivist definitions of research. In more recent composition research, the more salient distinction may turn on whether or not the patterns that are perceived within empirical data are presumed to exist a priori or whether they are described as socially constructed by the participants and by the researcher.

In their excellent overview of empirical research in composition, Lauer and Asher include the following types of research studies: case studies, ethnographies, surveys, quantitative descriptive studies, experimental designs, and meta-analyses. Rather than using methodologies as a means of sorting research, North classifies researchers and their characteristic "modes of inquiry" into four categories: experimentalists, clinicians, formalists, and ethnographers. Both books treat research as something different from the activities of practitioners, historians, philosophers, and critics, the activities typically associated with traditional humanistic teaching or scholarship. Not surprisingly, North directly attributes much about his way of working to Deising book entitled Patterns of Discovery in the Social Sciences.

North claims that practitioners are mainly interested in the question, "What do we do?" and scholars with "What does it mean?" Researchers, according to North, are supposed to ask "What happened (or happens)?" Such divisions are artificial, of course. Teachers ask questions about meaning and observations, scholars in English compositionare concerned with practice and observation, and researchers care about meaning and interpretation as well as practice. Using current deconstructionist theories, one could further argue that all science embeds certain philosophical and hermeneutical assumptions.

More recently, particularly with the borrowings of poststructuralism that we have seen in the late 1980s, there have been some shifts toward more speculative "research" that places less faith in observation or measurement, but for the most part such work is still called "conceptual" or "scholarly," and not "research." Typically, certain kinds of historians, philosophers, and critical theorists who work in composition are excluded from "research" by North, Lauer and Asher, and others because of the way the profession continues to limit its definition of research.

In most composition research, the heritage of social science is clearly revealed in the positivist belief that truth resides in reality and that reality can be observed, measured, and described so as to reveal that truth. These beliefs are clearly evident in quantitative studies, but they are also at the heart of some qualitative or ethnographic studies that interpret observations, artifacts, and relationships. The difference between a positivist and an interpretive or hermeneutical tradition for research lies in the degree to which researchers allow themselves to speculate about meaning and bias.

The facts must speak for themselves in hypothesis-testing research; either something is true or not true, within a certain range of probability. Of course, the researcher has to interpret the findings, but he or she is aware that this is mainly a descriptive task (i.e., describing all the possible influences on variables). Naturalistic researchers in a more hermeneutical tradition write "thick descriptions" and self-consciously interpret their data within a set of meanings that they often describe as "socially constructed." Many see their interpretations as constructive, rather than descriptive. They are aware of the shifts in interpretation that can take place from one observer to another. In the section on methodology, I discuss some of the ways naturalistic researchers deal with multiple realities. Of course, not all naturalistic researchers belong to the "social construction" school of thought, so it is possible that some researchers might import positivist assumptions about truth into their ethnographies, for example.

Another assumption of our research heritage is the faith that engaging in composition research can lead to solutions of certain problems, particularly pedagogical problems. This gives it a decidedly "applied," rather than "basic" or "pure" emphasis. (See Bridwell and Beach for a list of the challenges to a current-traditional pedagogy that have come from the research community.) Even during the 1970s when we heard repeated calls for "basic research" on composing processes, the emphasis was on discovering and applying this knowledge within the context of composition classrooms.

These ways of sorting out "types" of composition research and their shared assumptions, like all dichotomies and systems of classification, are often simplistic, reductive, and misleading. The members of the groups are almost never "pure types" incapable of change. All possible combinations exist (e.g., a feminist who produces critical theory as research/scholarship, but who gathers empirical data to get funds for her writing program). Some would argue that all research is conceptual and dependent upon a philosophical position, whether stated or unstated; that all reporting is interpreting; and that all research designs include both inductive and deductive reasoning, hypothesis-testing, and hypothesis-generating. Even those who want to maintain the boundaries between types may argue for an eclectic position that uses knowledge gained from all sources. Witte, for example, argues that composition can embrace qualitative studies and their "logic of discovery," as well as quantitative studies and their "logic of validation."

Nevertheless, because these distinctions exist within the literature on composition research, it is useful to understand the dialectical tensions they set up within the field and their consequences for emerging knowledge and theory within composition studies. More recent trends suggest that these systems of classification are becoming less rigid; for example, we routinely describe historical work as "research," especially that which examines empirical data such as writing textbooks (Connors). Moran and Lunsford, writing out of a tradition broader than many of the books on research cited above, include rhetoric, philosophy, psychology, and literary theory in their collection entitled Research in Composition and Rhetoric.