to home

The Chinese Character for Big Bird
The Language Instinct
By Steven Pinker

A Review by Alex Gross

Would-be language experts like to talk about Chinese, whether they know any or not. Steven Pinker, the latest Chomskian apologist, is no exception. His recent best-seller The Language Instinct contains five assertions about the Chinese language and its speakers. Unfortunately, all of them are incorrect.

Chinese is a language that few Westerners ever learn well enough to make meaningful generalizations about. We all know that four years of college French or German make no one an expert in either language, much less a translator. But three or four years of college Chinese barely bring students to the level of being able to use a Chinese-English dictionary. It takes at least another two years before one can approach the riches of a Chinese-Chinese dictionary, the equivalent of a Larousse or a Brockhaus (which even beginning students of French or German can consult).

Chinese characters are of course the reason—there is nothing comparable in Western languages. It is therefore no surprise that Pinker mistakenly concludes that this "writing system has served the Chinese well, despite the inherent disadvantage that readers are at a loss when they face a new or rare word." (1)

This statement is entirely incorrect. No "disadvantage" is necessarily present, it is certainly not "inherent," and readers of Chinese are on the whole far less at a loss with new or rare words than readers of alphabet-based languages. Take the following English sentence as an example:

Cassowaries were a problem that year.

This sentence can make no sense at all to readers unless they know exactly what a "cassowary" is (or have picked up the Malay word it comes from). I do not know the Chinese word for "cassowary," nor do I need to know it. But I do know that I would immediately have a far better understanding of this sentence in Chinese than in English. That is because the Chinese word for cassowary, whatever it may be, is certain to contain either the radical niao for "long-tailed bird" or the radical zhui for "short-tailed bird." Which is what cassowary means: a big ostrich-like bird, whether its tail is visualized as long or short. Many comparable examples could be found and are typical of Chinese.

Elsewhere Pinker relies on the faulty assumption that Chinese is "an isolating language" simply so he can celebrate his discovery that it can also "create multipart words such as compounds and derivatives." (2) Here again he shows his ignorance, for despite the quaint prejudice that Chinese is monosyllabic, in actual fact it does little other than create multi-syllabic compounds.

He also falls into the trap that Chinese essentially encodes "morphemes" (units of meaning) in its "logograms" (characters) (3). But in point of fact Chinese often also encodes monosyllabic sounds, though not as efficiently as its students might wish. This dual approach—attempting to convey both meaning and sound—leads to the paradox familiar to all advanced students and even some native speakers: in Chinese it is possible to know how to pronounce a character without understanding what it means. Alternately, one may know what a character means without having a clue about its pronunciation. Finally, one may have some idea of a character's overall domain without knowing either its sound or its precise meaning. This is a truly unique experience for most linguists. Such are the lesser mountain ranges one must surmount before one can remotely approach the summit of translation.

Pinker makes further errors in explaining the function of Chinese tones (4) and in asserting that "Chinese college students tend to have more scientific training than American students." (5) This last claim opens such a mammoth can of cross-cultural worms that it is best passed over at present. But readers are entitled to wonder, given the author's poor score on Chinese, how much truth there can be in the rest of his book or in the Chomskian doctrines defended here and in a recent three-hour PBS series. Perhaps this can be dealt with in a further article.


(1) Steven Pinker: The Language Instinct, Harper Perennial, 1994, p 191.
(2) Ibid., p. 239.
(3) Ibid., p. 189.
(4) Ibid., p. 164.
(5) Ibid., p. 67.

Replies to two critics of the two pieces on Chomskian theories:

This following text contains two replies to criticisms of the pieces on Chomskian linguistics. Those familiar with this field may be aware that attacks on these theories can meet with fierce resistance from their advocates, sometimes bordering on religious intolerance. Due to copyright problems, the criticisms themselves are not reproduced here, but their tenor and contents should be to some extent clear from the nature of the replies.

Reply to first critique:

To the Editor:

Through either oversight or misplaced self-censorship, the title of my piece on Chomskian linguistics was omitted. It was called "The Emperor's New Linguistics," and its deletion turned my description of a naked gentleman into near nonsense.

Jon Johanning's arguments are easily dealt with. If he is so deeply impressed with this "MIT linguistic scientist," why is he not equally impressed with the theories of W.V.O. Quine, another MIT savant whose views significantly differ from Chomsky's? He is also premature in rejoicing that he found no misstatements about Japanese in Pinker's book— one need scarcely be an expert in this language to realize that the author's statements were far too superficial for any such mistakes to occur.

Through the centuries, pseudosciences like phrenology, ether-based physics, physiognomy, and even eugenics have enjoyed enormous popularity even among the learned. Freud himself believed in biorhythms, and cold fusion still has its proponents. People love to hear explanations, even when they do not understand them, and even when the explanations explain very little. If Johanning believes Chomskianism does not belong to these other failed sciences, his argument does not lie with me alone. Much of George Steiner's After Babel is a refutation of these theories, and I.A. Richards was also deeply skeptical. Or let Johanning pick up and read any ten pages at random of Randy Allen Harris' book The Linguistics Wars, and he will quickly see how arbitrary, pretentious, and obscurantist this whole school of linguistics can truly be.

As for his statement that "this subject is of no interest whatsoever to translators or interpreters," he speaks for himself. Chomskian doctrines, which emphasize the triviality of linguistic differences, do a real disservice to translators and interpreters and are, furthermore, demonstrably false or in some cases themselves "trivial." Through the millennia, observers from Cicero and Roger Bacon to Tytler and Unamuno have joined in asserting that many linguistic phenomena do not fit into any system of "universal grammar."

Where Pinker's book appears glancingly persuasive, it is most often relying on earlier and better writing in the field, which Johanning does not appear to have read. Perhaps its gravest flaw lies in its flimsy last chapter, where the author ought to have been providing his deepest insights. Here he bases his summation on a quote by a fellow Chomskian (these experts usually talk only to each other), words he quite clearly regards as of major importance:

 "The thing is: I hate relativism. I hate relativism more
   than I hate anything else..."

Perhaps some one should gently remind these great "linguistic scientists" that there is one thing worse than "relativism." It is called "absolutism," an area they do seem to know a great deal about.

Alex Gross

The above letter was published in the January, 1996 ATA Chronicle

Reply to second critique:

This reply to two earlier critics centered on the Chinese piece alone, though some broader implications are also present. Both critics gratuitously assumed that the Chinese piece stood alone and betokened a limited acquaintance with Chomskian theories, when in fact it was only the first prong of the far more general attack contained in "The Emperor's New Linguistics." It was published in the August, 1995 Chronicle along with the more general piece and is reproduced here with permission from the Editors.

To the Editor:

If these two correspondents have now seen my most recent Chronicle piece, they should be aware that my skepticism about Chomskian linguistics is neither shallow nor sudden. I discussed this matter even more thoroughly in my contribution to "Computers in Translation: A Critical Approach" (Routledge, 1992). I believe that most translators would agree with me that these theories are incorrect, incomplete, and incompatible with any true understanding of language. In his informal survey, my colleague Per Dohler found that reactions of translators to linguistics in general range "between indifference and outright scorn" and that linguistics belongs to the "pseudosciences" or "fields that don't yet have their jargon under control." Certainly the Chomskians have contributed to such a perception.

My article on Chinese contained 600 words and addressed five points. Yet one correspondent has employed 700 words to address a single one of these points: whether or not Chinese is an "isolating language." He also claims Professor Chao Yuen Ren as an ally of Chomskianism, but if he had more carefully examined page 3 of Prof. Chao's book—which I first encountered fourteen years ago—he would have found these words on that very same page: "no transformational or generative grammar has as yet been fully worked out for any language." That observation, printed in 1968, is still true today, and Professor Chao was quite correct in analyzing Chinese not in terms of Chomskian models but on the basis of "immediate constituents," or IC's, something my own professor has called "subsyntactic particles."

This correspondent makes a crucial error in not distinguishing between spoken and written Chinese, as only the latter has a ghost of a chance of being called truly "isolating," though it too will fail such a test. Such a claim fails on mathematical grounds alone. Mandarin Chinese possesses a mere 400 spoken syllables. Multiply this by the four tones, and—allowing for the many missing sound-tone slots—you will still end up with fewer than 1200 distinct utterances. These are simply not enough to constitute a viable language devoid of ambiguity and do not even satisfy the minimum requirements of Shannon's Information Theory for differentiating message elements. While 1200 sounds might have served for stylized poetry or imperial proclamations in times past, they could not alone suffice to name the entities, processes, and descriptors for even a "limited" culture (and there are no "limited" cultures!).

Such would have been the fate of Chinese if it had become a truly "isolating language." The only possible escape from this predicament is to join two or more characters and/or sound-tones together to form compounds, in which case you no longer have an "isolating language." Unless someone wants to pull in Cantonese, Fujianese, or some other form of Chinese, I sincerely hope this discussion is now ended.

Alex Gross

This article is Copyright © 1995
by Alexander Gross. It may be
reproduced for individuals and for
educational purposes only. It may
not be used for any commercial (i.e.,
money-making) purpose without
written permission from the author.

to top
to linguistics menu
to home