Phonological similarity as an index of
short-term memory precision in
monolingual and trilingual speakers
M. Bouffier & S. Majerus
Conference on Multilingualism (COM)
16-18/12/2018
Introduction
Short-term memory (STM)
Often defined in terms of capacity (Miller, 1956; Cowan, 2010)
Limited number of items one is capable of recalling
Binary measure STM performance
Item recalled or forgotten
2
Introduction
STM precision
Resolution of representations in STM (Joseph et al., 2015)
Memory trace degraded, but not fully erased
Quality rather than quantity of representations (Ma et al., 2014)
Flexible resource allocation (Ma et al., 2014)
Mainly studied in the visuo-spatial domain (Bays et al., 2009;
Zokaei et al., 2011; Burnett Heyes et al., 2012; Klyszejko et al., 2014)
Less in the auditory-verbal domain (Joseph et al., 2015; Gilbert et
al., 2017; Clark et al., 2017) 3
Introduction
Use of a phonological similarity gradient between memory and probe items
In monolinguals (Study 1)
In trilinguals (Study 2)
Hypotheses
More errors with increased similarity
Performance impacted by language proficiency
Interdependence between language and verbal STM
Interindividual differences
Study 1
Phonological similarity for French words
Probe recognition task
Participants
60 French-speaking participants (30 women)
18-30 years ( = 22.63; = 2.840)
No neurological disorder or learning disability
5
Study 1: Methods
Accuracy (correct rejection+ hits) 7
Study 1: Resutls
0,4 0,5 0,6 0,7 0,8 0,9 1 1 2 3 4 5 6 A cc ur ac y (Co rrect reject io ns ) PositionStudy 1: Results
Individual differences and Precision scores
-3,5 -2,5 -1,5 -0,5 0,5 1,5 2,5 0 10 20 30 40 50 60 d' Participant ID
Diff d' (high-low similarity)
-3,5 -2,5 -1,5 -0,5 0,5 1,5 2,5 0 10 20 30 40 50 60 d' Participant ID
Diff d' (moderate-low similarity)
Interim summary
Effect of similarity gradient
Interindividual variability
What about precision for L2 and L3?
Varying language representations
Depending on language proficiency
Study 2: Methods
Participants:
35 Belgian German speakers (18 women)
18-30 years ( = 22.11; = 3.35)
Same task…
…in German (L1), French (L2) and English (L3)
Evaluation of receptive (Dunn & Dunn, 1997) and productive
vocabulary (Bonin et al., 2003)
Three testing sessions
10
Study 2: Methods
11
L1
L3
Study 2: Results
Interaction Language * Condition
12 0,7 0,75 0,8 0,85 0,9 0,95 1
Dissimilar Moderately similar Highly similar
A cc ur ac y (Correct reject io ns )
Study 2: Results
Correlation between L2 and L3 mean accuracy and L2 and L3 proficiency 13 0,6 0,65 0,7 0,75 0,8 0,85 0,9 0,95 1 -2 -1,5 -1 -0,5 0 0,5 1 1,5 2 M ean a c c uracy in French a nd Eng lish
Proficiency in French and English R = 0.626
Study 2: Results
Interindividual variability and precision scores
14 -3 -2,5 -2 -1,5 -1 -0,5 0 0,5 1 1,5 2 2,5 0 5 10 15 20 25 30 35 D' Participant ID German
Diff d' (high-low similarity)
-3 -2,5 -2 -1,5 -1 -0,5 0 0,5 1 1,5 2 2,5 0 5 10 15 20 25 30 35 D' Participant ID French
Diff d' (high-low similarity)
-3 -2,5 -2 -1,5 -1 -0,5 0 0,5 1 1,5 2 2,5 0 5 10 15 20 25 30 35 D' Participant ID English
Diff d' (high-low similarity)
Mean: -0.52
Study 2: Results
DE Diff d' (high-low similarity) FR Diff d' (high-low similarity) DE Diff d' (high-low similarity) — FR Diff d' (high-low similarity) -0.262 —ENG Diff d' (high-low
similarity) -0.043 0.012
* BF₁₀ > 10, ** BF₁₀ > 30, *** BF₁₀ > 100
15
General discussion and conclusion
Observed gradients
Phonological similarity
Language
Correlations with L2 and L3 proficiency
Interindividual variability
Correlation between precision scores?
General discussion and conclusion
Sensitive measure of STM performance
Interdependence between language system and verbal
STM (Baddeley et al., 1998; Cowan, 1999; Burgess & Hitch, 2006; Majerus, 2010;
Majerus, 2013) 17
Merci für die attention!
Backslides
Introduction
Some studies already conducted
20
Introduction
Main results
Load effect (Zokaei et al., 2011; Joseph et al., 2015; Clark et al., 2018)
Recency effect (Zokaei et al., 201; Burnett Heyes et al., 2012; Joseph et al., 2015)
Study 2: Results
German: Interaction Condition * Position
22 0,6 0,65 0,7 0,75 0,8 0,85 0,9 0,95 1 1 2 3 4 5 6 A cc ur ac y (co rrect reject io ns ) Position Dissimilar Moderately similar Highly similar BFInteraction: 0.043 BFCondition: 3.157e +8
Study 2: Results
23
French: Interaction Condition * Position
0,6 0,65 0,7 0,75 0,8 0,85 0,9 0,95 1 1 2 3 4 5 6 A cc ur ac y (co rrect reject io ns ) Position Dissimilar Moderately similar
Highly similar BFCondition: 6.005e +15
Study 2: Results
English: Interaction Condition * Position
24 0,6 0,65 0,7 0,75 0,8 0,85 0,9 0,95 1 1 2 3 4 5 6 A cc ur ac y (co rrect reject io ns ) Position Dissimilar Moderately similar