The management of cohesion in written narratives in students with specific language impairment: Differences between childhood and adolescence

HAL Id: hal-01950077

https://hal.archives-ouvertes.fr/hal-01950077

Submitted on 5 Dec 2019

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

The management of cohesion in written narratives in

students with specific language impairment: Differences

between childhood and adolescence

Monik Favart, Anna Potocki, Lucie Broc, Pauline Quemart, Josie Bernicot,

Thierry Olive

To cite this version:

Monik Favart, Anna Potocki, Lucie Broc, Pauline Quemart, Josie Bernicot, et al.. The management

of cohesion in written narratives in students with specific language impairment: Differences between

childhood and adolescence. Research in Developmental Disabilities, Elsevier, 2016, 59, pp.318-327.

�10.1016/j.ridd.2016.09.009�. �hal-01950077�

Contents lists available atScienceDirect

Research

in

Developmental

Disabilities

The

management

of

cohesion

in

written

narratives

in

students

with

specific

language

impairment:

Differences

between

childhood

and

adolescence

Monik

Favart

_,

_Anna

_Potocki

_,

_Lucie

_Broc

_,

_Pauline

_Quémart

_,

Josie

Bernicot

_,

_Thierry

_Olive

a_{CentredeRecherchessurlaCognitionetl’Apprentissage,UMR7295,UniversityofPoitiers,France} b_{Bases,Corpus,Langage,UMR7320,ESPE,UniversityofNiceSophiaAntipolis,France}

a

r

t

i

c

l

e

i

n

f

o

Received3February2016

Receivedinrevisedform13July2016 Accepted13September2016 Numberofreviewscompletedis2 Keywords:

ChildrenandadolescentswithSLI Narrativewriting

Communicativesituation Cohesiondevices

a

b

s

t

r

a

c

t

Thegoalofthisstudywastoinvestigatethemanagementofcohesionbychildrenand adolescentswithspecificlanguageimpairment(SLI)whenwritinganarrativeina com-municativesituation.TwelvechildrenwithSLI(from7to11yearsold)and12adolescents withSLI(from12to18yearsold)werechronologicalage-matchedwith24typically devel-oping(TD)childrenand24TDadolescents.Allparticipantsattendedmainstreamclasses: childreninelementaryschoolsandadolescentsinmiddleandhighschools.Analysesof cohesionfocusedonbothdensityanddiversityofconnectives,punctuationmarksand anaphors.ResultsattestedthatchildrenwithSLIweregreatlyimpairedintheir manage-mentofwrittencohesionandusedspecificformspreviouslyobservedinnarrativespeech suchasleftdislocations.Bycontrast,andnotexpected,themanagementofwritten cohe-sionbyadolescentswithSLIwasclosetothatoftheirTDpeers.Thecommunicativewriting situationwesetup,whichengagedparticipantstotakeintoaccounttheaddressee,also madepossibleforadolescentswithSLItomanagecohesioninwriting.

©2016ElsevierLtd.Allrightsreserved.

Specificlanguageimpairment(SLI)isadevelopmentaldisorderoflanguageacquisitionthatoccursintheabsenceofmentalretardation, neurologicaldamage,hearingdeficits,orenvironmentaldeprivation(Bishop,1992a,1992b;Leonard,2014).Languagedifficultiesare het-erogeneousandappearessentiallyatthephonological,morphosyntactic,andsemanticlevels.Thesedifficultiesgenerallyremainthrough adolescence(Snowling,Bishop,&Stothard,2000;vanWeerdenburg,Verhoeven,Bosman,&vanBalkom,2011).ChildrenwithSLIalsoare atriskfordevelopingproblemswiththeacquisitionofwrittenlanguage.Forinstance,childrenwithSLIaged9;5yearsoldproducemore spellingerrorswhencomposingalettertotheirbestfriendthantheTDage-matchedparticipants(Williams,Larkin,&Blaggan,2013;seealso Cordewener,Bosman,&Verhoeven,2012).Broc,Bernicot,Olive,Favart,andQuémart(2013)alsopointedoutthelexicalspellingdifficulties ofparticipantswithSLIagedfrom7to18yearsold.Theyhoweverobservedthatthesedifficultieswerelowerinanarrativecommunicative situationthaninanevaluativetask,i.e.adictationofisolatedwords.Inthisstudy,eachparticipantcomposedanarrativewiththepresenceof theexperimenterthroughouttheentirewritingprocess.Theexperimenterorallydeliveredtheinstruction:participantswereaskedtowritea storyaboutapersonaleventandtomakeitaspreciseaspossible,sothatshe(theresearcher)couldreallyunderstandwhathappened.Insuch asituation,performancesinthemanagementoflexicalspellingbyadolescentswithSLI(agedfrom12to18yearsold)werenotsignificantly differentfromperformancesoftheirTDage-matchedpeers.

∗ Correspondingauthorat:CeRCA(UMR7295),UniversitédePoitiers,MSHSBât.A5,5rueThéodoreLefebvre,TSA21103,86000Poitiers,France. E-mailaddress:monik.favart@univ-poitiers.fr(M.Favart).

http://dx.doi.org/10.1016/j.ridd.2016.09.009

TheaimofthepresentstudywastoinvestigatetheabilityofchildrenandadolescentswithSLItomanagecohesioninwrittennarratives. Cohesionensuresthestructurationofthetextasawholeandiscarriedoutusingspecificlinguisticdevices:connectives,punctuationmarks andanaphors.Thesedevicesplayamajorroleinwriting,drawinguprelationshipsbetweenideasgeneratedduringtextplanning.Inthisway thedevicesenabletheaddresseetoestablishtherepresentationofthetextasawholeandcontributetotextualcoherence(seeApotheloz, 1989;Favart&Passerault,1999forconnectives;Passerault,1991forpunctuationmarks;andReichler-Beguelin,1988foranaphors).

1. Theuseofcohesiondevicesduringwritingacquisitionoftypicallydevelopingchildren

Connectives,punctuationmarksandanaphorsarelinguisticdevicesthatcarryspecificfunctionsindiscourseprocessing.Asasyntactic function,theylinktogethertextualsegments,andassemanticandproceduralfunctions,theyprovideinstructionstotheaddresseeto accu-ratelyintegratetextualcontent.Thus,themainroleofcohesiondevicesistooperateasprocessinginstructions.Theycarryaproceduralcore meaning,inthattheyinstructtheaddresseehowtoconnectorseparatediscourseeventsandsohelpinghim/hertodrawupanaccurate representationoftheoveralldiscourse(Louwerse&Mitchell,2003).KarmiloffandKarmiloff-Smith(2003)pointedoutthecloseconnection betweentheachievementoftextualcoherenceandtheaccuratemanagementofcohesioninlanguageacquisition.

Thestudyofconnectivesinthecourseofwritingacquisitionrequiresspecificfocusonthewordand(seeFavart&Passerault,1999fora review).Usingandisveryconvenientforyoungstudentsasitcanexpresssingle-handedlyasubstantialvarietyofrelations.Itshugefrequency inchildren’swrittenstoriesatteststoanimmaturelocalmodeofplanning,carriedoutaccordingtostep-by-stepprocessing.Theconnective andissystematicallyrepeatedandtriggeredbetweentwocyclesofcontentgenerationtosupportthesubstantialburdencausedwithin workingmemorybythewritingactivity.Thechronologicalconnectives:puis,après,ensuite(after,then)canbeusedthesamewayasand betweentwogeneratedideastosustainthewritingactivityandtoavoidchildrentoexpressmoresophisticatedrelationships.

Fromtheageof10years,thedecreasingloadofhandwritingandorthographiccodingenableschildrentodevotemorecognitiveresources tolinguisticchoices(Berninger&Swanson,1994;McCutchen,1996).Atthesametime,childrengraduallygainastructuredknowledgeof thesystemofconnectives(Karmiloff-Smith,1985,1992).Thisenablesthemtodiversifyrelationshipsusingconnectiveswhicharemoreand moreconsistentwiththetextualstructure.

Specificconnectivescanbeselectedtotranslatemoreappropriatelynarrativerelationshipsandgraduallyincreasecoherenceinwritten stories.FavartandPasserault(1995)showedthatthedensityofand(relatedtothetotalnumberofusedconnectives)decreasedacross elementaryschool,mainlyingradefive(i.e.,in10yearsoldchildren).Theundifferentiatedconnectivethengavewaytomorediversified ones:mais(but),alors(then);temporalconnectivesquand,lorsque,soudain,toutàcoup(when,while,suddenly);andcausalconnectivescar, parceque(since,because).

Thepunctuationmarksusedbychildrenwhenwritingstoriesaremostly,evenexclusively,periodsandcommas.Fromtheageof7years, childrenusetheperiodpredominantlytodelimitblocksofinformation,justasthewordandoperates(Favart&Passerault,2000).Theperiod andtheconnectiveandfunctioncompetitively,aswellasinanadditionalway,topacetherepetitionofcyclesofcontentgeneration.The useofthecommaemergesinthirdgrade(i.e.,attheagesbetween8and9yearsold).Suchadiversificationopensupnewpossibilitiesbut simultaneouslycreatesnewproblems(Fayol,1997)becausethirdgradersconcurrentlyuseperiodsandcommas,sincetheyarenotyetaware ofthevalueoftheirrespectivebreak(period>comma).Finally,theuseofthecommagraduallyincreasesduringfourthandfifthgrades,andit progressivelybecomesappropriate.Justaswithconnectives,theuseofperiodsandcommaschangesattheageof10yearsold,i.e.,inthelast gradeofelementaryschool.Periods,quasi-exclusiveuntilfourthgrade,werejoinedinfifthgradebyincreasinguseofcommas,particularly withinthedescriptionofdetailedinformation.Itisnotbeforetheageof14yearsthatchildren’suseofpunctuationmarkscanbedetermined atatextuallevelinordertoappropriatelyclarifythetextualcontentfortheaddressee(Schneuwly,1988).

Specifically,anaphorsareusedtorefertoapreviouslymentionedreferent(Gernsbacher,1989),predominantlyasnominalizationspaired withdefinitearticlesorpronouns.Theyenabletheaddresseetoestablishrelationsbetweentextualsections,justasconnectives(Fraser, 1999).Connectivesadditionallyspecifythenatureoftherelationshipstotheaddressee(Stoye,2013),andtheirinterpretationisinalarge partdeterminedbytheaddressee’smentalrepresentationofevents(Moeschler,2002).Punctuationmarksbuildupahierarchicalsystem accordingtothestrengthofthebreaktheyinducebetweentextualunits.Theyalsoservetoorganizetextualsectionsinordertoenablethe accurateconstructionofthewholetextmeaning.

Developmentalstudiesconductedintypicallydevelopingchildrenattestedthatatabouttheageof10years,childrencanmanageaccurately anaphorsinwrittennarratives(Bartlett&Scribner,1982;Decool-Mercier&Akinci,2010;Pellegrini,Galda,&Rubin,1984).Thisskilldevelops bothquantitatively(Hickmann,2004)andqualitatively(Rutter&Raban,1982)untiltheendofelementaryschool.Childrenthencanaccurately usevarioussortsofpronouns(e.g.,personal,relative,demonstrative)andjugglewithpropernames,ornominalsubstitutionswithdefinite articles,tomaintainreferenceifnecessary.Attheageof10years,childrenalsocanaccuratelymanagereferentialambiguitywhenwriting storiesfrompicturesthatdisplayedtwocharactersofthesamegender(Favart&Passerault,1996).However,Lambert(2003)pointedoutthat thereferentialupholdingbypronounsstillcouldbeproblematicattheendofelementarygrades.Thus,theacquisitionandthemanagement ofcohesiondevicesintextcompositionisquitealaboriousprocess,evenintypicallydevelopingchildren(seeFavart,2005forareview).

2. NarrativewritingabilitiesinstudentswithSLI

OnlyafewstudieshaveinvestigatedwritingabilitiesinparticipantswithSLI,andmostofthesestudieswereconductedwithEnglish nativespeakers.Moreover,nostudyhasinvestigatedparticipants’overallmanagementofcohesioninwriting.Thefewavailableresults demonstratethatstorieswrittenbyparticipantswithSLIareshorterthanthoseofchronologicallyage-matchedparticipants(seeDockrell, Lindsay,Connelly,&Mackie,2007;Dockrell,Lindsay,&Connelly,2009;Fey,Catts,Proctor-Williams,Zhang,&Tomblin,2004;Mackie& Dockrell,2004;Williamsetal.,2013).AccordingtoDockrellandLindsay(2000),teachersreportthatlinguisticcodingisspecificallydifficult tomanageduringwritingforstudentswithSLI.Theirselectionoflinguisticunitsappearstobehinderedandreducedtoaminimum.Butthese findingsmainlyfocusedongrammaticalcoding.Sowhenwritingstories,9-to12-year-oldparticipantswithSLIwereattestedtoproduce moregrammaticallyunacceptablecomplexT-Units(i.e.,includingbothamainandasubordinateclause)thanage-matchedchildren(Gillam &Johnston,1992).Thegrammaticalcomplexityorbasicsentencestructurealsowaslower,and/ortheyproducedmoregrammaticalerrors thantheirchronological-agepeers(Dockrelletal.,2007;Feyetal.,2004;Mackie&Dockrell,2004;Scott&Windsor,2000).Thefewresearch

studiesanalyzingnarrativequalityconfirmthatchildrenwithSLIhaveapoorerwritingqualitythanwhatwouldbeexpectedattheirage, evenifitcouldimprovebetweensecondandfourthgrade(Feyetal.,2004),andahighlevelofliteracyimpairment(Freed,Adams,&Lockton, 2011).

Mostofthestudiesfocusingonnarrativecohesionhavebeenconductedinspeech.Bothcausalandtemporalconnectivesareproduced regularlyintypicalchildren’snarrativespeakingfromtheageof7or8years(Kern,2000).Hilaire-DeboveandRoch(2012)analyzedtheuse ofcausalandtemporalconnectivesinchildrenwithSLIagedfrom7to12yearsold.TheyobservedthatparticipantswithSLIproduceda seriesofsentencesregularlylinkedwiththeconnectiveandorwithchronologicalconnectives(87,6%oftheoccurrences)andonly10,4%of temporalorcausalconnectives.Moreover,theydidnotnoticeanyimprovementinconnectivesdiversificationbetweentheagesof7and 12years.AccordingtoBernard-BarrotandGéhard(2003),narrativespeakingofchildrenwithSLIincludesasmanyconnectivesastypically developingchildren’sbutmakesuseoflessdifferentiatedones.

Hilaire-DeboveandRoch(2012)alsoanalyzedtheuseofpronouns(boththeirmaintenanceandreintroduction)inchildrenwithSLIaged from7to12yearsold.Theynoticedtheequaluseofpronounsandofdeterminer+noundevice.ChildrenwithSLIstillusethelatterformuntil theageof11years,whereasitsuseismarginalaftertheageof5yearsinTDchildren.ChildrenwithSLIalsodifferfromTDchildrenintheir useofspecificconstructionssuchasleftdislocations(DeWeck&Jullien,2013;Jullien,2008).Suchdislocationsaresaidtopromoteaccessible referentstothestatusoftopicbyextractingthelexicalunitfromtheclausetoitsleftperipheryandthenco-indexingthisreferentwitha pronoun.TheyareexpressedsuchasC’estNPqui=It’sNPwho(i.e.,C’estungarc¸onquilepousse:It’saboywhopusheshim).Thesedislocations arevisibleinoralproductionofparticipantswithSLIbetweentheagesof4and11years,buttheydeclineinchildrenfromtheagesof10to11 years,whotendtoperformsimilarlytotypicalchildren(Jullien,2008).Moreprecisely,inthestorytellingonthebasisofpicturesproposedby DeWeckandJullien(2013),theproductionofdislocationswassignificantlyhigherinchildrenwithSLIthaninTDchildrenonlyattheages between6and7years.ThechildrenwithSLIproducedmoreleftdislocationsandfewerpresentationalconstructionsthantheirTDpeers. Accordingly,aleftdislocationseemstoreflectmoreamorphosyntacticdifficultyratherthanapragmaticdifficulty.Itcouldbeusedtoavoid theproductionofmorecomplexstructuresandthereforeenablethechildrentoexpressthereferenceatalessercognitivecost(DeCat,2004). Innarrativecomposition,onlyonestudyhasyetcodedpunctuationinwrittenproductsofchildrenwithSLI.BishopandClarkson(2003) askedparticipantsaged7.5to13yearstocomposenarrativesonthebasisoffivepictures.Aglobalscoreofintelligibilitywitha4-point scaleassessedaccuracyofthewrittenoutputintermsofspellingandpunctuation.Resultsshowedthatthisintelligibilityindexwasthemost sensitiveindicatoroflanguageimpairment,with16ofthe28childrenscoringmorethan1SDbelowthecontrolmean.Butpunctuationand spellingcarryverydifferentfunctionsinwrittendiscourseproduction.So,onthebasisofthisglobalintelligibilityindex,wecannotdraw anyconclusionaboutthemanagementofthelonesystemofpunctuationasaspecificindicatorofcohesion.Thesepreviousresults,however, stillindicatetheimpactoforaldifficultiesonnarrativecomposition.TheyevidencethelowerperformanceofchildrenwithSLIcomparedto typicallydevelopingchronological-agepeersinthisactivityandtheirlimitsintextgenerationprocessing.

LittleisknownhoweveraboutthedevelopmentofabilitiesofparticipantswithSLItomanagecohesioninwrittencomposition.Onlyone studyanalyzedthisaspectthroughchildhoodandadolescence.Dockrelletal.(2009)longitudinallyexaminedthewritingperformanceof58 studentsfrom8to16yearsold.Theyconcludedthat,untiltheageof16years,participantscontinuedtoexperienceproblemswithliteracy. Moreover,students’writingskills,suchashandwritingfluency,sentencestructure,andspelling,decreasedrelativetostandardizednorms fromtheageof11years.AccordingtoDockrelletal.(2009),fortypicallydevelopingchildren,increasinglanguageandliteracyskillscan supportlaterdevelopmentofwriting;conversely,forthosewithpersistentdifficulties,suchresourcescannotbeavailable.

3. Thepresentstudy:goalsandhypotheses

ThepresentstudyinvestigatedthemanagementofthethreesystemsofcohesiondevicesinparticipantswithSLIwhenwritinganarrative. Thestudywasconductedintwoagegroups:childrenandadolescents.TheperformanceoftheparticipantswithSLIwascomparedtotheone ofchronologicallyage-matchedtypicallydeveloping(TD)students.Allparticipantsattendedordinaryclassesandweretaughtaccordingto instructionsoftheFrenchcurriculum.Childrenattendedelementaryschool,andadolescentsattendedmiddleandhighschools.Wecompared theperformanceofthegroupwithSLItotheperformanceoftheTDgroupateachageaswellastheperformanceofthetwoagegroupswithin eachtypeofgroup.

WeexpectedparticipantswithSLItoencounterdifficultiesinthemanagementofeachsystemofcohesiondevicescomparedtotheirTD peers.Weassumedthattheirhighlevelofliteracyimpairment(Freedetal.,2011)andtheirrecurrentdifficultiesinlinguisticcoding(Dockrell &Lindsay,2000)wouldaffecttheirgenerationofcohesiondevices.Accordingly,theircapacitytochoosesufficientandadequatemarksto expresstextualrelationshipsshouldbelowerthanthatoftheirTDpeers.Inthatway,weexpectedstudentswithSLItousebothlessnumerous andlessdifferentiatedconnectivesthanTDstudents,andthattheywoulduseahigherdensityofandand/orlessdifferentiatedconnectives (seeFavart&Passerault,1995).ThispredictionalsoreliedonoralskillsofparticipantswithSLI,whoshowarecurrentuseoftheconnective andaswellasalackindiversificationcomparedtoTDchildren(DeWeck&Rosat,2003;Hilaire-Debove&Roch,2012).

Asregardspunctuationmarks,weexpectedtheperiodstobeusedpredominantlyassoonaschildrenwithSLIcouldimplementthe punctuationsystemintheircomposition.Themaindifferencebetweenthetwogroupsmightpreferentiallyaffecttheuseofthecomma, whichonlydevelopsattheendofelementaryschoolinTDstudents’narrativetexts(Favart&Passerault,2000).

Asregardsreferentialexpressions,basedonoralproduction,weexpectedparticipantswithSLItomainlyusenominalrepetitions( Hilaire-Debove&Roch,2012).UnlikeTDparticipants,weexpectedthemtoalsousespecificmentionssuchasleftdislocations(DeWeck&Jullien,2013; Jullien,2008).Moreover,thedifferencesobservedduringchildhoodbetweenparticipantswithSLIandTDparticipantsshouldcarryonandeven belargerinadolescents.AccordingtoDockrelletal.(2009),difficultiesofstudentswithSLIinlanguagegenerationandsentencestructuring decreaserelativetostandardizednormsfromtheageof11years.InagreementwithBerningerandSwanson’s(1994)developmentalmodel ofcompositionandwithMcCutchen’s(1996)capacitytheoryofwriting,Dockrelletal.(2009)explainedthatthepersistentdifficultiesof adolescentswithSLIinlanguageandliteracyskillscannotsavesufficientresourcestosupporttheirwritingdevelopment.IntheTDgroup,we expectedtoobservebetterperformancesineachaspectofwrittencohesionabilitiesinadolescentsthaninchildren.Anaccuratemanagement ofthethreesystemsofcohesiondevicescanworkinnarrativewritingattheendofelementaryschool.TheTDadolescentsallattendedmiddle andhighschools,sotheyshouldbeabletodiversifymorethanchildrenthemarkingofnarrativerelationsbyconnectivesandanaphors.A regularuseofbothpunctuationmarksshouldbeeffectiveonlyinthisgroup.

Table1

Numbersofparticipants,meanages(standarddeviations)andgenderbreakdownsinchildrenandadolescentswithSLIandintypicallydevelopingchildren andadolescents(TD).

Participants N Meanage(SD) Range Gender

SLI

Children 12 8.94(1.12) 7.66to11.00 10boys/2girls

Adolescents 12 14.33(1.77) 12.08to17.50 7boys/5girls

TD

Children 24 8.85(1.07) 7.50to11.16 20boys/4girls

Adolescents 24 14.05(1.41) 12.25to17.68 14boys/10girls

Tosumup,thefirstaimwastoinvestigatethemanagementofcohesioninchildrenandadolescentswithSLIbycomparingtheir

perfor-mancetothatofage-matchedTDstudents.Weexpectedthemanagementofconnectives,punctuationmarks,andanaphorstobeweakerin

participantswithSLIthaninTDparticipants,inbothagegroupsandatbothlevelsofdensityanddiversity.Thesecondaimwastocompare,in

bothSLIandTDgroups,themanagementofcohesioninchildrenandadolescents.Weexpectedaweakermanagementofcohesioninchildren

thaninadolescents.Moreover,thedifferencesobservedduringchildhoodbetweenSLIandTDparticipantsshouldcarryonandevenbelarger

inadolescents.

4. Method

4.1. Participants

Thesampleincludedatotalof24participantswithSLI.Twelvewerechildren(10boysand2girls)agedbetween7and11yearsold,and

12wereadolescents(7boysand5girls)agedbetween12and18yearsold.

ChildrenandadolescentswithSLIhadbeendiagnosedbytheCentreRéférentdesTroublesduLangage(ReferralCenterforLanguage

Disorders)oftheHenriLaboritHospitalinPoitiers,France.Thediagnosiswasbasedonamedicalassessmentaswellasneuropsychological

andpsycholinguistictesting.ParticipantswithSLIattestedadiscrepancybetweentheirleveloforallanguageandnonverbalabilities.

TheinclusioncriterionoftheseparticipantswithSLIwasascoreofatleast1.25SDbelowthemeanscoresonthefollowingstandardized

languagetests:

•For7-year-oldparticipants,theBilanInformatiséduLangageOralaucycle2(BILO–2,[ComputerizedAssessmentofOralLanguagefor5–7

yearsoldchildren],Khomsi,Khomsi,&Pasquet,2007)andtheNouvellesEpreuvespourlExamenduLangage(N-EEL,[NewTasksforthe

LanguageAssessment],Chevrie-Muller&Plazza,2001).

•Forparticipantsattheageof8yearsandolder,theBilanInformatiséduLangageOral3etauCollège(BILO–3C,[ComputerizedAssessmentof OralLanguagefor8–10yearsoldchildren,andatmiddleschool],Khomsi,Khomsi,Pasquet,&ParbeauGuéno,2007)andtheBatterieLangage OraletÉcrit,Mémoire,Attention(L2MA,[TestOralandWrittenLanguage,Memory,Attention],Chevrie-Muller,Simon,&Fournier,1997)

ThelanguagecomprehensionabilitiesoftheparticipantswithSLIfellwithinnormallimits,andtheyshowednocognitiveimpairment, withascoregreaterthan80ontheWechslerIntelligenceScaleforChildren,3rdEdition(WISC–III,Wechsler,1996).

TheparticipantswithSLIwerematchedtoagroupoftypicallydeveloping(TD)participants,accordingtoage,gender,andtheparental socialeconomicstatus.Twenty-fourtypicallydevelopingchildrenwerematchedtothechildrenwithSLI(agerange:7–11)and24typically developingadolescentswerematchedtotheadolescentswithSLI(agedrange:12–18).ItwasnotpossibletomatchparticipantswithSLIwith TDchildrenonthebasisoflanguageagebecausesuchTDchildrenwouldhavebeentooyoungtoperformthewritingtask.

BothTDparticipantsandparticipantswithSLIwereFrenchnativespeakersandattendedmainstreamclassesinandaroundPoitiers, France.Thechildrenattendedprimaryschool,andtheadolescentsattendedmiddleandhighschools.Noneofthemwassufferingfromany neurological,sensory,relational,orscholasticdisorder.Thenumberofparticipants,theirmeanages,andgenderproportionsareshownin Table1.

4.2. Taskandprocedure

Thewritingtaskconsistedofanarrativecomposition.ItwasinspiredbyBerman(2005),Berman,Ragnarsdóttir,andStrömqvist(2002), BermanandVerhoeven(2002)andhasbeenusedbyBrocetal.(2013)andbyBroc,Bernicot,Olive,Favart,andReilly(2014).Theexperimenter firstintroducedherselfasauniversitystudentcarryingoutaresearchprojecttocollectstoriesonthetopicofcontentioussituationsatschool. TheinstructionswereprovidedinFrench;forthisarticle,wehavetranslateditasfollows:“I’mauniversitystudentandIhavetobringaproject tocompletion.Tothisend,Ihavetocollectstoriesonthetopicofcontentioussituationsatschool.I’vealreadycollectedafewstories,butIstillneed somemore.Couldyouhelpme?Rightnow,I’mlookingforstoriesabouttheftorfightsituationsatprimary/middle/highschools(accordingtothe participant’sgradelevel).Iassumeyou’vealreadyhadtodealwithsuchasituationatprimary/middle/highschool.Socouldyoutellmethatstory? I’dlikeyoutowriteitasawhole,insuchawaythatIcanreallyunderstandwhathappened.Ifyoumakeanyerrororifyouwishtochangeanything, justcrossitoutandkeeponwriting.”

Eachparticipantthenwrotethestoryindividuallyandinacommunicativecontext.Thisdualsituationwasunfamiliaratschoolgiven thateachparticipantwasinstructedtocomposeastoryaddressedtotheexperimenterwhowasphysicallypresentthroughoutthewriting process.Notimelimitwasfixed.Afterareflectiontime,allparticipantsbegantowritespontaneously,andthewritingtasklastednomorethan 15min.Nootherinstructionsweregivenasregardslength,writingspeed,orcohesion.Oncethetexthadbeencomposed,theexperimenter askedtheparticipanttorereadhis/hertexttoavoidanyambiguityinthereadingofthewrittenwords.

Table2

Medians(andstandarddeviations)ofdensityanddiversityofconnectivesusedinchildrenandadolescentswithSLIandintypicallydevelopingchildren andadolescents(TD).

Participants

SLI TD

Children Adolescents Children Adolescents

Densityofconnectives 3.23(4.92) 8.48(4.07) 6.40(3.42) 6.61(3.02)

Diversityofconnectives 1(0.90) 2(1.78) 1(1.14) 2.5(1.44)

4.3. Textanalysis

Thenarrativeswerefaithfullytyped,complyingwithfeaturesoftheoriginalproducts:lower/uppercases,crossedoutwords,spelling

errors,andpunctuationmarksused.Participantswhowroteatextincludinglessthanoneclause,i.e.atotallackofverb,wereremoved

fromthesubsequentstatisticalanalysis.ThiswasthecasefortwochildrenwithSLI,whohadgreatdifficultiestoperformthewritingtask.

Therefore,thefourTDchildrenwhowerematchedwiththesetwoparticipantswerealsoremovedfromtheanalysis.

Analysisofthetextfirstfocusedonnarrativelength,measuredintermsofthetotalnumberofwordsproduced.Then,theanalysisof

cohesionfocusedonthedensityanddiversityofcohesiondevices.Densitywasexpressedasthenumberofeachmarkuseddividedbythe

totalnumberofwordsinthetext,multipliedby100.Diversitywasassessedbythenumberofdifferentconnectivesoranaphorsused.As

regardspunctuationmarks,onlyperiodsandcommaswereanalyzed.Onthisbasis,thedependentvariableswereDensityandDiversityof

Connectives,DensityofPeriodsandCommas,andDensityandDiversityofAnaphors.

Tocompletetheanalysesofdiversity,qualitativeanalyseswereperformedforconnectivesandanaphors,whichweresortedintocategories.

Connectiveswereclassifiedintolessversusmoredifferentiatedconnectives.SubclassesweredrawnuponthebasisofFavartandPasserault

(1999)semanticcategorizationofconnectives,specifictowritingacquisition.Whendrawingupthesecategories,wesometimesneeded tospecifytheliteralformusedinFrench,duetothedifficultytomatchEnglishandFrenchmeaningsoftheselinguisticdevices.Theless differentiatedconnectivesincludedtheandconnectiveandchronologicalonessuchasthen(inFrench,puis,après,ensuite,etpuis,etaprès). Themoredifferentiatedconnectivesincludedthosethatexpressedrelationshipsspecifictothenarrativeframe:theadversativeconnective but;temporalconnectivessuchasthen(inFrench,alors),when,andsuddenly;causalconnectivessuchasbecauseandfor;consequence connectivessuchastherefore;andgoalconnectivessuchasforandinorderto.Wealsocodedspatialconnectives(e.g.,atthetop,atthebottom, on)asdescriptivesequencescouldoccurwithinthestories.Anaphorswerecodedaccordingthefollowingcategories:nominalrepetitions(la surveillante/[thesupervisor]),nominalsubstitutions(unebandedejeunes/[agroupofyoungpeople]wassubstitutedbycespersonnes/[these persons]),firstnameorpropernamerepetitions,personalpronouns(distinguishingsubject,elle/[she];orobjectgrammaticalfunctions, lui/[her]),andrelativepronouns(qui/[who]).

5. Results

StatisticalanalyseswereperformedwithStatistica-7software.Fortextlength,werananANOVAwithgroup(SLIvs.TD)andage(7–11 vs.12–18)asbetween-participantsfactors.Duetonon-normalityofthedistributionsandunequalvariancebetweengroups,nonparametric testswereperformedonmediansofdensitiesanddiversitiesusingMann-Whitneycomparisonteststocompare(a)childrenwithadolescents withineachgroupand(b)SLIwithTDgroupsateachage.

5.1. Textlength

TextlengthinnumberofproducedwordssignificantlydifferedbetweenthegroupwithSLIandtheTDgroup,F(1,62)=6.05,p=0.016, ␩2_{=0.09:ThenarrativeswrittenbyparticipantswithSLI(M=33.23;SD=32.39)wereshorterthanthosewrittenbyTDparticipants(M=53.25;}

SD=35.09).Theeffectofagealsowassignificant,F(1,62)=15.17,p=0.0002,␩2_{=0.20:Thenarrativeswrittenbychildren(M=28.87;SD=20.52)}

wereshorterthanthoseofadolescents(M=61.33;SD=38.33).TheGroupxAgeinteractionwasnotsignificant,F(1,62)<1.

5.2. Connectives

5.2.1. Effectsofageandgroup

ThedensityanddiversityofconnectivesareshowninTable2.IntheSLIgroup,thedensitysignificantlyincreasedwithage,U=25,Z=2.3, p=0.02,r=0.49,aswellasthediversity,U=19,Z=2.8,p=0.005,r=0.59.IntheTDgroup,theconnectivesusedbyadolescentsalsoweremore diversifiedthanthoseusedbychildren,U=123.5,Z=2.8,p=0.005,r=0.42,buttheywerenotmorefrequent,U=232.5,Z=0.18,p=0.85.The differencebetweenthegroupwithSLIandtheTDgroupwassignificantonlyinchildren.TheconnectivesusedbythegroupwithSLIwere significantlylessfrequent,U=56,Z=1.97,p=0.04,r=0.36,andalsolessdiversified,U=57,Z=2,p=0.04,r=0.36,thantheconnectivesusedby theTDgroup.NosignificantdifferencewasnoticedbetweenscoresofadolescentswithSLIandTDadolescents.

5.2.2. Qualitativeanalysis

Thequalitativeanalysisofthediversityofconnectiveswasconductedintwosteps.Wefirststudiedhowmanycategoriesofconnectives participantsusedand,second,whichtypesofcategorieswereused.ThedependentvariableNumberofCategorieswassortedintothreelevels. Thefirstlevelidentifiedthepercentageofparticipantswhousednoconnectives,thesecondlevelthepercentageofparticipantswhoused oneortwocategoriesinwhichtheconnectiveswereusedonlytoconcatenateideasandthethirdlevelthepercentageofparticipantswho usedatleastthreecategories,i.e.withmorediversifiednarrativerelationships.TheresultsareshowninFig.1.Themajorityofchildrenwith SLI(60%)didnotuseanyconnective,whilethatwasonlythecasefor15%oftheTDchildren.Moreover,childrenwithSLIneverusedmore thantwocategoriesofconnectives(40%vs.70%ofTDchildrenwhousedoneortwocategories).Finally,15%ofTDchildrenusedatleast

Fig.1. PercentagesofSLIandTDchildrenandadolescentsusingno,oneortwo,oratleastthreedifferentcategoriesofconnectives.

Fig.2.Percentagesofand,chronological,andmorediversifiedconnectives(onthetotalnumberofconnectives)usedbySLIandTDchildrenandadolescents.

threecategoriesofconnectives.Unlikechildren,thepatternofdiversificationofconnectiveswascomparableinadolescentswithSLIandTD adolescents:Only8%ofadolescentswithSLIusednoconnectives,whichwasthesamepercentageasinTDadolescents,50%ofadolescents withSLIusedoneortwocategories(vs.42%inTDadolescents),and42%ofadolescentswithSLIusedthreeormorecategories(vs.50%inTD adolescents).

ThedependentvariableVarietyofConnectivesthenwassortedaccordingtothreecategories:and,chronological,andmorediversified connectives.Fig.2showsthepercentagesofeachvarietyofconnectivesonthetotalofconnectivesusedbychildrenandadolescents.We calculatedthepercentagesofeachtypeofcategoryinrelationtothetotalofconnectivesused.

Inchildren,theparticipantswithSLIonlyusedtheandconnectiveandchronologicalconnectives.Theandconnectivewasmainlyusedin bothgroups(70%and61.7%ofconnectivesinparticipantswithSLIandTDparticipants,respectively).Thechronologicalconnectiveswere30% oftheconnectivesusedbythechildrenwithSLI,and8%oftheconnectivesusedbytheTDchildren.OnlyTDchildrenusedmorediversified connectives(30%ofthetotalofconnectives)whichwerenotusedatallbychildrenwithSLI.FocusingontheusebyTDchildrenofthislatter category,butwasmainlyused(11%ofconnectives),thenthetemporal(8%)andcausal(6%)connectives.Finally,alorsandgoalconnectives amountedto2%oftheusedconnectives.

Inbothgroupsofadolescents,theuseofconnectiveswascomparableastheywerequiteequallysplitintwocategories:and,whichwasless frequentinadolescentsthaninchildrenbutstillfrequentlyused(49%inbothSLIandTDgroups).Themorediversifiedconnectivesalsowere quitefrequent(51%ofthetotalinadolescentswithSLIand45%inTDadolescents).Only6%ofTDparticipantsusedchronologicalconnectives, whileparticipantswithSLIusednone.

Aspartofthemorediversifiedconnectivescategory,adolescentswithSLIpreferentiallyusedcausalconnectives(15%),thenbutand temporalconnectives(9%),goalconnectivesandalors/then(6%),andfinallyspatialconnectives(4%).ThepatterninTDadolescentswasquite different,withcloseproportionsofuseforcausalconnectives(13%)andbut(11%),followedbyalors(7%),goalconnectives(6%)andtemporal connectives(6%),andfinallyaveryweakuseofspatialconnectives(1%).

5.3. Punctuationmarks

Onlydensitiesofperiodsandcommaswerecoded.Table3showsscoresateachageandineachgroup.IntheTDgroup,thedensity ofcommausesignificantlyincreasedwithage,U=118,Z=3,p=0.004,r=0.45,whereasthedensityofperioduseremainedstable,U=181.5, Z=1.4,p=0.16,r=0.21.NosignificantimprovementwasnoticedforpunctuationintheSLIgroupbetweenchildhoodandadolescence,U=47.5, Z=0.8,p=0.42forperiodandU=45,Z=1.3,p=.35forcomma.

ThedifferencebetweenthegroupwithSLIandtheTDgroupwasnotsignificantinchildren,U=90,Z=0.44,p=0.65forperiodandU=83.5, Z=1,p=0.47forcomma.Inadolescents,thedensityofcommausewassignificantlyhigherinTDparticipants,U=77.0,Z=2.3,p=0.02,r=0.38, whereasnodifferencewasnotedforperioduse,U=128.5,Z=0.5,p=0.61.

Table3

Medians(andstandarddeviations)ofdensityofperiodsandcommasusedinchildrenandadolescentswithSLIandintypicallydevelopingchildrenand adolescents(TD).

Participants

SLI TD

Children Adolescents Children Adolescents

Densityofperiod 3.33(6.61) 5.67(3.41) 4.65(3.61) 6.50(3.44)

Densityofcomma 0(0.92) 0(2.60) 0(4.81) 2.95(3.18)

Table4

Medians(andstandarddeviations)ofdensityanddiversityofanaphorsusedinchildrenandadolescentswithSLIandintypicallydevelopingchildrenand adolescents(TD).

Participants

SLI TD

Children Adolescents Children Adolescents

Densityofanaphors 5.88(4.11) 9.03(6.02) 8.33(6.73) 8.22(6.28)

Diversityofanaphors 0(.95) 2(1.48) 1(1.04) 3(1.22)

Fig.3.PercentagesofSLIandTDchildrenandadolescentsusingno,oneortwo,oratleastthreedifferentcategoriesofanaphors.

5.4. Anaphors

Anaphorswerecodedintothefollowingcategories:nominalrepetition,nominalsubstitution,firstnameorpropernamerepetition,

personalpronoun(distinguishingsubjectorobjectgrammaticalfunctions),andrelativepronoun.

5.4.1. Effectsofageandgroup

ThedensityanddiversityofanaphorsareshowninTable4.IntheSLIgroup,thedensityofanaphorssignificantlyincreasedwithage,

U=29,Z=2,p=0.04,r=0.42,unlikethediversity,U=31,Z=1.9,p=0.059.Conversely,intheTDgroup,diversitysignificantlyincreasedwith age,U=113,Z=3.1,p=0.002,r=0.47,butnotdensity,U=221.5,Z=0.4,p=0.66.Itisimportanttonoticethatanaphorswereverypoorly diversifiedinchildrenwithSLI.

ThedifferencebetweenthegroupwithSLIandtheTDgroupwassignificantonlyinchildren.TheparticipantswithSLIusedlessfrequent, U=47.5,Z=2.4,p=0.02,r=0.44,andalsolessdiversified,U=53.5,Z=2.2,p=0.03,r=0.40,anaphorsthantheTDgroup.Conversely,in ado-lescents,thedensityanddiversityofanaphorswerenothigherinTDparticipantsthaninparticipantswithSLI,U=131.5,Z=0.4,p=0.68and U=98,Z=1.6,p=0.13,respectively.

5.4.2. Qualitativeanalysis

Thequalitativeanalysisfocusedonthediversityofanaphors.Asforconnectives,theanalysiswasconductedaccordingtothenumber ofcategoriesusedandthetypesofcategoriesused.ThedependentvariableNumberofCategoriesalsowassortedintothreelevels:the percentageofparticipantsusingnoanaphors,usingoneortwocategories,andatleastthreecategories.Resultsoftheseanalysesinchildren andadolescentsareshowninFig.3.Noanaphorswereusedby60%ofchildrenwithSLIversus25%ofTDchildren.MostoftheTDchildrenused oneortwotypesofanaphors(60%)ascomparedto20%ofchildrenwithSLI.Finally,20%ofchildrenwithSLIand15%ofTDchildrenusedthree categoriesormore.Inadolescents,contrarytoconnectives,thequalitativepatternsofanaphorsstayedrelativelydifferentinparticipantswith SLIandTDparticipants:33%ofadolescentswithSLIstilldidnotuseanyanaphors(vs.8%intheTDgroup).Theintermediatescore(oneor twotypes)concerned42%ofadolescentswithSLIand29%ofTDadolescentsandonly25%ofadolescentswithSLIusedatleastthreetypesof anaphors(vs.63%ofTDadolescents).

Next,weanalyzedthedifferenttypesofanaphorsusedbyparticipantsbycalculatingthepercentagesofeachtypeofanaphorinrelation tothetotalofanaphorsused(seeTable5).InchildrenwithSLIonlypronouns(personalandrelative)wereusedwhereasinTDchildrenthe distributionwasmuchmorediversified.TheuseofnominalanaphorswasobservedonlyinTDchildrenandinterestingly,thesechildrenused relativelyfewrelativepronouns.ThescoresofrelativepronounsinchildrenwithSLI(45.45%vs.11.94%,inTDchildren)mightseemquite

Table5

PercentagesofcategoriesofanaphorsusedinchildrenandadolescentswithSLIandintypicallydevelopingchildrenandadolescents(TD).

Participants

SLI TD

Children Adolescents Children Adolescents

Nominalrepetitions – 18.52% 5.97% 6.25%

Nominalsubstitution – 14.81% 7.46% 17.5%

Namerepetition – – 22.39% 11.25%

Subjectpersonalpronoun 27.27% 35.18% 37.31% 28.12%

Objectpersonalpronoun 27.27% 12.96% 14.92% 21.25%

Relative 45.45% 18.51% 11.94% 15.63%

surprising.Amorefine-grainedinspectionshowed,however,thattherelativepronounqui/whowasusedby80%ofchildrenwithSLIaspart

ofleftdislocations,suchas“c’est qui”/“itis who”

Inadolescence,thepatternofanaphorswascomparablebetweentheSLIandTDparticipants.Personalandrelativepronounswereused

atasimilarfrequencybyparticipantswithSLIandbyTDparticipants.Moreover,wedidnotnoticeanyoccurrenceofrelativepronounas

leftdislocationinadolescentswithSLI.Asregardsnominalanaphors,onlytheTDadolescentsusedfirstnamerepetitions;theproportionof

nominalrepetitionswasslightlyhigherinadolescentswithSLIthaninTDadolescents;butthetwogroupsofadolescentsusedatasimilar

ratenominalsubstitutions.

6. Discussion

ThepresentstudyinvestigatedhowchildrenandadolescentswithSLImanagecohesioninwrittennarrativesbycomparingtheparticipants

withSLItochronologicalage-matched,typicallydevelopingpeers.Participantsallattendedmainstreamclasses:childreninelementaryschools

andadolescentsinmiddleandjuniorhighschools.Specifically,wequantitativelyandqualitativelyanalysedconnectives,punctuationmarks,

andanaphorsinwrittennarratives.Cohesiondevicesarecrucialindicatorsofthewriter’scapacitytoconsidertheoverallmeaningofthetext.

Asthesedeviceslinkideasinatext,theyserveaskeyprocessinginstructionstoimprovetheaddressee’scomprehension,sotheiruseisin

directrelationtopragmaticabilities.Onthisbasis,wechoseacommunicativenarrativesituation,whichpragmaticallyengagedparticipants

inthetasksothattheycouldtakeintoaccounttheaddresseeasmuchaspossible.Thissituationseemedappropriatetosupportlinguistic

codingattheleveloftextgeneration,whichisparticularlydeficientinparticipantswithSLI(Dockrelletal.,2009).

Thecoreofthisresearchwastocomparethescoresoffourdifferentgroups:childrenwithSLI,adolescentswithSLI,TDchildrenandTD adolescents.Toanswertothefirstaim,wecomparedperformancesofparticipantswithSLIandofTDparticipants.First,textswereshorter inthetwogroupsofparticipantswithSLI(childrenandadolescents)comparedtotheirpeers,attestingtotheiroveralldifficultieswithtext generation.Inspiteoftheirwell-establishedmanagementofperiods,thechildrenwithSLIfailedintheiruseofbothconnectivesandanaphors: bothdensityanddiversitywereweakerthaninTDchildren.MostofthechildrenwithSLIdidnotuseaconnectiveand/oranaphor,noneof themusedspecificconnectives,andtheirmarkingofnarrativerelationshipswasrestrictedtotheundifferentiatedandand/orchronological connectives.Bycontrast,almostathirdoftheTDchildrenexplicitlyexpressedadversative,temporal,orcausalrelationshipsspecificto narratives.ThewaychildrenwithSLImanagedcohesioninthiscommunicativewritingsituationwascomparabletothemanagementof cohesionattestedbystudiesonnarrativespeech,withalackofdifferentiationofconnectives(Bernard-Barrot&Géhard,2003),seriesof sentencesbeingcommonlylinkedwiththeconnectiveand(DeWeck&Rosat,2003),andnoimprovementinconnectivesdiversification betweentheagesof7and12yearsold(Hilaire-Debove&Roch,2012).

Asregardsanaphors,previousresultsinnarrativespeechshowedthatchildrenwithSLIpersistentlyusenominalrepetitions( Hilaire-Debove&Roch,2012).Unexpectedly,inwriting,theyonlyusedpronouns,halfrelativeandhalfpersonal,whereasTDchildrenusedabouthalf variednominalreferencesandhalfpronouns.Nevertheless,amorefine-grainedanalysisindicatedthatahighmajorityofrelativepronouns (80%)wereusedwithinleftdislocations.Thesespecificstructureswereobservedinpreviousstudiesinvestigatingtheoralproductionof childrenwithSLI(DeWeck&Jullien,2013;Jullien,2008)andtheyarethesignsofanimmaturereferentialmanagement.Accordingto Leonanduzzi(2008),therelativepronounisinthatcaseincludedinasplitsentenceinordertoputtheemphasisonthereferent.Theaimand thefunctionofsuchformsaremainlypragmaticandtheyareusedtostressthefocusonthereferent.Ourresultsshowcomparableoutputs inwriting,perhapsbecauseofthecommunicativesituationwesetup.

Thesecondaimwastocomparethetwoagegroups(childrenandadolescents)inparticipantswithSLIandinTDparticipants.Children producedshorternarrativesthanadolescentsandthedifferencewascomparableinthegroupwithSLIandintheTDgroup.Moreover,the densitiesofconnectivesandanaphorswerehigherinadolescentswithSLIthaninchildrenwithSLI,asamajorityofthelatterdidnotuseany connectivesand/oranaphors.Diversityisindeedmorerelevantthandensitytoassesscohesion(seeBernard-Barrot&Géhard,2003orFavart &Passerault,1995forconnectives;Decool-MercierandAkinci(2010)orKarmiloff-Smith,1985,1992foranaphors).

Resultsondiversitywereverydifferentcomparedtoresultsondensity.Asregardsconnectives,adolescentsusedmorediversified con-nectivesthanchildreninboththegroupwithSLIandtheTDgroup.Bycontrast,anaphorsandpunctuationmarksweremorediversifiedin adolescentsintheTDgrouponly.MostoftheTDadolescentsusedatleastthreedifferenttypesofanaphors,andthedensityofcommasalso washigherinthisgroupcomparedtoTDchildren.Asinpreviousstudies(Favart&Passerault,2000),theperiodwaspredominantlyusedand stronglyimplementedfromthebeginningofnarrativewriting,whiletheimplementationofthecommawaspostponedandonlydeveloped inTDadolescents.InlinewithSchneuwly(1988),onlyTDadolescentswereabletomanagethepunctuationmarksatatextualleveland diversifiedthestrengthofbreakingsbetweentextualunitsinordertoappropriatelyclarifythetextualcontentfortheaddressee.

ThemostsurprisingandnoteworthyresultofthisstudyreferredtotheperformanceofadolescentswithSLI.Whereastheuseofconnectives andanaphorswasimpairedinchildrenwithSLI,theirusebyadolescentswithSLIwasclosetotheoneoftheirTDpeers:Theyusedasmany connectivesandanaphors,whichwereasmanydiversified.ThisoutcomediffersfromDockrelletal.’sstudy(2009),inwhichhugedifficulties inwritingnarrativeremainedandevenincreasedforparticipantswithSLIthroughoutadolescencerelativetostandardizednorms.

OnepossibleexplanationfortherelativelygoodperformanceinthemanagementofcohesionofoursampleofadolescentswithSLImight lieinthespecificinstructionweprovidedforthenarrativecompositionandinthecommunicativeinteractionwesetup.Indeed,because

individualswithSLIhavepragmaticabilities(Katsos,Roqueta,ClementeEstevan,&Cummins,2011;Wetherell,Botting,&Conti-Ramsden, 2007),theymayhavetakenadvantageofthistypeofsituation.PreviousworkbyBrocetal.(2013,2014)demonstratedthatspellingerrors arelessdominantinchildrenandadolescentswithSLIinacommunicativecontextthaninamoreevaluativesituation(i.e.,dictation).In comparison,Dockrelletal.(2009)setupamoreregularwritingsituation,astheirparticipantswereinstructedtowritealetterdescribing theiridealhouse.Theabilitytotakeintoaccounttheaddresseeiscrucialforanaccuratemanagementofcohesion,butthisisadifficulttaskfor youngwritersinsofarasitrequiresthecapacitytoactivatetherepresentationoftheaddresseeinworkingmemory.AccordingtoMcCutchen (1986)andMartlew(1983),takingintoaccounttheaddresseecanonlybeeffectiveinTDstudentsattheageofmiddleschool.Thegood performanceoftheadolescentswithSLImayindicatethattheyalsoindeedtookintoaccounttheaddressee.InlinewithBrocetal.(2013, 2014),acommunicativenarrationcouldenableindividualswithSLItofullyexpresstheirpotential.Thisisonlypossibleinsofarasstudentscan takeintoaccounttheaddresseeinwriting,andthiscapacityseemedtobeeffectiveinbothmiddleschoolstudentswithSLIandTDstudents. Thisfindingisworthtakingintoaccount,inparticularineducationaland/orremediationcontexts,asitsuggeststhatwhenchildrenwithSLI areprovidedwithappropriatesupport,theirwritingabilitiesmayimproveuntiladolescence.

Insum,onlythesystemofpunctuationmarkswasaffectedinbothSLIchildrenandadolescents.Thisiscertainlyduetothebelated developmentofthiscohesionsystem,whosetextualmanagementinTDstudentsisnotcompletelyreachedbeforetheageof14yearsold (Schneuwly,1988).Exceptforpunctuation,frommiddleschool,studentswithSLIwereabletomanagecohesionaswellasTDstudents,in spiteoftheirrepeatedorallanguagedisorders.WeagreewithDockrelletal.(2009)whointerpretthenatureofthewritingdifficultiesin termsoflackoflanguageandliteracyresources,andourresearchfurtherindicatesthatparticipantswithSLIcandealwithsomeaspectsof writingwhenpragmaticcuesareprovidedinthewritingsituation.

References

Apotheloz,D.(1989).Aspectscognitifsdesprocéduresdelacohésiontextuelle[Cognitiveaspectsoftextualcohesion].Duisburg:L.A.U.D.

Bartlett,E.J.,&Scribner,S.(1982).Textandcontent:Aninvestigationofreferentialorganizationinchildren’swrittennarratives.InC.H.Frederiksen,&J. F.Dominic(Eds.),Writing:Thenature,development,andteachingofwrittencommunication,vol.2:writing:Process,developmentandcommunication(pp. 153–168).Hillsdale,NJ:LawrenceErlbaumAssociates.

Berman,R.A.,&Verhoeven,L.(2002).Crosslinguisticperspectivesondevelopingtext-productionabilitiesinspeechandwriting.WrittenLanguageand Literacy,5,1–44.http://dx.doi.org/10.1075/wll.5.1

Berman,R.A.,Ragnarsdóttir,H.,&Strömqvist,S.(2002).Discoursestance.WrittenLanguageandLiteracy,5,255–290.

http://dx.doi.org/10.1075/wll.5.2.06ber

Berman,R.A.(2005).Introduction:Developingdiscoursestanceindifferenttexttypesandlanguages.JournalofPragmatics,37,105–124.

http://dx.doi.org/10.1016/j.pragma.2004.08.003

Bernard-Barrot,C.,&Géhard,S.(2003).Lerécitoral:Comparaisond’enfantsprésentantunedysphasieetd’enfantssanstroublesdulangageoral.Mesures linguistiquesetnarratives[Oralnarratives:Comparisonofchildrenwithdysphasiaandofchildrenwithoutlanguagedifficulties.Linguisticandnarrative measures].Lyon,France:Mémoired’orthophonie.N◦_1216.

Berninger,V.W.,&Swanson,H.L.(1994).ModifyingHayesandFlower’smodelofskilledwritingtoexplainbeginninganddevelopingwriting.InJ.S. Carlson,&E.C.Butterfly(Eds.),Advancesincognitionandeducationalpractice.vol.2:children’swriting:Towardaprocesstheoryofthedevelopmentof skilledwriting(pp.57–81).Greenwich,CT:J.A.I.Press.

Bishop,D.V.,&Clarkson,B.(2003).WrittenLanguageasawindowintoresiduallanguagedeficits:Astudyofchildrenwithpersistentandresidualspeech andlanguageimpairments.Cortex,39,215–237.http://dx.doi.org/10.1016/s0010-9452(08)70106-0

Bishop,D.V.M.(1992a).Theunderlyingnatureofspecificlanguageimpairment.JournalofChildPsychologyandPsychiatryandAlliedDisciplines,33,3–66.

http://dx.doi.org/10.1111/j.1469-7610.1992.tb00858.x

Bishop,D.V.M.(1992b).Theunderlyingnatureofspecificlanguageimpairment.JournalofChildPsychologyandChildPsychiatry,3,1–64.

http://dx.doi.org/10.1111/j.1469-7610.1992.tb00858.x

Broc,L.,Bernicot,J.,Olive,T.,Favart,M.,Quémart,P.,Reilly,J.,etal.(2013).LexicalspellinginchildrenandadolescentswithSpecificLanguageImpairment: Variationsindifferentwritingsituations.ResearchinDevelopmentalDisabilities,34,3253–3266.http://dx.doi.org/10.1016/j.ridd.2013.06.025

Broc,L.,Bernicot,J.,Olive,T.,Favart,M.,Reilly,J.,Quémart,P.,etal.(2014).Évaluationdel’orthographedesélèvesdysphasiquesensituationdenarration communicative:Variationsselonletyped’orthographe,lexicaleversusmorphologique[Assessmentoflexicalandmorphologicalspellinginyoung SLIchildrenwithacommunicativenarrativetask].EuropeanReviewofAppliedPsychology,64,307–321.http://dx.doi.org/10.1016/j.erap.2014.09.004

Chevrie-Muller,C.,&Plazza,M.(2001).NouvellesEpreuvespourl’ExamenduLangage(N-EEL)[Newtasksforlanguageassessment].Paris:EditionsduCentre dePsychologieAppliquée.

Chevrie-Muller,C.,Simon,A.M.,&Fournier,S.(1997).BatterielangageoraletecritMémoireattention(L2MA)[Testoralandwrittenlanguage,memory, attention].Paris:EditionsduCentredePsychologieAppliquée.

Cordewener,K.A.H.,Bosman,A.M.T.,&Verhoeven,L.(2012).Characteristicsofearlyspellingofchildrenwithspecificlanguageimpairment.Journalof CommunicationDisorders,45,212–222.http://dx.doi.org/10.1016/j.jcomdis.2012.01.003

DeCat,C.(2004).Afreshlookathowchildrenencodenewreferents.InternationalReviewofAppliedLinguistics,42,111–127.

http://dx.doi.org/10.1515/iral.2004.005

DeWeck,G.,&Jullien,S.(2013).HowdoFrench-speakingchildrenwithspecificlanguageimpairmentfirstmentionareferentinstorytelling?Between referenceandgrammar.JournalofPragmatics,56,70–87.http://dx.doi.org/10.1016/j.pragma.2013.05.009

DeWeck,G.,&Rosat,M.F.(2003).Troublesdysphasiques.Commentraconter,relater,faireâgiràl’âgepréscolaire[Dysphasics’difficulties:Howtotell,recount, andmakeactingatpreschoolages].Paris:Masson.

Decool-Mercier,N.,&Akinci,M.A.(2010).Lefonctionnementdesanaphoresdanslestextesorauxetécritsenfranc¸aisd’enfantsbilinguesetmonolingues[The functioningofanaphorsinoralandwrittentextsofmonolingualandbilingualchildren].In2èmeCongrèsMondialdeLinguistiqueFranc¸aise.Paris:EDP Sciences[p.102].

Dockrell,J.E.,&Lindsay,G.(2000).Meetingtheneedsofchildrenwithspecificspeechandlanguagedifficulties.EuropeanJournalofSpecialNeeds Education,15,24–41.http://dx.doi.org/10.1080/088562500361682

Dockrell,J.E.,Lindsay,G.,Connelly,V.,&Mackie,C.(2007).Constraintsintheproductionofwrittentextsinchildrenwithspecificlanguageimpairments. ExceptionalChildren,73,147–164[1177/001440290707300202].

Dockrell,J.E.,Lindsay,G.,&Connelly,V.(2009).Theimpactofspecificlanguageimpairmentsonadolescents’writtentexts.ExceptionalChildren,75, 427–446.http://dx.doi.org/10.1177/001440290907500403

Favart,M.,&Passerault,J.-M.(1995).Evolutiondurôlefonctionneldesconnecteursetdelaplanificationdurécitécritchezlesenfantsde7à11ans [Changesinfunctionalroleofconnectivesandofplanninginchildrenfrom7to11].RevueDePhonétiqueAppliquée,198–212[115-116-117].

Favart,M.,&Passerault,J.-M.(1996).Functionalityofcohesiondevicesinthemanagementoflocalandglobalcoherence:Twostudiesinchildren’s writtenproductionofnarratives.InR.Rijlaarsdam,H.vandenBerg,&T.Jechle(Eds.),Theories,modelsandmethodologiesinwriting(pp.349–365). Amsterdam:AmsterdamUniversityPress.

Favart,M.,&Passerault,J.-M.(1999).Aspectstextuelsdufonctionnementetdudéveloppementdesconnecteurs:Approcheenproduction[Textual aspectsoffunctioninganddevelopmentofconnectives:Approachinlanguageproduction].L’AnnéePsychologique,99,149–173.

http://dx.doi.org/10.3406/psy.1999.28552

Favart,M.,&Passerault,J.-M.(2000).Aspectsfonctionnelsdupointetdelavirguledansl’évolutiondelaplanificationdurécitécrit[Functionalaspectsof fullstopsandcommasinchangesofplanningofwrittennarratives].Enfance,2,187–205.http://dx.doi.org/10.3406/enfan.2000.3176

Favart,M.(2005).Lesmarquesdecohésion:Leurrôlefonctionneldansl’acquisitiondelaproductionécritedetexte[Cohesiondevices:Theirfunctional roleintheacquisitionoftextproduction].PsychologieFranc¸aise,50–53,305–322.http://dx.doi.org/10.1016/j.psfr.2005.05.006

Fayol,M.(1997).Onacquiringpunctuation:AstudyofwrittenFrench.InJ.Costermans,&M.Fayol(Eds.),Processinginterclausalrelationships:Studiesin theproductionandcomprehensionofText.Mahwah,NJ:LawrenceErlbaumAssociates.

Fey,M.E.,Catts,H.W.,Proctor-Williams,K.,Zhang,X.,&Tomblin,J.B.(2004).Oralandwrittencompositionskillsofchildrenwithlanguageimpairments. JournalofSpeech,Language,andHearingResearch,47,1301–1318.http://dx.doi.org/10.1044/1092-4388(2004/098)

Fraser,B.(1999).Whatarediscoursemarkers?JournalofPragmatics,31,931–952.http://dx.doi.org/10.1016/s0378-2166(98)00101-5

Freed,J.,Adams,C.,&Lockton,E.(2011).Literacyskillsinprimaryschool-agedchildrenwithpragmaticlanguageimpairment:Acomparisonwith childrenwithspecificlanguageimpairment.InternationalJournalofLanguageandCommunicationDisorders,46,334–347.

http://dx.doi.org/10.3109/13682822.2010.500316

Gernsbacher,M.A.(1989).Mechanismsthatimprovereferentialaccess.Cognition,32,99–156.http://dx.doi.org/10.1016/0010-0277(89)90001-2

Gillam,R.B.,&Johnston,J.R.(1992).Spokenandwrittenlanguagerelationshipsinlanguage/learning-impairedandnormallyachievingschool-age children.JournalofSpeechandHearingResearch,35,1303–1315.http://dx.doi.org/10.1044/jshr.3506.1303

Hickmann,M.(2004).Coherence,Cohesion,andContext.Somecomparativeperspectivesinnarrativedevelopment.InS.Strömqvist,&S.L.Verhoeven (Eds.),Relatingeventsinnarrative:Typologicalandcontextualperspectives(pp.291–306).Mahwah,NJ:LawrenceErlbaumAssociates.

Hilaire-Debove,G.,&Roch,D.(2012).Laconduitederécitchezl’enfantdysphasique[Narrativesinchildrenwithdysphasia].LesEntretiensDeBichat, 179–195.

Jullien,S.(2008).Constructionssyntaxiquesetdiscours:Lesintroductionsderéférentsdanslesnarrationsproduitespardesenfantsprésentantdes troublesspécifiquesdulangageoraletdesenfantstout-venant[Syntacticalconstructionsanddiscourse:Theintroductionofreferentsinthe narrativesproducedbychildrenwithandwithoutSLI].TravauxNeuchâteloisDeLinguistique,48,7–24.

Karmiloff,K.,&Karmiloff-Smith,A.(2003).(1sted.).CommentlesenfantsentrentdanslelangageHowchildrenenterinlanguage](2001)Paris:Retz.

Karmiloff-Smith,A.(1985).Languageandcognitiveprocessesfromadevelopmentalperspective.LanguageandCognitiveProcesses,1,61–85.

http://dx.doi.org/10.1080/01690968508402071

Karmiloff-Smith,A.(1992).Beyondmodularity:Adevelopmentalperspectiveoncognitivescience.Cambridge,MA:MITPress.

Katsos,N.,Roqueta,C.A.,ClementeEstevan,R.A.,&Cummins,C.(2011).Arechildrenwithspecificlanguageimpairmentcompetentwiththepragmatics andlogicofquantification?Cognition,19,43–57.http://dx.doi.org/10.1016/j.cognition.2010.12.004

Kern,S.(2000).JunctionandsegmentationinFrenchchildren’snarratives.PsychologyofLanguageandCommunication,4,47–63.

Khomsi,A.,Khomsi,J.,Pasquet,F.,&ParbeauGuéno,A.(2007).BilanInformatisédeLangageOralaucycle3etauCollège(BILO-3C)[Computerizedassessment oforalLanguagefor8–10yearsoldchildren,andatmiddleschool].Paris:EditionsduCentredePsychologieAppliquée.

Khomsi,A.,Khomsi,J.,&Pasquet,F.(2007).BilanInformatisédeLangageOralaucycle2(BILO-2)[Computerizedassessmentoforallanguagefor5–7yearsold children].Paris:EditionsduCentredePsychologieAppliquée.

Lambert,M.(2003).Cohésionetconnexitédansdesrécitsd’enfantsetd’apprenantspolonophonesdufranc¸ais[Cohesionandconnexityinthewritten narrativesofchildrenandofPolishlearnersofFrench].MargesLinguistiques,5,106–121.

Leonanduzzi,L.(2008).Dislocationsàgauchesetantépositions:Desvariantesendistributioncomplémentaire?[Leftdislocationsandantepositions:Variantsin complementarydistribution.]ActesduCongrèsdelaSAES.pp.121–142.Saint-Etienne,France:Publicationsdel’UniversitédeSaint-Etienne.

Leonard,L.B.(2014).Childrenwithspecificlanguageimpairment(2nded.).Cambridge,MA:MITPress.

Louwerse,M.M.,&Mitchell,H.H.(2003).Towardsataxonomyofasetofdiscoursemarkersindialog:Atheoreticalandcomputationallinguisticaccount. DiscourseProcesses,35,199–239.http://dx.doi.org/10.1207/S15326950DP35031

Mackie,C.,&Dockrell,J.E.(2004).ThenatureofwrittenlanguagedeficitsinchildrenwithSLI.JournalofSpeechandHearingResearch,47,1469–1483.

http://dx.doi.org/10.1044/1092-43882004/109

Martlew,M.(1983).Thepsychologyofwrittenlanguage:Developmentalandeducationalperspectives.Chichester,UK:JohnWileyandsons.

McCutchen,D.(1986).Domainknowledgeandlinguisticknowledgeinthedevelopmentofwritingability.JournalofMemoryandLanguage,25,431–444.

McCutchen,D.(1996).Acapacitytheoryofwriting:Workingmemoryincomposition.EducationalPsychologyReview,8,299–325.

http://dx.doi.org/10.1016/0749-596X(86)90036-7

Moeschler,J.(2002).Connecteurs,encodageconceptueletencodageprocédural[Connectives,conceptualencodingandproceduralencoding].CahiersDe LinguistiqueFranc¸aise,24,265–292.

Passerault,J.-M.(1991).Laponctuation:Recherchesenpsychologiedulangage[Punctuation:ResearchinPsychologyoflanguage].Pratiques,70,85–106.

Pellegrini,A.D.,Galda,L.,&Rubin,D.L.(1984).Contextintext:Thedevelopmentoforalandwrittenlanguageintwogenres.ChildDevelopment,55, 1549–1555.http://dx.doi.org/10.2307/1130025

Reichler-Beguelin,M.J.(1988).Anaphore,cataphoreetmémoirediscursive[Anaphor:Cataphoranddiscursivememory].Pratiques,57,15–43.

Rutter,P.,&Raban,B.(1982).Thedevelopmentofcohesioninchildren’swriting:Apreliminaryinvestigation.FirstLanguage,3,63–75.

http://dx.doi.org/10.1177/014272378200300704

Schneuwly,B.(1988).Lelangageécritchezl’enfant[Thewrittenlanguageofchildren].Paris:Delachaux&Niestlé.

Scott,C.,&Windsor,J.(2000).Generallanguageperformancemeasuresinspokenandwrittendiscourseproducedbyschool-agechildrenwithand withoutlanguagelearningdisabilities.JournalofSpeech,LanguageandHearingResearch,43,324–339.http://dx.doi.org/10.1044/jslhr.4302.324

Snowling,M.,Bishop,D.V.M.,&Stothard,S.E.(2000).IsPreschoollanguageimpairmentariskfactorfordyslexiainadolescence?JournalofChild PsychologyandPsychiatry,41,587–600.http://dx.doi.org/10.1017/s0021963099005752

Stoye,H.(2013).Lesconnecteurscontenantdesprépositionsenfranc¸ais:Profilssémantiquesetpragmatiquesensynchronieetdiachronie[Connectives containingprepositionsinFrench:Pragmatiqueandsemanticprofilesinsynchronyanddiachrony].Berlin:DeGruyter.

Wechsler,D.(1996).Echelled’intelligencepourenfants(3èmeédition)[Intelligencescaleforchildren,3rdedition].Paris:EditionsetApplications Psychologiques.

Wetherell,D.,Botting,N.,&Conti-Ramsden,G.(2007).NarrativeskillsinadolescentswithahistoryofSLIinrelationtoNon-verbalIQscores.Child Language,TeachingandTherapy,23,95–113.http://dx.doi.org/10.1080/13682820601056228

Williams,G.J.,Larkin,R.F.,&Blaggan,S.(2013).Writtenlanguageskillsinchildrenwithspecificlanguageimpairments.InternationalJournalofLanguage andCommunicationDisorders,48,160–171.http://dx.doi.org/10.1111/1460-6984.12010

vanWeerdenburg,M.,Verhoeven,L.,Bosman,A.,&vanBalkom,H.(2011).Predictingworddecodingandwordspellingdevelopmentinchildrenwith SpecificLanguageImpairment.JournalofCommunicationDisorders,44,392–411.http://dx.doi.org/10.1016/j.jcomdis.2010.12.002