Session 1C

Previous research has shown that the percentage of words learners know in a text, i.e. lexical coverage, has an important effect on second language (L2) reading comprehension and lexical inferencing. It has been argued that a high percentage of lexical coverage is needed for successful comprehension of a text, with figures ranging from 90% - 98% depending on the level of comprehension required (e.g., Hu & Nation, 2000; Laufer & Ravenhorst-Kalovski, 2010). These lexical coverage figures are often used for the selection of reading materials and for setting vocabulary learning targets in language courses. However, no previous studies have examined the effect of lexical coverage on vocabulary learning. In addition, little is known about the effect that the amount of unknown word density has on the cognitive effort involved in reading. These estimates are based on offline measures, i.e. post-reading tests, which do not provide direct evidence of the potential processing and decoding difficulties that readers have when reading at different levels of lexical coverage. The present study addresses these gaps by examining the effect of different percentages of lexical coverage on vocabulary learning and cognitive effort, using a combination of online and offline measures, i.e. vocabulary test and eye-movement recordings. The main objectives of the present investigation are twofold: 1) to investigate the effect that different percentages of lexical coverage have on vocabulary learning; 2) and to examine the effect of different lexical coverages on reading fluency and cognitive load, as measured by readers’ eye movements. Seventy-nine L2 advanced speakers of English were asked to read a text with one of four levels of lexical coverage (90%, 95%, 98%, or 100%) while their eye movements were recorded (EyeLink 1000+). The four different versions of the same text with different levels of lexical coverage were created by replacing words with unknown pseudowords in the first three versions. Participants were randomly assigned to one of these four reading conditions. After the eye-tracking experiment, participants in the first three groups (90%, 95%, and 98%) completed a vocabulary test (meaning recall). Global eye-movement measures were examined, including number of fixations, number of saccades and average fixation duration. Regarding offline data, results of Kruskal Wallis tests showed that learners’ vocabulary learning was significantly higher in the 95% than in the other two conditions. Concerning learners’ eye-movements, results of Kruskal Wallis and one-way ANOVA tests showed that there were no significant differences across conditions in any of the eye-movement measures examined. All percentages of lexical coverage examined led to similar online reading behaviour. This combination of offline and online measures provides the best account to date of the cognitive effort that is required by learners when reading texts of different lexical coverages, as well as its effects on vocabulary learning. Pedagogical implications will also be discussed. References: Hu, H. & Nation, P. (2000). Unknown vocabulary density and reading comprehension. Reading in a Foreign Language, 13(1), 403-431. Laufer, B. & Ravenhorst-Kalovski, G. (2010). Lexical threshold revisited: Lexical text coverage, learner’s vocabulary size and reading comprehension. Reading in a Foreign Language, 22, 15–30

Introduction: Aspects of a second language (L2) that differ from the native language (L1) pose a particular challenge for learners. Cross-language congruency effects are well attested in the realm of multiword units, such as in L1-L2 incongruent collocations (i.e., conventional phrases such as “gastar una broma” ‘spend a joke’, equivalent to ‘play a joke’). Research has shown that L1-L2 incongruent collocations pose challenges in second language processing (Wolter & Gyllstad, 2011), production (Nesselhauf, 2003) and learning (Peters, 2016; Pulido & Dussias, 2019). But some have proposed that the ability to regulate competition from the L1 may improve learning. A recent behavioral study investigating the learning of incongruent collocations (Pulido & Dussias, 2019) showed that experience rejecting L1-related distractors resulted in significantly higher recall rates, relative to learners who saw unrelated distractors. By using brain event related potentials (ERPs), this study aimed to examine the specific mechanisms associated with successful learning of L1-L2 incongruent collocations. In particular, it addressed the following questions: 1. What ERP components are elicited by retrieval Practice in conditions of low interference (unrelated distractors) or high interference (when L1-related distractors are presented)? 2. Is the neurophysiological activity elicited during Practice associated with recall rates in post-tests? Methodology: L1-speakers of English (N=45) who were intermediate learners of Spanish participated in a learning study of verb-noun collocations. The materials consisted of 28 L1-L2 incongruent collocations (“gastar una broma” ‘play a joke’, literally ‘spend a joke’) and 14 congruent collocations (“jugar un partido”, literally ‘play a match’). Learners were first familiarized with the collocations and then completed a Practice procedure. They were presented with two verbs followed by a noun and were required to respond manually by pressing a button corresponding to location of the verb (to the left or right of a fixation cross). Immediate feedback was provided after each response. In critical trials, learners in the “Unrelated” condition (N=22) saw unrelated distractors, e.g. ordenar – gastar – broma (correct response “gastar” ‘spend’), while learners in the “Interference” condition (N=23) had to correctly discard the L1-equivalent verb, e.g. jugar – gastar – broma (i.e., discard “jugar” ‘play’ to choose “gastar” ‘spend’). Order of presentation of the verbs was counterbalanced (50% target-distractor-noun, 50% distractor-target-noun). Success of learning was assessed with gap-fill and translation tests, which were completed immediately after practice and after a one-week delay. ERPs were time-locked to the presentation of the noun in each trial. Results: ERPs for the “Unrelated” group presented an N400 effect for incongruent collocations, followed by a right frontal negativity associated with response selection (600-800 ms). Critically, the “Interference” group showed a reversed N400, indicating inhibition (a more negative N400) for congruent collocations. Individual ERP means showed a significant correlation between inhibition of congruent items (i.e., more negative N400) and higher accuracy in recall in the immediate Translation test (r = 0.31, p < 0.05), and marginally significant in the gap-fill test (r = 0.28, p < 0.07). The results support the hypothesis that native language regulation plays a role in foreign language learning. References: Nesselhauf, N. (2003). The use of collocations by advanced learners of English. Applied Linguistics, 24(2), 223–242. Peters, E. (2016). The learning burden of collocations?: The role of interlexical and intralexical factors. https://doi.org/10.1177/1362168814568131 Pulido, M. F., & Dussias, P. E. (2019). Desirable difficulties while learning collocations in a second language: Conditions that induce L1 interference improve learning. Bilingualism. https://doi.org/10.1017/S1366728919000622 Wolter, B., & Gyllstad, H. (2011). Collocational links in the L2 mental lexicon and the influence of l1 intralexical knowledge. Applied Linguistics, 32(4), 430–449. https://doi.org/10.1093/applin/amr011

Mastering formulaic sequences is a crucial part of second language acquisition due to the processing advantage they imply (Myles & Cordier 2016) and the communicative functions they perform (Wray 2002). At the same time, formulaicity is recognized as a problematic dimension of L2 acquisition, for statistical, structural and pragmatic reasons (Pawley & Sider 1983; Read & Nation 2004). Among other factors, formulaic sequences’ ubiquity in the language makes it unviable to learn them explicitly, while purely incidental conditions result in slow and ineffective learning, due to the single sequences’ high dispersion (Conklin & Schmitt 2012). Such features point to the need to speed up incidental learning, while not hampering desirable elements, such as the focus on meaning, which is a requirement for the creation of implicit knowledge (Paradis 2004). The present study set out to achieve this aim through enhanced incidental learning, i.e. by adding input enhancement to incidental learning conditions, such as reading while listening. Therefore, the experiment had two main goals, on both a pedagogical and a psycholinguistic level. From a pedagogical perspective, it addressed the effectiveness of enhanced incidental learning for the acquisition of formulaic sequences. From a psycholinguistic angle, it investigated the amount of consciousness involved under enhanced incidental learning conditions, as well as the nature (implicit or explicit) of the knowledge gained. 83 Chinese learners of Italian L2 were exposed to reading-while-listening to a graded reader including seven occurrences of each target idiom (input flood). Participants were randomly assigned to four experimental groups and exposed to the target items with (i) visual, (ii) aural, (iii) visual+aural, or (iv) no enhancement. A control group performed the tests with no treatment. Learning was assessed through both offline (receptive and productive assessments of form and meaning) and online (self-paced reading) tests, which were performed immediately after the treatment and again three weeks later. In order to investigate the relationship between level of consciousness at the point of learning and kind of knowledge gained, a subsample of participants had their eye movements recorded at the process level. In addition, stimulated recalls provided information about participants’ awareness of the enhancement device and of the learning task. Significant knowledge was gained, and subjects exposed to enhancements outperformed the no-enhancement group on both online and offline measures except for the self-paced reading delayed posttest. Eye-tracking measures displayed a learning effect (decrease in the number and length of fixations) as the encounters increased for subjects exposed to enhancements, but not for the no-enhancement group. Visual and aural enhancements resulted in similar patterns, despite the fact that according to the stimulated recalls, subjects noticed visual enhancement but not aural enhancement. Such finding can contribute to the debate about the relationship between awareness and learning. Conklin, K. & Schmitt, N. (2012). The processing of formulaic language, Annual Review of Applied Linguistics, 32, 45-61. Myles, F. & Cordier, C. (2017). Formulaic sequence (FS) cannot be an umbrella term in SLA. Studies in Second Language Acquisition, 39, 3-28. Paradis, M. (2004). Neurolinguistic Theory of Bilingualism. Amsterdam: John Benjamins. Pawley, A. & Syder, F.H. (1983). Two puzzles for linguistic theory: nativelike selection and nativelike fluency. In J.C. Richards & R.W. Schmidt, eds., Language and communication. New York: Longman, 191-226 Read, J. & Nation, I.S.P. (2004). Measurement of formulaic sequences. In N. Schmitt, ed., Formulaic Sequences. Amsterdam: John Benjamins: 23-35. Wray, A. (2002). Formulaic language and the lexicon. Cambridge: Cambridge University Press.