JAC Volume 15 Issue 1

Over the past ten years or so, a significant number of scholars in composition and rhetoric have argued in support of what has come to be known as antifoundationalist theory. Although these scholars emphasize different aspects of antifoundationalism and its implications for teaching writing, they might agree with the basic distinction between foundationalism and antifoundationalism made by Stanley Fish in an article entitled "Anti-Foundationalism, Theory Hope, and the Teaching of Composition." In that piece, Fish defines foundationalism as "any attempt to ground inquiry and communication in something more firm and stable than mere belief or unexamined practice"; that is, in such "objective" things as "God, the material or 'brute act' world, rationality in general and logic in particular, a neutral-observation language, a set of eternal values, and the free and independent self” (342-43). Antifoundationalism, on the other hand, is the theory that

Foundationalists argue that despite a number of epistemological problems we can rely to a certain extent on such things as the "real world," the referential nature of language, and human rationality and intersubjectivity. After all, most of us live and act as if the real world exists and that we can depend on it; we never seriously doubt that in the foreseeable future the sun will continue to come up in the cast, and we depend on the fact that our homes will still be there when we come back from vacation-or if they are not, we assume that there will be clear signs as to why they are not: ashes, for example. Moreover, most of us rely on some sort of stable references for certain words. When we ask family members, friends, and waiters to bring us a glass of water, there is hardly ever any problem with what we mean: we are usually brought a glass of water. And when we argue certain points or express our opinions, ouraudiences often knowingly nod their heads or they get a fire in their eyes and try to interrupt us with opposing arguments. So it would seem obvious by everyday-ordinary common sense that "a material or 'brute act' world" exists and that some words in our language refer to objects in this world. And it would seem equally obvious that we share some form of rationality, some form of subjectivity, with our fellow human beings.

However, antifoundationalists argue that our understanding of the material world is mediated through our language and our individual historically conditioned and culturally determined world-and-life views and that as a result there is no necessary connection between the words of our language and what they refer to. And because our individual world-and-life views are so historically conditioned and culturally determined, we cannot assume that all forms of human reason and all forms of intersubjectivity have a great deal in common. As a result, antifoundationalists often seem to imply that our language and knowledge of the world is an infinite hall of mirrors, in which diction and syntax, rules of grammar, the neurological structure of the human mind, and what we call reality reflect upon one another with no ultimate source for all the shifting mirror images. Pressed to account for how we do seem to understand one another to some extent, antifoundationalists often rely on the idea of an interpretive community or social convention to explain how we know and understand and how we use language.

Now the discussion of foundationalism and antifoundationalism in our discipline strikes me as curious for two reasons. The first reason is the sheer one sidedness of the argument: scholars in composition and rhetoric are overwhelmingly antifoundationalist--this in spite of the fact that common sense would seem to tell us that antifoundationalism is extremely problematic. Indeed, scholars who study rhetoric in departments of Speech Communication have long argued the implications of a much wider range of philosophical positions than we have in composition and rhetoric (see for example Cherwitz).  These positions include realism, the belief that we can know and rely on our sense of an objective material world. The main arguments these scholars advance in favor of realism are the apparent givenness of our sensory experience and the necessity of a knowable objective reality in order for us to have any coherent sense of truth at all (Hikens).

In addition, scholars in literary theory, such as Gerald Graff and Kathleen McCormick, have pointed out the difficulties of relying on interpretive communities as a basis for a theory of knowledge: such a concept does not sufficiently distinguish between the kinds of strategies that people may use in understanding; nor does it explain how individuals within a community acquire these strategies or how they may move from community to community and develop new strategies. In addition, the philosopher Donald Davidson has made a convincing case against the very idea that conventions explain anything about language. And yet, despite its many conceptual problems, antifoundationalists have carried the day in our books and journals without any organized or serious opposition.¹ 

The second curious thing that strikes me about all the antifoundationalist talk in our discipline is that its relevance to the teaching of writing has never been very clear. Two early arguments using forms of antifoundationalism as a rationale for specific pedagogical practices-for teaching a self conscious- ness about the forms of discourse and for using collaborative teaching methods-have either been severely qualified or they have been shown to be very problematic. The latest version of antifoundationalism to be presented in our journals-an argument based on the work of Donald Davidson-suggests that writing cannot be taught in any systematic way, especially in schools. We might wonder then whether antifoundationalism has any relevance to the teaching of writing.

One way antifoundationalism has been used to justify a particular pedagogical practice in writing is exemplified in the work of Patricia Bizzell. In a 1982 essay entitled "Cognition, Convention, and Certainty: What We Need to Know about Writing," Bizzell argues for "teaching about academic discourse (the 'theoretical perspective on situationality') as a way of demystifying academic discourse and giving students more control over their use or rejection of academic discourse" ("Response" 243).² Fairly or unfairly, Fish seized upon Bizzell's article as one of a number of arguments which claim a natural affinity between antifoundationalism and "a process-oriented or rhetorical approach to composition," and which Fish in turn identifies with such things as "process over product, the replacement of a standard of correctness by the fluid and dynamic standard of effectiveness, the teaching of strategies rather than of rules and maxims" ("Anti-foundationalism" 346-47).

However, Fish argues that an antifoundationalist epistemology cannot be used to justify a process pedagogy because whatever we believe, we will always be situated in a particular context which will determine the language we use and how we communicate. If we subscribe to an antifoundational epistemology, we can only increase our self-consciousness of our own situatedness; we cannot transcend it:

It follows, according to Fish, that teaching our students about discourse conventions and writing strategies rather than, say, rules of grammar and conventional genre formats will make no difference in their ability to learn how to write because any particular piece of information has to be contextualized, and information about discourse conventions and writing strategies will in the end be just another kind of knowledge which our students will have to apply to particular situations. If antifoundationalism is correct in asserting that all knowledge is contextual, then even the awareness that we have to apply conventions and strategies in context is contextual and our awareness of that fact will not be helpful. In the words of the Nike ads, we just have to do it.

Thus, according to Fish, all instructors can do to teach writing is help students work through the problem of communicating in specific situations:

It is noteworthy that Bizzell now accepts this argument and agrees that "talk about academic discourse cannot be metadiscursive," although she does hold out the possibility that teaching students the conventions of academic discourse can "help students to understand it better, practice it more fluently, and work to change it more creatively, without having to claim that this kind of perspective on the discourse constitutes critical ‘distance' of the kind Fish disallows" ("Response" 243-44). In other words, Bizzell admits that she can no longer use antifoundationalism to privilege her pedagogy.

Another theorist, Kenneth Bruffee, uses antifoundationalism to provide a philosophic rationale for collaborative learning. In two major articles "Collaborative Learning and the 'Conversation of Mankind... and "Social Construction, Language, and the Authority of Knowledge: A Bibliographic Essay," Bruffee, relying heavily on the work of Fish, Thomas Kuhn, Richard Rorty, and Clifford Geertz, argues that "thought is internalized public and social talk," that writing is "internalized social talk made public and social again," and thus that knowledge is "socially justified belief" ("Collaborative" 651). Bruffee concludes that only some form of collaborative learning as a pedagogical method can capture the social nature of language and knowledge, that collaborative learning ought to be intrinsically better than other methods in teaching people to write. However, as I have pointed out elsewhere, because Bruffee argues that all language is social, it does not follow that collaborative learning is inherently more social or more able to capture the social nature of experience than any other pedagogical method. In Bruffee's own terms, lecturing is social and recitation classes are social, all methods of teaching writing are social, and we must therefore find additional reasons for supposing that collaborative learning is a superior way to teach writing.

Finally, another antifoundationalist, Thomas Kent, in a series of articles, argues that we may not be able to teach writing systematically at all. In "Paralogic Hermencuties and the Possibilities of Rhetoric" Kent follows Donald Davidson in arguing that all communication is a matter of interpretation. To Kent--and Davidson--each act of interpretation is so dependent on the idiosyncracies of a particular context that generalizations and systematic descriptions of how language works cannot help interpreters do their job. Interpretation is primarily a matter of sympathy, luck, and skill, not general knowledge:

In "Beyond System: The Rhetoric of Paralogy" Kent argues that the idea of convention cannot adequately account for interpretation. Again relying heavily on the work of Davidson, Kent demonstrates that no convention can identify completely how a particular sentence is employed, whether for example it is an assertion, a joke, or a question. Likewise, no convention can identify the intention of the speaker of an utterance or clarify whether that intention is sincere. As a result, according to Kent:

And in "Externalism and the Production of Discourse" Kent again follows Davidson's lead and argues that notions about thought and mental processes are less important to a theory of communication than the nature of "public and external communicative interaction" (66). Such public interaction involves Davidson's concept of charity, the presupposition that "an interpreter will assume that we intend him or her to hold our sentences true, and in a reciprocal manner, we will attribute beliefs to the interpreter that agree with our own" (65); his idea of triangulation, that public forms of interaction are the basis for internal states; and his idea of a passing theory, that communication requires "on the spot interpretation that cannot be reduced to a schematic cognitive process or to any kind of epistemological system" (67).

How then, according to Kent, should we go about teaching writing? In "Paralogic Hermeneutics" he resorts to Fish's advice that students just have to practice: "A student may only learn how to employ his background knowledge--and learn how to expand it at the same time--by entering into specific dialogic and therefore hermeneutic interactions with others' interpretive strategies" (37). These dialogic interactions cannot be artificial. Teachers cannot be mock audiences, for they do not have enough background in specific areas to hold up their side of the dialogue. Likewise, case study methods are inadequate because they are too detached from the conversations in specific disciplines and contexts. Real dialogic interactions must involve, Kent says, an opportunity for the writer to engage in collaboration with others in the context of specific disciplines. However, discipline-specific courses are not enough: "no practical writing course like technical writing can teach an engineering student to write" (39). This seems to suggest that students can only learn to write like engineers by being engineers, an idea which may be true but which begs the question of how students learn to write as engineers.

Finally, in "Externalism" Kent opposes process-oriented instruction and dialectical methods that lead students to "the truth." Such methods, he says, can be used to teach a body of knowledge, such as grammar, methods of analysis, and theories of communication, but they cannot be used to teach writing. His only suggestion is that we "drop our current process-oriented vocabulary, and begin talking about our concrete social and public uses of language" (70). The problem, of course, is that talking about such social and public uses of language may become another form of knowledge, which Kent has ruled out ahead of time as being useful for learning how to write. We seem to be back to Stanley Fish's position: students can only learn to write by doing it, albeit in the context of some specific real-world domain for which academic disciplines are not an adequate substitute. One possible interpretation of Kent's position is that writing cannot be taught in schools.

However, we may contrast Kent's conclusions with another compositionist sympathetic with Davidson's theories, Reed Way Dasenbrock, who reaches a conclusion much different from Kent's. From studying Davidson, particularly Davidson's notion of triangulation, Dasenbrock concludes that learning the public forms and conventions of discourse, presumably in a systematic way, is absolutely essential. Dasenbrock quotes Davidson--"thought is necessarily part of a common world"--and goes on:

So according to Dasenbrock, we must teach rules, conventions, received forms--the knowledge of these things--and rather rigorously and systematically at that, if only to give our students the choice of either following the rules or breaking them. It would seem then that even among dedicated antifoundationalists there is no unanimity about the implications of the theory for particular practices.

What then are we to make of all this? Is antifoundationalism convincing when it argues for a hall of mirrors and against the reliability of the material world, human rationality, and the referential nature of language? And whether it is convincing or not, does antifoundationalism have any practical implications for the teaching of writing?

The major difficulty in discussing these matters is that the conversation about antifoundationalism in our discipline has been greatly oversimplified. We have been presented with only a few of the many arguments involved in a major philosophical problem that goes back to ancient times and is still intensely controversial in current philosophical circles. The reasons for this state of affairs in composition/rhetoric, I think, are two: the relative newness of our discipline and our overwhelming desire and need to be practical, to apply philosophical positions to the business of teaching writing. As a result, we do not have the tradition of philosophical inquiry that exists, say, in departments of Speech Communication, and the tradition we do have is skewed by our desire for practical applications. In our desire to apply philosophical theory to the teaching of writing, we tend to look for metaphors for instruction rather than to analyze a particular theory for its strengths and weaknesses.

Although I am sympathetic to what we generally call antifoundationalist positions, it seems to me that antifoundationalist theory faces a number of major difficulties, difficulties with which it has not yet come to grips. For one thing, it has not yet given an adequate account of how we rely on the material world and how we learn and use language to refer to that world in ways which are consistent and useful. As Davidson has put it, ". . . there must be some connection [between belief and truth] if we are to relate the truth of utterances to their use. 'Be question is what that connection can be" ("Structure" 305). For another, if antifoundationalism is to have any relevance to composition and rhetoric, it must offer some convincing suggestions about how we ought to teach writing, suggestions which seem to be organic or integral to the theory.

So in the interest of furthering the conversation about antifoundationalism, I would like to address both of these issues: First, the problem of relating what we know, our beliefs, to the material world, using language. Davidson's theories are an excellent point of reference--it is no accident that he has been getting so much attention in composition and rhetoric--because his work is at the cutting edge of current philosophical thinking on matters of language, knowledge, and belief.  What I find encouraging about Davidson's work--but what has not been sufficiently pointed out, however--is the degree to which Davidson tends to blur the distinction between foundationalism and antifoundationalism and seek solutions that transcend these pigeonholes. Nor has much attention been given to the problems in Davidson's theories. Although Davidson has made a strong case for relying on our sense of an external reality--primarily by arguing against traditional empiricism and the notion that we have internal conceptual schemes that mediate reality--he has yet to go the next step and specify more concretely what the connection can be between belief and truth. So I would like to point out the difficulties of Davidson's position and still note that his solution may be the best we can hope for.

Secondly, I would like to explore how we use philosophical theory in composition/rhetoric as a model or metaphor for instruction. Since there may be no other way to use theory in praxis, I offer Davidson's theory of triangulation as a more fruitful model in ways that Kent or Dasenbrock have yet suggested. And I note that Davidson's reliance on theories of language acquisition to explain his theory of triangulation may be a profitable suggestion for us in continuing to think about these matters.

Davidson's Theory of Triangulation

Davidson is not the only contemporary philosopher who is trying to account for how we know about "the material or 'brute act' world"-hence, such positions as internal realism, unlimited realism, metaphysical realism, and any number of theories as to how language reflects both the material world and our beliefs (see for example Putnam, especially chapters 1-3). The arguments for or against the various positions have such a long and complicated history that most philosophers settle on a position as lightly as a bee on a flower, carefully noting the difficulties of their chosen position and delicately assaying its strengths and weaknesses.

Among contemporary philosophers, however, Davidson is somewhat unique in that his theories finesse the entire problem of how much we can trust our sense of the material world. Although he calls himself a "nonfoundationalist" ("Myth" 167), Davidson assumes that the world is in some sense knowable-it is just not knowable in any way that will convince a skeptic--and that epistemic views--that is, views in which our perception and language constitute the world--are false ("Structure" 304).

The main theory which Davidson offers to explain how we know the world is what he calls triangulation:

In short, triangulation is a theory of how human beings come to use words to refer to a common reality in a similar way: only by using words in response to a common "stimuli" can they get any sense of how another person might be using the same words in the same way. 

Now Davidson's notion of triangulation carefully sidesteps two major theoretical problems. First, by asserting that language is one way we respond in common to external stimuli, he avoids the problem of reference, the precise way in which language reflects reality. Note that Davidson does not say that the objects of our perception correlate with any particular aspect of our language, that there is any necessary correspondence between the stimulation of our senses and any objects in the material world. He only argues that human beings share a way of talking about common stimuli in common situations. Thus, in Davidson's theory, when we use language we may be using words to refer to vastly different things in vastly different ways. All we know for certain is that we share a common language in certain situations and that this common language gives us some sense of the material world. Just what that sense is Davidson does not say, although he says we can rely on it.

Secondly, in developing a holistic theory about common stimuli in common situations, Davidson avoids the problem of empiricism and how we develop concepts, our "picture of the world," out of more primitive sense data. In effect, Davidson rejects the entire empirical tradition of grounding our knowledge in what we receive through our senses. He will only assert that our language is based in common experience, not in sense data. Thus, Davidson also rejects the notion that we constitute the world by building up our own internal beliefs as if we were putting together a grand mental structure using external sense data as the basic pieces. In fact, he argues the opposite view: that our internal beliefs and the language we use are somehow directly related to the beliefs and language of others. Again, just how our language and beliefs are related he does not say, but he is adamant that we can assume they have something in common because of how we learn language in the first place and use language together in common circumstances.

One great advantage of Davidson's theory is that he can refute an argument made by a number of antifoundationalists that if the world is independent of our minds, it is theoretically possible that all our beliefs could be false and or that there may be "one, complete, true description of the world against which all other accounts are to be measured" (Crumley 363). On the contrary, Davidson argues that we hold the beliefs we do because, as Jack Crumley puts it, "at least some of those beliefs are conditioned to events and objects in the world" and this conditioning is reflected in our language (365). Crumley goes on:

Another great advantage of Davidson's theory is that it rejects any notion that something comes between, as it were, our experience of the world and our beliefs about that world. No socially-conditioned conceptual scheme or individual power of interpretation or self-awareness mediates between our experience of the world and our understanding of it. All we need for an adequate theory of language is the notion of triangulation: that the way we learn language in common situations reflects something truthful about the world.

Anyone with thoughts, and so in particular anyone who wonders whether he has any reason to suppose he is generally right about the nature of his environment, must know what a belief is, and how in general beliefs are to be detected and interpreted. These being perfectly general facts we cannot fail to use when we communicate with others, or when we try to communicate with others, or even when we merely think we are communicating with others, there is a pretty strong sense in which we can be said to know that there is a presumption in favor of the overall truthfulness of anyone's beliefs, including our own. So it is bootless for someone to ask for some further reassurance; that can only add to his stock of beliefs. All that is needed is that he recognize that belief is in its nature veridical. ("Coherence" 314)

However, the fundamental weakness of Davidson's theory, as Richard Rorty is at great pains to point out, is that it does not give an adequate description of how language and the material world are related. Davidson takes truth as "primitive" ("Coherence" 308). Rorty glosses Davidson's argument this way:

On the other hand, if Davidson is correct--and I think his solution to the problem of the relation of truth and belief is very attractive--we need not be as dogmatic about our antifoundationalism as Fish, Bruffee, and Kent suggest. As Wittgenstein pointed out earlier, we can choose to play a believing game or a doubting game, and no amount of evidence can convince a person who insists on playing the doubting game. Indeed, there is a school

of what David Ray Griffin and others call "constructive postmodern philosophers"--Peirce, William James, Bergson, Whitehead, and Hartshorne--who argue that we should trust our basic beliefs about the world primarily because there is little practical reason to doubt them:

This is not foundationalism in Fish's definition by any means--it does not guarantee ultimate "objectivity,” whatever that might be--but it does suggest that our "mere beliefs"' are more firm and stable than radical antifoundationalists suggest. Although such a position may not be sufficiently "grounded" to convince skeptics, it may be all we can ask of any theory of the relationship between belief, language, and the material world. The theory of triangulation does not provide a necessary and sufficient ground for us to trust our language, our reason, and our sense of the world, but it does suggest we have good reasons for doing so.

Triangulation as a Metaphor for Instruction

The primary way in which theorists in composition/rhetoric seem to use philosophical theories is as models or metaphors for how we ought to teach. Each of these theorists argue that if a certain theory of language and belief is convincing in accounting for how we communicate, then we ought to teach writing using methods modeled after that theory:

Now it is important to recognize that even if a theory seems to accurately describe some aspect of the way language works or the way language reflects the material world, it does not necessarily follow that the best way to teach is to model instruction after that theory. A certain state of affairs may not necessarily be the best model for how to teach others about that state of affairs. Similarly, a certain kind of behavior may be a goal of instruction, but simply having students engage in that behavior may not necessarily be the best way to teach it.

For example, Fish may be right that in an ultimate sense all knowledge is contextual, that we cannot escape being "situated." But it does not follow that because all knowledge is contextual and we cannot escape being situated, ho particular awareness of our situation will ever help us in learning to write. Our knowledge of our particular situation may be limited, but what we know and how we apply that knowledge will have consequences, and we can prefer and choose one situation over another.

Take for instance, Fish's famous example of the sign on the door at the Johns Hopkins University Club: PRIVATE MEMBERS ONLY (Text 275). If we were to hear these words spoken aloud with the usual failing intonation after the word "private," most of us would spontaneously understand them to mean that only certain members of the Hopkins faculty are allowed in the room behind the door. And even without that spoken intonation, I suspect that most of us, passing quickly down the hall at Hopkins and glancing at the door, would understand the sign in the same way. But with a little training and practice in noting multiple meanings, especially in written language where we do not have the help of intonation, we can start to understand things differently. We begin to notice the ambiguity, the indeterminacy of the language we use all the time--with the same spontaneity that we used to understand things with single meanings. In the case of the sign on the door at Hopkins, we can understand the sign in any number of more suggestive and comic ways.

Now Fish would argue that in learning to understand and notice the indeterminacy of language, we have not transcended our situation; we have simply substituted one situation for another. And this is true. But I would argue that being able to recognize the ambiguities and indeterminacies of language is a superior position to be in. Not all situations are of equal value; knowing that language is indeterminate and often ambiguous can help us to recognize indeterminacies and ambiguities. And in many cases, for many people, it is better to be able to recognize such ambiguities than not. For many people in many situations, it is better to know more than less; it is better to recognize a range of possible meanings than to be limited to one-o course, not for everyone in every situation, but for many.

The same argument holds for Bizzell's teaching her students to be aware of the conventions of academic discourse. Such talk may not be "meta-discursive" or privileged in any ultimate sense, but it may produce the results she wants: students who are more self-aware about what they are writing and who have a better sense of how to go about manipulating the conventions of academic discourse for their own purposes. Bizzell's only difficulty may have been that she felt compelled to justify her instruction on something metaphoric of Fish's theory in the first place.³

Similarly, Kent's notion that writing cannot be taught systematically may be the result of Kent's trying to apply Davidson's theory of radical interpretation a little too literally. Even if we do learn to write in ad hoc and unsystematic ways, it does not necessarily follow that we should not teach writing systematically. Indeed, it seems intuitively obvious that systematic instruction might make unsystematic learning easier simply by providing a Wider range of helpful information, guidance, and response over time. There is no reason why we cannot learn a host of things we bring to every writing situation in a "systematic" way, among them the way the language works and a knowledge of what we might call the principles of communication: the importance of context, various strategies for both producing prose and understanding certain kinds of discourse. And we can learn how to apply 'What we know systematically by practicing how to use different rhetorical strategies in different situations.

In short, we can insist on applying particular philosophical metaphors With such rigor that we ignore other aspects of our experience. The trick may he to keep our metaphors as rich and complex and true to our experience as We can make them. One of the major conclusions of Fish's antifoundationalism is this: If we know things to be true "only within the precincts of the contexts or situations or paradigms or communities that give them their local and changeable shape" (344), then arguments cannot be conducted primarily by referring to some independent  “objective" reality; rather arguments can only he conducted by using persuasion, appealing to the "local and changeable" Warrants and beliefs of those we want to persuade. When we choose a particular method of teaching writing, then, the question naturally arises: what makes a particular method convincing? If we decide with Bizzell that our students need to know the "meta-discourse" of academic prose, if we decide with Bruffee that our students need practice in collaborative learning, 1 t we decide with Kent that our students need to write primarily as practitioners of a particular discipline, on what grounds do we make these decisions?

Obviously in any particular situation we make these decisions for a host of reasons, dependent on our background and experience, our situation, the nature of our curriculum, the kinds of students we are teaching, the resources ht our disposal. But it may be useful to think that we use any particular teaching method primarily for one of two reasons. First of all, we may find The theory on which our teaching is metaphorically based to be convincing; that is, we intuitively recognize that, say, Fish's description of our situatedness is inherently true to our experience, and we want to capture the truth of that experience in our teaching. And secondly, we apply a particular method over time, and after a while we recognize that in some way it works; that is, we recognize that it actually produces good writers or good writing. What strikes me about the scholarship in favor of antifoundationalist pedagogics is the remarkable lack of much discussion about whether these methods of instruction actually produce good writers or good writing.

Now I am not suggesting that antifoundational writing theorists need to justify their pedagogics by using the empirical methods that were popular in our discipline in the 1950s and 60s. I recognize that empirical studies are certainly not "objective" and have their own theoretical problems. But it seems to me that ever since Stephen North, among many others, pointed out the conflicting assumptions and methods of various schools in our discipline--including the "lore" of practitioners who recognize in their students' work the success of how they have been teaching-we have been very uncomfortable talking about what makes our instruction successful, preferring to talk about philosophical metaphors rather than how we recognize that our students have become better writers or that their writing has improved. I sense that Bizzell, Bruffee, and Kent know that their methods are effective. They just have not yet discovered a language to talk about their successes with broad acceptance in our discipline.

I wonder if philosophical metaphors are any more appealing to the wide range of practitioners in composition/rhetoric than various empirical methods of research or even practitioner lore, indeed any language that talks about how we recognize the success or failure of our instruction. In any case, it seems to me that sooner or later we will have to again bring back the success or failure of our instruction as a major part of the arguments in favor of various pedagogics in our discipline, whether they are antifoundationalist or not. After all, philosophical theories can be applied in many different ways with many different results, and any number of theories may produce methods of instruction which seem to those involved to be successful. The question is whether we can develop a way to talk about our successes and failures which appeals to a broad spectrum of scholars and practitioners.

Here again I find Davidson suggestive. Davidson has looked to theories of language acquisition to buttress his theory of triangulation:

Needless to say, the whole story cannot be this simple. On the other hand, it is hard to believe that this sort of direct interaction between language users and public events and objects is not a basic part of the whole story, the part that, directly or indirectly, largely determines how words are related to things. ("Myth" 163)

Most theories of language acquisition now emphasize the nature of the active learner who develops internal hypotheses about how language works and constantly tests them against her experience. Young learners--or learners of a second language, for that matter--must work out their own private sense of the way language works, gradually approximating the language of adult speakers. Many composition theorists, such as Frank Smith and James Britton, have applied theories of language acquisition to how we learn to write.

What is interesting about Davidson's use of language-acquisition theories is that he invokes the empirical tradition to suggest that his theory of triangulation may be valid and useful. In doing so, he suggests that we too may combine the best of our knowledge and experience from a wide range of philosophical assumptions and points of view in thinking about how writing should be taught. For example, both the theory of triangulation and theories of language acquisition suggest that what will be most helpful in learning to write and in teaching writing is attention to those aspects of the world, those aspects of human behavior, those conventional aspects of discourse which we can name together and use in common ways. Thus, our instruction ought to pay overt and systematic attention to the ways we correlate our writing behavior with "the events and objects [we find] salient in the world" (Davidson, quoted in Kent, "Externalism" 65) and the common conventions of language and written discourse we find salient in our communication.

If this way of looking at Davidson seems useful, we may think of the concept of triangulation as philosophical support, a useful metaphor, for justifying the instructional methods of a number of composition theorists in the empirical tradition: those promoting sentence-combining (see Kinneavy) and those using what George Hillocks calls the environmental mode and focus on inquiry. Each of these methods focuses on activities which clarify a target concept or an aspect of written discourse-an object or event salient in the world or a convention salient in discourse-before providing for controlled application and practice using the target concept or aspect of writing. And like Davidson, we need not worry about whether these target concepts are "objective"; we need only recognize that such objects and events are what other people refer to when they communicate, that such aspects of discourse are indeed what people use when  they communicate, and that novice writers need to have these things made salient for them so that they too can participate in the process of triangulation, of developing shared concepts and practices.

One major reason why the pedagogics noted by Kinneavy and Hillocks have not received more broad-based support in composition/rhetoric may be that their persuasive appeal is based on research in the empirical tradition of pre- and post-tests, a language which many of us view with suspicion. Perhaps if these theorists based their appeals on philosophical metaphors, they might make their pedagogics more attractive.

The larger question for our discipline to address is this: What kind of evidence would we accept that a particular pedagogy was successful? Any theory of language and communication will have its strengths and weaknesses; the point is to recognize the value of any particular theory and how it attempts to capture the truth of our experience. Ways of talking about the success and failure of our instruction, theories of evaluating writers and Writing, are by no means objective and privileged, but they do provide us with one way to get at the heart of our instruction: whether it works. I doubt if any Argument in favor of a particular method of instruction will win broad-based support in our discipline if it does not provide both a philosophical justification and evidence of its successful application.

Conclusion

The fascination of compositionists with antifoundationalism may or may not prove to be very useful for the teaching of writing in the long term. I suspect that most of us do not feel lost in a hall of mirrors: we feel comfortable in the material world and trust that our perceptions about it are generally true. We alsomanage to understand one another quite well enough, thank you, and whatever the quality of our instruction, our students do manage to learn to write, although of course some of them learn to write better than others. Although I too am fascinated by the arguments in favor of antifoundationalism, I believe that as a discipline we need to seriously examine the claims and arguments of the wide range of antifoundational theories and we need to evaluate their relevance to the teaching of writing. Most of the varieties of antifoundationalism are still a matter of lively debate in professional philosophy. They ought to be matter of lively debate in composition/rhetoric too.

In addition, I think we need a great deal more discussion of the metaphors we use to guide our instruction and the ways in which we can evaluate their effectiveness. Fish refers rather sarcastically to "theory hope," the belief that theory can help us out of the hall of mirrors. Davidson's theory of triangulation may not provide a necessary and sufficient rationale for believing in the objectivity of a material world, the language we use to talk about that world, and our own rationality; but the theory of triangulation does give us reason to trust the experiences we have in common, to believe that there is something there for our language to capture, and to trust in some small way the language we develop to talk about those situations. The question remains whether the theory of triangulation or any antifoundational philosophy can provide us with a useful language for talking about writing or provide us with convincing reasons for teaching a particular way in the hall of mirrors.