Theodore Dalrymple has written affectingly about dysfunctional aspects of the culture of the underclass and the tendency of elites to indulge or romanticize them. But in seeing the world through that lens, he has missed the point of my book The Language Instinct. The book did not make the mad assertion that language can be learned without input from other people, as the title of the first chapter ("An Instinct to Acquire an Art") makes clear. Nor did it claim that all people are equally articulate or all dialects equally effective at conveying all ideas.

Its aim was to introduce readers to language as a topic of scientific understanding. Many phenomena look very different when they are judged by moral, political, or aesthetic criteria and when they are taken apart to see how they work. Dog breeders have strong opinions on the merits of breeding practices, but those opinions should not be confused with principles of mammalian genetics. Every linguist knows that the biggest roadblock to explaining language is that audiences are locked into the mindset of judging it rather than analyzing it, leading to ignorant beliefs such as that Black English is Standard English with errors, that some languages have no grammar, and that English is deteriorating because people split infinitives. The quotations that Dalrymple misinterprets as assertions of political correctness are just attempts to get readers to stop scorning other people's language long enough to learn something about it.

With his preference for dudgeon over analysis, Dalrymple simply refuses to think about language methodically. He considers it obvious that children learn language by imitating their parents. Yet if what parrots do is imitate, we need a very different word for what children do, and (as Dalrymple hopes his readers won't notice), the children in his vignettes acquire the language of their peers, not of their parents. He is quick to diagnose his uncle's inarticulateness (which he compares to that of a stroke victim) as a product of a lack of schooling in standard English, as if all people with that lack (say, Robert Burns or Abraham Lincoln) spoke like they had brain damage, and as if there were no need to rule out other explanations, such as genetic and developmental conditions. He reproduces verbatim a transcript of the speech of a teenage mother, failing to note that a verbatim transcript of anyone's speech is filled with self-interruptions, while being tone-deaf to the considerable grammatical complexity even in her speech.

Dalrymple writes, "Everyone, save the handicapped, learns to run without being taught; but no child runs 100 yards in nine seconds, or even 15 seconds, without training." Good analogy. That is why it would be fatuous to vilify a biologist studying the biomechanics of locomotion just because he wasn’t training people to run faster. Of course children should be taught to read, write, and express themselves clearly in the standard language of their community. But people with intellectual curiosity might want to learn something about the remarkable mental faculties that make these accomplishments possible in the first place.

Dr. Dalrymple does us all a service by highlighting two separate problems. The first lies within linguistics, and has to do with the dual function of language as both a means of communication and also as a vehicle for articulated thought. The second has to do with the way in which reasonably sound findings within any theoretical academic discipline may be abused in the broader realm of applied disciplines and beyond.

To address the first, there is a tension within linguistics about the extent to which grammar (meaning only syntax) can be understood without reference to meaning. Hardliners, and Pinker is closer to this end of the spectrum, point to the formal properties of syntactic structures, which appear to require no significant reference to content. This works to a point, but limits much of what one might plausibly acccount for. By introducing a distinction between competence (a presumed mastery of rules) and performance (the messy product of speaking or writing), Chomsky unwittingly armed such theorists against many attacks. Sloppy, inarticulate speech may sometimes be dismissed as reflecting performance issues, while theorists remain confident that given optimal circumstances, a linguistic utterance might have been produced which was consonant with a set of formal rules. Many other linguists disagree, and, to name but one, the approach of Cognitive Linguistics takes as its starting point a rejection of the separation of form and content. That said, most linguists would sign up to some such statement as "all languages/dialects studied to date exhibit comparable formal complexity and regularity." This seems relatively uncontentious. This is not to say anything whatsoever about the second, and largely distinct, role of language, that of organizing and articulating thought. Linguists typically ignore this aspect of language, while some philosophers, psychologists and sociologists are quite aware of it. Learning to think, to bring desires, emotions and attitudes, to a pertinent expression, is obviously a skill that can and should be honed. This appears to be what is lacking in the cases mentioned in the article. Linguists simply have nothing to say on the matter (to their shame).

Now to the second. Even if there were harmony among linguists, and all subscribed to some such statement, this can (and frequently has been) grossly misunderstood once it moves to the applied world and beyond. This is not surprising. The above statement refers to the formal properties of strings of symbols. Despite its simplicity, it is actually quite a technical statement. It does not assert that all forms of expression are equal, nor that all people are capable of articulating equally sophisticated thoughts. Regrettably, an inheritance of a classical approach to grammar, and a belief in the need for an absolute yardstick, led in the past to some appalling practices in teaching language. Pinker's colorful passage about Plato and the Swineheard, is an understandable counterreaction. It is a genuine discovery that there are no debased languages. But his good intentions are easy prey for the relativistic forces among us that would deny difference where possible.

Thank you, Dr. Dalrymple, for this wonderful article. Linguists will have much to chew on.

The main thrust of this article is the notion that, contrary to the assertions of most cognitive scientists, some dialects of English (and therefore some languages) are more expressive than others. The evidence is a series of anecdotes in which, according to the author, speakers of various working-class dialects were chronically unable to express certain thoughts.

The notion that there should be a direct link between the expressiveness and the phonology of a dialect (say, whether a person says "human" or "uman") or between expressiveness and syntax (whether a person can or cannot say "I might could go"; whether they say "different from" or "different to") is clearly quite implausible in the absence of any empirical evidence.

If dialect differences in pronunciation and grammar do not affect expressiveness, there remains only one possibility: that the words in a person's vocabulary might be in some way linked with their ability to express certain thoughts. A very cursory web search reveals a number of uncontroversial scholarly articles linking vocabulary size with socioeconomic status (and many proposed explanations). The crucial claim would then be that smaller vocabulary prohibits clear expression of many important thoughts; the working-class would then indeed be at a disadvantage, but NOT because they use grammatical features like negative concord ("There isn't none"), common across the world's languages but stigmatized in English by historical accident. The teaching of Standard English, whatever its merits, would, on its own, see no improvement in the supposed expressiveness problem. (It should also be clear that hesitations marked by "just," "maybe," "like," "you know," etc. are common to all dialects of human language.)

Personally, I do believe that every new word a human learns probably has some subtle shade of meaning which the old ones lack, and that in order to do science we need to develop new concepts and thus new vocabulary. Whether having a smaller vocabulary causes people to be "inarticulate" is another question; but we clearly need not turn to this extreme view to understand why, when a man is educated while his brother is not, only one of them can "discourse philosophically." Again, although vocabulary must eventually enter into any answer, there is certainly no way to invoke the differences in pronunciation and grammar which mark dialects. Vocabulary, however, would certainly not be the main factor; education would, leaving us with the unenlightening claim that education has an effect of some kind. Scientists would call this confusion of two issues (education and vocabulary) a "confound"; "missing the point" would be a more down-to-earth way to put it—admittedly less precise, but, I think, just as "articulate."

I admit that there are a few linguists who would take the
position Dalrymple attacks as such, but they are very few,
and I doubt they include Pinker.

The problem is that Dalrymple is confusing different issues. When linguists say that all languages are equal, we are saying simply that they are all rule-governed in the same basic way. It is more or less saying that all human arms (except those that are defective because of genes or accident) are the same. That does not mean that one is not stronger than another, or that one cannot be trained to be stronger, do finer work on certain tasks. Also it doesn't mean that appearance can't be different, the skin darker or lighter, hairier or more freckled. It just means that the structure is essentially the same. The point is that "he be fienin' " is not in any essential way more defective than "he constantly acts in an uncontrolled manner." It all depends on that bugaboo of conservative essentialists: context. For example, "He be fienin' " works better in a rap battle. "He constantly acts in an uncontrolled manner" works better in, say, a medical report.

That is why few thinking linguists would take the examples that their dealings with authority gives as contradicting our basic point on the equality of language. All they show is that in "their dealings with authority" as he says "they are at a huge disadvantage." Now, I suspect that there are many domains, perhaps a rap battle, bargaining in a souk, negotiating a high-level business deal, where I and perhaps Dalrymple will "be at a huge disadvantage" compared at least to some of his patients.

Yes, you do go against trendy egalitarian tree-hugging theories held by many idealistic and misguided educators, which are that we must not improve on inadequacy (for whatever reason) because that would imply that the system is highlighting such inadequacies. Let's pretend that everyone is the same and let's ignore the obvious disdavantages that an ineloquent person will endure. We think in words and phrases, don't we? Without an adequate language to experience these thoughts and then to communicate them to others, a person is severely handicapped. How right you are to quote Montaigne! I remember someone once pointing out to me that without the differences between people, they would all be the same. Yep, a truism if ever I heard one. Sounds like the basis for lyrics for a folk song. Something as pathetic as "Imagine."