ERIC ED577098: Middle School Learners' Use of Latin Roots to Infer the Meaning of Unfamiliar Words

2016·0.65 MB·English

by  ERIC

Cognition and Instruction
ISSN: 0737-0008 (Print) 1532-690X (Online) Journal homepage: http://www.tandfonline.com/loi/hcgi20
Middle School Learners' Use of Latin Roots to Infer
the Meaning of Unfamiliar Words
Amy C. Crosson & Margaret G. McKeown
To cite this article: Amy C. Crosson & Margaret G. McKeown (2016) Middle School Learners'
Use of Latin Roots to Infer the Meaning of Unfamiliar Words, Cognition and Instruction, 34:2,
148-171, DOI: 10.1080/07370008.2016.1145121
To link to this article: http://dx.doi.org/10.1080/07370008.2016.1145121
Published online: 16 Mar 2016.
Submit your article to this journal
Article views: 114
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=hcgi20
Download by: [Pennsylvania State University] Date: 16 August 2016, At: 16:43
COGNITIONANDINSTRUCTION
,VOL.,NO.,–
http://dx.doi.org/./..
MiddleSchoolLearners’UseofLatinRootstoInfertheMeaningof
UnfamiliarWords
AmyC.CrossonandMargaretG.McKeown
LearningResearchandDevelopmentCenter,UniversityofPittsburgh,Pittsburgh,Pennsylvania,USA
ABSTRACT
This study investigated how middle school students leverage information
aboutboundLatinroots(e.g.,vocinadvocateandvociferous)toinfermean-
ings of unfamiliar words, and how instruction may facilitate morphological
analysis using roots. A dynamic assessment of morphological analysis was
administeredto29sixthgraders(n=17interventionstudents)and30seventh
graders(n=18interventionstudents).Qualitativeanalysesofanalyticstrate-
giesrevealedpatternsofmorphologicalproblemsolvingthatincludeddirect
(i.e.,directapplicationofrootstoanalyzeunfamiliarwords)andindirectroutes
(i.e., use of known words that carry the roots to analyze unfamiliar words).
Interventionstudentsappliedadirectrouteathigherratesthancontrolstu-
dents.Correlationalanalysesrevealedasmallbutsignificanttreatmenteffect
onestablishingmeaningmemoryrepresentationsforrootsandasignificant,
positivetreatmenteffectforuseofrootstoinferunfamiliarwordmeanings.
OverallresultsshowpromiseforuseofboundLatinrootsformorphological
problemsolving.
Asstudentstransitionfromelementarytomiddleschool,theyareexpectedtolearnfromcontent-area
texts that are not only conceptually dense but also are more linguistically complex than those they
encounteredduringtheelementaryyears(cf.CarnegieCouncilonAdolescentLiteracy,2010;Nagy&
Townsend,2012;Snow&Uccelli,2009).Inparticular,thevocabularydemandsofacademictextspresent
asignificanthurdletosuccessfulreadingcomprehensionandcontentlearningformanystudentsinmid-
dleandhighschool(cf.CarnegieCouncilonAdolescentLiteracy,2010;Corson,1997;Townsend,Fil-
lipini,Collins,&Biancarosa,2012).NagyandAnderson(1984)estimatedthatthereareover225,000
wordsinschooltexts(grades3–9).Howcanstudentspossiblyacquireknowledgeofsomanywords?
Oneexplanationisthatstudentsacquiremetalinguisticknowledgeaboutwords’morphologicalcon-
stituentsthatenablesgenerativewordlearningprocesses.NagyandAnderson(1984)foundthatofthe
daunting225,000words,about170,000ofthesearemorphologicallycomplexwords.Theyassertedthat
well over three quarters of derived words can be easily inferred (i.e., hunter can be easily inferred by
childrenwhoknowthemeaningofhunt)orwith“reasonablehelp”fromcontext(gunnercanbeinferred
fromgunandacontextaboutsomesortofmilitaryoperation).Nonetheless,thisleavesover40,000mor-
phologicallycomplexwordswhosemeaningsarenoteasilyinferred,presentingaformidablechallenge.
Yetthereisanaspectofmorphologythathasbeenlargelyoverlookedintheresearchliteraturethat
mayequipstudentstoinfermeaningofalargeportionofwordspreviouslynotconsideredcandidates
formorphologicalproblemsolving.TheorywouldsuggestthatboundLatinrootscouldhelpstudents
overcomethelargeacademicworddemandstheyface.Inthepresentstudy,weinvestigateaheretofore
CONTACT AmyC.Crosson [email protected] LearningResearchandDevelopmentCenter,UniversityofPittsburgh,O’Hara
Street,Pittsburgh,PA.
©Taylor&FrancisGroup,LLC
COGNITIONANDINSTRUCTION 149
littleinvestigatedaspectofmorphologicalproblemsolving—theuseofmorphologicalinformationabout
boundLatinrootstoinfermeaningsofunfamiliarwords.
BoundLatinrootsandmorphologicalproblemsolving
Morphologyhasgarneredagreatdealofattentioninrecentyears,asthereisnowextensiveevidencethat
morphologicalknowledgeplaysanimportantroleinliteracyacquisitioninEnglish(cf.Carlisle,2010;
Nagy, Carlisle, & Goodwin, 2014). In English, morphemes are the smallest units of a word that carry
meaning.Morphemescanbeinflectional(e.g.,suffixesthatchangetenseornumber,suchastheplural
marker–saddedtoschools),derivational(i.e.,prefixesandsuffixesthatalterformand/ormeaning,such
asun-and–edaddedtounschooled),ortheycanbelexical(i.e.,freestandingunitssuchasschool,or
boundrootssuchasliterinliterate).
Morphologicalawarenessisthemetalinguisticinsightthatwordsaremadeupofthesemeaningful
units. We adopt Carlisle’s definition of morphological awareness as the “ability to reflect on, analyze,
andmanipulatethemorphemicelementsinwords”(Carlisle,2010,p.466).Morphologicalawareness
is linked to several components of literacy for native English speakers, such as positive associations
with decoding (e.g., Singson, Mahony, & Mann, 2000), word identification (e.g., McCutchen, Logan,
&Biangardi-Orpe,2009),andspellingability(e.g.,Kemp,2006;Leong,2000;McCutchen&Stull,2015).
Thereisalsoconvergingevidencethatmorphologicalawarenessholdsstrong,positiveassociationswith
vocabularyknowledge(Anglin,1993;Carlisle,2000;Nagy,Berninger,&Abbott,2006)andpredictsread-
ingcomprehension(Carlisle,2000),evenpredictinguniquevarianceincomprehensionaftercontrolling
forvocabularyknowledge(Nagy,Berninger,&Abbott,2006;Nagy,Berninger,Abbott,Vaughan,&Ver-
meulen, 2003; Wagner, Muse, & Tannebaum, 2007) or prior comprehension performance (Foorman,
Petscher,&Bishop,2012).
BoundLatinrootsasonetypeofmorphologicalknowledge
Studiesofmorphologicalknowledgeandinstructionhavefocusedonstem-wordsandderivationalpre-
fixes and suffixes (e.g., Baumann, Edwards, Boland, Olejnik, & Kame’enui, 2003; Baumann, Edwards,
Font,Tereshinski,Kame’enui,&Olejnik,2002;Kirk&Gillon,2009),oftenreferringtothemanipulated
unitsas“roots.”Thisillustratesachallengeinunderstandingtheroleofmorphology—useofconsistent
terminology.Wehavechosentousethefollowingterms:
(cid:129) Weuse“stem-word”torefertothemorphologicallysimplestderivationofawordthatisfreestand-
ing.Forexample,vocalisthestemofvocalizeorvocalization.Othershaveusedthetermssuchas
“baseword”or“rootword”torefertothisconcept.
(cid:129) Weuse“boundroot”torefertoroots,mostoftenfromLatin,thatarethesemanticbasisofEnglish
wordsbutarenotfree-standingwordsinEnglish.Forexample,vocmeaning“speak,”istheLatin
boundrootinadvocateandvocalize.
(cid:129) When referring to relationships between words via bound roots, we use “root-related word” to
indicate freestanding words that share a bound root. For example, the following are all root-
relatedwords:vocal,vocalize,vocalization,vocabulary,advocate(astheyallcontaintheboundroot,
voc).
Why is it important to distinguish bound roots from stem-words? Words that students analyze
in elementary grades are often of Germanic origin and are made up of stem-words to which pre-
fixes and suffixes can be cleanly and efficiently added and subtracted (Bar-ilan & Berman, 2007).
However, academic words rarely contain such transparent, freestanding stem-words (Tyler & Nagy,
1989); in fact most words on the Academic Word List (Coxhead, 2000) are Latinate (Lubliner &
Hiebert, 2011), with their main semantic components being bound roots. As such, morphological
analysis of words likely to appear in academic texts may call on knowledge of Latin bound roots.
Morphological analysis of bound roots could be highly generative given that roots often carry sub-
stantial information about a word’s meaning. Consider that the word illuminate contains the root
lumin, from Latin for light; but if a learner does not know the meaning of lumin, knowledge about
150 A.C.CROSSONANDM.G.MCKEOWN
the word’s derivational affixes il and ate are not likely to be helpful. Knowledge of the bound root
meaning, on the other hand, might support accessing the meaning of words such as luminous and
luminary.
Theoreticalframework:Theroleofboundrootsinwordlearning
OurstudydrawsprimarilyonSchreuderandBaayen’s(1995)modelofmorphologicalprocessing,which
accounts for how morphemes come to be represented in the lexicon. The model proposes that over
thecourseofmultipleencounters,learnersdevelopconnectionsbetweenorthographicstringsandtheir
correspondingmeanings.Whenthelearnerdetectsaredundantrelationshipbetweenformandmean-
ing, a “concept node” is created; each time the unit is encountered, the strength of the representa-
tion becomes stronger and accumulates additional semantic and syntactic information. For example,
when the morpheme nov meaning “new” is encountered in multiple words such as novelty and reno-
vate,learnersmaycreateaconceptnodefornovanditsmeaningthatwouldbecomeincreasinglyspec-
ified and robust with each encounter. According to this model, when an unfamiliar morphologically
complexwordisencountered,thelearneractivatesrelevantconceptnodes(i.e.,themeaningsindexed
by the morphological constituents) to hypothesize how the morphemes might be combined to infer
word meaning, checking this hypothesis against other semantic and syntactic information from the
context.
Although in this model, both whole, freestanding words and morphemes have the potential to be
representedbyconceptnodes,theframeworkisnotadequatetoexplainalearner’suseofboundmor-
phemes. For Schreuder and Baayen (1995), frequency and transparency are key to a learner’s ability
to activate and strengthen nodes, yet bound roots are not always redundant and transparent enough
to activate their own nodes. Indeed, several dimensions may affect the accessibility of the morpho-
logicalconstituent.First,theboundroot’ssemantictransparencywilllikelyaffectitsaccessibilitysuch
1
thatopaqueboundrootswillprobablynotbeaccessible. Second,theboundroot’sphonologicaltrans-
parency may affect its accessibility; bound roots with more than one phonological form may take
longer to see than bound roots with a single form (e.g., there is a vowel shift in novice versus ren-
2
ovate, which may affect the accessibility and stability of its representation). Finally, the bound root’s
orthographicstabilityanduniquenessmayaffectthemorpheme’saccessibility,asrootsthatshiftortho-
graphically(e.g.,vidandvisbothmeaning“see”)arelikelytobelessaccessible,androotsrepresented
by common orthographic strings (e.g., no meaning “know” as in “notify”) will be more difficult to
3
detect.
Thusaframeworkfortheroleofboundmorphemesneedstobeaugmentedbyfurtherconstructs.
OnesuchconstructismetalinguisticawarenesssuchasNagyandScott(2000)describeasnecessaryto
understandingmorphologicalrelationshipsthatareobscuredbychangesinspellingandpronunciation.
Thus we hypothesize that learners need both established memory representations (i.e., concept nodes)
forboundrootsandtheirmeaningsandthemetalinguisticinsightabouthowtousethisinformationfor
problemsolving.
Moreover, given the characteristics of bound roots that learners need to bring to bear in order to
accesstherelevantroot(semantictransparency,phonologicaltransparency,andorthographicstability
and uniqueness), our theory of morphological problem solving using bound roots also emphasizes a
third construct, flexibility in applying root meanings to infer new word meanings. Such flexibility is a
specificcaseofawell-documentedaspectofcognitiveprocessing,cognitiveflexibility,whichinvolves
consideringmultipleaspectsofthoughtatonce(Deak,2003;Scott,1962).Forexample,successatusing
boundLatinrootsrequiresbeingabletoswitchone’sfocusfromasingle,preciseorthographicformof
theroottoconsideringotherpossibleorthographicrepresentations(vidandvis).Italsorequiresflexibil-
ityforlearnerstoapplytheirunderstandingoftherootmeaningasitmakessenseinaparticularcontext,
asitisrarelysufficienttobluntly“plugin”aroot’smeaning.Similarly,vocabularyresearchersidentify
flexibilityasneededforlearnerstoadjusttheirfocusbetweenwordmeaningsandcontext,andtocon-
siderpossiblealternatemeaningsinordertounderstandthefitofawordinanovelsentence(Deane,
2014;Pearson,Hiebert,&Kamil,2012).
COGNITIONANDINSTRUCTION 151
TheroleofboundLatinrootsinwordlearningandprocessing:Indirectordirectroute?
Despitethetheoreticalbasisforassertingthatknowledgeofboundrootsshouldsupportmorphological
analysistoinferlearningnewwords,empiricalevidencethatboundrootsarehelpfultostudentsisthin.
Someevidencesuggeststhatstudentsleveragetheirknowledgeofroot-relatedwordsformorphological
problemsolving.Inthiscase,theboundrootisnotinvokedexplicitly,butisthecommonmorphological
constituentusedforinferringmeaningoftheunfamiliarword.Forexample,studentsrecognizetherela-
tionshipbetweenanunfamiliarword(e.g.,astronomy)andafamiliarroot-relatedword(e.g.,astronaut),
andusethemeaningofthefamiliarwordtoshedlightonthemeaningoftheunfamiliarword.While
theboundrootastro(star)isnevercalledonexplicitly,itisthesharedmorphologicalconstituentthat
gaveinsightintotheunfamiliarword’smeaning.Wecallthisanindirectrouteformorphologicalproblem
solvingwithboundroots.
Alternatively,studentsmightfollowadirectrouteformorphologicalproblemsolving,suchthatthey
explicitlyapplythemeaningsofboundLatinrootstoanalyzeunfamiliarwordsthatcontaintheroots.
Forexample,whenencounteringtheunfamiliarwordastronomy,studentsmightexplicitlycallonthe
meaningoftheboundrootforinformationaboutthewordmeaning(e.g.,somethingtodowithstars).
Morphologicalprocessingstrategiessimilartotheindirectanddirectrouteshavebeenobservedin
previousresearch,butmoreoftenforderivationalaffixesthanforboundroots.Forexample,inhissemi-
nalstudyofmorphologicalproblemsolvingwithstudentsintheelementarygrades,Anglin(1993)found
somestudentsfollowedastrategyakintotheindirectrouteformorphologicalproblemsolvingwhen
theywerepresentedwithamorphologicallycomplex,unfamiliarword.Hecalledthisstrategy“analogy.”
However,Anglin’sparticipantsemployedthisstrategyonlyrarely,andtheytendedtotapderivational
morphologyratherthanboundroots(e.g.,useofpiglettoproblemsolvetreelet),suchthatevidenceof
anindirectrouteusingboundrootswassparseandevidenceofthedirectrouteusingboundrootswas
nearlyabsent.
Inastudywithadolescentlearnersinthemiddleschoolgrades,PachecoandGoodwin(2013)found
someevidencethattheseolderparticipants(manyofwhomwereEnglishlearners)usedboundrootsin
morphologicalproblemsolving,includingstrategiesthatseemtomap(partially)ontoboththedirect
andindirectroutes.Evidenceofadirectrouterelatestotheresearchers’observationthatstudentsused
information about multiple morphological constituents (called “parts to whole”). In fact this was the
mostcommonstrategyappliedbymiddleschoolparticipants.However,itisimportanttonotethatthe
morphological constituents included use of larger morphological units beyond bound roots for mor-
phologicalproblemsolving(e.g.,analysisofarcheo+astronomerinarcheoastronomerasanexampleof
parts-to-whole).Asaresultwearenotcertainhowoftenandhoweffectivelystudentswereabletouse
boundrootsspecificallyintheirmorphologicalproblem-solvingprocess.
The researchers also found evidence related to an indirect route. Following Anglin, Pacheco and
Goodwin(2013)used“analogy”todescribeuseofwordwithasimilarmorphologicalformtoproblem-
solve an unfamiliar word (e.g., use spectator to problem-solve the meaning of introspective). While
this strategy was rarely employed by students (used in only 4% of problem-solving episodes), the
researchers argue that it was highly effective when used. As such, Goodwin and colleagues (Good-
win, Gilbert, Cho & Kearns, 2014; Pacheco & Goodwin, 2013) have argued that knowledge of root-
relatedwordsmaybeakeymechanismbywhichmorphologicalconstituents’orthographic,phonolog-
ical,andsemanticrepresentationsareactivatedinsolvingnewwordmeanings.However,Pachecoand
Goodwinincludedderivationalaffixesinthe“analogy”classification(e.g.,useofpediatriciantoprob-
lemsolvethemeaningofclinicianasanyexampleof“analogy”).Thus,theroleofboundrootsspecif-
icallyinmorphologicalproblemsolving,whetherstudentsmoreeffectivelyandefficientlypursueone
routeortheother,andhowinstructionmightaffectprocessingofboundrootsareissuesthatremain
unresolved.
Two additional studies also suggest that students may use bound roots for inferring the meanings
ofunfamiliarwords,butsimilartothestudyabove,theydonotfocussquarelyonboundroots.First,
inacross-sectionalstudywithfifthandeighthgraderstoexaminethemechanismbywhichmorpho-
logicalawarenesslinkstoreadingcomprehension,McCutchenandLogan(2011)investigatedstudents’
152 A.C.CROSSONANDM.G.MCKEOWN
abilitytousemorphologicalanalysistofigureoutthemeaningsoflow-frequencyunfamiliarwordsdur-
ingreading.Thetaskrequiredstudentstoreadawordandchoosethecorrespondingdefinitionfrom
threeoptions.Thetargetwordsdesignedtobemorphologicallyaccessiblecontainedmorphologicalcon-
stituentsthatgaveinformationaboutthetargetwordmeanings.Theirfindingssuggestthatstudentsdo
usemorphologicalanalysistoinferwordmeaning,andwhiletheirassessmentwasmostlyfocusedon
derivationalmorphology,itdidincludesomeitemsthatcalledonknowledgeofboundroots.Forexam-
ple,oneitemaskedstudentstomakeasemanticconnectionbetweentalkativeandverbose,whichderives
fromtheboundLatinrootverb(fromverbum),meaning“word.”McCutchenandLogan’sfindingssug-
gestthatmorphologicalanalysisusingboundrootsmaybeonestrategythatchildrenusetoinferthe
meaningsofunfamiliarwordsduringreading.
Second,aninterventionstudybyBowersandKirby(2010)investigatedwhetherfourthandfifthgrade
studentscanlearntousemorphologicalanalysistofindinformationaboutwordmeanings.Theinstruc-
tionaddressednotonlyderivationalaffixesbutalsobothstem-wordsandboundroots—bothofwhich
werecalled“bases.”Theauthorsinvestigatedwhethersuchinstructionleadstogainsinvocabularylearn-
ing, and whether this knowledge transfers to inferring meaning of low-frequency words that contain
taught“bases”andderivationalaffixes.Controllingforinitialvocabularyknowledge,BowersandKirby
notedthatinterventionstudentswerebetterabletodefinenotonlyinstructedwordsbutalsonovelwords
that included “bases” taught during the intervention when compared to a control group. Bowers and
Kirbyconcludedthatteachingmorphologicalanalysishelpsstudentslearnvocabularybeyondthewords
taught.However,whiletheinterventionincludedinstructioninhowtoparseboundrootsfrommorpho-
logicallycomplexwords(e.g.,ruptindisrupt),theiroutcomemeasuregaveequalcreditforextracting
stem-wordsandboundroots.Thustheimplicationsforboundroots,whilepromising,areunclear.In
thepresentstudy,ourapproachprivilegesadirectroutebyinvestigatinganinstructionalinterventionin
whichstudentsweretaughttouseboundroutestoanalyzewordmeanings.Further,ourstudyexamines
whatoccurswhenstudentsattempttouseboundrootstoinferwordmeaningsandtheprocessbreaks
down.
Effectsofmorphologyinstruction
Whiletherelationshipbetweenmorphologicalknowledgeandliteracyisstrong,theeffectsofproviding
instructioninmorphology,althoughpositive,aresomewhatlesscompelling.Interventionshavedemon-
strated that morphology instruction has a moderate effect on vocabulary learning (Bowers, Kirby, &
Deacon,2010;Carlisle,2010;Goodwin&Ahn,2010,2013).Inameta-analysisofstudieswithstudents
intheelementaryandmiddleschoolgrades,GoodwinandAhn(2013)foundastatisticallysignificant
meaneffectofmorphologyinstructionforvocabulary(meand=0.34).Forexample,Baumannetal.
(2003)taughtstudentstousemorphologicalknowledgetoinfermeaningsofunknownwordsandfound
alargetreatmenteffectforinstructionaboutprefixesandstem-words(e.g.,prejudge,disloyal,rehire),such
thatstudentsweremoresuccessfulat“unlockingthemeaningsofmorphemicallydecipherabletransfer
vocabulary”(p.464).
Contextofthestudy
Thepresentinvestigationwasconductedinthecontextofa2-yearinterventionwithmiddleschoollearn-
ers,RobustAcademicVocabularyEncounters(RAVE;McKeown,Crosson,Artz,Sandora,&Beck,2013).
Five classrooms of students participated in the study from the beginning of sixth grade until the end
ofseventhgrade.Eachyear,threeclassroomsreceivedtheinterventionandtwoclassroomsservedas
controls. The intervention focused on instruction for 99 words from the Academic Word List (AWL;
Coxhead,2000)insixthgradeand96AWLwordsinseventhgrade.Theinterventionwasbasedoncon-
sensusofprinciplesofeffectiveinstructionderivedoverseveraldecadesincludingpresentationofdefi-
nitionalandcontextualinformation,encountersinmultiplecontexts,andactiveprocessing(Baumann,
Kame’enui,&Ash,2003;McKeown,Beck,Omanson,&Pople,1985;NRP,2000;Stahl&Fairbanks,1986).
COGNITIONANDINSTRUCTION 153
Table.ExamplesofLatinrootsandrelatedwordsfromRAVEmorphologylessons.
Root Rootmeaning RAVEword Root-relatedwords
Equ Equal Adequate equation,equator
Loc Place Allocate locate,dislocate,local
Form Shape Conform uniform,transform,deform,
reform
Sequ Follow Consequences sequel,sequence
Min Small/less Diminish minimize,minus,miniscule,
mince,miniature,minor,
minute
Tract Drag Extract tractor,attractcontract
Spect See/look Prospect specimen,spectacles,inspect,
prospector,respect
Fer Carry Transfer ferry,prefer,odoriferous
Fy Tomake Unify terrify,beautify
RAVEpresentedtargetwordsinmultiplenon-fictioncontexts,andstudentswereguidedtointegrate
wordmeaningwithindifferentcontexts.AccordingtoBowersandcolleagues(Bowers,Kirby,&Deacon,
2010),ifmorphologicalinstructionistotransferfromthesublexicaltowordlevelandultimatelysupport
comprehensionatlevelofconnectedtext,“thistransferislikelytobefacilitatedbyinstructionalmethods
thatintegratemorphologicalinstructionwithotheraspectsofliteracyinstruction”(p.149).Morphology
instruction in RAVE equipped students with knowledge and problem-solving techniques to connect
boundrootstothemeaningofmorphologicallycomplexwordsinwhichtheyappear.
Researchquestions
Ourresearchquestionsareasfollows:
1. Whatarethepatternsofmorphologicalproblemsolvingwithboundrootsthatallowedstudents
toinfermeaningofunfamiliarwords?
2. Howdidanacademicvocabularyinterventionwithamorphologicalanalysiscomponentaffect
students’approachestosolvingunfamiliarwordsusingboundroots?
a. Didtheinterventionsupportstudentstoestablishmeaningrepresentationsinmemoryfor
boundroots,asindicatedbytheirabilitytorecognizerootsthathadbeentaughtandrecall
themeaningsoftheseroots?
b. Didtheinterventionsupportstudentstodevelopthemetalinguisticinsightthatrootscarry
informationaboutwordmeaning,asindicatedbystudents’abilitytouserootsformorpho-
logicalanalysistoinferwordmeaningsandcomprehendthemincontext?Doesthisinsight
requireestablishedmemoryrepresentationsforboundroots?
Method
Intervention
TheRAVEinterventioncomprised12unitsofsevendailylessonseachinsixthgrade,and16unitsof
sevendailylessonseachinseventhgrade.Nineunitsinsixthgradeandfiveunitsinseventhgradecon-
cluded with a lesson called “Becoming Aware of Language,” whose focus was morphological analysis
usingboundLatinroots.
BetweentwoandfourrootsweretaughtintheBecomingAwareofLanguagelesson(seeexamplesin
Table1).Therootschosenforinstructionwereselectedfromtheacademicwordsthathadbeentaught
inthatunit.Atotalof34rootswastaughtoverthecourseoftheintervention(18insixthgradeand16
inseventhgrade);30ofthesewereLatinrootsandfourwereGreek.Atotalof95root-relatedwords
weretaught(52insixthgradeand43inseventhgrade).Asmallnumberofprefixeswerealsotaughtin
conjunctionwiththerootswhentheywerekeytowordmeaning.Teachersusedtheterm“wordparts”
154 A.C.CROSSONANDM.G.MCKEOWN
totalkaboutmorphologicalconstituentsingeneral,andtheyalternatedbetweentheterms“wordpart”
and“root”whenreferringtotheboundrootsduringinstruction.
Instructionwasdesignedtoteachstudentstouseananalytical,problem-solvingstancetowardmor-
phologicalinformation.Thegoalsofinstructionweretoteachthat:
(cid:129) aboundrootisfoundwithinawordanditoftenprovidesinformationabouttheword’smeaning;
(cid:129) becausetheserelationshipshaveevolvedovertime,therelationisnotalwaysreadilyapparent;
(cid:129) boththespellingandthepronunciationofrootscanvary;and
(cid:129) thereare“falsealarms,”thatis,justbecauseyouseearootdoesnotmeanitisnecessarilythere.For
example,valcarriestheLatinrootmeaning“strong”invalue,butitdoesnotholdanyetymological
relationshiptovalve,valleyormedieval.
Instructionproceededasfollows.Thefirststepwastoexplaintheoriginoftheboundroot,provideits
meaning,andinvitestudentstoanalyzehowtheroot’smeaningconnectstothemeaningoftheacademic
word they had studied. For example, to introduce the root, spect, the teacher wrote the RAVE word
containingtherootontheboard(prospect)andunderlinedtheboundroot.Shethenexplainedthatthe
boundrootcamefromaLatinwordmeaning“tolookat”andthat“look”isthemeaningoftheroot.
Finally,shewrotethemeaningbelowtheunderlinedrootontheboard.
The second step was to engage students in analyzing how the root’s meaning is connected to the
meaningofotherwordsthatcarrythesameroot(i.e.,root-relatedwords).Ininstruction,teacherscalled
these“extensionwords.”Forexample,afteranalyzingtherelationshipbetweentheideaof“look”andthe
meaningofprospect(i.e.,becauseaprospectissomethingthatmighthappen,it’slikelookingforwardinto
thefuture),studentswereaskedtoexplaintherelationshipbetweentherootandextensionwordssuch
asspecimenandinspect.Instructionalsoincludedteachingstudentsthatorthographicandphonological
variationisafeatureofboundrootsbyaskingstudentstolookatwordssidebysideandnotedifferences
inhowtheboundroots’pronunciationsandspellingsvaried.Inthiscase,studentswereguidedtoseethat
theboundrootcanbewritten“spec”or“spect,”andthe“c”canbeahardorsoftsound(e.g.,specimen,
spectator).
Atranscriptexamplefromsixthgradeillustrateshowstudentswereguidedtodirectlyrelatethemean-
ingoftheboundrootmin(small)tothemeaningsoftheroot-relatedwords,whichwere,inthiscase,
familiarinmeaning,minusandminor.Thusstudentswerenotbeingguidedtoinferwordmeaning,but
toseethedirectroutefromthemeaningofmintothemeaningsofthetworoot-relatedwords,akey
featureofourRAVEmorphologyinstruction.
Ms.H.:Okaywhataboutminus.Takemethroughyourthoughtsforminus.
Tyler:Makeitgetsmallerandsmaller.
Ms.H.:Yeahandhowdoyoudothat?
Tyler:Bysubtracting.
Ms.H.:Yeah,soyou’retakingsomethingawayandmakingitsmallerandsmaller. …Whatabout
minor?
D’wayne:Likeasmallpartofsomething.Yeahit’slikeasmallpartofsomethingorlessofsomething.
Ifit’saminorhockeyleagueit’s …it’sless.It’smaybenotascompetitiveormaybeit’slessskillfulthan
thepeoplewhoplaythemajorleagues.Ifyou’reminor,likeyou’reminorswhenyou’relessthan18years
old.
It is important to note that no further instruction about each of the bound roots was provided in
subsequentRAVElessons.Whilewecannotbecertain,webelievethatthetypeofmorphologicalanalysis
withboundrootsdidnotspilloverintoothercontent-areasorbeyondthespecificinstructions.Teachers
notedonseveraloccasionsthatwhiletheyconsideredthiscontentworthwhileandchallenging,itwas
newtothemastheyhadneverstudiedLatinandhadbeenlargelyunawareofthewords’rootmeanings.
Controlcondition
Insixthgrade,theinterventionandcontrolgroupshadthesamebasiclanguageartsinstructionfrom
abasalreader,except thatthedailyvocabularylessonsfromthebasalreaderwerereplacedbyRAVE
instruction for the intervention group. The basal lessons included minimal attention to morphology
COGNITIONANDINSTRUCTION 155
in the form of applying common prefixes and suffixes to transparent, high frequency stem-words. In
seventhgrade,theRAVEandcontrolgroupshadthesamebasiclanguageartsinstruction(theyreadthe
samenovelsandcompletedthesameassignments),exceptthattheinterventiongroupreceivedRAVE
vocabularylessonswhilethecontrolgroupreceivedlessonsfromanalternativevocabularyprogramwith
nooverlappingwordsandnomorphologycomponent.
Participants
Inthefirstyearofthestudy,participantswere29sixthgradersselectedfromthefiveclassesparticipat-
ingintheRAVEstudy,whichtookplaceinapublicschoolwithinaworkingclasscommunityinthe
northeasternUnitedStates.SeventeenoftheparticipantsweredrawnfromtheRAVEclassroomsand
12fromthecontrolclassrooms.AllstudentswerenativespeakersofEnglish.Studentsfromeachclass
weresampledtorepresentlow-middle-tohigh-middle-performingstudentsontheGates-MacGinitie
(MacGinitie,MacGinitie,Maria,&Dreyer,2000)standardizedreadingpretest.Anindependentsamples
t-testconfirmednosignificantdifferenceinpriorreadingachievementbetweenthetwogroups(t=.351,
p=.363).
Inthesecondyear,participantswere30seventh-gradestudents,18whohadparticipatedinRAVE
in both sixth and seventh grades and 12 who were in the control group for two years. There was no
overlap of participants in sixth and seventh grades. In seventh grade, one teacher taught all five sec-
tions of English Language Arts offered at that grade level. Four of the sections were heterogeneously
grouped; two of these were assigned to the intervention and two were assigned to the control group.
Fromthesefoursections,participants(n=12RAVEandn=12control)wereselectedtorepresentlow-
middle-tohigh-middle-performingstudentsontheGates-MacGinitie(MacGinitie,MacGinitie,Maria,
&Dreyer,2000)standardizedreadingpretest.Inadditiontothesefoursections,onesectionwasdesig-
natedasanhonorsclass.Thissectionwasassignedtotheintervention(n=6HonorsRAVE),hereafter
calledHonorsRAVE.Aone-wayANOVArevealedasignificantdifferenceontheGatespretestbetween
groups,F(2,27)=11.95,p<.000.Post-hocpairwisecomparisonsconfirmedthat,asexpected,Honors
RAVEhadsignificantlyhigherGatespretestscoresthanstudentsinthecontrolandRAVEgroups,mean
diff=26.76,p<.000andmeandiff=24.167,p<.000,respectively.However,therewasnosignificant
differencebetweenstudentsinthecontrolandRAVEconditions(meandiff=2.583,p=.586).Thusall
analyseswereperformedonthethreegroups:HonorsRAVE,RAVE,andcontrol.
Whileweareawareofevidencethathighandlowskillreadersdemonstratedifferentpatternsinmor-
phologicalproblem-solvingskill(Carlisle,2003;NagyandHiebert,2011),oursamplewastoosmallto
makeinferencesaboutreadersofdifferentabilitylevels,thuswedidnotseektotesthypothesesabout
readingskillandmorphologicalanalysis,butratherincludedasufficientrangeofstudents(excluding
both struggling and highly advanced readers) so that we could be confident that our findings would
likelygeneralizetoarangeofreadersratherthanagroupataparticularskilllevel.
Dynamicassessment
Students’abilitytoengageinmorphologicalanalysiswasmeasuredusingaresearcher-designed,dynamic
assessment.Dynamicassessmentwasselectedbecauseitprovidesrichinformationaboutstudents’think-
ingprocesses,especiallyforskillsandknowledgethatarestillforming(Burton&Watkins,2007;Poehner
&Lantolf,2010).Assuch,itrevealsnotonlywhetherastudentcansuccessfullyengageinmorphological
analysisofunfamiliarwords,buthow,aswellaswhereandhowtheprocessbreaksdown.Thusthistask
assessedtheprocessofapplyingmorphologicalinformationratherthanjusttheknowledgeneededtodo
so.Similarmeasureshavebeenusedtoinvestigatemorphologicalproblemsolvingofunfamiliarwords
byAnglin(1993)aswellasbyPachecoandGoodwin(2013).
Dynamicassessmentallowedustosee:(1)whenstudentsdecomposewords,dotheyrecognizethe
boundLatinrootasoneoftheunits;(2)whetherstudentslinkmeaninginformationtothisunit;and
(3) whether students can integrate orthographic/phonological information and semantic information
associatedwiththeboundroottounlockthemeaningofanunfamiliarword.Forcontrolstudentswho
156 A.C.CROSSONANDM.G.MCKEOWN
Table.Dynamicassessmentitemsadministeredinsixthandseventhgrades.
Root Sentencepresented
Sixthgrade
Crit Laurawassurprisedbyteacher’scritique.
Min Mostoftheirconversationswereabouttheminutiaeofdailylife.
Form Themiddleschoolyearsareformative.
Mit Theanimalemittedastrangesound.
Sent JanetsharesMaria’ssentimentsaboutsurpriseparties.
Equ Somepeopleequateicecreamfrozenyogurt.
Uni Thechildrencountedfromonetoteninunison.
SeventhGrade
Voc Outonthestreettherewasavociferouscrowd.
Sed/sid Thepartywassurprisinglysedate.
Vert/vers Theengineerstriedtodiverttheriver.
Mon Thedoctoradmonishedtheboyabouthisdiet.
Val Thewomanreceivedpublicityforhervalor.
Var Thewritertriedtocomeupwithavariantending.
Circum Themanalwaystriedtocircumventtherules.
Clude Thetunneltotheplaygroundwasoccluded.
Vid/vis Thehousesoldquicklybecauseofitsamazingvista.
Note.Targetwordsareunderlined.
had not been taught the bound roots, the question was whether—when given the information about
rootsneededformorphologicalanalysis—theyweretheyabletousethisinformationaseffectivelyas
interventionstudents,orwhethertheRAVEstudentsdemonstratedgreaterpotentialtouseinformation
aboutboundrootsformorphologicalanalysisfollowingtheirexperienceintheintervention.
Inthistask,theassessoraskedstudentsaseriesofquestionstopromptmorphologicalanalysisofan
unfamiliarwordpresentedinasentence-levelcontext.Thetaskcomprisedsevenitemsinsixthgradeand
nineitemsinseventhgrade.Foreachitem,studentswerepresentedwithasentencethatincludedatarget
wordthatcontainedoneoftherootstaughtintheRAVEintervention.Thetargetworditself,however,
hadneverbeenusedforinstruction.Sentencesweredesignedasneutralcontextsforthewords,soasnot
toprovidehintsabouttheword’smeaning.Forexample,theboundrootmin,meaning“small”or“less”
wastaughtwiththeRAVEword,diminish.Thedynamicassessmentformincontainedthetargetword,
minutiae:“Mostoftheirconversationswereabouttheminutiaeofdailylife.”Acompletelistofitemsis
presentedinTable2.AsamplescriptisfoundinAppendixA.
Wordselection
Targetwordswereexpectedtobeunfamiliartosixthandseventhgraderspertheirclassificationas8th
gradeorhigherontheLivingWordVocabularyinventory(Dale&O’Rourke,1979).ConfirmingLiv-
ingWordVocabularyclassifications,alltargetwordshadStandardFrequencyIndex(SFI)scores(i.e.,
logtransformationofU-score)ofless than45 per theZeno corpus(Zeno,Ivens, Millard& Duvvuri,
1995).Regardingmorphologicalfamilysize,thefrequencyofboundrootswasdeterminedusingBecker
andcolleagues’(Becker,Dixon,&Anderson-Inman,1980)corpusofmorphographicunitsinthe26,000
highestfrequencywordsinEnglish.Theboundrootstestedwereestimatedtobeinthetopfifthofmost
4
frequentmorphographsinEnglish. Thisisimportantinlightofthefactthatfrequencyofencounters
withanymorphologicalconstituentimpactswhetherandhowquicklythatconstituentwillberecognized
inotherwords(Nagy,Anderson,Schommer,Scott,&Stallman,1989;Nagy&Hiebert,2011).Wordand
root-levelstatisticsarepresentedinTable3.
Taskinstructionsandprocedure
All tasks were administered individually by a member of the research team in a quiet room in the
school,andallsessionswereaudiorecorded.Thetaskwasintroducedtothestudentsbytellingthem,
“we are interested in understanding more about what middle school students know about how lan-
guage works, so we’re going to talk with you about that today. We’re going to look at some sentences
and I’ll ask you some questions.” During the task, the assessors were directed to use the term “word