Preface to the Second Edition
Linguistics, strangely, did not stand still after this book first came out. Much has happened in the field, so much has changed with this new edition. But most of those linguistic happenings have direct, if not exactly smooth, trajectories out from the battle over Deep Structure that sparks this story. They follow a clear and coherent narrative; in fact, two clear and coherent and mutually informing narratives, each of them international in scope, though each radiating from one of the two American coasts.
From the one coast, with Noam Chomsky’s MIT-rooted programme of generative grammar, the story grows through multiple, somewhat looping iterations of a framework with an atomistic picture of meaning, a procedural ethos, and the computational allegiance that has always been Chomsky’s, and MIT’s, calling card. It grows toward an evolutionary framework of the sort many people thought it should have had all along, all the while somehow maintaining its fundamental idealism, and has now adopted the label Biolinguistics. From the other coast, out of a Berkeley hot house teeming with figurative, encyclopaedic, latticed approaches to meaning based in the moment-to-moment creativity of everyday speakers, grows a narrative arc that flourished into a framework known as Cognitive Linguistics. In most of these overlapping developments, one finds a guy with perhaps the greenest thumb in linguistics, George Lakoff.
While these two trajectories might still look like the two solitudes suggested by the acrimony from which they sprung, I am hopeful. “Every discourse on metaphor,” Umberto Eco has said, perhaps thinking mostly of the arguments that could be heard around Berkeley, against the backdrop of MIT,
originates in a radical choice: either (a) language is by nature, and originally, metaphorical, and the mechanism of metaphor establishes linguistic activity . . . or (b) language (and every other semiotic system) is a rule-governed mechanism, a predictive machine that says which phrases can be generated and which not, and which from those able to be generated are ‘good’ or ‘correct’, or endowed with sense; a machine with regard to which the metaphor constitutes a breakdown, a malfunction, an unaccountable outcome. (Eco 1983: 218)
But there is a plot development shared by our two narratives that indicates theories of language may not have to live always and exclusively either in Eco’s (a)-world or Eco’s (b)-world: the long, slow, attenuation of hostilities. The productivity of our two stories, and the great energy that linguists have poured into each of them, and the (alas, not universal) decline of antagonism in linguistics bodes promisingly for the Chomskyan and Lakovian frameworks to become as mutually informing as their narratives. Languages are creative junctions of our personal and cultural lives, rich with figures and lattices of meaning, under daily renewal. They are also systematic and predictive instruments. It’s time the twain shook hands.
So, the book has changed a good deal. The reception the first edition got has been wonderfully valuable in making these changes. Uniformly, on a personal level, the reception has been encouraging. I can’t go to a linguistics conference without a handful or so of kind strangers asking me if I am the author of The Linguistic Wars and then telling me how much they enjoyed it. Often in recent years they have also taken to asking about whether I plan to write about Minimalism or Biolinguistics or Frame Semantics or Construction Grammar or the Cognitive Linguistics program more generally, because of their perceptions that these developments are natural outgrowths of the wars. (I have also been asked if I will write about Chomsky’s and/or Lakoff’s politics. No.)
On a scholarly level, there have been two phases of the book’s reception, the initial reviews and its uptake into academic and journalistic accounts of Chomsky, Lakoff, and linguistics. I have to say that my favourite example of its uptake in the academic literature is “Idiostylistic Peculiarities of Fabricated Knowledge in the Monograph The Linguistics Wars by R.A. Harris” in the Russian Linguistic Bulletin, with its irresistible compliment, “The author makes the most use of fantastical knowledge in these spheres to create a persuasive system of images to realize his cognitive and communicative intention” (Menshakova 2016:1). But my love of that article does not diminish my appreciation of the others, even the ones that have used swear words like post-modern and Foucauldian of my work. I am especially grateful to the comments that have provided correctives to the book or pushed me into fuller perspectives of the history. I have responded to them as best I can, often quietly incorporating them but sometimes taking them up in the text or its notes.
One especially good thing in the Linguistics Wars literature is that Noam Chomsky and George Lakoff (who did not like my book much at all the first-time round, and who I have no reason to believe will have any more affection for it this time) have each now published a few accounts of the Deep Structure battles and their fallout. The reasons for their disagreement with my telling of those developments, therefore, is now public; Chomsky has also commented in a few places directly on the book itself. I have incorporated their accounts, though not uncritically, and they can be checked out independently. The three versions, Chomsky’s, Lakoff’s, and mine are all quite discrepant. They are likely all true (mine certainly is, to the best of my ability, and I give them both the benefit of the doubt). I will only plead that their truths are personal. Mine is historical.
The book has changed, but in some senses not enough. The major regret I have with this edition is that the scholarship on the history of linguistics has grown impressively in the decades after the first edition, giving a much rounder picture of both the Bloomfieldians, whom I largely caricatured in that edition, even more than I realized at the time, and also of Chomsky’s rise. In an early draft of this edition I incorporated the insights and findings of that scholarship, but the manuscript ballooned up so alarmingly (to the publisher) that I had to trim it way back, and all of that material ended up on the cutting-room floor, along with some bits and pieces from the post-bellum coverage.
I know there are people who may feel an apology is due to them with my rewriting of The Linguistic Wars, some for neglecting what may have been better advice than I thought it to be at the time, some for overlooking or forgetting their critiques, but some also for my being ‘hard’ on them in the text. It is a solid bet that I have offended at least one particular person. But it is a tough business being responsible to facts and conclusions that may be inconvenient to some people without being unkind, which I have tried very hard to do. If I have not always succeeded it was not for lack of effort. It was, rather, that the scales pulled too completely toward accuracy, clarity, and the strength of the interpretation that emerged from those facts and conclusions. For anyone who feels ill-treated, I offer a blanket, sincere, but rather shapeless apology that the record, as I saw it, got in the way your sense of self. But I also want to offer one highly specific apology as well: sorry, Jamin, Ms. Piggy did not make the final cut.
I also apologize to the many people who have helped me with this project who I am now slighting in this preface. There are just far too many of you to name, or even remember, but I am grateful for our discussions and correspondence. There are a few, however, whom it would be criminal to omit. I repeat my profound thanks to Haj Ross, Paul Postal, George Lakoff, Noam Chomsky, Ray Jackendoff, and the lamentably late Jim McCawley, as well as the many others catalogued in the first-edition preface, listed at the head of the Works Cited list, or scattered among the pages of the book, who gave me so much support when I was finding my way into this story. I thank my Oxford referees for their generosity in giving me such extensive and detailed suggestions on a swollen and lumpy earlier draft of this edition, including the one who seems to have been very angry with it. I thank the lamentably late John Ohala, and Manjari Ohala, for the wonderful photograph on the cover. It may not go especially well with the “edgy” and “confrontational” design that OUP hopes will give the book more shelf appeal, but it is such a perfect image of Chomsky in the period it had to be included. Among my many interlocutors on these topics over the years, I need to single out Noam Chomsky (again), John Goldsmith, John Joseph, Robin Lakoff, Paul Postal (again), and Geoff Pullum, and to single out from among those singulars Frederick J. Newmeyer. Beyond all his specific, individualized help, Fritz’s continually inspiring scholarship has set the bar in this field for rigour and lucidity.
The first printing of this edition contained an error concerning claims in Noam Chomsky’s Undergraduate Thesis. I would like to thank Bruce Nevin and Geoffrey Pullum for pointing that error out to me, and Peter Daniels (2010) for the scholarship I overlooked when reproducing that error, which Oxford University Press has promptly corrected.
Much has happened in linguistics since the first edition of this book; so, too, in my rather smaller corner of the cosmos, life-shaking and life-making happenings. Their names are Galen and Oriana, the bright one and the golden one, though both are golden, both brighter than bright.
It all starts, and will never end, with Indira.
Milton, Ontario
R.A.H
Preface to the First Edition
Approach
There is, most unfortunately, a widespread ignorance and trepidation about linguistics—peculiarly so, since language is unutterably fundamental to our human- hood. Look at this odd comparison Herbert Spencer offered in the nineteenth century (when linguistics was making spectacular advances):
Astonished at the performances of the English plough, the Hindoos paint it, set it up, and worship it; thus turning a tool into an idol. Linguists do the same with language. (1865 [1852):33)
The comparison, in addition to inventing some mythically idiotic Hindoos, has it exactly backwards. Linguists would take out their wrenches and screwdrivers, pull the plow apart, and try to figure out how it works. The ignorant worship and fear. Scientists worship and investigate. But the attitude Spencer displays—something we might call linguiphobia if that term didn’t conjure up a fear of certain sexual practices or pasta cuts—has not abated. Continuing to pick on the British, we can cite one Auberon Waugh, who finds contemporary linguists to be so evil as to play into the hands of Neanderthal conservatives:
Linguistics [has been] reduced by Chomsky and his disciples to a positively mind-boggling level of stupidity and insignificance. If ever [the Prime Minister] wants an excuse to close down a university, she has only to look at its department of linguistics. (1988)
Ignorance is a kind word for Waugh, who is in need of kind words, but his isn’t a unique ignorance.
This book—a “popular science” look at linguistics by way of narrating an influential dispute in the sixties and seventies—attempts to clarify what linguists do, why they do it, and why everyone else should care about what they do.
My hope is that linguists will find this book useful, since many of them have a shaky or partisan view of their own recent history, but my greater hope is that non linguists will find an entertaining and informative account of the science of our most profound and pervasive human attribute, language.
Technical Notes
The bibliographical references in the text are quite standard. Citations are given by author, year of publication, and page number—usually in the main text—with the Works Cited accordingly organized primarily by author and year of publication. (A lowercase letter is suffixed to the year of publication, in the text and the Works Cited, when there is more than one bibliographical entry in the same year by the same author.) But there are two wrinkles.
First, there was a great deal of underground literature circulating during the period of interest, only some of which eventually surfaced, so that time-of-composition is often more important than time-of-publication, and I have tried to provide a bit of a road map here by including year-of-composition, in brackets, after year-of-publication. For instance, George Lakoff’s claim that “a generative semantic theory may well be simpler and more economical than an interpretive theory” shows up as “Lakoff (1976a [1963]:50).” Consistency called for me to maintain this style for most other authors as well, leading at times to some awkward looking citations like “James (1981 [1907]),” for his Pragmatism lectures. The only authors for whom I have avoided this style are the ancients—preferring the conventional “Aristotle (Rhetoric 1358b)” to the ugly and overly specific “Aristotle (1954 [c355 B.C.]:32).”
Second, the research for this book blurred at times into an oral-history project, generating hundreds of pages of interview transcripts, hundreds of pages more of letters and e-mail printouts, and almost as many pages of telephone bills. When quoting from this material—which I do quite extensively in the later chapters there is no citation at all. An embedded quotation, then, like “In Ross’s terms, Lakoff ‘is fearless, absolutely fearless’,” without further attribution, is by default a remark made directly to me by Ross.
Acknowledgments
I would like, very gratefully, to acknowledge the contributions and assistance of a great many people to this book.
First, I thank the many scholars who responded to my letters, e-mail, phone calls, and hallway ambushes with (mostly) good will, candor, and generosity: Dwight Bolinger, Guy Carden, Wallace Chafe, Bruce Derwing, Matthew Dryer, Joseph Emonds, Charles Fillmore, Susan Fischer, Bruce Fraser, Gerald Gazdar, Allan Gleason, John Goldsmith, Joseph Greenberg, Geoffrey Huck, Robert Kirsner, E. F. Konrad Koerner, Yuki Kuroda, Robin Lakoff, Judith Levi, George Miller, Stephen Murray, Greg Myers, Margaret Nizhnikov, Gary Prideaux, Geoffrey Pullum, Wallace Reid, Sally Rice, Sarah Grey Thomason, Don Walker, and Anna Wierzbicka.
Second, and with even more gratitude, I thank the scholars who set aside valuable time to speak with me at length, also with (mostly) good will, candor, and generosity: Thomas Bever, Noam Chomsky, Jerry Fodor, Morris Halle, Ray Jackendoff, Jerrold Katz, Jay Keyser, Susumu Kuno, George Lakoff, Howard Lasnik, Robert Lees, James Mccawley, Frederick Newmeyer, Paul Postal, Haj Ross, Jerrold Sadock, and Arnold Zwicky.
Bolinger, Chafe, Chomsky, Gleason, Jackendoff, Lakoff, Lasnik, Mccawley, Murray, Myers, Newmeyer, Postal, Prideaux, and Ross deserve still more gratitude, for reading and commenting on various earlier bits and pieces of this book, as do Victoria Bergval, Allen Fitchen, Michael Halloran, Lou Hammer, Pete Matthews, Leo Mos, Terry Nearey, Cynthia Read, Nicolette Saina, and Jim Zappen. Cynthia, my editor, gets another large measure of gratitude for cheerfulness beyond the call of duty. Many thanks also to Susan Hannan, Kim Honeyford, Sylvia Moon, Jean Mulder, and Philip Reynolds.
Not all of these people agree with what this book says, of course, or how it says what it says, and most will happily insist that any flaws are mine alone, but two people disagree so violently with the substance of this book as to require special notice, Chomsky and Lakoff. Both had very extensive comments on the same previous incarnation of the book, comments which I found mostly very profitable, and for which I remain extremely grateful, but both had very strong negative responses to the overall arrangement and orientation. Their responses were essentially inverse, Lakoff finding me to have sided with the interpretive semanticists, Chomsky finding me to have told the generative semantics version, both feeling that I slighted or ignored their own impressions or interpretations of the dispute. I should stress that the version they saw is very different in many ways from the one in your hands, but I have reason to believe that neither of them will be much more pleased with this version (their displeasure, in fact, may very well increase, since some of the latent elements that they found objectionable in the earlier version are stated a little more directly here; my correspondence with them sharpened my judgments on several matters, sometimes in directions neither of them would have preferred). Indeed, Chomsky even objects strenuously to my characterization of him in this preface; he sees no “symmetry” between his and Lakoff’s opposition to the book. I am naturally distressed by their negative reactions, but it would have unquestionably been impossible to satisfy both; perhaps by satisfying neither, I am closer to neutrality than either of them believe.
I also thank the interlibrary loan people at the University of Alberta, and especially, at Rensselaer Polytechnic Institute, for above-and-beyond-the-call assistance; Madam X, for diverting to higher education a little time, money, and photocopying resources that the State of New York had not allocated to higher education; the many people at Bell-Northern Research who supported my work, especially Bill Hosier, John Phillips, and Andy Sutcliffe; the Heritage Trust Fund, Rensselaer Polytechnic Institute, the Social Sciences and Humanities Research Council of Canada, and, particularly, the estate of Dorothy and Izaak Walton Killam, for financial assistance; the editors and publishers of Historiographia Linguistica, Rhetoric Review, Rhetoric Society Quarterly, College English, Word, and Rhetoric and Ideology (Kneupper, 1989) for permission to reprint the portions of this book that have appeared elsewhere; and the staff and students of the department of linguistics, at the University of Alberta, for stimulating discussions of my work and theirs.
Lastly, and therefore most importantly, I thank my wife and guiding light, Indira Naidoo-Harris.
Waterloo, Ontario
R.A.H.
1
But from what cause does it happen, that a good speaker no sooner conceives what he would express, than the letters, syllables, and words arrange themselves according to innumerable rules of speech, while he never thinks of these rules? He means to express certain sentiments; in order to do this properly, a selection must be made of the materials out of many thousands. He makes this selection without any expense of time or thought. The materials selected must be arranged in a particular order, according to innumerable rules of grammar, logic, and rhetoric, and accompanied with a particular tone and emphasis. He does all this as it were by inspiration, without thinking of any of these rules, and without breaking one of them.
—Thomas Reid (1827 [1788]:496)
Language is the liquid that we’re all dissolved in. Great for solving problems, after it creates a problem.
—Isaac Brock (2004)
Language is the strangest and most powerful thing ever to exist on this planet. All the other, more mundane and less powerful things, like nuclear weapons, quantum computers, and antibiotics, would be literally unthinkable without language. Even volcanoes and tsunamis pale in comparison. But it is only noise and thought. Molecules pelt ear drums. Cascades of electro-chemical charges fan out into a brain, into consciousness, evoking thought. Or alternatively, photons pelt retinas with abstract pictures of noises (letters). Clacking tongues or clacking keyboards make material impressions that trigger mental impressions and, somewhere in there, we’re thinking.
This book tells the story of a few remarkable linguists, of the coöperative and cantankerous pursuit they share, linguistics, and therefore, of the way we clack and the way we think. From now on, we’ll mostly refer to the activity at the clacking end of the process as form; at the thinking end, meaning.
Figuring out the network of correspondences that get us from form to meaning is the job of linguists. They sort out why it is that one set of forms (say, the ones that might be uttered at the breakfast table conventionally represented as “Pass me the herring, will you please, Ragnar?”) seamlessly coordinates action but another set (say, “The me you herring pass Ragnar please will”) makes people think you’ve had a stroke. We have a grammar, linguists say, a largely submerged shared awareness of how our languages work, or we couldn’t use them. In seamless-coordination events, no one notices grammar; in wonky-language events, the grammar seems to be broken. But what does that grammar look like? That’s at the heart of our story.
It’s a tough job, linguistics. Language is full of conventions, so it’s social. Language resides in brains and percolates through minds, so it is neurocognitive. Evolutionarily, it showed up along the way and made us the peculiar sorts of symbol-using animals we are, so it’s biological. Language is creative and communicative, emotive and informative, full of music and full of data . . . making linguistics a tough job.
Our story centres on a merry band of linguists who, setting out to work on these problems in the famously fractious decade of the 1960s, determined the course of twenty-first century language studies. At the center of it all is Noam Chomsky—in one common epithet, the “Einstein of linguistics”—with assorted students and colleagues. Chomsky and this band were for an exuberant while working together, then at cross-purposes, then in embittered opposition, and ultimately in different silos. In a gross caricature of the dispute, but one that often characterized it at the time: One side thought form drove meaning, that getting the herring from Ragnar begins with the rules that determined “Pass me the herring” was good piece of English; the other, that meaning drove form, that everything started with concepts like HERRING and PASS and ME.
The form-in-the-driver’s-seat side of the dispute flew a number of flags, but we’ll mostly call that program Interpretive Semantics because its model of semantics (the what-does-that-mean? part of linguistics) only came into play after the forms were already lined up nicely in place. For them, syntax was generative. Chomsky was the intellectual and spiritual leader of the Interpretivists, and also forged most of their weapons.
The meaning-in-the-driver’s-seat side rallied enthusiastically round one flag, Generative Semantics, because in that model, meaning generated form. It was mostly leadership by reluctant committee on that side, but one guy was the most eager flag-bearer, someone who grew from a Chomskyan acolyte to an opponent of such proportions that he now rejects, point-by-point, virtually every aspect of Chomsky’s work: George Lakoff. As we will see, Chomsky has done a lot of work, contributing to a lot of fields—psychology, philosophy, mathematics, computer science—and Lakoff has repudiated virtually all of it. He has also become perhaps the most influential linguist of this current century, much of that influence orbiting around a notion quite alien to Chomsky’s framework: metaphor.
So, our brains, our culture, our evolution, even our machinery are all commingled among the coöperating and competing views that shaped the Linguistics Wars and defined the course of linguistics since, and we will get to all of them in the course of our chronicle. But it all starts with Chomsky. In the decade before the dispute he pulled together several converging lines of thought in linguistics, logic, psychology, and mathematics, splicing them into a radically new framework for studying language and understanding us humans. The clash that defines our story would have been inconceivable without Chomsky’s dominating synthesis; and the linguistics we now have would be equally inconceivable without the frayings and the branchings of that framework through ceaseless renovations and periodic rejections of that framework, over more than sixty years, starting with the Generativist Heresy.
Chomsky’s version of our story is strikingly anomalous: Generative Semantics “threw a war” but he didn’t go; or, rather, okay, when urged by a former student, he reluctantly answered a few of their arguments, in a solitary surgical strike, flying in and out of Texas for a conference in 1969. Other than that, he was wholly uninvolved and out of the loop. So far from opposing Generative Semantics, he says, he actively supported its leading figures. Besides, he was too busy opposing something else, the Vietnam War, to find the time or the energy for such a trifling matter, even if he wanted to, which he didn’t, so there. Any claims to the contrary are just laughable, absurd, delusional, “fabricated,” he says, “by enthusiastic postmodernists” (Hughes 2006 [2001]:86). This account is the one he gives to anyone who asks, and occasionally it gets published as fact. Here is how it looks in his most notorious biography:
I was never really part of it. . . . My one participation in the debate was in 1969, at a conference in Texas, where I flew in and flew out immediately at the impassioned request of a former student there, Stanley Peters, who wanted me to make some public response to the by then rather hysterical tone of the generative semanticists, all pretty childish in my opinion. (Letter to Barsky, April 3, 1995; Barsky 1997a:151)
Chomsky is virtually alone in this version of events,1 and this book charts out a quite different story from the one he tells, but we can let it stand for a few chapters, while we fill in the necessary background.
Lakoff, too, has an alternate version, more intricate and extended than Chomsky’s, but no less personal: lots and lots happened, and he was everywhere, making it happen. He invented Generative Semantics having little more than bored awareness of Chomsky’s work, he says, from one uninspiring undergraduate course he took as part of his English literature degree at MIT. He invented it as a story grammar for his 1962 undergraduate thesis. It was a flame he fed through the years as more data and more ideas came to him, as collaborators and enemies appeared, as opportunities and barriers arose, as errors and omissions and solutions and insights and shifting hierarchies of values continually defined and redefined his understanding of language, the mind, and being human. Chomsky, barely a bit player at the start of Lakoff’s version, enters as a looming malevolence once Generative Semantics goes public, the villain whose machinations drive the plot through to its climax, in the 1970s, when Lakoff disproved both Generative and Interpretive Semantics himself, and turned toward forging Cognitive Linguistics, with its distinctive integration of literary and rhetorical notions, like metaphor and metonymy.2
I don’t mean to say that the accounts of neither Chomsky nor Lakoff can be believed. Chomsky certainly had other things on his mind, and his published output reflects mostly a theoretical disinterest in Generative Semantics work after 1969. Lakoff, meanwhile, was everywhere, a whirligig of theoretical, methodological, and data-mongering energy. Neither version serves very well to characterize the field overall, in my view, serving largely to structure each man’s personal self-image. But keep them both in mind as we proceed, and when we get to the end, you can map them against the account I lay out in the chapters ahead.
Rhetoric is part of our story as well, not just because of the way its notions enter into the substance of the dispute, but because of all the rhetoric—the arguments, the appeals, the patterning of issues—that constitutes the dispute.
The first people who drew a paycheck for studying and teaching about language—the ancient grammarians—were aware of the mutual intellectual debts rhetoric and grammar owe each other (as well as to the shared debt both fields owe philosophy; all three of these fields show up throughout our story). In Suetonius’s Lives of Illustrious Men: Grammarians and Rhetoricians, there are no masters of one who are not also scholars of the other, and the truly accomplished can be called “a rhetorician among grammarians and a grammarian among rhetoricians” (Suetonius 1998 [c.119].2:395). In Suetonius, one also finds, rather as we find today, that the Grammatici were generous, collegial, scrupulous; mean, paranoiac, underhanded; indolent and industrious; loyal and treacherous. Staberius Eros was so fearlessly open-hearted that he taught the children of assassinated state enemies for free. Lucius Orbilius Pupillus of Beneventum, intemperate to rivals and students alike, and even to the parents of students, earned himself the nickname plagosum, “the flogger.” Lucius Crassicus, at the height of his fame and wealth, was overcome by humility and disbanded his school to become a Stoic.3 One finds among the grammarians savants and cranks, confederates and adversaries, patrons and disciples, friends and enemies, benefactors and crooks; one finds, that is, people. We find them in our story too.
Linguistics and linguists are the subject of this book, but so is language. That gives us a special problem. Language is the method and the material, the chisel and the stone—points which, language being language, often slip noiselessly away while more immediate matters occupy us, but which, language being language, also crash in from the shadows when we least expect them. Sometimes we seem to look right through language, hear right past it, and apprehend directly the ideas behind the writing on the page, the words in the air; sometimes we can’t get it out of the way. When we’re actually in the business of languaging, as you and I now are, we want a frictionless flow, an invisible medium connecting us, head to head. Not linguists. Not when they’re working. They come at language from the other end. Their job depends on making language visible, everything from the tiniest features of single sounds to meanings and intentions.
And they come at it obstreperously, Emmon Bach tells us in an echo of Suetonious. “Linguists,” Bach said before our dispute had even begun, “are notoriously cantankerous creatures” (1965b:128). Bach would know. He was one, an exceptionally good one who shows up in the thick of our story a few times. But for now, let’s just take his point: linguists don’t always agree, even about the most fundamental matters. We’re not just talking about Chomsky and Lakoff here, who came to occupy the North and South Poles of Planet Language, but the whole lot of them. Since many of these disagreements arise in the story of the Linguistics Wars, maybe we can air them with a quick survey of the dramatis personae. Here are the principal characters, with two of whom you are already acquainted:
Noam Chomsky (1928–): Hero of the Cognitive Revolution, architect of a theory of language and mind that defined the battleground of the Linguistics Wars, the most influential linguist of his, or any, generation; a legend of epochal proportions.
Jerrold (Jerry) Katz (1932–2002): A philosopher, among the earliest scholars to join Chomsky’s project, helping to craft its first theory of meaning; an antagonist of both camps after the Wars broke out.
Paul Martin Postal (1936–): An early ally of Chomsky, a trenchant and abrasive advocate who worked closely with Katz to integrate semantics into the developing program, but who soon found himself on the outs and became an intellectual leader of Generative Semantics; its most resolute theorist.
George Lakoff (1941–): An eager but transient Chomskyan, coiner of the term Generative Semantics; a tireless and ebullient, also restless and protean, evangelist for the movement; its most wide-ranging and creative force; husband of Robin.
Robin Lakoff (1942–): A Chomskyan enthusiast whose extension of his program to the depths of Latin diverged from his interests in a way that pushed the grammatical theory envelope toward Generative Semantics; wife of George.
James D. (Jim) McCawley (1938–1999): One of Chomsky’s students who became an early and forceful advocate for Generative Semantics, producing some of its most compelling arguments.
John Robert (Haj) Ross (1938–): One of Chomsky’s students, whose work was immensely influential on both sides of the debate, a close confederate of Lakoff’s in sparking what was soon seen as the Generative Semantics revolt.
Ray Jackendoff (1945–): Yes, yet another one Chomsky’s students, only a couple years after Ross and McCawley but, coming on line in the wake of the Ross-Lakoff revolt, he became the chief articulator of Chomsky’s technical shifts and one of his strongest partisans in the Wars.
Chomsky, Chomsky, Chomsky; the common denominator among our personae. One of them is Chomsky. The others all have trajectories that move from reflection to rejection of Chomsky in variously differing but utterly fundamental ways, though—and please don’t miss this part—various lines of agreement were also maintained. Nor, ahem, has Chomsky’s trajectory been particularly uniform, with Chomsky1967 and Chomsky1980 and Chomsky2002 and Chomskyhe-may-not-be-done-yet variously reflecting and rejecting, as well as deflecting, earlier versions of his own research program.
All of these people, and more, drive the story of the Linguistics Wars, and all of them figure very prominently in its wake, which has determined the current shape of the field—that is has given us our best theories of what language is, how it works, and where it comes from, the web of correspondences between forms and the meanings they (somehow) evoke.
Forms, most immediately, are linguistic sounds but also the letters of written languages (like English), the characters of pictographic languages (like Mandarin), and the gestures of signing languages (like American Sign Language). They are observable, measurable material things: noises, marks, movements.
Meaning is where clarity ebbs. It runs the gamut from logical and grammatical concepts (like negation and subject/predicate relations) to the blurrier domains of implication and suggestion—like getting Ragnar to pass you the herring not by requesting it, but by expostulating “Jeez, Ragnar! You gonna hoard all those herring to yourself?!” At this end of the correspondences, we have the private/collective interface of understanding. It’s not just our knowledge of the code that matters in this expostulation (what meanings are tethered directly to forms like hoard and herring, or what we know about how sentences work), but our knowledge of the situation and even the personality of the speaker. Perhaps it’s a threat (you regularly box Ragnar’s ears for herring-hoarding), or a reminder of some kind (mother is coming and she dislikes the herring anywhere but the center of the table); maybe there’s irony or anger or desire in play.
But even if we filter off the personal and the situational (as most linguists tend to do) and ratchet things back only to bare forms, the various things going on are still pretty tricky, vastly trickier than the people who make linguistic noises suspect. Most native English speakers give a little puff of extra air when they make the sound at the beginning of the word kill, but they don’t puff for what seems to be the the very same sound—it sounds the same to me, my mouth seems to be doing the same things, we write it with the same letter, k—when it shows up in the closely related word, skill. And here’s where it gets really weird: no one will believe you that there is a difference in air-puffage without elaborate demonstrations, but swap things around and they hear the difference plainly. You can do the demonstrations yourself, in the comfort of your own home, but you need some special equipment for the swapping. Here’s the demonstration: Say kill with a match or a Kleenex close to your lips. The match will flicker or go out. The Kleenex will move. Now say skill. There will be much less, or no, action at all with your props. Voila, puffage! For the swap-test, you need digitally slice the two sounds out of their home ground and patch them into the other word. You take the first sound of kill (let’s call it [kh] and second sound of skill (plain old [k]) and then trade them for each other. You’ll hear what sounds like gill for the puffless-k version of kill and Marilyn Monroe sounding hyper breathiness for the puffed-up-k version of skill. Weird, weird, weird, but it’s the slight violations of these hidden practices that give people accents in foreign languages.
It’s a tough job, linguistics. This kill/skill example is one very minor and mundane illustration with two words in one language, not even the tip of an iceberg; more like a chromosome in the cytoplasm of a bacterium on the head of a flea on the head of a mouse on the tip of an iceberg. The languages of the world are an utter tumult (or, if you’re a linguist, a swollen treasure chest) of such puzzling correspondences. Describing them is far more difficult than describing the trivial structures of something like a galaxy or a computer chip. It’s no wonder linguists don’t all do it all the same way.
We won’t be dealing very much with sound in this book. We’ll mostly be dealing with words and rules; or, as linguists put it for a long time, with dictionaries and grammars. That distinction will eventually blur under the attention it gets in this book, but in prototypical terms, “Grammar deals with the general facts of language, and [dictionaries] with special facts.” That was the great Danish linguist Otto Jespersen talking to us. He elaborates:
That cat denotes that particular animal is a special fact which concerns that word alone, but the formation of the plural by adding the sound -s is a general fact because it concerns a great many other words as well: rats, hats, works, books, caps, chiefs, etc. (Jespersen 1965 [1924]:32)
He might have added the observation that cat can follow the word black with or without that -s but can only follow two when dressed in that plural. That’s also a general fact in these terms, a fact of grammar; black cat is fine, in other words, but two cat is ungrammatical. (Actually, linguists have a special way to signal ungrammaticality, an asterisk, so the proper way to express it around them is *two cat.) All of the ways that forms combine with other forms are points of grammar in this view. Grammar is combinatoric.
Keep in mind, too, that we’re talking about the knowledge and the use of language when we’re talking about words and their combinatoric rules. The relevant dictionary is not the one published by Oxford or Merriam-Webster. The relevant dictionary is in your head. Same with the rules. That’s why Jespersen also says words and rules are “part of linguistic psychology or psychological linguistics” (1965 [1924]:29). . . . though, er, not all linguists would agree, and even when they do agree, they might agree for different reasons.
But it’s not a war of all against all in linguistics. They’re cantankerous, but they aren’t idiots. (I use, of course, the technical sense of idiot here—separate, distinct, isolated—that we see in words like idiosyncracy.) There are two basic approaches that chart the way different groups of linguists deal with words and rules: a what-goes-with-what orientation and a how-did-we-get-here orientation; or, as historians of linguistics, John Goldsmith and Geoff Huck, put it, the distributional orientation and the mediational orientation (Goldsmith & Huck 1991; Huck & Goldsmith 1995:5–58). Distributional tribes are preoccupied with the black-cat-but-*two-cat issues, sorting out the grammatical expressions from the ungrammatical ones. The mediational trbes are preoccupied with how it comes to be that when you pair forms like black and cat you can make the concept of a furry little mouser (that is, a Felis catus domesticus) show up in other people’s heads with the property of blackness. Tribes is not the best word, I admit. Distribution and mediation are tendencies, allegiances, for linguists, not all-consuming, opposing gestalts they fly as battle standards. It’s not that linguists can only see the faces of distribution or only see the candlestick of mediation. All linguists care deeply about the mediational job of language, about mapping the road between form and meaning. And all linguists care deeply about the distributional vehicles. But, well, they do sometimes go to war, as we will see, over the distributional “surface” correspondences and the mediational “deep” correspondences in language.
Conventionally, those correspondences run from the hard currency of sound to the mysterious payoff of meaning, from patterns of bouncing molecules to patterns of sparking neurons. Phonetics is the field that studies the acoustic dimensions of linguistic sound (speech noises). Phonology studies the clustering of those acoustic properties into recognizable sound units, corresponding very loosely to letters. Our kill/skill example operates in the phonetics/phonology borderlands. The two different sounds we make (with the puff of air and without the puff of air, [kh] and [k]) are phonetic data. The fact that we think of them as the same sound, so long as they obey certain placement rules, is a fact of phonology. Morphology studies the clustering of those cues into meaningful units (affixes and words—so, kill and skill but also the -er of killer and the un- of unskilled; in Jespersen’s example, it covers both cat and -s). Syntax studies the combination of those meaningful units into expressive sequences (phrases and sentences—so, the herring and Pass me the herring). Semantics studies the composite meaning of those sequences (claims, propositions—the fact that I’m hungry basically means that the speaker wants food). Pragmatics studies how those composite meanings do their daily situational work of cooperation and competition; that I’m hungry might be an announcement, or a complaint, or yet another way of getting Ragnar to pass the herring, depending on when and where and by whom it is uttered.
For anyone unfamiliar with linguistics, those terms and definitions may remain murky for a while. As they become relevant to our story, they should become clearer. Again, sound (so, phonetics and phonology) are peripheral to our story, but as a crash course in the other four, consider these sentences:
Morphology concerns the units, the words and meaningful subwords—that language is a noun, that -s marks the plural of that noun; that speak is a verb which can also take an -s (but this time, it is an agreement marker, not a plural; I told you languages were tricky), and has yet another form, spoken (technically, a past participle); and so on. That bridges us to syntax because words build into larger form-meaning units, phrases, and sentences. Syntax concerns the arrangement of words into phrases, and phrases into sentences—that speaks two languages is a verb phrase in 1a, that two languages is a noun phrase that ‘belongs’ in the verb phrase, that Everyone is another noun phrase, that the whole thing is a sentence. Syntax also tells us that lots of sentences have associated sentences—here, that 1a and 1b are the same basic words in systematically different arrangements. We see the same systematically different arrangements in these sentences:
Semantics concerns the proposition expressed by sentences—for instance, that 1a is true if and only if every member of the population of a place called ‘Yalis Island speaks two languages and that 2b is true if and only if a guy named Ragnar performed some action with the herring that satisfies meaning conditions of pass. Pragmatics, . . . well, let’s wait a moment on that one; James McCawley will help us with it shortly.
The events at the heart of this book—the work of Noam Chomsky, the semantic extensions it triggered, the rebellion those extensions begot, and the impact the resulting wrangle had on modern linguistics—have everything to do with thought and being human. The story begins with Chomsky’s compelling arguments that fundamental aspects of human behavior (linguistic creativity, in particular, and language acquisition) are inaccessible without his innovations. It develops further when his followers, principally the Lakoffs, McCawley, Ross, and Postal push this work deeper into the territory of thought than Chomsky was comfortable going. And it erupts into open warfare when Chomsky, soon abetted by the work of a newer rank of followers (headed up by Jackendoff), retrenches aspects of his approach to head off such extensions and to repudiate the extenders. How it ends, even if it ends, is controversial, but the Chomskyan view is that George Lakoff and company were routed for their irrationality and errors, and that linguistics is much the better for their defeat. Perhaps. Generative Semantics became something of a snide joke in linguistic orthodoxy for a generation or two. The movement deserves some of its shame; Chomsky’s camp, some of its glory. But one of the aims of this book is also to show that Generative Semantics was not without nobility or virtue, and that its lasting impact on linguistics was valuable; nor, inversely, was Chomsky’s camp engaged in pristine error correction with exclusively wholesome results for linguistics.
The events at the heart of this book also have everything to do with borders; more specifically, with border disputes. The borders involving phonetics and phonology saw very little action in the debate, but those among morphology, syntax, and semantics—the provinces more closely affiliated with meaning—were all flashpoints. At one end, some Generative Semanticists argued that language was one big schmoosh, with no place at all for borders, even in principle; sound was at one end of the linguistic continuum, meaning at the other, and a small group of uniform rules knit them together. At the other end, Chomsky’s camp, the Interpretive Semanticists, seemed to be boundary fetishists, redrawing their borders daily; one day a piece of data was syntactic, the next day morphological; one day it was semantic, the next syntactic. Each saw the other side as perverse, and each opened its guns on the perverted.
Such internal border wars are largely a matter of pitting one theory against another, much the same as disputes in other fields between a cloud-like subatomic model and a mini-solar-system model, between fixed continents and drifting ones, between Darwinian and Lamarckian evolution, between quantum theory and string theory. One of the ways science gets done is in the crucible of dispute. Generative Semantics wanted to leave the language pie pretty much as a whole, describing its shape and texture noninvasively. Interpretive Semantics wanted to slice it into more manageable pieces. But as the battle became fiercer another border war arose, an extra-theoretical one, concerning the definition of the entire field, the scope of language study, the answer to the question, What is linguistics?
Every science needs to slice their pies somewhere, though, or it can’t produce coherent explanations. Nature is in dizzying flux, abuzz with colliding, chaotic, bleary events; it is a realm of infinitesimal detail and immeasurable vastness. Our senses have adapted to this by tuning to only a tiny range of those events, the ones most relevant to our survival and propagation. We see only a certain narrow band of optical frequencies (light), hear only a small range of aural frequencies (sound), smell and taste and feel only the grossest of data. Everything else is filtered off by our incapacities. Sciences do exactly the same thing. Collectively they have overcome many organic limitations, augmenting some of our capacities astonishingly well with telescopes and microscopes and sonographs. But they also make far more exclusive choices than our sensual machinery. Even in the outlandishly general schemes of physicists working away on a Theory of Everything, only the narrowest of phenomena would be covered; a Theory of Everything would not explain, for instance, a moth drawn to a flame, a wolf baying at the moon, a physicist writing a grant proposal. Nor should it. Science, like any other form of apprehending the world, would be impossible without its self-imposed limits.
Chomsky argued forcefully that in linguistics such limits should be drawn between our knowledge of language and our use of language. Take the difference between knowing how to play chess and making a specific move. The first is relatively tidy—the rook goes horizontally and vertically, the bishop goes diagonally, the knight does a buttonhook. But actually playing your rook or your knight depends on a welter of ephemeral conditions—past moves, adversarial skill, emotional state, even the amount of light on the board or the cough of a spectator. The first can be described comprehensively by a body of rules. The second can only be described broadly, and never predicted with anything approaching certainty by an observer; among other things, that’s what makes it a game. In language, this kind of basic knowledge is relatively stable after childhood acquisition, though vocabulary and conceptual knowledge grow and decay. Use is altogether different. It is subject to all the vicissitudes of life—stress, distraction, altered states of consciousness. A speaker who knows the pronunciation of two and martinis might still claim to have had only “tee martoonies” if pulled over for erratic driving, especially in a golden era TV sitcom. Chomsky and the Interpretivists felt the only way to isolate tractable problems for linguistics to solve was to focus on knowledge and filter off the haphazard factors of use.
Generative Semanticists started from the same place, but soon found this approach absurd and arbitrary, regarding accounts of linguistic knowledge to be completely misleading when separated from the application of that knowledge. McCawley’s analogy for the Interpretivist separation of form and function was to a theory of the stomach that ignored digestion. And, of course, language is neither chess nor digestion. It is far messier and far less exact than chess, far more ramified than digestion, though perhaps not so messy. Nor is separating knowledge from use all that easy. However common the language-as-chess analogy is—and it is one of linguists’ most venerable topoi—‘knowledge of chess’ is not quite as uniform as it suggests. Garry Kasparov’s knowledge of chess is very different from my knowledge of chess and it results in different play, no matter what the light conditions are. Generative Semanticists, some of them, said there was no defensible separation at all. Responding appropriately to “Hot enough for you?” was equally as important for them as matching pronouns to their antecedents. For their part, Interpretive Semanticists seemed to shift their definitions daily. Yesterday’s knowledge was today’s use; today’s use, tomorrow’s knowledge—whatever allowed them to defend their own claims and attack their enemies’. Each side saw little more than depravity in the other’s methodological proclivities.
The story, then, is in large measure about how much is too much, about how big a bite of language is more than linguistics can chew. Chomsky charged the Generative Semanticists with trying to swallow every conceivable phenomenon with the most oblique relation to language. The return accusation was that Interpretive Semanticists took only conservative, tasteless, nutritionless little nibbles from the immense, immensely challenging, human phenomenon, language.
By language, our linguists initially pretty much meant just sentences when they started to diverge, and their focus was on grammaticality. What does it take for some assemblage of words to be a sanctioned sentence of some language (be grammatical)? What does it say about a given language, or the whole notion of language, if assemblage X belongs to a language (is grammatical), but assemblage Y doesn’t belong (is ungrammatical)? Our ‘Yalis Island sentences (1a and 1b) both belong to English. They’re grammatical. One is an active construction (1a), the other is a passive construction (1b). But now take a gander at this thing, courtesy of McCawley:
Whatever this clump of words is, at first pass it certainly doesn’t look like a grammatical sentence.
The issues and appeals surrounding 1a and 1b are very narrow, highly technical, and concern a single question about a single principle—should it stay or should it go? So, let’s set them aside for a while as good little English sentences with curious implications for theoretical machinery. But the issues and appeals surrounding 3a are very wide, largely informal, and go way beyond a single principle or a piece of theory.
By the time 3a surfaced as linguistic data, about six years after the first volleys, the Interpretive–Generative Semantics differences had outgrown and exhausted the term debate. What began as a compact, in-house disagreement over a single principle within Chomskyan linguistics had mushroomed to foundational proportions. Both sides saw the relevance of 1a and 1b very clearly, and both sides saw a resolution within reach. But the years of acrimony and diverging arguments between those sentences and the wonky-looking 3a had altered their vision. Interpretive Semanticists didn’t even see 3a as data (not without an asterisk, it’s not!), and regarded its invocation by the other camp as clear and damning evidence they were no longer even doing linguistics; Generative Semanticists saw the dismissal of data like 3a by the other camp as clear and damning evidence that they were practicing a brand of linguistics so sterile and ingrown as to be vacuous. Even the political and whimsical elements of 3a (Spiro was Spiro Agnew, Richard Nixon’s much ridiculed vice president) defined the chasm that had grown between the former associates.
When 1a and 1b were relevant to our story, all the arguers agreed closely about their consequences for theory, and about what sort of enterprise linguistics should be; with 3a, there was so little agreement that the word arguers hardly applies. But we are only catching the vaguest image of the conflict here, the shadows on the wall. In the mid-1960s, when the Linguistics Wars began, Interpretive and Generative Semanticists agreed with one another about how to study language, certainly, but they stood arm in arm against their immediate predecessors, whom we’ll call the Bloomfieldians. Chomsky’s early polemics had painted that generation as confused and bumbling amateurs, not real scientists, from whom his Transformational-Generative theory would rescue linguistics. By the mid-1970s, with Spiro’s laxative conjecture, the Bloomfieldians were still anathema to Chomsky’s immediate charges, the Interpretivists, but the Generative Semanticists now began to find points of agreement with pre-Chomskyan linguists. This shifting ground of agreements—that is, history—forms not only the defining backdrop for the Interpretive–Generative Semantics dispute, but for all the whys, whats, and hows of language study.
Before we get into that story, though, our Spiro “sentence” is still hanging around waiting for some sense to be made of it. McCawley’s point that when you take 3a in isolation, as Chomsky always prefers his data, it is just a hopelessly jumbled clump of words, ungrammatical because the verb, conjectures, doesn’t have the right companions. Conjectures, you see, has certain basic requirements that must be satisfied to earn a grammaticality seal of approval; don’t meet them, you get an asterisk. Conjectures requires an embedded sentence and can’t immediately precede a proper noun (like Ex-Lax). So, 3a is screwy (it’s got a proper noun), but 4 is perfectly fine (an embedded sentence):
In short, the combinatorics of 3a are awry. Conjectures should be combining with an embedded sentence, not something else. That’s all true for isolated data.
But, if you take 3a out of isolation, find it a context (the way Generative Semanticists grew to prefer their data), put it into some real-world use of words like conjectures and Ex-Lax, rather than just pinned into some textbook or academic argument about what sorts of verbs need what sorts of companions, then 3a is perfectly okay; namely, as a response to a question like 3b:
