Language Identity

Language Identity

Given plausible empirical assumptions, we can argue on conceptual grounds for two counterintuitive theses: (1) there is only one human language, and (2) this language is not learnable. It will turn out that these statements are not as counterintuitive as they seem. In fact, they follow from well-known considerations advanced by Noam Chomsky, and widely received.[1] I will not argue for these positions, though I will articulate them in ways that may not be familiar. The first claim is to be understood on the model of theoretical identities in science, like “heat is molecular motion” or “Hesperus is Phosphorus”. Thus, we have “English is identical to Spanish”, despite appearances. There is the language, on the one hand, and its expression in speech on the other: the former is an abstract structure realized in the brain; the latter is the sensorimotor externalization of that abstract mental structure. The structure is identical from speaker to speaker—it is a human universal—while the external expression varies. There could be a language distinct from this common human language, perhaps spoken by Martians (or whales), but in fact our human language is a fixed commodity. By common consent, what distinguishes it are grammatical peculiarities: recursive rules, hierarchical structure, digital discreteness, finite uniform lexicon, infinite potential, and perhaps other features. What this structure sounds like when coupled with an articulatory system is beside the point. You can see the logic here by recalling that even in what we call a single language there is wide variation in pronunciation and vocabulary—as in different dialects. Mutual comprehension is not a necessary condition of language identity. What is called “English” is itself a combination of other languages (Greek, Latin, German, Anglo Saxon); the classification is conventional and pragmatic. It isn’t a scientific natural kind. What if I were to decide to alter my native English by swapping words around—using “hot” to mean cold and “cold” to mean hot, and likewise across the language? Or pronouncing English words in a strange way that no one else can understand. I would still be speaking English, intuitively, though sounding completely foreign. And what of bilingual speakers who speak in a mixture of English and Spanish—what language are they speaking? You can say what you like, but what is clear is that they are speaking human—not Martian (or whale). The commonality would be much more salient to us if we were surrounded by aliens speaking radically different types of language (non-recursive, linear, analogue, finite, etc.) Compared to that we are all speaking the same language, just using different dialects of that shared language. The same could be said of the concept of race: we humans are all the same race as compared to Martians and Venusians or Neanderthals. Language is like anatomy: there is a universal shared human anatomy, despite some superficial variations (compare water). In the same way we all share an identical conceptual scheme (despite claims to the contrary). Human language is basically innate, and what is innate to the species is a universal. We all have the same language instinct; there is no English language instinct different from the Spanish language instinct. We all have the same basic body plan, genetically fixed, and we all have the same language plan, genetically fixed. That is the Chomskian position, and it generates the thesis that all human languages are identical—English is Spanish, deep down. That language can be articulated differently, depending on where you live, but it is one and the same from instance to instance. English can be spoken in many ways (not all of them acoustic), and written in many ways, and among these ways is what we call Spanish, paradoxical as that may sound; just as English speakers are speaking Spanish when speaking English. For the languages in question are literally identical. You can learn different dialects of this language, as when an English speaker learns the Spanish dialect (or a southerner in England learns the Geordie dialect)—that is, you employ different sounds when you speak. You learn a different articulation of the same language—because that is what the science of linguistics has discovered (as we are assuming). We have an a posteriori identity claim, backed up in the usual way. We don’t now think that ice and steam are not water because they are not liquid, and likewise we shouldn’t think that Spanish isn’t English because it sounds different from a common form of English (that spoken in the British Isles). Water is H2O whatever form it may take, and English is a certain sort of cognitive structure in whatever outer form it may assume. Language proper is in the head and humans have the same type of head (viewed scientifically). The rest is practical and conventional—like not saying “I spilled water on the floor” when you upset the ice tray. If the neuroscientists investigate the brains of different types of speakers and find a common neural structure, corresponding to linguistic competence, they will have confirmed the Chomskian hypothesis—that language mastery is a human universal rooted in the brain. The philosophical point I have wanted to make is that this is an instance of scientific identity claims—“heat is molecular motion”, “pain is C-fiber firing”, and the like. We have discovered a posteriori that English is identical to Spanish (French, Russian, Swahili, etc.). In the jargon, these terms are rigid designators of identical referents that generate true a posteriori identity statements–necessarily true statements to boot. Not only is English identical to Spanish; it is necessarily identical to Spanish. In no possible world is English not identical to Spanish (though the claim is epistemically contingent). It might indeed turn out that “Spanish” speakers do not speak the same language as English speakers—for it may be that their brains house a radically non-human language structure—but we have good reason to believe that Spanish-speaking people have normal human brains. I like to think this identity thesis might help international relations—fundamentally, we all speak the same language. The identity thesis is good politics as well as sound philosophy of language. Compare meeting a broad Geordie for the first time and thinking he must be speaking a foreign language, perhaps a variant of Old German, and then discovering he is speaking English after all—wouldn’t that be a nice thing to know? Different sounds are coming out of his mouth, but inside his head the same machinery is chugging away. Ditto for the Spanish speaker. We just have to let go of the prejudice that people who don’t sound like us speak a different language. We can already see this on the assumption that there is a universal language of thought, but now we see that even the spoken language is universal—just not the way it is spoken. People with different idiolects can speak the same dialect, people who speak different dialects can speak the same language, and people who speak (what look like) different languages can speak the same underlying language. All humans speak Human; and that is the only languagethey speak—the rest is variations on that language. This is what the science is telling us.

What about the second thesis—that human language is not learnable? How can that be true, you may ask, given that people have demonstrably learned to speak English etc. But be careful: it doesn’t follow from the fact that someone knows something that he learned it. You do not learn what is already in you. What is innate is not learned, so the innate human language is not learned—only its sensorimotor expression is. The Geordie accent is learned but not the language onto which that accent is directed. Speakers of natural human language don’t learn that language, only dialects of it. But that doesn’t answer the question at issue—whether that language can be learned. And surely it can be, in this sense—a suitably intelligent person could learn it without already possessing it innately. A Martian, say, could learn it as a second language: first he knows Martian innately; then he learns the different language Human as an additional language. This can’t be too difficult for a sophisticated intelligence not already innately equipped with Human (Mr. Spock would find the task a piece of cake). However, that doesn’t answer a slightly different question: is a fundamentally different language learnable by a normal human child? And the answer to that question must surely be no, because the human child can only learn a particular human language by exploiting his or her innate linguistic capacity, so won’t have the equipment to learn a different type of language, say Martian. By the same token, a Martian child will not be able to learn Human. So, we can say that human language is not learned by a human child and is not learnable by a Martian child—though it is probably learnable by a smart Martian adult. Maybe this latter mastery will be stilted and unnatural, unlike normal human mastery, so it won’t be learnable in the same way a human learns it. In other words, it is really not true that the language of humans is learnable: it is not learned by human children, and it cannot be learned by other children with a different innate endowment. It is either possessed by being innate, as in the human case, or it is not possessable at all, as in the Martian case (or only partially learned by the intelligent adult Martian). In short, languages proper are not learnable at all, though they can obviously be possessed. The reason they are not learned is that they are innate, and their being innate is connected to their being singular: there is one human language and it is innate, and therefore not learned. Thus, contrary to tradition, there are not multiple human languages that are learned by people; there is one human language and it is not learned. The error comes from identifying language with speech, competence with performance. That is, it comes from assuming a behaviorist view of knowledge of language: linguistic behavior is both variable and learned—not so linguistic competence. It is constant and unlearned, universal and a priori (in one sense of that term). Everyone speaks the same language, and no one learned that language. The language faculty is a species-wide instinct that admits of no learning. It is necessarily shared (by humans) and necessarily not acquired—like human anatomy and instincts in general.[2]

[1] These are well-worn themes of Chomsky’s and hardly need restatement. For recent discussion, see his What Kind of Creatures Are We? (2016).

[2] The necessity here is nomological not metaphysical—in some possible worlds, humans might not speak Human, because of brain re-wiring. The point is that the language instinct is like other instincts: there is just one sex instinct, say, not multiple sex instincts depending on your country of origin; and no one learns to have a sex instinct (by definition). The instinct can be variably expressed, depending on the surrounding culture, but it would be misplaced behaviorism to suppose that the instinct itself were thus plastic. It is loose talk to speak of “human languages” (plural); similarly, for talk of “language learning”. Ordinary language is out of step with scientific language here—not for the first time. Commonsense linguistics must give way to theoretical linguistics.