The Parts, Kinds, and Purposes of Grammars

     People study languages for four main reasons. When they are babies, they study a language to learn to talk. We do not ordinarily think of this as "studying," but more learning occurs at this stage than at any other. Beyond this stage, many people study a language either to become -- or to appear to be -- educated. Language always has been and probably always will be a social yardstick. People really are offered or refused jobs, accepted into high society, or, at times, left in the gutter, all on the basis of how they speak or write. Because of this, some people have found it possible to study a language to make a living, whether it be by teaching English to an Englishman, by introducing a Frenchman to Chinese, or by writing the texts that others can use in their teaching. Finally, a few people study languages simply because they find them interesting.

     What do people who study a language do? We usually think of them as studying vocabulary and grammar, but a closer look reveals that there is little agreement about the definition of "grammar." Most people think of grammar as the rules of a language, but what are "rules" and what is a "language"? Very simply, a "language" is a system of symbols, either oral or written, that is used to convey information. We will return to the question of "rules" later, but notice that we already have the potential for at least two different sets of rules -- rules related to the sound system, and rules related to the written system. Actually, there are several different systems that work together, and therefore overlap, in every language. There is, moreover, no absolute agreement on how these systems should be distinguished from each other: as people try to answer different questions about language, they define terms differently. Although here we are primarily concerned with syntax, a brief glance at some of the other systems may clarify how the rules in different systems have different kinds of validity.

     Phonetics, the study of the sound system of speech, concerns how sounds are produced and distinguished from each other. Henry Higgins was interested in phonetics. Many of the rules of phonetics concern voiced and unvoiced consonants, stops, fricatives, nasals, liquids, and glides. I mention these concepts to indicate that phonetics is a complicated science with rules of its own. Phonetics is not concerned with the meaning of sounds; phonemics is the science of sounds that distinguish meanings. The Japanese, for example, have trouble distinguishing "lap" from "rap" because in their language "l" and "r" do not convey a difference in meaning.

     Morphology is the study of word-formation. A morpheme is the smallest part of a word that has meaning: in English "-s" and "-ed" are morphemes since they change the meanings of words. The study of roots, prefixes and suffixes is part of morphology. Phonetics, phonemics, and morphology are primarily empirical sciences, comparable to geology or biology: linguists collect examples from actual writing and speech and form rules by generalizing about the examples.

     Semantics, the study of meaning itself, is more philosophical. Since any language is a system of symbols, the semanticists explore the relationships between the symbols and the things symbolized. Is there, for example, an ideal "chair," perhaps in the mind of God, to which all our individual uses of the word refer? Or is every individual’s concept of "chair" simply derived from repeated associations between certain objects and the word? In the first case, we would all mean exactly the same thing when we used a word; in the second, we would mean something at least slightly different. Semantic rules are not inductive generalizations; they are systematic hypotheses, similar to philosophical theories.

     Syntax, the system with which we are primarily concerned, concerns the way words are related to each other in sentences. English syntax allows us to distinguish "The dog bit the boy" from "The boy bit the dog." The rules of syntax are simple and relatively logical. In many ways they are comparable to the rules of a moderately difficult branch of mathematics: more difficult than the rules for multiplication and division, they are simpler than the rules of algebra. They can be broken only at the risk of being misunderstood. No one can say "The dog bit the boy" and be understood as meaning "The boy bit the dog."

     The rules of usage, on the other hand, are more a matter of social convention. They are the rules of etiquette. No one would misunderstand either "He ain’t done nothing" or "He has not done anything." Rules about contractions, double negatives, and the use of ain’t are social, rather than logical conventions. Usage, therefore, involves the study of how different groups of people use words, syntax and sounds. People who wish to appear to be educated must learn and use the conventions of the educated. 

     Usage, syntax, semantics, morphology, phonemics, phonetics -- the list could be increased --  are all parts of what is generally termed "grammar." Since they have different interests, some linguists use the word "grammar" to refer to usage; others, to syntax; still others, to morphology. When the average person on the street speaks of "grammar," he probably has in mind a combination of usage, syntax, and morphology. This confusion about "grammar" results in confusion about the "rules." Most of the rules of phonetics, for example, are simply descriptions. It is impossible to break them. No one can produce a voiced consonant without vibrating vocal cords, because, by definition, a voiced consonant describes sounds made when the vocal cords vibrate. If they do not vibrate, the consonant is unvoiced. The rules of syntax and usage, on the other hand, can be broken, but even here the results are different. Breaking the rules of syntax risks saying something other than what was meant; breaking the rules of usage usually results in being understood, but being considered uneducated. The history of the writing of grammars indicates that many grammarians and teachers have not been aware of these differences in the rules.

     Their lack of awareness is caused, in large part, by a failure to distinguish the kinds and purposes of grammars. What we have looked at thus far are parts of what is generally referred to as "grammar." Grammars can also be of two kinds. Synchronic grammars describe a language at one period in time, i.e, English in the twentieth century. Diachronic, or historical grammars, on the other hand, develop rules of generalizations about how languages change in time. Currently only pronouns and possessive nouns have case endings: he, his, him, Bob, Bob’s. In Old English most nouns had such endings. Similarly, whereas we currently distinguish singular and plural, Old English had singulars, duals, and plurals. (Thus one could tell from the ending of the word whether someone waved with one hand or two.) Changes in languages follow comprehensible, but at times complicated, patterns, and it is these patterns that the historical grammarians attempt to discover and describe.

     The various purposes for which grammars are written fall into three main groups. Descriptive grammars explain how a language is actually used. Prescriptive grammars try to tell people how to use it. Theoretical grammars attempt to answer questions about the study of language. Suppose, for example, that we know the grammar of language X in the eighth century and in the twelfth century. Since the grammar has changed, a historical linguist might wish to develop a theoretical grammar of language X for the tenth century. Many of the most interesting theoretical grammars attempt to answer questions about how the mind uses language: what does the mind do as it arranges words into sentences?

     These distinctions among parts, kinds and purposes of grammars are important background for a historical review. Before the nineteenth century there were no theoretical or historical grammars -- all the early grammars are synchronic and either descriptive or prescriptive. Nor did the early grammars develop the distinctions among the parts of grammar -- these distinctions mainly arose from the work on theoretical grammars. Since the early grammars did not distinguish the parts of grammar, they did not distinguish the validity of the rules -- a rule was a rule, and so it continues to be in most of our English classrooms. The descriptive/prescriptive distinction likewise was not understood before the nineteenth century. Grammars in the schools had always been -- and in many schools still are -- entirely prescriptive even though, as I hope to show in the following survey, many of the rules are inappropriate for English.
 

From Ancient Grammars to the Eighteenth Century

     Like so much in our culture, the history of English grammars can be traced to the Greeks. Greek, like Latin, Russian, and German, is an inflected language. In inflected languages, the function of words -- the relationship between the word and the other words in the statement -- is usually indicated by word endings. English, a relatively non-inflected language, has a few inflections, such as the "-ed" for past tense, the "-s" to mark third person verbs, and the pronoun case endings such as "he," "his," and "him." In inflected languages, however, almost every noun, pronoun, adjective and verb may have ten or more endings. Because it is an inflected language, Greek grammarians focused on morphology -- the study of word forms. They developed rules of declensions for nouns and of conjugations for verbs in which words were arranged in tables according to their form and function. Greek grammarians had two main purposes. Some were intellectually curious and simply wanted to understand their language better. Others wanted to make their living by teaching. Since they attempted to describe the language of the prestigious and educated, Greek grammars were synchronic and primarily descriptive. The Romans, because of the similarities of Latin and Greek, were able to adopt all the major concepts and assumptions of the Greek grammarians. Thus Latin grammars focused on morphology and the individual word. 

     Two problems, problems that are still with us, arose when the first English grammarians adopted the Latin and Greek models for English in the eighteenth century. The first is that since English is a non-inflected language, the Graeco-Roman models did not fit. The Graeco-Roman focus on the word is inappropriate for a language like English, in which the sentence should be considered as the primary unit. The ancient emphasis on parts of speech, on declensions and on case -- the emphasis on categorizing words outside the context of a sentence -- gives a distorted, useless picture of English grammar. A simple example of this is the word "like," which can be an adjective, preposition, conjunction, noun, or verb, depending on the context of the sentence. In inflected languages, a single form such as "like" cannot normally have such a wide variety of possible functions -- suffixes would differentiate the noun, verb, adjective, and preposition and/or conjunction. Unfortunately, this lack of fit has still to be fully understood -- much less has it been applied to our teaching. Across the country, millions of third and fourth graders are making lists of nouns, adjectives, and verbs.

     Not only did the inappropriate fit lead to an overemphasis of grammatical concepts important to Latin and Greek but secondary in English, it also resulted in inadequate development of concepts important to English syntax. It is noteworthy, for example, that English has no generally accepted names for even the basic sentence patterns. As Francis Christensen noted:

I believe that one can do very little to produce mature and interesting and natural sentences within the traditional grammatical categories of simple, complex and compound or the rhetorical categories of loose, balanced, and periodic, and yet I have never seen a textbook that went much beyond these types. (18)

     Students are frequently expected to learn the differences among transitive, intransitive and linking verbs. These are categories of words, based on the Graeco-Roman models. Students study definitions and then attempt to memorize lists, especially of the linking verbs. Since there are numerous exceptions to the lists, in actual analysis, the only way to distinguish the verbs in English is to see how they function within a given sentence: the sentence pattern tells us whether a given verb is transitive, intransitive, or linking. English grammar would better reflect English structure (and would be much easier for students to learn and apply) if, instead of teaching transitive, intransitive, and linking verbs, we switched to a focus on transitive, intransitive and linking sentence patterns. Traditional grammar has never been suitably adapted to English structure -- it is no wonder, therefore, that students find it frustrating, and that many people have claimed that studying grammar does not improve students’ writing.

     In addition to the poor fit, the second problem that arose when the first English grammarians adopted ancient models was that respect for Latin led to a prescriptive, rather than descriptive emphasis. In fairness, it should be noted that the early English grammarians were not blundering fools. Latin had long been the dominant language of Europe and was still learned by almost all the educated. It deserved respect. European philosophy, moreover, was still dominated by what Arthur O. Lovejoy calls the belief in the "great chain of being," the belief that the universe is essentially logical and shares a common pattern. The early English grammarians, therefore, probably never considered the possibility that English and Latin could be radically different.

     Finally, English was still in a state of transition and confusion: there were wide differences in spelling and usage. Naturally, the grammarians attempted to clear up the problems in usage and paid little attention to the fact that noun declensions of Latin are largely irrelevant to English. The prescriptive emphasis on usage was further supported by the social mobility of the English. As people became prosperous, many wanted to speak like the Lords in Parliament: many grammarians gave them the prescription. The distinction between syntax and usage not having been made, no one recognized the inadequacy of Latin grammar for English syntax. To this day, the fundamental problems in the teaching of English grammar can be traced to the adoption of Graeco-Roman models: the Graeco-Roman focus on individual words has become a focus on individual constructions, but little has been done to teach students how constructions fit together in good sentences.

From the Eighteenth Century to Today

     The problems have persisted so long because of other events in the eighteenth century, which marks the turn from a Classical world-view to the Romantic. Whereas Classicism emphasizes the eternal, non-changing, ideal norm, romanticism explores the historical and the individual. In language studies, this meant a shift in focus from the supposedly universal and ideal Latin and Greek, to the individual national languages. This shift resulted in the new science of linguistics and a separation between pedagogical and linguistic grammars. One of the events leading to the development of linguistics, for example, was the discovery that Sanskrit is historically related to Latin, Greek, and German.

     This discovery, among others, initiated two linguistic sciences: comparative linguistics and historical linguistics. Whereas the Romans and Greeks had concentrated on describing one language as it was currently used, the new sciences of linguistics explored the relationships between languages, changes in languages, and the philosophy and psychology of language. In essence, linguistics pulled many good minds away from the pedagogical problems of English grammar. Since many attempts were made to apply the discoveries of linguistics to grammar in the schools, we should look briefly at the history of both.

The Linguistic Path

     Linguistics in the nineteenth century was primarily devoted to developing the theories of comparative and historical grammar. Great advances were made in such fields as phonetics, and grammarians traced the development of individual languages from ancient ones: French, Italian, and Spanish from Latin; Russian, Polish, and Slovak from proto-Slavic; German, Swedish, and Dutch from Teutonic. Tracing the relationships in words, sounds and structure among all these languages, plus discovering the rules of their historical development, took years of intensive labor, but little, if any, of this new knowledge affected the teaching of native languages in the schools.

     The early twentieth century witnessed the development of structural linguistics. Having studied the major languages of the world, linguists became interested in the other languages, many of which had no written form. Since these languages were quickly disappearing, the linguists wished to record and preserve them. The structural grammars developed for this purpose were an amazingly successful accomplishment. They work in such a way that one can write a grammar for a language that one does not understand! Having phonetically recorded hundreds of conversations in the target language, linguists would then examine the recordings for recurrent patterns. Had someone done this with English, for example, he would have started by noticing the frequent occurrence of "the," "a," "is," "of," "in," and words that end in "-ed." These words would help the linguist learn about others. "The" and "a," for example, are usually followed by nouns and pronouns. Gradually the linguist could uncover the basic patterns of the language. The patterns formed the grammar. Obviously, each grammar required months, if not years, of work, but structural linguists achieved their goal: numerous languages were recorded and provided with grammars.
     Although attempts are still being made to use structural grammars for other purposes, it soon became apparent that they could not answer all the linguists’ questions. Still, the success of the structuralists demonstrated that a language does not have a grammar, but rather grammars. Thus several structural grammars were written to describe English.

     Since this discovery, linguists have been exploring a variety of grammatical theories, the most important of which is "transformational/generative." The beginning of transformational grammar is usually associated with the publication of Noam Chomsky’s Syntactic Structures in 1957. The purpose of transformational grammar differs radically from all earlier grammars. Previous grammars, from Panini’s on Sanskrit to those of the structuralists, had attempted to describe languages; Chomsky, on the other hand, attempted to explore how the mind functions as it uses language: how does the mind "generate" a sentence? Transformational rules are very complex, and transformationalists repeatedly note that they make no claim to describing what the mind actually does. Their grammars are analogies, or models of the logical operations that the mind might perform. Transformational grammar is a very young science, as is linguistics itself, but it has already contributed to such things as talking computers. The rapid advances in linguistics naturally attracted most of the linguistically oriented into further research. The young scholars in college naturally became caught up in the enthusiasm of their professors, and both the formation of new departments of linguistics and the tremendous increase in college enrollments resulted in jobs for the new scholars as college professors of linguistics. Interested in their research, few of these people considered the applications of linguistics for pedagogical grammars.

The Pedagogical Path

     In the schools, the inadequacy of the early English grammars was not felt for a long time for several reasons. First, English does share many constructions with Latin and Greek. Second, students, who learn more out of school than they do in, absorbed the essentials of English grammar from hearing it, speaking it, and reading it. Third, education was largely limited to the children of the educated, i.e., children who came from homes where they learned "proper" English. As education was expanded to include more of the population, the problems began to surface.

     Attempts to remedy the problems flow in four trends. Historically the first, and probably still the dominant, is the prescriptive focus on "grammar" without a distinction between usage and syntax. In the late nineteenth century, a new trend --sentence diagramming -- came close to solving the major problems. The spectacular advances in linguistics naturally led to the third trend -- attempts to apply the discoveries of linguistics to the teaching of grammar in the schools. Frustration has resulted in a fourth trend -- the abandoning of any attempt to teach grammar.

     The prescriptive focus of the eighteenth century has remained with us, although it would now be more properly called "proscriptive," since it often emphasizes "don’t" more than "do." Whereas linguists quickly abandoned all thoughts of prescriptivism, many school teachers have clung to it. One reason for their tenacity is that the essence of "usage" is the belief that one way of saying something is better than another. Prescriptive approaches are usually based on either or both of two procedures -- correcting errors and memorizing definitions. In the former, teachers focus on typical errors and students fill in blanks with "who" or "whom," "lie" or "lay." Gleason notes that this emphasis was often labeled "functional grammar." (14) This label again indicates the problems that English teachers have with terminology: "functional" in the sense of "correcting errors" is quite different from "function" in the sense used throughout this book, a sense based on Jespersen’s concept of "function" -- the function of words and constructions in relation to each other. 
     The failure of the "correction of errors" approach is easily understood: students are asked to do exercises to correct errors that they may or may not make themselves. The exercises are somebody else’s sentences, and, usually, these sentences are syntactically very simple. It should be no great wonder not only that most students are bored by the exercises, but also that they are frustrated. Their own writing is often much more complicated than that in the exercises, and their problem is not in distinguishing whether it should be "[Who, whom] did they see?" but in figuring out whether [who, whom] is the object of "about" or the subject of "was supposed" in "The children argued about who was supposed to wash the dishes."

     The other prescriptive procedure involves memorization of definitions and exceptions. Although I have already suggested that many of these definitions are largely irrelevant to English, it should also be noted that many others are inconsistent and/or poorly formulated. Many a traditional teacher defines an independent clause as a "clause that can stand alone," much to the confusion of students who then consider "Mother knows best" as the main clause in "Everyone knows that Mother knows best." Such poor formulations survive because the definitions are treated as ends in themselves -- things to be memorized and then ignored. Anyone who attempts to use this definition, to apply it to an analysis of a paragraph, soon finds it to be inadequate. Its survival is simply another indication that traditional grammar has not been thoroughly or systematically applied to syntactic analysis in our schools. 

     Prescriptive approaches have failed because so much time and energy is devoted to terminology and exceptions that teachers, as well as students, lose track of the primary patterns of the language. In the last quarter of the nineteenth century, Alonzo Reed and Brainerd Kellogg introduced a technique of sentence-diagramming which came closer than anything before or since to solving students’ problems. Very simply, sentence diagramming creates a visual pattern for every sentence:
 

From Alton Morris, et al. College English: the first year. 4th edition. NY: Harcourt, Brace & World, Inc. 1964. 774.

The technique was successful because it addresses syntax -- the interrelationships of words in a sentence, rather than usage or word categories. Millions of people learned basic English sentence structure through diagramming.

     If diagramming was so successful, why did it ultimately fail? The primary reason is that diagramming entails learning an additional set of rules -- the rules for which line goes where on the diagram. Simple sentences are easy to diagram, but more complicated sentences require complex diagrams with complex rules - as well as large sheets of paper. (Note that in our seventeen-word example, the authors omitted the articles from the diagram in order to simplify it.) Another factor contributing to the failure was that many teachers became so interested in the diagrams that they forgot to explain to students what their purpose is. Like the prescriptive definitions, diagramming became an end in itself. This is most evident in the fact that rarely were students asked to diagram any sentences other than those in the grammar textbook. Like so much else in our educational system, the students’ objective in diagramming was to discover the "right" answer, the one in the teacher’s answer book. As a result, students began to perceive diagramming as busywork, didn’t relate it to either their reading or writing, and continued to have problems. Many teachers continue to use diagramming, and there are even a few who invite students to diagram their own sentences. Most teachers, however, were insufficiently trained in grammar to have the confidence necessary for such freedom of exploration: students diagramming their own sentences come up with all sorts of questions that the teacher cannot answer. Since Education departments focus more on method than on the subject matter to be taught, many, if not most, teachers are poorly trained in their subject matter and, intimidated, are also afraid to admit their lack of knowledge, especially to students. It is much safer, therefore, to stick to the sentences in the grammar book, the sentences for which the answers are supplied. Diagramming could be much more effective than it is, if it were used as a tool for the students’ exploration of sentence structure, rather than as a dead end in itself.

     Although it is comparatively recent, the third major trend -- attempts to apply the discoveries of linguistics --has not proven very promising. Its failures have resulted from applying grammars to purposes for which they were not intended. Several structural grammars were written for the schools, one of the best of which is Paul Roberts’ Understanding English. Structural grammars failed primarily because they divorce grammar and meaning. Students, if they want to learn grammar at all, usually want to learn it so that they can say or write what they mean more clearly and effectively. Like their traditional counterparts, however, structural grammars do not address the practical applications of the concepts they present. The concepts are different, and students are invited to see some of the theoretical problems of linguistics, but the texts still present "grammar" as an end in itself, not as a tool for the students’ use in exploring language. Roberts, for example, spends half a page discussing whether "moving" in "moving van" is a regular adjective or a verbal, a question of no practical importance to students and probably of no interest either. In many ways the structuralist textbooks are comparable to the new graduate assistants in college who bring into their Freshman seminars all the concepts they are learning in their own graduate courses -- the ideas are more important to and appropriate for the teacher than the students.

     Although it may not have been the direct cause, structural linguistics, with its emphasis on sentence patterns, surely increased the popularity of textbooks that present sentence patterns. Perhaps the best example of such a text is The Art of Styling Sentences by Marie L. Waddell, Robert M. Esch, and Roberta R. Walker. The trouble with these texts is that they present one construction at a time with examples that are too simple. The typical pattern book explains subordinate clauses, for example, and then gives several sentences with one, perhaps two, subordinate clauses. Students rarely have trouble with these sentences in the first place. Their problems arise when they try to combine two or three subordinate clauses plus a gerund or gerundive, and, perhaps, an infinitive, all in one sentence. Pattern books cannot possibly present all possible combinations of syntactic structures, and although they usually make a valiant attempt to suggest to students that the patterns presented can be combined in infinite ways, students rarely go beyond seeing the patterns as molds into which sentences are to be fit rather than as concepts for the exploration of sentence structure. 

     Attempts to import transformational grammar to the classroom involve the same problems as the attempted importing of structural grammars. An example is Jean Malmstrom’s Grammar Basics: A Reading/Writing Approach. These attempts have likewise failed simply because the primary purpose of transformational grammar is to figure out how a five-year-old knows what he already knows, and not to teach him something new. Transformational grammars, for example, develop complex rules and procedures for explaining the relationship between "John closed the door" and "The door was closed by John." The rules, moreover, change, depending on which transformational grammar one reads. Another approach, adopted by some teachers, is to abandon all pretense of teaching grammar and to teach a general course in linguistics instead. Other teachers have opted for sentence-combining.

      Sentence-combining is based on the transformational concepts of kernel sentences and embedding. Transformational theory has shown that most sentences are composed of several simple sentences embedded in one main one. Thus, "That is an old red house" is a combination of "That is a house," "The house is old" and "The house is red." In an attempt to get students to write longer, more complicated sentences, researchers and teachers developed the sentence-combining approach. Students are given exercises in which they are to combine two or more simple sentences into one:

Most of the exercises are more complicated than my example and include five or six sentences to be combined into one. When it was first developed, instruction included some explanation of grammatical terms. Since the publication of Frank O’Hare’s Sentence Combining: Improving Student Writing without Formal Grammar Instruction in 1973, many sentence-combining texts have been published that include exercises and avoid explanations.

     While it is true that sentence combining has been slightly successful in increasing the complexity and length of students’ sentences, there is no indication that it provides students with an understanding of what they are doing or why. A major objection to the approach, therefore, is that it treats students as if they were Skinner’s rats -- just as a rat in a Skinner box is trained to push a lever, students are trained to combine sentences. The premise of the approach (that longer is better) is also questionable. It is based on the work of Kellogg Hunt, who demonstrated that the length of the average "T-unit" (i.e., main clause) is a good measure of syntactic maturity. Most of Hunt’s research, however, involved students in grades three through seven, and Hunt himself suggested that the same yardstick might not be appropriate to students in high school and college, especially since instructors at the latter levels often have to encourage students to simplify, rather than complicate, their sentence structure.

     Once again terminology has troubled the profession. When O’Hare uses the word "improving," he means increasing the average length of T-units. On this scale, James, Faulkner, and Joyce are masters; Hemmingway needs a lot of remedial work. It is a very mechanical scale, a scale with which many of us in the profession would not agree. But the profession, for better or worse, is very broad and complex --instructors teach both writing and literature. Our interests vary. Some of us are mainly interested in literature, and even among those of us who are primarily teachers of writing, grammar is only one of a number of the things we teach. We try to cover all the bases, but some of us are interested in invention, others in organization, others in cohesion, etc. Thus, when we hear that an approach "improves writing," we may gravitate toward it without fully considering that one person’s "improves" may not be ours.

     Out of plain frustration, many teachers have simply stopped teaching grammar. Because they are much more free to choose what they will and will not teach, and because they often feel that students should have learned how to write decent sentences in high school, many college professors openly admit this. Few school boards would allow elementary or high school teachers to stop, but the confusion over approaches and effectiveness has surely resulted in some rather mechanical teaching. Research claiming to prove that knowledge of grammar does not improve students’ writing has furthered this trend, a trend officially supported by NCTE’s resolution against the teaching of grammar. 

     If the situation seems chaotic, it is; hopeless, however, it is not. For the past several years, articles in the professional journals have noted and bemoaned the separation between the "towers" and the "trenches." Researchers and most of the contributors to the journals have argued against the teaching of traditional--or any other--grammar, but the teachers in the trenches have refused to listen. One such teacher, Frederica Davis, attempted to improve her teaching by taking courses in linguistics and by studying the research. (See Bibliography.) Still, she felt that her students should be taught grammar. In her research, she found some people, such as Martha Kolln, who strongly argued for grammar, but looking for additional support, she wrote to Noam Chomsky, the founder of transformational/generative grammar, to ask his opinion. In his response to her, a letter printed after her article, Chomsky responded that he is not an expert in pedagogical grammar, but that he thought students should probably learn a traditional grammar like that of Otto Jespersen. Chomsky’s response is remarkable in that he did not suggest that students should study transformational/generative grammar. All the new pedagogical grammars -- structural and transformational -- have been attempts to apply the newest ideas in linguistics to the students’ problems in the classroom. They are all, more or less, comparable to the graduate assistant’s thrill of discovery being carried into his Freshman writing course. Chomsky, on the other hand, views the question from the point of view of the students’ probable needs, and, although claiming not to be an expert, suggests Jespersen’s traditional grammar.

     To say that I was pleased by Chomsky’s reply would be an understatement. For several years I had been developing a new approach to teaching traditional syntax. The impetus for that approach had been the desire to combine what I considered to be the simplicity of Paul Roberts’ Understanding Grammar with the applicability of Reed and Brainerd’s sentence diagramming. Having studied various structural and transformational grammars, I was slowly discovering ways in which Roberts’ traditional concepts could be even further simplified. I read Chomsky’s reply, just when I was also reading Jespersen’s The Philosophy of Grammar and discovering that Jespersen had already developed some theoretical justifications for many of "my innovations." Chapter Two explores some aspects of Jespersen’s theories as well as further modifications in traditional syntax that are suggested by structural and transformational concepts. Before we give up hope and abandon the teaching of grammar, perhaps we should see if traditional Graeco-Roman syntax can be usefully modified to fit English syntax.

1. What textbook (or textbook series) is being used by the school system to teach grammar?

 a) Is it prescriptive or descriptive?  Traditional, structural, transformational,
 or a hybrid?
 b) Why was this book (or series) chosen?

2. What is the difference between syntax and usage? Do teachers make that distinction clear to the students, or do they claim that a rule rule is a rule? In teaching usage (such as "Don’t use ‘ain’t’") do teachers make it clear that this is a rule of social etiquette, or do they pretend that the rule is universally applicable (thereby, perhaps, scorning the language that the child may hear at home)?
3. When teachers have students do grammatical exercises (such as underlining subjects and verbs, or identifying clauses) do they supplement the textbooks with the students’ own writing? If not, why not?

4. A large part of this chapter was devoted to the parts, kinds, and purposes of grammars. My intention is to suggest that "teaching grammar" is not an "either/or" proposition. There are many choices to be made, and many alternatives. But I have also suggested that this may be news to many teachers, and the teachers themselves are not to blame for it. (They cannot teach what they have not been taught.) What preparation do the teachers have? What kind of preparation would they have liked to have had? What can be done about it?