The Language Instinct: How the Mind Creates Language
English, like free-word-order languages, has free ordering in strings of prepositional phrases, where each preposition marks the semantic role of its noun phrase as if it were a case marker: The package was sent from Chicago to Boston by Mary; The package was sent by Mary to Boston from Chicago; The package was sent to Boston from Chicago by Mary, and so on. Conversely, in so-called scrambling languages at the other extreme, like Warlpiri, word order is never completely free; auxiliaries, for example, must go in the second position in a sentence, which is rather like their positioning in English.
English, like ergative languages, marks a similarity between the objects of transitive verbs and the subjects of intransitive verbs. Just compare John broke the glass (glass = object) with The glass broke (glass = subject of intransitive), or Three men arrived with There arrived three men.
English, like topic-prominent languages, has a topic constituent in constructions like As for fish, I eat salmon and John I never really liked.
Like SOV languages, not too long ago English availed itself of an SOV order, which is still interpretable in archaic expressions like Till death do us part and With this ring I thee wed.
Like classifier languages, English insists upon classifiers for many nouns: you can’t refer to a single square as a paper but must say a sheet of paper. Similarly, English speakers say a piece of fruit (which refers to an apple, not a piece of an apple), a blade of grass, a stick of wood, fifty head of cattle, and so on.
If a Martian scientist concludes that humans speak a single language, that scientist might well wonder why Earthspeak has those thousands of mutually unintelligible dialects (assuming that the Martian has not read Genesis 11; perhaps Mars is beyond the reach of the Gideon Society). If the basic plan of language is innate and fixed across the species, why not the whole banana? Why the head-first parameter, the different-sized color vocabularies, the Boston accent?
Terrestrial scientists have no conclusive answer. The theoretical physicist Freeman Dyson proposed that linguistic diversity is here for a reason: “it was nature’s way to make it possible for us to evolve rapidly,” by creating isolated ethnic groups in which undiluted biological and cultural evolution can proceed swiftly. But Dyson’s evolutionary reasoning is defective. Lacking foresight, lineages try to be the best that they can be, now; they do not initiate change for change’s sake on the chance that one of the changes might come in handy in some ice age ten thousand years in the future. Dyson is not the first to ascribe a purpose to linguistic diversity. A Colombian Bará Indian, a member of an outbreeding set of tribes, when asked by a linguist why there were so many languages, explained, “If we were all Tukano speakers, where would we get our women?”
As a native of Quebec, I can testify that differences in language lead to differences in ethnic identification, with widespread effects, good and bad. But the suggestions of Dyson and the Bará put the causal arrow backwards. Surely head-first parameters and all the rest represent massive overkill in some design to distinguish among ethnic groups, assuming that that was even evolutionary desirable. Humans are ingenious at sniffing out minor differences to figure out whom they should despise. All it takes is that European-Americans have light skin and African-Americans have dark skin, that Hindus make a point of not eating beef and Moslems make a point of not eating pork, or, in the Dr. Seuss story, that the Star-Bellied Sneetches have bellies with stars and the Plain-Bellied Sneetches have none upon thars. Once there is more than one language, ethnocentrism can do the rest; we need to understand why there is more than one language.
Darwin himself expressed the key insight:
The formation of different languages and of distinct species, and the proofs that both have been developed through a gradual process, are curiously parallel…. We find in distinct languages striking homologies due to community of descent, and analogies due to a similar process of formation…. Languages, like organic beings, can be classed in groups under groups; and they can be classed either naturally, according to descent, or artificially by other characters. Dominant languages and dialects spread widely, and lead to the gradual extinction of other tongues. A language, like a species, when extinct, never…reappears.
That is, English is similar though not identical to German for the same reason that foxes are similar though not identical to wolves: English and German are modifications of a common ancestor language spoken in the past, and foxes and wolves are modifications of a common ancestor species that lived in the past. Indeed, Darwin claimed to have taken some of his ideas about biological evolution from the linguistics of his time, which we will encounter later in this chapter.
Differences among languages, like differences among species, are the effects of three processes acting over long spans of time. One process is variation—mutation, in the case of species; linguistic innovation, in the case of languages. The second is heredity, so that descendants resemble their progenitors in these variations—genetic inheritance, in the case of species; the ability to learn, in the case of languages. The third is isolation—by geography, breeding season, or reproductive anatomy, in the case of species; by migration or social barriers, in the case of languages. In both cases, isolated populations accumulate separate sets of variations and hence diverge over time. To understand why there is more than one language, then, we must understand the effects of innovation, learning, and migration.
Let me begin with the ability to learn, and by convincing you that there is something to explain. Many social scientists believe that learning is some pinnacle of evolution that humans have scaled from the lowlands of instinct, so that our ability to learn can be explained by our exalted braininess. But biology says otherwise. Learning is found in organisms as simple as bacteria, and, as James and Chomsky pointed out, human intelligence may depend on our having more innate instincts, not fewer. Learning is an option, like camouflage or horns, that nature gives organisms as needed—when some aspect of the organisms’ environmental niche is so unpredictable that anticipation of its contingencies cannot be wired in. For example, birds that nest on small cliff ledges do not learn to recognize their offspring. They do not need to, for any blob of the right size and shape in their nest is sure to be one. Birds that nest in large colonies, in contrast, are in danger of feeding some neighbor’s offspring that sneaks in, and they have evolved a mechanism that allows them to learn the particular nuances of their own babies.
Even when a trait starts off as a product of learning, it does not have to remain so. Evolutionary theory, supported by computer simulations, has shown that when an environment is stable, there is a selective pressure for learned abilities to become increasingly innate. That is because if an ability is innate, it can be deployed earlier in the lifespan of the creature, and there is less of a chance that an unlucky creature will miss out on the experiences that would have been necessary to teach it.
Why might it pay for the child to learn parts of a language rather than having the whole system hard-wired? For vocabulary, the benefits are fairly obvious: 60,000 words might be too many to evolve, store, and maintain in a genome comprising only 50,000 to 100,000 genes. And words for new plants, animals, tools, and especially people are needed throughout the lifespan. But what good is it to learn different grammars? No one knows, but here are some plausible hypotheses.
Perhaps some of the things about language that we have to learn are easily learned by simple mechanisms that antedated the evolution of grammar. For example, a simple kind of learning circuit might suffice to record which element comes before which other one, as long as the elements are first defined and identified by some other cognitive module. If a universal grammar module defines a head and a role-player, their relative ordering (head-first or head-last) could thus be recorded easily. If so, evolution, having made the basic computational units of language innate, may have seen no need to replace every bit of learned information with innate wiring. Computer simulations of evolution show that the pressure to replace learned neural connections with innate
ones diminishes as more and more of the network becomes innate, because it becomes less and less likely that learning will fail for the rest.
A second reason for language to be partly learned is that language inherently involves sharing a code with other people. An innate grammar is useless if you are the only one possessing it: it is a tango of one, the sound of one hand clapping. But the genomes of other people mutate and drift and recombine when they have children. Rather than selecting for a completely innate grammar, which would soon fall out of register with everyone else’s, evolution may have given children an ability to learn the variable parts of language as a way of synchronizing their grammars with that of the community.
The second component of language differentiation is a source of variation. Some person, somewhere, must begin to speak differently from the neighbors, and the innovation must spread and catch on like a contagious disease until it becomes epidemic, at which point children perpetuate it. Change can arise from many sources. Words are coined, borrowed from other languages, stretched in meaning, and forgotten. New jargon or speech styles may sound way cool within some subculture and then infiltrate the mainstream. Specific examples of these borrowings are a subject of fascination to pop language fanciers and fill many books and columns. Personally, I have trouble getting excited. Should we really be astounded to learn that English borrowed kimono from Japanese, banana from Spanish, moccasin from the American Indians, and so on?
Because of the language instinct, there is something much more fascinating about linguistic innovation: each link in the chain of language transmission is a human brain. That brain is equipped with a universal grammar and is always on the lookout for examples in ambient speech of various kinds of rules. Because speech can be sloppy and words and sentences ambiguous, people are occasionally apt to reanalyze the speech they hear—they interpret it as having come from a different dictionary entry or rule than the ones that the speaker actually used.
A simple example is the word orange. Originally it was norange, borrowed from the Spanish naranja. But at some point some unknown creative speaker must have reanalyzed a norange as an orange. Though the speaker’s and hearer’s analyses specify identical sounds for that particular phrase, anorange, once the hearer uses the rest of grammar creatively, the change becomes audible, as in those oranges rather than those noranges. (This particular change has been common in English. Shakespeare used nuncle as an affectionate name, a recutting of mine Uncle to my nuncle, and Ned came from Edward by a similar route. Nowadays many people talk about a whole nother thing, and I know of a child who eats ectarines and an adult called Nalice who refers to people she doesn’t care for as nidiots.)
Reanalysis, a product of the discrete combinatorial creativity of the language instinct, partly spoils the analogy between language change on the one hand and biological and cultural evolution on the other. Many linguistic innovations are not like random mutation, drift, erosion, or borrowing. They are more like legends or jokes that are embellished or improved or reworked with each retelling. That is why, although grammars change quickly through history, they do not degenerate, for reanalysis is an inexhaustible source of new complexity. Nor must they progressively differentiate, for grammars can hop among the grooves made available by the universal grammar in everyone’s mind. Moreover, one change in a language can cause an imbalance that can trigger a cascade of other changes elsewhere, like falling dominoes. Any part of language can change:
• Many phonological rules arose when hearers in some community reanalyzed rapid, coarticulated speech. Imagine a dialect that lacks the rule that converts t to a flapped d in utter. Its speakers generally pronounce the t as a t, but may not do so when speaking rapidly or affecting a casual “lazy” style. Hearers may then credit them with a flapping rule, and they (or their children) would then pronounce the t as a flap even in careful speech. Taken further, even the underlying phonemes can be reanalyzed. This is how we got v. Old English didn’t have a v; our word starve was originally steorfan. But any f between two vowels was pronounced with voicing turned on, so ofer was pronounced “over,” thanks to a rule similar to the contemporary flapping rule. Listeners eventually analyzed the v as a separate phoneme, rather than as a pronunciation of f, so now the word actually is over, and v and f are available as separate phonemes. For example, we can now differentiate words like waver and wafer, but King Ethelbuld could not have.
• The phonological rules governing the pronunciation of words can, in turn, be reanalyzed into morphological rules governing the construction of them. Germanic languages like Old English had an “umlaut” rule that changed a back vowel to a front vowel if the next syllable contained a high front vowel sound. For example, in foti, the plural of “foot,” the back o was altered by the rule to a front e, harmonizing with the front i. Subsequently the i at the end ceased being pronounced, and because the phonological rule no longer had anything to trigger it, speakers reinterpreted the o–e shift as a morphological relationship signaling the plural—resulting in our foot–feet, mouse–mice, goose–geese, tooth–teeth, and louse–lice.
• Reanalysis can also take two variants of one word, one created from the other by an inflectional rule, and recategorize them as separate words. The speakers of yesteryear might have noticed that an inflectional oo-ee rule applies not to all items but only to a few: tooth-teeth, but not booth-beeth. So teeth was interpreted as a separate, irregular word linked to tooth, rather than the product of a rule applied to tooth. The vowel change no longer acts like a rule—hence Lederer’s humorous story “Foxen in the Henhice.” Other sets of vaguely related words came into English by this route, like brother-brethren, half-halve, teeth-teethe, to fall-to fell, to rise-to raise; even wrought, which used to be the past tense of work.
• Other morphological rules can be formed when the words that commonly accompany some other word get eroded and then glued onto it. Tense markers may come from auxiliaries; for example, as I’ve mentioned, the English -ed suffix may have evolved from did: hammer-did hammered. Case markers may come from slurred prepositions or from sequences of verbs (for example, in a language that allows the construction take nail hit it, take might erode into an accusative case marker like ta-). Agreement markers can arise from pronouns: in John, he kissed her, he and her can eventually glom onto the verb as agreement affixes.
• Syntactic constructions can arise when a word order that is merely preferred becomes reanalyzed as obligatory. For example, when English had case markers, both give him a book and give a book him were possible, but the former was more common. When the case markers eroded in casual speech, many sentences would have become ambiguous if order were still allowed to vary. The more common order was thus enshrined as a rule of syntax. Other constructions can arise from multiple reanalyses. The English perfect I had written a book originally came from I had a book written (meaning “I owned a book that was written”). The reanalysis was inviting because the SOV pattern was alive in English; the participle written could be reanalyzed as the main verb of the sentence, and had could be reanalyzed as its auxiliary, begetting a new analysis with a related meaning.
The third ingredient for language splitting is separation among groups of speakers, so that successful innovations do not take over everywhere but accumulate separately in the different groups. Though people modify their language every generation, the extent of these changes is slight: vastly more sounds are preserved than mutated, more constructions analyzed properly than reanalyzed. Because of this overall conservatism, some patterns of vocabulary, sound, and grammar survive for millennia. They serve as the fossilized tracks of mass migrations in the remote past, clues to how human beings spread out over the earth to end up where we find them today.
How far back can we trace the language of this book, modern American English? Surprisingly far, perhaps five or even nine thousand years. Our knowledge of where our language has come from is considerably more precise than the recollection of Dave Barry’s Mr. Language Person: “The
English language is a rich verbal tapestry woven together from the tongues of the Greeks, the Latins, the Angles, the Klaxtons, the Celtics, and many more other ancient peoples, all of whom had severe drinking problems.” Let’s work our way back.
America and England first came to be divided by a common language, in Wilde’s memorable words, when colonists and immigrants isolated themselves from British speech by crossing the Atlantic Ocean. England was already a Babel of regional and class dialects when the first colonists left. What was to become the standard American dialect was seeded by the ambitious or dissatisfied members of lower and middle classes from southeastern England. By the eighteenth century an American accent was noted, and pronunciation in the American South was particularly influenced by the immigration of the Ulster Scots. Westward expansions preserved the layers of dialects of the eastern seaboard, though the farther west the pioneers went, the more their dialects mixed, especially in California, which required leapfrogging of the vast interior desert. Because of immigration, mobility, literacy, and now the mass media, the English of the United States, even with its rich regional differences, is homogeneous compared with the languages in territories of similar size in the rest of the world; the process has been called “Babel in reverse.” It is often said that the dialects of the Ozarks and Appalachia are a relict of Elizabethan English, but this is just a quaint myth, coming from the misconception of language as a cultural artifact. We think of the folk ballads, the hand-stitched quilts, and the whiskey aging slowly in oak casks and easily swallow the rumor that in this land that time forgot, the people still speak the traditional tongue lovingly handed down through the generations. But language does not work that way—at all times, in all communities, language changes, though the various parts of a language may change in different ways in different communities. Thus it is true that these dialects preserve some English forms that are rare elsewhere, such as afeared, yourn, hisn, and et, holp, and clome as the past of eat, help, and climb. But so does every variety of American English, including the standard one. Many so-called Americanisms were in fact carried over from England, where they were subsequently lost. For example, the participle gotten, the pronunciation of a in path and bath with a front-of-the-mouth “a” rather than the back-of-the-mouth “ah,” and the use of mad to mean “angry,” fall to mean “autumn,” and sick to mean “ill,” strike the British ear as all-American, but they are actually holdovers from the English that was spoken in the British Isles at the time of the American colonization.