Baby Talk

Nature versus nurture: the debate still goes on; but the way to resolve such questions is to do good research, not to argue from an armchair in an information vacuum. This is what Alison Gopnik, Andrew Meltzoff & Patricial Kuhl have achieved, and described in their 1999 book.

For centuries, it was assumed that “blood would tell”: a nobleman’s son would have nobility in his blood. Then, at the dawn of the Age of Reason, empiricism was born, and it was assumed that babies were born knowing nothing; the 17th century Latin phrase was tabula rasa, a blank writing tablet. See (for example) John Locke: Essay Concerning Human Understanding (1690). This is ironic, because although the term is apt for the content of the theory (that human beings are born mentally blank, and that each acquires knowledge and skills only through experiences in the world), it could not be less apt as it is generally used now, applied to the method by which the theory was reached.

Nowadays, the term “empiricism” is generally used as a synonym for the scientific method, by which all theories are based on experiment. But until the 20th century, nobody bothered to collect any data, or to do any experiments, on what babies know and are capable of when they are born: Locke’s “empiricism” theory about babies was not itself based on any empirical evidence!

Observation of babies begins

All this has now changed. In the 1920s Jean Piaget, in Geneva, observed and documented the development of his own three children in immense detail; officially Piaget was an expert on crustaceans, which kept him busy in his day job, so the work he and his wife did at home was unofficial, and on too small a scale to be conclusive. But their work inspired three American psychologists to begin in the 1970s a major program of observation and experimentation, which they have documented not only in the scientific journals but in a fascinating recent book. It is How Babies Think by Alison Gopnik, Andrew Meltzoff & Patricial Kuhl (Weidenfeld and Nicolson, 1999). (Note: this is the UK title; for reasons best known to the US publishers, the original title in the USA, was The Scientist in the Crib — which neatly takes us back to the scientific method, because the authors expound very cogently the thesis that every human baby is born a scientist, and constantly conducts experiments and refines theories about how the world works — which I have held to be true for a long time, though without the means to prove it.)

The researchers comment in their introduction that only once before the 20th century had any philosopher tried an experiment about what children know: in the dialogue Meno, Socrates quizzed a slave boy about geometry, and tried to prove that the boy knew something he had never been taught. And they tell of one English philosopher they spoke to who admitted that “while he had, of course, seen children about, he had never actually spoken to one”!

At this point you may well be screaming: “So how on earth did they go about these experiments on the poor babies? What do they do to the poor little mites?!” Well, one fascinating aspect of the book is the descriptions of the experiments; I cannot here go in for details, but you can be sure the babies are kept safe, and have a lot of fun.

The method relies heavily on the use of video recording equipment — which is why the work could not easily have been done much earlier than it was. Most experiments consist of many sessions in which the baby sits comfortably with its mother (often on her lap, for very small babies), being entertained, with the whole thing recorded on videotape for later analysis.

This is done for large numbers of babies, for many sessions over the pre-school years of each baby, so that over the decades a large body of observations has been built up.

Several main areas of study are discussed in the book, including what children learn about things, about people, and about language.

Babies extrapolate linear motion

An example in the first category involves the baby being entertained by watching a moving object. Even newborns will follow a moving object with their eyes. In a carefully designed game of “peep-bo”, the object moves steadily from one side to the other of the baby’s vision; the watching baby follows its movement with interest until it disappears behind a small screen. After a certain age at which it learns to extrapolate linear movement, the baby will predict the trajectory of the object and look expectantly at the other edge of the screen, expecting the object to re-appear (assuming constant velocity).

Babies are born able to distinguish all speech sounds

An example in the third category involved another modern technology: computer speech synthesis. It tests (and disproves) the theory that each baby is born knowing nothing about language, and learns to distinguish the sounds of its mother tongue by hearing those sounds around it.

If a recording of a word being repeated is played to a baby, the baby will perk up initially, then get bored and ignore the sound. However, the repeated word is then suddenly changed, it perks up again briefly before losing interest again. The experiment used the fact that some languages do not rely on differentiation of sound pairs, while others do. Specifically, English has the consonants L and R, but Japanese does not, so natives of Japan typically cannot tell the two apart. A speech synthesizer was given sampled sounds of the two words “lake” and “rake”, and a tape was prepared in which one word was repeated many times, but slowly changed to the other. Young babies perked up when the tape started, lost interest, then perked up again at some point around the middle of the transition from L to R in the sound. This happened for both American and Japanese babies less than about a year old.

The researchers tried the tape on adults for comparison, asking them to signal when the repeated word changed. American adults indicated a point around the middle of the transition. When the researcher travelled to Japan and tested adults there, Japanese adults never reported any perceptible change! This is the phenomenon that gave rise to “flied lice” jokes. But very young Japanese babies noticed the change from “lake” to “rake”!

The same test was done with other sound pairs and other languages; a variety of languages (including Hindi and Kikuyu) contain sound pairs indistinguishable to normal native English speakers, and yet very young American babies could distinguish them!

These experiments indicate that babies are born with the ability to distinguish all sound pairs, but that somewhere usually between 6 and 12 moths old babies begin to specialize in the sounds of their own language and lose the ability to make fine distinctions that adults of their own language community cannot make. Thus we have the opposite result from that previously assumed by the “armchair philosophers”.

A lot more could be said about how children learn language, but there is no more space here! Read the book, which I highly recommend.