Research in second language acquisition has consistently demonstrated the importance of vocabulary knowledge in language learning (e.g., González-Fernández Milton, 2013; Nation, 2001; Rose et al., 2020; Staehr, 2008; Trenkic Urdaniz Vidal, 2011). Cognates (i.e., words that share similar form and meaning across languages) represent a lexical resource that can facilitate learning through cross-linguistic similarities (e.g., Cenoz et al., 2022; Engemann et al., 2024; Puimège Szabo, 2020). Despite this potential, research remains limited on how cognates are recognised and translated in underresearched language pairs such as English and Malaysian vernaculars. In addition, findings on the effectiveness of explicit instruction on cognates have been inconclusive, raising questions about whether cross-linguistic similarity alone is sufficient or whether learners benefit from targeted training to fully exploit this resource. This thesis sets out to address this gap by quantifying the proportion of cognates, examining their use in translation, and investigating the extent to which they are recognised by Malaysian ESL learners. A cognate list was developed for English-Malay and EnglishChinese language pairs in Studies 1 and 3 respectively, identifying cognates, false cognates and non-cognates within the 10,000 most frequent English word families (BNC/COCA; Nation, 2012). The analyses revealed a higher proportion of English-Malay cognates (approximately 26%) compared to false cognates (approximately 1%). While EnglishChinese cognates (approximately 2.1%) were less prevalent, these still outnumbered false his thesis sets out to address this gap by quantifying the proportion of cognates, examining their use in translation, and investigating the extent to which they are recognised by Malaysian ESL learners. A cognate list was developed for English-Malay and EnglishChinese language pairs in Studies 1 and 3 respectively, identifying cognates, false cognates and non-cognates within the 10,000 most frequent English word families (BNC/COCA; Nation, 2012). The analyses revealed a higher proportion of English-Malay cognates (approximately 26%) compared to false cognates (approximately 1%). While EnglishChinese cognates (approximately 2.1%) were less prevalent, these still outnumbered false. Although cognates can be identified through formal criteria and dictionary definitions, it remains uncertain whether Malaysian ESL learners perceive and process the identified cognates as such. To examine whether these cognates are known and used by learners, two bidirectional translation tasks were conducted with Malay-English bilinguals (N = 52) in Study 2 and Chinese-English bilinguals (N = 50) in Study 4. Participants completed a forward and backward translation task with 400 Malay words and 170 Chinese words respectively, which controlled for cognateness, word frequency, and word length. The results indicated that English-Malay cognates were translated more accurately than non-cognates in both translation directions, whereas the cognate advantage in translation accuracy for ChineseEnglish pair was detected only in the backward translation direction. Multiple regression analyses across these translation directions confirmed that cognateness and word frequency were significant predictors of translation accuracy, with cognateness explaining additional variance for both language pairs. Furthermore, there was a consistent tendency for participants to produce cognates as translation equivalents in both directions and across language pairs. These findings confirm prior studies that cognates are generally processed more easily than non-cognates in translation tasks (e.g., Jacobs et al., 2016; Lijewska, 2020; Lijewska Zheng et al., 2024) and also expand current knowledge within the Malaysian context. While previous studies have acknowledged the potential of cognates in supporting vocabulary learning (e.g., Bosch Mitchell et al., 2024; Peters Szabo, 2016; 2018), others argue that explicit instruction is necessary for learners to be aware of cognates, especially when the form and meaning overlap is partial (e.g., Aguinaga Echeverría, 2017; Frances et al., 2021; Otwinowska, 2016). This raises question as to whether Malaysian ESL learners notice the cross-linguistic similarities in cognates automatically, or is explicit instruction required to help them recognise and utilise these connections. Building on the findings of Studies 2 and 4, Study 5 employed a cognate form-meaning recognition study with a pre-post-test quasiexperimental design to assess whether L1 Malay and L1 Chinese ESL learners recognise cognates more accurately than non-cognates, and the extent to which explicit instruction can improve their ability to identify and utilise these similarities. Ninety L1 Malay and 112 L1 Chinese ESL secondary school students were assigned to either an experimental group (with explicit instruction provided) or a control group and completed a pre-test and a post-test. The results indicated that while explicit instruction significantly improved cognate awareness, no significant differences in the post-test performance were observed between the experimental and control groups. Across both language groups, participants recognised cognates more accurately than non-cognates. Generalised mixed-effects modelling showed that cognateness, word frequency, and English language proficiency were significant predictors of cognate recognition, underscoring the importance of both word frequency and cognateness in vocabulary processing. This confirms previous studies that cross-linguistic similarity supports vocabulary recognition and cognate recognition may occur automatically (e.g., de Vos et al., 2019; Molnár, 2010; Szabo, 2018; Puimège & Peters, 2019), although the benefits of explicit instruction may be limited to raising awareness rather than measurable improvements in recognition performance. Overall, this thesis introduces a novel English-Malay-Chinese cognate list to systematically identify lexical similarities across these language pairs based on both form and meaning. Serving as a foundation for subsequent empirical investigations, it offers a systematic basis for examining cross-linguistic similarities in bi-/multilingual contexts, enabling empirical exploration of how learners draw on their linguistic resources when processing and learning L2 vocabulary. This thesis presents original empirical evidence of the cognate facilitation effect in an understudied population, demonstrating the consistent facilitative role of cross-linguistic similarity in both translation and recognition tasks. It further shows that cognateness, word frequency, and language proficiency significantly predict vocabulary performance across tasks. These findings extend existing empirical research while also offering theoretical insights into models of bilingual lexical processing. The cognate list serves both as a research tool for further investigation into cross-linguistic transfer in the bi- and multilingual contexts, and as a pedagogic resource. From the pedagogical perspective, the findings highlight the potential of leveraging cognates and the value of integrating cognate awareness training to support vocabulary development and inform the design of language assessments, curricula, and instructional materials.
Koh Ying Xian (Sat,) studied this question.