Idiom Comprehension in Bilingual and Monolingual Adolescents

1
Idiom Comprehension in Bilingual and Monolingual
Adolescents

• Belinda Fusté-Herrmann, Ph.D., CCC-SLP
• ASHA 2008
• Chicago

2
Key Concepts For This Approach

• It is crucial to understand the language
  processing skills necessary for bilingual
  students to read more proficiently.
• Idiom comprehension provides a unique vantage
  point for examining the underpinnings of text
  comprehension because of shared
  cognitive-linguistic processes.
• Use of unfamiliar idioms supports the goal of
  cultural equivalence (Peña, 2007).

3
Idiom Comprehension as an Innovative Assessment
of Language Impairment

• Minimize cultural bias
• Control familiarity of idioms
• Diminish effects of prior knowledge
• ? By using novel idioms derived from Cain,
  Oakhill, Lemmon (2005)

4
Idiom Comprehension Measure
5
Conceptual Framework The Global Elaboration
Model

• Posits that 2 processes operate when interpreting
  unknown idioms semantic analysis and inference
  from context.
• Developmental Trend Piece by piece (non-global)
  language processing of idioms to global language
  processing.

5
6
The four principles of literal and figurative
(Idioms) reading comprehension

• Inference making from
• single word level
• to sentence level

4) Ability to choose contextual appropriate
meaning from various possible meanings
2) Ability to ignore contextually inappropriate
meanings in favor of contextually appropriate
ones
3) Ability to monitor ones comprehension
(Adapted from Levorato et al., 2004)
7
Purpose

• To investigate cognitive-linguistic processes
• shared by text and idiom comprehension in
• bilingual adolescents and their monolingual
• peers through
• The systematic evaluation of each of the Levorato
  et al. (2004) four principles.
• Simultaneous control of
• 3 variables that affect idiom comprehension
  (familiarity, context, and semantic transparency)
• Decoding ability

7
8
Method

• Mixed quasi-experimental design with between- and
  within-subject variables

• 4 Within - Subjects Variables
• Idiom comprehension measure
• (familiarity, transparency, context)
• b) a reading comprehension task
• c) a multiple meaning (synonym) vocabulary task
• and d) a comprehension monitoring task
• Student Language History Questionnaire for
  Bilinguals.

MONO
BI
All measures were administered in English.
8
9
Inclusion Criteria for Bilinguals
(Spanish-English) and Monolinguals (English Only)

• Have normal or aided hearing adequate for
  understanding oral directions
• Have adequate vision (i.e., normal or corrected)
  to read at least 12 point font
• Not be receiving or eligible for speech and
  language services
• Pass the WJ-III Word Attack (Nonword) Subtest at
  the 9th grade level
• Bilinguals met additional ethnicity and
  educational criteria.

9
10
Participants (N 62)

• Monolinguals (n 31)
• 14 males
• 17 females
• Ages
• 14 10 18 6
• (M 16 4)

• Bilinguals (n 31)
• 12 males
• 19 females
• Ages
• 14 9 17 8
• (M 16 8)

• Students were recruited from 3 rural, public
  schools in 2 districts in west central Florida.

10
11
Research Question Aim 1

• To what extent would each of the three variables
  (Reading Comprehension, Error Detection, Synonym)
  predict the criterion idiom comprehension
  accuracy?

12
Results Aim 1
12
13
Summary of Findings Aim 1

• 1) Idiom comprehension Significantly correlated
  with reading comprehension, error detection,
  synonym measures.
• 2) When groups were collapsed, reading
  comprehension, error detection, and synonym
  measures predicted 33 of the variance in idiom
  comprehension accuracy.
• 3) Error Detection Accounted for most variance
  in idiom comprehension scores.
• 4) Synonym task Best predictor of group
  membership.

13
14
Research Question Aim 2

• Would the performance outcomes on the idiom
  measure differ between the bilingual and
  monolingual adolescents?

14
15
Three-Way Interaction Familiarity,
Transparency, Context
16
Idiom Comprehension Accuracy Familiar Idioms
17
Idiom Comprehension Accuracy Unfamiliar Idioms
18
Results Aim 2

• 1. Everyone performed worse on unfamiliar
  (novel), opaque idioms.
• 2. Monolinguals performed better than Bilinguals
  on familiar idioms.
• 3. Everyone performed better when given
  contextual support, except for the monolinguals
  performance on familiar-transparent idioms
  (memorized).
• 4. Context did not benefit skilled reading
  comprehenders more.
• 5. Participants with poorer reading comprehension
  scores did not choose more literal responses.

19
Research Question Aim 3

• Would bilingual students who
• are more linguistically assimilated
• perform in a significantly different
• manner from bilingual students
• who are less linguistically
• assimilated?

20
Demographics Aim 3
14 (45) Simultaneous 17 (55) Sequential
20
21
Results Aim 3Questionnaire Results and Four
Predictor Variables

• F(4, 26) 3.109, MS 912.747, plt.05.
• Adjusted R Square for this model was .219
• Beta values for idiom comprehension (ß .411)
  and reading comprehension (ß -.472) were both
  significant
• Variable M SD 1 2 3 4
• Quest. 52.06 19.39 -.168 -.432-.365
  -.000
• Predictor variable
• 1. Syn 14.68 1.90 -- .288 .255
  .265
• 2. RC 35.52 3.05 -- .534 .528
• 3. ED 7.03 2.36 --
  .487
• 4. ICM 32.65 3.20 --
• p lt.05
• p lt.01

22
Performance Scores on All Four Measures for
Simultaneous and Sequential Language Learners
23
Summary of Findings Aim 3

• 4 measures predicted questionnaire scores
  (Adjusted R square .219).
• Idiom comprehension and reading comprehension
  were the best predictors of questionnaire scores.
• 3) Negative correlation with RC, ED, Syn no
  correlation with IC.
• 4) Synonym task was not correlated with any of
  the predictor variables or the criterion
  variable.
• 5) Simultaneous bilinguals performed more like
  English-only monolinguals Less variability.

24
DevLex Model of Bilingual Lexical Development
(Hernandez, Li, Mac Whinney, 2005)
L1
L2
L1
L2
Parasitic
Internal Resonance, Competition, Entrenchment
Lexical Breadth
Language Impairment
Lexical Depth
25
Directions for Future Study

• Expand the corpus of novel idioms used in the
  present study- could be a clinical diagnostic
  tool.
• Assess 4 linguistic domains, plus oral language
  proficiency, in both languages longitudinally in
  bilingual adolescents with and without detected
  language impairments.
• Assess implicit speech Indirect speech acts
  (Pinker, 2007) like veiled threats or veiled
  bribes to detect language impairments.

25
26
Conclusion

• The relationship between linguistic domains and
  reading comprehension appears to be mediated most
  powerfully by higher-order skills (e.g.,
  comprehension monitoring) Provide clinical
  support in this area.
• To some degree, amount of time spent talking in
  L1 and L2 acts as a knowledge barometer of
  English polysemy (Snow Kim, 2007) Integrate
  the two lexicons to reduce Big Problem Spaces.

26