De Villiers and Roeper: Relative De Villiers and Roeper: Relative clauses are barriers to whclauses are barriers to wh--

movement for young childrenmovement for young childrenmovement for young childrenmovement for young children

Sherrod WrightSherrod WrightSherrod WrightSherrod Wright

BackgroundBackgroundBackgroundBackground

Subjacency Principle:Subjacency Principle:Subjacency Principle: Subjacency Principle: Movement may not cross more than one Movement may not cross more than one bounding nodebounding nodebounding node. bounding node. Bounding nodes are IP(S), CP(S’), and NPBounding nodes are IP(S), CP(S’), and NPLanguages differ over whether S or S’ is a Languages differ over whether S or S’ is a bounding nodebounding node

These sentences show that there are These sentences show that there are constraints on movement that determine constraints on movement that determine which interpretations are grammatical or which interpretations are grammatical or not.not.

Study 1Study 1Study 1Study 1

Does the grammar of children ages 3Does the grammar of children ages 3--4 4 treat relative clause barriers/islands as a treat relative clause barriers/islands as a proper constraint for movement and will proper constraint for movement and will the children try to answer the relative the children try to answer the relative pronoun.pronoun.

Otsu(1981)Otsu(1981)Otsu(1981)Otsu(1981)

Showed that 3 and 4 year olds answeredShowed that 3 and 4 year olds answeredShowed that 3 and 4 year olds answered Showed that 3 and 4 year olds answered questions in a manner that violated the questions in a manner that violated the barrierhood of the relative clause or prepositional barrierhood of the relative clause or prepositional phrase 25phrase 25--30% of the time.30% of the time.

HypothesisHypothesisHypothesisHypothesis

Children ages 3 and 4 are obedient to the Children ages 3 and 4 are obedient to the relative clause constraint on movementrelative clause constraint on movementrelative clause constraint on movementrelative clause constraint on movement

MethodMethodMethodMethodSubjects Subjects –– 23 children ages 323 children ages 3--6 and an average 6 and an average jj gg ggof 4. English as a first language and no of 4. English as a first language and no handicapping conditions.handicapping conditions.StimuliStimuli –– short pictured stories that set upshort pictured stories that set upStimuli Stimuli short pictured stories that set up short pictured stories that set up appropriate conditions for the use of the relative appropriate conditions for the use of the relative clause and allowed for all possible construals of clause and allowed for all possible construals of the question being asked at the endthe question being asked at the endthe question being asked at the end. the question being asked at the end. Procedure Procedure –– six relative clause stories and six relative clause stories and questions were presented in randomized order questions were presented in randomized order

h hild l b f 20h hild l b f 20to the children as a larger battery of 20 to the children as a larger battery of 20 questions, none of which had relative clauses as questions, none of which had relative clauses as part of their structure. part of their structure. pp

StimuliStimuliStimuliStimuli

Question typesQuestion typesQuestion typesQuestion typesSubject Relative:Subject Relative:Subject Relative: Subject Relative:

HowHowii did the man who hurt his leg get home tdid the man who hurt his leg get home tii??Object Relative:Object Relative:Object Relative:Object Relative:

HowHowii did the man rescue the cat tdid the man rescue the cat tii who broke who broke her leg?her leg?ggExtraposed SR:Extraposed SR:

WhenWhenii did the man go home tdid the man go home tii who hurt his leg?who hurt his leg?WhenWhenii did the man go home tdid the man go home tii who hurt his leg?who hurt his leg?

ResultsResultsResultsResults

FindingsFindingsFindingsFindings

Relative clauses are respectedRelative clauses are respectedRelative clauses are respectedRelative clauses are respectedErrors do not violate RC island constraintsErrors do not violate RC island constraintsChild d th di l 0% f thChild d th di l 0% f thChildren answered the medial 0% of the Children answered the medial 0% of the timetimeThe younger children had more errorsThe younger children had more errors

Study 2Study 2Study 2Study 2

Attempts to investigate the changes in theAttempts to investigate the changes in theAttempts to investigate the changes in the Attempts to investigate the changes in the grammars of 3 year olds over the course of a grammars of 3 year olds over the course of a year with respect to the subject extraposed year with respect to the subject extraposed sentences.sentences.Also included were questions with medial Also included were questions with medial argument embedded questions to hopefully argument embedded questions to hopefully answer; do children who answer the medial answer; do children who answer the medial question in embedded questions also answerquestion in embedded questions also answerquestion in embedded questions also answer question in embedded questions also answer the relative pronoun in the RCthe relative pronoun in the RC

MethodMethodMethodMethod

SubjectsSubjects –– 12 children ages 312 children ages 3--44Subjects Subjects 12 children ages 312 children ages 3 44Stimuli Stimuli –– Subject extraposed questions Subject extraposed questions and embedded questionsand embedded questionsand embedded questionsand embedded questionsProcedure Procedure –– same methodology as Study same methodology as Study 1 t thi l it di l t d1 t thi l it di l t d1, except this was a longitudinal study 1, except this was a longitudinal study which subjects were brought in 3 times which subjects were brought in 3 times

th f 1th f 1over the course of 1 year.over the course of 1 year.

ResultsResultsResultsResults

ResultsResultsResultsResults

FindingsFindingsFindingsFindings

WhWh--island constraints are also not violatedisland constraints are also not violatedWhWh island constraints are also not violatedisland constraints are also not violatedChildren are able to distinguish whChildren are able to distinguish wh-- and and relative clausesrelative clausesrelative clausesrelative clausesSeems children are learning to interpret Seems children are learning to interpret th i l b tt th ith i l b tt th ithe main clause better as their grammar the main clause better as their grammar develops over timedevelops over time

Take Home MessageTake Home MessageTake Home MessageTake Home MessageChildren as young as 3 years do have the Children as young as 3 years do have the y g yy g ystructures available to them to recognize, make structures available to them to recognize, make sense of, and correctly respond to RC questionssense of, and correctly respond to RC questionsTh i l i b i i hTh i l i b i i hThe never violate constraints by associating the The never violate constraints by associating the whwh-- with the verb of the RCwith the verb of the RCUnderstanding of constructions with embeddedUnderstanding of constructions with embeddedUnderstanding of constructions with embedded Understanding of constructions with embedded questions increases over time as grammars questions increases over time as grammars adaptadaptChildren 3Children 3--4 don’t have the same grammars as 4 don’t have the same grammars as adults yet still don’t violate islandsadults yet still don’t violate islands