The Language Instinct: How the Mind Creates Language (48 page)

We know that this experience must include, at a minimum, the speech of other human beings. For several thousand years thinkers have speculated about what would happen to infants deprived of speech input. In the seventh century
B.C.
, according to the historian Herodotus, King Psamtik I of Egypt had two infants separated from their mothers at birth and raised in silence in a shepherd’s hut. The king’s curiosity about the original language of the world allegedly was satisfied two years later when the shepherd heard the infants use a word in Phrygian, an Indo-European language of Asia Minor. In the centuries since, there have been many stories about abandoned children who have grown up in the wild, from Romulus and Remus, the eventual founders of Rome, to Mowgli in Kipling’s
The Jungle Book
. There have also been occasional real-life cases, like Victor, the Wild Boy of Aveyron (the subject of a lovely film by François Truffaut), and, in the twentieth century, Kamala, Amala, and Ramu from India. Legend has these children raised by bears or wolves, depending on which one has the greater affinity to humans in the prevailing mythology of the region, and this scenario is repeated as fact in many textbooks, but I am skeptical. (In a Darwinian animal kingdom it would be a spectacularly stupid bear that when faced with the good fortune of a baby in its lair would rear it rather than eat it. Though some species can be fooled by foster offspring, like birds by cuckoos, bears and wolves are predators of young mammals and are unlikely to be so gullible.) Occasionally other modern children have grown up wild because depraved parents have raised them silently in dark rooms and attics. The outcome is always the same: the children are mute, and often remain so. Whatever innate grammatical abilities there are, they are too schematic to generate speech, words, and grammatical constructions on their own.

The muteness of wild children in one sense emphasizes the role of nurture over nature in language development, but I think we gain more insight by thinking around that tired dichotomy. If Victor or Kamala had run out of the woods speaking fluent Phrygian or Proto-World, who could they have talked to? As I suggested in the preceding chapter, even if the genes themselves specify the basic design of language, they might have to store the specifics of language in the environment, to ensure that a person’s language is synchronized with everyone else’s despite the genetic uniqueness of every individual. In this sense, language is like another quintessentially social activity. James Thurber and E. B. White once wrote:

Though speech input is necessary for speech development, a mere soundtrack is not sufficient. Deaf parents of hearing children were once advised to have the children watch a lot of television. In no case did the children learn English. Without already knowing the language, it is difficult for a child to figure out what the characters in those odd, unresponsive televised worlds are talking about. Live human speakers tend to talk about the here and now in the presence of children; the child can be more of a mind-reader, guessing what the speaker might mean, especially if the child already knows many content words. Indeed, if you are given a translation of the content words in parents’ speech to children in some language whose grammar you do not know, it is quite easy to infer what the parents meant. If children can infer parents’ meanings, they do not have to be pure cryptographers, trying to crack a code from the statistical structure of the transmissions. They can be a bit more like the archeologists with the Rosetta Stone, who had both a passage from an unknown language and its translation in a known one. For the child, the unknown language is English (or Japanese or Inslekampx or Arabic); the known one is mentalese.

Another reason why television soundtracks might be insufficient is that they are not in Motherese. Compared with conversations among adults, parents’ speech to children is slower, more exaggerated in pitch, more directed to the here and now, and more grammatical (it is literally 99 and 44/100ths percent pure, according to one estimate). Surely this makes Motherese easier to learn than the kind of elliptical, fragmentary conversation we saw in the Watergate transcripts. But as we discovered in Chapter 2, Motherese is not an indispensable curriculum of Language-Made-Simple lessons. In some cultures, parents do not talk to their children until the children are capable of keeping up their end of the conversation (though other children might talk to them). Furthermore, Motherese is
not
grammatically simple. That impression is an illusion; grammar is so instinctive that we do not appreciate which constructions are complex until we try to work out the rules behind them. Motherese is riddled with questions containing
who, what
, and
where
, which are among the most complicated constructions in English. For example, to assemble the “simple” question
What did he eat?
, based on
He ate what
, one must move the
what
to the beginning of the sentence, leaving a “trace” that indicates its semantic role of “thing eaten,” insert the meaningless auxiliary
do
, make sure that the
do
is in the tense appropriate to the verb, in this case
did
, convert the verb to the infinitive form
eat
, and invert the position of subject and auxiliary from the normal
He did
to the interrogative
Did he
. No mercifully designed language curriculum would use these sentences in Lesson 1, but that is just what mothers do when speaking to their babies.

A better way to think of Motherese is to liken it to the vocalizations that other animals direct to their young. Motherese has interpretable melodies: a rise-and-fall contour for approving, a set of sharp, staccato bursts for prohibiting, a rise pattern for directing attention, and smooth, low legato murmurs for comforting. The psychologist Anne Fernald has shown that these patterns are very widespread across language communities, and may be universal. The melodies attract the child’s attention, mark the sounds as speech as opposed to stomach growlings or other noises, distinguish statements, questions, and imperatives, delineate major sentence boundaries, and highlight new words. When given a choice, babies prefer to listen to Motherese than to speech intended for adults.

Surprisingly, though practice is important in training for the gymnastics of speaking, it may be superfluous in learning grammar. For various neurological reasons children are sometimes unable to articulate, but parents report that their comprehension is excellent. Karin Stromswold recently tested one such four-year-old. Though he could not speak, he could understand subtle grammatical differences. He could identify which picture showed “The dog was bitten by the cat” and which showed “The cat was bitten by the dog.” He could distinguish pictures that showed “The dogs chase the rabbit” and “The dog chases the rabbit.” The boy also responded appropriately when Stromswold asked him, “Show me your room,” “Show me your sister’s room,” “Show me your sister’s old room,” “Show me your old room,” “Show me your new room,” “Show me your sister’s new room.”

In fact, it is not surprising that grammar development does not depend on overt practice, because actually saying something aloud, as opposed to listening to what other people say, does not provide the child with information about the language he or she is trying to learn. The only conceivable information about grammar that speaking could provide would come from feedback from parents on whether the child’s utterance was grammatical and meaningful. If a parent punished, corrected, misunderstood, or even reacted differently to a child’s ungrammatical sentence, it could in theory inform the child that something in his growing rule system needed to be improved. But parents are remarkably unconcerned about their children’s grammar; they care about truthfulness and good behavior. Roger Brown divided the sentences of Adam, Eve, and Sarah into grammatical and ungrammatical lists. For each sentence he checked whether the parent had at the time expressed approval (like “Yes, that’s good”) or disapproval. The proportion was the same for grammatical sentences and ungrammatical ones, which means that the parent’s response had given the child no information about grammar. For example:

Brown also checked whether children might learn about the state of their grammars by noticing whether they are being understood. He looked at children’s well-formed and badly formed questions and whether their parents seemed to have answered them appropriately (that is, as if they understood them) or with non sequiturs. Again, there was no correlation;
What you can do?
may not be English, but it is perfectly understandable.

Indeed, when fussy parents or meddling experimenters do provide children with feedback, the children tune it out. The psycholinguist Martin Braine once tried for several weeks to stamp out one of his daughter’s grammatical errors. Here is the result:

Braine wrote, “Further tuition is ruled out by her protest, vigorously supported by my wife.”

As far as grammar learning goes, the child must be a naturalist, passively observing the speech of others, rather than an experimentalist, manipulating stimuli and recording the results. The implications are profound. Languages are infinite, childhoods finite. To become speakers, children cannot just memorize; they must leap into the linguistic unknown and generalize to an infinite world of as-yet-unspoken sentences. But there are untold numbers of seductive false leaps: