Session 3H

Elgort (2011) tested whether the deliberate, or intentional, learning of second language (L2) vocabulary items could lead to the acquisition of implicit as well as explicit knowledge. Using form, semantic, and masked priming tasks, Elgort found that deliberate learning could indeed lead to the acquisition of both implicit and explicit knowledge of L2 vocabulary items. In a replication of Elgort's 2011 study, Elgort and Piasecki (2014) found similar results employing bilingual (L1 and L2) flashcards. Although these studies provide insights into whether the form and meaning aspects of words can be acquired, they do not provide any information as to whether the knowledge that is developed through deliberate learning is available for online use in natural contexts (Elgort, 2019), such as sentence reading in self-paced reading (SPR). To date, no studies have examined the single-word processing of deliberately learned L2 vocabulary using SPR. Instead, previous studies have solely focused on the processing of deliberately-learned multi-(pseudo)word items and idioms using SPR (Obermeier, 2015; Obermeier & Elgort, 2021), reporting mixed findings of implicit vocabulary acquisition through deliberate learning. The present study addressed this gap between the two paradigms (online, contextualized/naturalistic processing versus online, decontextualized processing) and sheds light on the power of deliberate learning for explicit and implicit vocabulary acquisition. 
Twenty-five university students of advanced L2 proficiency from a range of L1 backgrounds participated in the present study. The materials were taken from Elgort and Piasecki (2014), which were adjusted from Elgort (2011). On the first day, the participants completed a learning session online to learn 48 English-like pseudowords from Elgort (2011) and Elgort and Piasecki (2014) as well as tests to determine their lexical processing automaticity and vocabulary size. The participants then completed a spaced-repetition study schedule using online flashcards to learn the pseudowords spanning over a week. On the eighth day, the participants completed several tasks to examine their implicit (form priming and semantic priming) and explicit knowledge (explicit knowledge test) of the learned pseudowords as well as the availability of this knowledge for online use in a natural context (SPR).
The results of the present study provide evidence to the existing research by showing that deliberate learning of single-word L2 vocabulary items can lead to the development of complete lexical representations that are both implicit and explicit in nature and are available for real-world language use. The form priming and semantic priming tasks from Elgort (2011) and Elgort and Piasecki (2014) showed evidence of the explicit and implicit acquisition of pseudowords, replicating the findings of previous studies. By adding SPR as a measure of naturalistic language processing, stronger evidence of acquisition was found through online use of the acquired knowledge in a natural context. In doing so, the present study provides evidence that the deliberate learning of vocabulary items can lead to the acquisition of both explicit and implicit knowledge that can be used during online, natural language processing.

The present study aims to investigate the effects of the degree of lexical coverage on viewing comprehension. Numerous studies have indicated that 98% lexical coverage is needed for adequate reading comprehension (Hu & Nation, 2000; Laufer & Ravenhorst-Kalovski, 2010). For listening comprehension, the figures indicate that to understand informal spoken texts a lower lexical coverage is needed, viz., 90-95% (van Zeeland & Schmitt, 2012; Noreillie, et al., 2018). Regarding viewing, the study by Durbahn, et al. (2020) indicated that there is an almost medium-sized correlation between lexical coverage and viewing comprehension (rs (94) = .39). To date, no study has indicated the lexical coverage needed to understand audio-visual texts. Because imagery may provide semantic support to understand the meaning of unknown words from context (Peters, 2019), 95% coverage is thought to be the level at which learners may have adequate comprehension (Rodgers, 2018); however, this has not been empirically tested yet. 
The present research follows the methodology used by van Zeeland and Schmitt (2011), who, in turn, based their work on Hu and Nation's (2000) design. In the present study, a counterbalanced within-participants design was used, in which different levels of lexical coverage were created by using non-words to replace 5%, 10% or 20% of the words in the video. The result was four viewing clips with the following levels of lexical coverage: 100%, 95%, 90% and 80%. The audiovisual texts used were four 2-minute clips of the first season of the documentary Planet Earth (2006), which is the same documentary used in Rodgers (2018) and Peters (2019). Documentaries were chosen because they have been found to contain more correspondence between oral word presentation and imagery than narrative television genres, making these excerpts appropriate for the present research study. The clips were combined considering the topic and the level of lexical coverage so that each version contained all the topics and all the lexical coverage levels (20 minutes approximately). Seventy-eight EFL university students (L1 = Spanish) viewed the four clips, after which they answered fourteen comprehension questions (seven literal and seven inferential questions). A Friedman test was run to determine whether there were differences in viewing comprehension scores throughout four different levels of lexical coverage, i.e., 100%, 95%, 90% and 80%. Pairwise comparisons were performed with a Bonferroni correction for multiple comparisons. Results indicated that viewing comprehension scores were statistically significantly different at the different points of lexical coverage, χ2(3) = 28.309, p < .001, which means that the degree of lexical coverage had an effect on viewing comprehension. Post hoc analyses revealed statistically significant differences in viewing comprehension scores from 100% (Mdn = 12) to 90% (Mdn = 11) (p < .001) and 80% (Mdn = 10.5) (p < .001) and from 95% (Mdn = 12) to 80% (Mdn = 10.5); all the other combinations were not significantly different. Implications for video-based pedagogy and future research will be interpreted in light of reading and listening research.

The importance of word frequency for vocabulary development is generally agreed on (e.g., Ellis, 2002; Webb, 2014): the more frequently learners encounter a lexical item, the more likely they are to acquire it. As a result, it has been argued that it is beneficial for language learning if vocabulary input is structured in a way that ensures repeated encounters with target vocabulary. In the foreign language classroom, where the input is limited, it can be difficult for the individual teacher to ensure that learners are encountering target words frequently enough (cf. Schmitt, 2019). Teaching materials can thus play a valuable role in the language classroom by being the source of structured vocabulary input, where target vocabulary is systematically recycled. Studies have evaluated teaching materials and their function as learning tools by investigating the amount of recycling in the books, using either the raw frequency of words or a type-token ratio. However, given that some words are inherently more frequent in the language, figures representing the extent of recycling may be misleading, as the recycled items may be words already known to the target students. 
This paper presents a corpus-based textbook analysis where recycling is investigated numerically and the nature of the recycled items is also considered, that is, how frequent the recycled items are in general discourse. The corpus encompasses the texts from five series of intermediate EFL materials used in Swedish secondary schools (years 7–9)­. Lemmas occurring ten or more times were considered frequently recycled and thus possible to learn from the exposure. As the target students are intermediate learners, particular attention was paid to the amount of recycled mid-frequency vocabulary (cf. Schmitt & Schmitt, 2014) in the materials.
The results indicate that although a considerable proportion of the lexical input is recycled ten or more times in the materials, the vast majority of the recycled items are high-frequency words, which students are likely to already know. Considering the target students, this means that the recycling figures cannot be considered favorable for vocabulary development. A conclusion that can be drawn from these results is that investigations of the recurrence of words may have to be complemented with analyses of their frequency profile to understand the vocabulary learning opportunities in language learning materials.


References:
Ellis, N. C. (2002). Frequency effects in language processing: A review with implications for theories of implicit and explicit language acquisition. Studies in Second Language Acquisition, 24(2), 143–188. https://doi.org/10.1017/S0272263102002024 
Schmitt, N. (2019). Understanding vocabulary acquisition, instruction, and assessment: A research agenda. Language Teaching, 52(2), 261–274. https://doi.org/10.1017/S0261444819000053
Schmitt, N., & Schmitt, D. (2014). A reassessment of frequency and vocabulary size in L2 vocabulary teaching. Language Teaching, 47(4), 484–503. https://doi.org/10.1017/S0261444812000018
Webb, S. (2014). Repetition in incidental vocabulary learning. In C. A. Chapelle (Ed.), The encyclopedia of applied linguistics (pp. 1–6). Wiley. https://doi.org/10.1002/9781405198431.wbeal1425

Mastering collocations is one of the most challenging aspects of vocabulary acquisition, especially for adult learners of a second language (L2). Research shows that frequency (exposure to L2 input), congruency (availability of a word-for-word translation in the L1), and semantic transparency (degree to which meaning can be inferred from the constituent words) are important factors that affect L2 collocational processing. Frequency is known to interact with both congruency and semantic transparency. However, congruency and semantic transparency have not been explored simultaneously. A recent study by Yamashita (2018) found that incongruent collocations (available only in the L2) also tend to be more semantically opaque than congruent collocations (available in the L1 and L2). Therefore, this study aims to investigate L2 collocational processing in proficient L2 English speakers (L1 Portuguese) by extending Gyllstad and Wolter (2016), who found that free combinations enjoy a processing advantage over congruent collocations, to explore the role of semantic transparency in conjunction with congruency. Specifically, it investigates whether free combinations (completely transparent items, e.g., drink water, 'beber água'), congruent collocations (kill time, 'matar tempo'), and incongruent collocations (e.g., kick habits; not available in Portuguese) differ in terms of processing by incorporating congruency into Howarth's Continuum Model (Howarth, 1998). A semantic acceptability judgement task was completed by 329 adults (122 Brazilian Portuguese, 108 European Portuguese, and 99 L1 English speakers). The reaction times and error rates were analyzed using Generalized Linear Mixed-effects Models.
The results (see figures below) showed that the L2 speakers processed free combinations faster and more accurately than congruent and incongruent collocations, and congruent collocations faster and more accurately than incongruent collocations. The L1 English group showed no significant processing advantages for free combinations versus congruent collocations, nor for congruent collocations versus incongruent ones. However, surprisingly, a significant processing advantage was found, only in terms of speed, for free combinations versus incongruent collocations. The results lend support to Gyllstad and Wolter's (2016) findings that there may be psychological validity to Howarth's continuum (1998).
Fig 1: RT Model.pdf    Fig 2: ER Model.pdf
Additionally, they seem to support Yamashita's findings that incongruent collocations may be inherently more opaque than congruent collocations. For congruency, this study lends support to the growing body of evidence that L2 speakers incur processing costs for incongruent collocations. This highlights the critical role of the L1 in L2 collocational processing. Based on these findings, recommendations are made for incongruent collocations to be explicitly taught in the L2 classroom. Future research should test incongruent free combinations with incongruent collocations to further isolate these effects.
Keywords: Collocations, Semantic Transparency, Congruency, L1 Portuguese, L2 English, Semantic Acceptability Judgement Task, Generalized Linear Mixed-effects Models
Selected References
Gyllstad, H., & Wolter, B. (2016). Collocational Processing in Light of the Phraseological Continuum Model: Does Semantic Transparency Matter? Language Learning, 66(2), 296–323. https://doi.org/10.1111/lang.12143
Howarth, P. (1998). Phraseology and second language proficiency. Applied Linguistics, 19(1), 24–44. https://doi.org/10.1093/applin/19.1.24
Yamashita, J. (2018). Possibility of semantic involvement in the L1-L2 congruency effect in the processing of L2 collocations. Journal of Second Language Studies, 1(1), 60–78. https://doi.org/10.1075/jsls.17024.yam