When we speak to babies and toddlers we alter our voices dramatically: listen to an adult talking to a child and you hear a voice forever swooping and soaring.1 But although baby talk – sometimes called ‘motherese’, ‘parentese’, or ‘infant-directed speech’ – may seem like an intuitive response to the challenge of communicating with a preverbal child, or even an example of how easily adults regress, it’s actually a highly sophisticated register operating on many different levels. Through motherese our voices swathe babies in warmth while helping them understand language and learn the rules of conversation. It’s hard to imagine another human faculty that could achieve so much developmentally in such an apparently effortless way.
First identified by linguists in 1964,2 baby talk has now been recorded and analysed all around the world. It has seven main features. Addressing a 2-year-old, we tend to use a much higher pitch – around an octave higher – than when we’re talking to an adult, or even a 5-year-old. Our pitch range is also wider than usual,3 and we use a ‘rising final pitch terminal’, raising our voices at the end of a sentence even when we’re not asking a question. We’re more inclined to whisper, too, as well as speak more slowly in shorter sentences, stressing two different syllables in a word that usually has only one.4 Remarkably, even American Sign Language has its motherese, involving signing more slowly and using exaggerated movements when addressing infants rather than adults.5 You don’t need to be a mother, a parent, or even an adult to use it.6 Children as young as 3 modify their speech when talking to younger children,7 and one child used its extra-high pitch in talking to her dolls when she was only 2 years and 3 months old herself.8 Baby talk is so sensitive to the characteristics of individual children that a mother was heard to alter her voice according to which 3-month-old twin son she was talking to, using a higher pitch and rising intonation contour when speaking to the less responsive baby.9
Reams of research points to the same fact: that, given a choice, young infants clearly prefer baby talk to adult-directed speech.10 There are many different possible reasons. One is simply novelty – babies are stimulated by auditory change, so the exaggerated sweeps of the mother’s voice and her wide pitch range probably excite them.11 (As we’ve seen, by the age of 3 to 5 days, infants can discriminate pitch and volume, and are more sensitive to high pitch.12) When everything has been stripped away from baby talk but its pitch characteristics, infants still prefer it to speech designed for adults.
Baby talk’s features are all geared to get the child’s attention so as to engage in conversation – its distinctive melodies make it more easily picked out against other background sounds.13 By going up at the end of a sentence, mothers are also introducing their babies to another skill essential for successful conversation – turn-taking. Their final rise announces, ‘Now it’s your turn.’
Babies, as we’ll see, mimic adult voices – it’s how they learn to speak. Motherese is the reverse: it gets adults to imitate babies’ voices. So ‘the baby talk of adults seems to have originated in the talk of babies to adults’.14
But babies already talk like babies, so what’s the point of parents doing the same?15 In motherese the difference between what the baby is capable of producing and what it hears from its mother is reduced.16 By borrowing the child’s own emerging sounds, adults when they baby talk are acknowledging and legitimating them.
Motherese is babies’ first dictionary, because they respond to intonation before they understand other aspects of language. A German researcher in the 1920s reported a child of 9 months who not only looked at the clock after hearing, ‘Where is the clock?’, but also did the same when the intonation pattern stayed the same only different words were used.17
Adult-directed speech, being a somewhat sloppy affair, is a poor way to learn language.18 When trying to articulate vowels and consonants we often overshoot our targets: in the rush of daily conversation it’s often hard to detect where one word ends and another begins. Speaking to babies slowly, though, gives them more time to process what they’re hearing,19 and over-enunciating and emphasising a couple of words in a sentence helps them divide up the speech stream into comprehensible units.20
Parents in the United States, Russia, and Sweden (whose languages contain quite different amounts of vowel complexity) all stretch their vowels when speaking to their infants, particularly ‘ah’, ‘ee’, and ‘oo’ – the three vowel sounds common to all the world’s languages.21 Exaggerating vowels in this way increases the acoustic distance between them, making each one more distinct from the next. ‘Beads’, when stretched, is much less likely to be confused with ‘bed’ or ‘bid’, ‘coooold’ is more easily distinguished from ‘kind’, ‘caught’, and ‘can’t’. And this also helps babies ignore meaningless variations in the way one individual or another says the word.
Mothers instinctively change their cadences and intonation as their child ages, to accommodate its growing capacity to comprehend.22 When we talk to newborns, for instance, we use our voice mainly to soothe and calm. By 3 months babies are more socially responsive, so the mother uses baby talk to encourage their social and emotional development.
Her pitch is at its highest when the baby is 6 months old.23 At 9 months both pitch and warmth decline because by now babies are listening to the sounds of their mother tongue more selectively, more like adults.24 As the baby reaches its first birthday, its mother’s voice becomes less emotional and more informational and directional.25 It changes again when the child is between the ages of 2 and 5, when her pitch and pitch range decrease once more. By the time a child turns 5, their attention span is so much better that subtler ways of communicating can be used.26
According to this argument, mothers, by instinctively using and modifying baby talk, are giving their infants a customised tutorial in language, a free grammar lesson. (With its talk of the ‘mommy linguist’, some of the literature on motherese is seriously patronising – as if mommies could never be qualified academic linguists, only intuitive ones.)
Noam Chomsky claimed that the ability to learn language must be innate because the child suffers from such ‘poverty of stimulus’ that it couldn’t possibly acquire such a complex system purely from exposure to the limited examples of speech in its environment. On the contrary, said others, the baby is exposed to a wealth of linguistic stimulus merely through motherese, and the modifications in a mother’s voice are as important as innate universal grammar in helping a child learn the rules of its language and therefore to speak.27
To the psychologist Steven Pinker this line of thinking is just ‘folklore’ – another example of American parents’ ceaseless exertions to prevent their ‘helpless infant from falling behind in the great race of life’.28 But Pinker tries to have it both ways: as a devout Chomskyist he subscribes to the ‘poverty of input’ theory, and then goes and blasts motherese for being terribly complex. Baby talk isn’t designed to guide infants through the language curriculum, he insists, while at the same time acknowledging that it does mark out sentence boundaries and highlight new words. Pinker is an evolutionary psychologist, and therefore much happier comparing motherese to the vocalisations ‘that other animals direct to their young’.29 Note the ‘other’.
It’s paradoxical that motherese became a hammer with which to bash the Chomskyist universalists on the head because motherese itself is now held by many to be innate and universal – and has been challenged in turn on precisely those grounds.
The list of explanations for baby talk never stops growing. Its acoustic patterns have been compared to those of poetry, with our ability to respond, at even a few weeks old, to the poetic features of language in motherese supposedly laying the foundation for our later ability to produce and appreciate literature.30
To others, baby talk is a kind of living, vocal fossil, with its origins in evolution. The increasingly large brains of early homi-nins (human ancestors), if they’d been allowed to mature in the womb, would have made childbirth more difficult. Instead human infants were born earlier, less mature, and so exceptionally helpless – without even the ability (which chimps develop at 2 months) to cling to their mothers.
Since human mothers also lacked body hair that their babies could easily grasp, and stood up on two feet, making any vestigial clinging ability that their babies possessed at birth less effective, they had to put their infants down when they foraged for food. Was prosody used by bipedal hominin mothers to compensate for this increase in distance and keep in touch with their baby? Did motherese act as a kind of ‘vocal rocking’, a way of reassuring babies that mother was near – an extension of the mother’s cradling arms?
Motherese, according to this theory, is genetically driven, the product of past selective pressures, and retains the melody and prosodic features that humans used before speech.31 Yet if motherese had an evolutionary function, wouldn’t it have died out in cultures where babies are carried around by mothers in slings or in those where, on the contrary, very small infants spend most of their time in nurseries, pre-school, or looked after by nannies and child-minders?
There’s also an emotional aspect to baby talk. As we’ve seen, long before the infant can process words, the melodies of its mother’s speech give it access to her feelings and intentions. One feature of motherese 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, motherese is a much better guide to the feelings of the speaker.32
Most research into motherese misleadingly compares speech addressed to a baby with that spoken to an adult acquaintance (rather than, say, to a lover). But when various similar emotions – love, comfort, fear, and surprise – expressed to a baby and to an adult were compared, there were astonishingly few differences between them. Baby talk turned out to be remarkably like lover talk: an adult saying lovingly, ‘Honey, come over here,’ to a baby sounds acoustically almost identical to them saying it to another adult.
So perhaps infant-directed speech isn’t a special register or tone at all, but just reflects our greater expressiveness and vocal freedom when we communicate with babies. Powerful social taboos prevent us from sounding loving or enormously comforting to adults with whom we’re not on kissing terms, but when we do express emotion to other adults, we use almost exactly the same acoustic features as when we speak to babies.33 Motherese is truly the melody of intimacy.
Interestingly, although motherese is cross-cultural, those few cultures where it doesn’t exist are also societies that frown upon uninhibited emotional speech: their adults aren’t only exercising acoustic restraint in communicating with children but also with other adults.
Baby talk’s main emotional purpose may be even more specific: to express happiness. When we speak to adults in a tone as positive as that used in baby talk, 6-month-old babies respond equally well. Even more telling, when babies have to choose between happy adult-directed speech and neutral baby talk, they prefer the adult-directed speech. By this reckoning, motherese is fundamentally a response to babies’ delight in happy voices.34
Is baby talk universal or shaped by culture? As well as English it’s been found in French, Italian, German, and Spanish-speaking societies.35 Arabic has it, and so too do Cocopa, Gilyak and Comanche, as well as Marathi,36 Greek, Hidatsa, Kannada, Kip-sigis, Latvian, Luo, Maltese, Marathi, Romanian,37 and Berber.
In tonal languages like Mandarin Chinese, as we’ve seen, a change in pitch can also change the meaning of words. Yet when Mandarin mothers address their babies, they use a much higher frequency than normal, slow down, and use longer pauses, just like other baby talkers.38 In fact they risk transgressing linguistic rules in order to use motherese.39
But then there’s the experience of the half a million people living in the western highland region of Guatemala who speak the Mayan language of Quiche. Quiche mothers, who are mainly responsible for the care of babies, keep their infants close to them at all times, either strapped to their back, in a cradle of rags near by, or beside them when they sleep. Since they interpret any sound from their baby as a sign that it needs feeding, vocal interaction between infants and parents is minimal. The idea of talking to babies to stimulate their linguistic development is quite alien here: infants are ignored in conversation until the arrival of their first two words – their mothers don’t even recite nursery rhymes or lullabies to them. Quiche parents don’t use baby talk, or any other simplifying register, and yet their babies grow into fully competent speakers of their native language (advantage Chomsky).
The Quiche experience suggests that culture does have a role in shaping motherese. It may be a worldwide, instinctive, unconscious practice with many near-universal features, but these are inevitably modified by a country’s social conventions about how children should be treated – even by differing cultural ideas of what babies are,40 as well as by a society’s rules governing the public expression of emotion. Hence the fact that middle-class Americans, who show the most extreme prosodic changes when they speak to children, are citizens of a country where emotional expressiveness isn’t only tolerated but actually expected, as compared with Asian cultures where there are strong social constraints against exaggerated vocal expression.41
Baby talk isn’t just for babies: we also speak in a high pitch to our dogs42 and rabbits,43 using pet talk or so-called doggerel.44 There are all sorts of factors that affect how and when pet talk is used. Women, for instance, pet talk to their dogs more than men do, but both men and women pet talk more to unfamiliar dogs.45
Despite the apparent absurdity of studying how pet owners speak to their pets, comparisons between baby talk and pet talk are actually quite useful: they can help elucidate what’s distinctive about talking to babies, and whether motherese really does have a linguistic function. An American linguist who claimed never to use baby talk suddenly realised that she used pet talk almost constantly to her cat. Clearly she wasn’t trying to teach it to speak English, only to express affection.46 On the other hand one study found that mothers exaggerated their vowels to their infants but not to their pets, confirming the theory that stretching the vowel or ‘hyper-articulating’ makes the distinctions between the different vowels sound larger so that babies can tell them apart.
Though women speak more affectionately to their children than to their pets, they use a high pitch to address both.47 On some level, we don’t differentiate between our child and our chihuahua.
Baby talk also has something in common with a vicar’s delivery of a sermon (slow, with clear enunciation),48 and with foreigner talk – the slow, distinct speech with frequent pauses and exaggerated stress that we think, perhaps misguidedly, makes us sound more like the person to whom we’re talking.49 Speech addressed to people with learning difficulties, hospital patients, old people, and even plants shares some of these features too and has been called ‘secondary baby talk’.50 In all these cases (except, perhaps, the plants) there’s an element of coaxing and wheedling to soften the directness of commands, providing a kind of smile in the voice for people we care about.51 Or think of as being less competent (babies, people with learning difficulties, old people, foreigners): secondary baby talk might also be an expression of condescension,52 a powerful way of infantilising people.53
In fact, in terms of their high pitch and pitch range, baby talk and secondary baby talk to elderly adults are paralinguistically the same. Secondary baby talk may not only be conveying to the elderly their childlike status but also reinforcing it.54 Again, if the evidence of our voices is to be believed, we regard the very old and the very young as indistinguishable.
Finally, the most charged question of them all: do fathers also speak motherese? Or is there something in the way that mothers address their infants – either intrinsic or learned – that’s different?
Some studies have found only minimal differences.55 In others, mothers talking to their babies increase their average pitch significantly more than fathers, who use the same intonation range for talking to babies as to adults.56
Indeed, when recordings of parents talking to their infants were fed into BabyEars, a machine designed to detect acoustic differences, it was able more accurately to classify which emotion was being expressed when it came to female speech.57 Was this because mothers are more skilful and practised in baby talk? Were the fathers more self-conscious engaging in baby talk in a laboratory? Or did they use different prosodic features to convey feelings?
‘Daddyspeak’ has begun to appear in newspapers and magazines. There are anecdotal accounts of macho men with normally bass booms protesting that they never baby talk, only to turn to their child and say, in a voice a full octave higher, ‘Do I, sweetie? No, I dooon’t talk baby talk to YOOOOU.’58 Of new fathers using diminutive words and exaggerated intonation, sometimes in inappropriate settings (i.e., to checkout assistants in supermarkets).
Do fondly self-mocking articles like this mark a new stage in male self-representation? Some men have begun to appropriate the stereotypes of dizziness previously pinned on to women, but are wearing them proudly as an expression of strength. We fathers, they seem to be saying, are so confident in our masculinity that we even use mushy voices.
‘The descent into Daddyspeak is inevitable’, claimed one. ‘Adjectives mutate into gibberish. Pronunciation dies a slow death’, to be replaced by rhyming sillyisms. This father claimed – in an argument with impeccable academic support – that, ‘Daddyspeak strengthens the tentative bond between father and child. For one, it helps level the playing field. My son can’t speak yet and so, to be fair, I act as if I can’t either … Also, while the adult male voice is often demanding and even scary, Daddyspeak is soft and inviting.’59
But it’s more complicated than this, for when white, middle-class mothers in suburban Atlanta speaking to their 2-year-olds were compared with fathers, both men and women were found to raise their pitch – the men even more than the women. Yet when it came to the older children, the fathers didn’t differentiate between 5-year-olds and adult listeners in terms of pitch. The mothers, on the other hand, were still raising their pitch significantly when they spoke to the older children.60
If the sole purpose of exaggerated intonation was to get the child’s attention, then why wasn’t everyone addressing all children of the same age in roughly the same way? Perhaps it’s a question of fathers’ cultural expectations: maybe men revert to the more limited prosody in order to avoid being tarnished by more stereotypically female speech patterns.61 Taken together, these studies suggest that, despite the exceptions, one of the major factors inhibiting men from using baby talk is still the overriding male imperative not to sound like a woman.
So is baby talk a way of getting infants’ attention, teaching them social and grammatical skills, expressing affection or condescension, or providing security while foraging for food? However many of these functions it manages to accomplish at the same time, there’s no doubt that motherese plays an important developmental function: our voices are shepherding children through their early years. Yet ultimately it remains an enigma – to adults and researchers, if not to babies.