The power of speech

The voice is one of our most powerful instruments, lying at the heart of the communication process. It belongs to both the body and the mind. It is shaped by our earliest infant experience and by powerful social conventions. It bridges our internal and external worlds, travelling from our most private recesses into the public domain, revealing not only our deepest sense of who we are, but also who we wish we weren't. It's a superb guide to fear and power, anxiety and subservience, to another person's vitality and authenticity as well as our own. You can't really know a person until you have heard them speak.

Although we are drenched in other people's cadences, most of the time we have no inkling of how they work on us or shape our comprehension. As a culture, until quite recently we have been barely voice-aware - a society that has blocked its ears. In western cultures' hierarchy of the senses, sound is often placed below sight in importance. We suffer from what Coleridge called "the despotism of the eye". Indeed, you could say that we often despise sound since we live in a culture where to love the sound of one's own voice is a term of abuse.

The voice is fundamentally just audible air. The process by which we breathe in order to live also, with minor changes, provides the energy for speech. Normally, when we exhale the air can't be heard because the vocal folds (two flaps of retractable muscle tissue) are pulled open, like a pair of curtains. When we speak, though, the vocal folds vibrate. After each opening to release the air, the elasticity of the vocal folds ensures that they return to their closed position, until the air pressure builds up once more and the cycle begins again. The sound of the human voice is produced by this rapid opening and closing of the vocal folds, hundreds of times per second, creating a succession of bursts of air that we hear as a buzz or hiss.

What we manage to do with this air is astonishing. With the exception of the muscles around the eyes, those of the human larynx have more nerves than any other muscles in the human body, including the hands and the face. Each can produce a different balance of forces in the larynx, generating a different pulse wave and sound quality. They're our vocal palette: through them we colour our voices with affection, bitterness, pleasure, disgust etc. With such a formidable range, we are in effect Leonardos of the larynx.

Though complex communication sounds are emitted by a whole range of animals, only three groups of mammals (humans, whales and dolphins, and bats), and three groups of birds (parrots, hummingbirds and songbirds) have to learn them. In all the rest they are innate. Cats' miaows and dogs' barks, for example, are instinctive and not learned. Vervets use a repertoire of 36 different calls to warn of the degree of danger from predators, but they don't go through a period of gradual vocal development.

Apart from humans, birds' vocalisations are vastly more complex and precisely modulated than those of any other animal. Most remarkably, songbirds too learn from experience. Some songbirds even have "dialects" - defined, localised, particular acoustic features that are culturally transmitted. And songbirds come to use their voices only if they're exposed to the communicative signals of adults of their species. Just as human infants who don't hear the human voice or speech don't develop normal vocal capacities of their own, so songbirds raised in isolation produce abnormal "isolate" songs.

The distinctiveness of the human voice is also anatomical in origin. The vocal tract of a newborn is incapable of producing the full range of speech sounds - indeed, it's more like that of an adult chimpanzee than an adult human. But all this changes between the ages of roughly two and six years old, when the gradual descent of the child's larynx makes possible a growing mastery of the sounds of speech.

In this respect humans are unique among air-breathing vertebrates: because our tongue and larynx lie so low in our neck and our pharynx is so long, we can turn our tongue and vocal tract into a huge variety of different, sound-altering lengths and shapes - the prerequisite of speech. Apes, dogs and monkeys, on the other hand, have mouths and tongues that allow them to swallow food without risk of blocking their larynx, and eat and drink while they breathe, but not talk. So this is the trade-off: talking makes us breathe and eat less efficiently. Though we can speak, we can also easily choke on our food. Apes do neither, which is why they don't need to learn the Heimlich manoeuvre.

But the story of the human voice begins before birth. Thanks to modern ultrasound techniques it has become clear that foetuses begin to react to some sounds as early as 14 weeks, and from about 28 weeks' gestation respond to auditory stimulation. Babies possess an astonishing prenatal sensitivity to the voice. Remarkably, although the foetus can't speak or understand speech, it is already able to recognise voices - especially its mother's voice. The mother's voice affects the foetus's heart rate, slowing it down significantly, according to some studies, which suggests that the soothing capacity is present even before the baby is born. In other studies foetuses got more, and not less, excited when they heard their mother speaking. Either way, they're able not only to discriminate between their mother's voice and other people's, but also seem to remember the maternal voice and respond to its familiarity. Even foetuses learn from experience.

Thirty years ago neonates were regarded by developmental psychologists as an almost alien species, lacking the vocal skills and abilities possessed by their parents and older siblings. But it is now clear that newborns come not only with a set of innate abilities that allows them to communicate from the very beginning, but also with a ferocious desire to connect vocally with other human beings, which, if it is reciprocated, turns them rapidly into quasi-speakers.

Immediately after being born, healthy newborn babies everywhere produce an almost identical reflex cry - one which, intriguingly, matches the frequency of the international standard tone for tuning musical instruments. So essential is that first cry as a sign of independent life that, if it is absent, a baby is often - still today - slapped into producing it.

From the very beginning, the baby is linked to its parents in an audio-phonic loop. A baby's cry is tuned to the 3,000 hertz band where its mother's ear is at its most sensitive. The cry is one of the few acts under the infant's control - its most powerful, and often only, means of communication.

Newborns can discriminate between their own cries and those of other babies. They get upset when they hear other babies cry, which is probably why, when one baby cries in a maternity ward, the others inevitably follow. It is always assumed that they are simply copying each other, but they are probably also pained by the sounds of distress. Already at birth, it seems, babies can discriminate vocally between me and not-me, and are most sensitive to the group that most resembles - but isn't - them. Human empathy develops early, and it is expressed vocally.

The infant learns to "speak" through its cry. Sometime during the second and third month it discovers that the sound it produces by itself has the power to conjure up its mother's presence - its voice can summon people, pleasure and comfort. From then on the infant's cry is purposeful, and acts as a kind of speech. Within a couple of months of being born, babies leave a space for a response to their cry: they've learned the art of turn-taking.

Even the newborn, it seems, knows what it wants to listen to - and mostly it's mother. Infants less than two hours old react and orient more to their mother's voice than to those of other women. Newborns younger than three days old also seem to prefer her voice to that of another female - even if they've spent most of their short lives up till then in a nursery and so have been barely exposed to it.

After birth babies smile when they hear their mother's voice (but not when they see her face). So attuned are small babies to their mother's intonation that they suck faster when they hear their mother if, and only if, the mother has recorded a message specially addressed to them. This early aptitude seems to result from their exposure in the womb to the distinctive prosody (pitch, volume, rhythm) of their mother as she expresses different feelings. When she talks angrily, her breathing, heart rate and degree of muscular tension, as well as the movement of her diaphragm, change, so that the baby not only hears its mother's angry voice prenatally but feels it too. And, astonishingly, this capacity to discriminate between the vocal sounds of different emotions and respond differently to them is evident within hours of birth.

In contrast, the few experiments on paternal voices that have been conducted have uncovered an almost shocking indifference in babies. While they can distinguish between their father's voice and that of an unfamiliar man, they don't prefer their father's. Even at four months, infants express no more interest in the voice of their father than in that of a male stranger.

Babies can start vocalising in response to the mother's voice as early as three days after birth. For the first two years of life, according to one study, babies' vocalisations and pauses last for almost exactly the same length of time as their mother's. Decoding its mother's prosody and interpreting her inflections is no mere idle pastime for a baby, but an issue of survival: from the very beginning, it needs to be able to secure for itself care and nurture, and voice-reading is a powerful aid. No wonder babies are so gifted at it.

When we speak to babies and toddlers, we alter our voices dramatically. Although babytalk may seem like an intuitive response to the challenge of communicating with a preverbal child, it's actually a highly sophisticated register. Through it, our voices swathe babies in warmth while helping them understand language and learn the rules of conversation.

Mothers instinctively change their cadences and intonation as their child ages, to accommodate its growing capacity to comprehend. When we talk to newborns, for instance, we use our voice mainly to soothe and calm. By three months, babies are more socially responsive, so the mother uses babytalk to encourage their social and emotional development. There's also an emotional aspect to babytalk. One feature of it that especially enthrals babies is the amount of emotion it expresses. Compared with speech addressed to other adults, where politeness and restraint are usually de rigueur, babytalk is a much better guide to the feelings of the speaker.

The human voice is an exquisite human barometer, sensitive to tiny shifts of feeling. Emotion can be recognised in segments of speech as short as 60 milliseconds. At its best the mother's voice serves as a container for the child, a sonic version of amniotic fluid. A major study found that, purely through the pattern of rhythm between mothers and their four-month-old babies, it was possible to predict which babies would be most securely attached at 12 months. By contrast, postnatal depression is audible in the mother's voice.

From six months, babies start to become skilled imitators of the intonations of the people around them. Babbling used to be thought to be a random collection of sounds, bearing no relation to words, but it is now considered a kind of rehearsal for speech, a playing with its components. Babbling is uniquely human. Babbles may be quasi-words, but they're phonetically similar in English, French, Thai, Chinese and Dutch - at this stage babies still have their own Esperanto. But by eight to 10 months, babbling has become much more language-specific.

Before then, though, babies reared in English-speaking households can still hear syllables that are distinct in Hindi, for example, but not in English - syllables that adult English speakers can no longer differentiate: their phonetic perception has been altered by linguistic experience. In some real phonetic sense, therefore, growing up entails loss. We become deaf to certain sounds: in order to master one language, we have to lose our sensitivity to all of them.

The behavioural psychologist BF Skinner believed that infants are a tabula rasa, and learn to speak in the same way as rats learn to push a bar. Noam Chomsky, on the other hand, proposed that they arrive with innate grammar. Increasingly, researchers into babies and speech argue for something more subtle - that language is innate but modified by experience, a form of "innately guided learning". What's innate isn't a universal grammar or phonetics but "inherent perceptual biases" that place constraints on perception and learning. Babies arrive as "citizens of the world", "with abilities highly conducive to the development of language", but the "neural commitment" to one language that they make very early on shapes the way in which they process linguistic information.

Long beyond infancy, the parental voice remains a major force in children's lives, with the capacity to create and sustain (but also destroy) closeness. Parents often talk purely to signal to their child that, whatever else they're doing at the time - washing up or tying the child's shoelace - they're still intimately involved with them. Their voice acts as a connective tissue.

Interpreting the human voice is one of our most essential human skills. Recent research suggests, remarkably, that children who have trouble decoding vocal prosody are judged less popular and lower in social status by other kids even before they have left nursery. By the time they have reached adolescence, boys with problems identifying anger in the voice show up more in those arrested for sexual offences. Perhaps coincidentally, prisoners given courses in oral skills are less likely to re-offend.

The voice-reading skills possessed by foetuses and finessed by infants and toddlers never really leave us. Pessimists, nostalgics and technophobes have tried to convince us that voice is no longer a central part of human culture, but has been displaced by the text and downgraded by the image. The eye, they insist, has prevailed over the ear, and technology rendered the voice superfluous. They're wrong. Our vocal skills haven't atrophied. On the contrary, they've proved fabulously resilient: every corner of our lives is animated by talk. In fact a skilful use of the voice and a talent for interpreting those of other people are today prized professional assets, ones that infinite numbers of experts, classes and manuals promise to teach. Politicians from Roosevelt to Hitler, Thatcher to Blair, have recognised the importance of the voice in a mass-mediated world, and exploited it to persuade and seduce. Actors from Brando to Streep have used their voices to transform themselves into someone else.

And yet, despite the past 15 years' fascination with conversation and dialect, language and the psychology of talk, with how men and how women sound, with not just what we say but also how we say it, the voice has continued to be marginalised. We have no shared, public language through which to speak about the voice, especially compared to the wide vocabulary that we've now developed to talk about visual images. We've lost a sense of the primacy of the human voice, our most stunning talent. We need to start a new conversation about this most critical means of communication that emerges so early and so urgently. We need to develop fresh ears.