This article presents descriptions of and examples from qualitative case studies of 3 high school classrooms in Norway and the United States. The focus is on how classroom discourse and writing interact with each other and provide an important and unique instructional resource. The teachers in 2 of the classrooms consistently elicited, overtly valued, and helped develop student opinions and ideas. In this process, authentic questions and uptake were common, and a great diversity of voices was heard. Bakhtin's and Rommetveit's dialogical framework is used as the basis of analysis, as is Lotman's theory about the functional dualism of texts. The main argument is that the interaction of oral and written discourse increased dialogicality and multivoicedness and therefore provided more chances for students to learn than did talking or writing alone. In this way, the texts, both oral and written, were used to generate thoughts and opinions.

The move from theorizing difference to dealing with difference in an intercultural collaboration creates generative conflicts for educators and students. This article tracks the conflicting discourses, alternative representations, and political consequences the construct “Black English” had for Black and White mentors, teenage writers, and instructors in a Community Literacy Center collaboration. Comparing the accounts offered by resistance, conversation, and negotiation theory, it examines the dilemmadriven process of constructing a new negotiated meaning in the face of conflicting forces, voices, and representations. Dealing with difference in such collaboration means not only interpreting diverse verbal and nonverbal signifying systems based on values, experience, and competing discourses but constructing a new negotiated representation in the face of conflict that offers an (at least provisional) ground for action.

The evolution of technical communication conventions in America is more anthropologically complex than the traditional linkage to the scientific plain-style tradition suggests. Analysis of leading ideas in early 20th-century engineering writing textbooks and other primary sources demonstrates that disciplinary discourse conventions develop from an intricate nexus of human motivations, beliefs, and social activity. This article explores currents in American social and intellectual history that explain this complex, sophisticated view of language, which combines a rhetorically sensitive formalism with the ideas of professional literacy and cultural reading to facilitate communication with various audiences and to reinforce the status and dignity of the emerging profession.

Student writing in history courses, graded evaluation of that writing, and faculty interviews all reveal a contradiction between the stated and implicit aims of historical discourse. The explicit definition of writing in history is “argumentation”; the implicit expectation, however, is for narrative. This apparent contradiction highlights what the author argues is the central function of academic historical discourse: the establishment of an autonomous subject of meaning who is always speaking from outside history about a distant and objectified past. Students are rarely aware of the importance of this voice, even at an unconscious level, because faculty themselves fail to articulate for students the distinctive nature of their genre or the function of historical discourse generally. This project thus builds on previous studies in rhetoric by using the work of theorists of history to identify more precisely what it is in historical discourse that is hidden from student view—the autonomous, transhistorical voice.

This article describes an investigation in which I explored an impression that I had developed in earlier work that the grammatical subjects in scientific discourse are markedly long. An examination of a sample of scientific discourse produced evidence that makes a fairly strong case that on the average the grammatical subjects in the sample are markedly long. A stronger case can be made that many of the specific subjects in the sample are very long indeed, probably long enough to draw some attention to themselves in most any kind of discourse. I identify three pressures that I believe operate on scientists to produce very long grammatical subjects: The pressure to be precise, the pressure to be concise, and the pressure to be efficient and progressive in constructing a set of claims that will remain true within a framework of knowledge that has been built up over time. I conclude by exploring some possible connections between both the grammatical subjects in and the overall style of the sample of discourse and what Jerome Bruner calls the paradigmatic mode of thought.

Many authorities have come to recognize the critical importance of the Greek notion of kairos (right timing and due measure) in contemporary rhetoric. But Aristotelian scholars have generally ignored or demeaned Aristotle's use of kairos in his rhetoric, often contrasting it especially to Plato's full treatment in the Phaedrus. This lack of attention has been partially due to faulty indexes or concordances, which have recently been corrected both by Wartelle and programs like PERSEUS and IBICUS. Secondly, no one has hitherto attempted to go beyond the root kair- and examine the concept as expressed in other terms. This article will attempt to meet both of these concerns. It will first examine carefully the 16 references to kairos in the Rhetoric and show that the term is an integral element in Aristotle's own act of writing, in his concept of the pathetic argument, and in his handling of maxims and integration. There are also important passages using kairos in his treatment of style, often in conjunction with his use of the notion of propriety or fitness (to prepon). Possibly the two most important indirect uses of the concept of kairos can be seen in his definition of rhetoric and in his treatment of equity in both the Rhetoric and in the Nicomachean Ethics, probably the two most important treatments of the concept in antiquity.

Technical books for women of the English Renaissance provide a microcosm for studying connections among the emergence of technical writing as a genre, the rise of literacy, expansion of knowledge and technology, and replacement of orality by textuality as a result of increasing knowledge. These books on Renaissance technologies such as cooking, carving, household “physick,” home management, silkworm production, farming and estate management, midwifery, medical self-diagnosis, and gardening exhibit some differences from technical books written for men. Books for women are shorter and less detailed, but their style is similar to that of books for men. The style does not suggest writers believed that their women readers possessed an inferior reading comprehension level. Content differences seem to suggest that women's work was different from men's with many skills taught by oral transmission. The increasing complexity of the styles of technical books for women during the 16th and early 17th centuries suggests that women's reading skills increased as knowledge increased. Thus the oral style of the early 16th-century technical books disappeared with the need for an analytical style that would better convey growth of knowledge in the English Renaissance.

The style of discourse underlying writing instruction in this country, which has been termed essayist literacy by Scollon and Scollon and others, is grounded historically and culturally in the development of Western civilization. This style of discourse is the register of English used in academic situations, and it also has been found to be characteristic of some educated (especially male) mainstream speakers in other contexts. Because this register often differs from the naturally acquired discourse styles of students from nonmainstream groups, many such students face difficulties in writing instruction that mainstream students do not face. Given the importance of the essayist literacy register in this society, it is important (a) to make the characteristics of this discourse style explicit in order to increase the likelihood that writing instruction will be clear and available to all students, and (b) to learn about other discourse styles that are already known and used by students from a range of communities. A conceptual framework from the ethnography of communication is presented for studying verbal performances in different cultural contexts, and two examples of persuasive oral performances from ongoing research among Mexican immigrants are analyzed within this framework.

Rhetorical studies on experimental research articles in science have focused predominantly on introductions and discussions. The contextual nature of Results sections—the empirical heart of a scientific article—remains largely unexplored, however. What is known about the content of these sections comes from prescriptive style guides, which define Results as purely expository, leaving the argumentation to other sections of the article. This study examines one eminent biochemist's publications over time and a sampling of current articles authored by other biochemists. Six rhetorical moves were identified: (a) justifications for methodological selections, (b) interpretations of experimental results, (c) evaluative comments on experimental data, (d) statements citing agreement with preestablished studies, (e) statements disclosing experimental discrepancies, and (f) statements admitting interpretive perplexities. This investigation demonstrates that biochemists explicitly argue for the validity of their experimental data by employing certain rhetorical moves. Moreover, the findings challenge the traditional lore that Results sections engage in only simple, factual reporting.

This article proposes a method for examining how disciplinary differences in knowledge making are created or reflected at the sentence level. The method focuses on the grammatical subjects of sentences as key indicators of disciplinary knowledge making. Grammatical subjects of all sentences in sample academic journal articles were classified by a system identifying (a) the kind of abstraction or particularism involved and (b) the ways in which the researcher may or may not have foregrounded research methods and warrants. Findings from the sample articles in subfields of psychology, history, and literature indicated that psychology articles were more likely to foreground research methods and warrants and least likely to be particularistic. History articles tended to be intermediate. Literature articles were most likely to be particularistic and least likely to focus on research methods and warrants.

This essay presents a critical case study of how shifts in the style and genre of written communication both reflect and influence historical shifts in political consciousness and action. The field of study is the discourse of environmental advocacy. With increased public support for actions that would forestall environmental degradation, environmental politics has diversified. Formerly a resistance movement directed toward influencing large-scale governmental or industrial actions through the rhetoric of polemical dispute, environmentalism has evolved into several distinct approaches, including a globalist movement and a grass roots movement that share an interest in policy and procedure, the traditional topics of instrumental discourse. A new genre built upon this proactive attitude—the green how-to book—currently dominates the popular literature on environmental problem solving. Capitalizing on the document designs of technical communication, these manuals recommend courses of action ranging from fixing the Environmental Protection Agency to fixing the toilet; they are directed to audiences ranging from the President of the United States to the ordinary householder. They have in common an attempt to break the paralysis of fear associated with realizations about the scale of environmental damage. But—because the instrumental genre tends to obscure relations of agent, action, and effect—covert political agendas may pass unnoticed into the personalist politics of the new literature.

The almost exclusive reliance on evidence developed from documentary analyses, specifically analyses of textbooks, in composition historiography has resulted in an agonistic, heroes-and-villains image of the history of writing instruction, whereby modern composition scholars have defined themselves in terms of their opposition to what has come to be called “current-traditional rhetoric.” This article promotes the use of oral evidence in composition historiography to guard against overgeneralization and simplistic reduction of composition history to binary oppositions. Oral interviews also can serve as a way of collecting information that would otherwise be lost, of exploring the thoughts, motivations, feelings, and values of informants, and of giving voice to those marginalized politically, socially, and professionally. This article also defends oral data against positivistic attacks on its reliability as evidence and argues that the evidentiary value of any piece of historical data depends not on some abstract ranking of different kinds of evidence but on the historian's understanding of the rhetorical context informing the production of that data.

On-line computer conferences have been of increasing interest to teachers of composition who hope to provide alternative forums for student-centered, collaborative writing that involve all members of their classes in active learning. Some expect them to provide sites for discourse that are more egalitarian and less constrained by power differentials based on gender and status than are face-to-face discussions. These expectations, however, are largely unsupported by systematic research. The article describes an exploratory study of gender and power relationships on Megabyte University, one particular on-line conference. While the results of the study are not definitive, they do suggest that gender and power are present to some extent even in on-line conferences. During the two 20-day periods studied, men and high-profile members of the community dominated conference communication. Neither this conference domination nor the communication styles of participants were affected by giving participants the option of using pseudonyms.

This article analyzes “Timestyle” in order to identify the persona in the Nation and World sections of Time, a weekly newsmagazine. The sample for the study consisted of articles by 30 staff writers, articles that were randomly selected from Time' s 1988 issues. These articles were analyzed in six categories adapted from Walker Gibson's Style Machine: (a) word size and familiarity, (b) subject words and pronouns, (c) verbs, (d) modifiers, (e) sentence length and subordination, and (f) other effects of tone. The analysis suggests that Time' s narrative voice is a powerful one. The narrator has the power to put us at ease, engage our feelings, secure our trust, and divert our attention.

Text transmitted electronically through computer-mediated communication networks is an increasingly available yet little documented form of written communication. This article examines the syntactic and stylistic features of an emergent phenomenon called Interactive Written Discourse (IWD) and finds that the concept of “register,” a language variety according to use, helps account for the syntactic reductions and omissions that characterize this historical juxtaposition of text format with real-time and interactive pressures. Similarities with another written register showing surface brevity, the note taking register, are explored. The study is an empirical examination of written communication from a single discourse community, on a single topic, with a single recipient, involving 23 experienced computer users making travel plans with the same travel advisor by exchanging messages through linked computers. The study shows rates of omissions of subject pronouns, copulas, and articles and suggests that IWD is a hybrid, showing features of both spoken and written language. In tracing variable use of conventions such as sentence initial lower case and parentheses, the study shows that norms are gradually emerging. This form of written communication demands study because, as capabilities expand, norms associated with this medium of communication may come to influence or even replace those of more traditional writing styles.

The authors recount their attempt to analyze a case study in terms of two conflicting rhetorics: a collectivist rhetoric that values most the contributions individuals make to an ongoing collective project and an individualist rhetoric that values most the original and autonomous voice. These two rhetorics conflict in the experience of one writer working concurrently in a literature seminar within a university English department and in the public relations office of a reproductive services agency. This conflict, centering on different rhetorical ethics, had less to do with competence than with commitment: the writer's commitment to the individualist ethics practiced in the writing she did in the literature seminar prevented her from valuing the writing she did at the agency that worked toward a collectivist end. The authors then examine how this analysis is problematized by alternative interpretations of this case that demonstrate that the collectivist rhetoric practiced by researchers and theorists of writing itself involves the interaction of conflicting individualist assertions. This analysis suggests that the most useful theoretical insights any case might provide into the question of how writing ought to be taught are embodied in the exchange of interpretations that case provokes and in the confrontation of diverse arguments that emerge from that exchange.

This study analyzed the persuasive essays of high school juniors and seniors to determine the specific rhetorical and linguistic features that contributed to raters' holistic judgments about the overall quality of the essays. Essays written by a random sample of an ethnically, socially, and economically diverse population of high school writers were analyzed using an array of rhetorical and linguistic measures: overall quality, use of a five-paragraph structure, coherence, three types of persuasive appeals, and sentence-level errors. The relationships between the variables and the holistic scores were examined using a correlation analysis. A forward stepwise regression analysis was also used to estimate the amount of variance contributed by each variable. Results indicate that use of logical appeals, five-paragraph structure, coherence, and number of words were strongly correlated with the overall quality ratings.

Personal voice in writing is currently an all-too-subjectively understood notion. Different authors, Coles and Elbow, for example, have drawn appropriate attention to the voice phenomenon, but objective definitions and practical understanding are still lacking. One step toward understanding the workings of voice can be taken, however, by a linguistic analysis of structures that observably cause perception of a personal voice. Examining a limited set of data from professional writing reveals that one clear source of voice is appositive and parenthetical structures. These structures are produced “paragrammatically” by being inserted into a sentence, interrupting its normal flow, with the effect of creating a personal voice. They have a commentative function associated with a second-order “reflective mentality” and can be classified into at least three structural subtypes—displacements, equivalents, and interruptives—correlating with particular commentative functions. This analysis suggests, in general, that distinguishing between a second-order reflective mentality and a first-order factive mentality is central to the perception of voice. The intuitions of compositionists are important in uncovering discourse properties relevant to composition studies, and linguistic analysis is important for successful description of the phenomena and as a basis for pedagogical application.

The environmental impact statement (EIS) was created by the National Environmental Policy Act in 1969 as a means of ensuring careful study of possible effects on the environment of projects involving public lands and as an aid to effective decisions regarding such projects. This article presents a case study involving the reading of several EISs produced by one government agency, the Bureau of Land Management. An analysis of these documents reveals that, to answer the leading question of rhetoricians in the field of technical writing—Is the document effective?—we must consider the social and cultural context of the EIS as well as the characteristics of the text, its organization and style. Simple notions of purpose and audience are ruled out. We must account for pragmatics as well as syntactics and semantics. The very category of “effectiveness” is conditioned by the historical and political forces that shape the EIS. An approach through genre theory is recommended.

Introspection suggests that both writers and readers experience auditory imagery of intonation, accents, and hesitations. The suggestion here is that certain important aspects of this “covert prosody” of written language are reflected in punctuation. In order to study systematically the degree to which punctuation reflects the covert prosody of written language, one would like to find independent ways of uncovering that prosody. Two such ways are explored here: reading aloud and “repunctuating” (inserting punctuation in passages from which the author's punctuation has been removed). The article focuses especially on the relation between “punctuation units” (stretches of language between punctuation marks) and the “intonation units” of speech, and the variations in this relation that are found among different authors and different styles. It explores the degree to which different pieces of writing are prosodically spokenlike, and the degree to which they capture the prosodic imagery of ordinary readers. “Close” and “open” punctuation are discussed, as are selected grammatical sites at which there is a discrepancy between punctuation and prosody. It is suggested that an awareness of prosodic imagery is an important ingredient of “good writing.”

In their freshman year in college, Puerto Rican students take composition courses in both Spanish and English. Although the rhetorical structure of the final product, the composition, may respond to national writing styles in the two languages, studies show the composition process to be similar. Writing instructors in either language find similar problems in student compositions, regardless of the language code used. One of the difficulties students have in both languages is blocking, or apprehension about writing. Although some aspects of the composition process may be universal, we assumed that in bilingual writers the source of writing block depended on the language used. This article presents the results of a questionnaire designed to determine the sources of bilingual students' apprehension in writing by considering three groups of bilingual writers: graduate students in English, freshman English composition students, and freshman Spanish composition students. The results suggest some insights on the nature of blocking in a native language (Spanish) and a second language (English), which may then lead to ways of helping bilingual students to overcome blocking.

Hedges, which signal writers' tentative assessments of referential information, were added to a passage from both a science and a social studies textbook. The hedges appeared in either personal or impersonal voice; in the first half, second half, or both halves of the passages; and in either a low-intensity condition or a high-intensity condition. A measure of what subjects learned from reading the passages showed that they learned most when the hedges appeared in personal voice, the second half of a passage, and low intensity. Some extensions of the implications of this work to practices in composition classes—particularly practices of evaluating whether or not material should be hedged—are recommended in order to broaden students' critical-thinking abilities and their views of language.

This article surveys and analyzes the contemporary reception of Plato's rhetorical theory in contemporary rhetoric and composition studies by examining the response from three current perspectives: (1) presenting Plato as completely against rhetoric; (2) leaving Plato out of rhetoric altogether; and (3) interpreting Plato's work as raising issues central to classical and contemporary rhetoric. The discussion of the first two responses to Plato's relationship to rhetoric reveals a reductive, or formulaic, presentation of classical rhetoric. The discussion of the third perspective shows that it is the most accurate interpretation. Plato's rhetoric is related to the traditional five canons that were prominent in Greek rhetoric and explicitly systematized in Roman rhetoric, beginning with the Rhetorica Ad Herennium. If Plato's extensive contribution to the last two of the classical canons of rhetoric, memory and delivery, were more commonly included in the historicizing of rhetoric, then the five canons would work in the fullness of their interaction, rather than as the three-part system (invention, arrangement, and style) that dominates much current interpretation of classical rhetoric. Examples of reintegration of Plato into classical rhetoric (the third perspective) leads to a conclusion that Plato's rhetoric is central to contemporary interpretations of classical rhetoric.

This article draws upon data collected in a five-month study of primary grade writers to illustrate dimensions of variation in how young children orchestrate or manage the complex writing process. The observed children, all members of an integrated urban public school classroom, varied in the degree to which they focused on the diverse message forming and encoding demands of the writing activity and in when they maintained that focus. These differences may have existed, in part, because of differences in how the children made use of the available sources of support for their composing; that is, they differed in the degree to which other symbolic media (pictures and talk) and other children shaped their individual writing efforts. The children's composing behaviors were consistent with their apparent intentions and with their styles as symbolizers and socializers in their classroom. Viewing differences in children's ways of composing from the perspective of linear or uniform conceptions of writing growth may mask the holistic sense of each child's behavior.

Two studies investigated the editing strategies used by college basic writing (BW) students as they went about correcting sentence-level errors in controlled editing tasks. One study involved simple word processing, and a second involved an interactive editor that supplemented the word-processing program, giving students feedback on their correction attempts and helping them focus on the errors. In both studies BW students showed two clearly different editing strategies, a consulting strategy in which grammatical rules were consulted and an intuiting strategy in which the sound of the text was assessed for “goodness” in a rather naturalistic way. Students consistently used their intuiting strategies more effectively; however, errors requiring consulting strategies showed a larger improvement after intervention by the interactive editor. Cognitive implications of the editing strategies are discussed in terms of the requisite knowledge involved in successful application of each strategy.

Introductions to research articles (RAs) have become an important site for the analysis of academic writing. However, analysts have apparently not considered whether RA introductions typically include statements of principal findings. In contrast, this issue is often addressed in the manuals and style guides surveyed, most advocating the desirability of announcing principal findings (APFs) in RA introductions. Therefore, a study of actual practice in two leading journals from two different fields (physics and educational psychology) was undertaken. In the Physical Review 45% of the introductions sampled contained APFs (with some increase in percentage over the last 40 years), while in the Journal of Educational Psychology the percentage fell to under 7%. These figures are at variance with the general trend of recommendations in primary and secondary sources. Thus preliminary evidence points to (a) a mismatch between descriptive practice and prescriptive advice and (b) diversity in this rhetorical feature between the two fields.

This essay examines narrative choices in experimental (interpretive) and traditional (analytical) ethnographies. The material covered includes probability in quantitative and qualitative research; ethnographic narratives as ways of knowing and telling about the world; perspective as a consequence of both narrative stance and narrative voice; and the economics of producing interpretations and analyses in academic prose. Underlying the argument is the assumption that decisions ethnographers make about what to tell and how to tell it are influenced by to whom they plan to tell it and under what circumstances. Hence the ethnographer's narrative dilemma glosses over the epistomological crisis that authorship raises for the social sciences, namely, whether the researcher or the research method is telling the story.

This pilot study investigated relationships between individual differences in levels of writing skills and proficiencies at simultaneous and successive cognitive processing. Data from a group of 46 subjects indicate that scores on successive processing tasks were able to predict final grades in an introductory English composition course (p<.01). This suggested both the possibility and importance of investigating further how simultaneous and (especially) successive processing relate to writing skills. With three subjects used for pilot data, low scores in successive processing showed relationships with sentence-level errors and with the ability to develop sequences of ideas in writing. Low scores in simultaneous processing correlated with an inability to indicate clear relationships between sentences and paragraphs. Planning, a third cognitive factor, was found to be a powerful influence in organizing content. In the interaction of planning and simultaneous processing, lack of planning ability may interfere with the writer's ability to survey and thus organize his or her material.

Two hypotheses are outlined about the reasons for obscurity in expository writing. Neither accounts adequately for the general results of an exploratory study of the writing of postgraduates nor for the individual cases I present by way of illustration. A crucial factor, I suggest, is a person's implicit model of expository writing. Many of our subjects assume that the purpose of technical writing is to compress thought. I argue, on the contrary, that clarity is achieved through expansion.

The primary hypothesis was that field independent subjects would produce discourse that would be judged more coherent than the discourse of field dependent subjects. A total of 44 subjects in their first term of college composition were selected from a group of 60 volunteers from two universities and a community college. Each subject was administered the Culture Fair Intelligence Test, the Sixteen Personality Factor Questionnaire, and the Group Embedded Figures Test. There were five research conditions: Three evoked oral responses, and one evoked a written response. A group of readers unaware of the nature of the research evaluated each response holistically, rating it in terms of a coherence scale. Coherence scores were then analyzed in relation to cognitive style classification. The primary hypothesis was supported by the data. Analysis of variance (ANOVA) indicated significant cognitive style effect, F(6,25) = 4.82, p <.0001. The correlation between cognitive style and coherence was significant, r(32) = .54, p <.002. The results suggest that cognitive style is a significant variable in explaining differences between good writers and poor ones.

The composition teaching we tend to associate with nineteenth-century schools was exemplified by A. S. Hill's courses at Harvard, which emphasized correctness, clarity, stylistic refinements, and organization. But there was also a “reform tradition” that stressed the importance of the student's interests and experience, and saw the writing task as based on observation, description, speaking, and listening. Inspired partly by romantic educational theories from the continent, this tradition grew out of the social and educational reforms of the 1830s and 1840s and provided the basis for the early progressive teaching of the 1890s.

Ability to vary one's style is an important skill of mature writing, and it would be useful to have tests of this skill. We developed a cloze test to measure writing flexibility, then asked college students (all good writers) to replace sentences that had been deleted from two short stories. The style of the cloze sentences, for students with experience in creative writing, more closely resembled the original story than the cloze sentences of less experienced students. Style differences, between experienced and inexperienced students, appeared in average sentence lengths, sentence types, and verb-adjective ratios. In another experiment, less experienced students were given explicit instructions to imitate story style; they showed virtually the same adaptability to style as the creative writing group in the first experiment. Thus we have evidence that the cloze test measures style differences between experienced and less experienced writers, and also that responsiveness to style features, distinct from the skill needed to change those features, is a significant component of experienced writing.

Experts on style agree that writers frequently have trouble using the unattended anaphoric this clearly. Few, however, have proposed explicit guidelines for sorting appropriate from inappropriate uses. This article examines the limitations of a recent classification proposed by Moskovit (1983), and then suggests an alternate classification relying on concepts from functional grammar. In particular, Moskovit's distinction between demarcational, syntactic, and semantic reference is found not to predict actual readers' judgments. In its place, the authors suggest a classification based on the functional notions of topic and focus. The unattended this is shown to be English's economical routine for moving the focus of a discourse from nominal topics to clausal predications relating those topics. Before deciding to employ this routine, however, writers are warned to evaluate its consequences on clarity and rhetoric.

This study attempted to describe the structural and content characteristics of actual employment rejection letters (following job screening interviews). Their impact on applicants' feelings about themselves (self-concept and self-satisfaction) and about letters (perceptions of letter clarity, “personalness” and appreciative tone) are assessed. Results provide a profile of the “typical” rejection letter and indicate that while few of the letter characteristics affected applicants' feelings about themselves, a number of these attributes were related to applicants' perceptions of the letters.

This article demonstrates the potential of discourse analysis for exploring cognitive processes that occur during writing. Discourse analytic studies and text comprehension studies are reviewed for their contribution to a cognitive process view of writing. Research is reported which combines discourse analysis with on-line pause data to determine how semantic propositions reflect sentence-level planning patterns. Results indicate that decisions regarding predicate relationships are central to sentence production. Some implications for a process model of writing are suggested.