Introduction
This essay provides a detailed QALMRI (Question, Alternatives, Logic, Method, Results, Inferences) summary of a study examining second language (L2) predictive processing, specifically whether native Korean speakers learning English can adopt native-like sentence-processing strategies involving verb bias and complementizer cues. The central question explores a fundamental issue in L2 acquisition: whether learners can transcend first language (L1) processing constraints to adopt strategies absent in their native language. English, with its Subject-Verb-Object (SVO) word order, allows early verb placement, facilitating predictive processing, whereas Korean, with its Subject-Object-Verb (SOV) structure, positions verbs sentence-finally, rendering such predictions less viable. This analysis is situated within the broader field of language processing, focusing on words, sentences, and discourse. It aims to outline the study’s research questions, alternative hypotheses, experimental design, findings, and implications for L2 processing theories. By doing so, it contributes to understanding how L2 learners adapt to new linguistic environments and whether sustained exposure and proficiency can reshape entrenched L1-based processing habits.
Research Question and Context
The study addresses both a broad and a specific question in L2 sentence processing. Broadly, it asks whether L2 learners can adopt sentence-processing strategies not used in their L1 (Clahsen and Felser, 2006). More specifically, it investigates whether native Korean speakers learning English can learn to use English verb bias and complementizer cues interactively and predictively during online sentence comprehension, mirroring native speaker behaviour. This question is significant because it tackles the issue of L1 transfer and its potential constraints on L2 acquisition. English’s SVO structure allows verbs to appear early in sentences, enabling speakers to predict upcoming syntactic structures based on verb bias—whether a verb typically takes a direct object (DO-bias, e.g., “read”) or a sentential complement (SC-bias, e.g., “admitted”). In contrast, Korean’s SOV structure delays verb placement until the sentence end, making such predictive strategies less practical in L1 processing (Kim and Osterhout, 2005). Therefore, the ability of Korean-English learners to adopt predictive processing represents a significant departure from their native parsing habits, testing the plasticity of adult language processing systems.
Alternative Hypotheses
The study considers three competing hypotheses regarding how L2 learners might process English verb bias and complementizer cues. The first, the Full Transfer Hypothesis, posits that L1 processing constraints dominate at all proficiency levels, predicting no interaction between verb bias and ambiguity (e.g., presence or absence of the complementizer “that”). Under this view, Korean learners would process sentences uniformly, ignoring verb bias cues due to their L1’s verb-final structure (Hawkins, 2001). The second, the Proficiency-Based Adaptation Hypothesis, suggests that only higher-proficiency learners will show native-like interactions between verb bias and complementizer cues, while lower-proficiency learners will not, reflecting a gradual adaptation process driven by L2 experience (Dussias, 2010). Lastly, the Partial Integration Hypothesis predicts that learners may show sensitivity to verb bias and complementizer cues as separate factors but fail to integrate them predictively during online processing, resulting in main effects without interaction. These alternatives frame the study’s exploration of whether and how L2 learners can overcome L1 constraints, providing a spectrum of possible outcomes based on proficiency and processing integration.
Logic of the Experimental Design
The experimental logic hinges on testing for a verb bias × complementizer interaction in reading times at disambiguating regions of temporarily ambiguous sentences, such as “The student read (that) the article…” Here, DO-bias verbs (e.g., “read”) typically predict a direct object, while SC-bias verbs (e.g., “admitted”) predict a sentential complement. The presence or absence of the complementizer “that” either resolves or maintains ambiguity until later in the sentence. If learners integrate verb bias and complementizer cues predictively, a garden-path effect—longer reading times indicating processing difficulty—should emerge when the verb bias conflicts with the actual structure (e.g., a DO-bias verb followed by a sentential complement without “that”). Conversely, if learners lack predictive integration, no systematic interaction should appear, even if they show sensitivity to individual cues. This design effectively isolates the role of predictive processing, distinguishing between mere sensitivity to linguistic cues and their integrated, anticipatory use in real-time comprehension (Pickering and Garrod, 2013).
Methodological Approach
The study employed a robust experimental design to test its hypotheses. Participants included a control group of native English speakers and two groups of native Korean L2 learners of English, categorised by proficiency: lower and higher proficiency. Groups were matched on biographical variables, though proficiency and length of residence covaried, reflecting typical patterns in L2 acquisition research. The materials consisted of 20 verbs (10 DO-bias and 10 SC-bias) embedded in 80 critical sentences across two conditions: ambiguous (without “that”) and unambiguous (with “that”). These sentences were controlled for verb bias strength, frequency, and noun characteristics to minimise confounding variables (Garnsey et al., 1997). The procedure used a self-paced moving-window paradigm with word-by-word non-cumulative presentation, a standard method for measuring online processing. Comprehension accuracy was monitored to ensure participant engagement. The dependent variable was reading times at disambiguating regions, with the primary analysis focusing on a three-way interaction among verb bias, ambiguity, and participant group. This methodological rigour ensures that differences in processing can be attributed to proficiency and predictive strategies rather than extraneous factors.
Results and Key Findings
The results revealed distinct processing patterns across participant groups. Native English speakers exhibited a significant verb bias × ambiguity interaction, with prominent garden-path effects following DO-bias verbs in ambiguous contexts, consistent with predictive processing (Trueswell et al., 1993). In contrast, lower-proficiency Korean learners showed no interaction, processing sentences uniformly regardless of verb bias or complementizer presence, suggesting reliance on L1-based non-predictive strategies. However, higher-proficiency learners demonstrated a significant interaction resembling the native pattern, indicating predictive sensitivity to verb bias and complementizer cue combinations. Indeed, the key finding was that proficiency significantly modulated L2 parsing strategies, with higher proficiency enabling native-like predictive cue integration. These findings suggest that sustained L2 experience can facilitate the development of processing strategies absent in the L1, challenging rigid transfer accounts.
Inferences and Theoretical Implications
Several important inferences can be drawn from these results. Primarily, higher-proficiency L1-Korean learners can acquire the ability to integrate verb bias and complementizer cues predictively during real-time processing, while lower-proficiency learners struggle to do so. This supports Experience-Based Models of L2 processing, which argue that sufficient L2 exposure can override L1 constraints (MacWhinney, 2005). Furthermore, the findings demonstrate predictive processing plasticity in adult learners, indicating that strategies absent in L1 can be acquired in L2 given appropriate conditions. They also highlight proficiency-dependent adaptation, showing that optimal parsing strategies develop gradually. However, the study’s inferences are limited: while it establishes proficiency-dependent acquisition of predictive processing, it does not confirm whether learners use identical cognitive mechanisms as native speakers or alternative pathways. This ambiguity suggests a need for further investigation into the underlying processes driving L2 parsing.
Future Directions and Broader Significance
The study opens several avenues for future research. Firstly, cross-linguistic generalisation to other verb-final L1 groups could test the universality of these findings. Secondly, identifying exposure thresholds for achieving native-like processing would help refine pedagogical approaches. Additionally, extending research to other predictive linguistic cues beyond verb bias—such as prosody or discourse context—could provide a more comprehensive understanding of L2 processing adaptability. Finally, exploring individual difference factors, such as motivation or working memory capacity, might reveal predictors of successful strategy acquisition. The significance of this study lies in its contribution to understanding how L2 learners transcend L1 limitations through experience, offering an optimistic view of adult language learning capacity while underscoring the importance of sustained exposure and proficiency development (Dussias and Sagarra, 2007).
Conclusion
In summary, this QALMRI analysis of L2 predictive processing of English verb bias demonstrates that native Korean speakers learning English can, with sufficient proficiency, adopt native-like predictive strategies involving verb bias and complementizer cues, despite conflicting L1 word-order patterns. The study’s findings challenge rigid L1-transfer accounts, supporting experience-based models that highlight the role of exposure in reshaping processing habits. While native speakers and higher-proficiency learners showed significant verb bias × ambiguity interactions, lower-proficiency learners did not, underscoring the gradual, proficiency-mediated nature of adaptation. Although limitations exist—particularly regarding the cognitive mechanisms underpinning L2 processing—the results affirm the plasticity of adult language systems. Future research should build on these insights by exploring generalisability, exposure thresholds, and individual differences. Ultimately, this study enriches our understanding of L2 sentence processing, suggesting that dynamic, experience-driven adaptation can enable learners to approximate native-like comprehension over time.
References
- Clahsen, H. and Felser, C. (2006) Grammatical processing in language learners. Applied Psycholinguistics, 27(1), pp. 3-42.
- Dussias, P. E. (2010) Uses of eye-tracking data in second language sentence processing research. Annual Review of Applied Linguistics, 30, pp. 149-166.
- Dussias, P. E. and Sagarra, N. (2007) The effect of exposure on syntactic parsing in Spanish-English bilinguals. Bilingualism: Language and Cognition, 10(1), pp. 101-116.
- Garnsey, S. M., Pearlmutter, N. J., Myers, E. and Lotocky, M. A. (1997) The contributions of verb bias and plausibility to the comprehension of temporarily ambiguous sentences. Journal of Memory and Language, 37(1), pp. 58-93.
- Hawkins, R. (2001) Second Language Syntax: A Generative Introduction. Blackwell Publishing.
- Kim, A. and Osterhout, L. (2005) The independence of combinatory semantic processing: Evidence from event-related potentials. Journal of Memory and Language, 52(2), pp. 205-225.
- MacWhinney, B. (2005) A unified model of language acquisition. In: Kroll, J. F. and De Groot, A. M. B. (eds.) Handbook of Bilingualism: Psycholinguistic Approaches. Oxford University Press, pp. 49-67.
- Pickering, M. J. and Garrod, S. (2013) An integrated theory of language production and comprehension. Behavioral and Brain Sciences, 36(4), pp. 329-347.
- Trueswell, J. C., Tanenhaus, M. K. and Garnsey, S. M. (1993) Verb-specific constraints in sentence processing: Separating effects of lexical preference from garden-paths. Journal of Experimental Psychology: Learning, Memory, and Cognition, 19(3), pp. 528-553.
