Russell A. Hunt
St. Thomas University

A Horse Named Hans, a Boy Named Shawn:
The Herr van Osten Theory of Response to Writing

[as published in Writing and Response: Theory, Practice, and Research, ed. Chris M. Anson. Champaign-Urbana: National Council of Teachers of English, 1989. 80-100.]


It was in 1981, at a session during a meeting of the Canadian Council of Teachers of English in Halifax, that I first learned the term "whole language" and encountered the work of a writer whom I know only as Shawn. One presentation there consisted of a series of overhead projector transparencies which amounted to a case study of Shawn's writing development, over his first year of school. (Shawn's development has since been described more fully by Judith Newman in The Craft of Children's Writing [1984].)

In Shawn's class, from the very beginning, a certain amount of time every day was set aside for writing. The children could write anything they wanted to; at the outset, of course, many of them copied material from the writing which surrounded them in the classroom -- bulletin boards, blackboards, posters, and so forth. In each case the teacher asked the child to read what he or she had written and then responded in writing, reading aloud as, or after, she wrote. She never corrected the child's writing; she never marked an error. She responded to what the child's writing said -- or, indeed, to what the child said it said, for in many cases only the child could read it.

Shawn's earliest productions were pretty distant copies of language from the classroom environment; sometimes he knew what they said, sometimes not. Shawn was a child who had not grown up in a very literate environment, and his acquaintance with written language was limited. Within a few sessions, however (under the pressure, it seems probable, of wanting to make contact with his teacher), he began to produce his own statements, about things like the earache he'd had the night before or the partridge his father caught in the greenhouse. Occasionally, outside the context of that situation, these statements were virtually incomprehensible to those of us looking at the overhead. For example:

my Father ct a prtr
he fat it in the grenhs
His teacher, however, accepted his reading of what he'd written, and invariably wrote a genuine response to it, like this one: And eventually Shawn began responding (this, in October, was the first occasion on which he did so): As the successive overheads were displayed, we followed Shawn's growth as a user of language. We watched his "discovery" of punctuation -- for instance, the grotesque and marvelous dot, a full eighth of an inch in diameter, which was his first period. We watched as the length of Shawn's utterances grew. We saw his spelling strategies develop. We watched him taking risks with his spelling, using words that he needed rather than restricting what he said to what he could spell. For instance, at one point, when he wanted to tell his teacher about the snake he'd found, he needed the words "aquarium" and "veranda" (I particularly admired the silent e's, as well as the sense of story signaled by his conclusion). The clearest patterns in Shawn's work -- patterns that jump out in even the most cursory examination of his year's worth of writing -- are the increasing length of his utterances, the way he becomes more engaged with what he's writing, and the growing sophistication of his strategies for spelling, punctuation, and other "mechanical" matters.

It is important to make clear, incidentally, that this is not to say that his spelling and punctuation became more conventional; rather, the strategies one might infer from his texts were becoming more sophisticated. On March 11, for example, he produced the following text:

To his teacher's written response ("Why did you miss the bus at the bus-stop Shawn?"), Shawn answered: The advance here is partly in the amount of thought which seems to be involved in the spellings ("mrss" for "Mrs.," for example) and even more in the growing complexity of the syntax, and the sheer length of the utterance. As Newman (1984) has pointed out, Shawn's concern for meaning shows in his willingness to take a stab at a word to "place-hold the meaning" in order to get his idea down, even though with reflection he might be able to spell the word conventionally (consider the two appearances of both "today" and "because" in this text, for instance).

Two things were profoundly illuminating to me in that presentation. One was watching a child learning language through using language -- and learning things I'd thought would normally be learned only when they were overtly taught. The second involved the apparent attitude of Shawn's teacher, and the other people at the session, toward what I might have been tempted to call Shawn's gross errors if I hadn't just read Mina Shaughnessy's Errors and Expectations: A Guide for the Teacher of Basic Writing (1977). The teachers there didn't treat errors either as sins to be corrected or as something cute to be chuckled over; they treated them as evidence of principled, strategic thinking on Shawn's part, and as a promise that Shawn could continue his constructive, rational, active learning of the principles of written communication. They treated them, in other words, as hypotheses in the process of being tested.

Most vividly, perhaps, besides Shawn's wonderful, graphically and syntactically adventurous writing, I remember a man who sat stolidly in the middle of the classroom, watching Shawn's struggles and victories with tolerant amusement. He asked at the end a question whose exact words I can almost remember: "This is all very interesting. But at what point do you start pointing out his errors so that he can understand the importance of correctness?"

At the time, I found it amazing that anyone could have so thoroughly missed what seemed to me the main point of Shawn's (and every child's) miracle: We don't learn language by having our errors pointed out and corrected; we learn as a by-product of using language in order to do things we care about doing. This principle is one of the fundamental axioms underlying what has come to be called "whole language" pedagogy. It seemed clear -- and has come to seem even more so since -- that this principle has profoundly important implications for the way I think about what I am doing as a teacher of writing and as a teacher of literature.

Whole Language

The principles of whole language pedagogy are derived from the ideas of an eclectic assemblage of writers about language and reaming, at least some of whose names will be familiar to anyone who has been keeping up with developments in research on composing, on learning to write, and on the teaching of writing. They include the reading theorists Kenneth S. Goodman and Frank Smith and the linguist M. A. K. Halliday; going further back, much seems to be based on, or to grow out of, the American pragmaticism of C. S. Peirce and his successors -- most centrally, John Dewey -- and the linguistic psychology of L. S. Vygotsky. Whatever their sources, however, the ideas seem fundamentally quite simple and even commonsensical; it is only as they are wholemindedly applied to practical learning situations that their implications become new, and dramatic.

Among recent publications that offer an introduction to, and overview of, what we might call the theory of whole language pedagogy is a description of a massive study called "Children, Their Language and World," by Jerome Harste, Carolyn Burke, and Virginia Woodward, which appeared under the title Language Stories and Literacy Lessons (1984). Another is a collaborative account of a group of whole language classrooms: Whole Language: Theory and Practice, edited by Judith Newman (1985). In the absence of the kind of massive empirical and anecdotal detail which undergirds the conclusions of the Harste, Woodward, and Burke report, I think Shawn's story is a good way to begin understanding the kinds of observations of children in learning situations that were involved. From the innumerable learners like Shawn observed by Harste, Woodward, and Burke, and by Newman and her colleagues, some fairly straightforward ideas emerge.

A primary principle -- indeed, almost a slogan for the "movement" -- is that literacy learning is a natural phenomenon. There is some difficulty, of course, with terming any element of human culture "natural"; what it seems to mean in this context is that it is a process, like the development of oral language, that occurs virtually universally in a context which supports it and, like the learning of oral language, without conscious, institutionalized cultivation.

An equally important principle is that language is learned in use.The pragmatics of language use motivate and shape the learning of language, both the development of first-language ability in infants and that of written-language ability in older children. That this is true could hardly be more strongly attested than it is by the work of Kenneth Kaye (1982) and others on the development of dialogue between infants and caregivers, by the observations of M. A. K. Halliday (1975) on the development of language in his son, or by more classically oriented developmental psychology, such as that of Elizabeth Bates (1976a, 1979) or Jerome Bruner (1977, 1982).

A consequence of accepting these principles is that we are moved toward thinking of language as primarily a "top-down" process. That is, those smaller units which we have traditionally thought of as the "more basic" elements of language are not in fact "basic" at all. The meaning -- indeed, the very existence -- of each element of language is determined by the larger elements surrounding it. This is a principle that operates through the whole hierarchy of language phenomena. Sounds within words are determined by the words (ng, for instance, represents a different sound in "linger" than in "singer"); the meanings of words are determined by their sentences ("bear" can be one of a range of quite different words depending on the sentence in which we find it); the meanings of sentences are determined by the discourse around them ("It's getting chilly in here" can be an observation or a request or even a command, depending on the context in which it's uttered); an entire piece of discourse can be embedded in a context that alters it utterly (Polonius's advice to Laertes sounds rather more sensible outside the context of Hamlet, where Polonius is presented as a pompous fraud).

This, in turn, suggests that analysis of language into smaller and smaller units, while it may be of use to the metalinguistic understanding of some patterns of language use, and may help a teacher in understanding the way language is structured, is not likely to be very functional from the language learner's point of view. All of this forces us toward the conclusion that language needs to be experienced and dealt with in a "whole" condition in order for language learning to occur effectively. If we separate words from the contexts by which they are determined, if we pull sentences out of discourses, if we disengage discourse from the context of use and human purpose, we tend to produce something I call "textoids," synthetic fragments of language which exhibit none of the complex richness of natural language (Beaugrande [1982] remarks the error of working with what he calls "fragmentary and inane" texts.) And it is precisely this richness that enables us to function as the miraculous learners of language we all are, to navigate as effortlessly and unselfconsciously around the hermeneutic circle of understanding as we all must in order to understand any system of signs.

What all of this amounts to is an inversion of what I had always thought of as the most commonsensical model of language and language development. Such a model entails an assumption that it is a structure built up from the bottom, assembled out of basic linguistic elements like phones and phonemes. On the contrary, what this new view would suggest is that it is human purposes, social and interpersonal -- "pragmatic" -- purposes, that are the basic elements of language. Perhaps even more important, it creates a model which is so rich and complex that it seems foolhardy to suggest that discursive analysis of it will facilitate learning in any simple, direct way.

The most superficial survey of the kind of work that has been done in child language development since the publication of Roger Brown's landmark A First Language (1973) will demonstrate the radical changes that have occurred in our view of what is "basic" about human language learning. Elizabeth Bates (1976b) has described how we have moved away from Chomsky's (1965) miraculous and mythical innate predisposition to learn grammar -- the so-called language-acquisition device, posited to account for children's development of syntax, within a system whose largest allowable element was the sentence. We have come a long way toward the Vygotskyan (1962, 1978) view that meaning and intention come first, that language develops out of the patterns of social transaction between the infant and the people around the infant. And it has become more and more clear how much there is to learn about the way children begin to use their entire, richly structured social environment as a sort of scaffolding within which to build their own language. The work of Kenneth Kaye and others (e.g. Kaye and Charney 1980) on the origins of dialogue in the relations between mothers and infants has begun to build a persuasive and detailed model of just how this works.


The richness and power of this sociolinguistic or pragmatic context has been dramatized in recent years in connection with the controversy over ape language learning. The most devastating attacks on the notion that chimpanzees like Sarah and Washoe had actually learned human language were firmly grounded on the argument that the apes were not really "using language," but were responding to inadvertent extralinguistic "cues" on the part of their trainers. For example, Herbert Terrace (1979, 1981) reported on his own (unsuccessful) attempts to teach a chimpanzee -- named "Nim Chimpsky" in honor, Terrace said, of a certain linguist -- to talk. He also analyzed a number of films of Nim -- and more "successful" chimpanzees -- "talking" with their trainers. He identified example after example of such non-intentional cueing of the chimps. The "Clever Hans phenomenon" had struck again, he said.

It is through this controversy that many people have recently come to hear the name of the celebrated horse who, at the turn of this century, amazed Europe by apparently demonstrating the ability to add and subtract (see Pfungst [1911] 1965; Sebeok and Rosenthal 1981). Der Kluge Hans (Clever Hans) held forth in Berlin, demonstrating over and over that he could "think in a human way." "What's seven and five?" Herr van Osten (his owner and teacher) would ask, and the horse would dutifully tap his foot twelve times. "If the eighth day of the month comes on Tuesday, what is the date for the following Friday?" Eleven taps. "How much is two-fifths plus one-half?" Nine taps, then ten: numerator and denominator. Most people, of course, were skeptical of the horse's skill and assumed some sort of "trick," a secret cueing system, as there is with other trained animals. Adding and simple computation would have been difficult enough to swallow; such elaborate mathematics made it virtually impossible.

But for a time the situation looked as though people would have to choke it down. The problem was that if Hans's performance was a trick, no one could figure out how it was done. Herr van Osten swore that he had taught Hans actually to perform the mathematical operations, and that he was not cueing him. He agreed to have the case investigated in order to establish that, indeed, Hans could do arithmetic. And in fact, none of a panel of experts, including animal trainers and psychologists, could discover any of the systems of cues that animal trainers use to make it seem that circus animals, for instance, can think or understand human language.

What finally happened, of course, was that in a classic and exhaustive investigation whose description still makes compelling reading, the well-known psychologist Oskar Pfungst ([1911] 1965) established what the "system" was. Hans was responding to inadvertent cues so subtle as to be virtually imperceptible -- reading, in other words, the body language of the people around him. Among other things, he was responding to the slight inclination of the head and body as the interlocutor (usually Herr van Osten, but sometimes others could work with Hans) began watching for the tapping to begin, and the slight increase of tension as the questioner waited to see whether Hans would stop at the correct number. Carefully separated from anyone who knew the answer to the question, the horse was unable to perform; but when he was in a situation where his interrogator knew what answer to expect, Hans could sense it, and stopped tapping when he had tapped out the appropriate number (that is, the one his questioner expected). In the end, and after great difficulties, Pfungst was able to teach himself to start and stop the horse's tapping, in the absence of any mathematical questions, by synthesizing these subtle movements.

For a traditional empirical psychologist, of course, Hans's story is a cautionary tale. Robert Rosenthal's long-standing interest in the case, for example, derives from its implications for research methodology (Sebeok and Rosenthal 1981). If the communicative relationship between animals and humans is so rich and subtle -- and there have been many subsequent confirmations of the hypothesis that it is -- how much richer and more subtle is that between human beings likely to be? How much of what we've "learned" about behavior in psychological experimentation is really nothing more than artifacts of experimenters unconsciously signaling their expectations, and subjects unconsciously picking up on those signals? As Rosenthal remarks in his introductory essay in Pfungst's work, "That many experimenters over the years may have fulfilled their experimental prophecies by unintentionally communicating information to their subjects may be a disquieting proposition" (Pfungst [1911] 1965, xxii). For Rosenthal, the moral of the story is this: In setting up your experiment, be cautious; try to find ways to make sure your experimenter is not unconsciously signaling the "correct" answers to your subject, or your experiment will be invalid.

Others have drawn other morals. Roger Brown (1958), for instance, uses the story as evidence that horses -- animals in general -- can't use language, that we should be skeptical of all claims for intelligence in animals. "The case is dramatic because our impression of Hans plummets. He had seemed to be more intelligent than most of us but has turned out to be a quite unremarkable animal. There is not even any evidence here of an ability to recognize verbal commands, let alone understand or reply to them" (1958, 175). And Michael Polanyi cites the case as a cautionary tale of yet a different kind: an instance of observers seeing what they want to see, experimenters making their results come out right (1958, 169-70).

Like any good story, the saga of Clever Hans affords many different readings, depending on who's telling it, to whom, and in what situation. The point of my retelling it here has to do with the nature of that "paralinguistic" communication between the horse and his master. It seems clear, for one thing, that the system had to operate at a preconscious level. Herr van Osten was certainly not conscious of the system; indeed, he himself appears to have believed he had taught the horse to perform the mathematical operations (he never accepted Pfungst's explanation of Hans's behavior, and indeed aborted the study when he finally understood what Pfungst was doing). Equally important, Pfungst reported that effective "cues" were extremely difficult to generate consciously and artificially; on the other hand, it was almost impossible, through conscious effort, to avoid cueing Hans if you knew the answer to the question he'd been asked. The cues, it seems, were a natural, organic, and inescapable part of a communication situation.

Heini Hediger, in an essay called "The Clever Hans Phenomenon from an Animal Psychologist's Point of View" (1981), has pointed out that the Clever Hans phenomenon has never been replicated. No one has deliberately taught a horse to respond in such a sophisticated way to cues so subtle and so natural. Clearly, however, it's not at all unusual for horses -- and many other animals -- to learn to respond to the inadvertent cues of humans in almost equally sophisticated ways in other sorts of situations. Anyone who has ridden or driven horses at all regularly, for instance, is familiar with the kind of rapport that can develop between human and horse (owners of many other kinds of pets testify to this sort of communication as well). What distinguishes the Clever Hans case, I believe, is that Herr van Osten is the only person who has ever actually believed he was teaching a horse to calculate. Everybody else who has engaged in training animals for such public display has known that they were teaching the animals to respond to cues, and thus their cues have been consciously manipulated and synthesized and relatively obvious and crude.

What are the implications of Hans's story told in this context -- and juxtaposed with Shawn's? It brings to the foreground, I think, the fact that a horse could use aspects of his sociolinguistic environment, so subtle that no one knew they existed, in order to learn; even more, it shows as well that he could only do so when the whole complex remained subordinate to other purposes. Herr van Osten was teaching a horse mathematics, not devising cueing systems; Hans was pleasing his master, not attending to the cueing systems which were his means of doing so. What is common to the stories of Hans and Shawn is that in both cases, virtually miraculous achievements of language learning occurred -- and only occurred -- when the semiotic system was in use for real purposes, and was thus "whole" in the sense that it was organically embedded in a real, functional context. The power of the systems of cues to which both Hans and Shawn were responding -- to which we all respond in language-learning situations -- resides precisely in the fact that they are not conscious, that they bypass our consciousness just as the surface features of language do (we can remember only with difficulty the exact words people have used, though we remember the gist and import of what they said with amazing facility).

What Hans's case dramatizes, in other words, is the power of the kinds of unconscious intersubjective cueing that surround the language to which we actually attend. (It is interesting, incidentally, that the growing study of pragmatics in language use seems to show us the same thing; consider, for example, the work of Erving Goffman [1974], William Labov [1972], Livia Polanyi [1979, 1985] and others on the social frames in which conversation takes place, or of Elizabeth Bates [1976b] and others on developmental pragmatics.) One thing we can learn from Clever Hans is how inextricable linguistic phenomena are from this rich context -- and how, when they are extricated, they change utterly. The whole context -- social as well as linguistic, physical as well as grammatical -- is an inextricable part of any given utterance and plays as important a role in its conveying of meaning and impact as, say, its syntactic structure.


The psychology of human development explains some of the reasons human language learning -- and, if there are important differences, other learning as well -- is so profoundly dependent on its social situation and the wealth of subtle cues it provides. Central to this understanding are the ideas of Vygotsky (1962, 1978), whose concept of the "zone of proximal development" seems particularly relevant to thinking about the kind of learning that can occur in such situations. Basically, Vygotsky maintains that our traditional views of learning have been derived by assessing what the learner can do in isolation, by her- or himself, and then assigning the learner to a particular level of cognitive development. Vygotsky argues, however, that what we ought to be concerned with is what the learner can do in a social situation, with the assistance of the people around him or her -- that in fact, such considerations define a "zone of proximal development" where learning (or, more broadly stated, significant cognitive development) is most likely to take place. His view of learning as a result of a dialectical process occurring between the individual and the context (most centrally, the social context) describes, it seems to me, exactly how Hans and Shawn learned: that is, they utilized the whole of the social context in which they were immersed in order to learn. Michael Polanyi points out (1958, 3) that in one way Hans's performance was perfectly adapted to his situation: he achieved what there was for a horse to achieve -- the reward of pleasing Herr van Osten. What he learned was a system of signs. What Shawn achieved was contact with his teacher; what he learned, as well, was a system of signs.

For a teacher, then, Hans's story should be the opposite of a cautionary tale, reminding us as it does of the resourcefulness of the learning organism. Hans's skill at reading a psychosocial situation seems to me at least as miraculous as any skill at addition or calendar manipulation would have been, and what is most remarkable is that Hans may well have been, after all, "a quite unremarkable animal," whose skill might actually be quite widespread. Such a skill is no less remarkable because it's common; quite the contrary. Even more important, it is for a teacher not a danger to be avoided, but a natural phenomenon to be utilized. If a horse can be so sensitive to the social situation around him, how much more sensitive might we expect human beings to be? How much does a human learn by responding to such cues? Shawn, and all his compatriots, offer us an answer.

Far from being concerned that students may be inferring from the people around them when they're right and when they're wrong, teachers should be delighted that children have such an ability, and eager to find ways to help them utilize it more consistently and effectively. We should be excited that such skill can be invoked in the service of -- for instance -- language learning.

As indeed it almost always is, of course. An infant learning to talk is virtually never overtly corrected, or needs to be, or ought to be. What happens is that the fantastically sensitive set of social antennae that each of us has -- far more sensitive than even the most intelligent horse's -- tells us when we've said something a little wrong, something that hasn't quite connected, tells us when there's noise in the system. We have a Clever Hans inside us who notices such things. What our Hans notices may be as crude as a "correct" response to an "incorrect" utterance: for instance, baby says, "Daddy go bye-bye," and mother responds, "That's right; Daddy's going bye-bye." Or Shawn writes of the snake from under the "vradae" in "a cwareame," and his teacher hesitates in reading it, or uses the conventional spelling of one of the words in her response.

Or it might be more subtle: an acquaintance's stiffening and distancing, communicated by body language and other, even less immediate and specific, responses, when you've used an expression that is perhaps a bit too familiar for the situation. Or, indeed, it may be genuine confusion resulting from misunderstanding. In each case, it seems clear that it is the Clever Hans in us that is largely responsible for the breathtaking rapidity with which we assimilate the unspeakably complex rules of our native dialect and the even more unimaginably complicated pragmatic rules which govern our conversational interactions.

What seems to me most important about this metaphorical Clever Hans we each carry around inside us is that it seems clear he doesn't need to -- and perhaps can't afford to -- become self-conscious; this kind of learning works by bypassing the sort of effortful, conscious apprehension and understanding we associate, for instance, with formal grammar instruction. Michael Polanyi's (1958) ideas about knowing are very important here. His theory of knowledge is organized around a distinction between knowledge that we consciously attend to (focal knowledge) and that far larger portion of what we know that we do not and usually cannot attend to (tacit knowledge). Virtually all our competence at using language lies in the realm of tacit knowledge. In his view, this is the sort of knowledge we acquire best in an apprenticeship situation, by "dwelling in" the activity involved. The more conscious many kinds of knowledge become, the less effectively they can be utilized. In Stephen Potter's (1971) memorable slogan designed to help the prospective game player win without actually cheating, "Conscious flow is broken flow."

It is at least interesting that in the introduction to Pfungst's book, Professor C. Stumpf makes it clear that Hans learned his "language" while everyone involved was attending to something else -- to the principles of arithmetic or the mechanics of training the horse to tap. And, as Hediger (1981) observed, the phenomenon has, after all, never been recreated. It may be that in order to invoke your Clever Hans you have to be attending wholemindedly to something else.

This may be why language learning isn't always the easy, natural process that it ought to be if this "Clever Hans model" of learning invariably corresponded to what happens in the real world. When we study language in a conventional, institutional setting -- in situations where we are deliberately learning a second language, for instance, whether a foreign language or a new dialect of our own, such as formal, academic written English -- our internal Clever Hans is rarely given a chance to work, because the whole process is quite deliberately rendered conscious. Not only this; in such situations our "errors" are regularly corrected by an external authority. The more this happens, the less we can depend on -- or even attend to -- our internal Clever Hans. He specializes in noticing signals so subtle that they are overridden and obliterated by corrections from an external authority.

A consequence of this process of externalization of authority is that eventually our means of testing the accuracy and effectiveness of our statements moves outside us. We wait for external, overt, crude, obvious corrections. If we don't get them, we assume we're all right. Our Clever Hans, in other words, has been smothered.

A good example of this is the way students in my writing classes often responded to my written comments on their writing projects. In a typical conference, toward the end of a writing project when the issue had begun to be a little more like editing than like composing, I would typically mark a few characteristic errors, discussing the ways in which they werecharacteristic, and explicitly tell the student to find and correct similar problems elsewhere in the paper. Very often the student came back to the next conference with every error I'd marked dutifully corrected, and nothing else touched. Students knew where the authority was, and were not about to make a change on their own because, unless the authority had pointed the error out, it must be OK. I hate to think that student's Clever Hans was dead, but it was certainly one sick horse.

Why should this happen? It seems a reasonable hypothesis that Clever Hans only works in real situations, where people really do care about the answers to the questions. If no one in the room had been listening to the question, or cared whether Hans could add, Hans would have been as much at a loss as when they didn't know the answer. Just so, the child learning to speak needs -- as everyone who has ever watched a child learn this miraculous and miraculously complex skill knows -- an interlocutor (a "significant other," in Halliday's phrase) who cares about and responds to what he or she says, an audience engaged in making meanings from the child's utterances. If the people around that poor infant were concerned about supplying grammatical information and judging the correctness of the utterances, while caring little or nothing about their meanings, it seems likely that the child would end up in a remedial talking class (perhaps labeled "dysoral"). And perhaps that child would spend life talking, not fluently and joyfully with his or her awareness focused on interaction and communication, but, rather, painfully and self-consciously, with attention focused on avoiding embarrassment, choosing what to say -- the way I tend to choose what to say in German, not according to what needs to be said but according to what I think I can say without making a humiliating faux pas. (It is worth considering the consequences of corrections and their unavoidable shunting aside of meaning. Suppose Shawn's teacher had responded to his first original writing ["Tomy has a new trucK"] by saying, "Very good, Shawn; but you've misspelled Tony's name and the K in truck should be lower case. C." Or suppose Herr van Osten had rewarded Hans's responses to cues rather than his [apparent] understanding of arithmetic?)


All of this adds up, it seems to me, to a set of ideas which help us to understand why so few adolescents find writing easy or comfortable, or are very competent at it. It also suggests some means by which we can create a situation in which students can begin moving toward greater comfort and competence. It is worth considering some of the ways these ideas have found practical implementation in the context of actual elementary classrooms.

Accepting, and wholeheartedly implementing, these ideas in the context of a conventional contemporary classroom entails some dramatic -- one might even argue, revolutionary -- changes. Details vary from teacher to teacher, of course, but some patterns are clear. Learning to read and write begins with what the children already know (Harste, Woodward, and Burke, incidentally, have shown us that children know far more than we used to think), instead of with the assumption that they are in essence blank slates. Trade books and "naturally occurring" published texts replace basal readers and other textoids assembled by committees from wordlists, according to formulas derived from "readability" tests. Reading for meaning replaces reading for demonstration of competence, especially for demonstration of isolated, "targetable" skills like phonics decoding, vocabulary analysis, "word-attack skills," and so forth. Writing develops in tandem with reading, just as speaking develops in tandem with listening, and ways are found to allow both writing and reading to be meaning- and intention-driven. The teacher's job becomes one of finding and creating occasions and situations in which writing and reading can serve real, instrumental purposes for the students. Evaluation of written work virtually vanishes, to be replaced by response to it, and (even more important) by use of it. Writing and reading are conceived of as tools rather than as foci of attention, and metalanguage -- discourse about discourse -- develops for each child, if at all, only as it is needed, rather than being the center of the curriculum. The inherent power of children's language learning capability is freed.

As usually happens in descriptions of educational innovation, it sounds as though I'm hailing the advent of the millennium, and one might well wonder that such a revolution could occur with so little visible fuss. As with poor Lemuel Gulliver trying to figure out why Smithfield wasn't ablaze with pyramids of law books the day after his Travels was published, one might well conclude that my assessment of the situation is a little skewed. And certainly whole language methods, like other new educational ideas, are susceptible of being watered down and compromised until they have no effect; they're open to being implemented mechanically and without understanding, simply because they seem trendy or innovative; they require more flexibility and change than many teachers are prepared to exercise and accept; and they are going to have a difficult time competing for the dollars that are likely to be shaken out of governmental trees by cries of "back to the basics" and blue-ribbon committee calls for educational rearmament. I suspect that, like John Dewey's ideas -- or Lemuel Gulliver's, for that matter -- they're unlikely to be tried on a really wide scale. All this does not, however, constitute evidence that they are anything other than genuinely powerful methods based on substantial and profound ideas, or that they do not have important implications for composition classes at the secondary and post-secondary level.

What is most important and powerful about the ideas I am describing is their ability to help teachers move in their own ways in the new directions they open up rather than to adopt the recipes of experts. Frank Smith (1982) regularly refuses point-blank to make any concrete suggestions for applying his ideas on the development of reading to classroom situations, on the grounds that the central skill of teaching is precisely the application of theory, and that the appropriate situation is one in which every teacher applies a theory in a different and unique way to different situations. Still, because I also believe that it is possible for examples to function heuristically at least as effectively as abstract theories, I want to offer a few examples of the ways in which whole language theory can concretely change some fundamental practices in writing instruction.


I wish to do no more here than sketch one or two examples, and indicate a couple of important directions in which we might move under the impetus of these ideas, particularly with regard to the ways in which we respond -- and do not respond -- to student writing. We should not expect, of course, that we can work miracles of change in a sixteen- or even a thirty-two-week period. What we should be striving toward is opening a door of understanding even a crack, so the student will understand there is a door there and that perhaps light shines through it occasionally. As with so many processes in education, we are often not around long enough to see the student fling the door wide and walk through.

Perhaps the most important change might be characterized as finding ways to make writing more instrumental, attempting, if you will, to make it "real," and thus whole. The most effective way to do this is to create situations in which student writing -- and the teacher's own -- is read for its meaning, for what it has to say, rather than as an example of a student theme to be "assessed" or "evaluated." One way to do this is to employ writing in place of more usual pedagogical methods -- instead of class discussion, for instance. One can create situations in which students write very frequent short pieces, in answer to specific questions or problems, often questions or problems arising out of previous (frequently written) discussions. The short pieces are then photocopied and distributed, sometimes to the whole class, sometimes to selected or relevant individuals. I regularly set up situations in which some sort of written response (to the ideas, not to the form) is appropriate; a written response which is often read only by the writer of the original. What is even more effective, however, is a practical response -- for instance, a situation in which the reader has to use the information or ideas in order to pursue his or her own work. Two things are crucial here: one is that the meaning of the writing serves purposes the student understands and at least to some extent shares; the second is that the response the writing brings is not in any sense evaluative, but is, rather, instrumental.

It is also worth noting that, in general, response to the paper does not come from the instructor. I rarely read, and never evaluate or mark, this sort of writing -- I participate in the process, of course, by writing and circulating my own work, but the main point about all this writing is that it is to be conceived of as a way of learning, not a performance to be judged. Because the language is being used for some real purpose, it can be whole -- and because it is whole, it can function as part of a language learning experience.

Let me give a concrete example of how this has, in fact, worked. In about the third week of my course in eighteenth-century literature, I send the class to the library to read about the last third of the seventeenth century in literary histories, textbooks, anthologies and collections, and political histories, and to bring back questions raised by that reading. The only criteria are that they must think the questions answerable with a reasonable amount of effort and within a reasonable length, and the answers likely to be interesting or important. The questions are weeded by small groups, each of which selects two or three to write on the blackboard. These questions are then edited down to a half dozen or so by the whole class, whereupon small groups choose questions to answer. Questions which have been fruitful in the past have included these:

Each member of the class then goes to the library (working from a collaboratively-compiled annotated bibliography of useful research tools) and writes a short essay in answer to the question. Back in groups (sometimes in class, sometimes not), the individual pieces of writing -- which I seldom see and never comment on or evaluate -- are edited into one short collaboratively-written essay, which is then transcribed by a member of the group and photocopied for the class (again, I do not comment on or evaluate this writing). These photocopied essays become a "chapter" of the course textbook.

The next stage in this cycle involves everyone in the class asking a question (or two or three) of each of the groups responsible for the separate reports. These questions are written out and brought to the next class session, where each group receives their questions, weeds them, answers (orally) those that can be answered or disposed of immediately, and organizes to research and write answers to one or two which seem particularly useful or important. These written answers are photocopied and appended to the original reports. Finally, all of this writing may form a springboard for a new cycle of research, writing, and publication.

Almost any subject about which someone might want to learn, and which it is possible to divide up into separate questions or issues, will afford this kind of activity. Descriptions, summaries, and synopses are particularly useful, especially at the beginning, because their utility is obvious to students engaging in a joint research project. An adequate abstract of an article, for example, will usually tell readers whether they need to look any further, and will often save a good deal of time. What is most important is that the students begin to think of writing as a tool, one which is useful for purposes that seem immediate and palpable to them, and begin to use it.

Responding to Student Writing

The implications such a model holds for current practices in response to student writing are, it should be clear, fairly radical. In the context of this sort of use of writing, traditional evaluative or analytic responses have no obvious role. Indeed, it is not clear whether responses by the course instructor are desirable at all. It's difficult, of course, at first to prevent students from responding to each others' writing the way they've become accustomed to seeing English teachers respond, with approving and condescending generalities, or corrections of grammar, style, or diction. But after a while they begin to figure out that for writing of this kind, such things matter only when they really matter, when they cause genuine confusion or misunderstanding. They also discover that I'm probably not going to look at many of the papers in any case -- and if I do, I will not mark them -- and thus they don't have to impress me with their knowledge of what constitutes "good English." As this happens, sometimes real exchanges of ideas are generated, the linguistic surface can become tacit and the meaning focal, and real language learning can begin. The language, in other words, begins to be whole.

I am aware of no compelling reason to believe that the process of language learning that college students are going through is different in any radical ways from the processes they went through as infants learning a first language, as toddlers learning a second (literacy), or as adolescents learning all the new languages -- slangs, dialects, argots -- needed to survive in a complex society like ours. The evidence is overwhelming that all those languages are learned through use if they are learned at all; that they are learned because the learner understands, or comes to understand, that the new language can serve her or his own real, immediate human and social purposes. When we disengage language from those purposes -- when we make written language into an artifact to be analyzed and evaluated -- we do something much more profound than merely making it something based on a fantasy: we take away what is most basic to language.

When language ceases to be whole, in fact, it's no longer language. It's still a complex, rule-governed behavior, but the language-learning resources which operate in Shawn, and Clever Hans, and in you and me, no longer work on it. They aren't, after all, processes that we can turn on and off at will; they are a central element of our human need to be part of a social group. They constitute a powerful force, but the only wagon you can harness them to is that need. What's wrong with offering students textoids and fragments of language, and asking them to create textoids and fragments, is precisely that, by definition, textoids and fragments cannot serve that need.

I see no reason not to embrace as a fundamental principle that any language used in any classroom should be, in the most thorough possible sense, whole. To keep language whole we need to treat it as whole -- we must, like Herr von Osten, believe our horse is learning mathematics. We must believe our Clever Hanses, our Shawns, are indeed clever. They are -- at any rate, the Shawns are.


Bates, E. 1976a. Language and Context: The Acquisition of Pragmatics.New York: Academic Press.

Bates, E. 1976b. Pragmatics and Sociolinguistics in Child Language. In Normal and Deficient Child Language, edited by E. M. Morehead and A. E. Morehead, 411-63. Baltimore: University Park Press.

Bates, E. 1979. The Emergence of Symbols: Cognition and Communication in Infancy. New York: Academic Press.

Beaugrande, R., de. 1982. The Story of Grammars and the Grammar of Stories. Journal of Pragmatics 6:383-422.

Brown, R. 1958. Words and Things. Glencoe, Ill.: The Free Press.

Brown, R. 1973. A First Language: The Early Stages. Cambridge: Harvard University Press.

Bruner, J. S. 1977. Early Social Interaction and Language Acquisition. In Studies in Mother-lnfant Interactions: Proceedings of the Lock Lomond Symposium, Ross Priory, University of Strathclyde, September 1975, edited by H. R. Schaffer, 271-89. London: Academic Press.

Bruner, J. S. 1982. The Formats of Language Acquisition. American Journal of Semiotics 1:1-16.

Chomsky, N. 1965. Aspects of a Theory of Syntax. Cambridge: MIT Press.

Dewey, J. 1938. Experience and Education. New York: Macmillan.

Goffman, E. 1974. Frame Analysis. New York: Harper and Row.

Goodman, K. S. 1967. Reading: A Psycholinguistic Guessing Game. Journal of the Reading Specialist 4:126-35.

Goodman, K. S. 1983. Language and Literacy: The Collected Writings of Kenneth S. Goodman. Edited by F. V. Gollasch. London: Routledge and Kegan Paul.

Goodman, K. S., and Y. Goodman. 1979. Learning to Read Is Natural. In Theory and Practice of Early Reading, vol. 2, edited by L. B. Resnick and P. A. Weaver. Hillsdale, N.J.: Erlbaum.

Halliday, M. A. K. 1973. Explorations in the Functions of Language.London: Edward Arnold.

Halliday, M. A. K. 1975. Learning How to Mean: Explorations in the Development of Language. New York: Elsevier North-Holland.

Halliday, M. A. K. 1978. Language as Social Semiotic: The Social Interpretation of Language and Meaning. Baltimore: University Park Press.

Harste, J. C., V. A. Woodward, and C. L. Burke. 1984. Language Stories and Literacy Lessons. Portsmouth, N.H.: Heinemann.

Hediger, H. K. P. 1981. The Clever Hans Phenomenon from an Animal Psychologist's Point of View. In The Clever Hans Phenomenon: Communication with Horses, Whales, Apes, and People, edited by T. A. Sebeok and R. Rosenthal, 1-17. Annals of the New York Academy of Sciences, vol. 364. New York: The New York Academy of Sciences.

Kaye, K. 1982. The Mental and Social Life of Babies: How Parents Make Persons. Chicago: University of Chicago Press.

Kaye, K., and R. Charney. 1980. How Mothers Maintain "Dialogue" with Two-Year-Olds. In The Social Foundations of Language and Thought: Essays in Honor of Jerome S. Bruner, edited by D. R. Olson, 211-30. New York: Norton.

Labov, W. 1972. Language in the Inner City. Philadelphia: University of Pennsylvania Press.

Newman, J. M. 1984. The Craft of Children's Writing. New York: Scholastic Book Services.

Newman, J. M., editor. 1985. Whole Language: Theory and Practice.Portsmouth, N.H.: Heinemann.

Peirce, C. S. 1955. Philosophical Writings of Peirce. Edited by J. Buchler. New York: Dover.

Pfungst, O. [1911] 1965. Clever Hans (The Horse of Mr. von Osten): A Contribution to Experimental Animal and Human Psychology, with an Introduction by Prof. C. Stumpf. Translated by C. L. Rahn. Reissued with an introductory essay by Robert Rosenthal. New York: Holt, Rinehart, and Winston.

Polanyi, L. 1979. So What's the Point? Semiotica 25:207-41.

Polanyi, L. 1985. Telling the American Story: A Structural and Cultural Analysis of Conversational Storytelling. Norwood, N.J.: Ablex.

Polanyi, M. 1958. Personal Knowledge: Toward a Post-Critical Philosophy.Chicago: University of Chicago Press.

Potter, S. S. 1971. The Complete Upmanship. New York: Holt, Rinehart, and Winston.

Sebeok, T. A., and R. Rosenthal, editors. 1981. The Clever Hans Phenomenon: Communication with Hones, Whales, Apes, and People. Annals of the New York Academy of Sciences, vol. 364. New York: The New York Academy of Sciences.

Shaughnessy, M. P. 1977. Errors and Expectations: A Guide for the Teacher of Basic Writing. New York: Oxford University Press.

Smith, F. 1982. Understanding Reading: A Psycholinguistic Analysis of Reading and Learning to Read. 3d ed. New York: Halt, Rinehart, and Winston.

Terrace, H. S. 1979. How Nim Chimpsky Changed my Mind. Psychology Today 13:65-76.

Terrace, H. S. 1981. A Report to a Committee, 1980. In The Clever Hans Phenomenon: Communication with Horses, Whales, Apes, and People, edited by T. A. Sebeok and R. Rosenthal, 94-114. Annals of the New York Academy of Sciences, vol. 364. New York: The New York Academy of Sciences.

Vygotsky, L. S. [1930] 1978. Mind in Society: The Development of Higher Psychological Processes. Edited by M. Cole, V. John-Steiner, S. Scribner, and E. Souberman. Cambridge: Harvard University Press.

Vygotsky, L. S. 1962. Thought and Language. Edited and translated by E. Hanfmann and G. Vakar. Cambridge: MIT Press.