6 Talking the talk. 1. An ear forlanguage

Transcription

1 142 How Children Learn Language Children make relaively few comprehension misakes. And he few errors.ha do occur reveal much abou how hey learn o inerpre senences. Iniially, i seems, children rely on small word-based rules ("he noun o he lef of read refers o he doer, he noun o he righ refers o he undergoer"). Laer, bigger generalizaions (like he Canonical Senence Sraegy and he Minimal Disance Principle) emerge, dramaically increasing he child's inerpreive powers. Someimes, as we have seen, hese generalizaidns are TOO powerful. Bu he occasional misakes hey yield are manageable and disappear in ime, as children become beer a dealing wih excepions and special cases. So far in his book, we've alked abou words, senences, and heir meaning, bu we've said nohing abou pronunciaion. The nex chaper looks a how children come o perceive and produce he sounds oflanguage. 6 Talking he alk The firs sound ha a child makes is a shrill cry as air eners his lungs a birh. Cooing noises begin around he age of wo or hree mon~s. The. producion of speech-like sounds under he guise of babbling begms a abou he same ime or a lile aferward and is usually fully developed by he age of six monhs or so. The child's firs real words ofen saro show up around he age ofen or welve monhs. Age range birh o one monh wo o hree monhs four o six monhs seven o welve monhs en o eigheen monhs Typical sounds! crying, burps, gruns cooing squealing, yelling, growling, rills made wih he lips, marginal babbling full-fledged babbling firs words Long before a child ries o say anyhing, hough, he lisens. How children perceive speech is also a good place o begin hinking abou how hey come o learn he sound sysem of heir language. 1. An ear forlanguage Children seem o be especially designed o lisen o language. In fac, hey don' even wai unil hey are born o sar. Speech can be heard in he womb-no clearly enough o discern individual sounds, bu wih enough clariy o make ou he inonaional conour and oher feaures of he speaker's voice. Bu how can we know wheher his prenaal experience has an impac? Because infans exhibi an inborn endency o urn heir heads oward he source of a sound, i is possible o idenify heir early 143

2 144 How Children eam anguagr lisening preferences by keeping rack of when hey urn heir head and how long hey 100k. 2 I's even possible, wih he help of an elecronic pacifier, o each hem o selec he ype of audiory simulaion hey prefer (for example, a ape of one language versus anoher).l These echniques have led o a number of imporan findings. For insance, newborns prefer heir moher's voice o oher female voices." They prefer he language of heir parens over oher languages. ') And hey are able o recognize ha an English senence does no sound he same as a French senence, for example. 6 Bu i doesn' sop here. In one sudy, mohers-o-be read aloud a sory every day during he las six weeks of pregnancy. 7 Two days afer birh, heir infans were esed o see wheher hey found he sory more soohing han an unfamiliar sory. And indeed hey did _ even when he sory was read by someone oher han heir moher. Evidenly hey had been able o pick up enough ofhe sory's rhyhm while in he womb o recognize i afer hey were born. Tha sudy is rumped by anoher experimen in which moherso-be recied a children's rhyme ou loud every day over a four-week period,s The familiar rhyme (bu no an unfamiliar one) was found o bring abou a decrease in hear rae in 37 -week-old feuses - furher evidence ha children are sensiive o he sound paern of language even before hey are born. Consonans in he cradle Prenaal speech percepion is apparenly limied o general rhyhmic and inonaional feaures of language, so here are sill many deails lef o learn afer birh. Foremos among hese is he language's invenory of consonans and vowels. Here again, work wih very young infans calls for special echniques. We know ha an infan's hear rae and sucking rae change when he is exposed o a new ype of simulus. (Special equipmenincluding an elecronic nipple - makes i possible o measure hese hings.) Talking he alk 14 'i So, in order o find ou wheher an infan can hear he difference beween "b" and "p," experimeners playa series o1'''b'' sounds over a period of several minues, followed by a "p" sound (... balm ba ba ~m!. If here is change in hear rae or sucking rae righ afer he pa IS mroduced, we know ha he noices he difference beween "b" and "p."g Infans aged one o four monhs are able o hear he difference beween pairs of consonans. By he ime hey are six monhs old, hey can hear he differences among vowel sounds as well. 10 Infans are even able o disinguish beween sounds in languages oher han he one hey hear a home. In one experimen, six- o eigh-monh-old infans who were being raised in Englishspeaking homes could hear conrass among consonans in Hindi and Nhlakampx (a Naive Indian language spoken on he wes coas of Washingon sae and Briish Colombia). By he ime children are en o welve monhs old, hough, hey are already beginning o have difficuly hearing disincions no found in heir own language, II which seems o have become heir primary focus. 12 By he way, infans aren' he only ones who are gifed a hearing he difference among speech sounds. Various ype of animals have shown similar abiliies, including chinchillas, macaques, and even quail.! l This suggess ha a leas some of he percepual mechanisms relevan for speech have heir roos deep in he mammalian audiory sysem and beyond. 2. Can you hear ha? By he ime a child is a year old, he has begun o learn he words of his language. A his poin, sounds are no jus sounds anymorehey are pieces of words. And lisening involves using hose sounds o hear and idenify words. This is quie differen - and harder - han jus noicing ha wo sounds are no quie he same.!4 Noneheless, children do remarkably well from a very early age. A simple echnique for sudying children's abiliy o disinguish among words was developed by he Russian linguis N. Shvachldn in he 1 940s. I involves showing children pairs of picures whose names differ by jus one sound. (Such words are called millimal pairs.)

3 146 How Children Leanz z,anguage In some varians of his mehod, he names are made-up words like /Jok and pok (Shvachkin himself did i his way). In oher sudies, real words are used (e.g., sca andfre) 1 S or a combinaion of real words and made-up words (like car and gar).1 b Using Shvachlun's mehod, we can deermine wheher a child can hear he difference beween "k" and "g" by showing him he following pair of picures and saying "Show me he coa" (or "Show me he goa"). o o If he poins o he righ objec, we have an indicaion ha he can hear he difference beween he "k" and "g" sounds. (Of course. he es has o be done more han once o make sure ha he didn' ge he righ answer jus by guessing.) When echniques like hese are employed. he resuls show ha children can use phoneic differences o disinguish among words long before hey are able o make he disincions in heir own speech.17 Indeed, some sudies repor ha by age wo or earlier, children can disinguish among words almos as well as aduls can. IX I said "fis, " no "fis" Being able o HEAR a sound is no he same hing as being able o MAKE i, of course. And here is ample evidence ha children have Talking he alk 147 rouble pronouncing many of he phoneic conrass ha hey are able o perceive. Here's jus one example. One of us... spoke o a child who called his inflaed plasic fish a /is. In imiaion of he child's pronunciaion, he observer said: "This is your /is?" "No," said he child. "my [1s." He coninued o rejec he adul's imiaion unil he was old, "Tha is your fish." "Yes," he said, "my fi "19 s. Noice how he child rejecs he experimener's imiaion of his "fis" pronunciaion. I's obvious ha he is able o hear he difference beween "s" and "sh" in adul speech even hough he can' ye make i himself. 20 Do children perhaps make phoneic disincions in heir own speech ha aduls canno hear? Could i be ha he "fis boy" has wo "s" sounds in his speech ~ one for words like fish and he oher for words like miss- and ha he can hear he diiterence beween hem even hough aduls can'? Somehing lh;:e his does someimes happen. In he course of learning he conras beween "" and "d," for insance, a leas some children go hrough a sage where hey have wo differen "d" sounds, one of which is a bi more lil(e "" han he oher. 21 Aduls can' hear he difference, bu i is picked up by he machines ha phoneicians use o analyze speech. In general. hough, children aren' very good a hearing suble disincions in heir own pronunciaion. For insance, when hreeyear-olds lisen o ape recordings of heir own speech and of adul speech. hey can inerpre only abou 50 percen of heir own pronunciaions - compared o nearly 100 percen of he adul pronunciaions. 22 How do you spell ha? An inriguing indicaion of children's sensiiviy o speec h sounds comes from an unexpeced source - heir early aemps a spelling. Preschool children who have learned he alphabe bu have no ye been augh o spell in he convenional way someimes engage in sponaneous aemps o wrie. The resul is a naural sysem of

4 148 How ClzildrCl! Leam LilIlguage "phoneic ranscripion" ha reveals jus how aware children are of he way language is pronounced. 1l Take words such as pencil, open, and kien, for insance. If you lisen carefully o your normal "fas -speech" pronunciaion ofhese words, you'll noice ha he second syllable doesn' conain a vowel. I consiss jus of he "I" or "n" consonan. Children's invened spellings reflec his - hey frequenly wrie PESL for "pencil," OPN for "open," and KITN for "kien." Orake he pronunciaion ofhe "" sound when i occurs beween vowels, as in prey, leer, and boom. In Norh American speech, he "" is pronounced as a sor of rapidly ariculaed "d." (Linguiss call his sound a "flapped D.") Children pick up on his oo, which is why we find spellings lilce PRImE for "prey," LADR for "leer," and BODOM for "boom." Children are also aware of he fac ha he English pas ense ending has hree separae pronunciaions, depending on he preceding sound. For example, i's "" in lauglzed, "d" in sayed, and "id" in waied. Which is why children wrie LAFFT for "laughed," STAD for "sayed," and WATID for "waied." Children's abiliy o perceive he subleies of pronunciaion is indeed impressive, bu i's only half of wha i akes o learn he sound sysem of language. Ariculaion - being able o acually pronounce he sounds-- is he oher (and harder) half. We'll urn o ha nex. 3. Babbling Children show an ineres in how speech sounds are produced before hey can make any hemselves. Infans aged en o sixeen weeks old prefer o wach he image of a speaker whose lip movemens are in synchrony wih his speech o one whose lip movemens are accompanied by a voice delay of a half second. 24 Evidenly, hey realize ha mouh movemens are responsible for sounds and expec o see he wo coordinaed wih each oher. The firs sign of speech-like sounds in children comes when hey begin o babble, ypically a around four o six monhs of age. Mos Tillking lze alk 149 early babbling consiss of repeaed (or "reduplicaed") syllables such as dada, mama, balm, and he like. As we saw in chaper 1, wih a lile parenal imaginaion and encouragemen, some of hese forms evenually ake on a meaning and become he child's firs words. Babbling increases in frequency and complexiy unil he age of abou welve monhs. 2 S I may overlap wih he producion of real words for several weeks before evenually dying ou. Noneheless, babbling doesn' seem o be crucial for he laer developmen of real speech. Children who are unable o babble (because hey have o breahe hrough a ube in heir hroa during he babbling sage) subsequenly acquire normal pronunciaion. 26 Moreover, even deaf children babble, alhough no as much or as fluenly as children wih normal hearing. 1! Babbling across languages Babbling is no easy o sudy. Children's early sounds are ofen poorly ariculaed and run ogeher wih each oher. (Par of he problem is ha an infan's speech organs are quie differen from an adul's - everyhing is smaller, he voice box is higher in he hroa, here are no eeh, and he palae is no as high:~h) Neverheless, some good progress has been made in undersanding is basic properies. I was once hough ha babbling children produced all he possible sounds of human language. 19 Bu ha's jus a myh. In fac, in mos cases, children produce far fewer sounds han are found even in heir own language - le alone all he languages of he world. (Tha doesn' mean ha hey can' hear he difference beween speech sounds from oher languages; we're alking abou speech producion now, no percepion.) There's quie a bi of similariy among children he world over in erms of he sounds ha are mos likely o show up during babbling. One sudy of fifeen languages (including English, Thai, Japanese, Arabic, Hindi, and Mayan) uncovered he following endencies among babbling children.

5 1')0 How Children Lcnm Language (' and 1lI1COllll11011 sollnds eluring lze babbling phase i() Frequenly found consonans p b TIl d n k g s h w y lnfreq uen ly found consonans v h sh ch j I r ng Are babbling children influenced a all by he language hey hear around hem? They mus be, since heir babbling bears some resemblance o he language being learned, especially in is inonaion. In one experimen, samples of babbling from children learning French, Arabic, and Canonese were presened o adul speakers of French, who were asked o pick ou he children learning ha language. 11 The aduls were able o make he righ choices abou hree-quarers of he ime for six-monh-old infans, apparenly by paying aenion o he inonaional melody of heir babbling. Anoher feaure of babbling is even subler, bu every bi as fascinaing. Two researchers did an analysis of children learning English, French, Japanese, and Swedish, aping heir speech a four differen imes. when hey had no words a all and hen when hey had a vocabulary of four words, of fifeen words, and of weny-five words. They found ha he proporion of various sounds in children's babbling is very similar o he proporion of hose same sounds in adul speech. For example, if we look a he relaive frequency of sounds produced by he lips ("p," "b," "m," "f," and "v") in adul speech, i urns ou ha French uses hese sounds more han English and ha English uses hem more han eiher Swedish or Japanese. These differences me refleced in he way children babble. French infans produce a higher percenage of hese sounds in heir babbling han do American infans. And American infans produce Talking lze alk 1 ') 1 more han Swedish or Japanese infans. 12 In hese respecs a leas, he children me jus lil(e heir parens. 4. Early vowels and consonans From around age hireen or foureen monhs, give or ake a few monhs in eiher direcion, children sar producing recognizable words. There are basically wo ways o rack heir abiliy o produce sounds during his period. One is o record and ranscribe children's sponaneous speech in ordinary siuaions (play, meals, and so on). The oher is o elici speech from hem, usually by playing a naming game wih picures such as he following. 11 (I's also possible simply o have he child repea words ha are said by he experimener, 31 bu he naming echnique may give a more accurae idea of wha he can do on his own.) ca ("k" sound) lire ("f' sound) ) s leaf ("I" sound) I's no easy o sudy children's early pronunciaion. Some children do no ariculae heir words c lemly, and i is no uncommon for a child o be inconsisen in his pronunciaion of paricular words. In he space of jus a few weeks, for example, one child used he following pronunciaions for he word lze. 1'i duh, deh. uh. zuh, luh. dl, dee. he

6 1 ')2 How Childrcn Learn Language Differences in he pronunciaion of he same word can occur even wihin he space of seconds. 1b Noneheless, here are a few general endencies o repor, wih he imporan cavea ha hese are jus endencies and ha here is considerable variaion from child o child. Early vowels and consonans Five vowel sounds are ypically acquired quie early-- "ee," "ah," "00," "oh," and "uh." 17 These vowels are found in words such as he following. (When linguiss discuss sounds, hey usually do so wih he help of special symbols from he Inernaional Phoneic Alphabe. In order o keep hings as simple and sraighforward as possible, I'll avoid ha pracice here. However, you can find hese symbols and a descripion of he sounds hey represen in Appendix 2 a he end of he book.) bee ("ee") op ("ah") moo ("00") low ("oh") nu ("uh") The mos common early consonans in word-iniial posiion are "b," "d," "m," "n," and "h." 1H By age wo, a ypical English-speaking child can produce he following consonan sounds. 1'i Typical consonan invenory a age wo P Ie b d g m n There's a fascinaing correlaion beween hese endencies and he disribuion of paricular sounds in he world's languages he sounds ha are acquired early are generally found in more languages han he sounds ha are acquired lae. Cerain sounds w Talking he alk 1 ') 3 are evidenly more basic and easier for he human vocal rac o produce. Even he easy consonans are more likely o be heard a he beginning of a syllable han a he end. As we'll see shorly, children ofen delee consonans a he end of a syllable, pronouncing gum as" guh" and nose as "no." The consonans mos likely o be pronounced a ha end of a syllable in children's early words are "p," "," "k," and H n,"40 By age four, he child's invenory of sounds has grown considerably. All he vowel sounds have usually been acquired by his ime, as have he consonan sounds lised below. 41 (I've used boldface o indicae he sounds ha are likely o have been added afer age wo.) Typical consonan invcnory a age jour p b m v ch d n s z k g ng sh ~ ~-~-- ~ ~-~- -~ Sill o be acquired a his age are he "h" sounds in words like hing and ha. 5. Making adjusmens This brings us o he quesion of wha children do when hey ry o say a word conaining sounds ha hey can' ye pronounce. Someimes, hey avoid saying he word alogeher... 2 Bu more ofen hey eiher drop he ough sound (deleion) or replace i by an easier one (subsiuion). This can happen A LOT: more han 90 percen of he early words produced by some children show he effecs of deleion and/ or subsiuion." 1 Le's look a some concree examples. (If you have a young child a home, you'll probably be able o add dozens of addiional examples of your own.) Deleion Word-final consonans are prime candidaes for deleion, especially in a child's early speech. Iniial consonans, in conras, w y r

7 1 54 How Cili1 ir 'll Learn Langllage are ypically reained, especially if hey come righ in fron of a vowel dog.. "dah" bus~.. "buh" boo,. "boo" delee delee delee Deleion in hese cases creaes a syllable ha consiss of a single consonan followed by a single vowel (Linguiss call his a "CV Syllable.") CVC CV dog -+ "dah" delee CV paerns are iniially favored by children he world over, regardless of wha language hey are learning. They are also he syllable ype mos widely found in human language in general Deleion is also very common when wo or more consonans occur in a row wihin he same syllable. As he following examples indicae, regular rules deermine which consonan goes and which one says. When a consonan a he beginning of a word is followed by an "I" or "r," drop he "I" or "r." blanke -+ "banke" or "bankie" ry -+ "ie" crumb -+ "kum" or "gum" bring -+ "bing" from -> "fum" sleep --'> "seep" "Vhen a nasal consonan ("m," "n," or "ng") is followed by a "p," "," or "k" a he end of a word, delee he nasal. bump ~,. "imp" en -+ "e" or "de" Talking ile alk 1 55 VI/hen an "s" is followed by anoher consonan, drop he "s." sop,. "op" small.. " mah" desk,."dek" You may noice anoher sraegy for some words in which an "s" precedes anoher consonan heir order is cbanged. ask ~+ "aks" spaghei ~--'> "pas-ghei" This is a less drasic soluion o he phoneic hardship creaed by "sk" and "sp," since i keeps boh consonans. Bu i sill makes hings beer for he child, since an "s" is easier o pronounce a he end of a syllable han a he beginning if i's nex o anoher consonan. SlIbsillion Consonans ha survive he deleion process aren' necessarily given an adul -like pronunciaion. If hey are among he sounds ha children have difilculy making, here's a good chance ha hey'll be replaced by an easier sound. Once again, heir fae is deermined by fairly regular processes. Tile sopping process: Replace a consonan ha is produced wih a coninuous flow of air (e.g.. "s," "z," "sh," or "h") by a consonan ha is produced by compleely cuing off he How of air (e.g., "" or "d"). Word Child's pronunciaion Subsiuion sing "ing " s -+ see "ee" s-+ zebra "debra" z-+d hing "ing " h -+ his "di" h~-+ d, s -+ shoes "ood" sh -+. z -+ d =

8 ] ::; (j How Childre1l Leam La1lguage (You can verify for yourself wheher a paricular consonan is produced wih a coninuous flow of air by aemping o prolong is pronunciaion. You'll noice ha consonans such as "s," "z," "sh," and "h" can be prolonged indeliniely, bu ha consonans such as "" and "d" canno.) The gliding process: Replace "I" or "r" by "y" or "w." Word Child's pronunciaion Subsiuion lion laughing look rock sory "yine" "yaffing" "wook" "\voc!\:" "S[(llVY" 1-+ y I Y 1-> w r > \IV r-'> w The dellasalizaion process: Replace a nasal sound ("m" or "n") by a non-nasal sound ("b" or "d"). Word Child's pronunciaion Subsiuion jam room spoon "dab" "woob" "bood" m.b m-+b n -'> d The ji'oning process: Replace a consonan by a sound ha is made more oward he fron of he mouh. Word Child's pronunciaion Subsiuion humb ship jump chalk go "fum" "sip" "dzump "salk" Hdoe" " h -+ [ sh --+ s j dz ch -+ s g-+d Talking he alk 1::; 7 The ligure below summarizes where various consonan sounds are produced in he mouh, saring a he lips. (For furher deails and a diagram of he vocal rac, see Appendix 2.) Fron Lips Fron ('('h Alv('olar ridg(' (bdlind h(' fron keh) p. b, m h, d, n Lv s,z Assimilaion Blick Hard palak Sof pala(' sh, ch. j k. g, ng There's a leas one more imporan ype of sound change in children's speech. You can hink ofi as a ype of subsiuion, since i involves one sound being replaced by anoher. However, unlike he subsiuions we've alked abou so far, he oucome is deermined by a neighboring sound. Pu simply, a sound ries o become more like is neighbor in some respec. (Linguiss refer o his as assimilaion.) One very common ype of assimilaion involves a sound being modified so ha i is produced a he same place in he mouh as is neighbor. A good example of his occurs in words like impaien, where he "n" of he prefix in- (compare inacive, indirec, ec.) is changed o "m" because of he following "p." (Like "p," "m" is made wih he lips; "n" is made wih he ip ouhe ongue.) in + possible "n" becomes "m" o make i more like "p" V- m Assimilaion no only makes words easier o pronounce by reducing he number of differences among neighboring sounds, i is apparenly aracive o he ear as well. Infans as young as four monhs of age prefer o lisen o nonsense words such as limber compared o wzber. 44 Wha's he difference? Umber shows he effecs of assimilaion, since "m" and "b" are boh pronounced wih he lips.

9 1 'i 8 How Childre/l LmrII Lallguage Anoher common ype of assimilaion involves vocal-cord vibraions (or voicillg, o use he echnical erm). The vocal cords vibrae during he pronunciaion of all vowel sounds, as you can see for yourself if you place a finger on your neck as you say "ee," "ah," or "oh." Some consonans ("z" and "v," for insance) have accompanying vocal-cord vibraions, while ohers (like "s" and "[') do no. When a consonan ha shouldn' have vocal-cord vibraions occurs in fron of a vowel in children's early speech, i ofen ends up "caching" he vocal-cord vibraions from he vowel. Tha is, i becomes "voiced," urning ino an enirely differen sound. The able below conains some consonan conrass ha are based on vocal-cord vibraions (see Appendix 2 for addiional informaion). No vocal-cord vibraions P Ie f s Vocal-cord vibraions In he firs wo examples below, a "p" picks up he vocal-cord vibraions from he following vowel and urns ino a "b." A similar sor of hing happens when "" urns ino "d" (he hird example). and "s" urns ino "z" (he final example). b d g v z Word Child's pronunciaion Change pig push E'l1 soup "big" "bush" "dell" "zoop" p--:.b p -+ b --:. d Talkillg he alk 1 'i 9 Someimes, we even find oal assimilaion, which makes a sound idenical o a nearby sound. The resuls of his ype of change can be seen in he following words. Word doggy "goggy" self "fejr' Kahleen "KaIdE'E'n" baby "beebee''' 6. Sress is good Pronunciaion Change d -'> g. because of he nearby "g" S-r r. because of he nearby "f' h.,. k. because ofhe nearby "Ie" a -+ ee. because of he nearby "ee" (wrien as!j) One of he mos noiceable feaures of children's early speech is ha hey ofen drop enire syllables, especially when hey ry o pronounce longer words. Adul word Child's word j) giraffe' musache goodnigh away faff ass na way This creaes a shorer word ha is easier o pronounce, of course, bu we need o know why children delee he firs syllable in hese examples raher han he second one. (Children don' pronounce giraffe as "gi" or 11lusacize as "mus.") The explanaion for his seems o lie in he way children perceive longer words. Take he word alva!], for example. If you pronounce i o yourself. you'll noice ha he second syllable (he "way" par) is pronounced more forcefully han he firs. This is wha linguiss call "sress." English disinguishes hree levels of sress primary sress, secondary sress, and unsressed. A syllable wih primary sress is

10 16{) How Clzildrell Lmrll Lallguage pronounced louder han any oher syllable in he word, and is vowel is fully ariculaed. In conras. an unsressed syllable' is far less audible, wih a shor and weak VOWE'1 ha linguiss call a "schwa." The firs syllable of a word like' abou and il' las syllable' of soja conain a we'ak VOWE'1 of his ype'. primary sress abo l' unsressed primary sress so fa unsressed InermediaE' bewe'e'n a syllable' wih primary sre'ss and an unsressed syllable are syllables wih secondary sress. They are no pronounce'd as srongly as syllables wih primary sre'ss, bu hey don' have a weak vowel eiher. The se'cond syllable' in he words slowly and veo has secondary sress. In a word like alligaor, he firs syllable has primary sress, he hird syllable has secondary sress, and he second and fourh syllables are unsressed. Le's use a double underline' for primary sress and a single unde'riine for secondary sress. primary sress al secondary sress Ii 1@ or unsressed Under he spoliglz N ow you can probably guess wha's going on in our earlier examples. Because syllables bearing primary sress are more audible han heir unsressed counerpars, children end o ZE'ro in on hem whe'n hey are firsle'arning o speak (remember he "spolighs" discusse'd in chaper 2). For his reason, he sresse'd syllable is usually no dropped. 46 Talking ile alk ] 61 In E'xrE'mE' case's, il' sre'ssed syllable' may be il' only par of il' word ha is pronounce'd, even in words conaining hre'e' or more syllable's. (RE'membE'r ha a double' unde'rline' marks primary sress and a single underline indicae's se'clildary sress.) INord hlp po P~21 a mus kan ga~q spa gl~!li Child's pronunciaion 17 pahs woo ge Someimes, hough, children pronounce only he syllable' wih se'condary sress, especially if i's closer o he end of he' word han he syllable wih primary sress. Word ~li1@ or ea er IJli lar 1:' Ie IJhon~ Child's pronunciaion{s gay pi fo And someimes, syllables wih eiher primary or secondary sress are spared, bu unsressed syllables are' dropped. cl9 Word ill Ii 1@ or a~!"lum hd i.'o'qq er Child's pronunciaion III agay quarium alka I's as if children adoped a "Hear no schwas" policy. ljnsressed syllable's wih weak vowds are ofen jus no audible enough o make i pas he hreshold of percepibiliy when a child is geing sared on English.

11 162 How Children Learn Lm1llllalll' There is one excepion o his hough - children ofen keep he fillal syllable of a word, even when i's unsressed. This is probably because he ends of words are easier o noice and remember since hey're he las hing a child hears. Tal/dllil lle alk 16 3 Noice ha Joshua's words all have he same "profile" or "silhouee" as he adul words, since he number of syllables is he same. Bu generally only he sressed syllable resembles he corresponding par of he adul word. Word ba na na comr~er de phan <.!nimai ~ Ii g;-j or Child's pronunciaion \ I ana puer elfun amui gayda However. even when an unsressed syllable is pronounced, i's ofen missing one or more sounds. You can see ha in he pronunciaion ofhe las syllable of elephall, which is missing he "1. "'12 Some children use a compleely differen sraegy [or dealing wih long words. They produce he same number of syllables ha are found in he adul word, bu usually only one of he syllables (ofen he one wih primary sress) sounds anyhing like a par ofhe adul word. Tweny-monh-old Joshua was like his. In he examples ha follow, I've used boldface o mark he syllables in Joshua's pronunciaion ha seem o be based on syllables in he corresponding adul word. S1l11l11lillllllP Children come ino he world already able boh o recognize he language of heir parens and o make disincions among he sounds of any language. By age wo hey have figured ou which sounds heir own language uses, and hey are well on heir way o being able o pronounce hose sounds hemselves. Pronunciaion problems - and here are a few, as we have seen - are deal wih in sysemaic and predicable ways by judicious deleions and subsiuions ha allow he child o ge by unil full fluency arrives. The sudy of how children come o perceive and pronounce speech sounds, like he sudy of every oher aspec oflanguage acquisiion, raises he quesion ha has puzzled researchers for decadeshow do hey do i? We'll ake a closer look a his quesion in he nex and final chaper of his book. Word 9u~!lY cio ver iger ~uij i'2 zer kan ga roo!l~ ero Qhone Joshua's pronunciaion \ 1 ba bi dodo a da boo duhduh dada wu rna wuhwuh wa wa wi dah bee buh