The Atoms of Language: The Mind's Hidden Rules of Grammar
by Mark C. Baker
Basic Books, 2001
250 pp.; $28

Academic linguists are used to laymen assuming that we are professional polyglots—we are given to mimicking the eternal question "How many languages do you speak?" with affectionate exasperation. Many people also naturally assume that the linguist is a steward of the grammatical "pitfalls" one is taught to avoid, such as Billy and me for Billy and I—friends often joke that they feel like they need to "watch their grammar" around us.

Yet in reality, many linguists speak only one language, and not only are the tricky "blackboard grammar" rules irrelevant to what we do, but the very theoretical underpinnings of our field render them laughable little hoaxes. Mark Baker's The Atoms of Language is a welcome introduction to what many linguists are actually engaged in every day, helping to fill a glaring gap in the popular nonfiction literature until recently occupied only by Steven Pinker's bestseller, The Language Instinct.

Pinker built his presentation around the question of whether humans' ability to use language is innate. The book was masterfully written and provided an introduction to some fundamental tenets of modern linguistic theory. But the breadth of topics germane to the innateness argument, including data on infants' speech and language disorders, allowed for only brief consideration of the conception of grammatical structure with which Noam Chomsky revolutionized the linguistics field in the 1960s. Baker's is the first book aimed at a general readership that outlines the nuts and bolts of one of the main courses of current linguistics training and research—what is called the "Principles and Parameters" school.

The languages of the world differ in structure much more than is obvious from the typical American experience, where the languages we usually learn are relatives of English, thus built on a game plan familiar to us. For example, we all know that words for a given concept differ from language to language. But in the Native American language Mohawk (Baker's specialty), the entire sentence "He made the dress ugly for her" is expressed in a single word, Washakotya'tawitsherahetkvhta'se'.

Yet Baker's main point is that underlying the almost netherworldly appearance of a word like this are basic principles (i.e., those indicated in Principles and Parameters) that are constant across all of the world's languages, and that languages differ only according to a small set of binary choices as to how those principles will be expressed. A given pair of such alternate "settings" is a parameter in the Principles and Parameters framework.

For example, Baker shows that languages like Mohawk are built upon a parameter that requires that the entities that a verb in a sentence is driven by (i.e., the subject) or affects (e.g., its object) be hung on the verb itself rather than expressed as separate words as in English. Thus in the Mohawk word above, the -shako- part near the beginning is a prefix that means roughly "he acting upon her," incorporating what English would express with the separate pronouns he and her. The core of the whole word is the -hetkvht-, a verb meaning "to 'uglify,'" while the -tya'tawi- in the first half of the word means "dress." The whole word in a sense translates as "He dress-uglied her." While in English, words like baby-sit instead of sit the baby are relatively rare, in languages like Mohawk objects are regularly combined with verbs in this way. One "house-buys" rather than buying a house, and thus one "dress-uglifies" rather than uglifying the dress.

This parameter is one of a number which, while simple in themselves, can render languages radically distinct in appearance. In Japanese, "John read the letter in Tokyo" is John-ga Tokyo-ni tegami-o yonda, which comes out as "John Tokyo in the letter read." That seems unnatural to a native English speaker, but the difference between the Japanese and English sentences all boils down to a parameter choice. In Japanese, words with what we might call the semantic "juice" come after their "minions" rather than before them as they usually do in English. This means that verbs come after the objects that they affect (and thus "read" comes after "letter"), and prepositions come after what they apply to (thus "in" comes after "Tokyo"). Thus while to the layman, Japanese word order looks like an imposing scattering of random differences from English, in fact the difference all hinges on one difference in how the two languages set the parameter determining which side "minion" words fall on.

Baker shows that parameters get even more interesting in that they are not merely an unconnected collection but instead appear to constitute a flow-chart hierarchy: a given parameter setting leads to other parameter choices that a language that chose the other setting will never encounter. For example, one "principle" is that when the brain first produces what will become a sentence, a tense marker like the past-tense -ed is generated separately, and is "linked" with a verb like walk before the sentence is uttered, with the tense marker and the verb then united as "walked" when the sentence is pronounced.

Yet there are many languages that do not have tense endings like -ed, and instead express tense with separate words. In Mandarin Chinese, what marks the past in the sentence for He ate the meal is a separate word, le:

Ta He

chi eat

fan meal

le "-ed"

In languages like this, then, there is no need for a verb to be linked with a tense-marking suffix. These languages have set what Baker calls the Verb Attraction parameter on "negative."

This seemingly ho-hum observation becomes significant in light of the fact that many of the world's languages allow two or more verbs to be strung together in a single sentence. For example, in the African language Edo, to say Ozo will push the pot down one says:

Ozo Ozo

gha will

sua push

akhe pot

de fall

Here "push" and "fall" are squeezed into one sentence in a way that we would never dream of in English. Where we have two verbs in a sentence, one of them is an infinitive—I wanted him to know. In Edo, neither verb is an infintive—both are as "live" as the wanted in I wanted him to know. Edo exemplifies the Serial Verb parameter. Interestingly, all languages with this parameter are ones like Chinese with no tense-marking suffixes like -ed. In itself, there would seem to be no reason that it would be contrary to some cosmic sense of propriety to cram two verbs into one sentence even if both had tense suffixes on them.

But Baker shows that there is a systematic reason languages like this do not exist; e.g., why we cannot say Ozo pushed the pot fell in English. In a language with tense-marking suffixes, before the sentence is uttered, the universal principle dictates that the suffix must be linked to a verb. But that automatically means a sentence can have only one verb—any second verb that tried to "horn in" would have no suffix to "hook up" with (Baker cutely describes this as the second verb having "hurt feelings"). Thus you only get second verbs in languages like Chinese where there is no suffix to hook up with, the tense indicated instead with a separate word.

This demonstrates how parameters are "nested" in one another: if a language chooses the Verb Attraction parameter (that is, has suffixes to mark tense), then it is barred from opting for the Serial Verb parameter (that is, having sentences like Ozo pushed the pot fell). Only languages that have no Verb Attraction can set the Serial Verb parameter set on "positive." Baker ultimately ventures a "flow chart" of parameter settings that any language in the world can be situated upon. The genius of the chart is in the fact that there are only a dozen or so parameters, these simple binary alternations determining the seemingly vast differences between languages like English, Japanese, Chinese, and Mohawk which, on the surface, give the appearance of having arisen in different universes. The task of identifying these parameters and investigating the interactions between them occupies many professional linguists, and Baker's book is a valuable exposition of this project for the interested layman.

Yet in the end, The Atoms of Language does not quite hit home for a general audience to the degree I wish it did. Part of the problem lies beyond Baker himself. Pinker's book is driven by the conviction that "the language instinct" is a product of natural selection, a point he has argued widely since in general public venues. Baker, however, articulately argues against the idea that parametrical alternations, in particular, could have been somehow beneficial to the propagation of Homo sapiens sapiens, and in this he is, in my view, quite correct.

Then many linguists would attribute the differences between languages to cultural traits: here, presuppositions traceable to linguistics' deep roots in anthropology are reinforced by modern academia's enshrinement of cultural relativism. Yet Baker deftly dismisses this paradigm as well. Obviously certain aspects of languages reflect cultures—Japanese has special verbs and verb endings according to the social rank of the person spoken to or about, and this clearly responds to the rigidly hierarchical structure of (traditional) Japanese society. But in any given language's structure, things like this are but a sliver: there is nothing common to the cultures of Japan, India, Ethiopia, Germany, and New Guinea that would somehow leave verbs at the end of sentences rather than in the middle. Languages with the Serial Verb parameter are found in West Africa, China, Cambodia, and the Caribbean, among other areas; yet we search in vain for a cultural trait that would explain why preliterate villagers in southern Nigeria have this parameter choice while culturally similar preliterate villagers in northern Nigeria do not; and so on.

But having refuted these two arguments, Baker is left with no explanation as to just why human language appears to be built on parametrical alternations. He ventures that the answer may only come to us via dramatic paradigm shifts in linguists' approach to their subject, as yet inconceivable. The intellectual honesty in this is laudable, but it must be admitted that it deprives the book of a certain payoff. By the end, the general reader cannot help but seek a reason to care about the fact that, well, languages he has never heard of are less different from one another than they seem. Baker makes some dutiful genuflections to the idea that this realization will illuminate how the world's peoples can build bridges of understanding, but let's face it—one neither writes nor reads 200-plus pages about the likes of the Verb Attraction parameter and Washakotya'tawitsherahetkvhta'se' with this foremost in one's mind.

Baker also suggests that awareness of the diversity among the world's languages heightens one's appreciation of the diversity of human cultures and their perspectives upon life and the world. But this would seem incommensurate with the unifying thesis of the book, especially given his previous refutation of just such reflexive culture-based arguments. The truth is that syntacticians study parameters because it's neat, period. Of course if that alone could stand as the primum mobile of books for the general public, many more academics would go to the trouble of writing them. Baker cannot help trying to hang the book on something larger, but the subject does not make this easy, and thus the problem remains.

A more serious problem with the book is that Baker is not consistently successful in ushering the lay reader into concepts that can be initially rather obscure. It is notoriously tricky to impart modern linguistics' conceptions of grammar to the uninitiated. While hardly as inherently intricate and counterintuitive as, say, the tenets of astrophysics, these grammatical categories and assumptions are much less transparent to the novice than is often apparent to the academic linguist who has spent decades living and breathing them. At too many places in the text, Baker casually uses terms like "embedded clause," "ergative language," and "antipassive" as if he were addressing a graduate seminar in syntax, having not yet explained them or, in some cases, never explaining them at all. Terms like this are Hittite to the novice, and they shouldn't have passed unchallenged by an editor. There is a glossary, but some of the terms are not in it, many readers will not consult it, and many of those who do will lose the thread of the argument being made.

This is especially true with concepts where a careful explanation within the text is the only really useful approach. For example, only someone who has already had at least one university course on linguistics will benefit from learning that ergative means "a case marker that attaches to the subject noun phrase only in transitive sentences." Thus it is difficult to see just what audience this book is best suited for. It is pitched too low for linguistics students, but—happy as I was that Baker, a linguist of imposing reputation, had written such a book for a general audience—I could not help feeling that many lay readers would feel overwhelmed around the middle of the third chapter.

This is not to say that the book is a thicket of opaque jargon. On the contrary, Baker makes a sincere effort to communicate, and the reader will benefit if aware that the spoon-feeding will have a limit. The Atoms of Language clues the interested reader into what many linguists are up to, given that we are neither interpreters nor policemen of "good grammar." The modern syntactician is embarked on a quest to find order in chaos, and Baker's book is a useful progress report on the endeavor.

John H. McWhorter is associate professor of linguistics at the University of California at Berkeley and the author most recently of The Power of Babel: A Natural History of Language, which will be published in January by Times Books.

Copyright © 2001 by the author or Christianity Today/Books & Culture magazine.
Click here for reprint information on Books & Culture.