9/7/2016

Linguistics001Lecture21FirstLanguageAcquisition

Linguistics00121:FirstLanguageAcquisition
Stagesoflanguageacquisitioninchildren

home

Innearlyallcases,children'slanguagedevelopmentfollowsapredictable
sequence.However,thereisagreatdealofvariationintheageatwhichchildren
reachagivenmilestone.Furthermore,eachchild'sdevelopmentisusually
characterizedbygradualacquisitionofparticularabilities:thus"correct"useof
Englishverbalinflectionwillemergeoveraperiodofayearormore,startingfrom
astagewherevebalinflectionsarealwaysleftout,andendinginastagewhere
theyarenearlyalwaysusedcorrectly.

schedule

homework

Therearealsomanydifferentwaystocharacterizethedevelopmentalsequence.
Ontheproductionside,onewaytonamethestagesisasfollows,focusing
primarilyontheunfoldingoflexicalandsyntacticknowledge:
Stage
Babbling

Typicalage
68months

Description
RepetitiveCVpatterns

Onewordstage
(betteronemorpheme
918months Singleopenclasswordsorwordstems
oroneunit)
orholophrasticstage
Twowordstage

1824
months

"minisentences"withsimplesemantic
relations

Telegraphicstage
orearlymultiword
stage
(bettermulti
morpheme)

2430
months

"Telegraphic"sentencestructuresof
lexicalratherthanfunctionalor
grammaticalmorphemes

Latermultiwordstage

30+months

Grammaticalorfunctionalstructures
emerge

Vocalizationsinthefirstyearoflife
Atbirth,theinfantvocal
tractisinsomewaysmore
likethatofanapethan
thatofanadulthuman.
Comparethediagramof
theinfantvocaltract
shownontheleftto
diagramsofadulthuman
andape.
Inparticular,thetipofthe
velumreachesoroverlaps
withthetipofthe
epiglottis.Astheinfant
grows,thetractgradually
reshapesitselfintheadult
pattern.
Duringthefirsttwomonths
oflife,infantvocalizationsaremainlyexpressionsofdiscomfort(cryingand
fussing),alongwithsoundsproducedasabyproductofreflexiveorvegetative
http://www.ling.upenn.edu/courses/Fall_2011/ling001/acquisition.html

1/7

9/7/2016

Linguistics001Lecture21FirstLanguageAcquisition

actionssuchascoughing,sucking,swallowingandburping.Therearesome
nonreflexive,nondistresssoundsproducedwithaloweredvelumandaclosedor
nearlyclosedmouth,givingtheimpressionofasyllabicnasaloranasalized
vowel.
Duringtheperiodfromabout24months,infantsbeginmaking"comfortsounds",
typicallyinresponsetopleasurableinteractionwithacaregiver.Theearliest
comfortsoundsmaybegruntsorsighs,withlaterversionsbeingmorevowellike
"coos".Thevocaltractisheldinafixedposition.Initiallycomfortsoundsarebrief
andproducedinisolation,butlaterappearinseriesseparatedbyglottalstops.
Laughterappearsaround4months.
Duringtheperiodfrom47months,infantstypicallyengagein"vocalplay",
manipulatingpitch(toproduce"squeals"and"growls"),loudness(producing
"yells"),andalsomanipulatingtractclosurestoproducefrictionnoises,nasal
murmurs,"raspberries"and"snorts".
Ataboutsevenmonths,"canonicalbabbling"appears:infantsstarttomake
extendedsoundsthatarechoppeduprhythmicallybyoralarticulationsinto
syllablelikesequences,openingandclosingtheirjaws,lipsandtongue.The
rangeofsoundsproducedareheardasstoplikeandglidelike.Fricatives,
affricatesandliquidsaremorerarelyheard,andclustersareevenrarer.Vowels
tendtobelowandopen,atleastinthebeginning.
Repeatedsequencesareoftenproduced,suchas[bababa]or[nanana],aswell
as"variegated"sequencesinwhichthecharacteristicsoftheconsonantlike
articulationsarevaried.Thevariegatedsequencesareinitiallyrareandbecome
morecommonlateron.
Bothvocalplayandbabblingareproducedmoreoftenininteractionswith
caregivers,butinfantswillalsoproducethemwhentheyarealone.
Nootheranimaldoesanythinglikebabbling.Ithasoftenbeenhypothesizedthat
vocalplayandbabblinghavethefunctionof"practicing"speechlikegestures,
helpingtheinfanttogaincontrolofthemotorsystemsinvolved,andtolearnthe
acousticalconsequencesofdifferentgestures.

Oneword(holophrastic)stage
Atabouttenmonths,infantsstarttoutterrecognizablewords.Somewordlike
vocalizationsthatdonotcorrelatewellwithwordsinthelocallanguagemay
consistentlybeusedbyparticularinfantstoexpressparticularemotionalstates:
oneinfantisreportedtohaveused
toexpresspleasure,andanotheris
saidtohaveused
toexpress"distressordiscomfort".Forthemost
part,recognizablewordsareusedinacontextthatseemstoinvolvenaming:
"duck"whilethechildhitsatoyduckofftheedgeofthebath"sweep"whilethe
childsweepswithabroom"car"whilethechildlooksoutofthelivingroom
windowatcarsmovingonthestreetbelow"papa"whenthechildhearsthe
doorbell.
Youngchildrenoftenusewordsinwaysthataretoonarrowortoobroad:"bottle"
usedonlyforplasticbottles"teddy"usedonlyforaparticularbear"dog"usedfor
lambs,cats,andcowsaswellasdogs"kick"usedforpushingandforwing
flappingaswellasforkicking.Theseunderextensionsandoverextensions
developandchangeovertimeinanindividualchild'susage.
Perceptionvs.production

http://www.ling.upenn.edu/courses/Fall_2011/ling001/acquisition.html

2/7

9/7/2016

Linguistics001Lecture21FirstLanguageAcquisition

Cleverexperimentshaveshownthatmostinfantscangiveevidence(forinstance,
bygazedirection)ofunderstandingsomewordsattheageof49months,often
evenbeforebabblingbegins.Infact,thedevelopmentofphonologicalabilities
beginsevenearlier.Newbornscandistinguishspeechfromnonspeech,andcan
alsodistinguishamongspeechsounds(e.g.[t]vs.[d]or[t]vs.[k])withinacouple
ofmonthsofbirth,infantscandistinguishspeechintheirnativelanguagefrom
speechinotherlanguages.
Earlylinguisticinteractionwithmothers,fathersandothercaregiversisalmost
certainlyimportantinestablishingandconsolidatingtheseearlyabilities,long
beforethechildisgivinganyindicationoflanguageabilities.
Rateofvocabularydevelopment
Inthebeginning,infantsaddactivevocabularysomewhatgradually.Hereare
measuresofactivevocabularydevelopmentintwostudies.TheNelsonstudy
wasbasedondiarieskeptbymothersofalloftheirchildren'sutterances,while
theFensonstudyisbasedonaskingmotherstocheckwordsonalisttoindicate
whichtheythinktheirchildproduces.
Milestone

Nelson1973
(18children)

Fenson1993
(1,789children)

10words

15months
(range1319)

13months
(range816)

50words

20months
(range1424)

17months
(range1024)

Vocabularyat24months

186words
310words
(range28436) (range41668)

Thereisoftenaspurtofvocabularyacquisitionduringthesecondyear.Early
wordsareacquiredatarateof13perweek(asmeasuredbyproductiondiaries)
inmanycasestheratemaysuddenlyincreaseto810newwordsperweek,after
40orsowordshavebeenlearned.However,somechildrenshowamoresteady
rateofacquisitionduringtheseearlystages.Therateofvocabularyacquisition
definitelydoesaccelerateinthethirdyearandbeyond:aplausibleestimate
wouldbeanaverageof10wordsadayduringpreschoolandelementaryschool
years.
Sexdifferencesinvocabularyacquisition
Againstabackgroundofenormousindividualvariation,girlbabiestendtolearn
morewordsfasterthanboybabiesdobutthedifferencedisappearsovertime.
SvetlanaLutchmaya,SimonBaronCohenandPeterRaggat("Foetal
testosteroneandvocabularysizein18and24monthinfants",InfantBehavior
andDevelopment24:418424,2002)foundthatinasampleof18montholds,
boys'averagevocabularysizewas41.8words(rangefrom0to222,standard
deviation50.1),whilegirls'averagewas86.8(rangefrom2to318,standard
deviation83.2).By24months,thedifferencehadnarrowedtoaboys'meanof
196.8(range0to414,standarddeviation126.8)vs.agirls'meanof275.1(range
15to415,SD=121.6).Inotherwords,thegirls'advantageinaveragevalueshad
shrunkfrom86.8/41.8=2.1to275.1/196.8=1.5.
Astimepasses,thedifferencedisappearsentirely,andthenemergesagaininthe
oppositedirection,withmalesshowinglargeraveragevocabulariesduringcollege
years(thoughagainagainstthebackgroundofwithingroupvariationthat'smuch
largerthantheacrossgroupdifferences).Here'stable6fromJanetShibleyHyde
http://www.ling.upenn.edu/courses/Fall_2011/ling001/acquisition.html

3/7

9/7/2016

Linguistics001Lecture21FirstLanguageAcquisition

andMarciaC.Linn,"GenderDifferencesinVerbalAbility:AMetaAnalysis",
PsychologicalBulletin,104:15369(1988).

Perceptionvs.productionagain

Benedict(1979)askedmotherstokeepadiaryindicatingnotonlywhatwords
childrenproduced,butwhatwordstheygaveevidenceofunderstanding.Her
resultsindicatethatatthetimewhenchildrenwereproducing10words,they
wereestimatedtounderstand60wordsandtherewasanaveragegapoffive
monthsbetweenthetimewhenachildunderstood50wordsandthetimewhen
(s)heproduced50words.
Allofthesemethods(maternaldiariesandchecklists)probablytendto
underestimatethenumberofwordsaboutyoungchildrenactuallyknow
something,althoughtheyalsomayoverestimatethenumberofwordstowhich
theyattributeadultlikemeanings.

Combiningwords:theemergenceofsyntax
Duringthesecondyear,wordcombinationsbegintoappear.Novelcombinations
(wherewecanbesurethattheresultisnotbeingtreatedasasingleword)
appearsporadicallyasearlyas14months.At18months,11%ofparentssaythat
theirchildisoftencombiningwords,and46%saythat(s)heissometimes
combiningwords.By25months,almostallchildrenaresometimescombining
words,butabout20%arestillnotdoingso"often."
Earlymultiunitutterances
Insomecases,earlymultipleunitutterancescanbeseenasconcatenationsof
individualnamingactionsthatmightjustaswellhaveoccuredalone:"mommy"
and"hat"mightbecombinedas"mommyhat""shirt"and"wet"mightbe
combinedas"shirtwet".However,thesecombinationstendtooccurinanorder
thatisappropriateforthelanguagebeinglearned:
1.Doggybark
2.Kenwater(for"Kenisdrinkingwater")
3.Hitdoggy
Somecombinationswithcertainclosedclassmorphemesbegintooccuraswell:
"myturn","inthere",etc.However,thesearetheclosedclasswordssuchas
pronounsandprepositionsthathavesemanticcontentintheirownrightthatis
nottoodifferentfromthatofopenclasswords.Themorepurelygrammatical
morphemesverbalinflectionsandverbalauxiliaries,nominaldeterminers,
complementizersetc.aretypicallyabsent.
http://www.ling.upenn.edu/courses/Fall_2011/ling001/acquisition.html

4/7

9/7/2016

Linguistics001Lecture21FirstLanguageAcquisition

Sincetheearliestmultiunitutterancesarealmostalwaystwomorphemeslong
twobeingthefirstnumberafterone!thisperiodissometimescalledthe"two
wordstage".Quitesoon,however,childrenbeginsometimesproducing
utteranceswithmorethantwoelements,anditisnotclearthattheperiodin
whichmostutteranceshaveeitheroneortwolexicalelementsshouldreallybe
treatedasaseparatestage.
Intheearlymultiwordstage,childrenwhoareaskedtorepeatsentencesmay
simplyleaveoutthedeterminers,modalsandverbalauxiliaries,verbalinflections,
etc.,andoftenpronounsaswell.Thesamepatterncanbeseenintheirown
spontaneousutterances:
1."Icanseeacow"repeatedas"Seecow"(Eveat25months)
2."Thedoggywillbite"repeatedas"Doggybite"(Adamat28months)
3.Kathrynnolikecelery(Kathrynat22months)
4.Babydollridetruck(Allisonat22months)
5.Pigsayoink(Claireat25months)
6.Wantladygetchocolate(Danielat23months)
7."WheredoesDaddygo?"repeatedas"Daddygo?"(Danielat23months)
8."Cargoing?"tomean"Whereisthecargoing?"(Jemat21months)
Thepatternofleavingoutmostgrammatical/functionalmorphemesiscalled
"telegraphic",andsopeoplealsosometimesrefertotheearlymultiwordstageas
the"telegraphicstage".
Acquisitionofgrammaticalelementsandthecorrespondingstructures
Atabouttheageoftwo,childrenfirstbegintousegrammaticalelements.In
English,thisincludesfiniteauxiliaries("is","was"),verbaltenseandagreement
affixes("ed"and's'),nominativepronouns("I","she"),complementizers("that",
"where"),anddeterminers("the","a").Theprocessisusuallyasomewhatgradual
one,inwhichthemoretelegraphicpatternsalternatewithadultoradultlike
forms,sometimesinadjacentutterances:
1.She'sgone.Hergoneschool.(Domenicoat24months)
2.He'skickingabeachball.Herclimbinguptheladderthere.(Jemat24
months).
3.IteasingMummy.I'mteasingMummy.(Hollyat24months)
4.Ihavingthis.I'mhaving'nana.(Oliviaat27months).
5.I'mhavingthislittleone.Me'llhavethat.(Bettyat30months).
6.Mummyhaven'tfinishedyet,hasshe?(Oliviaat36months).
Overayeartoayearandahalf,sentencesgetlonger,grammaticalelementsare
lessoftenomittedandlessofteninsertedincorrectly,andmultipleclause
sentencesbecomecommoner.
Perceptionvs.productionagain
Severalstudieshaveshownthatchildrenwhoregularlyomitgrammatical
elementsintheirspeech,neverthelessexpecttheseelementsinwhattheyhear
fromadults,inthesensethattheirsentencecomprehensionsuffersifthe
grammaticalelementsaremissingorabsent.

Progressbackwards
Oftenmorphologicalinflectionsincludearegularcase("walk/walked",
"open/opened")andsomeirregularorexceptionalcases("go/went",
"throw/threw","hold/held").Inthebeginning,suchwordswillbeusedintheirroot
form.Asinflectionsfirststartbeingadded,bothregularandirregularpatternsare
http://www.ling.upenn.edu/courses/Fall_2011/ling001/acquisition.html

5/7

9/7/2016

Linguistics001Lecture21FirstLanguageAcquisition

found.Atacertainpoint,itiscommonforchildrentoovergeneralizetheregular
case,producingformslike"bringed","goed""foots","mouses",etc.Atthisstage,
thechild'sspeechmayactuallybecomelesscorrectbyadultstandardsthanit
wasearlier,becauseofoverregularization.
Thisoverregularization,likemostotheraspectsofchildren'sdeveloping
grammar,istypicallyresistanttocorrection:

CHILD:Myteacherholdedthebabyrabbitsandwepattedthem.
ADULT:Didyousayyourteacherheldthebabyrabbits.
CHILD:Yes.
ADULT:Whatdidyousayshedid?
CHILD:Sheholdedthebabyrabbitsandwepattedthem.
ADULT:Didyousaysheheldthemtightly?
CHILD:No,sheholdedthemloosely.

MoreInformation
Agoodstartingpointformoreinformationaboutchildlanguageacquisitionisthe
CHILDESwebsiteatCMU,whereyoucanfindoutaboutdownloadingtheraw
materialsofchildlanguageresearch,andalsosearchaspecializedchild
languagebibliography.
ArecentarticleintheNYTMagazine(PaulTough,"Whatittakestomakea
student",11/26/2006)discussesatlengthsomewellknownstudiesaboutsocial
classdifferencesinlanguageacquisition(BettyHartandToddRisley,"Meaningful
DifferencesintheEverydayExperienceofYoungAmericanChildren"(1995)
BettyHart,"ANaturalHistoryofEarlyLanguageExperience",TopicsinEarly
ChildhoodSpecialEducation,20(1),2000"TheearlyCatastrophe:the30Million
WordGap",AmericanEducator,27(1)pp.49,2003).Theabstractfromthe2003
paper:
Byage3,childrenfromprivilegedfamilieshaveheard30millionmore
wordsthanchildrenfromunderprivilegedfamilies.Longitudinaldata
on42familiesexaminedwhataccountedforenormousdifferencesin
ratesofvocabularygrowth.Childrenturnedouttobeliketheirparents
instature,activitylevel,vocabularyresources,andlanguageand
interactionstyles.Followupdataindicatedthatthe3yearold
measuresofaccomplishmentpredictedthirdgradeschool
achievement.
42isnotaverylargesample,andtherearemanyotherquestionstoask,butthis
worksuggeststhatweshouldbeconcernedaboutpossiblelastingeffectsof
culturaldifferencesinchildren'slinguisticenvironment.
Another,morerecent,studysuggestingthesameconclusionisMarthaJ.Farah,
etal.,("Childhoodpoverty:Specificassociationswithneurocognitive
development",BrainResearch1110(1)166174,September2006).Prof.Farah
andhercoworkers"administeredabatteryoftasksdesignedtotaxspecific
neurocognitivesystemstohealthylowandmiddleSES[socioeconomicstatus]
childrenscreenedformedicalhistoryandmatchedforage,genderandethnicity".

http://www.ling.upenn.edu/courses/Fall_2011/ling001/acquisition.html

6/7

9/7/2016

Linguistics001Lecture21FirstLanguageAcquisition

Fig.1.Effectsizes,measuredinstandarddeviationsofseparation
betweenlowandmiddleSESgroupperformance,onthecomposite
measuresofthesevendifferentneurocognitivesystemsassessedin
thisstudy.Blackbarsrepresenteffectsizesforstatisticallysignificant
effectsgraybarsrepresenteffectsizesfornonsignificanteffects.
AlltheparticipantsinthisstudywereAfricanAmericangirlsbetweentheagesof
10and13.Asthegraphaboveindicates,thedifferenceinperformanceonthe
"Language"partofthetestbatterybetweenmiddleSESandlowSESgirls
representedaneffectsizeofabout0.95.
Thereweretwolanguagerelatedtasks:
PeabodyPictureVocabularyTest(PPVT)
Thisisastandardizedvocabularytestforchildrenbetweentheages
of2.5and18.Oneachtrial,thechildhearsawordandmustselect
thecorrespondingpicturefromamongfourchoices.
TestofReceptionofGrammar(TROG)
InthissentencepicturematchingtaskdesignedbyBishop(1982),the
childhearsasentenceandmustchoosethepicture,fromasetoffour,
whichdepictsthesentence.Itslexicalsemanticdemandsare
negligibleasthevocabularyissimpleandapretestensuresthat
subjectsknowthemeaningsofthesmallsetofwordsthatoccurinthe
test.
ThisfindingisconsistentwithalastingeffectofdifferenceslikethoseintheHart
&Rislystudy,,thoughotherexplanationsarealsopossible.

[coursehomepage][lectureschedule][homework]

http://www.ling.upenn.edu/courses/Fall_2011/ling001/acquisition.html

7/7