UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Toward a rhetoric of reading Brent, Douglas Allan 1989

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-UBC_1989_A1 B73.pdf [ 11.68MB ]
JSON: 831-1.0098233.json
JSON-LD: 831-1.0098233-ld.json
RDF/XML (Pretty): 831-1.0098233-rdf.xml
RDF/JSON: 831-1.0098233-rdf.json
Turtle: 831-1.0098233-turtle.txt
N-Triples: 831-1.0098233-rdf-ntriples.txt
Original Record: 831-1.0098233-source.json
Full Text

Full Text

TOWARD A RHETORIC OF READING By DOUGLAS ALLAN BRENT B.A., The University of British Columbia, 1974 M.A., Carleton University, 1976 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF ENGLISH We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April, 1989 (c) Douglas Allan Brent, 1989 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada DE-6 (2/88) ABSTRACT Rhetoric is a process not only of persuasion but also of so c i a l l y constructing knowledge. A rhetorical model of reading accounts for the role of reading in this epistemic process by expanding the canon of invention to explain not just how one can find arguments but also how, through being persuaded by the texts one reads, one can invent the propositions that one wishes to argue. Thus i t complements theories of rhetorical production with a theory of rhetorical consumption. The model combines the insights of traditional and modern rhetoric, reader-response l i t e r a r y theory and discourse processing theories of comprehension. Reading is described as a constructive process guided by the reader's prior knowledge and values (his repertoire) and by the matrix of questions that he wishes to ask of the text (the rhetorical situation). However, to be rhetorical, reading must not be f u l l y free; the writer must be able to predict the reader's response. This predictive a b i l i t y results from the public, shared meanings of words, from commonalities among discourse communities, and from the expectation that the reader w i l l attempt to build a coherent, unified interpretation. Rhetorical reading also involves judging which elements of which texts to accept into one's own belief system. This process is accounted for by an expanded notion of the three rhetorical pisteis: logos, pathos and ethos. The concept of the enthymeme is adapted to describe how readers find logical connections between the new i i perspectives presented by texts and their own prior knowledge (doxai). The reader may also be persuaded by emotional connections (pathos) and by her response to the writer's implied character (ethos). The c r i t i c a l potential of the model is illustrated by an analysis of an intertextual dialogue concerning the value of James Kinneavy's taxonomy of discourse. The model provides terms of reference that highlight the ways in which the participants in the dialogue construct and judge Kinneavy's and each others' works. "Rhetorical reading can also inform composition pedagogy, notably the teaching of the research paper, by providing an account of mature performance. The Electric Monk was a labour-saving device, like a dishwasher or a video recorder. Dishwashers washed tedious dishes for you, thus saving you the bother of washing them yourself, video recorders watched tedious television for you, thus saving you the bother of looking at i t yourself; Electric Monks believed things for you, thus saving you what was becoming an increasingly onerous task, that of believing a l l the things the world expected you to believe. Unfortunately this E l e c t r i c Monk had developed a fault, and had started to believe a l l kinds of things, more or less at random. Douglas Adams, Dirk Gently's Holistic Detective Agency iv TABLE OF CONTENTS Abstract: i i Intr eduction: What Is a Rhetoric of Reading and What Can It Do? 1 Chapter 1: Starting Points 1.1: The Traditional Framework 7 1.2: Rhetoric in a Constructionist Age 18 1.3: Reading as a Special Form of Invention 26 1.4: The Shape of a Rhetoric of Reading 29 1.5: An Interdisciplinary Approach to a Rhetoric of Reading 31 Chapter 2: Reading as Construction; Reading as Communication .2.1: Sources for a Theory of Interpretation 34 2.2: Reading as a Constructive Act 43 2.3: Materials of Construction: The Repertoire 51 2.4: Materials of Construction: The Rhetorical Situation 61 2.5: Constraints on Construction: The Writer and the Text 72 Chapter 3: From Evocation to Belief 3.1: Sources for a Theory of Evaluation and Synthesis 84 3.2: What Does It Mean to be Persuaded 87 3.3: How Texts Persuade: The Logical Dimension 97 3.4: How Texts Persuade: The Emotional Dimension 110 3.5: How Texts Persuade: The Ethical Dimension 119 3.6: The Outline of a Rhetoric of Reading 134 v Chapter 4: The Rhetoric of Reading as a C r i t i c a l Technique 4.1: Rhetorical Criticism and the Unending Conversation 137 4.2: The Influence of the Rhetorical Situation 147 4.3: The Influence of Values and Assumptions 151 4.4: Furthur Turns in the Conversation: Responses and Rebuttals 170 Chapter 5: Implications and Conclusions 5.1: Implications for Teaching 182 5.2: Implications for the Art of Rhetoric 205 Works Cited 215 v i D. Brent 1 Toward a Rhetoric of Reading Introduction: What is a Rhetoric of Reading and What Can It Do? Rhetoric and reading have traditionally been seen as related but separate arts. Rhetoric is the art of communication seen from the point of view of the producer of discourse; i t s function i s to inform the rhetor as to how best to develop and manage arguments, arrangement and style to produce persuasive discourse. Reading, on the other hand, i s the art of communication seen from the consumer's point of view. The function of a theory of reading i s to inform the audience as to how best to interpret discourse in order to appreciate what a writer has produced. Recent revisions in our view of knowledge and of rhetoric have created a theoretical framework that permits, or rather, impels, an expansion of the scope of rhetoric and a much more integrated view of the composition and interpretation of discourse. The social constructivist view of knowledge holds that knowledge is not made simply through individual encounters with the physical world. Rather, knowledge exists as a consensus of many individual knowers, a consensus that i s negotiated through the medium of discourse in an unending conversation that involves a l l humanity.1 . What makes rhetoric interpenetrate a l l aspects of this conversation ^Kenneth Bruffee provides a useful overview of social constructionist thought in "Social Construction, Language, and the Authority of Knowledge: A Bibliographical Essay." In the f i e l d of rhetoric, the implications of this view can be seen most f u l l y worked out in Wayne Booth's Modern Dogma and the Rhetoric of Assent. D. Brent 2 is the negotiated nature of knowledge, which implies a competition for belief. Knowledge i s not what one has simply been told. Knowledge i s what one believes, what one accepts as being at least provisionally true. The process of symbolic negotiation i s thus a process in which competing propositions attempt to establish claims to be believed; only when such a claim i s established can a proposition be elevated to the status of knowledge.2 Thus persuasion, the essence of rhetoric, l i e s at the heart of this endlessly recursive process of producing and consuming discourse. This implies that rhetoric i s epistemic—that i t participates not just in the conveying of knowledge already formulated but also in the making of knowledge through symbolic interaction. It also implies that the consumption of discourse through hearing or reading is also epistemic. To attend to a discourse is not simply to absorb another person's meanings but also to participate in the creation of new knowledge through the process of symbolic negotiation. The specific task undertaken here i s to explore the implications of embedding reading in a larger process of rhetorical negotiation. The inquiry w i l l be confined to a specific kind of reading: not the reading of literature for the purpose of aesthetic enjoyment, but the reading of non-literary sources for the purpose of building knowledge. Aside from posing some obvious terminological problems ("What is 2See Michael Polanyi's Personal Knowledge for a highly influential account of the way knowledge is constructed communally, particularly in s c i e n t i f i c discourse. Polanyi notes that knowledge cannot continue to exist without being shared: "In order to be satisfied, our intellectual passions must find response. This universal intent creates a tension: we suffer when a vision of r e a l i t y to which we have committed ourselves is contemptuously ignored by others. For a general unbelief imperils our own convictions by evoking an echo in us. Our vision must conquer or die" (150). Thus arises what Polanyi ca l l s the "persuasive passion": the drive to s o l i d i f y one's own beliefs by persuading others to share them. D. Brent 3 'literature' as opposed to 'non-literature'?" "What do we mean by 'knowledge'?"), this inquiry encloses a number of more specific and closely related sub-questions: 1. What aspects of rhetoric, both traditional and modern, can inform a theory of reading? 2. How can we best describe within a rhetorical framework how the writer, text, reader and rhetorical situation co-operate to generate meaning? That i s , how does a reader know what propositions the author of a text is trying to persuade her to believe? 3. How does the reader evaluate the propositions presented by individual texts and decide which to be persuaded by? 4. How does the reader negotiate among the claims of various texts in order to develop a unified system of knowledge? The inquiry is thus f i r s t a philosophical one that enlarges and redefines the theoretical horizons of both a c t i v i t i e s with which i t deals. But i t has a number of more practical consequences as well. F i r s t , i t can provide a new direction for rhetorical analysis by enabling a more complex discussion of the relationships among texts. A concern with intertextuality is not new to rhetorical criticism, but most rhetorical criticism of an intertextual nature concentrates on lin g u i s t i c echoes—borrowings of phrases, tropes, stock appeals. 3 Such analyses are not intended to account in detail for the ways in which writers' participation in a larger conversation builds their beliefs as well as their language structures. By examining texts from the point of view of 3For a particularly good recent example of this form of rhetorical criticism, see Keith D. Miller's recent intertextual analysis of Martin Luther King's rhetoric, "Martin Luther King, Jr. Borrows a Revolution: Argument, Audience and Implications of a Secondhand Universe." D. Brent 4 a larger rhetorical conversation, we can focus not just on relationships between persuasive techniques, but also on relationships between the propositions argued by texts in the context of rhetorical inquiry. In other words, we can focus not just on why a particular author phrases his arguments in just such a way; we can also inquire into the development of the ideas themselves. This entails a focus on the ways in which different authors appear to interpret the same written sources in different ways, thereby suggesting how particular configurations of personal beliefs, goals and prior knowledge influence the reading process. In short, then, the rhetorical analysis of texts can be refocussed from a description of persuasive methods to a description of the ways in which beliefs are constructed and negotiated in discourse. Second, a rhetorical model of reading can also inform composition pedagogy. Aside from providing models for imitation or prompts for expressive writing, reading in the composition class appears most often in the context of the "research paper." The expanded definition of reading argued for here implies that the "research paper" is not just a peripheral genre characterized by mechanical operations such as note-taking and documentation.4 Rather, because i t requires the student to synthesize beliefs by locating, interpreting and evaluating written sources, the research paper is a classroom form of one of the most fundamental human a c t i v i t i e s . The obvious d i f f i c u l t i e s that most beginning writers have with this form are symptoms not only of a lack of familiarity with the form i t s e l f but also with the larger processes involved in joining a knowledge-producing conversation. By suggesting **The argument that many modern composition handbooks treat the research paper in this manner w i l l be made in chapter 5 of this dissertation. D. Brent 5 what i t re a l l y means to perform this task successfully, a rhetoric of reading can provide a firmer foundation for instruction. Finally, an understanding of non-aesthetic reading has implications much more general than textual analysis or pedagogy. In The Pursuit of  Signs, Jonathan Culler argues that the study of literature should be more than a mere piling-up of interpretations: "To engage in the study of literature is not to produce yet another interpretation of King Lear but to advance one's understanding of the conventions and operations of an institution, a mode of discourse" ( 5 ) . The same can be said for a rhetorical analysis of reading. It is not simply a means of understanding particular ways in which people read or f a i l to; i t is also a means of understanding one of the primary mechanisms of understanding i t s e l f . If we can advance our understanding of the conventions and operations of reading as a rhetorical mode of building belief, we may be able to advance our understanding of the divisions in society that result from diversity of belief. It is always dangerous to make broad claims for the a b i l i t y of a new theoretical perspective to make sweeping changes in the conduct of l i f e . The everyday world has a way of turning as before, apparently ignoring the implications of new theoretical accounts of i t s operations. But i t is nonetheless tempting to speculate that i f we can understand the sources of division in society we may be better able to heal them. In Modern Dogma and the Rhetoric of Assent, Wayne Booth sets such a goal for rhetoric: D. Brent 6 If i t is good for men to attend to each others' reasons—and we a l l know that i t i s , for without such attending none of us could come to be and questions about value could not even be asked—it is also good to work for whatever conditions make such mutual inquiry possible. (137) This is a high goal indeed, but perhaps by increasing our understanding of how beliefs are formed through reading—that i s , of one of the conditions that make mutual inquiry possible-—a rhetoric of reading may be able to help us work to improve those conditions. D. Brent 7 Toward a Rhetoric of Reading Chapter 1: Starting Points 1.1. The Traditional Framework The f i r s t of the four questions posed in the introduction is essentially a question of definition: the question "What aspects of rhetoric, both traditional and modern, can inform a theory of reading?" is another way of asking, "What exactly is a rhetoric of reading, as opposed to any other theory of reading?" To put the question another way, what can rhetorical theory contribute to reading theory that i t does not already have? In order to answer this question, we must f i r s t identify the essential features of rhetoric i t s e l f . Perhaps the most fundamental defining feature of rhetoric i s that articulated by Aristotle when he defines rhetoric as "the faculty of discovering in the particular case what are the available means of persuasion" (7). This definition, however, is helpful only i f we can in turn define persuasion. George Kennedy suggests that the class i c a l concept of "persuasion" carried a considerable range of meaning: The ancient world commonly thought of this purpose [i.e., the purpose of rhetoric] as persuasion, but meant by that something much looser and more inclusive than persuasion as understood by a modern social scientist. Purposes cover a whole spectrum from converting hearers to a view opposed to that they D. Brent 8 previously held, to implanting a conviction not otherwise considered, to the deepening of belief in a view already favorably entertained, to a demonstration of the cleverness of the author, to teaching or exposition. (Classical Rhetoric 4) Aristotle himself gives no formal definition of the term, but the types of speeches that he discusses under the heading of rhetoric indicate the range of acts that he considered examples of persuasion. For Aristotle, "persuasion" could mean inducing an audience to act, as in a speech intended to convince Athenians to undertake a p o l i t i c a l action such as a war. It could equally mean simply inducing the audience to believe something, as in the case of a speech intended to convince an audience that a particular man is honourable or dishonourable, guilty or not guilty. Thus classical rhetoric was not in any way limited to attempts to induce overt action by the audience. This broad use of the term "persuasion" has in some ages been narrowed considerably. By postulating that the understanding and the wi l l were separate parts of the human mind, eighteenth-century faculty psychology tended to separate the faculty of believing from the faculty of acting. Under the influence of this theory, eighteenth-century rhetoricians tended to reserve the word "persuasion" for discourse affecting the w i l l — t h a t i s , persuasion to act. Discourse intended to affect belief was placed under the head of "conviction" rather than "persuasion."1- However, with the breakdown of the s t r i c t divisions of the mind envisioned by faculty psychology, twentieth-century rhetoric has tended to conflate persuading and informing even more than did Aristotle. ^See for instance George Campbell, The Philosophy of Rhetoric, chapter 1, and Richard Whately, Elements of Rhetoric, part I, chapter 1 and part II, chapter 1. D. Brent 9 In Rhetorical Dimensions in Criticism, for instance, Donald C. Bryant defines rhetoric as "the rationale of the informative and suasory in discourse." This definition, according to Bryant, implies two distinguishable but closely entangled dimensions of discourse as rhetorical, and i t implies others which are not. Perhaps i t dodges or circumvents the problems of genre, but I think rather that i t recognizes pure genres as fictions and implies that most artifacts of discourse exhibit various dimensions, the informative-suasory of which comprise the province of rhetoric. (14) By using a compound term such as "informative-suasory," Bryant acknowledges that i t i s impossible to reliably distinguish discourse intended to change belief from discourse intended to prompt action. To persuade a person,to act, you must convince him that the proposed action is good or desirable; to convince him of a proposition, you must persuade him to treat i t as being true. Thus persuasion can be seen at the heart of even the most apparently neutral informative discourse, for such discourse is an attempt to persuade the audience to alter, i f ever so sligh t l y , their beliefs as to the way the world is or ought to be. In short, then, "persuasion" can be seen as a v i t a l defining feature of rhetoric without limiting rhetoric to discourse that seeks to influence overt behaviour. 2 "Persuasive," says Aristotle, "means persuasive to a person" (11). 2Showing that there is an element of persuasion in informative discourse does not, of course, mean that there is never a motive for making a rough distinction between discourse that is overtly persuasive and discourse that i s not. For works that make such a distinction, see for example James Kinneavy, A Theory of Discourse; Robert J. Connors, "The Rhetoric of Explanation"; Katherine E. Rowan, "A Contemporary Theory of Explanatory Writing." D. Brent 10 Defining rhetoric as the art of influencing behaviour and belief implies a faith that discourse is a reasonably reliable means by which one person can affect another. If discourse has no power to connect rhetor and audience, persuasion can have no meaning. But discourse must do more than simply communicate; i t must enable the rhetor to control the responses of the audience in more or less predictable ways. This further entails a faith that human nature is s u f f i c i e n t l y universal that one can predict audience response with some degree of accuracy. This belief in predictability does not necessarily mean that the rhetor exerts complete control over his audience. In Plato's Gorqias, this extreme Sophistic view is satirized in.Polus' claim that rhetors can "act like tyrants and put to death any one they please and confiscate property and banish any one they've a mind to" (27). Plato does, however, acknowledge a connection between the form and content of a discourse and i t s effect, for the competent rhetorician must "discover the kind of speech that matches each type of nature" (Phaedrus 72). Aristotle's more practical rhetoric develops this idea in more detail, showing how one's arguments must be chosen according to whether the audience is young or old, rich or poor, good or e v i l . Today, after twenty-five centuries of largely unsuccessful attempts to develop algorithms to predict behaviour, we might be more cautious about such generalizations. Yet behind even the most cautious attempts to suggest what w i l l and w i l l not affect an audience l i e s a basic assumption that underlies a l l rhetorical precepts: human beings act not at random, but rather for reasons that the rhetor can predict and use. This assumption can never be f u l l y abandoned i f rhetoric is to be true to i t s mandate as an art of persuasion. D. Brent 11 Rhetoric must, however, deal with more than the basic "rhetorical triangle" of discourse, rhetor and audience. Rhetoric involves discovering not just the means of persuasion, but the means of persuasion "in the particular case." This implies that rhetoric must be able to account for the ways in which means of persuasion vary not just with the rhetor and the audience: rhetoric must be able to explain how the means of persuasion vary according to where the speech is being presented, for what reason, and under what circumstances. For Aristotelian rhetoric, this requirement generated a division of rhetoric into three specific kinds: forensic, epideictic, and deliberative. The specific division is no longer appropriate, for modern society provides far more diverse opportunities for rhetorical discourse than did the class i c a l world. However, as Lloyd Bitzer reminds us in "The Rhetorical Situation," rhetoric must s t i l l take account of the occasion in the form of the exigency, which Bitzer defines as "a defect, an obstacle, something waiting to be done, a thing which is other than i t should be" (386). Rhetoric is not just persuasion in the abstract, but persuasion enacted in order to bring about a change in a specific aspect of the surrounding context. At the same time, rhetoric i s also a highly general art. Aristotle characterizes i t as an art that belongs to no one subject area: The better our choice of propositions, the more we imperceptibly glide into some discipline other than Dialectic and Rhetoric; for i f we light upon true s c i e n t i f i c principles, the art is no longer Dialectic or Rhetoric, but is the discipline based upon those principles. (16) This does not mean that rhetoric does not concern i t s e l f with D. Brent 12 specific occasions. With i t s concern to find ways of persuading particular audiences of particular propositions, i t is in one sense an art of specific occasions par excellence. However, we must distinguish between the art of rhetoric i t s e l f and the uses made of i t . The art i t s e l f i s not concerned with l i s t i n g in detail the specific methods of arguing for each occasion in each discipline; such a l i s t i n g would be so immense as to be impossible in practice, and occasions change and develop new characteristics at such a rate that such a program could never be completed. Rather, i t is an art of general principles which individual rhetors can use to analyze specific occasions and to develop appropriate arguments as necessary. Along with this general definition of rhetoric as an art of persuasion and the assumptions thereby entailed, traditional rhetoric supplies some more or less standard terms for discussing the means of persuasion. These terms are not s u f f i c i e n t l y universal to be written into a definition of rhetoric. However, they are so commonly associated with rhetoric that we can reasonably expect them, or at least the ways of seeing that they imply, to influence any art qualified by the term "rhetorical." One of the most fundamental of these sets of terms is the division of the art into the five canons of invention, arrangement, style, memory and delivery. The last two canons have dwindled in importance with the rise of written discourse, but the f i r s t three reflect a fundamental concern with two different aspects of persuasion: persuasion by arguments (invention) and persuasion by form (arrangement and style). These two elements of persuasion have risen and fallen in relative importance throughout the ages. However, the art of rhetoric has always carried D. Brent 13 with i t the implicit assumption that human beings are persuaded to believe what they believe by means of two related agents: the arguments themselves, and the manner in which they are produced and received. Rhetorical descriptions of discourse also tend to divide the means of persuasion into categories, the most famous of which are the Aristotelian modes of proof: logos, pathos, and ethos. The relative emphasis placed on these modes has, like the relative emphasis placed on form and matter, varied over the ages. Some rhetoricians, such as Plato, are concerned purely with the operations of the reason, rejecting pathos in particular as being too closely associated with the dark horse of unreason. Other rhetoricians, particularly after the eighteenth-century reappraisal of "the passions" as v i t a l moving forces of the human mind, favour a more equal division among the modes of proof. 3 But despite differences in opinion concerning the relative merits of different modes of appeal, traditional rhetoricians agree that there are different modes of appeal—that human nature is not unitary and that appeals can be profitably classed according to the aspects of human, nature to which they refer. Finally, rhetoric carries with i t an implied pedagogical commitment. Philosophical rhetorics do not always provide specific precepts that can be taught to neophyte rhetoricians in order that they may be better able to compose discourse. In particular, they do not always show a concern 3 I n chapter 7 of The Philosophy of Rhetoric, for instance, Campbell argues that in order to persuade hearers to act, a rhetor must move the w i l l by appealing to a l l of the four other faculties: reason, memory, imagination, and passions. Whately argues this case in further detail in part I, chapter 1 of Elements of Rhetoric. Wayne Booth is perhaps the most representative modern example of the revalidation of ethos and pathos, with a particular emphasis on ethical proof; see in particular Modern Dogma 154-64. D. Brent 14 for specific methods of teaching. Their concern is with developing the outlines of the subject i t s e l f , not of the methodology of inculcating i t . Aristotle, for instance, provides both a definitional analysis of the subject of rhetoric and a'substantial l i s t of specific rhetorical precepts; however, he says nothing of how these precepts are to be imparted, evidently assuming either that simply informing rhetoricians of these precepts w i l l enable them to put them into practice, or that his task does not extend beyond developing the precepts themselves. Plato is even more remote from the details of rhetorical education, concerning himself with defining the art and probing i t s moral significance rather than with teaching i t as a s k i l l . However, even the most philosophical rhetoric carries with i t the seeds of a practical rhetoric. Defining the art of rhetoric, analyzing the human mind and the ways in which i t can be influenced through words, always presupposes that the practical orator can make use of this information to design discourse that works. Even the indirect instruction favoured by much twentieth-century composition pedagogy is founded on a prior understanding of rhetorical transactions, even though that knowledge is transferred t a c i t l y rather than through overt instruction. 4 What does this sketch of rhetoric imply about a rhetoric of reading? F i r s t , i t implies that a rhetoric of reading must see reading as more than merely a transaction between text and reader. It must see reading as a transaction between writer and audience by means of a text. This in turn implies that there is some connection, however tenuous, between the 4 I n "Rhetorical Traditions and the Teaching of Writing, for instance, C. H. Knoblauch and L i l Brannon argue that rhetorical precepts as such have no place in the composition classroom, yet they also argue that an understanding of rhetorical processes can have considerable value in informing the composition teacher (102). D. Brent 15 meaning the writer intends and the meaning the reader interprets the text as having. Whether or not i t actively focuses on the problem of intentionality, any theory of reading that ca l l s i t s e l f rhetorical must account for interpretation in a way that at least allows for this transfer of meaning. A rhetoric of reading must also account for this transfer of meaning in the context of persuasion. If, as pointed out earlier in this chapter, i t is impossible to separate the informative element of discourse from the suasory, then a l l discourse carries within i t some element of persuasion. Even the most coldly informative written discourse presents not just information but a certain world-view, a complex of beliefs held, or presented as being held, by the author. A description of the digestive organs of a frog i s not a transparent window on r e a l i t y but a description of r e a l i t y as the author believes i t to be— even i f i t is such a basic description of sensory data that there is absolutely no reason to doubt i t s validity. Reading such discourse involves not simply a passive uptake of information, but the act of accepting as true the view of r e a l i t y presented, an acceptance resulting at least in part from the suasory power of the discourse. Thus reading is an active attempt to find in discourse that which one can be persuaded is at least provisionally true, that which contains elements worth adding to one's own worldview. A rhetoric of reading must therefore account not just for the way a reader decodes meanings from texts, but also how she decides what meanings to accept, what meanings to be persuaded by. A rhetoric of reading, like any rhetoric, should also be founded on principles that are not domain-specific. The specific application of those principles may be highly particular; readers in one f i e l d may be D. Brent 16 convinced not just by different arguments but by entirely different kinds of arguments from those that persuade readers in another f i e l d . 5 Yet the general principles that govern the infinite number of specific transactions, the lenses that allow us to read these transactions and account for them in a rational manner, cannot be tied to any one type of transaction. That i s , we should be able to develop a general rhetoric of reading, not just a rhetoric of s c i e n t i f i c reading, a rhetoric of philosophical reading, a rhetoric of historical reading, and so on. Like the specific applications of the art of rhetoric, the specific applications of a rhetoric of reading should be subsequent to the general principles of the art i t s e l f . Finally, we would expect pedagogical implications to be at least latent in a rhetoric of reading. In a more philosophically oriented exploration of reading we would not necessarily expect these implications to be developed into formal precepts—"Do this and you w i l l be a better reader." But we would expect that an analysis of the process of reading, of the ways in which the reader interprets a text and is or is not persuaded by i t , of the ways in which the author's intentions are or are not carried out, can ultimately be translated into ways in which novice readers can be helped to become expert readers. Beyond these general philosophical principles, we can apply to a rhetoric of reading at least some of the traditional rhetorical terminology. Terminology, as Kenneth Burke points out, is not just a reflection of r e a l i t y but a way of seeing, a way of making distinctions sThis is the argument upon which Kinneavy bases his analysis of the different logics of different types of discourse in A Theory of Discourse (63). Stephen Toulmin makes a similar point when he distinguishes f i e l d -dependent and field-independent aspects of argument in The Uses of  Argument (15). D. Brent 17 that would be made differently i f a different terminology were employed (Language as Symbolic Action 46). By pressing the act of reading to see i f i t w i l l yield some of the distinctions suggested by the language of traditional rhetoric, we can highlight those features that are the most sp e c i f i c a l l y rhetorical; that i s , those features that are the most tied to rhetoric's grounding in persuasion. For instance, the traditional distinction between the modes of proof w i l l lead us to investigate the degree to which psychological processes other than pure intellection are involved in deciding what meanings to be persuaded by. How can ethos, the speaker's character, play a significant role in reading a text whose author is not physically present? What place does pathos, the emotional reaction to a text and the propositions presented by i t , have in the building of a belief system based on written sources? Traditional rhetoric thus plays two compatible roles in a rhetoric of reading. F i r s t , i t plays a definitional role. The traditional definition of rhetoric i t s e l f sets an irreducible minimum of features that a rhetoric of reading must have i f i t is to merit i t s label as a branch of rhetorical investigation. Second, i t plays a heuristic role. We need not be bound by the terminology i t supplies, but using that terminology to the greatest extent possible directs our attention to certain distinctions and raises certain questions that might not be asked, or not asked as directly, by a theory of reading that is not e x p l i c i t l y situated in a rhetorical framework. D. Brent 18 1.2. Rhetoric in a Constructionist Age Thus far we have sketched the features of traditional rhetoric that separate a rhetoric of reading from any other theory of how texts are interpreted. We have not yet clearly articulated why such an account is needed. What is i t about the process of producing and consuming discourse that demands that i t be accounted for by an oxymoronic "rhetoric of reading," when the two separate arts have provided satisfactory accounts for twenty-five centuries? The answer to this question is tied to an increasing tendency to see rhetoric not just as a tool for transmitting ideas and persuading others to believe them, but as a communal medium in which thought grows. This change reflects a s h i f t in overall focus: rather than concentrating on the specific task of composing a piece of discourse that w i l l persuade the audience to accept the rhetor's propositions, many modern rhetoricians concentrate on the larger issue of how rhetoric is part of a di a l e c t i c a l process of developing new knowledge by discussion. This s h i f t in focus makes a rhetoric of reading both more possible and more necessary than in the traditional framework, and also suggests additional features of the rhetorical transaction that a rhetoric of reading must account for. Wayne Booth's rhetoric i s a particularly good example of this new focus. In Modern Dogma and the Rhetoric of Assent, Booth argues for a rhetoric that can heal the s p l i t between two "modern dogmas": an objectivist, rationalist view of s c i e n t i f i c "facts" and a t o t a l l y r e l a t i v i s t i c view of values. Booth claims that because of this "fact-value s p l i t , " i t is commonly believed that one can find certainty only in D. Brent 19 the realm of science. Science and logic refer to worlds that are in theory completely knowable and grounded in objective reality, values, on the other hand, are completely ungrounded and completely personal. There is no principled basis for one person to hold to what she believes as opposed to what another believes. Therefore there is no principled basis for sharing values, and collisions between opposed values can result only in apathy, violence, or at the very best an agreement that both parties w i l l go their separate ways. In short, "we have lost our faith in the very po s s i b i l i t y of finding a rational path through any thicket that includes what we c a l l value judgments" (7). This fact-value s p l i t is founded on a s p l i t between an objective world, in which facts are grounded, and a personal, subjective world that gives rise to values. To heal this s p l i t , Booth articulates a different view of knowledge based on a different view of the s e l f . Knowledge is not something discovered by the isolated self in i t s interaction with the physical world. For Booth, knowledge arises through collaboration. Booth begins with a proposition that, he argues, everyone in fact operates under: "It is reasonable to grant (one ought to grant) some degree of credence to whatever qualified men and women agree on" (101). From this starting point he develops a communal view of knowledge in which one knows the world "through a willing assent to the process of making an i n t e l l i g i b l e world with my fellow creatures" (105). Even s c i e n t i f i c knowledge is constructed socially: D. Brent 20 This is in formal structure—as Michael Polanyi among others has shown—the process of validation used even by scientists for a great share of their s c i e n t i f i c beliefs. No scientist has ever performed experiments or calculations providing more than a tiny fraction of a l l the s c i e n t i f i c beliefs he holds; the whole edifice of science depends on faith in witnesses, past and present—on testimony and tradition. . . . Thus science i s , in i t s larger structures, validated by the same social processes that I am arguing for in " a l l the rest." (108-09) Thus both facts and values are grounded in the same social processes, and both can be the subject of knowledge in a principled way. J This view of knowledge i s intimately connected with a view of the sel f . For Booth, the self is not an isolated entity that finds knowledge of facts through individual interaction with nature and knowledge of values through individual probing of i t s e l f . Rather, the self is "a f i e l d of selves": It i s essentially rhetorical, symbol exchanging, a social product in process of changing through interaction, sharing values with other selves. Even when thinking privately, " I " can never escape the other selves which I have taken in to make "myself," and my thought w i l l thus always be a dialogue. (126) This use of the term "rhetoric" for this process of building a world through symbolic interaction represents an extension of i t s meaning. Under Booth's definition, rhetoric is not merely an art of persuasion. It i s "a whole philosophy of how men succeed or f a i l in discovering together, in discourse, new levels of truth (or at least agreement) that D. Brent 21 neither suspected before" (11). It is a way in which such slippery and amorphous things as values may be shared and good reasons found by which men may be persuaded to change their minds, an art not limited to any particular discipline but rather an art that includes "the grounds for warrantable or corrigible assertion in any subject" ( l l n ) . This definition of rhetoric overlaps with but is by no means coterminous with the traditional definition; i f one developed a l i s t of works that deal with the subject thus defined, "only a small part of i t would consist of the thousands of works that s p e c i f i c a l l y mention 'rhetoric' . . . [and] many works ostensibly about rhetoric would not be in the l i s t " ( l l n ) . Yet this reshaped definition of rhetoric as an art of inquiry does not mean that i t i s no longer grounded in persuasion. It simply changes persuasion from an end to a necessary means: The supreme purpose of persuasion in this view could not be to talk someone else into a preconceived view; rather i t must be to engage in mutual inquiry or exploration. In such a world, our rhetorical purpose must always be to perform as well as possible in the same primal symbolic dance which makes us able to dance at a l l . (137) Kenneth Burke also locates persuasion within a context of mutual inquiry and knowledge creation. For Burke, symbols in general, and verbal symbols in particular, are in many ways more important than our senses in building our view of reality. In the opening essay of Language  as Symbolic Action, Burke defines man as "the symbol-using animal": D. Brent 22 Take away our books, and what l i t t l e do we know about history, biography, even something so "down to earth" as the relative position of seas and continents? What is our "reality" for today (beyond the paper-thin line of our own particular lives) but a l l this clutter of symbols about the past combined with whatever things we know mainly through maps, magazines, newspapers, and the like about the present? . . . And however important to us is the tiny sliver of reality each of us has experienced firsthand, the whole overall "picture" is but a construct of our symbol systems. (5) In the essay "Terministic Screens,11 also in the Language as Symbolic  Action volume, Burke expands on the way in which language influences our view of reality. We not only are dependent on symbols for information about events that are physically beyond the reach of "the paper-thin line of our own particular lives"; our symbol systems or terministic screens also organize the way we perceive "immediate" reality: "Even i f any given terminology is a reflection of reality, by its very nature as a terminology i t must be a selection of reality; and to this extent i t must function also as a deflection of reality" (45). A l l our observations, in other words, are profoundly affected by the terminologies in which those observations are made. And because terminologies are communal rather than individual, our reality is made communally: "The human animal, as we know i t , emerges into personality by f i rs t mastering whatever tribal speech happens to be its particular symbolic environment." (53) However, like Booth, Burke sees rhetoric as being epistemic because of, not in spite of, its grounding in persuasion: "The dramatistic view of language, in terms of 'symbolic action,' is exercised about the D. Brent 23 necessarily suasive nature of even the most unemotional s c i e n t i f i c nomenclatures" (45). Thus Burke comes, by a somewhat different route, to a view of rhetoric that is in many important respects similar to Booth's. Rhetoric is epistemic—that i s , i t is part of the process by which we create as well as pass on knowledge.6 Burke makes this point especially clear in The Philosophy of Literary Form> in which he links drama, dialectic, and rhetoric in his powerful image of the "unending conversation": Imagine that you enter a parlor. You come late. When you arrive, others have long preceded you, and they are engaged in a heated discussion, a discussion too heated for them to pause and t e l l you exactly what i t is about. In fact, the discussion had already begun long before any of them got there, so that no one present is qualified to retrace for you a l l the steps that had gone before. You l i s t e n for a while, un t i l you decide that you have caught the tenor of the argument; then you put in your oar. Someone answers; you answer him; another comes to your defense; another aligns himself against you, to either the embarrassment or gratification of your opponent, depending upon sRobert L. Scott also enunciates this view in "On Viewing Rhetoric as Epistemic," and c l a r i f i e s i t in "On Viewing Rhetoric as Epistemic: Ten Years Later." Scott's argument is that r e a l i t y i s not absolute but rather is soc i a l l y constructed through a process of persuasion. "In human a f f a i r s , " he argues, "rhetoric . . . is a way of knowing; i t is epistemic" ("On Viewing" 17). In "How Shall We Say, 'Reality is Socially Constructed through Communication'," C. Jack Orr addresses squarely the obvious problem generated by social r e l a t i v i s t views of knowledge: the implication that r e a l i t y is t o t a l l y founded on consensus and therefore not r e a l l y capable of being argued about. Orr distinguishes between "knowledge," which reflects what we know about reality, and "truth," which is an objective r e a l i t y beyond the s e l f . The assumption that the latter exists provides a standard for the former, and provides grounds for reasoned rhetorical debate about whether one consensual body of knowledge does in fact reflect objective r e a l i t y better than another. D. Brent 24 the quality of your ally's assistance. However, the discussion is interminable. The hour grows late, you must depart. And you do depart, with the discussion s t i l l vigorously in progress. (110-11) Thus for Burke, rhetoric, dialectic, and the drama of human symbolic interaction are inextricably intertwined. Dialectic i s simply a particular category of the drama of human symbolic interaction, a brand of interaction "concerned with the maieutic, or midwifery, of philosophic assertion, the ways in which an idea is developed by the 'cooperative competition' of the 'parliamentary'" (107). The metaphor of dial e c t i c as drama blends easily with a metaphor of dialectic as conversation, for what is drama but conversation placed on a stage for inspection? 7 This metaphor emphasizes the fact that the process of building each others' minds is a cumulative process that occurs over a vast stretch of time. Each individual act of rhetoric is but one element in a larger conversation, and that conversation is part of a vast discourse whose origin theoretically coincides with the development of symbolization i t s e l f . Whatever we collectively and individually know now is the result of an i n f i n i t e l y extended process of sharing and negotiating knowledge, of d i a l e c t i c a l l y testing and improving assertions in a l l domains of thought. To refocus attention on the function of rhetoric in this unending human conversation is to effect a major re-evaluation of the theoretical boundaries of the discipline, and most particularly of the canon of "'Not a l l of those who use the metaphor of human discourse as unending conversation see i t as primarily d i a l e c t i c a l . See for instance Michael Oakeshott (198) and Richard Rorty (318) for the argument that inquiry, the testing of propositions, is only incidental to a conversation whose overriding purpose is simply to converse. D. Brent 25 invention. Invention is traditionally seen as a forward-looking process. The rhetor uses his knowledge of the audience and of the occasion to develop arguments that w i l l be effective when the rhetorical process culminates in the delivery of a speech or the composing of the f i n a l draft of a written discourse. The process in effect funnels out from the single rhetor toward the audience and moves forward in time from the i n i t i a l framing of a discourse to i t s delivery. When rhetoric is situated in an epistemic conversation, however, we can see that i t also involves another movement, from the rhetor back into the vast network of conversation that precedes in time that particular exchange. In other words, a f u l l account of rhetoric must take account of the fact that the rhetor is himself an audience. Before he comes to the point of attempting to create belief in others, he has created belief in himself through interaction with countless other selves. Thus invention becomes not a private but a social a c t , 8 not a matter of the rhetor deciding for himself what arguments w i l l persuade an audience, but a matter of the rhetor developing knowledge in collaboration with countless others. But pointing out that invention is a social act is only a starting point. This expansion of the theoretical horizons of rhetoric requires us to ask a new set of questions about how invention operates. We must consider not only the role of subject and audience, of topoi, of rhetorical scene; we must also ask exactly how i t is that the rhetor taps the resources of the unending conversation of which any particular discourse is only a part. In short, how does the rhetor look back before—or as well as—looking forward? A rhetorical BKaren Burke LeFevre coins this apt phrase in her book Invention as  a Social Act, an illuminating study of the complex role played by various audiences in the development of a writer's ideas. D. Brent 26 system that does not provide principles for answering this question is necessarily incomplete, for the question is entailed by the epistemic perspective to which modern rhetoric has become committed. 1.3. Reading as a Special Form of Invention We have seen how a constructive epistemology expands the notion of invention. But symbolic negotiation of knowledge is not limited to symbols printed on paper. Symbolicity need not even be verbal. As Kenneth Burke points out, "Nonverbal conditions or objects can be considered as signs by reason of persuasive ingredients inherent in the 'meaning' they have for the audience to which they are 'addressed'" (Rhetoric of Motives 161). Anything that can be consciously manipulated by man can be read as a symbolic language in which beliefs are formed and shared. Why then do we need a rhetoric s p e c i f i c a l l y of reading rather than simply an investigation of the social aspects of invention in general, an investigation that would yield general principles that are not tied to the communicative code involved? Moving from one communicative mode to another—from speech to print to architecture to music—changes in non-trivial ways both the problems that must be addressed and the sorts of solutions that can be found. Print i s not just speech written down. It is a medium with i t s own history, i t s own constraints, i t s own psychological peculiarities. In particular, i t is a very d i f f i c u l t medium in which to have a conversation. It has become a truism of works on written rhetoric that the reader has no access to the incidental aids that the hearer of speech takes for granted—nuances of voice and faci a l expression, the a b i l i t y to D. Brent 27 ask for c l a r i f i c a t i o n , the subtle sources of feedback that allow even large-scale public speaking to be more intimate than the relationship between a reader and a book. But these distancing factors are incidental to the major difference between reading and hearing—the reader's il l u s i o n that she is simply absorbing information from a text rather than conversing with, and being persuaded by, another human being. The particular challenge of a rhetoric of reading is to account for the way in which, despite this i l l u s i o n of isolation, the reader is actually doing more than absorbing information from disembodied texts. This entails a study of an intricate series of transformations. How does a set of texts that can be held in the hand, texts which proclaim diverse and often contradictory views of the universe, become transformed into a reasonably consistent set of beliefs in the mind of the reader? To put i t another way, what are the special characteristics of participating in a "conversation" with others who are absent and in many cases long dead? To a certain extent, a rhetoric of reading thus conceived is simply a rhetoric of composition considered from another point of view. If a rhetoric of composition is designed to t e l l a writer how to persuade an audience, i t seems simple enough to turn those precepts, by a mere semantic f l i p , into information for the reader about how writers attempt to persuade her. However, any rhetorical art that goes beyond being a technical manual must be more than a l i s t of means of persuasion. It must be an analysis of the mechanism of persuasion i t s e l f and the decisions that go into performing the act. That mechanism and those decisions look very different when we move around the rhetorical triangle to focus on reader rather than writer. The rhetor's task involves an outward movement; she begins with D. Brent 28 certain basic propositions that she wants to convey, and then discovers arguments that w i l l support them and a form in which they can be convincingly argued. The discourse thus radiates outward from a more or less unitary centre to a more or less diffuse audience, an audience that may consist of hundreds or even (in the case of published writing) millions of individual entities. The rhetor is thus faced with a task of generalization; however well she may know a particular audience, she must be able to use a general knowledge of human nature and of discourse structures in order to predict and to some extent control the responses of that audience. The reader's task is in certain respects the opposite. The audience is not diffuse, but highly particular—herself. She does not have to ask how to frame propositions that w i l l have the best chance of convincing the largest number of hearers. Rather, she must ask the question that Booth uses as his touchstone in Modern Dogma: "When should I change my mind?" (12) This question is synthetic rather than analytic. It does not require the asker to take apart a vast audience and consider what characteristics they may possess as individuals; i t requires her to take a disparate group of claims made by individuals, each with his own perspective on the world and his own reasons for seeing i t as he does, evaluate them, and actively construct a single view satisfactory to herself. Thus while a rhetoric of discourse production can t e l l us some specific reasons why people are persuaded, a rhetoric of discourse consumption must develop an account of how readers sort through the bids made for their assent. D. Brent 29 1.4. The Shape of a Rhetoric of Reading In sum, then, a rhetoric of reading w i l l have the following features. In common with any theory of reading, i t must be able to account for the f i r s t stage of the reading act: the creation of meaning from symbols on paper. However, to qualify as a rhetoric of reading, i t must be able to account for this process in the context of the rhetorical framework. This means that the interaction between reader and text must be seen as being in the service of a larger process: making contact with the mind of another human being. Moreover, a rhetoric of reading must take into account the place of reading in the epistemic conversation of mankind. This means that i t must account not only for the proximate goal of perceiving another person's meanings, but also the ultimate goal of updating a belief system or "worldview," a theory about the way the world operates and about the way in which the believer can and should operate within i t . To use Booth's terms, i t is a synthesis of "facts" and "values," a set of beliefs about what is and what ought to be. This belief system w i l l in turn contribute to the building of further discourse as the reader takes his turn in the conversation. This ultimate goal implies a series of intermediate steps. The reader uses texts to supplement the "paper-thin line" of his own experience, to provide additional windows on what is and what ought to be in the world. However, texts refer not just to the world but to a world-view; not to an unmediated state of existence but to the author's perception of the state of things. There are therefore two steps between the world and the reader's perception of i t : the author's interpretation D. Brent 30 of the world, and the reader's interpretation of the author's text. The process of reading, then, i s not just the interpretation of a text but the interpretation of another person's worldview as presented by a text. The meaning of a text must not only be interpreted, but evaluated for the power of i t s persuasive claims; the reader must decide not only what the text says, but i f and to what degree what i t says is worth believing. As Perelman and Olbrechts-Tyteca note in The New Rhetoric, a rhetor does not simply persuade or f a i l to persuade an audience; rather, "What is characteristic of the adherence of minds is i t s variable intensity" (4). Seen in the context of reading, this v a r i a b i l i t y of intensity implies that a reader may find different elements of a text persuasive to varying degrees. A rhetoric of reading must answer the question of how readers are induced to assign this variable adherence to the various propositions presented by a text. Finally, a text never contributes to a belief system in isolation. It w i l l be considered in conjunction with other texts making some similar and some different claims for belief. In some cases the claims w i l l be incompatible, forcing the reader to decide which texts have the greatest persuasive claim. This is not, however, merely a matter of the reader's picking the text that is the most persuasive and believing i t , as i f a system of belief were an indivisible unit that can simply be imported from a source and made the reader's. Rather, the reader must synthesize belief from the contributions of the various texts with which he is presented, taking more elements of some and less of others. The structure of belief that results from this process w i l l be one in which the elements of the source texts are submerged in new structures that cannot be identified absolutely with any one of the contributing sources. D. Brent 31 The challenge of a rhetoric of reading, in short, is to discover the mechanisms of interpretation, evaluation and synthesis by which each individual creates for himself a structure of beliefs that is unique to him, influenced by but not under the control of the texts on which i t is erected. Because this process of re-creation of belief is the ultimate goal of the epistemic conversation of which reading is a part, accounting for i t is the most important challenge of a rhetoric of reading. 1.5. An Interdisciplinary Approach to a Rhetoric of Reading A rhetorical model of reading cannot be generated independently by either a theory of reading or a theory of rhetoric. A theory of rhetoric alone cannot account for the series of acts by which discourse is consumed; a theory of reading alone cannot account for the fact that reading is embedded in a larger process of rhetorical construction of belief. A rhetoric of reading w i l l combine elements of both disciplines in a new synthesis that accounts for a larger process overarching both the production and the consumption of discourse. The contributions of these two areas of study correspond roughly to the division of the process into the interpretation of texts and the evaluation and synthesis of the theses offered by them—that i s , into the proximate goal of making contact with another mind and the ultimate goal of using that contact to build one's own mind. The process of interpreting sources has traditionally been dealt with by theories of reading rather than by theories of rhetoric. In such theories we would expect to find insights regarding the ways in which readers build a meaning from a text; in particular we would expect to find an explanation D. Brent 32 of the fact that readers interpret texts differently, remembering them differently and reaching different conclusions as to what their authors intended to convey. In short, in theories of reading we would expect to find an answer to the second of the questions posed in the introduction: "How can we best describe within a rhetorical framework how the writer, text, reader and rhetorical situation co-operate to generate meaning? That i s , how does a reader know what propositions the author of a text is trying to persuade her to believe?" On the other hand, evaluation of sources for potential inclusion in a system of beliefs i s a much more sp e c i f i c a l l y rhetorical problem, for i t i s a restatement of the question that overarches the entire concept of rhetorical reading: "When should I change my mind?" Thus i t is in a more direct revision of the rhetorical framework that we w i l l seek the answers to the third and fourth questions implied by a rhetoric of reading: "How does the reader evaluate the propositions presented by individual texts and decide which to be persuaded by?" and "How does the reader negotiate among the claims of various texts in order to develop a unified system of knowledge?" In short, then, the distinction between the two terms that make up the phrase "rhetoric of reading" can shape the search for insights contributed by reading theory and rhetorical theory. However, i t is v i t a l to remember that the two terms are not separate but are two parts of a single concept. Yoking them together has important implications for methodology. In searching for insights through various disciplines that study discourse, we w i l l be looking not just for separate answers to separate questions, but for the materials with which to build a single theory. We must look, then, for what is common, or at least for what is D. Brent 33 compatible, among the theories. This necessity will act as an important f i l t e r . What we take from reading theory will be that which complements and enhances our understanding of reading as a rhetorical process. What we take from rhetorical theory will be that which complements and enhances our understanding of rhetoric as a process whose inventional aspect involves reading. Thus we will not be presenting a comprehensive survey of the disciplines that inform the inquiry; rather we will be sampling from a l l of them to determine what insights they can offer that bear on our central question, a question that is not in fact central to any one of the individual disciplines. Although the first stage of our inquiry will emphasize reading theory and the second rhetorical theory, we will be continually shuttling between the two domains, using one as a gloss on the other. The goal is not a theory of reading followed by a theory of rhetoric—we have those already—but an integrated rhetoric of reading that subsumes both into a single account that, despite our divisions of convenience, describes a single process: rhetorical reading. D. Brent 34 Toward a Rhetoric of Reading Chapter 2: Reading as Construction; Reading as Communication 2.1. Sources for a Theory of Interpretation We have argued that rhetorical reading is a process of updating a unique system of personal beliefs by interpreting, evaluating and synthesising written sources. This chapter w i l l attempt to build an account of the f i r s t stage of that process, interpreting a text. More specifically, i t w i l l seek to answer the question, "How does a reader know what propositions the author of a text is trying to persuade her to believe?" As previously suggested, theories of reading are particularly appropriate sources for such an account. Reading theories can supply the elements of a description of the reading act that w i l l help account for two linked but opposing characteristics of belief systems built from sources: the way in which they are unique to the individual believer, and at the same time connected in principled ways to the sources on which they are erected. We could assume that a l l of the personal elements in a belief system are a result of the reader's individual evaluation and synthesis of the source material; that i s , we could assume that a l l readers reading the same texts see the same things but evaluate them differently and base different conclusions on them. But there is an alternate hypothesis: that the divergence of beliefs begins at the level of interpretation. D. Brent 35 Under this hypothesis/ people reach different conclusions partly because they interpret their sources differently. This would imply that reading is not merely an algorithmic process that a l l competent readers will perform in the same way. Rather, it is a more complex and constructive process that can result in a variety of "correct" readings without one or another reader!s necessarily being in error. If we are building a rhetorical account, this second hypothesis is the more attractive. The domain of rhetoric is not the domain of the exact and the algorithmic, of propositions about which a disagreement is a sign of error. It is the domain of the contingent, of propositions that, in Aristotle's words, "appear to admit of two possibilities," for "on matters which admit of no alternative, which necessarily were, or will be, or are, certainties, no one deliberates" (11-12). The more we can characterize interpretation as a contingent process, the more we can build a model of reading interpenetrated at a l l points by the rhetorical viewpoint, as opposed to one that is rhetorical only in its later stages of evaluation and synthesis. Accordingly, this chapter will examine reading theories for insights into the ways in which reading can be seen as a rhetorical process from the very beginning: that is, for an account of the ways in which we are persuaded that a text presents certain propositions and not others. There are four main variables in the rhetorical situation: the classic rhetorical triangle of the writer, the text and the reader, plus the external context or rhetorical situation in which the act is situated. If divergence of belief is to be explained in part by divergence of interpretation, the reader must be seen as making an active contribution to the reading process; otherwise every competent reader D. Brent 36 would take the same meaning from a given text written by a given writer in a given situation, for there would be no other reason, aside from errors of decoding, for the interpretation to vary. On the other hand, the reader cannot be seen as having complete control over meaning, for that would sever the rhetorical connection between writer, reader and text. A l l four elements must be seen as co-operating to produce an interpretation. Thus arises the more specific question that commands this chapter: "How can we best describe within a rhetorical framework how the writer, text, reader and rhetorical situation co-operate to generate meaning?" Rhetorical theory can not only provide the general framework for this account but also some specific insights, particularly as to the ways in which readers build meaning in response to specific elements of the rhetorical situation. Rhetoric has always been an art that generates accounts, not of "persuasion" in the abstract, but of the possible means of persuading a particular audience of a particular thesis in a particular context. Rhetorical theory, especially accounts of the rhetorical situation such as Lloyd Bitzer's "The Rhetorical Situation" and accounts of rhetorical strategies such as Perelman and Olbrechts-Tyteca's The New Rhetoric f can help us understand the ways in which meaning i s influenced by elements exterior to the reader himself. However, because rhetorical theory was developed to account for the production rather than the consumption of discourse, i t has never been centrally concerned with the question of interpretation. In an account of this aspect of the reading act, more emphasis must be placed on reading theories that inquire dire c t l y into the matter of how readers build meaning. D. Brent 37 There are two main classes of reading theory that can offer an account of this co-operative generation of meaning: discourse processing theories of comprehension and l i t e r a r y theories of audience response. The former, represented by the work of researchers such as Smith, De Beaugrande, Rumelhart and Ortony, and Spiro, span a number of more specific disciplines such as linguistics, education, developmental psycho-social linguistics, cognitive psychology and cognitive science (Anderson 416). x The common ground in these theories i s that they are not centrally theories of language, of code systems, or of texts; rather, they are theories of mental processes, using empirical data to generate models of the ways in which the human mind generates and represents meaning when confronted with symbol systems such as written texts. These theories are sometimes applied to li t e r a r y texts, but often are not concerned with specific types of texts; indeed, they are often not concerned with texts as such at a l l , but with processes that operate at the level of the single sentence. Their chief concern is not textual but cognitive; they seek an understanding primarily of the mind as a language processing instrument, rather than of the language that i s processed. Literary theories, on the other hand, are much more text-specific. They are concerned with the special problems of interpretation posed by lit e r a r y texts, and tend to concentrate more on the properties of li t e r a r y texts and on general problems of interpretation than on producing detailed models of the mind. Each of these classes of theory can contribute different insights to a rhetoric of reading. Literary theory has a long history of dealing -^See De Beaugrande, "Design Criteria for Process Models of Reading," for a survey of ten major theories. D. Brent 38 with multiple interpretations, for i t has had to take account of the fact that l i t e r a r y works typically give rise to interpretations so different from one another that i t sometimes seems as i f readers have read different texts. Attempts to put interpretation on a more " s c i e n t i f i c " footing by defining more exactly the c r i t i c a l task and by attempting to document in detail the effects of various l i t e r a r y devices have only compounded the problem. One has only to review the incredible variety of interpretations produced by members of the New C r i t i c a l movement to realize how impossible i t is to devise a single consistent meaning for any l i t e r a r y work.2 As a result, l i t e r a r y theory has tended to be highly tolerant of multiple interpretation and to seek explanations for i t instead of trying to eliminate i t . In l i t e r a r y theory, then, we w i l l find one important ingredient of a rhetoric of reading: highly developed accounts of how meaning varies from reader to reader, and of how meaning is constrained so that multiple interpretation can be validated without assuming that any meaning is as valid as any other. Of the wide range of l i t e r a r y theories available, the category that bears most closely on this inquiry is "reader-response" or "audience-oriented" theories such as those of Stanley Fish, Louise Rosenblatt, and Wolfgang Iser, c r i t i c s who are particularly concerned with the ways in which the reader constructs meaning. Fish, Rosenblatt and Iser are not, of course, the only c r i t i c s who explore the act of constructing meaning from texts. Structuralist theorists such as Tzvetan Todorov and Umberto Eco also ask how the reader actively interprets a text to create from i t 2See for instance the contradictory interpretations of Wordsworth's "A Slumber Did My S p i r i t Seal" by New C r i t i c s Brooks and Bateson, cited in Rosenblatt 116. • L D. Brent 39 a meaning.3 However, the structuralist orientation focuses more on the features of the text as a coding system rather than on the constructive processes of readers as recipients of a rhetorical transaction. Poststructuralism, with i t s emphasis not on the reader's response to a text but on the text i t s e l f as a self-deconstructing system, 4 is even further from the domain of this study. The c r i t i c s dealt with here are thus put forward not as the only theorists who espouse a constructive view of the reading act, but as representatives of a branch of theory that is particularly useful here because of i t s emphasis on what the reader brings to the transaction. To explain the need to refer as well to discourse processing theory, we must digress for a moment to explore some of the limitations of l i t e r a r y theory. Literary theory is by definition concerned with l i t e r a r y texts. Such texts can be seen as rhetorical in that they are a means by which the writer contributes to the building of the reader's value system. Wayne Booth's A Rhetoric of Fiction, in fact, i s entirely devoted to arguing for and exploring the implications of this thesis. Literary texts, however, form only a portion of the sources contributing to any given person's value system, and in many domains of inquiry, they may play no overt part at a l l . If a rhetoric of reading i s to be a general theory enunciating principles that can explain acts of reading in various domains, we must be able to say with confidence that those principles are not limited to the reading of li t e r a r y texts. This means that we must consider whether there is any crucial difference between 3See Todorov, "Reading as Construction," and Eco, The Role of the  Reader, especially the introductory essay. 4See for instance Culler 43. D. Brent 40 l i t e r a r y and nonliterary reading, for only i f the answer to this question is "no" can we confidently generalize a rhetoric of reading from l i t e r a r y theory alone. It may be argued that there is no way to distinguish meaningfully between l i t e r a r y and nonliterary texts. In The Reader, the Text, the  Poem, for instance, Louise Rosenblatt claims that the crucial distinction between literature and nonliterature is not between types of texts but between li t e r a r y and nonliterary, or as she c a l l s them, aesthetic and efferent, reading acts. Efferent reading is reading "in which the primary concern of the reader is with what he w i l l carry away from the reading"; aesthetic reading is reading in which "the reader's attention is centered directly on what he is l i v i n g through during his relationship with that particular text" (24-25). The difference between literature and nonliterature is therefore centred in the reader, in a "shift of attention" that divides l i t e r a r y from nonliterary reading. This focus on reading acts rather than on types of texts, however, makes the generalization we are after more rather than less problematic. A rhetoric of reading is not a theory of texts but a theory of reading acts. The a c t i v i t y of the rhetorical reader is an a c t i v i t y in which the efferent mode overwhelmingly predominates, for he is reading not for the aesthetic experience of l i v i n g through a relationship with a text but for the efferent purpose of updating his system of beliefs. Thus we have no basis for automatically generalizing from a description of l i t e r a r y reading to a description of rhetorical reading. Moreover, i t is not s t r i c t l y true that there are no meaningful differences between l i t e r a r y and nonliterary texts. Although in theory any piece of discourse can be read in either the efferent or the D. Brent 41 aesthetic mode, in practice the vast majority of aesthetic reading acts involve texts which are not representations of the world in the same way as texts that trigger efferent reading acts. The efferent reading act may be seen as a chain of representations: the author represents the world to herself in a particular way, unique to herself because of her personal viewpoint. She then creates a text that conveys certain aspects of this representation. Her reader in turn infers this representation from the text and decides whether or not to use the data i t provides to add to his own representation of the way the world is or ought to be. Though each link in the chain involves complex inferential processes rather than a direct transfer of information, the ideal toward which the process aspires is an unobscured handing-on of representations of selected aspects of the world, representations that are intended to be taken as being l i t e r a l l y "true" in detail. The aesthetic reading act, on the other hand, is typically triggered by certain signals—textual clues, l i t e r a r y conventions, genre, t i t l e or author—that indicate to the reader that the events represented are not to be taken as l i t e r a l l y true. At i t s simplest, this means that even i f there were once a man named Hamlet who did roughly the things described by Shakespeare, the conventions of drama t e l l us that the playwright did not intend to portray the l i t e r a l events of his l i f e as being true in deta i l . If some of the details of the play happen to match the details of Hamlet's l i f e , this fact i s irrelevant to our enjoyment of the play in ways that i t would not be i f we were reading a biography or news report. As Wolfgang Iser, for instance, points out in considerable detail, the "object" described by a l i t e r a r y text i s , in the typical case, D. Brent 42 imaginary. s The purpose of the play Hamlet is to convey certain aspects of the playwright's view of the world by creating an imaginary object that symbolically conveys certain of his values and perceptions, not to represent his view of r e a l i t y in i t s particular details. It therefore makes sense to speak of li t e r a r y and non-literary texts as well as of aesthetic and non-aesthetic reading acts, for in practice the type of act is closely linked to (although not absolutely controlled by) the type of text and the way i t attempts to represent the world. Because of these differences between the aesthetic and efferent reading acts and the texts that trigger them, there i s room for doubt as to the generalizability of the aesthetic reading act. Therefore we cannot confidently erect a general rhetoric of reading on l i t e r a r y theory alone. We need to turn for corroboration and extension to a body of theory that has. a history of dealing with efferent reading—that i s , with the act of reading texts for which there is no assumption of f i c t i v i t y . This i s the sort of reading act that discourse processing theories attempt to account for. Rand J. Spiro characterizes the goal of reading as seen from the discourse processing point of view: When you are reading the latest installment of Newsweek about the energy c r i s i s , i f you have been following i t in the past you w i l l probably not endeavor to form a complete insular representation of the a r t i c l e as your goal of understanding. Rather, your goal w i l l probably be to integrate what you are reading with what you already know of the subject, with special 5See especially The Act of Reading 54-62, in which Iser refers to speech-act theory to distinguish between f i c t i v e and non-fictive language. Todorov makes a similar distinction, though he elaborates i t differently; see "Reading as Construction" 67-68. D. Brent 43 attention to information that i s new. That i s , your goal of reading i s to update your knowledge. ("Constructive Processes" 270) This process of seeking out the new information in texts and amalgamating i t with a pre-existing mental structure i s precisely the sort of reading act that a rhetoric of reading i s designed to account for. Discourse processing theories of comprehension, then, can provide the corroboration we need. Where l i t e r a r y theory offers insights into the mechanisms of generating and constraining meaning, and discourse processing theory offers parallel insights based on studies of the ways in which subjects interpret nonliterary texts in an efferent mode, we can assume that the insights offered are su f f i c i e n t l y general that they can be included in a theory of reading that is not confined to li t e r a r y texts. When in addition those insights help answer the overarching question of how readers are persuaded to form beliefs, we can use them to build a rhetorical theory of reading. Accordingly, the methodology of this chapter w i l l not be to provide exhaustive surveys of either l i t e r a r y or discourse processing theory. Rather, i t w i l l be to look for areas in which the three relevant bodies of theory—literary theory, discourse processing theory, and rhetoric—are in, or can be brought into, agreement. 2.2. Reading as a Constructive Act For the purposes of this inquiry, one of the most important points of agreement between li t e r a r y theory and discourse processing theory i s that our i n i t i a l hypothesis, that meaning does not reside exclusively in D. Brent 44 the text, is essentially correct. Both bodies of theory see the reader as actively constructing the meaning of a text rather than passively receiving i t . 6 In Is There a Text In This Class? Stanley Fish argues that reading is not so much a matter of discerning what is there as i t is a matter of "knowing how to produce what can thereafter be said to be there. Interpretation is not the art of construing but the art of constructing. Interpreters do not decode poems; they make them" (327). He illustrates this point with an anecdote from one of his classes. Given a reading assignment consisting of a l i s t of names of c r i t i c s , and told i t was a seventeenth-century poem, Fish's students were able, he alleges, to produce a meaning commensurate with their understanding of seventeenth-century poetry, complete with metaphysical conceits and rich religious significance. He rejects the possible argument that the "poem" reading was simply an interpretation forced on a text that contained the "natural" meaning of an assignment: An assignment no more compels i t s own recognition than does a poem; rather, as in the case of a poem, the shape of an assignment emerges when someone looks at something identified as one with assignment-seeing eyes, that i s , with eyes which are capable of seeing the words as already embedded within the institutional structure that makes i t possible for assignments to have a sense. The a b i l i t y to see, and therefore to make, an 6This argument is in fact so deeply entrenched in both l i t e r a r y and discourse processing theory that there is really no need to argue that reading is not passive. Rather, this survey of constructionist approaches to reading is intended to provide a sense of the extent to which various theorists in s i s t that reading is constructive, and an overview of the various kinds of constructive processes that have been proposed. D. Brent 45 assignment is no less a learned a b i l i t y than the a b i l i t y to see, and therefore to make, a poem. Both are constructed artifacts, the products and not the producers of interpretation. (330-31) This particular example has been described as strained and some of Fish's conclusions c r i t i c i z e d as excessive. 7 But the basic argument that the reader is an active creator rather than passive recipient of meaning finds wide support among reader-response c r i t i c s . For Louise Rosenblatt, for instance, the text is only one of the elements of the reading process. The other i s the reader, who actively organizes out of the inert materials of the text a personally constructed, personally experienced "poem" that she describes as an "event in time": It is not an object or an ideal entity. It happens during a coming-together, a compenetration, of a reader and a text. The reader brings to the text his past experience and present personality. Under the magnetism of the ordered symbols of the text, he marshals his resources and crystallizes out from the stuff of memory, thought, and feeling a new order, a new experience, which he sees as the poem. (12) Thus the "poem," as opposed to the black marks on paper that constitute the "text," i s variable. It is an event created by "a specific reader and a specific text at a specific time and place: change any of these, and there occurs a different c i r c u i t , a different event—a different poem" (14). Her term for the creation of this event i s "evoking" the poem. 7See the chapter of Robert Scholes' Textual Power entitled "Is There a Fish In This Text?" for a particularly trenchant critique of Fish's theories, including this anecdotal proof of his constructive hypothesis. D. Brent 46 To describe the interaction of text and reader, Rosenblatt borrows from Dewey and Bentley the term "transaction," which she describes as "an ongoing process in which the elements or factors are, one might say, aspects of a total situation, each conditioned by and conditioning the other" (17). She finds this term particularly useful because unlike other terms such as "interpretation," which suggests that the reader acts on the text, or "response," which suggests that the text acts on the reader, "transaction" implies no directionality. Rather, i t suggests a more or less equal contribution by both parties. It is this suggestion of balanced contribution, a balance that is essential to the rhetorical view of language (see chapter 1, 10), that makes Rosenblatt's concept of the "transaction" especially useful here. This concept of an abstract entity produced by the interaction of text and reader is repeated in the work of other reader-response c r i t i c s . In The Act of Reading. Wolfgang Iser works from a similar conviction that the work of art does not have a single "meaning" that can be extracted from i t . For Iser, texts " i n i t i a t e 'performances' of meaning rather than actually formulating meanings themselves" (27). This "performance" of meaning, like Rosenblatt's "poem," is not identical either with the text or with the reader: [It] must be situated somewhere between the two. It must inevitably be virtual in character, as i t cannot be reduced to the r e a l i t y of the text or to the subjectivity of the reader, and i t is from this v i r t u a l i t y that i t derives i t s dynamism. (21) Iser's term "virtual work" is particularly apt for our inquiry here because i t captures the abstract, transient quality of the meaning that D. Brent 47 the reader constructs, but unlike Rosenblatt's term "poem," does not suggest a necessarily l i t e r a r y process. Fish, Rosenblatt and Iser d i f f e r greatly in the ways in which they elaborate their theories. However, they are essentially in agreement on an important theoretical principle: that in a theory of reading the focus of inquiry must be not upon the text i t s e l f as a st a t i c object, but upon a much more dynamic and more important entity—the virtual work that is constructed in the reader's mind. The text is only one, and not necessarily the most important, stimulus for the construction of this virtual work. It remains to be asked whether this constructive account of reading can be extended beyond l i t e r a r y criticism into other sorts of reading. Until f a i r l y recently, discourse processing theories have been dominated by the assumption described by Ernest T. Goetz and Bonnie B. Armbruster as "the implicit assumption that text structure and content are inherent in the text" (202). However, this position is rapidly shifting, and many recent studies are taking a constructive approach similar in many respects to that taken by reader-response criticism. It is these studies that w i l l provide the insights most useful in a rhetoric of reading. Frank Smith's 1971 work Understanding Reading is one of the earlier attempts to take a more constructive view of reading. Smith argues against the prevailing assumption that readers f i r s t decode a text into sounds on a character-by-character basis, assemble those sounds into words, and then comprehend text as linear sequence of words thus decoded. The physical limitations of eye and brain, he argues, simply do not permit the e f f i c i e n t processing of the incredible amount of raw information that would be required i f reading proceeded in such a step-D. Brent 48 by-step fashion. Rather, Smith argues, the reader relies heavily on nonvisual information such as his knowledge of language and of the world to predict what the next few letters, words, or phrases w i l l probably be. He checks these predictions not by attending to a l l the visual features of the text but merely by sampling i t to confirm or deconfirm the meaning that he has predicted. The far greater efficiency of this model accounts for the fact that practiced readers can comprehend far more text in a given space of time than their a b i l i t y to process discrete units of data would seem to allow. Smith asks different questions than do the reader-response c r i t i c s : not "Why do people understand Hamlet in so many different ways?" but "How do people know that the last word in 'He put the book on the table' i s 'table' and not 'anteater' or 'green'"? In other words, he is more interested in l i t e r a l comprehension than in the construction of a virtual work that embodies the overall significance that a text has for a particular reader. Yet he is in fundamental support of the most basic assertion of Fish, Rosenblatt and Iser: that meaning depends not just on the text but on a transaction between text and reader. It is not just the shape of the letters but the shape of the letters acting in concert with the reader's prior knowledge of language that t e l l s him that the last word in the above sentence must be a noun and not a verb; i t is the shape of the letters acting in concert with his prior knowledge of the world that t e l l s him that i t is not l i k e l y to be "anteater." Thus Smith corroborates the reader-response view that meaning is constructed from a transaction between the stimulus of the text and what the reader brings to the text—that reading is not simply a matter of translating a code but a matter of "bringing meaning to print" (2). D. Brent 49 An even more constructive approach to meaning has been driven by the need to account for the comprehension of entire texts, especially ambiguous texts that raise some of the interpretive problems that l i t e r a r y criticism has long wrestled with. In a seminal study entitled "Frameworks for Comprehending Discourse," Richard C. Anderson et a l . report a study using the following text: Every Saturday night, four good friends got together. When Jerry, Mike, and Pat arrived, Karen was s i t t i n g in her l i v i n g room writing some notes. She quickly gathered the cards and stood up to greet her friends at the door. They followed her into the l i v i n g room but as usual they couldn't agree on exactly what to play. Jerry eventually took a stand and set things up. Finally, they began to play. Karen's recorder f i l l e d the room with soft and pleasant music. Early in the evening, Mike noticed Pat's hand and the many diamonds. As the night progressed the tempo of play increased. Finally, a l u l l in the ac t i v i t i e s occurred. Taking advantage of this, Jerry pondered the arrangement in front of him. Mike interrupted Jerry's reverie and said, "Let's hear the score." They listened carefully and commented on their performance. When the comments were a l l heard, exhausted but happy, Karen's friends went home. (372) Confronted with this text, subjects typically do not simply report one or other of the two most obvious interpretations: a night of playing cards or playing music. Rather, they report a wide variety of interpretations, none of which is in obvious conflict with the text. Interpreting this study in a later paper, Robert F. Carey and Jerome C. D. Brent 50 Harste conclude that "reading comprehension is neither exclusively nor predominantly text dependent" (199). They describe the reading process using a transactional terminology e x p l i c i t l y borrowed from Rosenblatt: These findings were consistent with the notion of the reading process as semantic transaction. This conceptualization provides for reading as a dynamic and v i t a l process that denies the supremacy of the author while anticipating shades of subtlety among the interpretations made by readers. In this view, comprehension is much less monolithic than conventional wisdom might suggest. Thus . . . each reading act constitutes a unique transactional event. (198) Rand Spiro goes even further in giving the reader a role in creating meaning. Arguing from similar ambiguous passages, Spiro concludes that language does not create meaning: What language provides is a skeleton, a blueprint for the creation of meaning. Such skeletal representations must then be enriched and embellished so that they conform with the understander*s preexisting world views and the operative purposes of understanding at a given time. ("Constructive Processes" 245) Discourse processing theories of comprehension, then, offer a view of reading that corroborates for efferent texts the transactional viewpoint developed by li t e r a r y theory. Our general hypothesis i s therefore supported. We do not have to explain divergence in belief systems solely in terms of differing evaluations of sources; we can explain at least some of this divergence in terms of different ways of constructing the meaning of the texts themselves. When readers come to D. Brent 51 such different conclusions about what they read that they seem to have read different texts, there i s a sense in which this i s true: they have looked at the same characters on paper, but have seen quite different meanings. To combine terms from Rosenblatt and Iser, they have read the same texts but evoked different virtual works. 2.3 Materials of Construction: The Repertoire It is not enough, however, to conclude that meaning is constructed. For a rhetoric of reading to be useful, i t must be able to explain how writer, text, reader and situation co-operate to produce meaning. This question can be„further subdivided. F i r s t , i f the reader contributes to the evocation of the virtual work, what specific materials does he bring to i t s construction and how are they organized? Second, how can we account for the fact that not only do different readers evoke different works, but the same reader w i l l evoke a different work on a different occasion? Finally, having accounted for sources of variant interpretation, can we find sources of s t a b i l i t y and predictability that w i l l allow this process to be characterized as rhetorical—that i s , as a means by which a rhetor can influence an audience with at least some degree of assurance that his intentions w i l l be realized? In answer to the f i r s t question, Rosenblatt claims that although the general meanings of words are reasonably stable and common to a l l speakers of the language^ these shared "dictionary" meanings form only the basic substratum of meaning on which the reader erects his more personal construction. The reader w i l l bring to the words of the text a wealth of personal associations drawn from an interplay of shared social D. Brent 52 and uniquely personal experience. He w i l l have learned the denotational meaning of a word "in specific life-situations and in various specific verbal contexts, spoken and written" (53). These contexts, both verbal and physical, endow each word with a set of associations that cannot be exactly the same as those of any other reader. It is these responses to the author's words, not the dead words on the page themselves, that provide the materials out of which the reader weaves the virtual work, his own personal and transient image of the text and the meanings that i t presents. Stanley Fish, while never denying that the reader's response to a text i s profoundly affected by the personal associations he attaches to words, strikes a very different balance between the personal and the social. The most original and most important observation made in Is  There a Text In This Class? i s that interpretations are made, not just by the reader as an individual, nor even by the reader in co-operation with the text and the writer, but by the reader.in co-operation with the institutions in which she is embedded. The meanings we make from any symbol system, whether written prose, poetry, gesture, abstract art, or computer programming language, are regulated by communally constructed and communally negotiated rules of interpretation. The rules that Fish's students followed in building a seventeenth-century poem from a l i s t of names were those created collectively by seventeenth-century poets and their audiences, re-interpreted through two centuries, handed on through university English classes, and f i n a l l y activated by Fish's assertion that the contents of his blackboard was a poem. The alternate construction of those characters, as a reading assignment rather than a poem, was guided by an alternate set of rules, less formally passed on D. Brent 53 but equally formed by communal expectations, that govern what an assignment i s , what a l i s t i s , et cetera. These sorts of rules, by which any set of symbols is given meaning and without which no set of symbols can have meaning, define various overlapping communities that are constituted by interpretive rules shared by those who belong to them. These associations of individuals sharing interpretive standards Fish designates "interpretive communities." Fish does not suggest that these rules are fixed. On the contrary, he claims that they are in continual flux, and that one of the major functions of the c r i t i c a l enterprise is to renegotiate the rules of interpretation. The interpretive communities set limits on interpretation, but there are no theoretical limits on the possible shape that interpretive communities can take. Interpretive strategies that might strike us as nonsense (such his own example of Blake's "The Tyger" interpreted as an allegory of the digestive tract) strike us as nonsense only because we do not belong to an interpretive community that can legitimize such an interpretation. Such a community could grow up at any time, impelled either by external or internal forces: i t may be discovered that Blake had a preoccupation with the digestive system, or i t may happen that interpreters begin moving spontaneously in the direction of such a set of interpretive conventions in s u f f i c i e n t l y large numbers that they can begin to constitute their own interpretive community and perhaps persuade others to join. There are a number of d i f f i c u l t i e s with Fish's approach to meaning. Fish does not provide a very satisfactory explanation, for instance, of how the members of a given interpretive community could begin to persuade others to join them, given that their interpretations and their entire D. Brent 54 mode of discourse would, by his own definition, be unintelligible to those outside the community. Nonetheless, the concept of interpretive communities is a useful one. It explains not only differences in interpretation but also simil a r i t i e s , accounting for the ways in which interpretations, though always .differing from one interpreter to the next, tend to cluster in groups and even form formal schools of interpretation. Most importantly, i t explains the influence of communal warrants of interpretation not just in terms of shared denotative meanings of words but also in terms of shared larger structures. Such structures allow social forces greater control over the way interpretations are made. Iser provides yet another perspective on the ways in which the materials of interpretation are organized, a perspective that represents yet another way of mixing the private and the public aspects of interpretation. Although Iser argues that the meaning of a text is evoked by the reader, this " i s not the same as saying that comprehension is arbitrary, for the mixture of determinacy and indeterminacy conditions the interaction between text and reader" (24). One of the most important sources of determinacy in this mixture, the prime ingredient in the instructions for the building of the situation, is what Iser c a l l s the "repertoire": "The repertoire consists of a l l the familiar territory within the text. This may be in the form of references to earlier works, or to social and historical norms, or to the whole culture from which the text has emerged" (69). This repertoire affects the evocation of the virtual work by forming "schemata," that i s , pre-existing patterns which condition the way meaning is formed out of the individual experience of the reader: D. Brent 55 The text mobilizes the subjective knowledge present in a l l kinds of readers and directs i t to one particular end. However varied this knowledge may be, the reader's subjective contribution is controlled by the given framework. It is as i f the schema were a hollow form into which the reader is invited to pour his own store of knowledge. (143) The chief difference between Iser's repertoire and Fish's interpretive communities is that the repertoire is associated with the text rather than with a community of readers. The repertoire, however, is in a sense no less a possession of the reader than are Fish's interpretive rules; the "familiar territory within the text" must be familiar to somebody—the reader—and i t must be familiar because the reader is part of a community that understands the social norms and contemporary and l i t e r a r y allusions inscribed in the work. Like the virtual work i t s e l f , these schemata can be seen as hovering between the text and the reader, connecting them by means of the familiar. The repertoire is in some ways more general and less flexible than Fish's interpretive structures; i t is a structure of general knowledge that informs the interpretive process rather than a set of specific rules as to how interpretation is to proceed. It is therefore a much larger and more generally available set of materials, and much less prone to variation. Knowledge can be added to and modified, but not rewritten in the way that a set of rules can. S t i l l , the repertoire functions in much the same way as an interpretive community. By providing a hollow form into which the reader's personal store of knowledge is poured, the schema formed by the repertoire acts as a structure of constraints, giving public form to the reader's private associations and accounting for the D. Brent 56 influence of social forces on the individual interpretive act. This representative sample of l i t e r a r y approaches to construction provides a l i s t of the various factors that the reader brings to bear on the text during the act of evoking from i t the virtual work. These factors range on a continuum from relatively public and stable to relatively private, personal, and unpredictable. At the most public end are the denotational meanings of the words themselves, the common property of a l l speakers of the language. Next come the structures of general knowledge that make up the readers' repertoire. Part of this is lit e r a r y knowledge such as the understanding of what a poem i s , what turns of plot could legitimately be expected to come next, the way text is formatted to represent dialogue, together with a knowledge of li t e r a r y allusions. Another part is world knowledge, a fund of information necessary for understanding the basic dramatic situation, the social structures that f i l l in the reasons for characters' actions, the information that helps the reader label one possible interpretation as l i k e l y and another as nonsense.8 Closely related—perhaps the same thing seen from another angle and with another emphasis—are the interpretive structures that t e l l the reader how far he can push allegorical interpretations, what sort of symbolic patterns he can expect to see, whether or not he is allowed to commit intentional and affective f a l l a c i e s . These public conventions can a l l be summarized under the heading of "expectations." By enabling the reader to predict what the text ought to or might legitimately be able to mean, they narrow the po s s i b i l i t i e s of eRichard Beach and Robert Brown deal with this matter in some detail in their a r t i c l e "Discourse Conventions and Literary Inference: Toward a Theoretical Model." D. Brent 57 meaning and provide shape for the fi n a l level of materials: the purely private and subjective associations that are personal to the writer and provide the most basic materials of response. 9 The balance between these various materials is seen differently by different critics—Rosenblatt emphasizes the reader's private associations, Fish the readers' shared interpretive structures, Iser the reader's knowledge of l i t e r a r y conventions and cultural norms—but there is general agreement on the fact that what the reader brings to a text is a mixture of public and private, shared and unsharable materials. This entire set of elements brought to the reading act by the reader may be called, in an expansion of Iser's terminology, the reader's repertoire. We have established that we cannot automatically generalize from li t e r a r y theory to a more general rhetoric of reading (see 39-43). We must therefore ask whether these perspectives on the materials of construction find any corroboration from discourse processing theories before we can enter them into our model of efferent reading. In fact, the insights of lit e r a r y criticism are very closely paralleled by discourse processing models of comprehension. The key factor in comprehension, according to such studies, is knowledge. At the grossest level, the reader's knowledge enables him to make inferences about what the text could mean in order to f i l l in comprehensional gaps. Walter Kintsch cites the following text as example: "John owns a lovely Kashan. The weft is made of blue s i l k " (11). This text might be f a i r l y easy to comprehend i f you already knew that a Kashan is a type of Persian rug. 9This is essentially the same strategy of prediction and confirmation that Smith identifies as operating in a much more fine-grained way in a l l comprehension from the level of discourse meaning down to letter recognition. See especially Smith 168-75. D. Brent 58 However, even i f you did not, you would be able to infer that a Kashan must be made out of some sort of text i l e ; your world knowledge, knowledge that might include information about wefts and almost certainly about s i l k , would not lead you to suppose that i t is something to eat. Moreover, this knowledge interacts with another sort of knowledge, text knowledge. It is not just your knowledge of wefts and s i l k but also your knowledge of the function of the definite a r t i c l e and of general principles of coherence that would lead you to connect the "weft" in sentence 2 with the rug in sentence 1. This rather general perspective on "knowledge," however, is not the only contribution of discourse processing theory. This discipline also provides a powerful model of how this knowledge is organized. Iser's term "schema," adopted from gestalt psychology by way of Gombrich's theory of a r t , x o is also widely used in discourse processing theory to designate a very similar concept. 1 1 Schemata are structures that 1 0See The Act of Reading 90-92. Iser borrows from Gombrich the idea that the a r t i s t must use schemata to organize the chaotic materials of perception and so reduce the contingency of the world into a coherent set of expectations. 1 1The concept of "schemata" is one of the most commonly accepted models of knowledge representation. Others exist, including Minsky's "frames," which emphasise the sta t i c and stereotyped nature of data, and Schank's "scripts," which emphasise the way knowledge tends to be stored as stereotyped sequences of events. In Discourse Analysis, G i l l i a n Brown and George Yule provide a useful summary of these concepts (238-50), and point out that they are closely related; they represent not so much competing theories as alternative metaphors for the way knowledge is organized (238). There are other approaches to comprehension; an example is that of Teun A. van Dijk and Walter Kintsch, who emphasise the decomposition of text into atomistic propositions (Strategies of  Discourse Comprehension). However, the knowledge-structural approach is adopted here because i t focusses dire c t l y on the question of how readers use their repertoire of prior knowledge to comprehend discourse. This is a particularly important question for a rhetoric of reading because the rhetorical approach sees the text as a channel between two human beings in. possession of different representations of the world. It is therefore D. Brent 59 organize the reader's knowledge so as to make i t accessible when needed. In his attempt to conceptualize the schema more closely, Asghar Iran-Nejad admits that "the concept of the schema remains, theoretically, disturbingly vague" (110), and there is considerable disagreement among researchers as to exactly what schemata are, how they function, or even whether they are best described as structures or processes. But there is general agreement as to what schemata do and what phenomena they are able to explain. As David E. Rumelhart explains in his overview of the concept, schemata are used not just to account for linguistic input but to explain how the mind functions in "interpreting sensory data (both li n g u i s t i c and nonlinguistic), in retrieving information from memory, in organizing actions, in determining goals and subgoals, in allocating resources, and, generally, in guiding the flow of processing in the system" (34); in short, in managing knowledge. Rumelhart describes a schema as a means of packaging knowledge in memory so that i t is not just raw data but an organized set of information, not a specific memory of experience but a generic concept that is derived from the set of a l l the knower's experience. "It is useful," explains Rumelhart, "to think of a schema as a kind of informal, private, unarticulated theory about the nature of the events, objects, or situations that we face. The total set of schemata we have available for interpreting our world in a sense constitutes our private theory of the nature of r e a l i t y " (37). The chief purpose of these sets of information is to help assimilate and interpret new knowledge. It is by f i t t i n g a text to known schemata important to know how those representations affect the way people understand others' texts. D. Brent 60 that a reader is able to understand i t . Deliberately ambiguated texts such as Anderson's card game/music night text highlight the way interpretation depends on finding a suitable schema. But this is only a matter of heightening for observation a process that takes place unconsciously whenever sensory or linguistic data are being interpreted. A l l discourse is ambiguous, and as the questions asked about a text become more complex—as we move from trying to decide between a music night and a game of cards to trying to decide whether Kant thinks that a l l experience is imaginary or not—the ambiguity of language rises dramatically. The schema provides a way of resolving that ambiguity and constructing a meaningful work. 1 2 In the concept of the schema, then, we again have confirmation that the general outlines of the reader-response view of interpretation can be applied to efferent as well as to aesthetic reading. Literary theory and discourse processing theory agree that what the reader brings to the work, what makes interpretation possible, i s the familiar: the conventions of discourse, the world knowledge, the linguistic knowledge, the personal associations, that allow him to evoke a meaning from l i f e l e s s text. Thus we have answered one part of the question posed at the beginning of the previous section: we have outlined the types of materials that the reader brings to the process of being persuaded to interpret a text as presenting a particular meaning. While the text in the hands of this or that reader w i l l always contain the same characters in the same order—the same "visual information," to use Smith's terminology—each reader w i l l bring to i t a different repertoire of "See also Rumelhart and Ortony, and Anderson for detailed discussions of the structure and function of schemata. D. Brent 61 nonvisual information, that i s , a different set of interpretive conventions, linguistic and world knowledge, and personal associations. We have thereby accounted at least in part for variations in interpretation from one reader to another: each rhetorical transaction with a different reader w i l l involve a different repertoire. 2.4. Materials of Construction: The Rhetorical Situation We have now suggested why interpretations vary from reader to reader. We have not, however, f u l l y explained why they can vary from transaction to transaction, even when successive transactions involve the same reader. Such variations occur not only in l i t e r a r y reading, in which one may completely reverse one's appraisal of a work upon rereading; they also happen in efferent reading. One may, for instance, come back to a book or a r t i c l e read months or even days before and see quite different things in i t . Complete reappraisals of a t e x t — evocations of different virtual works—can even occur between one act of thinking about a text and another, without a physical rereading. If we are to explain how readers interpret texts in the course of building their system of beliefs, we must be able to explain why interpretations s h i f t in this fashion. If we are content to assign these variations entirely to performance errors, positing for instance that the reader inevitably "misses" some aspects of a text on a f i r s t reading and must r e c t i f y these errors on subsequent readings, we leave an element of random variation in our account. Though such errors no doubt occur, to base an explanation of interpretive variation heavily on them would eat away at the assumption of predictability upon which a rhetorical system D. Brent 62 is founded. This would render more problematic any attempt to use our account as a means of illuminating the way knowledge is b u i l t through rhetorical interchange, for we would be less able to describe the ways in which belief systems are linked to the texts which inform them. However, our model as developed thus far cannot provide a more principled account of such variations. The materials we have enumerated are a l l aspects of the reader's store of knowledge, which can be expected to change only slowly as new information gradually forms new schemata and modifies old ones. 1 3 Because of this relative s t a b i l i t y over time, the reader's store of knowledge cannot entirely account for changes in interpretation that may be both sweeping and sudden. In order to account for these variations in a principled way, we must see differences in reading as dependent on factors other than just the reader's total repertoire. Iser provides a clue to one of those factors in his concept of the "wandering viewpoint." Iser uses the term "theme" to refer to the view of the work that the reader is involved with at a given moment; the other potential viewpoints, which continue to affect the reader but are not currently focal, constitute the "horizon." The perspectives by means of which the reader views the theme are supplied by this horizon, the traces of a l l the previous themes with which the reader has been involved. But the reader cannot stand s t i l l in contemplation of a single theme. As her viewpoint moves through the text, the present theme must become horizon as another theme becomes focal. Thus reading is a dynamic act of constantly shifting perspectives. These perspectives evoke a vi r t u a l " F o r a discussion of the ways in which schemata are changed through accretion, modification and restructuring, see Rumelhart 52-54; for a more radical view of schema change, see Iran-Nejad. D. Brent 63 work that does not have a single shape but constantly changes as the moving viewpoint brings into focal attention different themes viewed from the vantage point of different horizons. In "Text, Attention, and Memory in Reading Research," Robert De Beaugrande supplies an account of this aspect of reading from the perspective of discourse processing. De Beaugrande focuses not so much on changing perspectives as on the physical limitations of the reader's working memory. He defines this working memory as "a store that (in contrast to short- or long-term memory) can be defined not by range, but by i t s actions; working memory addresses ongoing input that demands processing in order to constitute comprehension" (22-23). The working memory can hold only a small portion of the text at any one time. To allow the reader access to the entire text, the information from this working store must be added to a larger mental representation (called by De Beaugrande a "text-world model") that is available for consultation but which i s being constantly updated from the "ongoing input." This internal representation consists not just of text passed into memory in raw form but of interpreted text; i t is the reader's estimate of what the text means, not of i t s surface features. In this respect i t is a close analogue of Iser's "virtual work," and like that virtual work i t is not stat i c but an event, constantly shifting as the reader's attention—the "wandering viewpoint"—addresses f i r s t one text segment and then another. Here we have an additional element of variation that can explain differences between one reading and another. As Iser points out, the reader's wandering viewpoint w i l l never wander the same way through the same work twice. Even i f , as Iser argues, the goal of efferent reading D. Brent 64 is to reduce uncertainty and evoke a stable meaning from the t e x t , 1 4 this goal can never be completely f u l f i l l e d . As each new theme comes into view, the new perspectives provided by old themes (now become horizons) w i l l demand that i t be read in a more or less different way. "What has been read," claims Iser, "shrinks in the memory to a foreshortened background, but i t is being constantly evoked in a new context and so modified by new correlates that instigate a restructuring of past syntheses" (111). Thus the effect of previous readings w i l l dictate that the wanderings of the reader's viewpoint w i l l evoke a new virtual work from a rereading, even i f the reader's eyes take the same linear path through the text. Iser's term "wandering viewpoint" is apt in that i t conveys the impression that the reader's viewpoint is not a stable one. However, the word "wandering" conveys the unfortunate impression that the movement of the viewpoint is more or less undirected, that the text establishes a landscape and leaves the reader's viewpoint to find i t s own way through i t . In fact this is not entirely true to Iser's theory; he sees the wandering viewpoint not as undirected, but as directed by a special feature of l i t e r a r y texts: textual "blanks," indeterminate areas in the texts that can only be f i l l e d with the reader's projections of meaning. These blanks break up the process of evoking a coherent, unified virtual work, forcing an alternation of theme and horizon and impelling the reader toward "ideation." According to Iser, this "ideation" is maximized in l i t e r a r y texts that are maximally impeded, forcing the reader to engage in maximal ideation in order to construct the work. In short, in Iser's formulation the l i t e r a r y work is rich, vivid and 1 4 I s e r 111, 185. Smith agrees with this assessment, 158. D. Brent 65 effective in proportion to the degree to which the wandering viewpoint is forced by blanks into ever new constructive activity. For Iser's purposes i t is satisfactory to leave the model at this point, for in a l i t e r a r y theory of reading such as his a high degree of uncertainty can be seen not as a drawback but as a virtue. It is not primarily important that the reader be lead to a specific form of ideation to f u l f i l l specific persuasive goals; rather, the s h i f t in perspectives provoked by the text i s a sufficient goal in i t s e l f . 1 3 In a rhetoric of reading, however, we cannot leave the model here, for a rhetorical theory demands not just ideation, but ideation that is controlled in principled ways. To work toward a more rhetorical account of the wandering viewpoint, we must reformulate the concept in rhetorical rather than l i t e r a r y terms so that, while preserving Iser's valuable insight, we can ask more rhetorical questions of i t . In The New Rhetoric, Perelman and Olbrechts-Tyteca develop the concept of "presence," a principle of v a r i a b i l i t y that in some ways parallels Iser's wandering viewpoint from a rhetorical point of view. Perelman and Olbrechts-Tyteca define presence as a variable in persuasion that acts in concert with the selection of arguments. It is not enough, they warn, for the rhetor to select the data and the arguments that he feels w i l l be persuasive to the audience. Those elements must be made present to the audience; the audience must be made to pay attention to them, to feel them as immediate, and thereby to be affected by them. The "Rhetorical c r i t i c s such as Wayne Booth, however, violently disagree. In A Rhetoric of Fiction (see especially chapter 13), Booth insists that the reader should not simply be turned loose on a text to engage in ideation, that the author must take responsibility for making his values clear and for leading the reader in the directions he sets. However, that is precisely what sets rhetorical theories apart from other li t e r a r y theories that do not emphasise this link between writer and reader. D. Brent 66 challenge of the rhetor, then, is "to make present, by verbal magic alone, what is actually absent but what he considers important to his argument, or, by making them more present, to enhance the value of some of the elements of which one has actually been made conscious" (117). Using this concept, we can enrich our conception of the wandering viewpoint. The text provides a landscape through which the reader's viewpoint wanders, but that landscape is not fixed. It varies according to the degree of presence with which each of i t s features is endowed. The elements of the text that are endowed with the most presence for a particular reader at a particular time are those to which he w i l l naturally pay the greatest attention, those to which the wandering viewpoint w i l l return most frequently in order to contemplate their significance. This perspective allows us to rephrase our question about what guides the wandering viewpoint as a question about what endows certain elements of the text with more or less presence. From Perelman and Olbrechts-Tyteca's point of view, i t is the rhetorical techniques used by the rhetor—the selection of arguments, arrangement, and style that the rhetor uses to make her discourse maximally effective—that allow the rhetor to control presence. These factors, utterly central to any rhetorical theory, are clearly ones that must be taken account of in a rhetoric of reading. But these are means by which the writer controls effect by controlling the text. We w i l l deal with this matter in the following section; at this point we are asking what i t is that accounts for v a r i a b i l i t y from one transaction to the next. In this case, we can r consider the text as a constant. This means that the writer is also a constant, for her influence is entirely limited by the way she shapes the D. Brent 67 text. In addition, as already noted, we must also consider the general materials that the reader brings to the transaction, her repertoire of knowledge, strategies, associations, interpretive moves, as being more or less constant, for the total store of a reader's knowledge does not change abruptly. If we are seeking to explain variation in interpretation under these circumstances, then, we must ask what variable remains that could account for changes in the relationship between these constants. What is i t that brings different text segments into focal position in different ways, that activates different parts of a relatively stable repertoire, so that constants such as the text and the repertoire behave like variables to create a variable entity, the virtual work? The answer must l i e with the fourth factor in the rhetorical transaction: the rhetorical situation. In his seminal a r t i c l e "The Rhetorical Situation," Lloyd Bitzer defines the rhetorical situation as a combination of the audience, the external constraints that govern decision (such as beliefs, documents, facts, traditions, and the l i k e ) , and the exigence, the imperfection in the current state of af f a i r s that gives rise to the need for discourse in order to set i t right. In order to apply this perspective to the question at hand here, we must remember the larger rhetorical situation in which, as argued in chapter 1, the act of efferent reading is situated. The reader reads not just for the proximate goal of constructing a meaning from a text, but for the ultimate goal of participating in a conversation in order to update and modify both his own knowledge and that of others. In this context the exigence that gives rise to this participation in a conversation is not as precisely defined as the ones Bitzer, focusing on specific acts of D. Brent 68 discourse production, has in mind: i t is not necessarily an "imperfection marked by urgency" (386). Rather i t is a general need to make knowledge through discourse, a need marked not always by a specific discourse situation, but by a general awareness that one's structure of beliefs undoubtedly does not perfectly reflect the world as i t really is and ought to be. 1 6 The rhetorical exigence, then, consists largely of a question or questions. 1 7 Seen from the context of a rhetoric of reading, this 1 6 S p i r o illustrates this point when he notes that many experiments in discourse comprehension present only a text with no purpose. "One of the main reasons in everyday l i f e for relating new knowledge to old is negated: selectively processing information in order to update one's knowledge (that i s , keeping the knowledge 'current') of issues which are personally interesting or important. It would be foolish to update one's knowledge with the useless, isolated, and probably false information usually found in experimental prose" ("Remembering Information" 140). This observation not only undercuts such experiments but underscores the importance of rhetorical exigence as a factor in interpretation. The rhetorical exigence also helps refine the schema as a model of comprehension. In Discourse Analysis, G i l l i a n Brown and George Yule point out that one of the main objections to knowledge structural models of comprehension is that they do not explain how the subject avoids instantiating a l l the schemata and subschemata that remotely bear on the discourse, a massive task that would bring comprehension to a halt (238-53, passim). The rhetorical perspective suggests an answer: since reading is goal-oriented, the reader instantiates only as much of her repertoire of schemata as is necessary to produce a satisfactory reading. The definition of "satisfactory," in turn, is determined by a l l components of the rhetorical context, and may vary greatly from one individual or occasion to another. (This connection was pointed out to me by M. J. Powell, personal communication.) 1 7Considered in this way, the rhetorical exigence can be seen as an extension of the traditional rhetorical concept of "stasis." This concept has a long history: i t was f i r s t systematized by Hermogenes in the second century and elaborated in the Roman rhetorics of Cicero and Quintilian, but has antecedents in Aristotle's interest in the questions properly addressed by the various branches of rhetoric. Accordingly i t has been individuated in a large number of specific rhetorical schemes; perhaps the most well-known is Cicero's division of the point at issue into questions of fact, definition, quality and jurisdiction. However, such individual incarnations of this theory are specific to their time and situation (chiefly forensic rhetoric), and are not dire c t l y relevant to this inquiry. What is important in such theories is the insistence D. Brent 69 perspective highlights the fact that when seeking out information and points of view in order to decide what to believe, the reader w i l l actively search for specific pieces«of material that relate to the questions he is asking. To return to Perelman and Olbrechts-Tyteca's formulation, from the reader's point of view presence is conferred on textual elements and on entire texts, not just by an interaction between text and reader, but by their relevance to the current rhetorical exigence—what the reader wants to know.xs Under the impetus of a new rhetorical situation, texts read casually and in a linear fashion months or years previously may be sought out and reread, often piecemeal, as the reader mines them for relevant information. Passages that once were skimmed over because they did not obviously relate to any purpose of the reader's now become luminous with meaning. While reading one work, the piece of text in current focus—the "theme," to use Iser's term—may trigger an association with another piece of the same text or a different one, read last week or last month. With no direct textual stimulus at a l l , pieces of remembered text—that that the starting point of a rhetorical transaction is not just the rhetor's thesis, but that thesis in the context of the question that must be addressed in order to reach a judgement. As John Gage notes, "Conclusions can be discovered, knowledge can be created in rhetorical discourse, only in the context of an issue which requires deliberation to answer" (158). "LSThe reader's repertoire i t s e l f , of course, w i l l be one of the factors controlling the questions asked, for her values and beliefs w i l l obviously influence what she is interested in knowing about the world. The "rhetorical situation," then, cannot be limited to external factors, but must be extended to include certain aspects of the reader herself. Thus the four rhetorical factors we are considering, though treated separately as a matter of theoretical convenience, are in the actual event almost inextricably intertwined. See van Dijk, "Relevance Assignment," especially 122, for an account of this phenomenon from the perspective of discourse processing theory. D. Brent 70 i s , pieces of the virtual work—rise to mind at the dinner table, in the shower, at red lights, and are contemplated for their relevance to the current project. The entire constellation of relevant texts becomes a vir t u a l work within which the reader moves freely, causing theme and horizon to alternate repeatedly. The questions that drive the activation of the repertoire and the movement of the wandering viewpoint are themselves unstable. The very act of acquiring answers, or partial answers, to some questions inevitably throws up new ones. It has long been recognized that there is a "research cycle" in which the reader, armed with only a very general question, explores sources to find answers that modify and refine the question, leading him to different sources and back into the same sources with a new focus. However, this constantly shifting research question determines not just which texts the reader goes to, but also directs the attention of the reader's wandering viewpoint as i t moves through the virtual work that is being evoked. For in Perelman and Olbrechts-Tyteca's formulation, "presence" is dynamic. Textual elements do not just either have or not have presence; they may have i t in greater and lesser degree and may acquire i t and lose i t according to the progress of a rhetorical transaction. This means that as the questions that form the rhetorical exigence s h i f t , the degree of presence conferred upon different aspects of the text shifts with i t , causing the network of interrelationships that comprises the virtual work to assume ever new forms. The reader's repertoire can be similarly activated by the rhetorical situation. Considered as a total store of schemata, this repertoire may be relatively stable. However, the schemata that matter are the ones D. Brent 71 made present for the moment by the rhetorical situation—the ones that appear to bear on the question at issue and on the text segments that are currently being contemplated in search of answers to that question. This much smaller store of relevant schemata can s h i f t much more rapidly than the entire structure of the repertoire, for changing i t is not a matter of learning new knowledge and accommodating to i t ; i t is simply a matter of shifting attention. Moreover, the act of reading i t s e l f i s a force that modifies the structure of present schemata, for the knowledge gained while reading must be considered part of the repertoire. However, i t is not simply submerged in the great ocean of knowledge that forms the more or less stable mass of the total repertoire. Endowed with relatively greater presence by relevance and temporal proximity, i t is more apt to contribute to the much smaller and more specific part of the repertoire that is brought to bear on the immediate inquiry. The relative impact of the new knowledge in this reduced context is therefore proportionally greater than i t s impact on the total repertoire. In short, then, while the total repertoire must be considered a constant, the part that is currently made present by the rhetorical exigence is much more variable. Because the rhetorical exigence is self-modifying, because the questions continually s h i f t under the pressure of their own partial answers, we have here at least a partial explanation for the way the evoked work can s h i f t so rapidly. The same text under the constructive gaze of the same reader whose store of questions has been modified w i l l be interrogated differently and a different work w i l l be evoked from i t . Thus relatively constant aspects of the rhetorical transaction such as the text and the reader's repertoire behave as variables as their degree of presence shifts with the shifting questions that guide the reader's D. Brent 72 inquiry. We now have in place an important part of a rhetorical model of reading. The model accounts for the empirical fact that interpretation varies a great deal from reader to reader and from reading to reading. However, this account does not surrender the overall rhetorical purpose of reading as defined here—reading conducted not for the general goal of aesthetic experience or ideation, but for the specific purpose of inquiring after the solution of a problem by contemplating solutions that other human beings have proposed. We have, in other words, come some distance toward explaining exactly how i t is that reading functions in a rhetorical conversation whose ultimate purpose is not aesthetic but epistemic. 2.5. Constraints on Construction: The Writer and the Text While identifying the materials and strategies that the reader and the rhetorical- situation contribute to the transaction, we have temporarily placed in abeyance the other part of our question: what do the writer and the text contribute to the construction of meaning? So far, we have considered the physical text as a constant in our discussion, for we have been focusing only on the act of reading from the reader's point of view. If we wish to include the writer in the transaction, we must remember that the view we have taken allows us to observe only a segment of a larger rhetorical conversation. At an earlier stage of this transaction, the text is not a constant. At the time of writing, the writer can shape the physical text by choosing as she wishes from a repertoire of textual strategies. In making this D. Brent 73 choice she is guided by her intention to persuade. The strategies she chooses w i l l be the ones which she hopes w i l l lead the reader, in spite of the sources of variation we have thus far discussed, to construe a meaning that w i l l in some way connect with the writer's view of the world, that w i l l lead the reader to modify his system of beliefs in ways that bring them closer to the writer's. If we are to see these choices as a l i v i n g force in the reading act even when the physical text has become a constant, we must ask whether they have any real influence on the shape of the evoked work. In other words, we must ask the question that has always haunted l i t e r a r y criticism, and which is made particularly pressing by the degree of freedom granted the reader by reader-response criticism: can a reader make a text mean anything he wants, and i f not, why not? 1 9 Some c r i t i c s answer this question with a qualified "yes." In Subjective Criticism, for instance, David Bleich e x p l i c i t l y denies that the author's original intention has any significant effect on the interpreter. Discussing Freud's analysis of Michelangelo's Moses, Bleich argues that Freud's interpretation "explains the effect on Freud, regardless of the ar t i s t ' s intention. That i s , the interpretation explains the interpreter's 'intention'—his perception of and response to the work of art" (89). For Bleich, then, the a r t i s t ' s intention has no 1 9 I n "The Construction of Purpose in Writing and Reading," Flower suggests a close interaction between the writer's purpose and the reader's construction of meaning. She argues that the writer's "purpose" is not just a global desire to persuade, inform, or the like, but rather a highly complex web of goals, plans and c r i t e r i a . The reader in turn constructs an ongoing representation of the writer's web of purposes and uses i t as an important source of information when constructing a meaning for the text. Flower asks how the reader uses this source of information. The remainder of this chapter asks a more general question: what does the writer know about the reader that allows bridges such as this to exist in the f i r s t place? D. Brent 74 relevance to the reader's interpretation; a l l that is relevant i s the interpreter's perception of this intention, which i s entirely internal to the reader. Likewise, Fish's theory of interpretive communities suggests that there are no absolute constraints on the meaning of a text. Although he argues that meaning i s constrained by the interpretive communities available to create i t , he poses no overall constraints on interpretive communities. Absolutely any form of interpretive community has the potential to come into being; there is no way for an author to know or even guess the rules of the game that w i l l be played when others come to construct the meaning of his work. Whereas Bleich simply dismisses the author's intention as irrelevant to the sort of analysis he i s interested in, Fish poses a theory that, by depriving the author of the a b i l i t y to predict his audience, severs the rhetorical link between author and reader. In short, i t makes rhetoric impossible. The severing of this link, however, i s not a necessary consequence of a constructive view of reading. In identifying the materials that the reader brings to the transaction, we have in fact implicitly identified one of the ways in which the author uses the text to retain some influence over the reader's construction of meaning: his a b i l i t y to predict the public and social elements of the readers' repertoires. As we have already argued, the repertoire is composed of both public and private elements. The writer cannot predict the most to t a l l y private of these elements, the personal associations that are features of the reader's individual biography. If one is solely concerned with the effects of those associations, as is Bleich, then Bleich is correct that there is no significant link between writer and reader. But other D. Brent 75 elements of the reader's repertoire are far less personal and far more predictable. The reader's schemata, for instance, may in one sense be viewed as personal contributions to meaning, since they exist within the reader and result from the reader's personal experience of the work and of discourse conventions. Yet the experiences on which they are based are not unique to the individual; they are l i k e l y to be similar in many important respects to those of others who share the reader's culture. The experience of what i t means to get together with friends for a game of cards or a night of playing music i s , in broad outline, available to everyone in a given culture. Moreover, the schemata that are abstracted from these experiences are not specific or r i g i d , for we could not possibly have enough r i g i d schemata to assimilate a l l of the diverse experiences we meet every day. Therefore, Rumelhart and Ortony argue, schemata must be quite general: "Schemata attempt to represent knowledge in the kind of flexible way which reflects human tolerance for vagueness, imprecision, and quasi-inconsistencies" (111). The specific details of experience tend to be assimilated into patterns of varying degrees of generality, some of which may be as general as schemata for receiving, breaking, asking, et cetera. Because of this abstractness, the author can depend on these structures to be reasonably well shared between himself and his readers. Only i f a culture were to s h i f t to such a degree that there were no points of contact between writer and reader could communication become impossible. It is d i f f i c u l t to imagine a definition of "culture" broad enough to admit of this degree of v a r i a b i l i t y . 2 0 2°It is this faith in highly generalized universals that, shifted from a schema-theoretical to a moral dimension, underlies Wayne Booth's argument that literature is rhetorical. In A Rhetoric of Fiction, Booth D. Brent 76 Rosenblatt points to another, related element of the repertoire that is not i n f i n i t e l y unstable. For her, an acceptable interpretation is one that contains nothing that contradicts the verbal meanings of the text. These meanings form the touchstone, the stable foundation on which the evoked work is erected, and i t achieves this s t a b i l i t y by virtue of the fact that the basic meanings of words in a language, however coloured by personal overtones and association, do not change from writer to reader to the extent that they become unrecognizable. The generally accepted meanings of words form an element of the reader's repertoire that is relatively stable, shared and predictable. 2 1 The writer, then, has the right to expect that the dictionary meanings of his words w i l l not change to the extent that his most basic meanings w i l l be obscured. The verbal meanings of the text in turn set limits on the possible interpretive communities that may arise. In discourse processing terms, comprehension is not only "concept-driven'' or "top-down," that i s , dependent on top-level schemata for the entire construction of meaning; argues that great works of art are founded on a solid base of universal human values with which only a maniac could disagree (141-43). Kenneth Burke supports his version of rhetorical criticism with a similar argument: The various kinds of moods, feelings, emotions, perceptions, sensations, and attitudes discussed in the manuals of psychology and exemplified in works of art, we consider universal experiences. . . . We c a l l them universal because a l l men, under certain conditions, and when not in mental or physical collapse, are capable of experiencing them. (Counter- Statement 149). 2 XTheories of natural categories and prototypes provide a linguistic account of this s t a b i l i t y by suggesting that the categories on which words are based are flexible but not t o t a l l y arbitrary. See Rosch, "Universals and Cultural Specifics"; Rosch et a l . , "Basic Objects in Natural Categories"; Pappas, "The Role of 'Typicality' in Reading Comprehension"; Mervis, "Category Structure and the Development of Categor i zat i on." D. Brent 77 i t is simultaneously "data-driven" or "bottom-up," dependent on the minutiae of the text, the verbal meanings on which Rosenblatt depends so heavily, to activate subschema and c a l l into being the high-level interpretive structures that organize meaning as a whole. 2 2 To the extent that interpretive communities are like top-level schemata, then, they are in part constrained by the bottom-up influence of the text. In short, the repertoire provides an element of interpretive variation, but i t s freedom of range i s limited and i t is to that extent predictable. The other factor that we have identified as a variable in the rhetorical transaction is the rhetorical situation, the reader's unanswered questions, the gaps in her knowledge that she undertakes the act i v i t y of reading in order to f i l l . This element is similarly predictable. We have been looking at the rhetorical situation from the reader's point of view, but the concept was originally developed to describe the rhetor's understanding of the context of his discourse, an understanding that he can use to systematize the rhetorical decisions that he must make while constructing his text. As imported into a rhetoric of reading, the concept of rhetorical situation does not entirely lose this character of being a set of circumstances about which the writer can have knowledge. A writer knows some of the sorts of questions that his text is intended to answer, for he knows something about the portion of the human conversation in which i t is intended to take i t s place. Each part of that conversation revolves about certain questions that occupy a certain discipline at a certain period of history, and the writer who understands the ongoing conversation in which 2 2See Rumelhart, 41-2, and Rumelhart and Ortony, 128-29, for discussions of this point. D. Brent 78 his work w i l l be read can predict—though without certainty—the general shape of the questions that readers w i l l be using his text to answer. Therefore the rhetorical exigency, which we have identified as a factor permitting v a r i a b i l i t y of interpretation, is nonetheless a p a r t i a l l y predictable factor that the writer can take account of in constructing his text. Like the repertoire i t s e l f , i t i s a mixture of public and private—the public conversation and the private route to knowledge that the reader takes as she contemplates various texts and text segments in the light of her personal inquiry after knowledge. In addition to the shared sections of the reader's repertoire and the rhetorical situation, the writer knows one more thing about the reader that perhaps more than any other factor enables the writer to predict the reader's evocation of the work. Overarching the entire project of evoking a virtual work is the principle of consistency. Rosenblatt sees this principle as providing the foundation for the organizing a c t i v i t y of the reader: underneath a l l of the business of evoking the poem as an event in time " i s the assumption that the text offers the basis for a coherent experience" (55). The work evoked from the text is not necessarily the same for every reader, but for every reader i t i s , or i t approximates, an "interrelated or interwoven "whole" or structure" (91). This provides the reader with a predictable goal in reading: to select from the text and from her personal repertoire, not just any elements that strike her fancy, but the elements that w i l l provide the basis for this unified, coherently structured virtual work. Similarly, Iser writes that "consistency-building is the indispensable basis for a l l acts of comprehension" (125). The reader constantly brings together perspectives to form gestalten that approach p. Brent 79 closure to the degree that they are internally consistent. Iser also argues that literature contains a countermovement in which consistency is systematically disrupted by the action of textual blanks. However, whether or not we accept this assertion as being true of literature, Iser himself does not claim that i t is true of expository texts; the expository text "aims to f u l f i l l i t s specific intention in relation to a specific, given fact by observing coherence in order to guarantee the intended reception" (185). For efferent reading at least, then, Iser essentially agrees with Rosenblatt that the reader strives not to build just any evoked work, but rather one that is maximally coherent. 2 3 The same search for consistency is posited by discourse processing theory. Schema are activated when they appear consistent with the text and with the subschema that are already active. An i n a b i l i t y to find such consistency triggers a search for schema that are more successful at providing i t . We can see this happening as subjects attempted to interpret the music night/card game text referred to earlier. Faced with ambiguous or peculiar representations of experience, subjects attempted to find schemata that would account for the elements of the text by building a unified narrative in which the ambiguous words ("score," 2 3There are, of course, other types of reading such as deconstructive reading, in which the goal is to search, not for consistencies, but for inconsistencies and logical gaps in a text. Such reading, however, is a special case which requires a reader to internalize a particular philosophy of language and to train herself to read according to i t s tenets. Reading for efferent purposes as part of a rhetorical conversation can be considered the more general case. The unspoken rules of this more general case of reading are captured in H. P. Grice's maxims of conversation, one of which is "Be Consistent." Unless given a good reason not to, the recipient of any communication w i l l automatically assume that the producer of the communication is trying to be consistent (see Sperber and Wilson 33-38). Accordingly she can assume that an internally consistent interpretation of the message is l i k e l y to be more representative of the producer's intention than would an inconsistent one. D. Brent 80 "recorder," "diamonds") were assigned denotations that a l l made sense as part of the same schema. Loose ends—unassimilated pieces of data that would not f i t the schema—caused consternation: Subjects tended to at least attempt to preserve the ambiguity of the original passages. When this did not seem possible, subjects were not reluctant to make their frustrations known: "I found this passage incredibly confusing as I f e l t the terminology was misleading and misused and therefore I had to reread many times and guess at what was trying to be said. Example: You don't 'listen to' a score." (Carey and Harste 197) This expectation of consistency-building is perhaps the most important means by which the writer can predict uptake and thereby retain his thread of rhetorical connection with the reader. Even i f he does not know the constituents of the reader's total repertoire, he knows that the ones that w i l l be activated are the ones that w i l l be the most consistent with the verbal meanings of the text and with each other. The reader w i l l i f possible discard inconsistent interpretations and favour those that can be brought into concord with each other. This means that the writer must look upon the three factors in the transaction—the textual elements he supplies, the predicted rhetorical situation of the reader and the elements of the repertoire that he can reasonably expect the reader to possess—as materials out of which a coherent whole must be fashioned. By manipulating the one element that is his to control, the text, he can adjust the sort of coherence that i t is possible to form. He can, in other words, introduce into the text elements which, in combination with the expected repertoire and D. Brent 81 situation, make i t relatively easier to form a coherent work that conveys the meaning he intends, and relatively more d i f f i c u l t to form other coherences. Unity building, in short, provides an organizational framework for a l l other materials of construction, a framework that is not i n f i n i t e l y indeterminate. We have thus identified the materials that allow the writer to maintain rhetorical control by means of the text. With this in mind, let us return to Fish's argument that interpretive communities have absolute authority over meaning. To demonstrate the unlimited range of interpretive p o s s i b i l i t i e s , Fish imagines some extreme examples of interpretive communities, including a community that finds Eskimo symbolism in the works of William Faulkner. Despite the d i f f i c u l t y of an author's predicting such off-the-wall interpretations (as Fish c a l l s them), a l l such interpretations must find something to work with in the text's solid core of verbal meaning, the reader's repertoire and the rhetorical situations that can reasonably be expected to occur. This suggests that they are not t o t a l l y unconstrained; there must be something in Faulkner's work that provides a warrant for Eskimo interpretations, and something in the structure of human nature and human culture that permits the evolution of an interpretive community that would allow such an interpretation. The reader cannot hallucinate the text; rather he must find points of contact with the author as another human being in order to derive an interpretation that is su f f i c i e n t l y meaningful to other readers that an interpretive community can begin to form. It is these points of contact upon which the author depends for his thin thread of communication and rhetorical influence. In sum, then, we have answered our second question, "How do we know D. Brent 82 what propositions the author of a text is trying to persuade us to believe," by answering the more specific question "How can we best describe within a rhetorical framework how the writer, text, reader and rhetorical situation co-operate to generate meaning?" In doing so we have identified the reading process as a delicate balance between sources of constructive freedom and forces that seek to constrain meaning. The former comprise the private and variable aspects of the reader's repertoire and situation, the personal associations of words and experiences and the individual path of the wandering viewpoint. The latter comprise the writer's attempt to f u l f i l l his persuasive intention by manipulating the text according to his best estimate of the reader's repertoire and situation, based on the public, communal and relatively stable aspects of those parts of the rhetorical transaction. It is at this intersection of public and private, of predictability and freedom, that we can be most certain that what we have found is rhetoric. This intersection is a restatement of the intersection between identification and division that Kenneth Burke points to as the natural home of rhetoric: Identification i s compensatory to division. If men were not apart from one another, there would be no need for the rhetorician to proclaim their unity. If men were wholly and truly of one substance, absolute communication would be of man's very essence. It would not be an ideal, as i t now i s , partly embodied in material conditions and partly frustrated by these same conditions; rather, i t would be as natural, D. Brent 83 spontaneous, and total as with those ideal prototypes of communication, the theologian's angels, or "messengers." (Rhetoric of Motives 22) The sources of unpredictability in discourse, the purely private and unique materials and strategies out of which multiple interpretations arise, are the result of the divisions between the members of the human race. But for a l l of this division, there is also identification through the regularities of the human situation on which the writer can rely for approximate, i f not ideal, communication. Thus we have put in place two essential elements of a rhetoric of reading: we have accounted for the vast range of individual interpretations and also for the rhetorical connection between writer and reader. It i s this inevitable mixture of predictability and unpredictability, of identification and division, that in spite of a l l the frustrations thereby created makes reading such a consummately rhetorical act. D. Brent 84 Toward a Rhetoric of Reading Chapter 3: From Evocation to Belief 3.1. Sources for a Theory of Evaluation and Synthesis We have now outlined a rhetorical account of the f i r s t stage in the reading process: the evocation of the virtual work i t s e l f . Guided by his rhetorical situation, the reader relates the symbols of the text to his own repertoire of linguistic and world knowledge to construct a virtual work that is as coherent and unified as possible. This virtual work, though a mental construct of the reader, represents the reader's best estimate of the propositions that the writer is attempting to communicate to him. It constitutes a rhetorical channel, a structure of identifications by means of which one human being makes contact with another. From the rhetorical point of view, however, contact is only part of the process of producing and receiving discourse. In order to participate f u l l y in the rhetorical interchange of meanings, the reader must not only construct but also be affected by the virtual work. He has listened to one previous turn of the unending conversation; now he must decide how much, i f any, of what he has heard is worth believing, and with what intensity of conviction. 1 In addition, he must sort out the ^-This stage is lo g i c a l l y subsequent to interpretation. However, i t must always be kept in mind that this does not imply a s t r i c t temporal sequence, for the various processes involved in reading may be inextricably intermingled. D. Brent 85 differing and often opposing claims of other works evoked from other texts, decide whether the theses presented have more merit than the ones that make up his current structure of beliefs, and f i n a l l y integrate the more meritorious theses into that structure. These two closely related aspects of the reading process generate our third and fourth guiding questions: "How does the reader evaluate the propositions presented by individual texts and decide which to be persuaded by?" and "How does the reader negotiate among the claims of various texts in order to develop a unified system of knowledge?" Whereas interpretation i s only rhetorical by extension, an important preliminary to a rhetorical process, the evaluation and synthesis of ideas i s fundamentally rhetorical. Rhetorical theories are ultimately addressed to the producer of a discourse in order to show him how to persuade most effectively; however, the more philosophical rhetorics derive their precepts not just from an ad hoc compilation of what works, but from a close inquiry into the nature of the receiving mind. George Kennedy makes this point clear: [A philosophical rhetoric] tends to deemphasize the speaker and to stress the v a l i d i t y of his message and the nature of his effect on an audience. . . . Philosophical rhetoric has close ties with dialectic or logic, and sometimes with psychology; i t s natural topic is deliberation about the best interests of the audience. (Classical Rhetoric 17) Because of this implicit focus on the interests of the audience as a means of generating both the methods of persuasion and the propositions that w i l l be presented, the art of rhetoric has always been centrally concerned with the subject about which we are inquiring here: the D. Brent 86 mechanisms by which people evaluate and come to believe the meanings presented by acts of discourse. However, there are limits to the a b i l i t y of rhetorical theory to inform a rhetoric of reading. As pointed out in chapter 1 (27-28), the processes involved in reading are not just the processes involved in writing considered from a different angle. Reading involves a different approach to propositions. Whereas the writer must select from a v i r t u a l l y limitless store of possible arguments according to the probability of their convincing a multifarious audience, the reader must select from a more limited store of propositions that have actually been presented, using as criterion not their likelihood of being persuasive but their fitness for inclusion in a single constantly developing worldview. Thus rhetorical theory can no more be imported wholesale into a rhetoric of reading than can any of the other disciplines that impinge on the questions addressed here. Because of i t s explici t focus on the consumption rather than production of discourse, discourse processing theory can help to supplement rhetorical theory by supplying an overarching conceptual framework within which to explore the persuasive effect of discourse. Whereas rhetoric seeks to answer the question, "How should I choose the strategies that w i l l persuade my audience," one of the important questions asked by discourse processing theory is "What happens when people do or do not add new knowledge to their developing world-views?" Discourse processing theory, therefore, can lend to our inquiry a more exact characterization of what i t means to be persuaded. This characterization i s v i t a l to answering our questions about the reasons for persuasion, for i t is hard to account for the operation of a D. Brent 87 phenomenon unless one has a clear description of what i t i s . It is with this characterization of persuasion from the reader's point of view, then, that we w i l l begin this section of our inquiry. 3.2. What Does It Mean To Be Persuaded? Texts that seek to persuade us do not simply erect a new structure of beliefs on vacant ground. They seek not just to create but to change our minds. This means that they start with structures of belief that are already in place. Each act of rhetoric modifies the world-view that we already have, convincing us to add certain features to i t , to subtract others as no longer useful or true to our current theory of the world, and to modify s t i l l others to bring about a better f i t with re a l i t y . In short, i t adds to our knowledge.2 Thus characterized, "being persuaded" f i t s exactly into the domain 2 I t i s important to keep in mind the fact that we are using words such as "knowledge" and "representation of the world" quite broadly. These terms are intended to refer to systems that include both what are commonly regarded as "facts," ie., mental representations taken to correspond to physical reality, and what are commonly regarded as "values," ie., beliefs regarding the desireable, the virtuous, the honourable. There are obvious intuitive distinctions between these types of belief, and the terms are sometimes carefully distinguished. In The  New Rhetoric, for instance, Perelman and Olbrechts-Tyteca make such a distinction: "We think i t convenient to divide these objects of agreement into two classes: the f i r s t concerning the real, comprising facts, truths, and presumptions, the other concerning the preferable, comprising values, hierarchies, and lines of argument relating to the preferable" (66). It is important to note that Perelman and Olbrechts-Tyteca describe this distinction as convenient, not as necessary. For the purposes of a more general study of knowledge formation such as this one, the distinction i s less convenient, for we are not attempting to link specific lines of argument with specific rhetorical goals. Accordingly, then, for the purposes of this study we w i l l follow the line of thought urged by Booth in Modern Dogma, conflating rather than distinguishing fact and value. Therefore terms such as "belief," "knowledge," "representation of r e a l i t y , " and the like can be taken as being functionally synonymous. D. Brent 88 of discourse processing theories of comprehension, theories that see adding to knowledge as a central goal of discourse comprehension.3 Such theories help to characterize what this process of updating knowledge means from the receiver's perspective. In "The Notion of Schemata and the Educational Enterprise," for instance, Richard C. Anderson summarizes the implications of the schematic perspective on knowledge for our understanding of information acquisition. Attempting to answer the question, "Why are persons in the emerging discipline of 'cognitive science' so excited about the concept of schema" (416), Anderson draws on the work of prominent investigators of discourse processing such as David Rumelhart, Rand Spiro and Bonnie Meyers to suggest that an understanding of schemata is fundamental to an understanding of how knowledge is created, stored, and modified. To describe the way schemata function, Anderson borrows two terms from Piaget: "assimilation" and "accommodation."'9 We have seen how schemata, the organized stores of knowledge built from the experience of a lifetime, help us to sort out sensory input and turn i t into meaning. Anderson c a l l s this process "assimilation": one's pre-existing schemata act as f i l t e r s , selecting the interpretations that can be placed on 3See for instance Spiro, "Constructive Processes" 270. In Relevance: Communication and Cognition, Dan Sperber and Deirdre Wilson similarly argue that the aim of the comprehension process is "to modify and improve the individual's overall representation of the world" (75). "•These terms are defined by Piaget in The Psychology of the Child 4-6. Anderson notes that "the reasons for employing his [Piaget's] terms go beyond deference to historical precedent; his analyses of these processes are among the most insightful to date" (419). In "Cognitive Development and the Acquisition of Concepts," Katherine Nelson also argues that Piaget's model is valuable because, founded as i t is on a complex notion of cognitive structures, i t can account for the formation not only of simple concepts but also of highly complex ideas (223). D. Brent 89 sensory input and t e l l i n g us how to make sense of i t . To make an analogy with Fish's terms, they provide the interpretive framework that breathes meaning into text. But schemata themselves cannot be invariant. If they were, they would force us to see the same things in a l l data, f i l t e r i n g out what is novel because i t cannot be assimilated to old schemata and in effect preventing the growth of new knowledge.3 Applied to the process of improving knowledge through rhetorical exchange, this view would mean that meaningful conversation would cease: we would only hear what we already know. There must therefore be a corresponding movement by which the schemata themselves are modified to take account of new information with which they are not already equipped to deal. Anderson, following Piaget, c a l l s this latter process "'accommodation." Schemata continually s h i f t and adjust to reflect new knowledge as i t becomes available, either through direct sensory input or reasoned prose. This process of updating one's worldview, of changing one's mind to take account of arguments presented to one, is the ultimate goal of reading considered as a rhetorical process. Thus we can characterize the process of being persuaded as a delicate push-pull balance between assimilation and accommodation. To the extent that we assimilate new knowledge to old, we w i l l f a i l to be persuaded in any meaningful way. We may add some units of data to our store of knowledge, but we w i l l not make any changes in the system of beliefs that organizes those data. Our repertoire of schemata w i l l 5 I n Discourse Analysis. G i l l i a n Brown and George Yule c a l l this the "strong" view of schemata: "In the strong view, schemata are held to be deterministic, to predispose the experiencer to interpret his experience in a fixed way" (247). However, Brown and Yule report that this deterministic view of schemata is not generally held. Rather, schemata are more commonly considered to be structures that condition expectations without t o t a l l y dictating the interpretation of discourse. D. Brent 90 remain undisturbed. To the extent that we accommodate our world-view to the new knowledge that is being offered us, we w i l l be persuaded. The extreme case of assimilation is a total failure to learn or to progress; the extreme case of accommodation, mental anarchy. To avoid either of these pathological extremes, rhetorical uptake must consist of a balance between assimilation and accommodation, between f i l t e r i n g our perceptions through old experience and taking account of new. This model of persuasion, based on Anderson's model of schema assimilation and accommodation, provides us with a more specific and more answerable version of our question, "What happens when we are persuaded?" The question is narrowed to a search for the factors that control this balance between assimilation and accommodation—that i s , the factors that answer Booth's question, "When should I change my mind?" We have already identified one of these factors in connection with interpretation: the urge to construct a system that is unified and internally consistent. A structure of beliefs i s like any other structure: i t is composed of parts that are in a mutually compatible relationship to one another, no one of which contradicts the other. When deciding whether or not to accommodate one's belief system to new perspectives offered by a text, one crucial test is whether or not the adjustments required would enhance or detract from the consistency of the entire system. However, this decision i s not a simple yes/no question. Presented with arguments that conflict with one's currently existing beliefs, one does not simply choose between rejecting the arguments or modifying the beliefs. One can modify one's belief system in more than one direction, incorporating into i t some elements of one text, some of another, and D. Brent 91 some that are entirely new constructions triggered by the discussion in which one is engaged but not e x p l i c i t l y found in any one of the texts that clamour for attention. Moreover, one's attitude to a particular text i s always influenced by implied comparisons, conscious or unconscious, with other texts in the conversation. An argument that might be f u l l y convincing by i t s e l f may appear in a completely different light when set beside other arguments that compete with i t . One's personal belief system, then, is a dynamic balance of a number of opposing forces, each attempting to alter the system's precarious consistency in a different direction. This perspective offers us a s t i l l more precise formulation of our basic problem. Changing one's mind is not a matter of choosing from a variety of texts the one that is the most persuasive and then accepting i t . Rather, i t is a matter of assigning the propositions presented by a l l texts an appropriate place in an intricate system. Some w i l l be rejected completely. Others w i l l be accepted in part: some features of the world-view they offer w i l l be incorporated into the reader's own system while others w i l l not. S t i l l others w i l l promote not so much an addition or subtraction of specific features of that world-view as a relative weakening or strengthening of the degree of confidence placed in those features. 6 Thus our last two guiding questions are v i r t u a l l y inseparable. The answer to the third question, "How does the reader evaluate the sAgain, Perelman and Olbrechts-Tyteca's observation is relevant: "What is characteristic of the adherence of minds is i t s variable intensity" (4). In Relevance: Communication and Cognition, Sperber and Wilson's construct a parallel argument from a discourse processing perspective: "Improvements in our representation of the world can be achieved not only by adding ju s t i f i e d new assumptions to i t , but also by appropriately raising or lowering our degree of confidence in them" (76). D. Brent 92 propositions presented by individual texts?" is inextricable from the answer to the last, "How does the reader negotiate among the claims of various texts?" For the persuasive power of a text is not simply an isolated feature of that text. It i s , as we have seen, a function of the degree to which i t is judged more or less suitable than others for inclusion in an integrated and maximally coherent system of belief. This is the judging process that Booth points to when he declares that the art of deciding to accept or reject others' assertions " l i e s in assessing degrees of r e l i a b i l i t y " (Modern Dogma 157). What then are the c r i t i c a l factors in the "logic of relative weight" (as Booth cal l s i t ) that readers use to decide which texts w i l l receive more and which less relative weight in the f i n a l synthesis? In terms of our model, that i s , what decides whether new material w i l l be assimilated to pre-existing structures or w i l l provoke an accommodation in the system, rearranging the schema through which the perceiver interprets reality? Discourse processing theories of comprehension offer us a useful model of the mental structures that are modified by persuasion. However, by their very nature as theories of comprehension, discourse processing theories address only tangentially the question of how the reader reaches this judgement. While offering f a i r l y well developed theories of how people decide what is being said, these theories are somewhat less helpful in explaining how people decide whether or not to believe i t — that i s , how they are persuaded to make significant accommodations in their structure of beliefs on the basis of new information. Anderson admits that the processes involved in answering the latter question, essentially a question of how schemata themselves change, are s t i l l D. Brent 93 poorly understood (424). In her paper "Cognitive Development and the Acquisition of Concepts," Katherine Nelson makes a similar admission: "Most of what we know about knowledge at any stage of development has to do with i t s content and structure rather than how i t is acquired" (215). The assimilation/accommodation model, in short, offers a valuable characterization of the general process of being persuaded, but does not f i l l in the details of what influences that process. In Nelson's words, i t "resists concretizing" (224). One way to concretize this process is to recognize i t as a version of the rhetorical problem of reaching a judgement. To reach a judgement on a question is to reach a decision as to whether a certain proposition or set of propositions is at least provisionally warranted as worthy of incorporation into one's own structure of beliefs. As already noted, the entire business of philosophical rhetoric is to understand persuasive processes in order to specify how the rhetor through words alone can influence the judgements of his hearers. Rhetorical theory thus offers a highly specific account of how i t is that hearers reach judgements under the influence of discourse. One of the most fundamental and long-standing features of this rhetorical account of judgement is the three pisteis or modes of proof, commonly called logos (persuasion based on reasoned arguments), pathos (persuasion based on appeal to the emotions), and ethos (persuasion based on the perceived character of the speaker). Aristotle is the f i r s t rhetorician to have presented these three forms of proof as a f u l l y developed and integrated system: 7 """Kennedy notes reference to ethos and pathos in many earlier rhetoricians (Art 91-103), but points out that in earlier rhetorics, appeal to emotion and character tended to be more a structural than an D. Brent 94 Of the means of persuasion supplied by the speech i t s e l f there are three kinds. The f i r s t kind reside in the character of the speaker; the second consist in producing a certain attitude in the hearer; the third appertain to the argument proper, in so far as i t actually or seemingly demonstrates. (Rhetoric 8) In one form or another this distinction persistently reappears in rhetorics down to the present day. Cicero, for instance, devotes considerable space to the means of composing legal arguments, but also claims that "A potent factor in success . . . is for the characters, principles, conduct and course of l i f e , both of those who are to plead cases and of their clients, to be approved, and conversely those of their opponents condemned" (De Oratore II.xliii.182). He also claims that the successful orator must master the means of speaking which "excites and urges the feelings of the tribunal towards hatred or love, i l l - w i l l or well-wishing, fear or hope, desire or aversion, joy or sorrow, compassion or the wish to punish" (De Oratore II.xliv.185). Quintilian continues the distinction with a particular emphasis on ethical proof, insisting that the f i r s t essential for a good orator i s that he should be a good man (Institutio Oratoria I.Pr.9). From Roman rhetoric the divisions are handed down to Christian rhetoric. In De Doctrina Christiana, Augustine points out the usefulness of logos: "In order that those things which are doubtful may be made certain, they must be reasoned out with the use of evidence" (IV.iv.6). However, emotional means of persuasion are also necessary: inventional feature. Ethos was used in the prooemium to conciliate the judges and pathos in the peroration to s t i r them to belief, but neither was truly an element of persuasion co-ordinate with reasoned arguments as i t is in Aristotle's system. D. Brent 95 But i f those who hear are to be moved rather than taught, so that they may not be sluggish in putting what they know into practice and so that they may fully accept those things which they acknowledge to be true, there is need for greater powers of speaking. Here entreaties and reproofs, exhortations and rebukes, and whatever other devices are necessary to move minds must be used. (IV.iv.6) For Augustine as for Quintilian, ethical proof is the most powerful motivator of a l l : "The li f e of the speaker has greater weight in determining whether he is obediently heard than any grandness of eloquence" (IV.xxvii.59). During the eighteenth and nineteenth centuries, the faculty psychology of the day saw the human mind as divided into discrete faculties such as reason, sympathy, the emotions and the will. This division provided a basis for rhetoricians such as Campbell and Whately to identify the modes of proof with more or less absolute divisions in the human psyche. Like Augustine, they see the address to the reason as being sufficient to create understanding, but in order to move the will, address to the emotions is required. This address can be achieved in part through the sympathy engendered by ethical proof—the recognition of the speaker's sincerity, honesty and integrity. The three pisteis are thus rearranged from being co-ordinate proofs into a more or less linear model in which the reason, the sympathies and the passions work in sequence to move the will. With the passing of faculty psychology, we are no longer obliged to treat—indeed, have no basis for treating—the traditional modes of proof as reflections of objective aspects of the human psyche. Nonetheless, D. Brent 96 Aristotle's original distinction continues to provide a useful means of discussing types of evidence. Booth, for instance, argues that to rehabilitate rhetoric, the modern fixation on logical proof must be tempered by a revival of pathetic and ethical proof (Modern Dogma 144). Richard Weaver, also arguing for a more humane and efficacious rhetoric, e x p l i c i t l y casts his argument back to classical rhetoricians such as Quintilian in order to argue that the human being is not merely a thinking machine, but rather a composite being that responds both to the ethical properties of the orator—a man who is "good in his formed character and right in his ethical philosophy"—and to his feelings, "the a c t i v i t y in him most closely related to what used to be called his soul" ("Language is Sermonic" 224). The three-part division of proof into three modes, then, comes to us sanctioned by a long history of both technical rhetorics such as Cicero's, in which i t is recommended because i t seems to work, and philosophical rhetorics from Aristotle to Weaver, in which i t is recommended because i t recognizes the complexity of the human being as a creature that naturally responds to reason, emotion, and social example. It is this latter recommendation that particularly commends i t to our attention here. A terminology obliges us to make certain distinctions when applying i t . By using the three modes of proof to describe the ways in which the reader evaluates beliefs and is influenced to change her own, we can focus attention on the process of being persuaded as an integrated a c t i v i t y that partakes of at least three different aspects of the human being: his cognitive capacity to apply rules of evidence, his tendency to be influenced by his emotions, and his pull in the direction of ideas that are presented by those whose character he values. D. Brent 97 Dividing our inquiry into these three parts w i l l prevent us from losing sight of any of these aspects of persuasion. Simultaneously, treating these divisions as parts of a single attempt to build a unified model of the persuasive process w i l l force us to account for ways in which these aspects of the human being work, not as separate forces, but together as an integrated process of improving the individual's structure of beliefs through, as Booth puts i t , "taking in other selves" (Modern  Dogma 114). 3.3. How Texts Persuade: The Logical Dimension Our f i r s t approach to the problem of how people determine the relative weight of arguments is to inquire how the rhetorical concept of "logos" can help add s p e c i f i c i t y to the conceptual framework developed so far. To do this we must f i r s t determine the most important characteristics of proof by logos, both in Aristotle's original formulation and in i t s modern reinterpretations. In Aristotle's rhetorical system, logos refers to persuasion by rational argument. Rational argument can be either inductive, arguing from particulars to generalizations, or deductive, arguing from generalizations to particular cases. Though he acknowledges the value to rhetoric of the inductive process, Aristotle devotes far more space to the deductive enthymeme, a concept that he claims is "the very body and substance of persuasion" (1). The enthymeme is the rhetorical equivalent of the syllogism, in which a generalization such as " A l l men are mortal" (the major premise) is connected to a specific conclusion such as "Socrates is mortal" by means of a particular fact that bridges the two, D. Brent 98 such as "Socrates is a man" (minor premise). When used rhetorically as an enthymeme, this form of logical argument may be expressed without one or the other of i t s premises being made exp l i c i t , for these premises may be obvious to the audience and repeating them would be merely tedious. However, the more important feature distinguishing the enthymeme from the syllogism i s the fact that i t s premises and therefore i t s conclusions may be only probable. The enthymeme is thus a device drawn from demonstrative logic but adapted to the needs of an art, not of absolute formal relationships, but of human interaction in the realm of the contingent. 8 Because i t is rooted in human interaction, an enthymeme must be constructed, Aristotle notes, "not from any and every premise that may be regarded as true, but from opinions of a definite sort—the opinions of the judges, or else the opinions of persons whose authority they accept" sThis view—that logos is essentially identical with the enthymeme and example but excludes ethos and pathos—is the traditional view as discussed by historians of rhetoric such as Kennedy (Art 95-102), Golden, Berquist and Coleman (55-60), and Corbett (Classical Rhetoric 50-93, esp. 50). In his monograph Studies in the Philosophy of Aristotle's Rhetoric, Fr. William Grimaldi has recently proposed a re-evaluation of this view of the enthymeme and the three pisteis. Basing his arguments on complex investigations of Aristotle's often d i f f i c u l t text, Aristotle's other works, and precursors to Aristotle's rhetoric, Grimaldi argues that the enthymeme is not limited to a single p i s t i s but instead overarches a l l three. The enthymene is not composed only of reasoned factual evidence (which Grimaldi c a l l s "pragma" in order to reserve the term "logos" for reasoned discourse as a whole). It is also composed of arguments drawn from emotion and character. Aristotle's complex discussion of these topics at the beginning of Book 2 is thus not an interruption in the discussion of topoi begun in Book 1, but rather a continuation of i t : the facts about character and emotion are special topoi no less than facts about government and law. This revision of the standard account, though far from being universally accepted, is attractive in that i t sees ethos and pathos as part of rather than parallel or subsidiary to reasoned discourse. However, whether or not we identify logos with pragma or with the enthymeme as a whole, i t is clear that Aristotle distinguishes between arguments based on quasi-syllogistic reasoning and the more general notion of appealing to emotion and character. D. Brent 99 (156). Unlike the syllogism, then, the enthymeme is not simply a logical device but a device by which the rhetor establishes contact with the audience by arguing from the opinions (doxai) that they already hold. The argument i t structures is rhetorical rather than purely logical because i t is structured for a particular audience using materials drawn from that audience in order to lead them to the judgement desired by the rhetor. 9 The enthymeme is a technique for connecting various general and particular facts about'the world to the case at hand. However, in order to make use of this technique, the rhetor must know how to find general and particular facts that w i l l provide the materials out of which he can build his enthymemes. Accordingly, Aristotle provides l i s t s of topics (topoi) to which the rhetor can turn for material. The special topics (eide) are those which are specific to certain fields, especially ethics, law and p o l i t i c s . Under these headings, Aristotle provides l i s t s , not of specific f a c t s — f o r that is the province of the individual arts themselves, not of rhetoric—but of areas of knowledge in which the rhetor must be informed. These include matters of government, national defence and economics, together with the basics of psychology such as the sources of human motivation and general ethical 9 I n some recent composition theory, this dependence of the enthymeme on doxai. together with Grimaldi's argument that the enthymeme includes ethos and pathos, is expanded into a claim that the enthymeme is not just a means of persuading an audience of a predetermined proposition but of exploring a subject in a co-operative fashion similar to that proposed by Wayne Booth. See in particular Gage 156; Lunsford and Ede, "Distinctions" 46-47; Hairston 62-69. Whether or not one presses to this extent the claim that Aristotle's enthymeme is a co-operative device, i t i s clear that i t , like any other rhetorical device, is s o l i d l y grounded in the need, not just to devise arguments in the abstract, but to devise arguments in the light of an audience's opinions and a speaker's persuasive intention. D. Brent 100 matters such as the sources of virtue. The general or common topics (koinoi topoi), on the other hand, are general principles of argument which can be used to generate discourse in any discipline. As Grimaldi puts i t , they are "forms of inference into which s y l l o g i s t i c , or enthymematic, reasoning naturally f a l l s " (134); they provide the logical shape to which the particular facts of the special topics may be applied in order to produce arguments for a particular occasion. 1 0 At the highest level of generality, Aristotle l i s t s four common topics: possible and impossible, more and less, past fact, and future fact. The common topic of more and less, for instance, embraces a l l arguments from comparison. When applied to the specific material of the special topics, the concept of more and less produces generalizations such as the following: Those things are greater in which superiority i s more desirable or nobler. Thus keen sight is more desirable than a keen sense of smell, since sight i t s e l f is more desirable than the sense of smell; and since i t is nobler to surpass in loving one's friends than in loving money, love of one's friends i s i t s e l f nobler than the love of money. (39) These highly general common topoi are supplemented by a l i s t of twenty-1 0The debate over whether or not the orator requires broad factual knowledge as well as s k i l l in constructing arguments has raged back and forth over the centuries. In De Oratore. Cicero argues that "no man can be an orator complete in a l l points of merit, who has not attained a knowledge of a l l important subjects and arts" (I.v.20). The eighteenth-century rhetorician Hugh Blair, while rejecting the topoi as an effective technique for supplying arguments, implicitly agrees with Cicero as to the rhetor's need for specialized knowledge: "What is truly solid and persuasive, must be drawn 'ex visceribus causae,• from a thorough knowledge of the subject, and profound meditation on i t " (2: 181-2). Quintilian, on the other hand, e x p l i c i t l y disagrees with the Ciceronian position, regarding i t as sufficient "that an orator should not be actually ignorant of the subject on which he has to speak" (II.xxi.14). D. Brent 101 eight valid and ten invalid common topoi that provide more focused lines of argument such as definition, division, precedent, consequences, and turning the opponent's words against him. 1 1 These two major components of logos—the enthymeme and the t o p o i — reappear in various forms throughout the history of rhetoric." 1- 2 In modern rhetoric, they are perhaps worked out most thoroughly in the work of Perelman and Olbrechts-Tyteca and of Stephen Toulmin. In The New Rhetoric, Perelman and Olbrechts-Tyteca provide a copious and detailed discussion of loci of argument such as identity, definition, ends and means, waste, analogy, and dissociation of concepts, a l i s t s trikingly like Aristotle's twenty-eight common topoi in general design i f not in d e t a i l . But unlike many topical systems that intervene between theirs and Aristotle's, Perelman and Olbrechts-Tyteca's is not reduced either to a mechanical system of generating material or a collection of l l r r h i s expansion from four topoi to twenty-eight is traditionally explained in terms of textual history; Solmsen (cited Kennedy, Art 101) suggests that these more practical and less philosophical topoi may be the remnant of an earlier stage of Aristotle's theory inserted into a later text by Aristotle or one of his students. Grimaldi, however, argues for a view of the Rhetoric as a completely unified text, and sees no disjuncture between the four and the twenty-eight common topics. Both, he argues, represent modes of inference by which one can reason from one thing to another. " F o r surveys of the vicissitudes of logos throughout history, see especially Corbett, "The Topoi Revisited," and Harrington, Rhetoric and  the Scientific Method of Inquiry. Both the topoi and the enthymeme reach possibly their lowest ebb during the eighteenth century. During this period, rhetoricians such as Hugh Blair and George Campbell soundly reject both the topoi (at least in their degenerate form as unthinking set-pieces) and deductive reasoning based on the syllogism. However, a more Aristotelian version of logos reappears almost immediately in Whately's Elements of Rhetoric, which discusses s y l l o g i s t i c argument and describes rhetoric in e x p l i c i t l y Aristotelian terms as "an off-shoot from Logic" (4). In his much-reprinted essay "Arts, Crafts, Gifts and Knacks," Richard Young surveys the debate between modern writers who recommend and those who reject the idea of incorporating clas s i c a l notions of logos into composition teaching. D. Brent 102 set-pieces to be used whenever the occasion presents i t s e l f . Rather, i t is a system in which the places of argument are closely tied to a carefully developed theory of the relationship between materials and structures of inference in the context of an audience and a rhetorical goal. In particular, Perelman and Olbrechts-Tyteca echo Aristotle's insistence that the premises of argument must be chosen from premises accepted by the audience: The unfolding as well as the starting point of the argumentation presuppose indeed the agreement of the audience. This agreement in sometimes on explici t premises, sometimes on the particular connecting links used in the argument or on the manner of using these links: from start to finish, analysis of argumentation is concerned with what is supposed to be accepted by the hearers. (65) Although they do not e x p l i c i t l y link their system of argumentation to a rhetorical form of the syllogism, Perelman and Olbrechts-Tyteca nonetheless recapture and redevelop the heart of the Aristotelian enthymeme: the linking of an argument to doxai, the pre-existing agreements of the audience. Stephen Toulmin similarly rejects the forms of s y l l o g i s t i c logic while preserving their essence: the validation of a claim by the use of premises that have already been accepted. In The Uses of Argumentf he adapts formal logic to the uses of practical argumentation in the realm of the contingent—that i s , although he never uses the word, he makes i t rhetorical. His chief contribution is to recast logical argument from the relatively simple three-part form of the syllogism into a more complex form that reflects more closely the complexities of practical D. Brent 103 argument. He begins with a three-part layout that roughly corresponds to the syllogism: specific data (minor premise) is connected to a claim (conclusion) by means of a warrant (major premise) that expresses the generalization that allows the two to be connected. The claim "Peterson is almost certainly not Roman Catholic," for instance, can be derived from the data "Petersen is a Swede" by means of the generalization "Almost no Swedes are Roman Catholics." This system differs from the classic syllogism in a number of ways. Fi r s t , i t contains an explicit place for the qualifiers (almost certainly, probably, possibly) that are invariably needed when arguing in the realm of the contingent. More importantly, i t includes within the structure of the argument a place for the backing of the warrant—further material evidence that establishes the generalization embodied in the warrant. 1 3 The backing for the warrant "Almost no Swedes are Roman Catholics," for instance, is whatever s t a t i s t i c a l surveys have been used to establish the percentage of Swedes belonging to the Roman Catholic church. The result of including this backing step in the layout of the argument is to find a place for induction within the structure of, rather than alongside of, the deductive argument. Toulmin distinguishes between "warrant-using" arguments and "warrant-establishing" arguments: " I t i s interesting to note that this structure in some ways reflects the structure of argument recommended by Cicero, who includes in the syllogism two additional steps that provide the proof for the major and minor premise (De Inventione I.xxxvii.67). Cicero's five-part syllogism, however, is more strategically motivated than Toulmin's layout. Cicero recommends the proof of the premises to make sure that an audience agrees with these premises. Toulmin's motivation i s more c r i t i c a l ; he wants to expose more clearly the differences between inductive and deductive premises in arguments in order to establish a more useful vocabulary for discussing and evaluating them. D. Brent 104 The f i r s t class [warrant-using arguments] w i l l include, among others, a l l those in which a single datum is relied on to establish a conclusion by appeal to some warrant whose acceptability i s being taken for granted. . . . Warrant-establishing arguments w i l l be, by contrast, such arguments as one might find in a s c i e n t i f i c paper, in which the acceptability of a novel warrant is made clear by applying i t successively in a number of cases in which both "data" and "conclusion" have been independently verified. In this, type of argument the warrant, not the conclusion, is novel, and so on t r i a l . (Uses 120) In short, then, Toulmin re-establishes the connection between arguments by example, in which the backing for the warranting generalization is made expli c i t , and enthymemes (or as he c a l l s them, "quasi-syllogisms"), in which the backing for the major premise is assumed. The most important feature of this way of looking at arguments is that i t distinguishes between types of argument on the basis of what is "on t r i a l . " In other words, i t raises to prime importance the question of what premises the audience has already accepted. If the warrant is accepted—if i t can be counted among the audience's d o x a i — then the argument is a warrant-using one, and can proceed to a conclusion from a single datum (such as "Petersen is a Swede"). However, i f the warrant is not accepted—if the audience does not already accept that "Almost no Swedes are Roman Catholics"—then the argument w i l l be a warrant-establishing one, and the inductive backing for the statement "Almost no Swedes are Roman Catholics" w i l l be highlighted. Toulmin's form of argument, then, must start from whatever point in the chain of D. Brent 105 propositions the audience w i l l accept as given. 1 4 The p i s t i s of logos, then, in both i t s original and i t s modern formulation, outlines ways in which the rhetor advances certain claims by using commonly accepted generalizations to form a bridge between the data at hand and the conclusion she wishes to establish. Because of the close relationship between the enthymeme and the syllogism, logos is closely a l l i e d to logic, but is separated from i t by an important distinction: as Toulmin puts i t , a rhetorical argument " i s not a string of propositions which may (or may not) be formally consistent: rather, i t is a human  interchange which is (or is not) substantively adequate" ("Logic" 396). Having said this much, our question now i s , what components of this rhetorical approach to logical argument can help explain how we judge the theses presented to us by the writers we read? That i s , how must logos operate in a rhetoric of reading? Recall that one of the primary features of the rhetorical model of reading is the role played by the reader's repertoire of experience, much of i t gained through prior interaction with texts. This repertoire acts as a f i l t e r , causing new knowledge to be assimilated to the schemata set up by the old. This implies the question of how new knowledge provokes the opposite reaction, causing the schemata to be accommodated to i t . How does new knowledge break through the barriers, the "terministic screens," to use Burke's term, that would seem to exclude i t ? 1 4Toulmin also re-establishes, in different terms, the distinction between special and common topoi. As with the Aristotelian enthymeme and common topoi, the form of Toulmin's argument tends to remain relatively constant across fields of inquiry. Like the special topoi, however, i t s backing, the material facts on which i t is based, changes greatly from one f i e l d to another (Uses 103-04). In fact, in a later a r t i c l e Toulmin e x p l i c i t l y states that the system developed in The Uses of Argument is a rediscovery, "though sleepwalkingly," of Aristotle's system of topics ("Logic" 395). D. Brent 106 The principles of logos suggest that new information is connected to old through bridging generalizations, both inductive and deductive. In induction (Aristotle's examples, Toulmin's warrant-establishing arguments) the generalization is novel and the argument shows how i t f i t s the previously-accepted data; in deduction (the enthymeme or quasi-syllogism), the generalization and data, the major and minor premise, are accepted and the novel conclusion is shown to be lo g i c a l l y inferable from them. An important point is that this account emphasizes the selective nature of the argumentative process. The reader's structure of knowledge may be a coherent and interrelated system, but that does not mean that i t is an inseparable monolith. Rather, i t is a mass of related but separate generalizations, each of which individually may become the premise for an argument when a piece of discourse gives i t presence by so using i t . Take, for instance, a person who is reading various authorities in pursuit of the answers to a particular problem (the "rhetorical situation" discussed in chapter 2). The principle of assimilation would suggest that the arguments that she w i l l find most persuasive w i l l be the ones that most closely connect with the opinions she already holds. As we have seen, those opinions as they stand are the systems that render the incoming argument i n t e l l i g i b l e . Considered as a whole, then, that system of opinions cannot be changed by an incoming argument any more than a man may pull himself up by his own bootstraps. But i f the opinions are considered as separable doxai rather than an unbreakable structure, i t then becomes possible for certain of them to be used as premises for an argument the conclusion of which involves the changing of other doxai. As Burke puts i t , "Some of their [the audience's] opinions are needed to support the fulcrum by which he [the rhetor] would move D. Brent 107 other opinions" (Rhetoric of Motives 56). This viewpoint allows us to explain more clearly not only why arguments work but also why they sometimes f a i l to work even when, from the rhetor's point of view, they are constructed as well as they possibly could be. l s In his a r t i c l e "Attitudes, Beliefs, and Information Acquisition," Robert S. Wyer, Jr. addresses this problem from the viewpoint of discourse processing theory: It seems l i k e l y that the implications of new information w i l l be resisted i f i t s acceptance would require a major cognitive reorganization, that i s , i f i t would require a change in a large number of other l o g i c a l l y related beliefs in order to maintain consistency among them. (264) The rhetorical perspective on logos allows us to put this another way. The important factor i s not simply the total degree of cognitive reorganization that would be required in order to accommodate t o — t h a t i t , be persuaded by—a given argument. Rather, a l l other factors being equal, an argument w i l l succeed to the extent that the opinions i t evokes as premises are more deeply-held than those that would be changed by i t s acceptance. "Deeply-held" can mean "requiring more reorganization of the entire system i f changed," a definition that makes the rhetorical view no more than a restatement of Wyer's. But i t can also mean more general, more fundamental to the human inferential system rather than dependent on specific field-dependent data. In other words, more deeply-held l sNote that Aristotle says of rhetoric, " i t s function is not to persuade, but to discover the available means of persuasion in a given case" (6). Thus in his very definition of the art, Aristotle acknowledges the fact that even arguments that seem irrefutable can frequently f a i l completely to convince. D. Brent 108 opinions, ones that can best be used as fulcra to move others, are the fundamental relationships among things described by the common topoi. The topic of possible/impossible, for instance, predicts that a reader who is shown that a particular belief is impossible in the light of other pieces of data that she accepts w i l l generally be induced to accommodate that belief even though i t involves changing a well-entrenched structure. This view of the ways in which rhetorical connections are established does not mean that the only connecting links are those which ultimately derive from formal deductive logic and which are expressed in the traditional topoi. As Toulmin's scheme makes clear, various forms of inductive reasoning (such as those Aristotle c a l l s examples and Toulmin ca l l s backing) can a l l take their place in a process that is broadly enthymemic. The important feature of the reasoning process developed here is not the specific forms of the connecting links, a form that Aristotle describes in terms borrowed from the syllogism but which could be described in other ways.3-6 The important feature is the way the X 6George Campbell, for instance, rejects formal logic as expressed in the syllogism, arguing that i t is no more than a way of rearranging the knowledge one already has, and that i t can be s t r i c t l y applied only in mathematics (Philosophy 61-70). In i t s place he sets up "moral evidence": the human mind comes to believe by virtue of experience, the accumulation of direct examples provided by the senses; by analogy, the transfer of relationships from one f i e l d of experience to another; by testimony, the weight of other's experiences delivered by report; and by mathematical computation of probability in the (relatively rare) circumstances amenable to such evidence. Campbell himself does not admit the enthymeme as such into his rhetorical system, perhaps because he was unable to separate i t from the syllogism as reduced by the schoolmen to an instrument of meaningless disputation. Yet he insists that logical reasoning as a means of addressing the understanding i s "the foundation of a l l conviction, and consequently of persuasion too" (61); reasoning, by addressing the understanding, forms the foundation upon which the other properties of rhetoric—address to the sympathy, imagination, and passions—must build in order to move the w i l l . It i s but a short step from this view to the one that we have been developing here. The forms of moral reasoning elaborated by Campbell may be seen as common topoi; that i s , general lines of reasoning that a l l rational beings use when D. Brent 109 argument taps into parts of the reader's mental construct in order to influence other parts. An argument thus becomes a good reason for changing one's mind, not just on i t s in t r i n s i c merits such as deductive validity, but also on i t s merits relative to the audience. It w i l l be a good reason i f the doxai i t activates as premises are powerful enough to allow the argument sufficient leverage against the convictions i t is attempting to change. Within this framework a refusal to be persuaded by an apparently irrefutable argument is not necessarily incompetent or pathological. It is rather the natural consequence of the interdependence of the beliefs that make up our theory of the world. Our reluctance to change such structures provides a stabilizing effect that ensures that we have always a solid base of belief to act as an interpretive framework without having to readjust the entire structure to accommodate every new input. The imperceptible gradient from conviction through obsession to neurotic delusion is not a simple matter of willingness or unwillingness to modify belief, but more a question of degree and of functionality: the system becomes pathological and maladaptive when new arguments that should make a conceptual reorganization appear the most desirable response do not in considering whether to be persuaded of a proposition, whether or not that proposition is formally phrased as any type of quasi-syllogism. This link between Campbell's moral evidence and Aristotle's lines of reasoning is noted in The Rhetoric of Western Thought by Golden, Berquist and Coleman, who suggest that "Campbell l i f t e d testimony from the i n a r t i s t i c plane described by Aristotle to the level of a r t i s t i c proof" (153). In other words, Campbell's novelty l i e s not in his notion of the way proofs work but in the kinds of proofs that he examines. See also Richard Weaver's integration of testimony and authority into the topoi in "Language Is Sermonic." Weaver suggests that we have four very general means of reading or interpreting the world of reality: being, cause, relationship, and testimony and authority. Here again the content of the topoi is reinterpreted without disturbing the operating principle. D. Brent 110 fact do so. Thus the art of rhetorical logos, at bottom a study of how arguments may be chosen, can also be used to describe the mechanism by which arguments are accepted or rejected. This account forces us to extend the "information-processing" model offered by discourse processing theory into an "argument-processing" model. Data does not come into the system simply as isolated bits. Rather, i t comes embedded in arguments which may or may not connect the input, the specific claims of the writer, to the doxai that make up the pre-existing system of beliefs. 1' 7 The reader may accept or reject these arguments or use them in ways not anticipated by the writer, according to the demands of the larger system into which they must be integrated. 3.4. How Texts Persuade: The Emotional Dimension Useful though the principle of logos may be in explaining rhetorical uptake, the traditional division of proof into three modes reminds us that i t i s not exhaustive. A rhetorical system, founded purely on logical proof could effectively persuade only people who base their beliefs exclusively upon reason at a l l times. Aristotle notes the limitations of this assumption: The same thing does not appear the same to men when they are friendly and when they hate, nor when they are angry and when 1 7Perhaps this feature of discourse i s one reason why Aristotle was more concerned with a r t i s t i c than with i n a r t i s t i c proofs. Inartistic proofs—material evidence that i s alleged to be persuasive merely by virtue of i t s existence—corresponds to isolated and inert units of data. The same evidence, the same data, comes to l i f e only when used " a r t i s t i c a l l y , " that i s , in an argumentative structure that connects i t to the audience's beliefs. D. Brent 111 they are in gentle mood; in these different moods the same thing w i l l appear either wholly different in kind, or different as to magnitude. (91) The emotions therefore act as a powerful influence on the ways in which proofs w i l l be received. Although he does not approve of deliberately attempting to win a case through irrelevant emotional appeals (which he likens to warping a carpenter's rule), Aristotle recognizes that i f rhetoric is to be an art of influencing human beings i t must take account of a l l , not just some, of the factors that cause human beings to make the judgements they do. He therefore provides a detailed analysis of the emotions and their connection to various general human types. The rhetorician who understands these aspects of the audience w i l l have the tools required to put them in the right frame of mind to receive his enthymemic proofs in a positive light. This need to deal with human beings' emotional nature has been variously reinterpreted throughout the history of rhetoric. Sometimes i t has been interpreted strategically, as mainly a technique to be used because i t works. Roman rhetoric is a good example of this sort of technical pathos; Cicero, for instance, recommends using pathos in the form of emotional arguments, language and even special effects such as showing scars or parading one's children in order to induce sympathy. Interpreted in this way, as a strategic technique, pathos tends to be separated from argument, often beginning to d r i f t entirely out of the office of invention and becoming entangled with arrangement. In De  Oratore, for instance, Cicero writes: In regard to the portions of a speech that in spite of proving no point by means of argument, nevertheless have a very great D. Brent 112 effect in persuading and arousing emotion, although the most appropriate place for them is in the introduction and the conclusion, nevertheless i t is often useful to digress from the subject one has put forward and is dealing with, for the purpose of arousing emotion; and accordingly very often either a place is given to a digression devoted to exciting emotion after we have related the facts and stated our case, or this can rightly be done after we have established our own arguments or refuted those of our opponents. (II.lxxvii.311-12) Cicero thus treats pathos as a digression to be superadded to the argumentative structure of the oration, not as an organic part of the proof. This technical form of pathos set the tone for Roman rhetoric in general, as Johnson notes in "Reader-Response and the Pathos Principle": Such attention in De Oratore (and in other classical works such as the Ad Herrenium and Quintilian's Institutio Oratoria) to definitions of various kinds of audience response and to inventional and s t y l i s t i c strategies for anticipating, circumventing, and co-opting audience predisposition in the service of persuasion awarded the pathos principle a central role in Roman rhetoric and in the long pedagogical tradition that the classical traditions inspired. (157) Although a technical treatment of pathos has never been absent from the art of rhetoric (modern advertising and propaganda might be the best example of the most strategic and least philosophical use of pathos), i t s u t i l i t y as a technique of persuasion has not been the only argument advanced for pathetic proof. The faculty psychology of the eighteenth century provided a new psychological basis for the principle of pathos. D. Brent 113 George Campbell, the most influential of the epistemological rhetoricians of the day, divides discourse on the basis of the four faculties of the human mind into four aims: enlightening the understanding, pleasing the imagination, moving the passions, and influencing the w i l l (Philosophy 1 ) . 1 B The last aim, the only one that Campbell identifies with persuasion as such, is seen as involving a l l the faculties—the understanding, the sympathy, the imagination, and the passions—for only through these other faculties can the w i l l be moved. This view of the relationship between rhetoric and the mind had the result of giving new impetus and new respectability to the use of pathos. Here at last was a s c i e n t i f i c rebuttal to the Platonic distrust of emotional proofs, a distrust v i v i d l y illustrated in Plato's image of the soul as a charioteer drawn up to heaven by the white horse of reason and down to earth by the black horse of the emotions (Phaedrus 38-39). The emotions could now be raised from the status of a necessary strategy to the status of a v i t a l means of motivating, a form of appeal that makes use of a v i t a l human faculty without which not only rhetoric but man himself would be motiveless. Although the distinction between various aims of discourse is s t i l l often maintained, especially in composition theory, 1 9 modern rhetoric has largely abandoned the identification of different aims of discourse with 1 Q E a r l i e r rhetoricians had distinguished between simply teaching and moving to action, and had identified pathos with the latter. See for instance Augustine's advice to appeal to the emotions " i f those who hear are to be moved rather than taught" (IV.iv.6). The eighteenth century rhetoricians, however, were the f i r s t to base this distinction on a s p e c i f i c a l l y psychological model of the human mind. 1 9See especially James Kinneavy, A Theory of Discourse, perhaps the outstanding twentieth century work predicated on the division of discourse by aim. D. Brent 114 different mental faculties. Rather, the human being is seen more ho l i s t i c a l l y , as a being that both thinks and feels. Richard Weaver states this viewpoint in "Language is Sermonic": The most obvious truth about rhetoric i s that i t s object is the whole man. It presents i t s arguments f i r s t to the rational part of man, because rhetorical discourses, i f they are honestly conceived, always have a basis in reasoning. . . . Yet i t is the very characterizing feature of rhetoric that i t goes beyond this and appeals to other parts of man's constitution, especially to his nature as a pathetic being, that i s , a being feeling and suffering. (205) It i s this more h o l i s t i c approach to the human being that enables conviction and persuasion, carefully separated by faculty-based rhetoric, to be reintegrated. As argued earlier (chapter 1, 7-9), "to persuade" can refer not just to an attempt to provoke action but also to an attempt to change belief. To do so, rhetoric must influence the understanding and the passions together, as parts of a united being, not as a linear sequence whose effects are passed from one faculty to the next. As a result, pathos can be returned to i t s original Aristotelian status as a mode of proof, as a source of evidence for knowing, not just a means of motivating. 2 0 As Booth puts i t in Modern Dogma, "every desire, every 2°Just how much of Aristotle's treatment of pathos can be treated as technical and how much h o l i s t i c is far from settled. Clearly, Aristotle attributes at least some of the need for pathetic proof to necessity; we must appeal to matters other than s t r i c t logic, he claims, because of "the sorry state of the audience" (184). Booth largely agrees, claiming that his own rhetoric of assent is not s t r i c t l y Aristotelian because "Aristotle i s much too interested in being s c i e n t i f i c . . . . There is a sense in which he seems to say: Oh, yes, indeed there are many other forms of proof besides the apodictic proof that s c i e n t i f i c demonstration affords, and I w i l l deign to give you a book about them; but isn't i t , after a l l , a pity that i t cannot a l l be done with greater rigor" (Modern D. Brent 115 feeling, can become a good reason when called into the court of symbolic exchange" (164). This i s the aspect of rhetoric that Burke captures in his term "courtship." Men and women are dissimilar beings who must be brought together through the emotional processes of courtship as well as the intellectual processes of argument; similarly rhetoric, because i t contains within i t s e l f the power of pathos, is a more powerful force of identification than is logic alone. Pathos leaps like a spark across the divisions that are inevitable in human society, creating the emotional identifications that allow rhetorical interchange to occur. Thus, says Burke, "Rhetoric remains the mode of appeal essential for bridging the conditions of estrangement "natural" to society as we know i t " (Rhetoric  of Motives 211-12). We can see the effects of this emotional source of proof even in the discourse of "hard" scientists, whose discipline is traditionally supposed to be maximally free of aesthetic and emotional processes. G. Nigel Gilbert and Michael Mulkay's Opening Pandora's Box, a sociological examination of s c i e n t i f i c discourse, presents a particularly fascinating account of the ways in which emotion shapes s c i e n t i f i c judgement. One of Gilbert and Mulkay's subjects offers this account of a particular instance of s c i e n t i f i c revelation: Dogma 144). In Rhetorical Criticism, Edwin Black also insists that Aristotle saw emotion as a purely secondary mode of proof, used only to "bias the judgment of auditors" (138). As previously noted, however (98), writers such as Grimaldi and others influenced by him see pathetic proof as utterly central to Aristotle's system of discourse, to the extent of largely equating Aristotelian and modern rhetoric (see also Lunsford and Ede, "On Distinctions between Classical and Modern Rhetoric"). Whether or not Aristotle is given f i r s t credit for i t , however, the point is the same: modern rhetoric moves pathos into a position of central importance not just as a means of motivating action but also as a source of evidence for developing knowledge. D. Brent 116 He came running into the seminar, pulled me out along with one of his other post-docs and took us to the back of the room and explained this idea that he had. . . . He was very excited. He was really high. He said, "What i f I told you that i t didn't take any energy to make ATP at the catalytic s i t e , i t took energy to kick i t off the catalytic site?" It took him about 30 seconds. But I was particularly predisposed to this idea. Everything I'd been thinking, 12, 14, 16 different pieces of information in the literature that could not be explained, and then a l l of a sudden the simple explanation became clear. . . . And so we sat down and designed some experiments to prove, test this. (47) Here we see a pattern of discovery opposite to the one normally implied by formal s c i e n t i f i c literature. In formal accounts, Gilbert and Mulkay point out, the theory is presented as i f i t followed inductively from s c i e n t i f i c data. Informal accounts, however, often reveal that the theory is arrived at f i r s t , in a flash of emotionally charged intuitive insight, and the s c i e n t i f i c results collected later by way of confirmation and documentation. Our conceptual model of belief modification, then, must be refined to take the operation of pathos into account. Hitherto we have treated i t primarily as a cognitive model, which i s not surprising given i t s debt to cognitive theories of discourse. But a rhetorical model of belief modification must be larger than this. As Booth points out, what we "know" is not just facts; i t is a complex of attitudes and perceptions that shape our selves and give us our personal identities. Accordingly, our model must describe beliefs not just as organized bundles of D. Brent 117 information but also as structures that include attitudes and values, features of the mind which are intimately connected to but not limited by cognitive processes. Once attitudes and values are added to the system, we precipitate a change in our account of the way in which new arguments gain admission to that system. Using the enthymeme as a paradigm of the cognitive processes by which arguments have their effect, we described how new arguments seek admission to the reader's belief system on the strength of their relationship to opinions—doxai—already accepted. To take account of pathos within this system, we must treat "opinion" as referring to more than just what is accepted as objectively "real," that i s , to factual knowledge. It must also refer to values, attitudes, emotional attachments. Beliefs as to what is preferable, honourable, morally or aesthetically superior—"better" in the largest sense of the term—can become the premises of an argument as easily as can factual statements and definitions such as " A l l men are mortal" or "Almost no Swedes are Roman Catholics." Therefore a proposition may be recommended to the reader both on the strength of i t s ties to her opinions about factual r e a l i t y and on the basis of i t s emotional ties to her more value-laden doxai. Thus Cicero's orator pointing to the defendant's war wounds may be playing on an emotional reaction that is technically irrelevant to the legally-circumscribed business of the courtroom; however, i t is not irrelevant to the larger human context in which the act of rhetoric is embedded. A premise such as "Much respect is due a man who has been wounded in his country's service" recognizes the social and emotional connections that give a culture i t s cohesion and human beings their D. Brent 118 humanness. To use Toulmin's terminology, i t is backed not by a set of inductive data but by social experience. Its use in court can be decried as cunning manipulation not because i t is i n t r i n s i c a l l y of an inferior order of proof but only because i t has been used in a context in which such evidence is not f u l l y appropriate. Likewise, the scientist's emotional apperception that his theory is worth pursuing could be expressed as an emotional premise to the effect of "An idea that is exciting i s more l i k e l y to be f r u i t f u l than one that is not." This premise is a warrant backed by a lifetime of personal experience in which the emotionally and the intellectually f r u i t f u l have generally been commingled. Like a l l rhetorical premises, i t is not certain, but i t is su f f i c i e n t l y probable to be a useful source of human knowledge. This form of proof is s u f f i c i e n t l y unlike what is normally considered logos that i t deserves i t s status as a separate p i s t i s . The sorts of premises constructed above to account for emotional proof are always just that: a r t i f i c i a l constructs to explain what is in practice communicated through unanalysed emotional connections, largely bypassing what are usually considered the logical faculties. To treat them as actually being processed through a formal structure derived from s y l l o g i s t i c logic is to stretch the definition of logic past the point of endurance. Yet from the point of view of our model, they function in essentially similar ways. Emotional evidence, like logical evidence, exerts pressure on the reader's entire structure of beliefs. It w i l l provoke an accommodation of that structure to the extent that i t connects with and uses as fulcra pre-existing doxai that are more deeply-held, more d i f f i c u l t to change, than the ideas that would be changed by D. Brent 119 accepting the argument.22-Our model of rhetorical reading, then, has been further elaborated to take account of the different sorts of claims that texts may make for their readers' belief. The reader's structure of beliefs can be pressed to change by two different types of connections between ideas: the sort that can best be described as l o g i c a l — t h a t i s , generalizations that connect specific data to previously accepted ones—and less data-oriented connections that can better be described as emotional. Only i f the model is extended in this way can i t do justice to the complex view of the human being, not as a "thinking machine," as Weaver puts i t , but as an entire being who can not only be motivated but also make judgements on the basis of values and feelings. 3 . 5 . How Texts Persuade: The Ethical Dimension. Thus far we have discussed how doxai, both logical and emotional, form connections between the writer's and the reader's point of view and help the reader reach a judgement as to whether to incorporate any part of the writer's beliefs into his own. However, the model as elaborated " T h i s is essentially the view that Grimaldi attributes to Aristotle: that i s , the view that pathos, ethos and pragma are co-ordinate forces within an expanded notion of the enthymeme. In "Toulmin on Argument," Ehninger and Brockriede espouse a somewhat similar view of pathos in their interpretation of Toulmin's system; they suggest that emotional proofs, like logical ones, can be used as the backing of a warrant ( 3 8 3 ) . Grimaldi's view puts considerable strain on the definition of the enthymeme, and Ehninger and Brockriede's puts considerable strain on Toulmin's system, for neither Aristotle nor Toulmin e x p l i c i t l y includes pathos within the structure of the enthymeme or quasi-syllogism. However, even i f Grimaldi and Ehninger and Brockriede press their sources rather farther than i t is possible to follow them, their views are valuable for the way in which they reflect the modern tendency to see emotional proof as part of a structure of argument, rather than as a digression from or reinforcer of the argument. D. Brent 120 to this point speaks of arguments and the materials that form them as i f they were disembodied, acting on their own to demand acceptance from the reader. The epistemological viewpoint upon which this entire inquiry is premised argues that this is not the case. Invention, we argued, is a social act (chapter 1, 25). We develop knowledge not just through an independent interaction with the facts of the universe but in social interaction with other people. The texts that those people create are just extensions of themselves, a means to an end: the interchange of selves upon which human knowledge depends. Thus our rhetorical model of reading must take account of one more factor in judgement: the perceived character of the speaker. This i s Aristotle's third p i s t i s , ethos: The character of the speaker is a cause of persuasion when the speech is so uttered as to make him worthy of belief; for as a rule we trust men of probity more, and more quickly, about things in general, while on points outside the realm of exact knowledge, where opinion is divided, we trust them absolutely. (8) As part of the art of rhetoric, the speaker's actual character is not direc t l y relevant: "This trust, however, should be created by the speech i t s e l f , and not l e f t to depend upon an antecedent impression that the speaker is this or that kind of man" (8-9). 2 2 Accordingly, Aristotle 2 2 A r i s t o t l e does not elaborate on the reasons for this statement, but one can assume that some combination of two factors led him to make i t : the fact that he is describing a speaker's art, and so is naturally more interested in what the speaker must do than in " i n a r t i s t i c " sources of proof such as the speaker's actual character; and second, the fact that a speaker who depends on antecedent impression may not succeed with an audience who is unfamiliar with the speaker. The speaker, then, who is in fact of good character must use his art to "make present" (to borrow Perelman and Olbrechts-Tyteca's term) his character to the D. Brent 121 provides a detailed description of tr a i t s of character, together with a description of the emotions and the kinds of virtue with which they are associated. The rhetor who understands these human tr a i t s w i l l be in a position to project himself as being a person of intelligence, sound moral character, and good w i l l . As Johnson points out in "Ethos and the Aims of Rhetoric," this essentially strategic view of ethos was continued in some but not a l l succeeding rhetorics. Cicero continued the Aristotelian tradition of strategic ethos but wedded i t , not to a knowledge of virtue, character, and emotion, but to style and delivery. The speaker, Cicero claims, can make himself appear to be a good man through his language and unruffled, eloquent presentation of his ideas. This strategic tradition i s continued in medieval handbooks of letter-writing in which i t is reduced to a specific part of the document, the benivolentiae captio, which i s devoted to a formulaeic securing of good w i l l . 2 3 The psychological rhetoric of the eighteenth and nineteenth century reinterpreted this concept in the light of the prevailing faculty psychology; Campbell and Whately, for instance, both link ethos with the faculty of sympathy and place i t in their linear framework as a necessary preliminary to s t i r r i n g the emotions and so moving the w i l l . As with pathos, then, ethos i s treated as strategic by the faculty rhetoricians, but strategic because i t evokes a v i t a l part of the human character. Johnson suggests that i t audience. Whether a speaker of bad character should use his art to make himself appear the opposite, Aristotle does not clearly say, though one can assume from his general concern with rhetoric as a means of defending truth that he would disapprove of such a strategy. "See Johnson, "Ethos" 105-06. In Rhetoric in the Middle Ages, James Murphy points out the close relationship between this division of the medieval letter and the function of the Ciceronian exordium, 225. D. Brent 122 is this strategic form of ethos that has influenced present-day-composition handbooks who enjoin the reader to project a favourable "persona" in order to engage the reader (113). There i s , however, another view of ethos that insists that i t must be rooted in the real character of the reader. Johnson traces this view of ethos to Plato's view that "the orator bears witness to the ideal Good by being an incarnation of virtue" ("Ethos" 103). This tradition is continued in classical rhetoric through Quintilian's insistence that the rhetor must not only be "a thorough master of the science and the art of speaking," but also "blameless in point of character" (Institutio  Oratoria I.Pr.18). For Quintilian, this requirement results not so much from factors intrinsic to the art of rhetoric i t s e l f as from the rhetor's position in society. Surely society has a right to expect not only technical knowledge but also moral stature from an orator who is not just a professional speech-maker but who also can r e a l l y play his part as a citizen and is capable of meeting the demands both of public and private business, the man who can guide a state by his counsels, give i t a firm basis by his legislation and purge i t s vices by his decisions as a judge. (I.Pr.10)' This requirement that the orator be personally good finds i t s natural home in Christian rhetoric, in which the orator is the guide not only of the state but of the soul. Augustine insists that the Christian orator must be truly wise in Christian doctrine, a wisdom that comes from sincere adherence to Christian principles and a prayerful study of doctrine; in fact, such an orator can be called eloquent even i f he f a i l s to win the assent of his audience (IV.xvii.34). We see a similar strain D. Brent 123 in eighteenth century rhetoricians who, despite their tendency to use faculty psychology as a foundation for strategic technique, also in s i s t on the need for sound doctrine when rhetoric is used for Christian purposes. 2 4 Yet this idealist view of the need for sound content is continually intermingled with a concern for effect. Blair, for instance, recommends against counterfeiting one's own reaction to a subject not just because i t is wrong to do so but also because i t probably won't work, for i t "betrays speakers into an unnatural manner, which exposes them to ridicule" (2: 55). 2 S Modern rhetoric shows the influence of both of these interpretations of ethos. Composition texts tend to recommend ethos as a strategy to capture the reader's good w i l l ; however, more philosophical rhetoricians such as Weaver and Booth focus more on the place of rhetoric in society as a reason for the importance of ethos. Because he sees us a l l as rhetoricians "affecting one another for good or i l l , " Weaver e x p l i c i t l y agrees with Plato and Quintilian that "the true orator is the good man, s k i l l e d in speaking—good in his formed character and right in his ethical philosophy" ("Language is Sermonic" 224). Booth is less Platonic in that he does not postulate the existence of an ideal order of goods that exists outside man. However, he is in f u l l agreement with Quintilian that the rhetor—by which he means every human being capable of symbolic interaction—has a right and a duty to use her power of discourse to shape the moral and social fabric of the culture in which 2 4See for instance Campbell 105 and Blair 2: 106. 2 S A similar argument has been advanced against modern handbooks of Business English that recommend an insincere attempt to capture goodwill at the beginning of a document. See Brent, "Indirect Structure and Reader Response." D. Brent 124 she lives. Particularly important to Booth's treatment of ethos is his conflation of facts and values. If what we are sharing in rhetorical interchange is not just knowledge of facts but representations of v a l u e s — i f we are being asked, that i s , to accept as part of ourselves at least part of the values of those with whom we are in communication—then i t becomes particularly important to be able to assess not just the "message" as a disembodied argument but also i t s source. The character of the speaker, her honesty, wisdom, and sincerity, become particularly important "good reasons" for belief. Ethical proof, which Booth describes as "taking in by contagion" (144), becomes a major source of evidence. Where do these views on ethos f i t into a rhetoric of reading? From a philosophical point of view, the neoPlatonic interpretation assumes great importance. If knowledge is evolved in a vast and interrelated conversation, the rhetor who puts forth as true belief propositions that are not based on the best opinions of the best sort of character is doing more than mislead his hearer. He is polluting at i t s source a vast system of interrelated ideas and beliefs that forms a structure throughout an entire society. Like a river-system, humanity's knowledge-making system may contain enough sheer volume of good reasons to dilute the bad into relative harmlessness. This does not, however, absolve the rhetor from responsibility for spreading the best ideas possible. It is no accident that the scientist who is found to have f a l s i f i e d data is disgraced; those who make insincere statements of a more value-laden quality would undoubtedly be equally disgraced i f only the charge were D. Brent 125 more easily proven. 2 6 If a l l human beings are rhetors, however, the statement that the rhetor should be of good character is too obvious to be helpful. Of course the world would be vastly improved i f we were a l l of good character and based a l l of our statements on that character; who could argue otherwise? In developing a rhetoric of reading, however, we are centrally interested in a more exact question: how does the reader use character as evidence for reaching a judgement on the trustworthiness of the representations of belief presented by texts? This question c a l l s upon us to apply the more strategic interpretation of ethos. F i r s t let us make some distinctions. There is what one might c a l l the " i n a r t i s t i c " consideration of testimony, that i s , judgements based on what we know of the speaker's character or of the character of the testimony in general, independent of the content of the discourse i t s e l f . This is not quite the same as the Platonic concern with the rhetor's moral character as such; rather i t is concerned with the effect of the reader's perception of that character, what Aristotle c a l l s "an antecedent impression that the speaker is this or that kind of man." Richard Whately deals with this form of evidence at some length, ca l l i n g i t "a kind of sign" and providing detailed discussions of how i t can be weighed on the strength of factors such as number of witnesses, character of witnesses, whether the testimony is substantiated by independent 2 S I n this context i t is interesting that hate-literature publisher William Zundel was successfully prosecuted on a charge of "knowingly spreading false news." D. Brent 126 witnesses, and so forth (58-76). 2 7 We s t i l l use such i n a r t i s t i c sources of evidence today, most obviously in courtroom procedure but in fact in a l l the dealings of our daily lives. We are clearly most apt to accept as worthy of belief statements that proceed from the greatest number of most reliable witnesses. 2 8 In scholarly research, this procedure is formalized in the institution of the footnote: statements of fact are supported by references to researchers who attest to them, and statements of opinion are supported by references to other writers who share them. 2 9 The r e l i a b i l i t y of those witnesses in turn is partly substantiated by the number of citations they receive in other literature. In short, scholarly cross-referencing i s simply a way of tapping into a community . consensus broken down into specific representatives whose testimony is cited as evidence. There is also a form of ethos that we might label " a r t i s t i c " , that 2 7 T h i s interest in testimony, as Golden, Berquist and Coleman point out, was occasioned partly by the contemporary concern with whether miracles could be substantiated by testimony, a question raised by David Hume. Campbell and Whately both attempted to refute Hume's negative answer to this question (Golden 139-43; 115-55; 164). However, Campbell also points out how testimony forms a major avenue of knowledge in " a l l the branches of philosophy, such as, history, c i v i l , ecclesiastic, and literary; grammar, languages, jurisprudence, and criticism" (56). 2 SAs Booth points out, both number and r e l i a b i l i t y are at stake. Simple weight of numbers in support of an i n t r i n s i c a l l y foolish proposition does not in i t s e l f constitute a good reason for accepting i t ; such affirmations must themselves be grounded in good reasons (Modern  Dogma 148). 2 9Whately makes a useful distinction between matters of fact which might conceivably be submitted to the senses (such as whether there is land at the south pole), and matters of opinion that cannot be (such as whether a certain act was blameworthy) (Elements 58-60). In both cases we are interested in the witness' a b i l i t y to form a judgement, the difference being in the relative weight we give to the witness* accuracy of observation as opposed to his intelligence at weighing observations. D. Brent 127 i s , the sort of ethos in which Aristotle is most interested, that inheres in the discourse i t s e l f . This sort of ethos is in many ways more interesting from a rhetorical point of view, because like a r t i s t i c logos, i t is illuminated and given l i f e by working as an in t r i n s i c part of a discourse under the direct command of a writer, rather than functioning externally to i t . Yet this sort of ethos is much more d i f f i c u l t to account for, because i t works invisibly. Our impression of the author behind the work is formed by subtle constructive process, not by the weighing of evidence that can be pointed to. In particular, the obvious problem with ethos as a feature of print conversations is that, unlike the Aristotelian rhetor, the author of a printed text is not physically present to give us an impression of his character. How can ethos affect persuasion when the author is known only through his works, and perhaps only through a single work? To answer this question, l e t us f i r s t ask whether the difference between listening and reading is r e a l l y as significant to the operation of ethos as the question implies. The main difference between the reader and the hearer from this perspective is the absence of delivery as a factor in persuasion. 3 0 This might be a problem for ethos according to the Ciceronian formulation, for Cicero places a great degree of emphasis on the way in which the rhetor must present himself through his delivery. Aristotle, however, barely mentions delivery, and claims that ideally i t 3°It i s , of course, possible to argue that there is a sort of "delivery" inherent in the printed text. Matters such as the date of publication, the prestige of the publisher and the format of the book (formal? scholarly? cheap-looking? remaindered a month after publication?) cannot help but exercise their influence on the writer's ethos. In Rhetoric in the Classical Tradition, Winifred Bryan Horner also argues that the mechanical presentation of a student essay and the knowledge of discourse conventions displayed are equivalent to oral delivery (378-79). D. Brent 128 should consist simply of not paining the hearer (183). Rather, Aristotelian ethos i s an aspect of invention. As one of the three p i s t i s , i t refers to the impression the rhetor gives of himself through his choice of words and arguments, based on his solid knowledge of the virtues, the emotions, and the facts of human character. The audience of the Aristotelian orator, therefore, must construct the rhetor's character from the clues provided by the discourse in very much the same way that a reader must, for the audience of a speech has no more direct access to the interior of the person who stands before them than does the reader. Wayne Booth confirms that i t i s possible—indeed inevitable—for the reader to construct a personality for the author and then to use i t as a source of evidence for taking in the values of that author. In A Rhetoric of Fiction. Booth argues that the novelist creates an identity for himself through his writing: As he writes, he creates not simply an ideal, impersonal "man in general" but an implied version of "himself" that is different from the implied authors we meet in other men's works. . . . Whether we c a l l this implied author an " o f f i c i a l scribe," or adopt the term recently revived by Kathleen Tillotson—the author's "second s e l f " — i t is clear that the picture the reader gets of this presence i s one of the author's most important effects. However impersonal he may try to be, his reader w i l l inevitably construct a picture of the o f f i c i a l scribe who writes in this manner—and of course that o f f i c i a l scribe w i l l never be neutral toward a l l values. Our reactions to his various commitments, secret or overt, w i l l help to determine our response to the work. (70-71) D. Brent 129 Both the reader and the hearer, then, to a very similar degree, must construct the character of the rhetor. Each must build, from clues in the text, not only an evoked meaning but also an evoked writer, a personality that l i e s behind the text and through the arguments he uses, the c r i t e r i a he demonstrates, the claims he asserts, projects a character that the reader w i l l admire to a lesser or greater extent. 3 3-Having established that i t is entirely possible for this " a r t i s t i c " sense of ethos to be applied to written texts, we are led to two more specific questions. F i r s t , why exactly is character persuasive—that i s , why do we trust "men of probity" more completely than others? Second, can we legitimately do so without committing the logical fallacy of argumentum ad hominem?32 To answer these questions we must remember that the propositions that we encounter in texts are not independent of each other; they are presented as systems of interconnected propositions. When manifested as texts we c a l l these interconnected systems "arguments"; conclusions are built upon sets of premises and presumed connections between premises, many of which may be unspoken. As with any structure, i t is possible to pick apart these systems. Presented with an argument, we may grant i t s conclusions but not a l l of i t s premises, or p a r t i a l l y grant some of the premises but organize them in ways different from those intended by the 3 : LFor other, related arguments that the reader constructs the author of a text, see also Rosenblatt, 50, 124-26; Iser, 35; Burke, Philosophy  of Literary Form 22-24. Booth, however, makes the strongest assertion that the reception of the work is governed by an impression of the author that can be generated entirely from the work i t s e l f . 3 2See Copi, 89: "This argument [argumentum ad hominem] is fallacious, because the personal character of a person is logically irrelevant to the truth or falsehood of what that person says." D. Brent 130 author (herein l i e s the creativity of human response.) But there is a natural tendency to consider the system as a package, to admit or reject the argument as a whole. When considered as a result of logical entailment, this tendency of arguments to hang together is clearly a feature of logos. 3 3 But logical entailment is seldom the only, or even the major, reason for arguments to be accepted as complete systems. Arguments are not just isolated logical systems presented by texts. Texts are representatives of people, and the systems of belief that readers see in texts represent samples of the systems of belief held by the authors of those texts. These systems, whose characteristics we have been elaborating throughout this inquiry, are not merely logical systems but delicately balanced schematic structures that incorporate the writer's personal experience of the world together with the attitudes and values that provide further organizing principles. Each system is the best estimate, of a particular human being at a particular time, of the way the world works. To the extent that such systems are seen as being attached to particular human beings, they can be seen as the foundation of "character." This ancient term can be taken to refer not just to a person's moral conduct—his willingness to help others and to pay his income tax—but rather to the entire system of beliefs that we perceive him to be operating under, as manifested by tokens such as actions and words. To admire a person's character is to admire the beliefs we perceive him as having. And i f we admire certain aspects of those 3 3See for instance Black's analysis of Newman's Apologia in Rhetorical Criticism. Black argues that Newman's letters, many of which are quoted in the Apologia, form an unbreakable logical chain. "Once a reader grants a single implication of the letters, he i s led to grant a l l their implications" (154). D. Brent 131 beliefs, we are more l i k e l y to admire others, not just because of an i l l o g i c a l "psychological transference," 3* but because we know from experience that beliefs are not isolated. Because we know that such systems are complete structures, albeit not always based purely on logical interconnections, we are encouraged to attend to them as systems rather than as heaps of independent propositions, and to accept or reject them as wholes rather than piecemeal. Why do we bestow this sort of admiration on certain human beings more than on others? In A Rhetoric of Motives, Burke answers this question in terms of identification: You persuade a man only insofar as you can talk his language by speech, gesture, tonality, order, image, attitude, idea, identifying your ways with his. Persuasion by flattery is but a special case of persuasion in general. But flattery can safely serve as our paradigm i f we systematically widen i t s meaning, to see behind i t the conditions of identification or consubstantiality in general. (55) Thus ethical proof depends on identifying two systems of belief, the reader's and the writer's. The power of this process i s , as Burke points out, not simply dependent on flattery. For i f beliefs form coherent systems, then i t follows that a system that we see as being congenial to ours in some respects—the respects that have already commanded admiration—is a system whose other aspects have a better chance of also being congenial, that i s , of f i t t i n g into our system, rectifying i t s gaps, correcting i t s inconsistencies, answering i t s unanswered questions, and generally improving and extending the system of knowledge and values 3 4See Copi 89. D. Brent 132 that is already in place. Whereas logical and pathetic proof depends on dividing the reader's belief system into separate doxai to use as fulcra to influence others, ethos depends on the opposite movement, the impulse to treat the writer's belief systems as a whole and to accept the whole on the basis of the character of the person responsible for i t . In Modern Dogma, Wayne Booth argues that this form of proof can often be stronger than logical proof: "We a l l excuse gaps in argumentative cogency i f we believe that the speaker or writer is essentially reliable in sharing values we share. And i t would be unreasonable not to" (157). Ethical proof therefore f i t s into our psychological model of belief systems. Because of our natural desire to maintain the coherence of our own system as economically as possible, we are attracted to systems associated with those who, because of our admiration for their "character" in the widest possible sense, seem to offer complementary rather than antagonistic perspectives. In extreme cases, of course, this effect can become pathological. Edwin Black notes that when a rhetor's arguments correspond with the audience's cluster of opinions, "this rhetor's word w i l l receive increasing credit from those auditors un t i l he becomes for them a prophet—that i s , his word alone w i l l be sufficient argument for them" (173-74). This phenomenon explains the power of some rhetors from Joe McCarthy to Adolph Hitler to command a fanatical audience merely by feeding back, and taking credit for, the opinions the audience already holds. In this case ethical proof short-circuits the way in which a conversation naturally modifies belief through mutual interchange, turning i t instead into a self-replicating c i r c l e in which no new knowledge can be b u i l t . But this i s merely a pathological extreme of the normal case. Far more often than not, our knowledge of the source D. Brent 133 of a communication provides v i t a l evidence for the value to us of believing i t . Here, then, is a legitimate psychological reason for the transfer of acceptance from the person in general to certain of the beliefs that he presents, and thus an answer to our second question. When considered not just in the context of formal logic but in the context of naturally occurring rhetorical interaction, argumentum ad hominem is not necessarily as fallacious as i t is sometimes made out to be. 3 5 Our perception of the quality of the source of an argument is relevant to the acceptability of the argument, for both are part of a single system. Different subjects of discourse w i l l , of course, result in different grounds for admiration being more or less relevant. A person whom we admire as an excellent basketball player cannot legitimately command more respect on the subject of nuclear physics than someone whom we do not. To the extent that we translate irrelevant aspects of a person's ethos into acceptance or rejection of particular beliefs, we do indeed d r i f t into a fallacy of relevance. But the most important aspects of ethos are 3 S I n fact, Michael Polanyi argues that s c i e n t i f i c communities may be so isolated from each other by their epistemic vocabularies, their terministic screens, that argumentum ad hominem is the only source of argument available. In the absence of a common ground for other forms of persuasion, the opponent "w i l l be made to appear as thoroughly deluded, which in the heat of the battle w i l l easily come to imply that he was a fool, a crank, or a fraud. . . . In a clash of intellectual passions each side must inevitably attack the opponent's person" (151-52). This scenario is also implied by Kuhn's view of incommensurate s c i e n t i f i c communities, although Kuhn allows for more pos s i b i l i t y of communication in his discussion of "translation" between s c i e n t i f i c paradigms (202). Both Kuhn and Polanyi suggest that persuasion i s the way to resolve this dilemma; however, their belief in the near-completeness of the divisions between s c i e n t i f i c communities reduces their a b i l i t y to see rhetoric as a force of identification and rational interchange. Under these circumstances ethical argument, in the absence of a clearly articulated principle of logos, can indeed become cancerous, eating up a l l other forms of symbolic interchange. D. Brent 134 generated from the most general aspects of a person's character—the Aristotelian triad of intelligence, moral character and good w i l l . 3 6 There is no problem, except perhaps one of t o t a l l y context-free formal logic, to which these aspects of personality are not at least tangentially relevant. Ethos, then, takes i t s place beside pathos and logos as a source of evidence that the reader can use in deciding whether or not to accept the beliefs offered by a given set of texts. The precise balance between these three sources of evidence is necessarily dependent on many different factors: the personality of the reader, the type of text and the sort of appeals i t s author overtly attempts to make, and the discipline in which the transaction takes place. But as we have seen, no discipline, including the sciences, manifests a decision-making process that is completely free of logical, emotional and ethical evidence. 3.6. The Outline of a Rhetoric of Reading. We have now outlined a rhetoric of reading that has the following specific characteristics: 1. From a rhetorical point of view, the act of reading is a process of being persuaded to modify one's system of beliefs to accommodate those presented by the authors of the texts one reads. "Being 3 SApproaching the same problem from the perspective of discourse processing theory, Wyer arrives at an essentially similar set of factors that may affect the reception of a communication: the source's personal likableness, prestige, power, expertise, and impartiality (273). It is interesting that according to Wyer, expertise is by far the most influential of these characteristics. Perhaps this reflects the modern valuation of specialized knowledge rather than general "probity" (277). However, Sherif p a r t i a l l y dissents from Wyer's conclusion (300). D. Brent 135 persuaded" means accommodating an organized system of beliefs to take account of the new perspectives that are being offered. Being persuaded begins with evoking the meaning of the texts, a meaning that can legitimately d i f f e r from reader to reader. Meaning is evoked as a transaction between the text i t s e l f and the reader's personal repertoire of associations and knowledge, the rhetorical situation, and the shared verbal meanings collectively attached to the words of the text by the reader's linguistic community. The overarching guide in this process is the reader's attempt to impose a coherent and unified meaning on the text. } Being persuaded continues with a process of evaluating the meanings that one has evoked not according to intr i n s i c merit but according to s u i t a b i l i t y for inclusion in a system of beliefs. The degree to which a system of beliefs w i l l be accommodated to new perspectives depends on three factors: a. The degree to which the new perspectives lo g i c a l l y connect with the most general, deeply-held or important ideas that make up the system already in place (logos); b. The degree to which the new perspectives emotionally connect with values and attitudes that contribute to the organization of the reader's systems (pathos), and c. The degree to which the writer's entire structure of beliefs, attitudes and values—his "character"—is amenable to the reader's own (ethos). A l l of these factors are highly interdependent, as are the reader's reactions to a l l the texts with which she is confronted. No text can be said to have " i n t r i n s i c " persuasive power even with respect to a D. Brent 136 particular reader; rather, texts f i t into a system held in dynamic balance by the three modes of persuasion. We can say that this outline of the reading process is a "rhetoric" of reading not just because i t employs traditional rhetorical concepts but because i t treats the transaction between reader, writer, text and situation not as a passive uptake of information but as a persuasive transaction. This view guides the ways in which insights from other disciplines such as discourse processing theory and li t e r a r y theory can be used to enhance the general model. To the extent that these disciplines see reading as an active transaction rather than as passive uptake, their insights are germane to this inquiry. By amalgamating such suggestions and searching out commonalities among them, we have been able to suggest some of the ways in which the act of reading i s part of a much larger movement: the building of communal knowledge through rhetorical interchange. D. Brent 137 Toward a Rhetoric of Reading Chapter 4: The Rhetoric of Reading as a C r i t i c a l Technique 4.1. Rhetorical Criticism and the Unending Conversation We have now outlined a model of the reading process that combines the general conceptual structure offered by discourse processing theory with the insights of l i t e r a r y and rhetorical theory. The most important feature of this model is that i t allows us to see each step of the reading process, from the evocation of the virtual work to the f i n a l decision of whether or not to be persuaded by i t , as being controlled by connections between the writer's and the reader's frame of reference. The writer's frame of reference, her doxai, the schema that she has constructed to make sense of the world, are implied in the text she constructs. Her reader infers from the text not only the specific message that the writer intends to communicate but also those background assumptions, using his own schema as a means of disambiguating the language through which the writer's intentions are conveyed. The virtual work that the reader constructs in this manner is a bundle of propositions that he can examine logically, evaluating i t according to his own prior assumptions. However, the virtual work is much more than this. It also implies a bundle of attitudes and values, an entire perspective on the world and on social values. The reader's emotional response to these aspects of the virtual work also contributes to his judgement of i t . Finally, the writer's character as inferred from D. Brent 138 her text is a criterion upon which the reader can base a judgement. Thus the reader's decision of whether or not to modify his worldview on the basis of the perspectives encountered in reading is influenced by three factors: a logical, an emotional, and a social response to the work. This is the triad that traditional rhetoric labels logos, pathos, and ethos. Because of this emphasis on the use of pre-existing logical, emotional, and social belief structures as bridging mechanisms between writer and reader, the rhetorical model of reading i s indebted to the s p i r i t i f not the precise form of the classical enthymeme. In addition, the model describes the reading act as situated in context: both the construction and the judging of the virtual work depends on the writer's and reader's repertoire of knowledge and values, their goals in participating in the dialogue, and the place of the dialogue in the larger conversation in which i t is embedded. It is this combination of attributes, not just the use of terminology borrowed from the rhetorical tradition, that makes this model of the reading process a particularly rhetorical one. What practical applications does this model of reading have? One possible application is to use i t to inform a methodology of rhetorical criticism. Perhaps the classic statement of the aims and methods of rhetorical criticism i s found in Edwin Black's Rhetorical Criticism: A  Study in Method, a work that seeks to define what rhetorical criticism i s , what i t s benefits are, and how i t can most profitably be conducted. 1 ^Tt should be noted that one of Black's major theses in Rhetorical Criticism is that neo-Aristotelian criticism is a limited and largely obsolete methodology and must be superceded (see especially 131). Because so much use is made here of Aristotle's pisteis and his theory of the enthymeme, i t i s important to point out that Black's work is being D. Brent 139 Criticism in general, Black asserts, has the aim of understanding; i t focuses on texts and the people who produce and receive them, not for the immediate purpose of understanding the people and events involved as such—considerations that are the proper study of the biographer, the historian, the journalist, the p o l i t i c i a n — b u t for the wider goal of revealing the processes that underlie the discourse. "Criticism," argues Black, " i s a discipline that, through the investigation and appraisal of the a c t i v i t i e s and products of men, seeks as i t s end the understanding of man himself" (9). In the case of rhetorical criticism, the a c t i v i t y investigated i s that of persuading others; the products investigated are the texts through which this a c t i v i t y is carried out. 2 From the point of view that we have now developed, the a c t i v i t y of persuading others is part of the larger a c t i v i t y of building knowledge through social interchange. The sort of understanding that we w i l l obtain from examining particular instances of this process, then, w i l l be an understanding both of discourse processes—reading and writing—and of knowledge building. Since we have been examining rhetorical interchanges principally from the point of view of the reader, we w i l l naturally apply used as classic statement of the goals of rhetorical criticism, but not necessarily of i t s methods. There is a good deal of wisdom in Black's work, but his rejection of Aristotle i s fundamentally at odds with the stance adopted here. 2 I n addition, of course, the c r i t i c w i l l necessarily evaluate the processes that he views. Weaver reminds us that i t is the duty of those who would examine human phenomena to arrive at and to share judgements of value. The c r i t i c , despite Black's attempt to achieve a disinterested pose, is not exempt from this duty. Just as the efferent reader in the general case evaluates the propositions presented by the texts he views, so the c r i t i c , focussing on the texts as records of a process, w i l l inevitably reach judgements regarding better and worse, successful and unsuccessful instances of that process. This w i l l be the focus of chapter 5, which w i l l treat as normative the description of rhetorical reading developed in this chapter. D. Brent 140 the principles we have developed mainly to an understanding of how reading f i t s into the process; however, we w i l l not be looking at acts of reading in isolation, but rather at acts of reading as they occur in a larger process of epistemic interchange. We cannot simply ask, then, how a particular person succeeds in interpreting, evaluating, and deciding whether or not to believe this or that particular text. Rather, because a rhetoric of reading i s implied in a rhetoric of writing, we must look at the relationship between the consumption and the production of discourse—at the way knowledge i s passed on from one person to another through the paired a c t i v i t i e s of reading and writing, with this knowledge being continually modified a l l the while. This requirement leads us to focus not just on particular texts but on sequences of related texts by various writers, asking of them not just how the writers marshall the available means of persuasion in the case, but also how they have assembled the knowledge that they are attempting to convey—that i s , how they have read as well as how they write. This project could be conducted psychologically through a close examination of. the personalities, the biographies, and the psychologies of the individual participants in the conversation. Such a study would involve close interaction with the individuals themselves as subjects of surveys, interviews, possibly interventionist data collection techniques such as think-aloud protocols, and perhaps even controlled experiments. 3 3One example of such a study is Christina Hass and Linda Flower, "Rhetorical Reading Strategies and the Construction of Meaning." Hass and Flower analyze the think-aloud protocols of ten readers attempting to understand a highly complex text. Their conclusions closely parallel one of the conclusions of the present study: "Rhetorical strategies include not only a representation of discourse as discourse but as unique discourse with a real author, a specific purpose, and actual effects" (178). The implied author, then, plays a v i t a l role in the construction and c r i t i c a l evaluation of meaning (chapter 3, 127-29). D. Brent 141 This form of study would have great value, for i t would give us clues to the mental processes involved, clues that could not be inferred as reliably from texts alone. However, i t is also possible to conduct such a project with methods more commonly associated with rhetorical criticism rather than empirical social science. We can learn a great deal of what we want to know simply by examining texts through the lens of our rhetorical model of reading. In doing so, we w i l l be attempting to observe in action the processes that have been outlined in the previous chapters: we w i l l attempt to observe f i r s t the construction of a virtual work through the interaction of the text, the reader's repertoire and rhetorical situation, and second the evaluation of that work through the interaction of logical, emotional, and ethical proof. This form of investigation, though involving a narrower range of data than a full-scale psychological study, is in many ways closer to the question we really want to ask. We have established that texts form a rhetorical channel between human beings. When one human being wants to communicate with another, to persuade him, to modify in some way the representation of the world that he possesses, the text is his means of doing so. Our question is about how these texts work, not as entities in themselves, but as connecting strands in the rhetorical network of texts that comprises the epistemic conversation of Burke's metaphor. How, we want to know, do ideas get passed from text to text, from writer to writer? How do prior texts that a writer has read affect subsequent ones she produces? How does a writer's frame of reference, her repertoire of values and ways of seeing the world, manifest i t s e l f in the way she represents in the texts she produces the texts she has read? In short, since the texts are the major, often the only, source of information that D. Brent 142 the participants in the conversation have about each other, we should be able to apply our model productively to the conversation by examining the texts through which i t is conducted. This application of the rhetorical model, then, is not s t r i c t l y empirical, but rather c r i t i c a l . The difference is that we are not direc t l y attempting to prove or disprove an empirical theory by referring to phenomena i t claims to predict. Rather, we are attempting to validate a c r i t i c a l model by seeing i f i t helps us ask and answer useful and interesting questions about the rhetorical impact of texts. We w i l l therefore be engaging in a special form of rhetorical criticism, a form that might be termed "dialogic c r i t i c i s m . " 4 This form of criticism involves examining closely the texts involved in a particular rhetorical interchange. Its goal is to observe in action the processes of rhetorical interpretation, evaluation, and belief that have been outlined thus far, in concert with the more traditional processes of composing persuasive discourse. It may therefore be thought of as a species of rhetorical criticism with a highly truncated focus on the offices of arrangement and style and a greatly expanded emphasis on the office of invention. This office is redefined to include not just the traditional methods of creating arguments but also the larger epistemic processes by which writers generate the viewpoints that w i l l become 4Here "dialogic" i s used in a somewhat special sense to mean not just a conversation between two participants, as the word would seem to imply, but a conversation among a limited number of participants on a specific subject. This dialogue, of course, consists of many specific instances of two-party dialogue, for each act of reading is an event involving exactly two participants: the writer and the reader. But the use of "dialogue" to refer to a subset of a conversation is to be understood simply as a device that allows the term "conversation" to be reserved for the entire ongoing social interchange, the "unending conversation" of Burke's metaphor. D. Brent 143 arguments: their constructive reading and evaluation of other texts. The principal question that the rhetoric of reading should help us answer, then, w i l l be, "Why do the conversants interpret and become persuaded by each others' texts in the way they do?" s Using the rhetorical model of reading, we w i l l seek explanations of the judgements that writers make and record in their texts by looking for two broad classes of phenomena: f i r s t , the ways in which their interpretations of source texts are influenced by the rhetorical situation and by their repertoires of knowledge, assumptions, and beliefs; and second, the ways in which their judgements of these texts are influenced by the factors we have labelled logos, pathos, and ethos. It i s important to understand, however, that i t w i l l not always be possible or profitable to make absolute separations between the operations of these factors. Although there is a theoretical advantage in discussing the processes of interpretation and judgement and the operation of the three pisteis as i f they were separate, in practice they are inextricably intertwined. A reader is simultaneously influenced by her logical, emotional and ethical reactions to a text, and the process sSuch a process w i l l naturally generate speculative statements concerning the personalities, assumptions and mindsets of the writers involved. This is of course extremely dangerous ground; i t is impossible to characterize a human being in propria persona by examining his writings, especially a very limited sample of his writings. But this is not the aim of such statements. They are intended to establish only the assumptions and personality of the author as i t appears in the text. It i s therefore an attempt to reconstruct the author as he w i l l be constructed by the other conversants in the dialogue: as a persona, an ethos that seems to characterise regularities in attitude that can be inferred from the text—the being that Booth labels the "implied author." That this persona is not necessarily the writer himself should go without saying; but i t does represent the writer in the form that is most salient when we are conducting not a psychological investigation but an investigation of rhetorical interrelationships between people through texts. D. Brent 144 of judging i t affects her interpretation of i t in an endlessly recursive cycle. The rhetorical model of reading w i l l not so much allow us to separate these processes as i t w i l l remind us that at a l l points in our analysis we can profit by watching for ways in which the factors we have liste d work in concert. To i l l u s t r a t e this sort of textual analysis we w i l l use a dialogue consisting of a series of c r i t i c a l analyses of James Kinneavy's A Theory  of Discourse. A Theory of Discourse, f i r s t published in 19717 had a profound effect on the discipline of English studies. Kinneavy's goal is nothing less than to bring the paired a c t i v i t i e s of composition teaching and textual analysis out of what Kinneavy, borrowing a term from Thomas Kuhn's The Structure of Scientific Revolutions, c a l l s a "preparadigmatic period," a condition in which "there has not yet been erected a comprehensive system of the discipline which has received some general acceptance and which could serve as a framework for research, further speculation, innovation, even repudiation" (2). Kinneavy attempts this monumental task by synthesising a theory of discourse from a l l available sources, ranging from classical rhetoric to modern rhetoric, l i t e r a r y theory, information theory, and language philosophy. He selects as his guiding structure the venerable communications triangle, a view of communication that characterizes i t as a signal linking three points: encoder, decoder, and reality. The complex taxonomy that Kinneavy derives from this foundation is based on four categories or "aims"— expository, expressive, persuasive, and literary—according to which of the four aspects of the communications triangle is emphasized in the text. This system has not been accepted unconditionally—it has never attained the status of an encompassing paradigm for the discipline of D. Brent 145 English in the way that, for instance, Aristotle's formulations have become a paradigm for traditional rhetoric. However, the complexity and detail of the taxonomic system, the depth and weight of the scholarship on which i t is founded, and perhaps most importantly the provocative p o s s i b i l i t i e s for organizing a discipline that was (and to a large extent s t i l l is) having d i f f i c u l t y finding a structure with which to define i t s e l f , has made Kinneavy's theory a profound presence i n English studies. It i s this dominating presence that makes A Theory of Discourse a particularly useful focus for a dialogic analysis. It has spawned a large number of other studies that argue the merits both of the book i t s e l f and of the attitude to discourse that i t represents. Some of these studies have in turn spawned others that respond e x p l i c i t l y to the points of view articulated in earlier studies as well as to A Theory of  Discourse i t s e l f . The intertextual web that results, then, is particularly rich in the type of dialogic relationships between texts that a rhetorical model of reading is designed to account for. The aim of this investigation w i l l not be to analyze A Theory of  Discourse i t s e l f , but rather to use i t as a hub that holds together a particular subsection of the larger conversation. The intertextual web involved embraces far more than Kinneavy's book and the analyses that mention i t by name; i t embraces a l l texts that discuss the same general topic, the taxonomizing of discourse, as well as a l l texts that have contributed to forming the ideas of the conversants. If we were to endeavour to take into account a l l relevant texts, we would quickly exceed the boundaries of our a r t i f i c i a l l y circumscribed dialogue and be attempting to deal with every text that has ever touched on the subject D. Brent 146 of discourse. This example of dialogic criticism is therefore not intended to be comprehensive. Rather i t i s intended to highlight a limited number of texts which anchor strategic points in the dialogue and which further the goal of understanding how in general knowledge is built through such interactions. These texts are those which most obviously cluster about the selected focal point: those that e x p l i c i t l y take a stand on the merits of A Theory of Discourse. 6 Of these, we w i l l highlight in particular the following: 1. Jim W. Corder and Frank D'Angelo's early summaries of Kinneavy's work in the f i r s t edition of Tate's Teaching Composition: Ten  Bibliographic Essays. 2. Four full-length critiques of Kinneavy, two of them positive—Timothy Crusius' art i c l e s "A Brief Plea for a Paradigm and for Kinneavy as Paradigm" and "Thinking (and Rethinking) Kinneavy"—and two negative—C. H. Knoblauch's "Intentionality in the Writing Process" and Paul Hunter's "'that we have divided/In three our kingdom'." 3. Crusius' response to Hunter's a r t i c l e in the "Comment and Response" section of College English, and Hunter's rebuttal. 6 I n addition to the ones dealt with in detail here, many other texts either centrally or peripherally deal with Kinneavy's taxonomy. Some of the more obvious examples are Edward Corbett's review of A Theory of  Discourse in Freshman English News; John O'Banion's "A Theory of Discourse: A Retrospective"; Richard Fulkerson's "Kinneavy on Referential and Persuasive Discourse"; Walter Beale's A Pragmatic Theory of Rhetoric; and Richard Larson's "Classifying Discourse." D. Brent 147 4.2. The Influence of the Rhetorical Situation In chapter 2, we argued that the reader's construction and evaluation of a virtual work is affected by two factors in addition to the words of the text i t s e l f : the rhetorical situation and the reader's repertoire of pre-existing beliefs. The f i r s t of these two factors is particularly helpful in accounting for the differences between two of the earliest texts in the dialogue on Kinneavy: Jim W. Corder and Frank D'Angelo's chapters in the f i r s t edition of Tate's Teaching Composition: Ten  Bibliographic Essays. The rhetorical situation, as defined by Lloyd Bitzer, consists of the audience, the external constraints that govern decision (such as beliefs, documents, facts, traditions, and the l i k e ) , and the exigence, the imperfection in the current state of aff a i r s that gives rise to the need for discourse in order to set i t right. In terms of the general relation between the text and the audience, the rhetorical situation of these two texts is the same; both are part of a survey of approaches to teaching composition, a rhetorical situation in which the audience expects a brief summary and critique of a number of the most well-known texts in the f i e l d . However, the rhetorical situation also consists of a specific exigence or, as we have reinterpreted Bitzer's definition here, a specific question or set of questions that a particular reader/writer is attempting to answer. Although both treatments of Kinneavy address the same general audience under the same circumstances, each does so under the influence of a different rhetorical exigence, a different set of questions. This difference causes each author to evoke a different representation of Kinneavy's text. D. Brent 148 Corder's chapter is called "Rhetorical Analyses of Writing." It is therefore not surprising that Corder focuses on the aspects of Kinneavy's scheme that are most relevant to this topic. After a brief summary of Kinneavy's four aims of discourse, Corder summarizes the uses to which he feels Kinneavy's scheme may be put: "This scheme of organization does not provide a rhetorical analysis. It does, however, raise a number of questions and open a number of possibilities that generate rhetorical analysis" (227). He then goes on to suggest the types of questions that Kinneavy's scheme might prompt a student to ask about an essay. The view of Kinneavy's work that Corder evokes in this chapter, then, is shaped by Corder's goal: identifying useful theoretical sources for rhetorical analysis. This does not, of course, mean that this i s the only view that Corder takes or ever could take of Kinneavy's text. As Rosenblatt points out, the evoked work is not an unvarying entity; i t is an "event in time" produced by a transaction between reader and text, a transaction that is transient and w i l l result in a different evoked work every time the text is read. Other rhetorical situations could result in the evocation of a quite different work. However, we are interested not in a l l the works that Corder might evoke under the influence of Kinneavy's text, but in the particular work evoked under the influence of this particular rhetorical situation. The effect of situation on Corder's view of Kinneavy, however, goes deeper than this rather superficial focusing of attention. Although one of Kinneavy's major innovations is the division of discourse by aim rather than by mode, Corder does not mention this aspect of A Theory of  Discourse. Rather, he highlights another key point in Kinneavy's text, his use of the communications triangle as the foundation of his D. Brent 149 classificatory scheme. This emphasis reflects the structure of Corder's discussion as a whole:-he uses A Theory of Discourse to il l u s t r a t e a governing assumption that "A conception of rhetoric is an enabling base for rhetorical analysis" (226). We can see this assumption illustrated in the c r i t e r i a that Corder uses when he compares Kinneavy with other theorists such as Richard Ohmann and Michael Halloran. Each such comparison is based on the way the theorist in question provides a starting point for rhetorical analysis. Corder thus identifies the salient points of Kinneavy's work—the points about which he constructs his evocation of his own virtual work—with reference to the aspects of i t that are most important with respect to the question he wishes to ask of i t : how i t can provide a basis for rhetorical analysis. Corder, then, does not simply analyze a constant and unchanging "text"; he analyses a virtual work that he has evoked for himself. During the course of this evocation, his wandering viewpoint has been drawn to those parts of the text upon which the rhetorical situation has, for him, conferred presence. The virtual work that Corder has evoked is not timeless but a unique, personal event in time, an event that represents a coming-1 together of Kinneavy's text and a particular rhetorical context. When we compare Corder's discussion with Frank D'Angelo's chapter on "Modes of Discourse," we can see more clearly the role that the uniqueness of the virtual work plays in the process of rhetorical reading. D'Angelo's discussion shares with Corder's a generally positive tone and an attention to potential uses of the theory, as one would expect from a chapter that shares the same general rhetorical context. However, his summary and discussion of Kinneavy differs from Corder's in some revealing ways. D. Brent 150 In this discussion, D'Angelo repeatedly uses the term "mode" rather than ''aim", despite Kinneavy's insistence that aims are different from modes and that A Theory of Discourse concerns the former (Theory 38). Is this r e a l l y a misreading of Kinneavy, and i f so, why does D'Angelo make this simple but important "error" of terminology? Perhaps D'Angelo shifts Kinneavy's terms partly as a matter of convenience; i f his subject is "modes of discourse," i t is certainly t i d i e r to use the term "mode" throughout. But this symmetrical terminology reflects a deeper symmetry. D'Angelo's chapter, based on a different classification system from Corder's, compares texts not on the basis of their usefulness as starting points for rhetorical analysis but on the basis of the ways in which they divide forms of discourse. D'Angelo's terminology f a c i l i t a t e s (or reflects, or creates) comparisons, not between Kinneavy and writers such as Ohmann and Halloran, but between Kinneavy and writers such as James Moffett and Alexander Bain. Moffett is the originator of a taxonomic system based on psychological distance between'the reader and the writer, and Bain, a nineteenth-century compositionist, is credited with inventing the "modes" of discourse (narration, description, exposition, and argumentation); both are frequently cited as examples of writers particularly interested in categorising discourse. 7 By comparing Kinneavy's work with that of others who are primarily remembered for their taxonomies, D'Angelo focuses attention more on Kinneavy's categories themselves than on the communicative interactions that give rise to them. In short, then, the views of Kinneavy's text presented by Corder and D'Angelo are influenced by the other texts with which they 7See Moffett, Teaching the Universe of Discourse, and Bain, English  Composition and Rhetoric. D. Brent 151 wish to compare i t , a decision in turn affected by the type of question they wish to answer ("What are useful starting points for rhetorical analysis?" as opposed to "What are useful ways of categorizing modes of discourse?"). Thus we see illustrated here an important point concerning the way people read: the works they evoke are influenced not just by the words of the text but by the place of that text in the larger epistemic conversation that the reader is immersed in, a conversation that generates certain questions about the text and places i t in relationship to other texts partly on the basis of those questions. Because of this influence of a broadly defined rhetorical situation, we can see that the act of reading is interpenetrated by rhetoric even when we consider only the interpretation of texts, without explici t regard to the readers' judgements of whether texts are worthy of inclusion in their own belief systems. 4.3. The Influence of Values and Assumptions The differences between Corder's and D'Angelo's brief discussions of Kinneavy can be accounted for mainly by reference to the differences in the rhetorical situations in which they wrote. When we move to later texts in the dialogue, we begin to encounter more extended texts that take a much stronger position for and against A Theory of Discourse. For instance, Timothy W. Crusius' two art i c l e s , "A Brief Plea for a Paradigm and for Kinneavy as Paradigm" and "Thinking (and Rethinking) Kinneavy," and C. H. Knoblauch's "Intentionality in the Writing Process: A Case Study," a l l originate in a rhetorical situation somewhat different from that of Corder's and D'Angelo's pieces. Corder and D'Angelo are both D. Brent 152 contributing to a survey of approaches to composition; this rhetorical situation encourages them to touch b r i e f l y on Kinneavy's work as one of many examples of methodologies that could prove useful to composition teachers. Crusius and Knoblauch, however, are not simply surveying methodologies; they are searching for a theoretical foundation for composition teaching. This rhetorical situation leads them to take definite stands on a much more d i f f i c u l t and highly-charged question: should the view of language and of composition teaching that underlies A Theory of Discourse be adopted as the paradigm for the entire discipline of composition? As a result, these texts evidence intense and complex differences in construal and judgement. To account for these texts we must examine more closely the ways in which A Theory of Discourse connects or f a i l s to connect with these writers' pre-existing values and assumptions—their doxai. By relating the conclusions expressed in these art i c l e s to the assumptions and values that we can infer from Crusius's and Knoblauch's texts, we can account for the significant differences between the two writers' manner of reading and evaluating Kinneavy's theories. In "A Brief Plea," Crusius e x p l i c i t l y states his argument in the form of a syllogism: i) As teachers of discourse, we must work from a theory of discourse—or else resign ourselves to ad-hoc measures and admit that we have no systematic ju s t i f i c a t i o n for doing anything. i i ) While there are no complete or adequate theories of discourse at present, Kinneavy's is the f u l l e s t developed. i i i ) Therefore, we should adopt Kinneavy's theory as the D. Brent 153 paradigm for our f i e l d , thinking with his concepts, developing his categories, extending his system in principled ways, un t i l i t either becomes adequate and complete or reveals innate shortcomings that c a l l for a new, differently conceived set of ideas. (1) The major and minor premises of this syllogism represent the most explicit assumptions on which the conclusion i s based. However, i t is possible to uncover more fundamental assumptions that l i e behind the ones that Crusius states. To use Toulmin's terminology, these assumptions form the backing for the premises of Crusius's syllogism. 8 Crusius' f i r s t premise, that we must work not just from theories but from a specific theory, is based on a doxa that i s evident throughout the entire discussion of Kinneavy in "Plea": the responsibility of the composition theorist to reason about her discipline. Not to seek an adequate theory of discourse "amounts to an evasion of our intellectual responsibilities. . . . Our duty is to reason about discourse" ("Brief Plea" 1). Such reasoning is essential to the field's "philosophical standing" and i t s status as an art rather than a knack. Behind these statements we can detect not only a powerful sense of professionalism, of the need to defend rhetoric as a discipline from directions of development that would weaken i t s intellectual stance; we can detect also an intense belief in the human being as a rational animal. This belief in the responsibility and duty of the teacher to search out and pass on true propositions, combined with a faith in human rationality as the proper tool for doing so, is the habit of mind, the 8See Toulmin, 94-113, passim, and chapter 3 of the present study, 102-04. D. Brent 154 cluster of fundamental assumptions, that Richard Young c a l l s " c l a s s i c a l . " It i s a faith in "the knowledge necessary for producing preconceived results by conscious, directed action" ("Arts, Crafts, Gifts and Knacks" 56). With i t is combined a strong desire for the sort of order that comes from synthesis. Its alternative, pluralism, is "too easy," too l i k e l y to s l i p into "facile eclecticism," too "chaotic and confusing" ("Plea" 1-2). The values revealed here suggest not just why Crusius believes in his major premise, the need for a single paradigm, but also why he believes in his minor premise, that Kinneavy's theory is the most l i k e l y candidate for that paradigm. Kinneavy's theory is profoundly synthetic, attempting to assemble a single theory based on the insights of hundreds of authorities on rhetoric, communication and philosophy. It is grounded in the "philosophical principles" that Crusius sees as necessary in order for a discipline to have coherence. In more practical terms, i t provides the "more or less fixed notion of what constitutes adult-level performance" that satisfies Crusius' desire for teaching methods founded on rational principles rather than on mere experience and exposure (2). From the point of view of a rhetoric of reading, then, we can describe the concordance of attitudes that we see here as an enthymemic connection between two points of view. The heart of the enthymeme is ,its use of shared doxai as warrants or major premises (spoken or unspoken) that render certain propositions acceptable to the audience. From the perspective of a productive rhetoric, the enthymeme is a device deliberately constructed by the rhetor in order to make his position acceptable. As we have redefined i t from the reader's perspective, however, the enthymeme is a term that identifies the rhetorical D. Brent 155 connections that after the fact can be seen to have produced a rhetorical judgement. The doxa that serves this function in this case is Crusius' belief in classical approaches to knowledge and to pedagogy: to order, rationality, and a unified point of view. Because Crusius perceives Kinneavy's paradigm as being constructed on these principles, he readily accepts i t as a good candidate for a paradigm for composition teaching. This set of attitudes forms a logical bridge, a point emphasized by the fact that Crusius opens his a r t i c l e with a syllogism. But i t also functions emotionally and ethically. Crusius obviously has a profound admiration both for Kinneavy's theories and for the sort of man implied by those theories, an admiration that goes deeper than a purely logical evaluation of those theories in the abstract. The Kinneavy that he constructs—synthetic, comprehensive, philosophically grounded, and p r a c t i c a l — f i t s these c r i t e r i a well enough that the advantages of the theory easily outweigh the limitations that Crusius freely admits exist. Thus Crusius' decision to be persuaded by Kinneavy's theories i s based on an identification between two sets of doxai: Crusius is persuaded because he values what Kinneavy values. In short, Crusius' judgement of Kinneavy's text has been guided by an interaction of logical and emotional response to a complex of ideas, attitudes and character—that i s , by an enthymeme based on an interaction of logos, pathos and ethos. Or, to s h i f t into Burke's terminology without shifting the fundamental insight involved, Crusius has been influenced by the threads of identification that he has identified as linking himself and Kinneavy. In Crusius' later a r t i c l e , "Thinking (and Rethinking) Kinneavy," we can see how both the factors that we have so far examined—the rhetorical situation and the reader's repertoire of knowledge and assumptions— D. Brent 156 interact in highly complex and recursive ways. In this a r t i c l e , Crusius evidences a similar general attitude to Kinneavy—that Kinneavy's theory ought to be used as the base for a comprehensive paradigm of discourse. 9 However, Crusius' repertoire of values influences not only his judgement of Kinneavy but also the very questions he sets out to ask. Crusius admits that Kinneavy's theory as i t stands is not an adequate pedagogical model because Kinneavy's abstract aims do not connect well with the ongoing experience of writing. Writers write not just to f u l f i l l general aims but to f u l f i l l immediate purposes, to persuade, to express, and to inform a particular somebody of a particular something in a particular context. However, Crusius defends Kinneavy's failure to distinguish aim and purpose: "In sum, on aim and purpose: Kinneavy does not distinguish them, nor should he have done so for the 9 I t i s interesting to note that this a r t i c l e is heavily influenced not just by the work of other scholars in the f i e l d , but by Crusius' own earlier text. Instead of concentrating on the positive aspects of Kinneavy's theory in order to argue that i t should be adopted as a paradigm, Crusius takes his earlier position as a given: "The following comments assume that Kinneavy should be our paradigm for teaching writing, as I have contended elsewhere" ("Thinking" 120). Thus "Thinking" forms part of the assumptions from which he works and which he no longer needs to argue in de t a i l . Partly because of the influence of this prior text, the emphasis in "Thinking" is quite different from that in "Plea." Freed from the need to argue that Kinneavy provides a good basis from which to work, Crusius can argue at length what was only a concession in the former a r t i c l e , the position that Kinneavy's theory needs extension. In fact, he is freed to make a number of individually quite negative comments on Kinneavy, secure that these negative comments are embedded in a larger discourse that has established a basically approving tone toward the subject. In a sense this i s not so much a matter of the kind of invention that we have been discussing as i t is a matter of audience adaptation: here we do not see Crusius developing his ideas on the basis of what he has read but rather accommodating his text to what he can assume his audience has read. However, the observation illustrates the general point that we have been making throughout this inquiry: texts are not just individual attempts to persuade an audience of a particular point of view, but rather parts of a complex network of textual interdependency. D. Brent 157 purpose of a general theory of discourse" ("Thinking" i24). For Crusius, Kinneavy's theory is adequate for the purposes that Kinneavy originally intended and only needs extension to become a dynamic theory with clear pedagogical application. Crusius's guiding question thus becomes "how exactly can we merge them [aim and purpose], taking advantage of their separate contributions, while getting them to work together?" (125) Here we see once more that the factors that influence the evocation and judgement of a text are not independent. In chapter 2, although we treated the rhetorical situation and the writer's personal assumptions and beliefs as two distinct factors in the evocation of a text, we also insisted that this treatment of them was a theoretical convenience only and that actually the two processes are inextricably intertwined (69n). Here we see this assertion confirmed: a v i t a l component of the rhetorical situation, the reader's guiding questions, is i t s e l f influenced by the reader's doxai. Crusius's synthesizing mentality leads him to ask how Kinneavy' concept of aim may be combined with the concept of immediate purpose; as a result, he highlights those parts of Kinneavy's text that are most amenable to such treatment, and having found them, judges Kinneavy's theories to be in the main worthy of qualified acceptance. Rhetorical reading is thus shown to be a great interconnected web of evocation and judgement driven by perceived logical, emotional, and ethical connections between the reader's and the writer's doxai—that i s , by our enlarged conception of the enthymeme. The processes of evocation and judgement that we have seen at work in Crusius's texts are thrown into sharp r e l i e f by comparison with another text that makes many of the same claims about the relation of aim and purpose but reaches diametrically opposed conclusions. In D. Brent 158 "Intentionality and the Writing Process," C. H. Knoblauch, like Crusius, points out the difference between generic "aims" as described by Kinneavy and the immediate rhetorical purpose that actually drives a given piece of discourse. However, he castigates rather than defends Kinneavy's failure to distinguish aim and purpose, and uses the pejorative term "blurring" to describe this failure. For Knoblauch, Kinneavy goes beyond f a i l i n g to distinguish aim and purpose: Kinneavy e x p l i c i t l y rejects the notion of operational purpose by invoking the intentional and affective fallacies, in which the "actual" intent of a discourse i s mistakenly assumed to be equivalent to the intent of i t s author or i t s impact on some intended reader. He argues that neither encoder nor decoder can re l i a b l y characterize actual intent, so that " i t seems better to find the aim which is embodied in the text i t s e l f . " (155) As a result, Knoblauch does not entertain the idea that Kinneavy's theory can be adapted to serve the needs of composition pedagogy. The entire argument of "Intentionality" i s that operational purpose and generalized aim are opposite notions and that one or the other (for Knoblauch, certainly, the former) must be chosen as an operating assumption for composition pedagogy. What is particularly interesting about these two accounts is that both writers cit e a very similar kind of inductive evidence. Knoblauch's a r t i c l e , subtitled "A Case Study," is founded on the author's practical experience of attempting to improve the writing of business executives by helping them c l a r i f y the purpose of their texts. During this enterprise, Knoblauch reports, i t was not general aim in Kinneavy's sense but D. Brent 159 "operational purpose"—that i s , a purpose tied to the specific rhetorical context of each piece of writing—that proved most helpful. Crusius admits more or less the same point, also based on teaching experience, but goes on to suggest ways in which Kinneavy's aims can be made to serve operational purpose rather than substitute for i t . Thus these texts pose a particularly interesting question for the rhetorical model of reading to answer: why do Crusius and Knoblauch reach such different conclusions based on essentially similar inductive evidence? The answer l i e s , again, in the doxai that provide the material with which the enthymemic process constructs—or in this case, f a i l s to construct—logical, emotional and ethical identifications between people. We have already noted that one of the most important doxai that underlies Crusius' evocation and judgement of Kinneavy is a powerful urge to synthesize. Knoblauch also admires synthesis, but he has a different notion of what i t means to synthesize. We can c l a r i f y his stand on the matter by digressing b r i e f l y to look at texts that stand outside the main line of the dialogue we are investigating here but which provide evidence for the doxai with which Knoblauch is operating. In Rhetorical Traditions and the Teaching of  Writing, co-authored with L i l Brannon, Knoblauch sets s t r i c t limits on the extent to which synthesis is profitable: The process of merger or bringing together is not always as well-suited to intellectual progress as i t is to p o l i t i c a l negotiation . . . . Intellectual consolidation is possible whenever concepts or problems can be interrelated within a D. Brent 160 single line of reasoning. But what i f two lines of reasoning oppose each other, and what i f evidence supporting one i s stronger than evidence supporting the other? (17) Where Crusius seeks every opportunity to synthesize, Knoblauch prefers to make distinctions and seek evidence in support of one point of view or another. In the case of Kinneavy's theories, he clearly believes that the evidence supporting an operational view of purpose is stronger than the evidence supporting a taxonomic view of aim, and therefore rejects the latter. This difference about what i t means to synthesize might be classed as a highly general doxa, a perspective on the world that would permeate any inquiry that either writer were to engage in, regardless of f i e l d . It i s , in short, the sort of doxa that might, in a classically-inspired c l a s s i f i c a t i o n system, be classed as a general topos ("analysis versus synthesis," perhaps). However, the two writers' views of Kinneavy's work are also affected by certain more special topoi, or as Toulmin would phrase i t , field-dependent warrants, such as their respective positions on pedagogical method. Crusius suggests that general discussions of aim can help students c l a r i f y the specific purpose of a given writing project. In other words, his rationalist viewpoint results in a belief that students can profitably transfer knowledge from an abstract model to the particular case. Knoblauch disagrees. In "Modern Composition Theory and the Rhetorical Tradition," an a r t i c l e published in the same year as "Intentionality," Knoblauch states that "No amount of theorizing about rhetoric, or grammar, or logic, or style w i l l change writers' habits of composing as rel i a b l y as i t improves their store of irrelevant precepts" (3). In Rhetorical Traditions. Knoblauch c l a r i f i e s the difference D. Brent 161 between theory for the teacher and theory for the student: Teachers familiar with the abstractions can better understand the nature of composing and therefore the processes their students are striving to control. As a result, they can more surely devise classroom a c t i v i t i e s to promote growth and can more readily define their roles as f a c i l i t a t o r s of growth. But what they know about modern rhetoric only informs their teaching; i t doesn't constitute class business. (102) Thus Knoblauch sets himself in opposition to the rationalist position. He does not discount the reason as an important component of the human being: the writing workshops that he and L i l Brannon recommend in later chapters of Rhetorical Traditions are surely examples of reasoned negotiation. However, he prefers the reason to act more or less unconsciously in a setting of collaborative social invention. For Knoblauch, direct address to the reason in the classical sense of specific discussion of compositional precepts i s not only inadequate but, because i t focuses conscious attention on mental operations that are best conducted t a c i t l y , positively harmful as a pedagogical tool. According to this perspective, no synthesis such as Crusius proposes is possible, for i t would require bringing together two pedagogical stances that are opposed, not complementary. We can therefore explain Knoblauch's rejection of Kinneavy on the basis of two related assumptions—that some lines of reasoning are incompatible (a general topos), and that the social inventional line i s preferable to the classical (a special topos specific to the discipline D. Brent 162 of pedagogy, particularly composition pedagogy). 1 0 Here we see the operation of a primarily logical form of judgement in which a reader (Knoblauch) rejects the propositions contained in a text (A Theory of  Discourse) because the dividing doxai outweigh any identifying doxai. However, the rhetorical model of reading reminds us that there are also emotional and ethical sources of judgement bound up with any reading act. It also reminds us that differences of judgement are partly based on different evocations of the virtual work and i t s implied author as well as on different evaluations of the work. These factors enter into an account of Crusius' and Knoblauch' texts, for the differences between the two accounts extend beyond a different evaluation of Kinneavy's text to a different construal of i t and of Kinneavy himself. It is these construals, particularly the latter, that in this case most clearly activate the emotional and ethical aspects of judgement. Recall that Crusius defends Kinneavy's failure to address operational purpose on the grounds that i t is not germane to Kinneavy's inquiry; Knoblauch, however, portrays a Kinneavy who actively rejects rather than simply f a i l s to emphasize operational purpose. Knoblauch's Kinneavy operates out of New C r i t i c a l assumptions which for Knoblauch are obsolete and which explain Kinneavy's interest in classifying texts as isolated entities rather than in the a c t i v i t i e s performed by writers in l o 0 f course, we could also examine the reasons for Knoblauch's belief in social inventional pedagogy, a belief that i t s e l f must inevitably be conditioned by the texts he has encountered. Like the conversation of humanity i t s e l f , dialogic criticism is capable of a complexity that approaches i n f i n i t e regress. The decision as to where to stop such an analysis must be made not on theoretical but on rhetorical grounds: one stops when one's account is satisfactory for the purpose intended. For the purpose intended h e r e — i l l u s t r a t i o n of the ways in which dialogic criticism can account for rhetorical interrelationships between texts—we need not peel Knoblauch's beliefs down any more layers. D. Brent 163 their construction ("Intentionality" 154-55). Crusius, who is not as concerned that Kinneavy neglects purpose, seeks no such explanation and does not include New C r i t i c a l assumptions in his characterization of Kinneavy. If Crusius does indeed have trouble accepting New C r i t i c a l assumptions, as one might expect from a writer working in an age whose c r i t i c a l climate i s firmly post-New-Critical, he does not extend that distrust to a negative construal of Kinneavy. What we see here i s not so much a direct cause-and-effect sequence as a complex interaction. It makes sense to say that Crusius and Knoblauch see Kinneavy differently because they like and dislike his theory respectively; that i s , that Knoblauch's di s l i k e of Kinneavy's conclusions gives greater presence to the aspects of the author that Knoblauch finds negative, thereby increasing the support for Knoblauch's judgement and helping to keep his view of Kinneavy consistent and coherent. However, i t makes equal sense to say the opposite: that Crusius's and Knoblauch's judgement of Kinneavy's theories do not just give rise to but are also supported by their emotional and ethical reactions to ah implied author who appears to one reader as a tremendous synthetic intelligence and to the other as a scholar who has been duped by his outworn New C r i t i c a l perspectives. The processes of construal and judgement, of logos,, pathos and ethos, are not linear but recursive, a l l working in concert as part of a complex enthymemic process to result in different readers' arriving at highly varied conclusions about the same text. Though the rhetorical model of reading predicts that a l l of these processes w i l l function to some extent in a l l texts, i t often happens that one or another of them w i l l be especially highlighted in a D. Brent 164 particular text. A case in point is Paul Hunter's a r t i c l e "'that we have divided/In three our kingdom': The Communication Triangle and A Theory of Discourse." This text is particularly interesting for the way in which Hunter extends his appraisal of Kinneavy's work to a highly conscious and deliberate attempt to characterize Kinneavy as a man. In doing so he not only foregrounds ethos, a mode of judgement that was more in the background in the work of Crusius and Knoblauch, but also highlights pathos as a mode of judgement through the way he reveals emotional responses to the values implicit in the Kinneavy he evokes. Hunter announces that he w i l l base his critique of Kinneavy primarily on what we might characterize as an argument from logos, arguing that the basis of Kinneavy's theories, the communication triangle, is more appropriate to speaking-listening than to writing-reading processes (280)."L:L However, a large portion of the a r t i c l e in fact comes to be dominated by a line of argument that Hunter c a l l s "ancillary": the argument that Kinneavy's theory, and by extension Kinneavy himself, is "essentially moralist" (280). This argument involves a construction of Kinneavy's character—not of his character in the sense of whether he beats his wife or tries to avoid paying his taxes, but rather in the sense of what we might c a l l his "textual " T h i s type of argument includes the strategy that Perelman and Olbrechts-Tyteca c a l l "dissociation of concepts," a technique in which what is originally assumed to be a single concept is resolved into two distinct concepts which can then be dealt with separately (411-15). In the case of Hunter's argument, the concept of "communication," treated by Kinneavy as a single concept with oral and written subvariants, is resolved into two related but distinct concepts: speaking-listening and writing-reading. This sort of argument, which Perelman and Olbrechts-Tyteca claim "brings about a more or less profound change in the conceptual data that are used as the basis of argument" (412), is an example of an argument from logos that addresses the rational faculties without e x p l i c i t l y using a format derived from the syllogism; D. Brent 165 character," the sum of the theories he espouses and the underlying attitudes and values implied by those theories. 1 2 This assessment of Kinneavy's textual character becomes Hunter's reason for refusing to accept Kinneavy's conclusions; in turn, i t becomes used in the productive aspect of Hunter's rhetoric, as an argument for Hunter's readers to reject Kinneavy's theories as well. Like Knoblauch, Hunter finds Kinneavy deeply entrenched in New C r i t i c a l assumptions, and opens his a r t i c l e with an assertion to that effect. However, he greatly broadens that i n i t i a l assertion to probe below the New C r i t i c a l attitude to discourse and expose what he sees as the underlying reason for Kinneavy's attraction to New C r i t i c a l perspectives: a belief in s c i e n t i f i c objectivity and a p o s i t i v i s t i c attitude to truth. For instance, Hunter traces Kinneavy's distrust of persuasive rhetoric to this p o s i t i v i s t i c outlook: "What we see in Kinneavy's examples . . . is the type of interpretation of texts produced by one who believes that logic and rhetoric are fundamentally at odds: logic concerned with truth, rhetoric concerned with manipulation" (281). 1 2 I n his response to Crusius' comment on "'that we have divided'" (to be dealt with in more detail later in this chapter), Hunter recognizes this distinction himself: "My complaint is with the text, not with the scholar who wrote i t . . . [It] should not be read as an attack on the man, one of the most helpful and influential teachers of my undergraduate years" (221). Hunter's view of Kinneavy "the man," that i s , Kinneavy as known from personal acquaintance, satisfies the traditional definition of positive ethos: Kinneavy appears as a person of intelligence and good w i l l , well suited to be a colleague and mentor. However, for the purpose of deciding whether to accept or reject a theoretical perspective, the more important ethos is that of the textual Kinneavy, who emerges, for Hunter, as a moralizing, intolerant po s i t i v i s t . This distinction underlines Aristotle's point that the ethos projected by the speech is more important to the reception of the speech than is any impression of character gained from prior acquaintance. It also reinforces the fact that for the purposes of a rhetoric of reading, the notion of character must be expanded to include the system of values and attitudes that a writer projects in his work (see chapter 3, 127-30). D. Brent 166 Because of Kinneavy's tendency to see the world in terms of this dichotomy, a dichotomy that Kinneavy, Hunter argues, treats as an objective fact, Kinneavy insists on devaluing persuasive communication and privileging s c i e n t i f i c discourse. In the second section of the a r t i c l e , the effect of ethos is further deepened and blended with a form of judgement by pathos, seen in Hunter's emotive reaction both to Kinneavy's theories themselves and to his textual character. From characterizing Kinneavy's theories as " p o s i t i v i s t i c " and "scientific,"" Hunter progresses to more pejorative terminology. He accuses Kinneavy not just of monism but of an "intolerant monism" (my emphasis), which he contrasts with an "authentic pluralism" (283). Chief among Hunter's evidence is Kinneavy's remark that those who espouse forms of a r t i s t i c interpretation other than the one he labels "objective" are prostituting art ("Divided" 280). From this remark and from Kinneavy's suggestion that people can be categorized by their ways of thinking ("Divided" 285), Hunter infers that Kinneavy betrays an "implicit categorization of people along the lines of Plato's Republic" (286). This tendency to categorize, combined with what Hunter sees as his " p o s i t i v i s t i c " belief in a single objective truth rather than a r e l a t i v i s t i c or p l u r a l i s t i c stance, "could lead to a perpetuation—even to a j u s t i f i c a t i o n — o f social stereotyping" (286). This is a stance that, for Hunter, " i s not a cure; i t is poison" (286). From the point of view of standard rhetorical criticism, Hunter's increasing swing to heavily emotional terms such as "intolerant" and "poison" can be identified as an important persuasive strategy. Certainly the identification of Kinneavy with concepts such as social stereotyping can be seen from the perspective of Ciceronian strategic D. Brent 167 rhetoric; writing as though he were an advocate for the prosecution, Hunter associates the accused with qualities that he knows his audience (liberal educated Americans) w i l l consider base, reinforced with emotional terms such as "poison." His use of this strategy is particularly expli c i t when, after characterizing Kinneavy as guilty of social stereotyping, he asks "how closely we as teachers of writing want to be identified with that" (283). Clearly he is not only recognizing the doxai of his audience—a deeply ingrained distrust of intolerance and stereotyping—but also using the American emotional response to raci a l discrimination by tangentially associating his opponent with this ultimate badge of ignoble character. Thus he is using pathos as an important mode of rhetorical proof. Yet our rhetorical model of reading suggests that these strategies are not only Hunter's attempts to persuade his readers, but also evidence for the ways in which Hunter himself has been persuaded, that i s , for pathos as a mode of judgement as well as pathos as a mode of proof. In this case, pathos is inseparably wedded to ethos, for Hunter's emotional reaction is not just to Kinneavy's theories but to Kinneavy himself. Clearly he sees Kinneavy's theoretical stance not as a set of separate theories but as an interdependent structure of beliefs that forms a coherent whole, as predicted by the rhetorical model of reading developed in chapter 3 (131-32). He is thus encouraged not only to consider some of Kinneavy's ideas as being predictable from others, but also to extrapolate from the beliefs that are manifest in the text others which are not, such as a tendency toward social stereotyping. Most important, this perception that theories and character are interdependent encourages Hunter to reject Kinneavy's entire structure of beliefs on the basis of D. Brent 168 its incompatibility, both logical and emotional, with doxai that Hunter holds with particularly deep conviction. Hunter's a r t i c l e not only shows us that pathos is an important factor in judgement, but also adds to our understanding of how i t functions. Terms such as "'monism,"' "moralizing" and "stereotyping" (negative) and "pluralism" (positive), according to the rhetorical model of reading, are not merely persuasive devices, though they certainly are that as well. 3- 3 They are also important mechanisms of pathetic judgement. They are "terministic screens" in Kenneth Burke's sense, terminologies that direct the attention in certain ways rather than others, that encapsulate certain attitudes to r e a l i t y — t h a t are, in short, indispensible tools with which to think and to f e e l . 1 4 In Hunter's text, we can see how Hunter teases out the assumptions that he feels underlie Kinneavy's approach to discourse, resolves them into a few 1 3 F o r instance, we may view this technique through the lens of Richard Weaver's concept of "god-terms" and "devil-terms." In "Ultimate Terms in Contemporary Rhetoric," Weaver points out that in every culture there exist terms of extremely high emotional value, both positive and negative. If a rhetor wishes to value or devalue a concept, much of the job is done i f he can associate i t with one or more of these terms. Hunter uses one of Weaver's devil-terms when he associates Kinneavy with r a c i a l prejudice. Another term with which he associates Kinneavy, "science," is listed by Weaver as a god-term but by the time of Hunter's a r t i c l e i t has, at least for the humanities-oriented readers of College  English, s l i d so far from i t s privileged status i t has become another devil-term. 1 4See Kenneth Burke's essay "Terministic Screens" in Language as  Symbolic Action: Not only does the nature of our terms affect the nature of our observations, in the sense that the terms direct the attention to one f i e l d rather than to another. Also, many of the  "observations" are but implications of the particular  terminology in terms of which the observations are made. In brief, much that we take as observations about "reality" may be but the spinning out of p o s s i b i l i t i e s implicit in our particular choice of terms. (46) D. Brent 169 key terms, and tests them by considering the values attached to those terms. It is because the values he attaches to "monism," "stereotyping," and "moralising" are a l l negative that he finds Kinneavy's theory of discourse wanting. Far from being a faci l e means of avoiding thought, such terms appear in this light as part of a process of thought, as a means of seeking what Booth would c a l l "good reasons" for holding a theory by testing i t not just against "reality" but also against the values implicit in i t . Key terms such as these, then, are important to the pathetic mode of judgement because of the ways in which they can summarize and juxtapose for comparison entire clusters of emotional a t t i t u d e s . 1 3 These four texts, then, reveal how writers' beliefs and values, their doxai, operate as part of an intertwined rational, emotional and ethical system of judgement. Hunter's evaluation of Kinneavy, from the perspective of our rhetoric of reading, is not simply a logical judgement. Like the evaluations of Crusius and Knoblauch, i t proceeds from an ethical view of a human being and his theories as a bundle of attitudes, attitudes which can be deduced from his texts and compared with one's own emotionally as well as rationally to produce good reasons for judgement. It i s , then, at least in part an argumentum ad hominem, a logical fallacy, but i t is not a rhetorical fallacy: the source of an l sNote that, just as Booth insists that there are both good and bad reasons for judgement, so there are god and devil terms which are philosophically grounded and ones that are perverse or sinister. The trick in using terms properly, according to Weaver, is not to avoid rhetorically loaded terms but to make sure that they are rhetorically loaded in some rational and justifiable manner ("Ultimate Terms" 111). If one's terms and the emotional response that one attaches to them are truly thoughtful, grounded in a d i a l e c t i c a l attempt to apprehend r e a l i t y and not just to concur with what is going on around one, then they can be the tools of genuine rhetorical inquiry. D. Brent 170 argument is a v i t a l part of a rhetorical decision to accept or reject a thesis presented for our assent. 4.4. Further Turns in the Conversation: Responses and Rebuttals Thus far this inquiry has been limited to a comparative analysis of rhetorical reading; that i s , we have been comparing various responses to Kinneavy's text as i f they were largely independent of each other. However, as argued in chapter 1, one of the most important features of the conversational model of rhetorical interchange is that i t draws attention to the ways in which texts form networks of interconnections (see especially 31). If we return to Burke's metaphor of the unending conversation, each conversant does not simply stand up, say his piece on the subject at hand, and then s i t down. The conversation ranges back and forth: "Someone answers; you answer him; another comes to your defense; another aligns himself against you, to either the embarrassment or gratification of your opponent, depending upon the quality of your ally's assistance" (Philosophy of Literary Form 110-11). Applied to the dialogue we are investigating here, this conversational model predicts that each of Kinneavy's c r i t i c s w i l l be influenced not just by Kinneavy's text but by countless other texts, in the forefront of which w i l l naturally be other critiques of Kinneavy. If we wish to explore the implications of the rhetorical model of reading to the fullest possible extent, then, we must go on to use i t to account for the ways in which the participants in a multi-party conversation judge not only Kinneavy's text but also each others'. Within the texts already cited there is some evidence of such D. Brent 171 interrelationships. Hunter, for instance, explicitly uses a refutation of Crusius' position as a take-off point: "I reject Timothy Crusius's proposition that 'we should adopt Kinneavy's theory as the paradigm for our field' as a dangerous and reactionary suggestion" (280). Such direct evidence of intertextual influence among the participants of this particular dialogue, however, is somewhat slight. It is more common for writers to refer to other writers who have proposed taxonomic schemes similar to Kinneavy's. Crusius, for instance, relates Kinneavy's taxonomy to those of Aristotle, Young, Becker and Pike, and Kenneth Burke; Knoblauch cites those of Richard Lloyd-Jones, James Britton, and James Moffett. In short, they tend to cite sources that are parallel to A Theory of Discourse rather than to their own critiques of i t . This is reasonable in view of the fact that Kinneavy's theory itself holds more inherent interest to these writers than do other critiques of i t . When one's rhetorical purpose is to comment on taxonomies of discourse, i t is reasonable to cite taxonomies of discourse rather than rival critiques. For our purpose here, however, the conversation becomes more critically interesting at a later stage when the participants begin specifically replying to each others' work. At this point, as writers attack each others' viewpoints and defend their own, when "someone answers; you answer him; another comes to your defense," we begin to see much more clearly how their perspectives differ and how they have been in some cases influenced by how they have read each others' texts as well as by how they have read Kinneavy's. The dialogue on Kinneavy is made particularly interesting by the fact that part of i t takes place in the "Comment and Response" section of College English. This section gives writers an opportunity to reply D. Brent 172 directly to specific articles in the journal and gives the writers of the original articles an opportunity for further rebuttal. These sequences provide a particularly explicit view of the ways in which writers read each others' texts. In the case of the dialogue on Kinneavy, the writers whose reading is displayed in this way are Timothy Crusius and Paul Hunter. Crusius' reply to Hunter is occasioned by Hunter's explicit description of Crusius' proposal as "a dangerous and reactionary suggestion" (280). His reply underlines many of the points that have been noted already, for he defends his attraction to Kinneavy's theories by restating and expanding on many of the arguments already made in "Plea" and "Thinking." Though by the time he writes "Comments," Crusius has modified the highly positive view of Kinneavy demonstrated in "Plea," he has not modified his belief in synthesis: "My position now is that Kinneavy makes the single greatest contribution to a synthesis of theories better than Kinneavy alone, or Moffett alone, and so forth" ("Comment" 214). In his response to Crusius, Hunter in turn clarifies and expands his earlier argument that the communications triangle is an inadequate basis for a theory of writing. In general, we see the familiar pattern of competing but internally coherent structures of belief: Crusius, backed by authorities such as Ong and Moffett, states the traditional view that literacy is based in orality, while Hunter, backed by post-structuralist authorities such as Foucault, Ricoeur, Bloom and Derrida, reiterates his earlier argument that writing should not be taught as an extension of speech. However, what is particularly interesting about these later texts in the dialogue is that they demonstrate how an evocation of a work and its D. Brent 173 implied author, which we have identified as an "event in time" rather than a stable entity, can shift under the influence of other texts. Though neither alters his basic position on Kinneavy, each writer is stimulated by other critics of Kinneavy, including each other, to create a new evocation of Kinneavy and a new perspective on his theories. This re-evocation deepens our understanding of the interaction between repertoire and rhetorical situation. In chapter 2 we argued that a chief component of the rhetorical situation is the questions that a reader asks of his material; these questions in turn affect the way the reader activates his repertoire of beliefs and values to evoke the virtual work from the inert words of the text. Here we see a further cross-connection of influence: other texts in the conversation form an important part of the rhetorical situation. By creating the need to answer and to come to others' and one's own defense, these texts have a profound influence on the way readers interrogate the original text (in this case, A Theory of Discourse), and thereby influence the way their repertoires of beliefs operate in practice to create a virtual work and its implied author. The body of critical literature that has grown up around Kinneavy since the publication of "Plea" creates a rhetorical situation that leads Crusius to undertake what is essentially a piece of dialogic criticism himself: he wishes to account for the resistance that many critics, Hunter included, feel toward Kinneavy's theories. Thus he constructs a new evocation of Kinneavy, this time from a new point of view and for a new purpose: not to improve his own representation of the world through apprehending Kinneavy's, but to help him understand the reaction of other critics to Kinneavy. To account for this reaction, Crusius evokes a D. Brent 174 Kinneavy whose textual character contains contradictions. This Kinneavy is influenced both by semiotics, which Crusius describes as "ahistorical and universalistic," and by traditional rhetoric, which "pulls him toward the relatively concrete domain of temporal process" ("Comment" 216). The result is a theory that inevitably pulls in contradictory directions— toward impossibly general questions answered by relatively context-specific examples. "No wonder," comments Crusius, "we feel somewhat i l l at ease" ("Comment" 216). Yet Crusius maintains his original position that the inadequacies in Kinneavy's theory are not insurmountable. By identifying for himself more clearly the problems generated by the structuralist side of Kinneavy's theory, he clarifies for himself the remedy: to refuse to reify the aims in the ways that a structuralist perspective (he claims) can lead one to do, and instead to treat the aims only as heuristic constructs that can help us "get on with the task of articulating the conventions of discourse that we are asking our students to write" ("Comment" 217). Crusius therefore has not been moved to accommodate his structure of beliefs to arguments such as Hunter's which are diametrically opposed to i t . Those opposed arguments, however, have had their effect. They have encouraged Crusius to re-evoke Kinneavy in a way that reshapes and modifies his response to A Theory of Discourse, making him more aware of the dangers of reification and of the internal contradictions within i t . Thus we see again the complexity of rhetorical inquiry. One participant does not necessarily bring another to believe as he does, but the enmeshing of ideas in a complex set of interchanges contributes to the growth of individual knowledge—in Booth's terms, the making of the D. Brent 175 self through the taking in of other selves (Modern Dogma 126). Because the rhetorical situation continually shifts when new texts and new perspectives join the conversation, each conversant is continually forced to re-evoke his view of other texts and of the reality they attempt to represent. A Theory of Discourse itself is not the only text that is evoked and re-evoked in this phase of the dialogue. In responding directly to Hunter's criticisms of Kinneavy, Crusius also evokes a representation of Hunter, an evocation crucial to the way Crusius both understands and judges Hunter's arguments. Crusius seeks to understand what he sees as Hunter's main charge against Kinneavy, the charge that Kinneavy is moralist rather than pluralist. To do so he constructs for himself and his readers the values that he sees Hunter attaching to key terms in the argument: Does Kinneavy's theory lead us away from "genuine pluralism"? We might ask first, What must pluralism be to be genuine? Hunter seems to believe that pluralism and a sense of what is right, suitable, appropriate (i.e., "moralism"?) are somehow at odds. Apparently for him a genuine pluralism must be open to everything, making no discriminations about relative value. ("Comment" 218) This passage can be interpreted as the rhetorical strategy of reiterating your opponent's stand in terms that make i t easy to refute (a strategy that, when taken to excess, becomes the dishonest "straw man" technique). Certainly to a certain extent this analysis must be correct: Crusius makes this construct partly for the benefit of the audience. But through the lens of a rhetorical model of reading, i t also appears as a D. Brent 176 reflection of Crusius' genuine attempt to understand Hunter, to create a "textual character" for him in order to facilitate a judgement. What Crusius finds at the bottom of Hunter's argument is a dichotomy between moralizing on the one hand and pluralism on the other, a dichotomy that Crusius finds unacceptable and which causes him to reject Hunter's argument. However, simply rejecting Hunter's argument is not satisfactory. In order to maintain the coherent structure of his own beliefs as well as to convince his audience, Crusius must find a satisfactory alternative to this dichotomy. To do so, he rearranges the terms of Hunter's argument to accentuate the values that he wishes to maintain. He relabels the attitude that Hunter calls "moralizing" with less negative terms such as "hierarchy of values," thus justifying his acceptance of Kinneavy's attitude. The term "pluralistic" he reanalyses into two terms: "eclecticism," which he associates with Hunter's values, and "genuine pluralism," which he associates with Kinneavy's.xs Here the division of "pluralism" into two concepts allows Crusius to remove the incompatibility arising from Hunter's opposition of "pluralism" and "moralizing," rejecting the concept of eclecticism while embracing a counterpart that is compatible with having a "hierarchy of values." Thus we see another aspect of our model fulfilled in practice. As noted in chapter 2 (78-80), the reader's evocation is influenced by the assumption that the text provides a consistent and internally coherent representation of the world. In turn, the reader also wishes to maintain 1 6Here again we see Perelman and Olbrechts-Tyteca*s technique of dissociation of concepts. It is worth noting that Crusius' use of this technique is exactly predicted by Perelman and Olbrechts-Tyteca: "The dissociation of concepts . . . is always prompted by the desire to remove an incompatibility arising out of the confrontation of one proposition with others, whether one is dealing with norms, facts, or truths" (413). D. Brent 177 the consistency of his own representation of the world. As a result, the enthymemic process becomes highly recursive. We have argued (chapter 3, 107-10) that a reader will be more likely to accept the propositions offered by a text i f they are consistent with his own doxai; reciprocally, a desire to accept certain propositions presented by a text provides a motivation to evoke the work in ways that as much as possible remove inconsistencies both within the work and between the work and the reader's own repertoire of doxai. To do otherwise would be to admit into one's own system of beliefs a set of perspectives that would degrade rather than enhance its coherence. Taken to an extreme, this process is simply a matter of seeing what one wishes to see, of justifying a prejudgement of a text by interpreting i t in the most favourable manner possible. But as with a l l other components of a rhetoric of reading, justifying a prejudgement is only a pathological extreme of an act that is an inevitable and positive part of the act of reading: of a recursive process of judging on the basis of an evocation and of evoking on the basis of a judgement. In this particular case, the re-evocation is motivated by another text (Hunter's), which gives presence to the potential inconsistencies which Crusius must remove. Hunter, for his part, undertakes a similar evocation of Crusius, for the same reasons: to understand and judge his arguments. The principle feature of his evocation of Crusius is to emphasize the gulf between them. "I wonder," ponders Hunter, " i f Crusius realizes how great the differences are between his language theory and my own" ("Responds" 220). In terms of the debate between them, this difference turns on a basic difference in the ways in which oral and written communication are viewed. Crusius sees writing as derived from speech, a view that renders D. Brent 178 acceptable the derivation of a theory of writing from the communication triangle; Hunter on the other hand sees the processes as quite different, writing being distinguished from speech by the way in which, unlike speech, i t combines the activities of encoding and decoding. It is this difference which lies at the foundation of his argument against a theory of writing based on the oral communication triangle. In his response to Crusius, Hunter is forced to make this point more explicit than he did in ''that we have divided'.11 Moreover, he condenses the differences between himself and Hunter to another pair of opposed terms, "structuralism" versus "poststructuralism" ("Responds" 220). By associating Crusius with structuralism (an association that Crusius himself never makes), Hunter builds an impression of Crusius' textual character that includes a l l of the value judgements inherent in the structuralist view of language as seen from the poststructuralist perspective. 1 7 Hunter is thereby creating an ethos for Crusius, an ethos that implies a bundle of attitudes which are incompatible with Hunter's. Here we see the circle of evocation and judgement working in the opposite direction, in the service of division rather than identification. To maintain the consistency of his own belief system, a belief system that includes a total rejection of most of Kinneavy's implied values, Hunter must either fend off Crusius' attempts to convince him of a different system or reorganize his own system completely. We have already noted how the operation of logos depends on connecting new perspectives with old ones, a technique that Burke refers to as using some of the 17Hunter never tells us exactly what these views are, but we can assume that they include the common charges that poststructuralism makes against the structuralist view of language. Of these, the one most obviously relevant here is that structuralism, by seeing writing as secondary to speech, systematically degrades i t . D. Brent 179 audience's beliefs to support a fulcrum by which other beliefs will be changed. To effect such a sweeping change in Hunter's belief system, Crusius would have to find a fulcrum of argument, a portion of Hunter's belief system with which he could identify the propositions that he wishes to promote and which is more deeply rooted than the ones he wishes to change. The strategy fails because Hunter, motivated by his desire to maintain the consistency of his own belief system, succeeds in evoking Crusius' argument in ways that make i t easy to resist by disassociating i t from any of his own more deeply-rooted perspectives. By using the familiar strategy of manipulating key terms, Hunter severs rather than builds enthymemic identifications between Crusius and himself. Here, then, we see again how the evocation of a text is bound up with the judgement of i t , this time from the point of view of rhetorical defense. In sum, though neither has been significantly moved from his original position, Crusius and Hunter emerge from the exchange having understood Kinneavy's system and their own reactions to i t from new perspectives. By evoking each others' positions, accepting parts and refuting others for the benefit of a third-party audience (the readers of College English and by extension a l l the other members of their discipline), Crusius and Hunter have not only reformed and deepened their own reactions to Kinneavy but have advanced a dialogue that started with the first reviews and summaries of Kinneavy (Corbett's, Corder's and D'Angelo's), a dialogue that is part of a larger conversation about discourse theory that started with Kinneavy's classical sources. Thus we see the power of rhetorical reading to build and rebuild knowledge, not just incrementally, by adding new pieces to an existing structure, but by continually provoking a reassessment of the old structure and its D. Brent 180 components.18 In turn, this dialogue illustrates the power of the rhetorical model of reading as a critical tool. The insights we have gained from this examination—the understanding of evocation and judgement, of logos, pathos and ethos, and the ways in which they operate in concert and in a recursive manner—could not have been gained from a productive rhetoric alone. That is, these insights could not have been gained by analyzing only the ways in which each writer attempts to persuade his audience. Rather, they require a perspective that subsumes this productive rhetoric into a larger system of social interchange. The arguments that the conversants choose, the strategic manipulations of terminology they engage in and the dichotomies they choose to accept or reject, must be accounted for as part of, and important evidence for, a process of re-examining and testing knowledge systems through social interchange. By examining their texts from this perspective, we catch the writers in the Boothian act of making themselves by taking in other selves. Conducted in this manner, a dialogic form of rhetorical criticism based on a rhetorical model of reading f u l f i l l s exactly Black's program for rhetorical criticism: by enabling a precise account of rhetorical interactions, i t promotes a more general understanding of human beings and the way they create knowledge through discourse. This explanatory "Here we see a microscopic look at a revisionary process that Kuhn describes on a macro level in The Structure of Scientific Revolutions. The remaking of knowledge through the sort of textual interchange we have seen here does not necessarily proceed as a series of revolutionary revisions of the sort that Kuhn argues occur in scientific progress (see Scientific Revolutions passim, especially 157). However, as with Kuhn's view of scientific progress, the intertextual conversation is not strictly incremental; i t is based not strictly on an addition of units of knowledge, but on a series of revisions and rejections of previous positions that, to use Spiro's term, '"update" one's view of the world (see chapter 2 of the present study, 42-43). D. Brent 181 power of a rhetoric of reading does not lie just in the labels that i t allows us to attach to features of the exchange. Rather, i t lies in the power of those labels and the general model of epistemic rhetoric that backs them to draw to our attention the textual features predicted by our model of the interpretive and judgemental processes. The terministic screens created by this model of reading thus direct our attention to evidence that the writers are not simply trying to persuade others of a predetermined point of view, but rather are creating their own representation of the world by looking for connections between their own repertoire of beliefs and those they perceive in their sources. It is because of this enthymemic search for connections between doxai, and by extension between the human beings whose minds are structured by those doxai, that the model of reading that we have used to account for the dialogue on Kinneavy can truly be called rhetorical. D. Brent 182 Toward a Rhetoric of Reading Chapter 5: Implications and Conclusions 5.1. Implications for Teaching Chapter 4 showed how a rhetorical model of reading could inform rhetorical criticism. Analyzed in the light of this model, the dialogue on Kinneavy shows how expert reader/writers are continually looking back to their sources and forward to their audience. They do not simply understand the words of their sources in isolation. Rather, they first interpret and then judge those sources in a complex logical, emotional and ethical interaction with other sources, their own structure of beliefs and the rhetorical situation. A description of a process performed by human beings automatically implies a description of the skills prerequisite for performing i t . To make this implied description explicit—to move, that is, from a descriptive to a normative account—we must take the description of what typically is done during the performance of the process and restate i t as a description of what one must be able to do in order to perform the process successfully. The rhetorical model of reading, then, not only suggests a number of features of reading considered as an abstract process; i t also implies the skills that successful readers must possess. First, because i t sees the reader's repertoire of prior knowledge and values as the matrix within which interpretation and judgement take place, the rhetorical model of reading implies that successful readers D. Brent 183 have both a richly stocked repertoire and the ability to access that repertoire in order to test the words of the sources against i t . Because i t bases the judgemental process on the model of the enthymeme, on connections between the reader's doxai and those that inform the writer's conclusions, the model suggests that readers can access not only their own repertoires but also those of others; that is, they can infer the patterns of assumptions that lead the writers of sources to their conclusions. Because i t sees judgement in terms of the three p i s t e i s — logos, pathos, and ethos—it implies that readers have learned (consciously or unconsciously) to trust not only their strictly logical analysis of a work as a set of propositions but also their emotional responses to the values underlying those propositions and their ethical responses to the implied character of the writer. Yet because i t sees reading as fundamentally rhetorical, as a process based on connections between writer and reader, i t implies that the emotional and ethical components of judgement are under a degree of constraint. The reader can use his own knowledge—his knowledge of language and the world, of what questions he is asking of the work, and of what constitutes unified and coherent discourse—as a means of constraining his subjective processes. Thus he avoids constructing totally idiosyncratic judgements that would sever the rhetorical connection between reader and writer. Finally, because i t sees reading as part of a larger rhetorical process that also includes the production of discourse, the model implies an ability to take one's own turn as a producer of discourse, to use effectively the double-edged power of writing both to refine one's own judgements and to pass them on to others. Thus, just as traditional systems of rhetoric are accounts of mature D. Brent 184 performance in the activity of choosing, arranging and expressing arguments for the purpose of persuading others,3- so a rhetorical model of reading is an account of mature performance in the activity of making knowledge by being persuaded by texts in an active and systematic fashion. And just as traditional systems of rhetoric suggest a body of knowledge and skills that can be imparted to the novice in order to help her become an expert,2 so a rhetoric of reading suggests a foundation for the systematic teaching of any activity that involves the making of knowledge through the processes that a rhetoric of reading describes— that is, through the juncture of discourse production and consumption, of persuading and being persuaded. In short, by clarifying and particularizing our understanding of the skills involved in making knowledge through rhetoric—in "taking in other selves," as Booth would put i t — a rhetoric of reading invites us to take a new look at how we teach knowledge-making activity. In its pedagogical aspect, a rhetoric of reading will not directly inform the teaching of reading as an isolated activity, for the very point of a rhetorical model of reading is that i t accounts for reading considered in a larger context. Reading, according to this model, is not just a transaction between an individual reader and a text. Rather, i t is an activity which, when paired with the complementary activity of ^Aristotle, for instance, defines his account of rhetorical technique as just such an account of mature performance: "When the practised and the spontaneous speaker gain their end, i t is possible to investigate the cause of their success; and such an inquiry, we shall a l l admit, performs the function of an art" (Rhetoric 1). It is such a systematic account not just of performance but of the reasons for successful performance that, as Plato argues more fully, characterizes an art as opposed to a knack (Gorqias 25 and Phaedrus 63-64). 2See chapter 1, 13-14 for this argument in more detail. D. Brent 185 writing, sustains a network of relationships among texts. From this perspective, reading is not just a consumptive but also an inventional process: i t is the first stage of a process of creating informative/suasory discourse as well as the last stage of consuming i t . This means that a rhetoric of reading cannot be applied effectively to reading as an isolated activity; i t must be applied to reading in the context of writing. Further, the rhetoric of reading that we have developed is specifically a rhetoric of non-literary reading, reading conducted for what Louise Rosenblatt calls "efferent" purposes. The efferent reader does not read for the aesthetic pleasure of the experience, the "living-through" of the perspectives created by the author, as Rosenblatt phrases i t . Rather, she reads to f u l f i l l two interrelated goals: to update her own knowledge, her own representation of reality, and to discover the materials that will enter into her own writing as she takes her turn in the conversation. The rhetorical model of reading, then, invites pedagogical application to the combined processes of efferent reading and writing—that is, to the activity performed by the conversants in the dialogue on Kinneavy. In the academic environment, this activity is typically labelled "research." In one of its senses, this term applies to what could more fully be described as "empirical" research: direct inquiry into phenomena through sensory observation, a form of inquiry that characterizes in particular the physical sciences. Yet as Booth points out, even practitioners of the physical sciences depend on others' testimony transmitted chiefly through reading (Modern Dogma 108-09). Throughout the academic disciplines, not just in the humanities, doing "research" D. Brent 186 means doing what Crusius, Knoblauch, and Hunter do when they attempt to assess discourse taxonomies in the light of Kinneavy's paradigmatic system: i t means examining not just the phenomena under discussion but also the phenomena as examined and interpreted by other writers and mediated by the texts those writers produce. It means interpreting and evaluating texts—reading rhetorically—as well as observing the world. The most consciously structured means of teaching this activity to novice researchers is the familiar composition exercise known as "writing the research paper." In Research Paper Instruction in the Undergraduate Writing Program, James E. Ford and Dennis R. Perry point out that the research paper is an extremely common form in the composition classroom: Research paper instruction is currently being offered in 84.09% of freshman composition programs and in 40.03% of advanced composition programs. In general this high level of occurrence is fairly uniform in a l l types and sizes of institutions and in a l l regions of the country.3 (827) As Thomas Kane suggests in The Oxford Guide to Writing, this exercise requires students to create their own interpretation of a phenomenon based in whole or in part on "secondary sources," and then to persuade others of their interpretation through the medium of writing: The challenge of the research paper lies in invention—in discovering topics to write about—and in organizing the material you find. Unlike the usual freshman composition, which you can write out of your own experience, the term paper 3These percentages, of course, can only be taken as definitive for the country in which Ford and Perry's survey was conducted, the United States. However, there is no reason to suppose that the research paper is significantly less emphasised in this country. D. Brent 187 requires considerable research. In some cases—in the sciences, for instance—this necessitates observation and experiment. More likely, however, invention largely arises out of reading. (519) The sort of reading that Kane associates with the student research paper, like the sort performed by expert researchers, is rhetorical reading, a complex process of listening to other voices and of being systematically persuaded by them. However, without a firmly articulated account of rhetorical reading on which to mount a pedagogical system, those who teach this sort of reading will have difficulty reaching beyond imparting what Plato calls a "knack" rather than an "art" (Gorgias 25). It will be a helpful collection of observations about what researchers often do and what often seems to work, but will not be informed by a solid sense of what the process i t is describing really is. Without an account of how knowledge is formed rhetorically to inform it—that is, as a result of its "knack" status—the teaching of the research paper tends to be consciously or unconsciously governed by an archival view of knowledge. In Successful Writing, for instance, Maxine C. Hairston distinguishes sources that involve some form of direct sensory observation (primary sources) from those that embody others* ideas (secondary sources). If phrased as a convenient pair of labels, this distinction between primary and secondary sources is a useful one. However, Hairston's language suggests that the first type of source is greatly to be preferred over the second. Secondary sources, she states, supply only "second-hand information" and are "really someone else's research on the same or a related subject" (195). Thus she suggests that reading is simply a vicarious means of obtaining sensory data. To judge D. Brent 188 sources, one must simply assess their reliability as to facts by using extrinsic criteria such as the date of publication, the author's reputation and the degree of scholarly treatment. The best sources, she implies, are those that function as camera-eye transmitters of experience, relaying to the reader data that he is not fortunate enough to have been able to experience first-hand. This view, typical of handbook treatments of secondary-source research, places the use of print sources outside the rhetorical act. In many important ways i t reflects Socrates' complaint in the Phaedrus: Writing, you know, Phaedrus, has this strange quality about i t , which makes i t really like painting: the painter's products stand before us quite as though they were alive; but i f you question them, they maintain a solemn silence. So, too, with written words: you might think they spoke as though they made sense, but if you ask them anything about what they are saying, i f you wish an explanation, they go on telling you the same thing, over and over forever. (69) According to this view, print is merely a repository of fossilized rhetorical acts that can no longer actively participate in the living process of rhetoric. Thus research involving secondary sources is only preliminary to, not part of, the process of creating new meaning. Few treatments of research writing, of course, explicitly denigrate the idea of writing from sources rather than from direct observation. Rather, this data-retrieval view of research is communicated indirectly through the relative emphasis placed on various parts of the process. Kane, for instance, describes the stages of writing the research paper as follows: D. Brent 189 The first step in constructing a research paper . . . is learning how to find what you want in the library. Next you take and arrange notes from the material you uncover, incorporate those notes into your text, and acknowledge your sources in appropriate footnotes and the bibliography. (519) For Kane, then, research is primarily a matter of finding and presenting data. Similarly, Michael Meyer's book-length treatment of the subject, the Little. Brown Guide to Writing Research Papers, is almost entirely devoted to using the library, taking notes, and documenting sources, together with some standard instructions on outlining and drafting. Again, the evaluation of sources is treated chiefly as a matter of measuring the writers' overall authority as witnesses. There is much of value to the student in such discussions. The processes of finding and analyzing sources, of taking useful notes and incorporating them smoothly into a text, of documenting with the needs of the reader in mind, are neither t r i v i a l nor mechanical. Yet the mastery of these skills is only a necessary, not a sufficient condition for becoming a good researcher. What novice research writers also need is a sense of how to perform the intricate rhetorical dance illustrated by the dialogue on Kinneavy; a sense of how to incorporate efferent reading into a process that is both rhetorical and epistemic. Some treatments of the research process already point in this direction. In their 1982 article, for instance, Schwegler and Shamoon argue that one of the most important goals of the research paper is to inculcate a mode of thinking, not just a mode of writing: In our conversations with instructors and in related studies of instructors' responses to students' research papers . . . i t D. Brent 190 became clear that college instructors view the research paper as a means to accomplish one of the primary goals of college instruction: to get students to think in the same crit i c a l , analytical, inquiring mode as instructors do—like a literary c r i t i c , a sociologist, an art historian, or a chemist. (820-21) Brenda Spratt's Writing from Sources, one of the best of the current handbooks on research writing, attempts to put this line of thinking into practice by taking a highly complex view of research writing: Writing an essay based on sources depends on a complex group of skills. . . . It is pointless to teach students how to compile an impressive bibliography i f , when the time comes to select materials and integrate them into a coherent essay, they simply produce the familiar string of end-to-end quotations, (v-vi) In order to give students a more complete understanding of how to write a research paper, Spratt's text works in stages toward the goal of showing students how to synthesize sources into what she calls a "multiple-source essay": Your own perspective is the framework for your essay, and you must use your understanding of the topic—which is more thorough and detached than that of any of your sources—to evaluate your materials. Review the hypothesis that you formulated before you began to read and analyze the sources. Decide whether that hypothesis is s t i l l valid or whether you wish to alter i t or abandon i t entirely in the light of your more detailed understanding of the subject. Then, having reaffirmed your hypothesis or chosen another, s i f t through a l l D. Brent 191 the statements and decide which ones are thoughtful and well-balanced, supported by convincing reasons and examples, and which are thoughtless assertions that rely on stereotypes, catch phrases and unsupported references. (262) It is just such a project as this, which asks students to decide which sources are "thoughtful and well-balanced" and which are "thoughtless assertions," that can most obviously benefit from a rhetoric of reading that seeks to answer Booth's question "When should I change my mind?" The most general benefit of a rhetoric of reading is to underpin projects like Spratt's with a principled alternative to the archival metaphor of research. Under the rhetorical view, the texts that intervene between the reader and the subject of his research are not merely repositories of data that "maintain a solemn silence" when questioned. Rather, they embody alternative ways of knowing that they will yield upon interrogation and which a reader can use to deepen his own understanding of the subject. They represent sample worldviews built upon a combination of their authors' sensory experience and their authors' previous reading. By interpreting and evaluating those worldviews, the reader as researcher can not only participate vicariously in others' experience; he can also participate in others' interpretations of experience and select portions of those interpretations to incorporate into his own worldview and ultimately pass on to others through writing. In short, through being selectively persuaded by texts and in turn persuading others by means of his own texts, the reader can participate in the epistemic interchange envisioned by modern rhetoric. If this view of mature successful performance in research is to inform teaching, however, we must be able to use i t not only to set D. Brent 192 general pedagogical goals but also to suggest specific ways in which instruction can help students attain those goals. At first glance, the rhetorical model of reading seems to suggest that the ability to read in a rhetorical mode is the result of processes not amenable to direct instruction. As outlined in chapter 2, the first stage of rhetorical reading is the erection of a virtual work based on a transaction between the text and the reader's repertoire of knowledge and values, a repertoire that can be described as a set of schemata that systematizes the reader's understanding of the world. By fitting the text into known schemata, the reader builds an understanding of the new material that is being presented and constructs an image of the world as presented in the text under scrutiny. This view of discourse comprehension can be taken to suggest that one cannot perform proper research until one has acquired a satisfactory stock of world knowledge with which to interpret new data. A logical conclusion would therefore be that students should not undertake research until late in their university careers when they have mastered the basic elements of their discipline. 4 On the other hand, i t is important to realize that the repertoire as defined in the rhetorical model of reading is not just constructed of disciplinary knowledge. Rather, i t is a complex system of linguistic information, general and disciplinary knowledge, and emotional associations and values. The schemata that a student can apply to a given research problem do not come into being solely through immersion in the discipline involved. Rather, such immersion ripens and gives depth 4An example of this argument is Stephen North's "Teaching Research Writing: Five Criteria," which argues that students often have "no depth of knowledge, no existing schema for the subject area in which they are writing" (18). As a result, North argues, the writing of the research paper should generally be taught only in higher-level courses. D. Brent 193 and substance to a repertoire that is already in place. Most importantly, this structure of associations and values is formed largely by the experience of meeting other worldviews through reading. We are, as Burke, Booth and other rhetoricians remind us, made in symbolic interchange with other selves—not just through passively receiving and storing up what others have to offer, but by interacting symbolically with them, participating in Burke's "co-operative competition of the parliamentary." If i t is true that, as we have argued, research is simply a formal, academically situated version of this symbolic interaction, then to delay immersing the student in research until his repertoire is formed is to deny him access to one of the most important of the processes that form i t . This is clearly a self-defeating proposition. The rhetorical model of reading does not t e l l us, then, that we cannot teach the research paper until the repertoire is in place. Rather, i t tells us that we must construct activities that help students learn to use the structures of knowledge that they currently possess as bridges to new and richer structures of more specifically disciplinary knowledge. In order to do so, they must be able to understand what i t means to engage in the social construction rather than the individualistic de-archiving of meaning. This emphasis on the social nature of knowledge construction points in the direction of the collaborative writing model espoused by composition theorists such as Karen Burke LeFevre and Kenneth Bruffee. Collaborative writing is in many ways the most promising attempt to base a composition methodology on an expanded view of rhetorical invention. In Invention as a Social Act, for instance, LeFevre draws one of her D. Brent 194 controlling metaphors from Michel Foucault: [Foucault] describes the beginning of a discourse as a re-emergence into an ongoing, never-ending process: "At the moment of speaking, I would like to have perceived a nameless voice, long preceding me, leaving me merely to enmesh myself in i t . . . . There would have been no beginnings: instead, speech would proceed from me, while I stood in its path—a slender gap—the point of its possible disappearance." Elaborating on this perspective, one may come to regard discourse not as an isolated event, but rather a constant potentiality that is occasionally evidenced in speech or writing. . . . Such perspectives suggest that traditional views of an event or act have been misleading when they have presumed that the individual unit—a speech or a written text, an individual hero, a particular battle or discovery—is clearly separable from a larger, continuing force or stream of events in which i t participates. For similar reasons Jacques Derrida has criticized literary theories that attempt to explain the meaning of a text apart from other texts that precede and follow i t . (41-42) Because she sees invention in this larger context, as a participation in a rhetorical interchange among texts rather than as an "atomistic" process of individual contemplation, LeFevre concludes that composition pedagogy should include co- and group authorship. Students, claims LeFevre, should be given opportunities to form research communities within the classroom. By collaborating on research projects, reading and evaluating texts and constructing new texts together, students will learn D. Brent 195 to form the sort of social relationships in which knowledge is typically-constructed outside of the classroom: The goal, as James A. Reither has put i t , would be to "immerse" students in discourse communities through a workshop approach in which "students and teachers function but as co-investigators, with reading and writing being used collaboratively to conduct the inquiry." (130) Kenneth Bruffee's work on collaborative writing also recognizes the relationship between reading/writing and the epistemic conversation. In his article "Collaborative Learning and the 'Conversation of Mankind,'" Bruffee develops an analogy between writing and conversation: Writing is a technologically displaced form of conversation. When we write, having already internalized the " s k i l l and partnership" of conversation, we displace i t once more onto the written page. (641) This conversational view of invention leads Bruffee, like LeFevre, to recommend teaching students to write in a collaborative mode: The inference writing teachers should take from this line of reasoning is that our task must involve engaging students in conversation among themselves at as many points in both the writing and the reading process as possible, and that we should contrive to ensure that students' conversation about what they read and write is similar in as many ways as possible to the way we would like them eventually to read and write. (642) Collaborative writing theory, however, does not pursue the implications of its social epistemology very far in the direction of accounting for reading as a rhetorical process. LeFevre's concrete D. Brent 196 examples of the writer at work, for instance, begin with the attempt to gather arguments for a particular discourse: To put this perspective into practice, let us imagine that we are observing a writer at her desk, at work on an article for a professional journal. In drafting notes that eventually become fuller texts, she perhaps subjects her ideas to the scrutiny of an imaginary reader, modifying them by means of internal dialogue. Later she asks a colleague to respond to a draft. As they discuss ideas, new thoughts begin to form. Back the writer goes to another draft, another imaginary reader, and yet another actual reader. Eventually the editor of a journal reads the article, suggesting changes and additions. A year later i t is published. Readers write responses to the editor; the writer composes her defense. The text is "finished"—except that the process has provoked a new idea that the writer may use in another article. Perhaps her article prompts someone else to develop ideas in a different direction. For both writer and reader, then, the closure of one series of talks and texts may end in invention. (42-43) This emphasis on personal interaction is a valuable corrective to an atomistic view of the writer as individual creator, yet i t enacts only part of the rhetorical philosophy on which i t is based. It takes account of the act of invention as a process that extends forward in time from the first attempts to formulate a text. However, we see no account of the unending conversation that precedes i t in time—the author's participation in the interaction of texts that gives rise to the propositions that will be further modified through the next turn of the D. Brent 197 conversation. Only at the end of the story, when the writer's article becomes itself an impetus to further texts, do we see an epistemic conversation developing. But this is a coda to LeFevre's story; the interactions between the writer and her collaborators is her main theme. Similarly, Bruffee recommends engaging students in conversation while composing, but maintains a focus on productive rather than consumptive rhetoric. This focus is fully appropriate in the light of what LeFevre and Bruffee are attempting to accomplish. They are, after a l l , composition theorists, and see their task as expanding our understanding of how discourse is composed.15 But the shift in focus imposed by the rhetorical approach to reading suggests ways in which the collaborative methodology sThis orientation of collaborative writing theory may also result in part from the sources to which i t has turned for inspiration. LeFevre, for instance, takes George Herbert Mead as the theoretical foundation for the collaborative perspective. For Mead, the objective universe only becomes meaningful to individual organisms when constituted in terms of a response. Further, meaning arises not just out of an organism's response to its environment, but out of one organism's response to another's symbolic gestures (verbal or nonverbal). Bruffee includes among his l i s t of theoretical influences Lev Vygotsky, an educational psychologist who argues that children first use speech instrumentally as a cognitive aid to a practical activity, talking through a problem to themselves rather than to others. This instrumental speech gradually splits into social speech, used for communication, and inner speech, a concept-forming and concept-handling activity that becomes the matrix of thought itself. Both of these perspectives on language are clearly relevant to the process of composition. If objects are constituted through social interaction, i f thought is internalized conversation, then conversation in the form of both literal conversation with peers and internal dialogue is an essential part of inventing discourse. The audience of discourse— both the final audience and the incidental audiences that come and go throughout the collaborative process—becomes not just the recipient of the rhetor's thought but an essential component of its development. However, these perspectives on language, derived as they are from psychology and sociology, naturally emphasise the role of language that is of most concern to those disciplines: language as a medium of personal interaction in everyday l i f e . This emphasis on immediate social interaction, while not precluding an interest in the discourse of print, provides no compelling reason to account in detail for the ways in which writers interact not only with their audiences but also with their sources. D. Brent 198 can be applied more specifically to the writing of research papers in order to emphasize the ways in which reading as well as writing is rhetorical and interactive. 6 If, for instance, a group of students is given the same text to understand, the rhetorical model of reading predicts that they will construct different and sometimes contradictory virtual works. A natural pedagogical reaction would be to try to show students how to synthesize these readings into a single definitive reading, eliminating erroneous or wrong-headed interpretations until a consensual virtual work can be constructed."7 A pedagogy based on a rhetoric of reading would shift this 6The fact that a rhetoric of reading provides an account of mature performance may seem to imply a sort of "top down" approach at odds with the collaborative model. The notion that the teacher should have a clear sense of the skills that she wants her students to acquire, that is, may seem at odds with a methodology of self-discovery in which students are set tasks that will encourage them to formulate their own notions of what i t means to read and write. However, this is a false antithesis. A collaborative model of instruction requires that students acquire skills by performing a series of tasks collaboratively, with or without explicit guidance, rather than simply being told how they should behave. However, it does not preclude the teacher's having a clear sense of what those skills should be, a sense that is in fact a prerequisite of good task design. Knoblauch and Brannon, for instance, are outspoken proponents of a collaborative pedagogy, but also stress the importance of teachers knowing what i t is they are facilitating: Teachers familiar with the abstractions can better understand the nature of composing and therefore the processes their students are striving to control. As a result, they can more surely devise classroom activities to promote growth and can more readily define their roles as facilitators of growth. (Rhetorical Traditions 102) •""This is the pedagogy implied by handbooks such as Anita Harnadek's Critical Writing Improvement. This book could well form the foundation of a rhetorically-based program of reading instruction such as outlined in this chapter, for its entire final section offers a wide selection of texts on the same general subject (the Detroit race riots) written by different authors from differing perspectives. Guide questions pose problems of interpretation such as, "Did the writer of article 46 appear to have more in mind than a straight news report? Explain." This sort of question could lead to a discussion of multiple interpretation; however, Harnadek goes on to provide an authoritative answer: "Yes. D. Brent 199 focus slightly. The rhetorical model of reading recognizes that there must be an element of coherence and predictability in a rhetorical transaction or the transaction ceases to be rhetorical at a l l ; i t is therefore reasonable to compare readings and eliminate some as being clearly too far from what one would normally expect to be the author's intention. Yet at the same time, groups of students could be set to the task of exploring what in their own experience of the world and their own values has led them to construct their varied interpretations. By detailed comparison of each others' readings, students can begin to get in touch with the processes by which they understand texts, and begin to grasp the relationship between different repertoires and different virtual works. Yet at the same time they can explore the factors that set limits on interpretation: the commonalities of human experience and of factual knowledge that the writer can use to predict response. In addition to the repertoire, the rhetorical model of reading identifies the rhetorical situation as a key element in the construction of the virtual work. This rhetorical situation is composed not only of extrinsic factors such as the audience and the setting, but also of the rhetorical exigence, the questions that the reader asks of the text. As the rhetorical transaction—which we, borrowing Rosenblatt's insight, have described as an "event in time"—proceeds, these questions will inevitably shift. As answers to some questions begin to form, new questions will arise. As a result, the virtual work will continually Aside from reporting a battle at Lawton and Carter, the writer tried to give a running account of the battle in a way intended to make the reader a witness. He also tried to convey the enormity of the sniper problem (paragraphs 4, 10, 15) and what the police were going through to end i t " (159). Such univocal answers clearly imply that the goal of reading is to uncover the correct interpretation. D. Brent 200 shift as the questions under which i t is interrogated mature. As applied to pedagogy, this insight implies that students must not only be encouraged to set specific questions when reading, but also to refine them as their research proceeds. A writer who begins with a question such as "What were the causes of the War of 1812?" may find that as her research proceeds she becomes involved with a slightly different question such as "What was the relationship between the revolutionary nations such as the U.S. and France and the non-revolutionary nations such as England?" This is more than the typical "narrowing" of a subject to make i t more "manageable"; i t is a recognition that increasing participation in an intellectual conversation leads one into different avenues of thought that make texts appear different. From the same set of words on the page, a different virtual work appears as the reader asks different questions. The corollary is that the student must be willing to reread works in context during various stages of researching and writing the paper. The notes that one would take under the influence of the first question would not look at a l l like the notes one would take from the same source under the influence of the second question. In terms of specific classroom practice, this means that teachers must allow sufficient time for students to refine their questions and give them step-by-step encouragement to do so. In their unpublished article "The Road Not Taken: How the Writing Context Influences Students' Choices," Jennie Nelson and John R. Hayes contrast two types of research strategy typically employed by students. One type, labelled "low-investment strategies" by Nelson and Hayes, involve techniques such as racing quickly through the first dozen books available to glean raw information in a single pass. "High-investment strategies," in contrast, D. Brent 201 involve techniques such as rereading sources in the light of refined questions. Nelson and Hayes claim that the students who choose high-investment strategies are typically those who are encouraged to do so by specific pedagogical techniques. A teacher who sets a research assignment well in advance, encourages students to record the progress of their ideas as they develop, and meets with them individually before they hand in their final drafts will have the opportunity to sound out their research strategies and motivate them to go back to their sources i f the direction of their inquiry seems to be changing. Only through such direct intervention will teachers be able to detect and remedy counterproductive assumptions about research, assumptions such as the opinion that you only need to reinterrogate sources i f you have not done a thorough job the first time.8 The rhetorical model of reading demonstrates that as well as constructing an abstract and ever-shifting virtual work, readers must also use a complex and highly personal set of criteria in order to judge that work and set i t in a useful relation to other works. Framing its view of the judgemental process in term of the classical pisteis, the rhetorical model of reading suggests that judgement involves not only a sAttempts to disabuse students of this idea are not helped by texts such as Audrey J. Roth's The Research Paper; Form. Process and Content. Roth's text is unfortunately not aberrant but rather is typical of texts on the research process when i t gives advice such as the following: If your notes are legible and accurate, they will probably also be as complete as they ought to be. It is frustrating to discover, while you are writing a paper late at night, that important information is missing because you failed to write i t down or that you need to look at a source again to clarify some point. (109-10) A note could only be complete on the first pass i f research consisted only of de-archiving information rather than interacting with texts. D. Brent 202 rational process of tying new beliefs enthymemically to pre-existing premises (logos), but also an emotional process of responding to values (pathos) and an ethical process of responding to the writer's character as implied in the text (ethos). It follows that the researcher must be able to use a l l three modes of being persuaded by a text in order to interact with i t fully. To apprehend fully the process of logos, students must work toward discovering how their own doxai allow them to evaluate a work at the same time as the work is persuading them to modify their own doxai. One student may reach a positive judgement of a particular writer's conclusions because the lines of argument the writer uses tie into the student's prior understanding of the world. Another may reject the work because its acceptance would require the rejection of too many deeply-held beliefs^without a corresponding gain in the coherence of his understanding of the world. By closely guided comparison of such judgements, students can work toward understanding how their own doxai affect their judgement. Once they understand how this process works, they will be better equipped to answer the more difficult question of whether they should reject the work or modify their doxai. Only by understanding which of their own beliefs are the most deeply held, as well as the reasons for this relative tenacity of belief, can students progress from an instinctive favouring of one text over another to a conscious, rational process of deciding when to change their minds. Students must also understand how they can use their emotional as well as their rational reactions to sources. It is commonly asserted that emotion has no place in academic research, particularly in the more scientific disciplines. Louise Rosenblatt, for instance, comments that D. Brent 203 in scientific reading, "the reader must adopt the attitude of mind, the stance, that will lead him automatically to reject or inhibit any personal associations activated by the symbols" (73). But as we have seen, this is only partially true. Logos and pathos are inextricably blended in even the most scientific reading: a reader's judgement of a source's value depends not just on the way i t connects with prior knowledge but also on the way i t connects with deeply held and deeply felt emotional values. Too often students are given exercises in separating "fact" from "opinion" with the explicit or implicit intention of encouraging them to value the former and discount the latter. 9 Rather, although students must learn to avoid being overwhelmed by powerful and immediate emotional judgements—they must not, for instance, automatically reject an assertion as untrue because they do not approve of its consequences—they must also learn not to attempt to expel emotional reactions from the process of judgement. In short, they must avoid what Booth calls the "fact-value split" by learning to use their emotional apprehension of value as one part of a complex judgemental process. To enable them to understand the interrelation of logical and emotional persuasion, students might, for instance, be divided into groups and given sets of source materials about which to form an opinion. 9See for instance Anita Brostoff and Barry K. Beyer, "An Approach to Integrating Writing into a History Course." Brostoff and Beyer describe how they encourage students to "distinguish evidence from hypotheses about evidence and statements of proveable fact from statements of opinion or value judgements" (43). Unfortunately, the tone of their discussion makes i t clear that they value the former over the latter. Despite Weaver's observation that there is no language without tendency, no discourse that is "a nonconductor of the current called 'emotion' and its concomitant of evaluation" ("The Power of the Word" 37), i t is not easy to break students, or teachers either, of the temptation not only to identify an author's "bias" but also to root i t out like a noxious weed. D. Brent 204 In discussing each others' reactions to the sources, students could be encouraged to probe themselves and each other for the reasons, the clusters of values and doxai, that lead them to have the reactions they do. The most important part of such an exercise would be to learn how to validate and to respond critically to emotional and value-laden reactions. The rhetorical model of reading suggests that the writer's ethos, the systematic complex of beliefs and values that we have called the writer's "textual character," is also a highly important factor in persuasion. Accordingly, while they are probing each others' reactions to sources, students should also be encouraged to construct and use the ethos of the writers of sources. Teachers can encourage students to ask the same questions that the participants in the dialogue on Kinneavy ask of Kinneavy and of each other: that is, they can encourage their students to attempt to characterize the philosophical mindset, the textual personality, that lies behind the conclusions that authors draw from their information. Here i t may be helpful to supply sets of material by the same authors. By attempting to trace consistencies in outlook among several texts by the same author, as we did with three of Knoblauch's texts in chapter 4, students can begin to understand beliefs as coherent structures through which readers see the world, not just as isolated "biases" that can be stripped away from the "facts." The rhetorical model of reading, then, can suggest a number of specific pedagogical approaches. It is important to remember, however, that these are only illustrations of some ways in which students can be guided in the direction of the mature performance described by the model. It is that account of performance itself, the exact specification of what D. Brent 205 expert readers do when confronted by multiple texts offering multiple interpretations of the world, that is the most important contribution of a rhetoric of reading. This account can further buttress the attempt, typified by writers such as Spratt and Schwegler and Shamoon, to move the pedagogy of the research paper in the direction of an art based on a thorough understanding of the process i t is designed to inculcate. As we have seen, the most important aspect of that process of rhetorical reading is that, like rhetoric itself, i t is a process of building knowledge through dialectical interchange. The goal of pedagogy must therefore be to help students learn to read, not just as retrievers of data, but as complex human beings in Weaver's sense; not just as logic machines, but as human beings who both judge and advise with their whole logical, aesthetic and ethical beings. 5.2. Implications for the Art of Rhetoric The i n i t i a l motivation for developing a rhetoric of reading was to expand the art of rhetoric in order to account more fully for new views of the role of persuasion in the creation of knowledge. Having developed such a rhetoric of reading and examined its implications for rhetorical criticism and for pedagogy, we can now return to the general implications of this model from a broadened and more detailed perspective. We can now ask not just why rhetoric needs to include a rhetoric of reading, but also what the inclusion of a rhetoric of reading does for rhetoric. In other words, we can ask what we now know about the rhetorical process that we did not know before developing a model of how reading can be understood as one of its subprocesses. D. Brent 206 First, the expanded definition of persuasion under which we have been operating has led to an expanded definition of invention. Persuasion has always been central to the art of rhetoric, an art defined by Aristotle as "the faculty of discovering in the particular case what are the available means of persuasion" (Rhetoric 7). However, modern rhetoricians such as Wayne Booth and Kenneth Burke emphasize the role of persuasion not only in the passing on of knowledge from one person to another but also in the development and perfection of knowledge, and in the creation of the self through intercourse with other selves. Burke's classic metaphor of the unending conversation illustrates the way in which we create and attempt to perfect knowledge by attempting to persuade others of the validity of propositions, a process that Burke calls the "'cooperative competition' of the 'parliamentary'" (Language as Symbolic Action 107). As Wayne Booth points out, this view of knowledge as created in rhetoric means that "the supreme purpose of persuasion . . . could not be to talk someone else into a preconceived view; rather i t must be to engage in mutual inquiry or exploration" (Modern Dogma 137). The office of invention, then, must be expanded to include not only the finding of arguments to support a position, but the finding of the position i t s e l f — a position that must be perceived as provisional, for attaining i t is merely a starting point in a process that will surely modify i t as the conversation proceeds. Therefore invention must be seen as a Janus-headed process. It looks backward to previous stages of the conversation that offer the rhetor others' provisional forms of knowledge; at the same time i t looks forward to the audience to which the rhetor's knowledge, her best estimate of the world at the time of D. Brent 207 composing the piece of discourse in question, will be presented. Having developed an account of rhetorical invention that takes account of this expanded view of the process, we not only know that, as Booth argues, we are creatures "made in rhetoric"; we know how. In a print-oriented society such as ours, an office of rhetoric that provides a link between receiving others' knowledge and presenting one's own also provides a link between reading and writing. It is by reading that one most clearly taps the resources of a conversation that now takes place largely in print, and i t is by writing that one most clearly offers to others the fruits of one's explorations. Only through a model of reading that sees i t as a rhetorical process can we fully account for the way in which these two related processes of reading and writing find a natural junction in rhetorical invention. A rhetoric of reading is therefore a component that modern rhetorical theory cannot do without and s t i l l remain true to its conviction that persuasion is a knowledge-making as well as a knowledge-transmitting activity. As we have sketched i t , the rhetorical model of reading begins with an important insight into the reading process drawn from both reader response critical theory and discourse processing theories of reading: the idea that what is read is not just a set of characters on paper, a static text that is the same for every reader, but a virtual work that the reader creates as an event in time. This work represents a transaction between the characters on the page, the shared conventional meanings attached to those characters, and the unique personal response of the reader as governed not only by her biography and personality but also by the rhetorical situation in which the reading event takes place. Rhetoric, then, governs the interchange of ideas at the most fundamental D. Brent 208 level: the understanding of others' structures of knowledge. A rhetoric of reading thus helps f u l f i l l Wayne Booth's search for "grounds for confidence in a multiplicity of ways of knowing" (Modern  Dogma 99). Again, a rhetoric of reading does not just assert that there is more than one way of attaining knowledge; i t suggests the mechanisms that both allow for and constrain this multiplicity. The close relationship between a person's particular configuration of doxai and the degree to which she will be persuaded by any given piece of discourse explains the wide diversity in human belief. It provides, that is, a set of principles that can answer the puzzling question of how different intelligent people can read the same sources and reach apparently contradictory conclusions. By grounding the process of reaching a conclusion in a highly variable process of constructing a virtual work, i t also accounts for the way different readers not only reach different conclusions about source texts but also appear to have read different texts. Yet the model also shows how commonalities in human experience limit the diversity of possible repertoires, and how the writer's expectations of the reader's rhetorical situation and his need to create a coherent interpretation of the work provide a mechanism by which the writer can predict and to a certain extent control the reader's response. The multiplicity of ways of knowing can therefore be shown to be a limited and predictable multiplicity, and therefore a multiplicity that does not collapse into arhetorical relativism. Confidence in a multiplicity of ways of knowing requires an understanding not only of how we construct but also of how we come to believe or reject the propositions presented by this virtual work. The rhetorical model of reading provides this account through its re-D. Brent 209 evaluation of the enthymeme, the most fundamental working principle of traditional rhetoric. In traditional rhetoric, the enthymeme is a structure of argument in which doxai, pre-existing opinions, provide connecting links between the rhetor and the audience. Rhetorical proof by logos depends on the rhetor's being able to connect his propositions with opinions of the audience. If those opinions are held with sufficient strength, the audience will be persuaded to change other opinions in order to accommodate those offered by the rhetor. Proof by pathos depends on a related process of establishing emotional connections between the values that the rhetor wishes to present and those of the audience. Finally, proof by ethos consists of establishing not just individual beliefs and values that connect with those of the audience, but an entire complex of beliefs and values, a total character presented in a text. Persuasion will depend partly on the degree to which the audience finds this character amenable. From the point of view of a rhetoric of reading rather than a rhetoric of composition, the principle of the enthymeme operates on what might be called an input rather than an output level. From the point of view of composition, the enthymeme is a method by which the rhetor as composer, by emphasising the connections between his and his audience's doxai, shapes the discourse in ways that will make i t maximally acceptable to the audience. However, from the point of view of reading, the enthymeme becomes a means by which the rhetor as reader selects the propositions that he will add to his own store of knowledge. The propositions which are accepted are those which make the most coherent logical, emotional and ethical connections with the reader's repertoire of knowledge. In short, because i t describes the mechanism by which D. Brent 210 knowledge is both received and transmitted, the enthymeme may be seen as the pivot upon which turns the entire process of creating knowledge through rhetorical interchange. However, the rhetorical model of reading also reminds us that, despite the attractiveness of electronic metaphors such as receiving and transmitting, the enthymeme is not a mechanical process. It is simply an abstract concept that helps account for a most humane process. The doxai that the enthymeme works with are the opinions of human beings— provisional, changeable, and, as the division of proof into the pisteis reminds us, influenced not just by logic but by emotion and character. By elaborating the principle of the enthymeme into a knowledge-making and knowledge-transforming as well as a knowledge-transmitting device, the rhetorical model of reading supplies a rich and detailed understanding of rhetoric as an epistemic process, just as traditional rhetorical theory supplies a detailed understanding of discourse understood as a process of transmission. By revealing in detail how doxai form bridges of identification between human beings, i t establishes a holistic view of knowledge; that is, i t emphasizes the way that human beings use not only their logical but also their emotional and ethical faculties in the process of coming to know. Thus i t f u l f i l l s and extends the view of man asserted by modern rhetoricians such as Richard Weaver, who argues that man is not a logic machine but a complex being whose knowledge is based not just on his powers of thought but also on his subjective reaction to his hopes, fears and special circumstances ("Language is Sermonic" 205). Despite the epistemological bias of this view of rhetoric, the art of rhetoric can never be reduced to just a means of attaining knowledge. D. Brent 211 Philosophical rhetorics have always been closely allied with systems of practical guidelines for persuasion. Because i t grounds knowledge in rhetorical interchange, the art of rhetoric as seen from the epistemic perspective is no less grounded in communication than i t was in Aristotle's day. This raises the question of whether this form of rhetoric with its vastly expanded inventional component can t e l l us anything new about how people can, in practice, communicate with each other more effectively. Are there any modern systems of practical guidelines that can be informed by this view of rhetoric? One example of such a set of guidelines can be found in the work of psychologist Carl Rogers.10 In his talk "Communication: Its Blocking and Its Facilitation," Rogers argues that communication tends to break down when the listener feels threatened. This sense of threat is reinforced by a communicator's tendency to evaluate, usually negatively, the opposing viewpoint. Rogers explains that Real communication occurs, and this evaluative tendency is avoided, when we listen with understanding. What does that mean? It means to see the expressed idea and attitude from the other person's point of view, to sense how i t feels to him, to achieve his frame of reference in regard to the thing he is 1 < 5In their 1970 textbook Rhetoric: Discovery and Change, in which Rogers' talk is reprinted, Young, Becker and Pike take Rogers' basic principles of mutual understanding and develop from them a mode of persuasion that they call "Rogerian Rhetoric." Their assertion that Rogerian rhetoric is totally different from traditional rhetoric has been severely criticised (see for instance Andrea A. Lunsford, "Aristotelian vs. Rogerian Argument: A Reassessment"). However, the basic principles that Rogers enunciates are closely related to the sort of dialectic inquiry on which the rhetorical model of reading is founded. Booth, for instance, though quarrelling with certain of Rogers' assertions, states that "this quibble does not affect the basic suggestiveness of his brief 'rhetoric'" (Modern Dogma xvii n.). D. Brent 212 talking about. (285) The deepened understanding of belief construction provided by a rhetoric of reading suggests ways of putting these principles into practice. The first principle of Rogerian rhetoric is to repeat the audience's ideas back to them in order to establish a basis of shared understanding; however, i t is often impossible to apply this principle because i t is impossible to hear another's ideas clearly. The divisions between human beings make every person's reading of a text a different virtual work. It is therefore beyond human power to achieve fully another person's "frame of reference in regard to the thing he is talking about." Yet an understanding of the relationships between text and virtual work, between doxai and persuasion, between logic, emotion and ethical influence, can help us find sources of identification and thereby at least partially f u l f i l l Rogers' ideal. If we can make educated guesses as to the sources of others' beliefs, if we can trace the relationships between others' texts and the doxai in which they are grounded, then we have a sounder basis for understanding and taking in opposing points of view. As with the teaching of the research paper, so with human communication in general, a rhetoric of reading suggests means by which identifications can be forced to grow out of the seemingly barren soil of division. Particularly important to this use of a rhetoric of reading are the insights that i t can provide into the balance between shared and unshareable components of discourse. It is essential to understand the role of individual personality, biography, and rhetorical situation in the construction of belief systems, for i t is these factors that enable us to account for the diversity of human belief. But as we have seen, i t D. Brent 213 is also important to understand the sources of constraint on belief that grow out of shared communal and universal values. These constraints—the conventional meanings of words, the similarities among interpretive communities that grow out of the commonalities of human experience, the assumption that systems of belief strive for maximum internal consistency—provide the basis for educated predictions as to how another will react to a given piece of discourse. 1 1 Conversely, they provide the basis for educated guesses as to the genesis and structure of others' belief systems. Thus a rhetoric of reading provides grounds for confidence not only in a multiplicity of ways of knowing but also in our ability to understand each others' ways of knowing. If we can understand the ways people can be lead to opposite conclusions by what appear to be the same facts, we have a basis for putting into practice the principles of mutual inquiry and exploration under which, in Wayne Booth's rhetoric, persuasion is subsumed. It is one of the fundamental, if often unspoken, assumptions of academic inquiry that to understand a process is to achieve some form of control over i t . If rhetoric can help us understand not just the way knowledge is transmitted but also, through an expanded account of invention, the way i t is made through verbal interaction, i t can help us understand how we build ourselves and our understanding of the world. If we can understand this, perhaps we can also understand how to sort through the seemingly infinite variety of jostling opinions that clamour x lAs pointed out in chapter 1, this assumption of predictability is an indispensible component of any rhetorical theory, for without i t there can be no reliable basis for one human being's verbal influence on another (9-10). D. Brent 214 for our assent and to identify in a principled way the ones that are worthy of having influence on our belief systems—to control, that is, in a non-arbitrary way the formation of our beliefs. In the conclusion of their treatise on practical argumentation, Perelman and Olbrechts-Tyteca suggest how far the implications of such understanding can reach: If freedom was no more than necessary adherence to a previously given natural order, i t would exclude a l l possibility of choice; and i f the exercise of freedom were not based on reasons, every choice would be irrational and would be reduced to an arbitrary decision operating in an intellectual void. It is because of the possibility of argumentation which provides reasons, but not compelling reasons, that i t is possible to escape the dilemma: adherence to an objectively and universally valid truth, or recourse to suggestion and violence to secure acceptance for our opinions and decisions. (514) That which Perelman and Olbrechts-Tyteca claim for argumentation—the production of arguments based on reasons—may also be claimed for its logical counterpart, a rhetorical account of the uptake of arguments. Only if we can understand how to base our beliefs on a logical, emotional and ethical response not just to propositions in the abstract but to others' belief systems can we hope to avoid the dilemma to which Perelman and Olbrechts-Tyteca draw our attention. The understanding provided by a rhetoric of reading can help us build belief systems based neither on absolute truth nor on arbitrary selection from equally valid propositions, but on good reasons and informed free choice. D. Brent 215 Works Cited Anderson, Richard C. "The Notion of Schemata and the Educational Enterprise: General Discussion of the Conference." Schooling and the Acquisition of  Knowledge. Ed. Richard C. Anderson, Rand J. Spiro, and William E. Montague. Hillsdale: Erlbaum, 1977. 415-31. Anderson, Richard C, et a l . "Frameworks for Comprehending Discourse." American Educational Research Journal 14 (1977): 367-81. Aristotle. Rhetoric. Trans. Lane Cooper. Englewood Cliffs: Prentice-Hall, 1932. Augustine, Saint. On Christian Doctrine. Trans. D. W. Robertson. Indianapolis: Library of Liberal Arts-Bobbs, 1958. Bain, Alexander. English Composition and Rhetoric. London: Longman, 1866. Beach, Richard, and Robert Brown. "Discourse Conventions and Literary Inference: Toward a Theoretical Model." Understanding Readers'  Understanding: Theory and Practice. Ed. Robert J. Tierney, Patricia L. Anders, and Judy Nichols Mitchell. Hillsdale: Erlbaum, 1987. 147-73. Beale, Walter H. A Pragmatic Theory of Rhetoric. Carbondale: Southern Illinois University Press, 1987. Bitzer, Lloyd. "The Rhetorical Situation." Contemporary Theories of  Rhetoric: Selected Readings. Ed. Richard L. Johannesen. New York: Harper, 1971. 381-93. Black, Edwin. Rhetorical Criticism: A Study in Method. Madison: U of Wisconsin P., 1978. Blair, Hugh. Lectures on Rhetoric and Belles-Lettres. Ed. Harold Harding. 2 vols. Carbondale: Southern Illinois UP, 1965. Bleich, David. Subjective Criticism. Baltimore: Johns Hopkins UP, 1978. Booth, Wayne C. Modern Dogma and the Rhetoric of Assent. Chicago: U of Chicago P, 1974. . The Rhetoric of Fiction. 2nd ed. Chicago: U of Chicago P, 1983. Brent, Douglas A. "Indirect Structure and Reader Response." Journal of  Business Communication 22 (1985): 5-8. Brostoff, Anita, and Barry K. Beyer. "An Approach to Integrating Writing into a History Course." Journal of Basic Writing 2 (1980): 36-52. Brown, Gillian, and George Yule. Discourse Analysis. Cambridge: Cambridge UP, 1983. D. Brent 216 Bruffee, Kenneth A. "Collaborative Learning and the 'Conversation of Mankind.'" College English 46 (1984): 635-52. "Social Construction, Language, and the Authority of Knowledge: A Bibliographical Essay." College English 48 (1986): 773-90. Bryant, Donald C. Rhetorical Dimensions in Criticism. Baton Rouge: Louisiana State UP, 1973. Burke, Kenneth. Counter-Statement. Berkeley: U of California P. 1931. . Language as Symbolic Action. Berkeley: U of California P, 1966. . The Philosophy of Literary Form. Berkeley: U of California P, 1941. . A Rhetoric of Motives. Berkeley: U of California P, 1950. Campbell, George. The Philosophy of Rhetoric. Ed. Lloyd F. Bitzer. Carbondale: Southern Illinois UP, 1963. Carey, Robert F, and Jerome C. Harste. "Comprehension as Context: Toward Reconsideration of a Transactional Theory of Reading." Understanding  Readers' Understanding: Theory and Practice. Ed. Robert J. Tierney, Patricia L. Anders, and Judy Nichols Mitchell. Hillsdale: Erlbaum, 1987. 189-204. Cicero. De Inventione. Trans. H. M. Hubbell. Cambridge: Loeb-Harvard UP, 1976. . De Oratore. Trans. E. W. Sutton and H. Rackham. 2 vols. Cambridge: Loeb-Harvard UP, 1976. Connors, Robert J. "The Rhetoric of Explanation: Explanatory Rhetoric from Aristotle to 1850." Written Communication 1 (1984): 189-210. "The Rhetoric of Explanation: Explanatory Rhetoric from 1850 to the Present." Written Communication 2 (1985): 49-72. Copi, Irving M. Introduction to Logic. 5th ed. New York: Macmillan, 1978. Corbett, Edward P. J. Classical Rhetoric for the Modern Student. 2nd ed. New York: Oxford UP, 1971. . Rev. of A Theory of Discourse, by James L. Kinneavy. Freshman English News 1 (March, 1972): 12. "The Topoi Revisited." Rhetoric and Praxis: The Contribution of  Classical Rhetoric to Practical Reasoning. Ed. Jean Dietz Moss. Washington: Catholic U of America P, 1986. 43-58. Corder, Jim W. "Rhetorical Analysis of Writing." Teaching Compositions: Ten  Bibliographical Essays. Ed. Gary Tate. Fort Worth: Texas Christian UP, 1976. 223-39. D. Brent 217 Crusius, Timothy W. "A Brief Plea for a Paradigm and for Kinneavy as Paradigm." Freshman English News 12.3 (1983): 1-3. "A Comment on '"that we have divided/In three our kingdom": The Communication Triangle and A Theory of Discourse.'" College English 49 (1987): 214-19. . "Thinking (and Rethinking) Kinneavy." Rhetoric Review 3 (1985): 120-30. Culler, Jonathan. The Pursuit of Signs: Semiotics, Literature,.  Deconstruction. Ithica: Cornell UP, 1981. D'Angelo, Frank J. "Modes of Discourse." Teaching Compositions: Ten Bibliographical Essays. Ed. Gary Tate. Fort Worth: Texas Christian UP, 1976. 111-35. De Beaugrande, Robert. "Design Criteria for Process Models of Reading." Reading Research Quarterly 2 (1981): 261-315. . "Text, Attention, and Memory in Reading Research." Understanding Readers' Understanding: Theory and Practice. Ed. Robert J. Tierney, Patricia L. Anders, and Judy Nichols Mitchell. Hillsdale: Erlbaum, 1987. 15-58. Eco, Umberto. The Role of the Reader: Explorations in the Semiotics of  Texts. Bloomingdale: Indiana UP, 1979. Ehninger, Douglas, and Wayne E. Brockriede. "Toulmin on Argument: An Interpretation and Application." The Rhetoric of Western Thought. James L. Golden, Goodwin F. Berquist and William E. Coleman. 3rd ed. Dubuque: Kendall, 1983. 377-86. Flower, Linda. "The Construction of Purpose in Writing and Reading." College English 50 (1988): 528-50. Fish, Stanley. Is There a Text In This Class? The Authority of Interpretive  Communities. Cambridge: Harvard UP, 1980. Ford, James E., and Dennis R. Perry. "Research Paper Instruction in the Undergraduate Writing Program." College English 44 (1962): 825-31. Fulkerson, Richard P. "Kinneavy on Referential and Persuasive Discourse." College Composition and Communication 35 (1984): 43-56. Gage, John. "An Adequate Epistemology for Composition: Classical and Modern Perspectives." Essays on Classical Rhetoric and Modern Discourse. Ed. Robert J. Connors, Lisa S. Ede, and Andrea A. Lunsford. Carbondale: Southern Illinois UP, 1984. 152-69. Gilbert, Nigel G., and Michael Mulkay. Opening Pandora's Box: A Sociological  Analysis of Scientists' Discourse. Cambridge: Cambridge UP, 1984. D. Brent 218 Goetz, Ernest T., and Bonnie B. Armbruster. "Psychological Correlates of Text Structure." Theoretical Issues in Reading Comprehension:  Perspectives from Cognitive Psychology, Linguistics. Artificial  Intelligence, and Education. Ed. Rand J. Spiro, Bertram C. Bruce, and William F. Brewer. Hillsdale: Erlbaum, 1980. 201-20. Golden, James L., Goodwin F. Berquist and William E. Coleman. The Rhetoric  of Western Thought. 3rd ed. Dubuque: Kendall, 1983. Grimaldi, William M. A., S. J. Studies in the Philosophy of Aristotle's  Rhetoric. Weisbaden: Hermes, 1972. Haas, Christina, and Linda Flower. "Rhetorical Reading Strategies and the Construction of Meaning." College Composition and Communication 39 (1988): 167-84. Hairston, Maxine C. "Bringing Aristotle's Enthymeme into the Composition Classroom." Rhetoric and Praxis: The Contribution of Classical Rhetoric  to Practical Reasoning. Ed. Jean Dietz Moss. Washington: Catholic U of America P, 1986. 59-78. . Successful Writing. 2nd ed. New York: Norton, 1986. Harnadek, Anita. Critical Reading Improvement. 2nd ed. New York: McGraw, 1978. Harrington, Elbert W. Rhetoric and the Scientific Method of Inquiry: A Study  of Invention. Boulder: U of Colorado P., 1948. Horner, Winifred Bryan. Rhetoric in the Classical Tradition. New York: St. Martin's, 1988. Hunter, Paul. "Paul Hunter Responds [to Timothy W. Crusius]." College  English 49 (1987): 219-21. "'that we have divided/In three our kingdom': The Communication Triangle and A Theory of Discourse." College English 48 (1986): 279-87. Iran-Nejad, Asghar. "The Schema: A Long-Term Memory Structure or a Transient Structural Phenomena." Understanding Readers' Understanding: Theory and  Practice. Ed. Robert J. Tierney, Patricia L. Anders, and Judy Nichols Mitchell. Hillsdale: Erlbaum, 1987. 109-27. Iser, Wolfgang. The Act of Reading: A Theory of Aesthetic Response. Baltimore: Johns Hopkins UP, 1978. Johnson, Nan. "Ethos and the Aims of Rhetoric." Essays on Classical Rhetoric and Modern Discourse. Ed. Robert J. Connors, Lisa S. Ede, and Andrea A. Lunsford. Carbondale: Southern Illinois UP, 1984. 98-114. "Reader-Response and the Pathos Principle." Rhetoric Review 6 (1988): 152-66. D. Brent 219 Kane, Thomas S. The Oxford Guide to Writing; A Rhetoric and Handbook for  College Students. New York: Oxford UP, 1983. Kennedy, George A. The Art of Persuasion in Greece. Princeton: Princeton UP, 1963. . Classical Rhetoric and Its Christian and Secular Tradition from Ancient to Modern Times. Chapel H i l l : U of North Carolina P, 1980. Kinneavy, James A. A Theory of Discourse. New York: Norton, 1971. Kintsch, Walter. "Contributions from Cognitive Psychology." Understanding  Readers' Understanding: Theory and Practice. Ed. Robert J. Tierney, Patricia L. Anders, and Judy Nichols Mitchell. Hillsdale: Erlbaum, 1987. 5-14. Knoblauch, C. H. "Intentionality in the Writing Process: A Case Study." College Composition and Communication 31 (1980): 153-59. "Modern Composition Theory and the Rhetorical Tradition." Freshman  English News 9.2 (1980): 3-17. Knoblauch, C. H., and L i l Brannon. Rhetorical Traditions and the Teaching of  Writing. Upper Montclair: Boynton/Cook, 1984. Kuhn, Thomas S. The Structure of Scientific Revolutions. 2nd ed. Chicago: U of Chicago P, 1970. Larson, Richard L. "Classifying Discourse: Limitations and Alternatives." Essays on Classical Rhetoric and Modern Discourse. Ed. Robert J. Connors, Lisa S. Ede, and Andrea A. Lunsford. Carbondale: Southern Illinois UP, 1984. 203-14. LeFevre, Karen Burke. Invention as a Social Act. Carbondale: Southern Illinois UP, 1987. Lunsford, Andrea A. "Aristotelean vs. Rogerian Argument: A Reassessment." College Composition and Communication 30 (1979): 146-51. Lunsford, Andrea A., and Lise S. Ede. "On Distinctions between Classical and Modern Rhetoric." Essays on Classical Rhetoric and Modern Discourse. Ed. Robert J. Connors, Lisa S. Ede, and Andrea A. Lunsford. Carbondale: Southern Illinois UP, 1984. 37-49. Mead, George Herbert. Mind. Self and Society from the Standpoint of a Social  Behaviorist. Ed. Charles W. Morris. Chicago: U of Chicago P, 1934. Mervis, Carolyn B. "Category Structure and the Development of Categorization." Theoretical Issues in Reading Comprehension:  Perspectives from Cognitive Psychology, Linguistics, Artific i a l  Intelligence, and Education. Ed. Rand J. Spiro, Bertram C. Bruce, and William F. Brewer. Hillsdale: Erlbaum, 1980. 279-307. D. Brent 220 Meyer, Michael. The Little, Brown Guide to Writing Research Papers. 2nd ed. Boston: Little, 1985. Miller, Keith D. "Martin Luther King, Jr. Borrows a Revolution: Argument, Audience, and Implications of a Secondhand Universe." College English 48 (1986): 249-65. Moffett, James. Teaching the Universe of Discourse. Boston: Houghton, 1968. Murphy, James J. Rhetoric in the Middle Ages: A History of Rhetorical Theory  from St. Augustine to the Renaissance. Berkeley: U of California P., 1974. Nelson, Jennie, and John R. Hayes. "The Road Not Taken: How The Writing Context Influences Students' Choices." Unpublished article read at the Conference on College Composition and Commmunication, St. Louis, 1988. Nelson, Katherine. "Cognitive Development and the Acquisition of Concepts." Schooling and the Acquisition of Knowledge. Ed. Richard C. Anderson, Rand J. Spiro, and William E. Montague. Hillsdale: Erlbaum, 1977. 215-39. North, Stephen. "Teaching Research Writing: Five Criteria." Freshman  English News 9 (1980): 17-19. Oakeshott, Michael. "The Voice of Poetry in the Conversation of Mankind." Rationalism in Politics and Other Essays. London: Methuen, 1962. 197-247. O'Banion, John D. "A Theory of Discourse: A Retrospective." College  Composition and Communication 33 (1982): 196-201. Orr, C. Jack. "How Shall We Say: 'Reality is Socially Constructed through Communication'?" Central States Speech Journal 29 (1978): 263-74. Pappas, Christine C. "The Role of 'Typicality' in Reading Comprehension." Understanding Readers' Understanding: Theory and Practice. Ed. Robert J. Tierney, Patricia L. Anders, and Judy Nichols Mitchell. Hillsdale: Erlbaum, 1987. 129-45. Perelman, Chaim, and L. Olbrechts-Tyteca. The New Rhetoric: A Treatise on  Argumentation. Trans. John Wilkinson and Purcell Weaver. Notre Dame: U of Notre Dame P, 1969. Piaget, Jean, and Barbel Inhelder. The Psychology of the Child. Trans. Helen Weaver. New York: Basic, 1969. Plato. Gprgias. Trans. W. C. Helmbold. Indianapolis: Library of Liberal Arts-Bobbs, 1952. . Phaedrus. Trans. W. C. Helmbold and W. G. Rabinowitz. Indianapolis: Library of Liberal Arts-Bobbs, 1956. D. Brent 221 Polanyi, Michael. Personal Knowledge: Towards a Post-Critical Philosophy. Chicago: U of Chicago P, 1958. Quintilian. Institutio Oratoria. Trans. H. E. Butler. 4 vols. Cambridge: Loeb-Harvard UP, 1980. Rorty, Richard. Philosophy and the Mirror of Nature. Princeton: Princeton UP, 1979. Rosenblatt, Louise. The Reader, The Text, The Poem: The Transactional Theory  of the Literary Work. Carbondale: Southern Illinois UP, 1978. Roth, Audrey J. The Research Paper: Process, Form, and Content. 5th ed. Belmont: Wadsworth, 1986. Rowan, Katherine E. "A Contemporary Theory of Explanatory Writing." Written  Communication 5 (1988): 23-56. Rosch, Eleanor. "Universals and Cultural Specifics in Human Categorization." Cross-Culteral Perspectives on Learning. Ed. Richard W. Brislin, Stephen Bochner, Walker J. Lonner. New York: Sage-Wiley, 1975. 177-206. Rosch, Eleanor, et a l . "Basic Objects in Natural Categories." Cognitive  Psychology 8 (1976): 382-439. Rumelhart, David E. "Schemata: The Building Blocks of Cognition." Theoretical Issues in Reading Comprehension: Perspectives from Cognitive  Psychology, Linguistics, Artificial Intelligence, and Education. Ed. Rand J. Spiro, Bertram C. Bruce, and William F. Brewer. Hillsdale: Erlbaum, 1980. 33-58. Rumelhart, David E., and Andrew Ortony. "The Representation of Knowledge in Memory." Schooling and the Acquisition of Knowledge. Ed. Richard C. Anderson, Rand J. Spiro, and William E. Montague. Hillsdale: Erlbaum, 1977. 99-135. Scholes, Robert. Textual Power: Literary Theory and the Teaching of English. New Haven: Yale UP, 1985. Schwegler, Robert A., and Linda K. Shamoon. "The Aims and Processes of the Research Paper." College English 44 (1982): 817-24. Scott, Robert L. "On Viewing Rhetoric as Epistemic." Central States Speech  Journal 18 (1967): 9-17. "On Viewing Rhetoric as Epistemic: Ten Years Later." Central States  Speech Journal 27 (1976): 258-66. Sherif, Carolyn Wood. "Comments on Chapter 8 by Wyer." Schooling and the  Acquisition of Knowledge. Ed. Richard C. Anderson, Rand J. Spiro, and William E. Montague. Hillsdale: Erlbaum, 1977. 289-304. Smith, Frank. Understanding Reading: A Psycholinguistic Analysis of Reading  and Learning to Read. 2nd ed. New York: Holt, 1978. D. Brent 222 Sperber, Dan, and Deirdre Wilson. Relevance: Communication and Cognition. Cambridge: Harvard UP, 1986. Spiro, Rand J. "Constructive Processes in Prose Comprehension and Recall." Theoretical Issues in Reading Comprehension: Perspectives from Cognitive  Psychology, Linguistics, Artifi c i a l Intelligence, and Education. Ed. Rand J. Spiro, Bertram C. Bruce, and William F. Brewer. Hillsdale: Erlbaum, 1980. 245-78. . "Remembering Information from Text: The 'State of Schema' Approach." Schooling and the Acquisition of Knowledge. Ed. Richard C. Anderson, Rand J. Spiro, and William E. Montague. Hillsdale: Erlbaum, 1977. 137-65. Spratt, Brenda. Writing from Sources. New York: St. Martin's, 1983. Todorov, Tzvetan. "Reading as Construction." The Reader in the Text: Essays  on Audience and Interpretation. Ed. Susan R. Suleiman and Inge Crosman. Princeton: Princeton UP, 1980. 67-82. Toulmin, Stephen E.. The Uses of Argument. Cambridge: Cambridge UP, 1958. van Dijk, Teun A. "Relevance Assignment in Discourse Production." Discourse  Processes 2 (1979): 113-26. van Dijk, Teun A., and Walter Kintsch. Strategies of Discourse  Comprehension• New York: Academic, 1983. Vygotsky, Lev Semenovitch. Thought and Language. Ed. and Trans. Eugenia Hanfmann and Gedrtrude Vakar. Cambridge: MIT, 1962. Weaver, Richard M. "Language is Sermonic." Language is Sermonic: Richard M.  Weaver on the Nature of Rhetoric. Ed. Richard L. Johannesen, Rennard Strickland, and Ralph T. Eubanks. Baton Rouge: Louisiana State UP, 1970. 201-25. "The Power of the Word." Language is Sermonic: Richard M. Weaver on  the Nature of Rhetoric. Ed. Richard L. Johannesen, Rennard Strickland, and Ralph T. Eubanks. Baton Rouge: Louisiana State UP, 1970. 33-55. "Ultimate Terms in Contemporary Rhetoric." Language is Sermonic:  Richard M. Weaver on the Nature of Rhetoric. Ed. Richard L. Johannesen, Rennard Strickland, and Ralph T. Eubanks. Baton Rouge: Louisiana State UP, 1970. 87-112. Whately, Richard. Elements of Rhetoric. Ed. Douglas Ehninger. Carbondale: Southern Illinois UP, 1963. Wyer, Robert S. "Attitudes, Beliefs, and Information Acquistition." Schooling and the Acquisition of Knowledge. Ed. Richard C. Anderson, Rand J. Spiro, and William E. Montague. Hillsdale: Erlbaum, 1977. 259-88. D. Brent 223 Young, Richard. "Arts, Crafts, Gifts and Knacks: Some Disharmonies in the New Rhetoric." Reinventing the Ehetorical Tradition. Ed. Aviva Freedman and Ian Pringle. Ottawa: Canadian Council of Teachers of English, 1980. Young, Richard, Alton L. Becker and Kenneth L. Pike. Rhetoric: Discovery and  Change. New York: Harcourt, 1970. Douglas Allan Brent PUBLICATIONS Brent, Douglas Allan. "Draft Editing in the Writing Centre." Writing and Reading/Theory and Practice Newsletter 2.5 (1983): 2-3. . "Audience, Role-Playing and the 'Non-Content' Course." The Writing Instructor 3 (1984): 37-45. . "Subverting Linear Structures." Inkshed 5.1 (1984): 2. . "A Comment on 'On the Neglect of Twentieth-Century Non-fiction'." College English 47 (1985): 80-82. — . "Indirect Structure and Reader Response." Journal of the  American Business Communication Association 22 (1985): 5-8. . "Using Generalizations: A Response to Rick Coe." Inkshed 4.4 (1986): 2-3. . "Trying Out Collaboration," Inkshed 5.5 (1986): 1-2. . "A Comment on 'Reading Nonfiction in the Classroom: From Theory to Practice. "' College English 48 (1986): 840-42. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items