grammatical profile(s) of monolingual spanish-speaking

GRAMMATICAL PROFILE(S) OF MONOLINGUAL SPANISH-SPEAKING CHILDREN WITH LANGUAGE DISORDERS
Gareth P. Morgan
Alejandra Auza
M. Adelaida Restrepo
Arizona State University, USA
Universidad Autónoma de Querétaro, México
Arizona State University, USA
Introduction
Research Questions
• Grammatical markers in the nominal and verbal
system are problematic for children with specific
language impairment (SLI) (i.e. Leonard, 1998).
• Articles, clitics, derivational morphemes and
subjunctives seem particularly difficult for Spanishspeaking (SS) children with SLI, but results are
inconclusive.
Articles:
• Omissions and substitutions in definite and indefinite
articles (i.e., Restrepo & Gutiérrez-Clellen, 2001; Anderson
1. Are there significant group mean differences
between monolingual SS children with SLI and agematch peers on the total score of an experimental
grammatical morphology measure when covarying
for MLU-W?
2. If so, do the two groups differ significantly on all
subsections?
3. Do any of the grammatical markers qualify as
clinical markers?
& Souto, 2005).
Clitics:
• More omissions than substitutions (i.e., Bosch & Serra,
1997; Jacobson & Schwartz, 2002).
• Gender and number substitutions out-weighed the
omission errors (i.e., Bedore & Leonard, 2005).
Subjunctives:
• Mood selection seems vulnerable for SpanishEnglish bilingual children (i.e., Restrepo & Kruth, 2000,
2001; Restrepo, 2003) although limited research is
available in this area for primarily SS children
(Gutiérrez-Clellen et al, 2000).
Derivational morphemes:
• Vulnerability observed cross-linguistically; children
demonstrated low accuracy in comprehension (i.e.,
Auza, 2001 & 2006; Roseberry & Conell, 1991) and
production of derivational suffixes (i.e.,Windsor and
Hwang, 1999; Ravid, Levy & Ben-Zvi, 2003).
Is there a Clinical Marker for SS SLI?
• Clinical marker is a “linguistic form or principle that
can be shown to be characteristic of children with
SLI” (Rice & Wexler, 1996).
• In this study, a clinical marker is defined as a
score for the whole SLI group of at least -1 SD
below the mean of typically developing children.
• Finding a clinical marker in SS SLI, would allow
investigators to:
• Develop and refine identification measures
• Explore the linguistic nature of the disorder
• Validate accounts of SLI cross linguistically
• However, some studies find variability in the profiles
•Limited grammatical productions and a simpler
linguistic system than their MLU peers (Ingram &
Moorhead, 2002)
•Processing limitations and inconsistent grammatical
errors (i.e., Grela, 2003; Leonard, et al, 2007)
• Procedural memory deficits (Ullman & Pierpont,
2005).
.
Procedures
• Two 30 min. sessions to complete testing.
• Audio-taped and transcribed for reliability.
• Tested by trained native graduate students and SLPs.
Sample Items
Method
Participants
• 49 monolingual SS children from Mexico.
• Attended a private school with no special services.
• 9 children with SLI and 9 age matches ± 3 months.
• Ages ranged from 49 to 83 months (M = 65).
• Non-verbal IQ > 75.
Target
Question
Expected Answer
Articles
¿Con qué chocó el coche?
With what did the car hit?
Con el / un árbol
With the/a tree
Clitics
¿Qué hace la niña con la sandía?
What does the girl do with the
watermelon?
Subjunctive
La casa estaba muy sucia. ¿Qué quería
la mamá que hiciera el niñ o?
The house was very dirty. What did the
mother want that the boy do?
Se la come/La muerde/
se la está comiendo
She eats/is eating/bites it
Que limpie/barra/recoja/
trapee/limpiara/barriera/r
ecogiera/trapeara la casa.
That he cleans/sweeps/
picks up/mops the house.
Está dormido.
He is asleep.
Derivational Este señor se durmió. ¿Còmo està?
Morphemes Está…
This man fell asleep. How is he? He is…
ANOVA of Total Score
**
• Individual variation in profiles indicated that at
least for this sample children with SLI do not
demonstrate a specific linguistic knowledge deficit
(i.e., Bedore & Leonard, 2005; Rice & Wexler, 1996).
• Subjunctives may seem to be a clinical marker in
this sample, however, some children with SLI scored
within 1 SD of the TD children.
F(1,18) = 23.90, p < .001, partial η² = .61
MANOVA
Follow-up ANOVAs
Instruments
• Estructura de Palabras (grammatical morphology) and
Repetición de Oraciones (sentence repetition) subtests
of the Spanish Clinical Evaluation of Language
(Semel, Wiig, & Secord, 1997).
**
• Results are consistent with the variability observed
in previous studies (i.e., Anderson & Souto 2005; Grela, 2003).
*
*
• Non-verbal Inventory of the Kaufman Assessment
Battery for Children 2nd edition (KABC-II; Kaufman
& Kaufman, 2004).
• Variability may be a result of a limited processing
capacity (i.e., Grela, 2003; Leonard, et al, 2007; Owen &
Leonard, 2006) or procedural memory deficit (Ullman
& Pierpont, 2005) impacting production.
Summary and Implications
Wilks’ Λ = .27, F(4,13) = 8.73, p < .01, partial η² = .73
• MLU and grammaticality from a story retell in
Spanish of “If You Give a Mouse a Cookie”
(Numeroff, 1985). Samples were transcribed into
SALT, segmented into C-Units and coded for
grammaticality.
• Experimental Grammatical Morphology Measure
• 40 items, 4 sections: Articles, Clitics,
Subjunctive, & Derivational Morphemes.
• CLOZE format using pictures for each target.
Discussion
• SS children with LI demonstrated deficits in
grammatical morphology compared to age matches
when covarying for MLU-W.
• Significant differences were observed on the total
score of the grammatical measure, but no single
section separated all SLI from TD children. This
evidences a grammar deficit, but doesn’t provide
evidence for a single grammatical clinical marker.
Results
Selection Criteria for SLI
• Scored less than 1 SD below TD sample mean on
subtests of the Spanish CELF.
• Percent of Ungrammatical sentences > 20% on a
narrative language sample.
• Clinical opinion of two native SS SLPs.
• Parent concern.
Selection criteria for TD
• Within normal limits on subtests of the Spanish
CELF.
• Percent of Ungrammatical sentences < 20% on a
narrative language sample.
• Clinical opinion of two native SS SLPs
• No parent concern.
TD SD
Subtest
TD Mean (SD) SLI Mean (SD)
Articles
85.60% (10.10) 54.40% (29.30)
.30
Clitics
87.80% (10.90) 41.10% (26.20)
.60**
Subjunctive
86.70% (12.30) 40.00% (34.60)
.48*
Derivational
Morphemes
73.30% (12.30) 47.80% (21.70)
.37*
.
* = p < .01, ** = p < .001.
Partial η²
This study found no clinical marker for SS SLI,
although it confirmed difficulty in grammatical areas
already identified in the group. This study adds to the
discussion on the nature of SLI, validating the
existence of heterogeneous grammatical profiles
related to the use of articles, clitic pronouns,
subjunctive verbs, and derivational suffixes. Further
studies finding a clinical marker for Spanish will allow
researchers to develop and refine identification
measures and explore the linguistic nature of the
disorder.