The nature of bilingual Specific Language Impairment SLI

1
The nature of bilingual Specific Language
Impairment(SLI)

• Anne E. Baker Jan de Jong
• University of Amsterdam
• (Collaborators A. Orgassa, N. Çavus F.
  Weerman)
• NWO grant 254-70-010

2
Overview of this talk

• Introduction bilingual SLI
• Two positions on bilingual SLI
• A review of research on bilingual SLI
• The Amsterdam study
• Gender marking in Dutch
• Verb placement in Dutch
• Subject-verb agreement in Turkish and Dutch
• A comparison of Turkish and Dutch
• Conclusions

3
Bilingual SLI
Introduction

• Bilingual SLI is SLI in both languages.
• SLI in only one language is by definition not
  possible
• But SLI may have different linguistic
  manifestations in both languages

4
Current interest in the topic
Introduction

• Social relevance
• Increasing role of multilingualism in society
• Diagnostic dilemmas in clinical practice (Genesee
  et al., 2004) sometimes lead to
• Missed identity
• Mistaken identity
• Theoretical relevance
• Insight into typical development and (S)LI how
  do the factors impairment and bilingualism
  combine?

5
Two theoretical positions (Paradis, 2007)
Introduction

• Generalized slowing hypothesis
• Bilingual SLI lt Monolingual TD in each language
• Bilingual SLI lt Monolingual SLI in each language
• Disruption within delay account
• (e.g. a deficit in tense and/or agreement)
• Bilingual SLI Monolingual SLI in both languages
  

6
Resulting questions
Introduction

• Do bilingualism and SLI have a cumulative effect?
  
• generalized slowing hypothesis
• Or does bilingual SLI resemble monolingual SLI in
  each of the languages?
• disruption within delay

7
Disentangling the effect of bilingualism and
SLIcomparisons
Introduction

• SLI effect
• L1- nonSLI L1-SLI
• L2
• effect
• L2- nonSLI L2-SLI

8
Monolingual SLI versus typical L2
Introduction

• SLI effect
• L1-SLI
• L2
• effect
• L2- nonSLI

vulnerable areas
Paradis Crago (2000) Hakansson (2001)
How to distinguish L2 and SLI effects?
9
Monolingual SLI vs bilingual SLI
Introduction

• SLI effect
• L1-SLI
• L2
• effect
• L2-SLI

Monolingual SLI often no research available
For English as L2 more serious morphosyntactic
problems in the bilinguals than in monolinguals
e.g. Crutchley et al, 1997
10
Monolingual SLI vs Simultaneous bilingual SLI
Introduction

• SLI effect
• L1-SLI
• L2
• effect
• L2-SLI

Mainly similarities between monolingual and
bilingual SLI Minority languages not yet studied
e.g. Paradis et al, 2003 Paradis et al,
2005
11
Bilingual SLI vs Bilingual typical
Introduction

• SLI effect
• L2
• effect
• L2- nonSLI L2-SLI

SLI effects in L1 and L2 symptoms resemble
those of monolingual SLI, e.g. omissions of
morphemes
e.g. Jacobson Schwartz, 2005 Salameh et
al, 2004 Rothweiler, 2007
12
The Amsterdam project Disentangling bilingualism
and SLI (1)

• SLI effect
• L1- nonSLI
• (Dutch) L2
• effect
• L2- nonSLI
• (Turkish-Dutch)

Reveals the effect of L2 in typical development
13
The Amsterdam project Disentangling bilingualism
and SLI (2)

• SLI effect
• L1-SLI
• (Dutch) L2
• effect
• L2-SLI
• (Turkish-Dutch)

Reveals possible cumulative effect of
bilingualism and SLI
14
The Amsterdam project Disentangling bilingualism
and SLI (3)

• SLI effect
• L1- nonSLI L1-SLI
• (Dutch) (Dutch) L2
• effect

Reveals the symptoms of monolingual SLI for the
domains concerned
15
The Amsterdam project Disentangling bilingualism
and SLI (4)

• SLI effect
• L2
• effect
• L2- nonSLI L2-SLI
• gtgt (L1 Turkish, L2 Dutch) ltlt

Reveals the SLI characteristics in
bilinguals Also allows for crosslinguistic
comparison
16
Amsterdam projectResearch design

• A cross-linguistic comparison of grammatical
  morphology
• - in the same bilingual children (thus avoiding
  heterogeneity in the SLI group)
• - whose two languages are typologically
  different
• The aims are
• To identify markers of SLI
• To identify markers of (typical) L2 development
• To gain insight in the interaction of SLI and L2

17
Typological differences between L1 and L2
Method
18
Linguistic variables and elicitation
Method

• Dutch
• Subject-verb agreement verb placement
• Adjectival inflection and gender selection
• How?
• Elicitation
• Frog story

• Turkish
• Subject-verb agreement
• Case marking and word order
• Case marking and definiteness
• Genitive-Possessive construction
• How?
• Elicitation
• Grammatical judgment
• Frog story

19
In this presentation
Method

• Dutch
• Subject-verb agreement
• Verb placement
• Adjectival inflection and gender selection

• Turkish
• Subject-verb agreement
• Case marking and word order
• Case marking and definiteness
• Genitive-Possessive construction

20
Selection of variables
Method

• Dutch
• Subject-verb marking is vulnerable in SLI in many
  languages, e.g. Dutch (de Jong, 1999)
• Adjectival inflection gender selection a
  well-known obstacle for L2 learners of Dutch
• Turkish
• No previous research on grammatical symptoms of
  SLI in (monolingual) Turkish ? explorative study.
  
• Given the findings for SLI in different
  (typologically different) languages, the focus is
  on grammatical morphology

21
Study on gender in Dutch Research questions
Gender Dutch

• Does bilingualism influence the acquisition of
  gender agreement?
• Does SLI influence the acquisition of gender
  agreement?
• Is there a cumulative effect in bilingual SLI ?

22
The Dutch determiner system
Gender Dutch

• Definite determiner (de/het)
• De fiets (the bike) - definite, singular, common
• Het kado (the present) - definite, singular,
  neuter
• Indefinite determiner (een)
• Een fiets (a bike) - indefinite, singular, common
• Een kado (a present) - indefinite, singular,
  neuter

23
Adjectival inflection in Dutch
Gender Dutch

• 1. Definite context (schwa)
• De grote fiets (the big bike) - definite,
  singular, common
• Het grote kado (the big present)- definite,
  singular, neuter
• 2. Indefinite context (schwa vs. ø)
• Een grote fiets (a big bike) indefinite,
  singular, common
• Een groot kado (a big present) indefinite,
  singular, neuter (the exceptional case)
• The rule is
• Always schwa, except for the exceptional case

24
Acquisition of gender agreement
Gender Dutch

• Lexical gender assignment Target-like around 70
• Early patterns and errors
• No determiner (ø mes ø hondje)
• De as a default form (de mes de hondje)
• Low use of het (het mes het hondje)
• Grammatical gender Target-like around 70
• Error schwa default (een grote hondje)

25
Subjects
Gender Dutch
26
Elicitation task 1 Lexical gender
Gender Dutch

• Kijk, een schoen. Waar is Kikker?
• Look, a shoe. Wheres Frog?
• Kikker zit in de schoen.
• Frogs in the shoe

• Dat zijn Kikker en Konijn.
• Here are Frog and Rabbit

27
Elicitation task 2 Grammatical gender
Gender Dutch

• Kijk, twee messen. Look, two knives
• Dat is een groot mes en dat is een klein mes.
  Thats a big knife and thats a small knife

• Waar is Kikker? Wheres Frog?
• Kikker zit op het grote mes.
• Frogs on the big knife

28
Lexical gender common ()De auto
Gender Dutch
29
Lexical gender neuter () Het paard
Gender Dutch
30
Grammatical gender (definite)de groene appel
het grote paard ()
Gender Dutch
31
Grammatical gender (indefinite) een groene
appel een groot paard ()
Gender Dutch
32
Summary of lexical and grammatical gender
Gender Dutch
Dutch L1 gt Dutch L1-SLI gt Bilingual typ gt
Bilingual SLI

• Dutch L1 typical gt Bilingual groups
• Dutch SLI gt Bilingual groups
• There is a cumulative effect in bilingual SLI
  group

33
The Dutch inflectional paradigm
Verb inflection Dutch

34
Dutch verb placement (SOV/V2)
Verb inflection Dutch

• In main clause
• Ik-SUBJ lees de krant-OBJ vandaag
• I read the paper today

35
Possible errors
Verb inflection Dutch

• -ø Hij lees boek (omission of
  marker)
• error in SLI and early TD
• -t De kinderen drinkt cola (substitution)
• error in SLI and early TD
• -en Jij lezen een krant (substitution)
• error in adult L2 learners
• -en Jij een krant lezen
  (infinitive, final) error in SLI and early TD

36
Subjects
Verb Inflection Dutch
37
Elicitation task(lezen, read)
Verb inflection Dutch
De vrouw leest een boek en de man leest een
krant The woman reads a book and
the man reads a paper
38
SLI-effect 1 Finite Lexical verbs ( correct)
Verb inflection Dutch
Significant group difference for inflection
Typical versus SLI
39
SLI-effect 2 Finite Lexical verbs ( correct)
Verb inflection Dutch
Significant group difference Typical versus SLI
40
L2-effect 1 Finite Lexical verbs ( correct)
Verb inflection Dutch
Significant group difference Monolingual versus
bilingual
41
L2-effect 2 Finite lexical verbs ( correct)
Verb inflection Dutch
Significant group difference Monolingual versus
bilingual
42
Summary
Verb inflection Dutch

• Verb placement high correctness rates in all
  groups ? evidence for knowledge of verb
  placement.
• Verb inflection
• ? SLI effects significant differences between
  SLI non-SLI groups
• ? L2 effects significant differences between L1
  and L2 groups

Dutch L1 gt Bilingual typ gt Dutch L1-SLI gt
Bilingual SLI
43
Turkish elicitation of subject-verb agreement
Verb inflection Turkish

• Target forms person marking after past tense -
  dI

44
The Turkish inflectional paradigm
Verb inflection Turkish

45
Elicitation task(sikmak, squeeze)
Verb inflection Turkish

• Anne ben portakal-i _____
  (sik-ti-m).
• Mummy I orange-ACC ____
  (press-PST.DI-1SG)
• Mummy, I have squeezed an orange.

46
Finite lexical verbs ( correct)
Verb inflection Turkish
47
Conclusion
Verb inflection Turkish

• Inflection is acquired fast in Turkish
• The SLI children, as a group, fall behind the
  typically developing children

48
Cross-linguistic comparison
Conclusions

• Bilingual children with SLI as a group
  perform worse than the typically developing
  bilinguals in both languages,
• BUT the SLI bilinguals are better at
  producing Turkish morphology than Dutch
  morphology.

49
Disentangling bilingualism and SLI
Conclusions

• The two factors can be distinguished
• Verb inflection is a marker of SLI
• Gender assignment and Adjectival agreement
  constitute a marker of L2
• But the difference between bilingual typical and
  bilingual SLI on the Dutch measures suggests that
  there is a cumulative effect at work in the SLI
  bilingual group

50
Conclusions

• (1) Does bilingual SLI show a cumulative effect?
  (generalized slowing hypothesis)
• Yes, given the differences between bilingual
  children with SLI and
• (a) bilingual children without SLI
• (b) monolingual Dutch children with SLI

51
Conclusions

• (2) Does bilingual SLI resemble monolingual SLI
  in each of the languages (disruption within
  delay)?
• This cannot be established for Turkish no
  information on Turkish monolingual SLI
• For Dutch No, there are differences between
  monolingual and typical SLI in problem areas (see
  gender) and severity of problem

52
What are the components of the L2 effect?

• Age of first exposure to L2
• Length of exposure to L2
• Quantity and quality of L2 input
• Language choice / dominance
• The population is heterogeneous, so the influence
  of these variables must be taken into account

53
Subject-verb agreement in Turkish and Dutch
within the SLI group ( gt90)
54
Language dominance?
55

• Cumulative effect of bilingual SLI is clear.
• The influence of language dominance and other L2
  factors have to be further explored.
• to be continued.

56
http//home.medewerker.uva.nl/a.e.bakerhttp//hom
e.medewerker.uva.nl/j.dejong1/

• a.e.baker_at_uva.nl
• j.dejong1_at_uva.nl