This general text for students in linguistics, and perhaps English majors, is ingeniously designed. After a couple of chapters on the nature of language and modern English phonetics, it begins a broad chronological account of English through the centuries, stopping at convenient stations—Old English in the time of Alfred, Middle English 1350, English of the King James Bible, and so on up to recent Americanisms. The ingenious device is that, at each station along the way, the authors treat the particular subject with a different method, chosen partly for its peculiar relevance but mainly to illustrate for the student the range of philological thinking. The place of English in the Indo-European languages is, necessarily, handled by the comparative philology that climaxed in the first half of the nineteenth century. After some necessary political history, Old English is treated morphologically, reasonably enough because of its inflexions, and this gives opportunity to try out the structural linguistics of the first decades of the twentieth century. Middle English is the occasion, not unreasonably, for illustrating linguistic maps. Early Modern English is picked on, rather arbitrarily and not too happily, to expound and apply full-blown structural linguistics, the grammar of kernel sentences and transformation—rules that are associated with the work of Noam Chomsky. The eighteenth century and Victorian times are used, inevitably but boringly, for a sensible but uninspired critique of “correct usage” and the theory of dictionaries. And the last chapter is a repository for vocabulary counts, etymologies, borrowings and coinings, slang, and other events that occur as relics in the contemporary lexicon.

It is a useful framework. It allows for the inclusion of a lot of intellectual and scientific thinking and general history, plenty for teacher and students to talk about. It avoids premature specialism and repeating material, yet it is not confusing and it is all about English. And the whole is fairly inclusive. The only surprising total omission—especially surprising since the authors are of the school of Sapir—is any discussion of English as world language, Basic English, etc.

The other weaknesses and omissions are endemic in the “linguistic” school itself, whether in its anthropological or logical-positivist tradition, and they will be the main subject of this review. It is the procedure of this school to abstract so rigorously to its one subject of “Language,” leaving out all nonsense about anything else, that finally one judges that its principles do not spring from anything, they are largely fiat and calculus. In this book, the “fundamental” functions of language, communication, and expression, are ritually touched in Chapter I and heard of no more; what they can possibly mean is not discussed, and nothing whatever is made to follow from their fundamentality. “Meaning” is what lexical units (“chair,” “swim,” or “yellow”) “have in a more usual sense,” but grammatical units (tense, pause, or dependent clause) “have only in a vague, subjective, and circular sense”; but the student is not notified that there are problems of semantics, and have been theorists, whether epistemologists, or social-psychologists, or linguistic philosophers. There are a few feeble remarks on style, but “the ordering of sentences that holds discourses together has not yet been satisfactorily described” (p.85); those who, from Aristotle to Auerbach, have labored in the structural analysis of poetics and rhetoric have labored in vain.1 Probably most serious of all, since it is assumed that “the proper subject of linguistics is not speech but language, though we may investigate a language only through its individual speech acts,” we hear very little of what it is to speak, though that interesting behavior must have some relation to language beside being its carrier; with a flourish, “Language Events” are boldly called “LE” and included in the calculus, but they prove to have no intrinsic properties or consequences; nothing in grammar, intonation, or lexicon is apparently affected by the act of speech, the circumstances of the speech, or the character of the speaker. And there is the usual listing of the phonemes—groups of similar sounds—of English, but no statistical count of the actual noises that English-speakers make. (Naturally there is a lengthy section on the physiology of speech, which has the same relation to the rest that the usual chapter on Physical Anthropology has to the theory of Culture, namely none.)

Being an ignorant man, I cannot judge much of the scholarship; where I know something, it is good enough. And the writing is better than in most text-books; it does not talk down. (My guess is that the more popular-sprightly and wooden passages come from the more anthropological of the authors; anthropological linguists are so involved in notating and de-coding hitherto inaccessible languages and translating them into our own code, that they tend to forget that it is as one of the humanities that language fulfills its functions.) It is too bad that this book on English does not breathe more love for English and demonstrate more finesse in using it. Let me single out one, unusually poor, passage:


If a distinction such as that held until recently between disinterested and uninterested breaks down, no irreparable harm has been done. If speakers need the distinction here, they will maintain it by using other words, such as impartial or indifferent [p. 355].

But do these words mean disinterested? How, with them, would we describe esthetic judgment, which as Kantians know is disinterested? Or how could we refer to the disinterested advice of a friend or an amicus curiae, who is certainly not indifferent and probably not impartial? I have the impression that it is not the word that has been lost but the meaning.


Let me turn to my main subject. Our title is “A Linguistic Introduction,” and the authors mean to include themselves in the aggressive and prestigious battalion of scientific structural linguistics. There are long and dogmatic methodological excursions in the three most explored areas of the school: the classification and parsimonious notation of phonemes, morphemes, and grammar; these provide the terminology throughout and give the book its cachet. The book has a not unpleasant schizophrenia: it wants to give an introduction to all thinking about language, and also to champion structural linguistics as the true way. The tension is noticed in the Preface as follows:

But as humans and humanists we must take a broader historical, more tolerant view of the goals of scientific endeavor, and must grant all serious attempts to understand the how’s and what’s and even the why’s of language the honorific title of linguistics. [p. vii].

I am moved by the tolerance for the poor relations. (I am glad I do not have to get a job in one of the departments.) The expression “even the why’s” promises not much along this line, and not much is performed. As I have pointed out, there is nothing on communication, meaning, speech-behavior, the social-psychology of speech, the facts of speaking, or speech as action. But let me stick to the structural linguistics exemplified by these authors. (I am quite happy, by the way, with structural analysis in general; I even believe in it myself.)

My deepest trouble with some structural linguists, including Drs. Bloomfield and Newmark, is this: They pretend to be descriptive but they define or really postulate important parts of the subject out of existence, and then call cases of them irrelevant. They finally mean by a language, that of which they have constructed a simple grammar; but they call it English. That is, they end up like the mathematical economists who think they are talking political economy (and advise the President).

There is a total-systematic or “formal” procedure that makes sense for self-defining and self-limiting games, and these may not be trival but major actions of the soul, like mathematizing or (almost) music-making. There might conceivably be such an autonomous action of language-making; I take it that this was Hjelmslev’s idea. But this is not at any point a descriptive procedure. Also, our linguists insist on defining “language” not by its own system but by its functions of communicating, expressing, etc. By and large, since the phenomena of speaking and the relation of speaking to language contain so much that is initiating and even creative, and so much that is historical and social, and no doubt so much that is chance, in my opinion the proper role of structural analysis is here quite different from total-system-making: it is to give a reasoned history of concrete acts of speech, omitting as little as possible from the description of the actuality, and making as much structural sense of it as possible, including novelty, unique appropriateness, modification, history, and accident, as expected structural factors. In The Structure of Literature I tried to show that only such a posteriori finding of structure in unique works makes for structural criticism of good poetry, whereas the a priori application of a theory of genres works better with hack-word. In linguistics the situation is less drastic; common language is more institutional, there is more predictability. Nevertheless, meaningful language, language that makes a difference by being spoken (and no other notion of communication will hold up), makes that difference by very often, perhaps usually (and perhaps the exceptions prove the rule) modifying the institutional system. If so, communicating, contacting, expressing, playing, creating, or whatever else the speaking accomplishes, can be structurally understood only in its occurrences. Of course to have a good prior organon of grammar, phonetics, and history is essential, to know what to look for and to recognize, but it is an organon and not a theory. My judgment is that many structural linguists are usefully telling us something about de-coding and the rudiments of translation or (not so usefully) about the meaningless speech of public spokesmen, bad journalists, and school teachers, but very little new, or especially true, about common speech, poetic speech, or philosophical speech.


“Language,” say our authors in their first sentence, “is fundamentally the means by which men communicate, etc.” Surely not language, but speaking (and writing) is the means by which we communicate. Or do the authors have some theory that there are “concepts” in the “mind” that are coded or otherwise imitated in language, incidentally accompanied by contingencies that belong to speaking? I doubt that they or I have experienced anything like this—although there is a common neurosis of blocked speakers who “think” and then put their “thoughts” into language. (This is not speaking.)

“Meaningful utterances [are] utterances with samenesses that can be interpreted” (p. 5) and the language is the system of these samenesses that “can be presented and understood” (p. 9). Literally, this is quite impossible; transmitting a mere sameness without a difference would convey no meaning at all—if the universe were all pea-soup green nobody would much notice the color, nor talk about it. What our authors probably mean is that the speaker puts a novel or not expected item in a constant grammatical-form as in a framework; but even in such a case, it is the difference that claims attention and carries most of the meaning. (Where the insert is also expected, who listens?) To convey the difference, his “meaning,” a speaker will alter the samenesses if necessary, by tonal emphasis, altering the word order, perhaps dropping most of the framework altogether. The communication—and expression, unless it is a repetition-compulsion—depends on the modifying, and this occurs in the speaking, not in the “language” apart from the speaking. (In hearing a report on a familiar subject, or seeking particular information, one has his structures of meanings pretty well decided beforehand, probably not linguistically but in his own characteristic or technical schemes. In the important cases of routine chit-chat, political party-speeches, etc., it is the speaking as such, not “what is said,” that communicates.) Naturally, an experienced critic can make a list of some of the likely expected modifications, but it is also likely that the next lively or careless or inventive speaker will choose outside his list or improvise outside his list. I don’t mean at all that there is no such thing as English or grammar, but that in both common speech and poetry it is the modifying of it that is the means of expressing and communicating, and that the value of a language is in its plasticity and open spaces as well as in its toughness and regularity. (I think this is what Whorf was trying to say, or ought to have meant, when he claimed that the language entirely determined the thought: if a speaker speaks only the linguist’s “language,” he cannot really communicate anything.) I may be unusually picky, but my experience is that in concerned interpersonal talk and in literature—but not in official reports, etc.—the grammar, and either the phonetics or metrics, can be exactly given only uniquely and a posteriori.

Indeed, the doctrine of the definiteness of meaning of the lexical units and the vagueness of the syntactical units is almost upside-down. If “meaning” is what is expressed, or communicated, or the effect that the speech has, then in common speech and literature most meaning is carried by the syntax, and the phonology, by the style. The “denotation of things” apart from attitudes, evaluations, and interpersonal games, is not a large part of gabbing. A remark of Chomsky on this subject (in Syntactic Structures, p. 108) is characteristically intelligent and revealing:

The notion of “structural meaning” appears to be quite suspect, and it is questionable that the grammatical devices available in language are used consistently enough so that meaning can be assigned to them directly.

That is, the consistency and simplicity of the explanation are now taken to determine the nature and semantics of the language, even though the explanations are supposed to be empirical descriptions! He goes on to say:

Nevertheless we find many important correlations between syntactic structures and meaning [which] could form part of the subject-matter of a more general theory of language concerned with syntax and semantics.

How true! The bother is that some linguists, like our present authors, but not Chomsky, imagine that they are talking “fundamentally” about the “means by which men communicate.” If this is what language is, it is hard to see how the relations of syntax, semantics, and style can be otherwise than controlling. Yet not a word of this is in the book.


To bridge the often gaping chasm between the theoretical language and the corpus of speeches, our authors use various devices; let us consider, typically, their use of ellipsis, transformation rules, and the concept of para-language.

In ellipsis, the sentence is irregular but an omitted part is understood. Now such a thing as true ellipsis exists. But when, as in ordinary rapid speech, there gets to be a striking preponderance of omitted “structural” elements, does it not seem plausible to assume that in some kinds of close speech-action, the structure is really different from the structure in more formal and detached situations? “Look“—“Where?“—“There!” Were there ever, or are there, any other words involved? The co-presence of the speakers, the shared ambience, non-linguistically fulfill grammatical functions, which are not “omitted” but have not yet come to exist. Indeed, when, in such close situations, we come suddenly to “spell out” all the subjects and predicates and “dot the i’s and cross the t’s,” the pedantic speech is not very good English, it is as if just learned. Again, what shall we say of the interpretation of modern English imperatives as containing an ellipsis of the second person pronoun, in order to save the Subject-Predicate form of all sentences? I submit that this is not how we speak, feel, or mean our imperatives, whether as children or adults (though an Anglo-Saxon adult might have spoken so). Why should we understand an unexpressed “you” when the imperative speech-action itself is so directly addressed to the other? It is truer to say that an expressed “you” would be a redundancy.

Our authors’ interpretation here seems to be determined by Chomsky’s calamitous dictum SS80):

When we try to set up for English the simplest grammar that contains a phrase structure and transformational part, we find that the kernel consists of simple, declarative, active sentences, and that all other sentences can be described more simply as transforms.

By the same token, Bloomfield and Newmark imitate Chomsky’s derivation of passive verbs from active kernels. Now again, there is no doubt that transformation exists in ordinary speech, for instance in immediate analogies and their analogies. But it is a very different matter when the chain of transformation contains many links that are not language at all, but schematic passage-dummies, e.g. (p.264):

X $$$ do $$$ thy will $$$ M $$$ thy will $$$ be $$$—Dpart $$$ do $$$ by $$$ X $$$—M.2

If we traverse such dummy passages, we have to assume that we have some prior idea of the goal that is directing our speech-desire unerringly through such barren territory; but if we have such an idea, why didn’t we invent something directly to express it? To my ear, it is plausible that a learned expression like “The man was bitten by the dog” is a transform; it is hardly English; yet it would never have been generated if there were not a need for a passive, probably already invented e.g. “She’s blessed with children.” Our linguists might say that there is here an ellipsis: “blessed (by God).” I doubt it; nobody means that. Consider, “I’m hurt.” Or maybe we are to take “blessed and “hurt” as adjectives and not passive verbs at all? Possibly, but not crashingly.

A question: what, except as a system for translation by computer, is the use of the “simplest grammar of English” if it is not the grammar I speak and if it distorts or omits much of what I feel to say? Another question: considering the efflorescent, babbling, artificing, and multi-situational acting and using of language, isn’t it improbable that the “simplest grammar” could be the grammar, any more than the assembly-line analysis of a manufacture is like the way we make the thing by hand? I’m damned (adjective?) if the commands, wishes, questions, reflexives, vocatives, interjections, periphrases, and curt phrases of English are in fact transforms of simple, declarative, active sentences at all!

A third way of saving a theory “that accounts economically for the observed linguistic phenomena” (p. 83) is to call the hard cases “para-language.” When our authors aver, “It is not at all clear that para-linguistic phenomena should be treated as parts of linguistic units,” we may be sure that is the last we shall hear about them. Consider the phonemes. I have read somewhere that more than three per cent of speech consists of floor-holding and embarrassment-allaying wells and growls. Most speakers have a good vocabulary of meaningful inbreaths, clicks, and burrs. Speed, slowness, hesitation, silence variously communicate. None of these will be listed as possible phonemes. (Indeed, sometimes we are told absolutely, as by Sapir and Whorf, that only the listed sounds and combinations are utterable by English speakers—so tightly is the actuality in the grip of the economical theory! Apparently, some linguists have met up with only obsessional-neurotics who are frightened not to “make sense” and dare not speak nonsense-syllables or imitate Chinamen.) Not much, surely, will be said about interjections, or about acts of speaking, “language events,” as such; or conversely, about stammering or keeping still. Most of these things, our authors tell us, are “better approached in terms of other kinds of description, e.g. psychological.” (Curiously, Roman Jakobson says truly that psychological explanation, e.g. of aphasias, is impossible without linguistics. Where should we begin?)

Of course, the reasonable justification for excluding phonemeon as para-linguistic or as unworthy of much attention, is that they would intolerably complicate the theory (and notation); the aim is to construct a simple grammar that is the grammar of the main body of the sentences. The procedure of exclusion or inclusion is admittedly circular: “The characterization of a sentence by linguistic pause is circular: a linguistic pause [unlike, e.g. a pause for breath] is exactly one which terminates a sentence. [p. 75] (Note that a metric analysis would have to be more subtile.) What is implied here is that it is the fullfledged meaningful discourses that really determine what shall be regarded as the sounds and forms that are parts of the language. This makes good sense; it is the same principle we use in poetics in deciding what are the “parts of a poem,” or, in psychology, what is “part” of the experience.

Unfortunately, the usual palaver belies this good sense. For instance, in his lapidary construction of language from the least parts to full discourses, in Structural Linguistics, Zellig Harris comes on as if he were describing actual unprocessed speeches. Is it not a bit of a hoax to accumulate in parallel the vast treasury of paralinguistic speech which also somehow communicates, expresses, etc.? I am even more bothered—perhaps politically!—by the economism and the technologism; why do they want to make our speech so ship-shape, instead of also honoring the anomalous, the passionate, the imaginative, the fumbling? Why bring on Occam’s razor when the plain facts are that we continually over-determine and are often idiosyncratic? Most important, however, is that the excluding spirit leads to a grammar that is not true enough. It is good sense to select the parts for the whole, but then the whole must be the greatest possible whole, open to the most various and worthwhile potentialities of speaking. Otherwise, a factitious body of average sentences is taken for English. To be candid, my own scientific bias is just the other way. Somewhat like Wordsworth, and somewhat like Jespersen (who is singularly slighted in this book on English), I would expect the truth about an initiating and creative function like language to be revealed best in spontaneous or high-powered deliveries, passionate or scientific excitement, high poetry and rhetoric, homely jabber and community lore.

The para-linguistic may make us understand the grammar. Let me illustrate this. In a remarkable sentence (SS67), Chomsky says, “In the simplest grammar of English, there is never any reason for incorporating ‘be’ into the class of Verbs…rather than be + Predicate.” This is a valuable insight. The contradictory cases—where “be” means exist—seem to be all rather special: “I think therefore I am”; “God was, and ever will be”; “I cannot but remember such things were” (Macduff); “To be or not to be.” But now let me speculate as follows: in modern English, speaking a sentence is an affirmation of existence: it is not neutral as to meaning.3 Therefore the copula does not have to carry the meaning. Then consider the important usual case: “You’re lazy”—“I’m not lazy”—“You are!” Without doubt we hear an ellipsis, “You are [lazy]”; but is it not even more true that “be” has re-emerged as a verb, since the mere speaking has not had enough force? Suddenly it is not the laziness but the inherence of it that is asserted—an interesting case of popular metaphysics. The next sentence will be, “You’re just like your father!” (Similarly, “Let the child be!” is emphatic and implies a previous sentence, perhaps unsaid.) Thus, the occurrence of a Language Event—which would be called paralinguistic—has consequences in the grammar.

Finally, the excluding spirit manifests not only an irrelevant parsimony but often an excessive purism.

Reasons from history, psychology, poetics, and rhetoric, are excluded from the rules, though they may be intrinsic to the behavior of speakers. For instance (p. 279), Bloomfield and Newmark reject “daily, weekly, monthly” (add “hourly, fortnightly”) as a sub-class because we cannot say “minutely, decadely.” Why not draw on the historical explanation for a rule? “Minute” and “decade” are later and Roman—presumably the older English did not know about minutes or reckon in decades. At first for strong social and political reasons, and later for stylistic reasons that still persist, there is a prejudice against Hybrids. And beautifully, the prejudice can also be overridden: “Kingly, queenly, knightly,” but “princely,” though not “dukely” or “baronly.” Such messes do not give the “simplest consistent grammar,” but they do cast light on the real one.


There is a peculiar fetishism of Language in modern linguistic theory; the human speakers and their occasions have quite vanished. In the extreme of the anthropological position—in the line of Whorf, Trager, and Edward Hall—the pattern of “communication” becomes the only pattern of culture, and communicating is nothing but ineluctably following the laws of the language, including the socialized roles and routines as part of the language; this is “meaning” and all the meaning there is. In the extreme of the formalist position, on the other hand, “the linguist sets up a general calculus in which all conceivable cases are foreseen—this calculus is deduced from the established definitions independently of all experience; linguistic theory cannot be verified by references to existing texts and languages” (Hjelmslev). Presumably the two extremes can be brought together by constructing the calculus that is in fact “applied” by the pattern of culture, the system of communication. This (plus Love) constitutes what Teilhard de Chardin calls The Noosphere, the next metamorphosis of man.

It does not seem to allow for much novelty, animality, political, or cultural change, does it? It is hard to see why people should bother to talk through the foreseen “conceivable cases.” But then, it is hard to understand why people run through many of the other paces that we do. Maybe it’s because of the language.

This Issue

May 14, 1964