What the F: What Swearing Reveals About Our Language, Our Brains, and Ourselves - Benjamin K. Bergen (2016)
Chapter 8. Little Samoan Potty Mouths
There’s a good chance you know what your first word was. Mine, for what it’s worth, was apparently tick-tock. I say “apparently” because, like you, I have no personal recollection of anything that happened when I was twelve months old. You and I, like everyone else, only know what our first words were because parents and relatives have told and retold the stories of these words like the revealing pieces of our identities they’re believed to be. After all, a child’s first word is a critical developmental milestone. But whereas there’s little to distinguish your first smile or first step from mine, a first word reveals something unique about the individual. We tend to believe that what a child says first tells us something about her burgeoning mental life, about her values or interests. A child who says tick-tock might be particularly interested in sounds or mechanical devices or might be in a hurry. A child who says the dog’s name first (like my son did) might have a future as a biologist or might think the dog is more interesting than his parents.
So we care a lot about children’s first words, especially when we are personally invested caregivers. Words for dogs and clocks are actually relatively rare first words for English-speaking children. Far more often, children first name one of their parents. A 2010 poll conducted in Great Britain found that a full 63 percent of parents reported that their children said some variant of dada or mama as their first word. And probably not in the proportions you think. In that poll, 25 percent of children were reported to have articulated a variant of mama (including mom and mommy) as their first word, but fully 38 percent started with some variant of dada (including dad and daddy).1
The fact that most English-speaking children ostensibly first produce the name of a parent, caregiver, or other family member makes intuitive sense to a lot of people because it comports nicely with our belief that these people are the most important parts of the child’s developing universe. But interest alone can’t fully account for what words a child produces first. After all, we have no reason to believe that children who say dada before mama are more interested in their fathers or love them more than their mothers. Frequency of exposure might also play a role. In those cases where a mother talks to the child more than a father does, she might say names for other people (like dada, for instance) more frequently than for herself. Higher frequency of exposure to dada than mama could make learning dada first more likely. And for that matter, maybe some words are easier than others for a one-year-old to pronounce.
All in all, it’s quite hard to say why a child articulates a particular word as his or her first. And when we move beyond English, we find that it gets even more complicated. What if I told you that there’s a place, an island, where children don’t say mama or dada first. Nor do they say the name of the dog or an older brother or sister. And what’s more, these children don’t vary wildly like English-speaking children do, with some saying one word first and others another. On this island, all children say the same word first. And what if I told you that word was shit? That would probably change what you think a child’s first word means and where it comes from.
In the late 1970s, University of California, Los Angeles, anthropologist Elinor Ochs recorded arguably the most surprising discovery ever made about how children acquire their first words, and she did it in the middle of the Pacific Ocean. Ochs was conducting research in Samoa, looking at how people there interact and use language.2 She spent time with locals, observing their daily routines and asking about their experiences. One question she asked mothers was what their child’s first word was. She doubtless expected something along the lines of patterns we’re familiar with from English and many other languages: names of (human or animal) members of the household or other nouns for common objects, like ball or bottle (or, as the British survey found, beer). She probably also expected a lot of variability. While over half of English-speaking kids do produce a name for a caregiver first, the distribution has a long tail. When I ask students in my classes, there are usually nearly as many first words as there are students.
But when Ochs asked the mothers in the families she was working with about their children’s first words, she got a completely unexpected response. Every single one of them reported the very same word. It did happen to be a noun, but it was a special one, used in a very specific way. It was the word tae, which, as suggested earlier, doesn’t mean “mommy” or “daddy.” It means “shit.” More precisely, it’s an abbreviation of the Samoan expression ‘ai tae, which means, “Eat shit.”
This startling fact turns what we thought we knew about children’s first words on its head. It means that a child’s first word is determined by more than just his or her internal values. Obviously, the children in these families didn’t value telling people Eat shit more than they valued their caregivers. No, something else must have been going on—maybe the children were exposed to specific language early on, or maybe something about the word’s pronunciation made it particularly easy for them to articulate. As we’ll see, nothing reveals better where first words come from and what they mean than the story of these particular Samoan children and their little potty mouths.
# $ % !
Why was shit the first word these children said? Let’s look first at the environment. Perhaps this word happened to be particularly frequent in the ambient language that the child was exposed to. Children, of course, are linguistic sponges during their second year of life. Words that you had no idea they were even paying attention to boomerang back at you in a tiny voice. And so, the words that children say often reveal, in an unsettling way, things that their parents don’t notice saying. It’s reasonable to presume that these Samoan children were telling people to ‘ai tae because their parents were casually tossing the expression about as well. They just got caught in the act by their little parrots.
Blaming the parents for the linguistic sins of the child is reasonable in principle, but in this particular case, the children probably didn’t learn the word from their parents. When Ochs asked the Samoan mothers, they rejected outright the idea that they could be the sources. They reported embarrassment that their kids were using words that the mothers themselves shied away from. And even if we retain some lingering suspicion—perhaps the mothers simply weren’t aware of what they were saying or were embarrassed to admit it to some strange woman from California or perhaps the fathers were the culprits—the story still seems unlikely. Imagine how many factors would have to line up. A derogatory epithet like ‘ai tae would have to populate the parent’s speech with such overwhelming frequency that it would take pole position in the child’s vocabulary ahead of other words denoting familiar people, objects, actions, or events. And this would have to be the case for every parent in every family. That’s the type of parental behavior a careful anthropologist like Ochs would have noticed.
So if we think it’s unlikely that parents were the sole source, then perhaps the children were learning this word from some other source in the environment. This is of course a common complaint among parents—you don’t have to travel to Samoa to find parents asking where a child learned to say something vulgar. Often the answer is other kids. Could the Samoan children have been parroting not parents but older relatives and neighbors?
Again, on the surface, it looks likely. These young Samoan children interacted with other children a lot. When Ochs looked closely at how young children in this village were reared, she saw a clear difference from what she was familiar with in North America. Most American children spend the majority of their time in the care of adults—parents, older relatives, nannies, babysitters, or teachers. But in the families she observed, parents primarily cared for only very young infants. After a certain period, they recruited their older children to take over the infants’ primary care. So a one-year-old might be supervised distantly by a mother and more closely by a six-year-old. Children were taking care of children.
So perhaps rather than parents, other juvenile caregivers were responsible for seeding little ears with profanity. It’s not hard to imagine six-year-olds telling each other and their infant charges to ‘ai tae—especially if that was the child caregiver’s own first word! The Samoan infants might have been hearing the scatological input that led to their first words not from other adults but from the children who were raising them—children who themselves may have learned to swear from the older kids who raised them, and so on.
This idea has legs. Although we like to imagine that adult caregivers provide the bulk of the linguistic guidance to young children, the fact is that children quickly reach a stage where they’re learning more from other children than from adults.
Developmental psychologists Paul and Lois Bloom encountered one very clear example of this when they conducted a little experiment on their own children. They wanted to see if they could trick their kids into not swearing by surreptitiously training them to use a made-up curse word. Thus began the short and underachieving life of the pseudo-swearword flep. When one of the elder Blooms stubbed a toe, flep! Broken dish? Flep! Another red light? Flep! As Paul Bloom reports, however, it was a total failure. Whenever the Blooms cried out flep! their kids looked at them like they were out of their flepping minds.3
Once they reach a certain age, kids actually learn most of their language from peers and older children, and they do a very good job of ignoring what they hear from their parents, as psychologist Steven Pinker points out.4 As a consequence, kids often come home with words that their parents don’t use and often don’t even know. They also come home using words the parents do know in ways that the parents would never imagine. Profanity is especially likely to be learned from peers, not only because it’s more likely to be said on the playground than at the dinner table but also because of what kids use it for. Profanity is different from mama and bottle and other words that kids learn from their parents in that children use it as a way to show who they are—to forge their own identity. And for most kids, a lot of their identity is wrapped up in their relations with their peers. As a result, while kids learn their earliest words from their caregivers, they tend to learn later words, including taboo ones, from other kids.
But like parental swearing, kids raising kids doesn’t provide a satisfactory explanation for the case of Samoan ‘ai tae. For one thing, getting the kind of consistency that Ochs observed would require an organized effort on the part of the child caregivers. Consider that merely 63 percent of British children reportedly say a variant of mom or dad first, despite the impassioned full court press that their parents apply. Certainly juvenile caregivers couldn’t pull something so extensive off with ‘ai tae, with or without malice aforethought.
In sum, environmental factors—the way people speak around kids—seem like a red herring. There must be another explanation for Samoan tae.
# $ % !
If ambient words alone didn’t lead these children, across the board, to say tae as their first word, then perhaps the cause has to do less with what the children hear than with what they’re able to articulate. Although children eventually develop mastery of the sounds of their language—by two or three years old, they’re often quite proficient at repeating most of the speech sounds they hear—they don’t start that way. Learning to speak is hard. Maybe tae is just easy for a one-year-old to pronounce.
Again, there’s a lot to recommend this explanation for the tae mystery. So let’s break it down into its component parts. First, it assumes that learning to articulate language is hard, which we know to be true. Suppose you’re a year old, and you’re about ready to pronounce your first word. People around you are flapping their lips, and sounds are coming out—in English, say, or Samoan—and you think this seems like an activity you could get into. But you have a lot to figure out. For one thing, you don’t know what the important sounds are—which differences are worth paying attention to. The first sound of tuck seems different from the first sound of truck (the latter sounds more like what we’d spell as ch). Is that important enough that you need to pronounce them differently? Furthermore, you can’t see much of what’s going on when people pronounce words because lips and skin hide from view most of the vocal tract (the lungs, the larynx, the velum, and so on). So you have to figure out what to move when via inference and trial-and-error exploration of your own vocal tract. And finally, learning to articulate words is hard because even if you knew which sounds were important and how to produce them, you’d still have to put in the work of actually training yourself to pronounce them. You have to get the different actions right, like opening your lips and engaging your larynx, and you have to perform them at just the right time. Delay your larynx a couple of milliseconds and you might accidentally produce a p instead of a b. And all of this is just for one sound; it’s even trickier to string together exactly the required sounds for a given word in the right order in real time. So we know that, as a consequence, kids systematically struggle to pronounce words early on.
The kids-say-tae-because-it’s-easy-to-articulate story also assumes that not all sounds are equally hard to pronounce. This is also true. Some sounds come more naturally to young children than others. By about six months, infants enter a developmental stage during which they babble—producing nonsense sounds and sequences of sounds, like dadada or bidubidubidu. And the particular sounds they articulate are remarkably consistent,5 not only within but across cultures and languages.6 Leading theorists have even proposed that babbling, including the specific sounds children articulate, is part of a universal, genetic program children are born with and that babbling, like secondary sexual characteristics or menopause, is an automatic part of the individual’s natural, scheduled maturation.7
The early sounds that infants around the world produce during babbling won’t surprise anyone who’s spent time around children. As far as consonants go, the ones you typically find in babbling include m, n, b, and d, and these are often followed by p, h, f, t, k, g, f, and w.8 As for vowels, across languages, children seem to be more proficient early on with vowels that you might transcribe as ee, ey, uh, and ah.9
Of course, as there is for just about every claim regarding innate predispositions driving human development, there’s conflicting evidence on this. For example, deaf children were once believed to babble just like hearing infants—which would be a pretty bulletproof piece of evidence for a universal genetic basis for babbling.10 But more thorough recent studies show that deaf children don’t start verbal babbling until about ten months, several months after their hearing counterparts.11 At the same time, deaf infants do display early manual babbling—repeated and stereotyped language-like movements of the hands.12 The fact that deaf children babble manually, together with the delay in their verbal babbling, suggests that input matters for the babbling that children do. What’s more, other evidence demonstrates that the language surrounding an infant can shape, if subtly, the sounds he or she produces.13
Nevertheless, early speech sounds—at least in hearing infants—are relatively predictable. And so are the syllables that children assemble those sounds into. As I discussed in Chapter 2, syllables are the rhythmic units that structure words. And they aren’t all equally easy for children. Syllables come in different types. In the simplest terms, you can think of any syllable as structured around one vowel and optionally one or more consonants that precede and/or follow it. So if we write out consonant and vowel sounds as C and V respectively, then the is a CV syllable. (Importantly, the isn’t CCV because we’re focusing not on spelling but on pronunciation. The word the has just one consonant and one vowel because the two letters th indicate a single consonant sound, produced by putting the tongue between the teeth to create a noisy baffle for air passing through. In Old and Middle English, in fact, this sound was written with just one single letter, þ.) The simplest syllable is a V syllable, like the words a or I. Children are pretty good at V syllables. They’re even decent at VC syllables, like it and up. But early on, they seem to prefer CV syllables, which are in evidence in words like mama and dada, each of which just repeats a single CV syllable twice.14
We know that CV syllables are the young child’s go-to because even when trying to imitate more complicated words, they often reduce them to fit this frame. A commonly encountered example is spaghetti, which rendered in syllables is CCVCVCV, as you can see on the next page.
That first syllable spa often trips children up, and they get around it by simplifying the word in their pronunciation. Often they omit a sound (paghetti) or even a whole syllable (ghetti). Reductions like this are so important for us because what the child says differs from what he or she hears. That tells us what the child can and can’t do at a given age. Generally, at around a year of age children are good at CV, V, and VC syllables, and they’re decently proficient at a range of sounds.
So we’ve established that learning to pronounce speech sounds is hard, but not uniformly so. Children follow largely similar pathways in developing their first sounds and syllables. This means that they’re more likely to articulate certain word-like sequences of sounds early in life and to simplify words of the ambient language in particular ways. Could it be that tae just happens to be assembled from easy, typically early sounds?
To answer this, we need to break down the word. The t seems like a good candidate for early children’s speech. But it’s not pronounced how you think it is. Spelling can be deceiving. Samoan words spelled with t are only pronounced like the English t in formal speech. In informal registers—for instance, when telling someone to ‘ai tae—it’s pronounced like k.15 (If that seems unreasonable, consider how much stranger English orthography is by comparison. Compare how t is pronounced in tree, nation, and the.) So the word begins with k, a sound that children around the globe are a little slow to master. K doesn’t usually come online until after m, n, b, and d. Then comes the ae. This is a diphthong. You should be familiar with diphthongs from English, which has a lot of them. They’re just vowels that start in one part of the mouth and end somewhere else. For instance, the vowels of English night and out are diphthongs. The diphthong in tae is similar to the vowel of night, but it ends with the mouth a little wider open. The upshot is that it’s a complicated sound. Not exactly what you’d expect in a first word.
In sum, tae seems like it would be moderately challenging for a child to pronounce. The t—that is, the k sound—is not well represented among children’s earliest sounds, although it does come along soon after. The diphthong vowel itself would not be easy. So it’s a stretch to think that children would regularly articulate precisely tae as part of their exploration of speech sounds. And even if they did, this would still be only one of many syllables they’d be articulating and typically not among their first.
So the question remains, why is this rare, somewhat challenging syllable interpreted as the Samoan child’s first word? The answer will bring us back to the parents. After all, English-speaking infants presumably also articulate something like this same syllable at points. And yet English-speaking parents rarely record their children’s first words as caw or kite. Although we originally exonerated the parents for tae, at least in modeling behavior, we have to return to them now to understand why, in the face of a flood of word candidates during babbling and then single syllables, they interpret this particular sound as their child’s first word.
# $ % !
Suppose that children across the world and across languages produce largely similar patterns of sounds through babbling and early word attempts, modulo some influence of the ambient native language. In that case, several factors will determine what a caregiver counts as a child’s first word. Not least of these are, of course, the words of the caregiver’s language. If your child produces something that sounds like dada and your language has a word that sounds something like that, then there’s a chance you’ll interpret it as that word.
But perhaps more important are your expectations about what sorts of words a child is likely to say. If a child says something like dada, a caregiver could interpret that as an approximation of dada or daddy. But because infants’ word-like vocalizations at around one year don’t quite use the specific sounds of the language around them, it might sound less like the father label dada and more like Dada, the European avant-garde art movement of the early twentieth century that’s pronounced slightly differently. The same vocalization might sound a lot or even more like dawdle or ta-ta or duh-duh. A caregiver might assume that this ambiguous sequence of sounds is an incipient approximation of dada in part due to the belief that a one-year-old will have no particular interest in or knowledge of Dadaism or dawdling. No, the caregiver presumes that the child is interested in Dad.
And these two factors are causally related. Paradoxically, the assumptions of the caregiver affect the words that a language has to choose from. Let me explain.
In Chapter 1, we saw that most words in spoken languages are arbitrary, with the exception of sound symbolism (or onomatopoeia). But I glossed over another exception at the time. The world’s languages show remarkable systematicity in their labels for parents. Consider the above list (compiled by Larry Trask in a delightful essay on the topic) of words for mom and dad in a host of largely unrelated or distantly related languages.16
Nearly every word for mother in this list has an m in it, and those that don’t have an n. Words for father are more variable but still use the same collection of sounds that we’re familiar with from babbling: b, p, d, and t, along with the vowel a. And these trends are representative of the world’s languages. In 1959, anthropologist George Murdock published a very large survey of terms for parents in nearly five hundred of the world’s languages.17 More than half (52 percent) had words for mothers that used m or n. Words for fathers again showed more variability, but more than half (55 percent) included one of four syllables: pa, po, ta, or to. This is clearly not random. Some process or processes must be conspiring to populate the languages of the world with words for mothers and fathers that sound similar.
The explanation isn’t sound symbolism: mothers don’t sound more or less like m or n than any other thing that a word could name. It’s also not due to some sort of linguistic founder effect, whereby the words used by humans 100,000 years ago have persevered to this day. That would be impossible. Given how quickly words are replaced in the languages of the world, words spoken at a time before all languages diverged from one another (if indeed that’s what happened, which the jury is still out on) would be totally unrecognizable by now.
No, the languages of the world have mama- and papa-like words because parents hear what they want to hear.
Influential linguist Roman Jakobson—an exile of the Bolshevik revolution in Russia who founded the influential Linguistics Circle of Prague in the early twentieth century before he was again exiled during World War II and found his way to a chair at Yale—explained this better than anyone has since.18 Parents, he argued, have expectations about what their children will say, and they specifically expect that their children will talk about them. Since mothers throughout history—for biological and cultural reasons—have tended to spend more time with infants than fathers, they’ve often been the closest on hand to deploy that assumption, and as children most commonly produce something like mama before dada or papa (contra the findings of the recent British survey), mama gets interpreted as referring to the mother.
This is especially true when a language already has a word like mama for mother, but it’s also true even when that’s not the case. Parents are quick to allow for deviations between how adults use language and how children approximate it. So it’s common for caregivers to accept children’s deviant words and pronunciations as proxies for adult equivalents. If I expect my child to call me daddy, but he’s saying something like dahdah, I’ll take it. But parents go farther. They don’t just categorize what the child says as an instance of an existing word; they frequently repeat, reuse, and reinforce it. At fourteen months, my son couldn’t (or just didn’t) pronounce the t sounds of night-night. (Notice that he predictably changed CVC syllables to CV.) So the rest of the household copied him and told him it was time to go nigh-nigh. At fifteen months, he couldn’t say lifeguard, and we followed his lead in referring to the gar-gar at the pool. Adults come in this way to adopt changes originated by children that the caregivers interpret as meaning specific things. As a consequence, the language changes. When enough adults use a word, the word effectively becomes a word of the language. And by this process, even if a language doesn’t have a word like mama to start with, it can gain that word in a generation.
This explains why so many languages have such similar words for mothers and fathers. And it also explains why, even when they’re dissimilar, words for mothers and fathers are simple for children to pronounce. They’re simple because of the interplay of evolution—which has endowed children with brains and bodies that are similar in their abilities and their proclivity to pronounce certain sounds at certain stages—with cultural expectations that caregivers apply to their children and with cultural change on the time-scale of years and decades, during which changes to a language take hold in a community.
And so, with that in mind, we can return to Samoa. Samoan does have words for mother and father, tina and tama, respectively (pronounced kina and kama). But Ochs’s study reported neither of these as a child’s first word. That, of course, was tae. So why doesn’t Jakobson’s explanation apply here as well? Why would mothers in Samoa not believe what their counterparts around the world do—that their children’s first interpretable utterances refer to them? It came as a surprise to me, though perhaps it wouldn’t to anthropologists more familiar with how cultures can vary around the world, that the Samoan parents Ochs was working with had specific beliefs about children that led them down a different logical path.
Here’s the way the Samoan villagers whom Ochs worked with thought about children, as explained by Ochs herself:19
From the Samoan point of view, the small child is heavily under the influence of amio [natural drives that lead people to act in socially destructive ways]. Infants and small children carry out such outrageous behaviors as running and shouting during a church service or formal chiefly council meeting, throwing stones at caregivers, hitting siblings and the like, because they are [believed to be] incapable of … suppressing amio.20
These Samoan parents considered children uncontrollable, unruly, and socially destructive. It’s not hard to see what would lead someone to such a belief. I have thirty pounds of insuppressible amio at home myself. The difference between these Samoan mothers and, say, my household isn’t that Samoans think children are out of control, whereas we don’t. It’s in the balance between this set of beliefs and the idea that children are incipient little communicators, seeking engaged interactions with another person. My family has a little more of the latter, in the balance, and the Samoan mothers, more of the former. And if you’re more prone to think children are ruled by amio, then your expectations about what they’re likely to say will reflect this belief, just as your belief that children are trying to connect through language with their parents will create your expectation that they’ll say a caregiver’s name first. Uncontrollable, unruly, socially destructive little people are more likely to tell those around them to ‘ai tae. And so, whereas an English-speaking North American might interpret something a child emits that sounds like ka as car or cat, the Samoan parents in question were apparently led by their expectations to believe that the child’s little amio was hurling an antisocial epithet.a
The Samoan mothers told Ochs something else, which she was kind enough to relay to me. They told her that, at some level, they liked their kids to be tough, for self-protection. Sticking up for yourself can be a useful survival tool. And so, as they explained it, “sometimes bad is good.” This belief might have contributed to making the parents more prone to hearing profanity.
And so ends the mystery of the diminutive Samoan swearers. Their parents were exceptional not because they happened to have diabolical children or because they let themselves incautiously swear in the presence of infants. Like caregivers around the world, they let their beliefs affect their expectations. In the mushy, imprecise vocalizations of their children, they heard what they anticipated hearing. In other words, in this particular case, the kids didn’t have potty mouths; the adults had potty ears.