Distributive Reading and Conceptual Plurality in Second Language Acquisition / J-SLA2024

Tamura, Y., Fukuta, J., & Kimura, T. (2024). Distributive reading and conceptual plurality in second language acquisition. The 24th International Conference of the Japan Second Language Acquisition. Osaka Kyouiku University, Osaka.

Transcript

1. June 23, 2024 The 24th International Conference of the Japan Second Language Association (J-SLA2024) Tamura Yu (Kansai University) Fukuta Junya (Chuo University) Takayuki Kimura (Utsunomiya University) Distributive Reading and Conceptual Plurality in Second Language Acquisition
2. Purpose of the Study • Exploring the processing of conceptual plurality in second language (L2) learners • Focus on distributive reading using the number judgment task 3
3. Background 4
4. Previous L2 research • Previous L2 research on the processing of conceptual number focused on distributive effect on number agreement production • Typical Sentence Completion Task ＋ red The label on the bottles ＋ self-paced 600ms 1800ms (beep) 5
5. Factors influencing the distributive effects Previous L2 research • Learner Proficiency • Hoshino et al. (2010) found that English-Spanish bilinguals showed sensitivity to both grammatical and conceptual information in L1, indicating that grammatical agreement is influenced by conceptual number • In L2, the influence of conceptual information on agreement errors was only observed in learners with very high proficiency levels • Learners’ L1 • Wei et al. (2015) found that the presence or absence of required subject-verb agreement in the L1 significantly affects the likelihood of observing a distributive effect in L2 English • When the L1 requires subject-verb agreement (like Uygur), the distributive effect is more likely to occur in L2 English. Conversely, when the L1 lacks subject-verb agreement (like Chinese), the distributive effect is less likely to emerge in L2 English 6
6. Factors influencing the distributive effects Previous L2 research • Conceptual number salience • The way in which the sentence fragments are presented can impact the distributive effect • Presenting pictures that emphasize the conceptual number of the subject phrase (as in Wei et al. 2015, Experiment 3 and Foote, 2010) led to a distributive effect, even in participants whose L1 lacks subject-verb agreement, compared to presenting the fragments alone Wei et al. (2015) Foote (2010) 7
7. Distributive reading A. Each of the men carried a large box. B. Together the men carried a large box. C. Each of the men carried some large boxes. D. Together the men carried some large boxes. 8
8. Both end with the same singular noun, but… Key Comparisons Each of the men carried a large box. Together the men carried a large box. Distributed reading Collective reading Multiple objects are represented 9
9. Patson & Warren (2010)’s Number Judgment Task Methodology to detect distributive reading • Design: Sentences ending with either a singular or plural noun (e.g., "The bartender served the beer to the man/men"). • Procedure: Participants read sentences presented in chunks and judged whether the final word was one-word or two-word. • Results: Participants were slower to judge that 1 word was on the screen when it was plural (M = 1,133 ms) compared to singular (M = 1,075 ms), confirming the interference hypothesis. 10
10. Patson & Warren (2010) Experiment 2 • Design: Sentences ending with either a singular or plural noun (e.g., "Each of/Together the men carried a large box/some large boxes"). • Procedure: Participants read sentences presented in chunks and judged whether the final word was one-word or two-words. 11
11. ns p. < .05 12
12. Importance of Investigating Distributive Reading in L2 • Goes beyond previous studies focusing on distributive effects in number agreement • Explores L2 learners' ability to access and use distributive readings • Identifies potential challenges unique to L2 learners. • Provides a comprehensive view of linguistic and cognitive mechanisms in L2 acquisition • Complements existing research on number agreement errors 13
13. The Present Study 1 4
14. Research Question • Can L2 learners of English access representations of conceptual plurality through distributive reading? 15
15. Participants • 97 Japanese learners of English • Mean Age: 18.8 (SD = 0.55) • Mean TOEFL score: 507.43 (SD = 26.88) • Male: 31, Female: 65 , Other: 1 16
16. Materials • Mostly based on the materials used in Patson & Warren (2010) • Revised difficult vocabulary • 36 target items (8 items for each) plus 64 distractors (100 items in total) A. Each of the men carried a large box. B. Together the men carried a large box. C. Each of the men carried some large boxes. D. Together the men carried some large boxes. 17
17. Materials (cont.) • Four counterbalanced lists • Random presentation • Simple T/F comprehension question after number judgment 18
18. Experimental Task • Online Stroop-like number judgment task using Gorilla Experiment Builder • One- or two-word chunks presented randomly • Participants judged the number of words in blue font • Four practice items before the main task • The main task lasted about 20-30 minutes 19
20. Analysis • Generalized linear-mixed effect model • Family: inverse.gaussian (identity function) • Response variable: Raw RT of number judgment • Explanatory variable: • distributivity (Each; -0.5, Together; 0.5) • number (sg; -0.5, pl; 0.5) • Interaction between distributivity and number 21
21. Covariates Analysis • Surface frequency • singular: frequency of singular nouns • plural: frequency of plural nouns • Base frequency (sum frequency of singular and plural) • Frequency information was obtained from SUBTLEX US Corpus (Brysbaert and New, 2009) • All the frequency measures are in Zipf scale: “log10 (frequency per million words)+3” (Van Heuven, Mandera, Keuleers, & Brysbaert, 2014, p.1179) 22
22. NOTE Analysis • 8 participants with less than 70% accuracy on comprehension questions were removed from the analysis (M = 84.15; SD = 5.28) • 1 participant with the accuracy of 50% on number judgment was removed (M = 98.8; SD = 2.3) • Due to the technical issue, only 36 of the 96 participants’ TOEFL scores were linked to the experimental data • Therefore, the data from only 36 learners were analyzed and reported today (If we have time, we can show results of the analysis using the data from all the learners) 23
23. Outlier Removal Analysis • By visual inspection, the 99th percentile (around 4000ms) was determined as the cut-off point • Responses with reaction times greater than the mean plus three standard deviations (M + 3SD) of individual RTs were removed • In total, 2.49% of the responses were removed 24
24. Results 25
25. GLMM Results • Model comparison: a) three-way interaction (distibutivity×number×proficiency) b) two-way interaction (number×proficiency) plus proficiency as a covariate • (b) showed a better fit (AIC: 18226 vs. 18231) 26
26. Results • Significant main effect of distributivity on RTs • Planned pairwise comparison: L2 learners judged singular nouns slower in "each" sentences than in "together" sentences (p = .027) • Suggests distributive reading is available to L2 learners • L2 learners conceptually represent multiple objects when processing singular nouns in distributive reading, interfering with number judgment 27
27. With TOEFL p = .179 p = .027 28
28. Results A. Each of the men carried a large box. B. Together the men carried a large box. C. Each of the men carried some large boxes. D. Together the men carried some large boxes. - A(each/sg) = C (each/pl) - B(together/sg) = D(together/pl) - A(each/sg) > B(together/sg) - C(each/pl) = D(together/pl) 29
29. Discussion 30
30. Proficiency Levels in Participants Discussion • Current study found that L2 learners can process conceptual plurality through distributive reading • Our participants' proficiency levels are likely lower compared to those in previous studies (e.g., Wei et al., 2015) • The influence of proficiency is crucial for understanding the effects on number agreement • The effects observed in previous research may be more related to agreement issues rather than purely conceptual number representation 31
31. Influence of L1 Discussion • Previous research that suggested L1 influence focused on number agreement (Hoshino et al., 2012; Wei et al., 2015) • The influence of L1 on interpreting "each" and distributive readings remains unexamined • Future research should include participants with different L1 backgrounds for comprehensive insights. 32
32. Conceptual Number Salience Discussion • Different from number agreement, conceptual number salience was suggested even without visual support, contrary to the explanation offered by Wei et al. (2015) • L2 learners may represent singular items as plural in distributive reading, Indicating the cognitive flexibility in representing number concepts in L2 33
33. Future Research Directions Discussion • Number Agreement and Representation • Number agreement remains a fascinating research area • Focusing on number representation offers valuable insights into cognitive processing in L2 34
34. Future Research Directions Discussion • Experimental Approaches • Our experiment and the number judgment task can be used in future L2 studies • These methods can help explore various aspects of number representation in different L2 contexts • Encourages further research to utilize these approaches for deeper understanding 35
35. Conclusion 36
36. Conclusion • L2 learners can access representations of conceptual plurality through distributive reading • Provides insights into the processing of conceptual number in L2 acquisition • Further research is needed to understand the underlying mechanisms and factors influencing conceptual plurality processing in L2 learners 37
37. References • Foote, R. (2010). Age of acquisition and proficiency as factors in language production: Agreement in bilinguals. Bilingualism: Language and Cognition, 13(2), 99–118. https://doi.org/10.1017/S136672890999040X • Hoshino, N., Dussias, P. E., & Kroll, J. F. (2010). Processing subject–verb agreement in a second language depends on proficiency. Bilingualism: Language and Cognition, 13(02), 87. https://doi.org/10.1017/S1366728909990034 • Hoshino, N., Kroll, J. F., & Dussias, P. E. (2012). Psycholinguistic Perspectives on Second Language Speech Production. In M. Sanz & J. M. Igoa González (Eds.), Applying language science to language pedagogy. Contributions of linguistics and psycholinguistics to language teaching (pp. 107–130). Cambridge Scholars Publishing. • Patson, N. D., & Warren, T. (2010). Evidence for distributivity effects in comprehension. Journal of Experimental Psychology: Learning, Memory, and Cognition, 36(3), 782–789. https://doi.org/10.1037/a0018783 • Wei, X., Chen, B., Liang, L., & Dunlap, S. (2015). Native Language Influence on the Distributive Effect in Producing Second Language Subject–Verb Agreement. Quarterly Journal of Experimental Psychology, 68(12), 2370–2383. https://doi.org/10.1080/17470218.2015.1014821 38
38. p = .074 p = .050 Without TOEFL p = .670 39
39. Results A. Each of the men carried a large box. B. Together the men carried a large box. C. Each of the men carried some large boxes. D. Together the men carried some large boxes. - A(each/sg) > C (each/pl) - B(together/sg) = D(together/pl) - A(each/sg) = B(together/sg) - C(each/pl) = D(together/pl) 40
40. Speculative Interpretation • Reaction times for A (each/sg) were slower than for C (each/pl). • Without considering distributivity, plural nouns typically result in slower judgments due to inherent complexity. • Both singular and plural nouns should induce conceptual plurality, potentially equalizing reaction times. Slower responses to singular nouns (A) might be due to participants' difficulty interpreting "each" + singular nouns as creating a distributive reading. 41