VOCABULARY ACQUISITION THROUGH CAPTIONED TV SERIES: ARE THERE ANY APTITUDE AND PROFICIENCY EFFECTS? Mª del Mar Suárez Ferran Gesa Imma Miralpeix Thanks to the Spanish Ministry (Funded project FFI2013-47616-P) Pre-doctoral research grant to second author (BES-2014-068089) GRAL Research Group 2 Theoretical Background Dual Coding Theory (Paivio, 1986, 2007) • Verbal and non-verbal systems • Independent functioning but interaction • Activation of one system stimulates the other • Greater depth of processing and better recall • Cognitive Load Theory (Chandler & Sweller, 1991; Sweller, 1994) • Brain’s limited cognitive capacity, should not be overloaded • Multimodality may increase cognitive load (CL) • Subtitles as a tool to reduce CL in language acquisition settings Cognitive Theory of Multimedia Learning (Mayer, 2002, 2009) “Students learn more deeply from a multimedia explanation than from a verbal explanation” (2002: 62) 3 Theoretical Background Bimodal input (text and sound) • Better learning Bird & Williams, 2002; Grañena, Muñoz & Tragant, 2015 Multimodal input (text and video) • Beneficial for SLA Price, 1983; Baltova, 1999; Markham et al., 2001; Danan, 2004 • Listening comprehension and vocabulary acquisition Vanderplank, 2010, 2016; Nagira, 2011; Rodgers, 2013; Montero Perez et al., 2013, 2014 Learners approach the task according to their abilities (Dörnyei, 2005) Theoretical Background Subtitled TV series • Simultaneous presentation of L1/L2 text + L2 sound + video • Verbal and non-verbal information • Real language input • Fun activity, range of multimedia materials available Theoretical Background L1 subtitles (standard subtitling) • Recommended for low levels Danan, 2004 • Improve listening comprehension Plass & Jones, 2005 • Foster automatic reading Peters et al., 2016 L2 subtitles (bimodal subtitling or captioning) • Positive effects Vanderplank, 2010 • Associate aural and written forms Borrás & Lafayette, 1994 • Develop segmentation abilities Charles & Trenkic, 2015 Theoretical Background • University learners Sydorenko, 2010; Etemadi, 2012 • One-off studies Yuksel & Tanriverdi, 2009 Few exceptions: Rodgers, 2013; Frumuselu, 2015 • Benefits not exclusive to advanced adult learners Rice et al., 1990; Koolstra & Beentjes, 1999 Theoretical Background • University learners Sydorenko, 2010; Etemadi, 2012 • One-off studies Yuksel & Tanriverdi, 2009 Few exceptions: Rodgers, 2013; Frumuselu, 2015 • Benefits not exclusive to advanced adult learners Rice et al., 1990; Koolstra & Beentjes, 1999 Very scarce research on: - Beginner and intermediate EFL learners - Sustained exposure to multimodal input - Classroom-based research with TV series Theoretical Background Best procedures for class use? • Good selection of videos and captions • Instructional support Higher benefits than simply viewing videos in class Vanderplank, 2010 Webb, 2015 9 Theoretical Background Multimodal Input + Vocabulary & Aptitude • Scarce research into sustained exposure to multimodal input + TV series class use • (Rather) scarce research into vocabulary learning and aptitude • Virtually no research into vocabulary learning through subtitles and language aptitude 10 Theoretical background Aptitude & Vocabulary • Aptitude is multicomponential (MLAT, LLAMA). • Little research on how each subtest (i.e. aptitude component tapped by the test) influences language learning rate. • Regarding vocabulary (lexical variety), using MLAT-EC/ES: inconsistent results (Rosa & Muñoz, 2013, Muñoz, 2014; Suárez, 2014) • Regarding vocabulary (lexis, collocations), using LLAMA: • • • Greater gains for higher aptitude (LLAMA B – vocab learning) in a lexical test of formulaic sequences (Serrano & Llanes, 2012) Positive significant correlations in highly advanced adult L2 learners (Grañena & Long, 2013) Negative correlations: word-monitoring task tapping automatic use of L2 knowledge (Grañena, 2012 – except LLAMA D – sound recognition) 11 Theoretical Background: LLAMA • According to Grañena (2013), LLAMA measures two kinds of language learning aptitude: 1. 2. Explicit learning aptitude (B, E, F): rote learning Implicit learning aptitude (D): implicit induction, memorization B: Vocabulary learning (word + image) D: Phonetic memory (no subtitles) E: Sound-symbol correspondence (subtitles in L2) 12 Research Questions 1. Does sustained exposure to subtitled TV series lead to vocabulary learning? 2. Does aptitude have an effect on vocabulary learning from subtitled TV series? 3. Do proficiency level and vocabulary size have an effect on vocabulary learning from subtitled TV series? 13 Methodology • Participants • 62 freshman students of Media Studies • N=39 allocated to the intervention group • N=23 allocated to the control group • 18-22 years old • Proficiency A2 to C1 (OPT) • Catalan / Spanish bilinguals 14 Methodology  Instruments: • Listening / grammar part of the Oxford Placement Test (Allan, 2004) • X_Lex / Y_Lex (Meara & Miralpeix, 2006) 15 Methodology  OPT – Listening and grammar (Allan, 2004)  X_Lex / Y_Lex (Meara & Miralpeix, 2006) 16 Methodology  Instruments: • Listening / grammar part of the Oxford Placement Test (Allan, 2004) • X_Lex / Y_Lex (Meara & Miralpeix, 2006) • LLAMA aptitude test (Meara, 2005) 17 LLAMA B: Vocabulary learning E: Sound-symbol correspondence D: Phonetic memory F: Grammatical inference 18 Methodology  Instruments: • Listening / grammar part of the Oxford Placement Test (Allan, 2004) • X_Lex / Y_Lex (Meara & Miralpeix, 2006) • LLAMA aptitude test (Meara, 2005) • I Love Lucy TV series: 8 episodes of 22 mins approx. = 3 hours of multimodal input • English audio + English subtitles (intervention) • 5 Target Words (TWs) and 3 Target Expressions (TEs) per episode • Total of 40 TWs and 24 TEs 19 Methodology INTERVENTION GROUP (N=39) CONTROL GROUP (N=23) 1. PRE-TEST 1. PRE-TEST (40 TWs + 24 TEs, form and meaning recall) (40 TWs + 24 TEs, form and meaning recall) 2. 8 VIEWING SESSIONS 2. 8 VIEWING SESSIONS 2.1. PRE-TASK 2.1. PRE-TASK 2.2. EPISODE (x8) 2.2. EPISODE (x8) 2.3. VOCABULARY POST-TASK 2.3. VOCABULARY POST-TASK (5 TWs and 3 TEs, form recall and meaning recognition) (5 TWs and 3 TEs, form recall and meaning recognition) 3. POST-TEST 3. POST-TEST (40 TWs + 24 TEs, form and meaning recall) (40 TWs + 24 TEs, form and meaning recall) 20 Methodology PRE- and POST-TEST 21 Methodology PRE-TASK 22 Methodology VOCABULARY POST-TASK 23 Results RQ1 • RQ1. Does sustained exposure to subtitled TV series lead to vocabulary learning? Forms of words in English (TWs L2) Pre-test Post-test Meaning of words in Catalan / Spanish (TWs L1) 24 Results RQ1 • RQ1. Does sustained exposure to subtitled TV series lead to vocabulary learning? Expressions in English (TEs L2) Pre-test Post-test Expressions in Catalan / Spanish (TEs L1) 25 Results RQ1: Descriptive statistics Pre-test Form words in L2 Post-test Meaning words in L1 Form express. in L2 Meaning express. in L1 Form words in L2 Interv. M 10.10 3.67 8.21 3.49 19.36 N 39 SD 5.365 3.444 5.449 3.493 Control M 7.26 1.83 N 23 SD 5.602 All M N 62 SD Meaning words in L1 Form express. in L2 Meaning express in L1 11.21 12.46 7.74 7.805 6.092 6.043 4.962 7.61 2.52 17.57 8.91 11.43 7.26 2.534 5.383 2.952 6.591 5.062 5.367 4.693 9.05 2.98 7.98 3.13 18.69 10.35 12.08 7.56 5.582 3.242 5.388 3.312 5.579 5.778 4.830 7.374 Pre-test  Intervention > Control in Form and Meaning of Words Intervention = Control in Form and Meaning of Expressions 26 Results RQ1: Post-test Group Intervention Control Form words in L2 Meaning words in L1 Form expressions in L2 Meaning expressions in L1 .000 .000 .000 .000 92% huge 205% huge 52% very large 122% huge .000 .000 .000 .000 142% huge 387% huge 50% very large 188% huge Mann-Whitney U Test Control vs. Intervention - No significant differences Form words in L2 Meaning words in L1 Meaning Form expressions expressions in L1 in L2 Post-test .246 .150 .377 .661 Gains .545 .468 .558 .572 27 Results RQ1 Increase not significant in size for the Intervention group. Words in L2 23% 77% Words in L1 TWs learned TWs to be learned Expressions in L2 18% 82% 19% 81% TWs to be learned Expressions in L1 TEs learned TEs to be learned TWs learned 19% 81% TEs learned TEs to be learned 28 Results RQ1 Increase not significant in size for the Control group. Words in L1 Words in L2 26% 74% TWs learned TWs to be learned Expressions in L2 18% 82% 18% 82% TWs to be learned Expressions in L1 TEs learned TEs to be learned TWs learned 16% 84% TEs learned TEs to be learned 29 Discussion RQ1 Does sustained exposure to subtitled TV series lead to vocabulary learning? • Yes, but so does exposure to TWs & TEs through the pre- and post-tasks only, with no multimodal exposure to them. • Other learning mechanisms come into play: learning strategies, memorization, note-taking, focusing on TWs and TEs only. • Deliberate / Intentional learning (Laufer, 2005, 2006; Nation, 2001; Schmitt, 2008; Webb & Kagimoto, 2011; Peters 2012) • Same behavior in the long run? • There was potential for much more learning in both cases. 30 Results RQ2 Intervention LLAMA B LLAMA D LLAMA E LLAMA F LLAMA TOTAL Gains TWs L2 .095 .133 .177 .255 .191 Gains TWs L1 .344* .126 .211 .100 .255 Gains TEs L2 .018 .023 -.083 -.056 .005 Gains TEs L1 .054 .201 .017 .020 .091 .016 *p 0.05 level – 2-tailed **p 0.01 level – 2-tailed 31 Results RQ2 Control LLAMA B Gains TWs L2 .295 Gains TWs L1 .231 Gains TEs L2 .289 Gains TEs L1 .345 LLAMA D -.047 LLAMA E .423* LLAMA F LLAMA TOTAL -.154 .251 .018 .392* .022 .079 .277 .032 .210 .295 .014 .431* .020 .207 .294 .208 .509** .007 *p 0.05 level – 2-tailed **p 0.01 level – 2-tailed 32 Results RQ2 Intervention • High (N=21) > Low (N=18) aptitude Only in LLAMA D (phonetic memory) p.050 for Meaning of TWs (Spearman correlation) Control • High (N=14) > Low (N=9) aptitude Only in LLAMA Total (B+D+E+F) p.004 for Meaning of TEs (Spearman correlation) 33 Discussion RQ2 Does aptitude have an effect on vocabulary learning from subtitled TV series? • LLAMA B (word + image) does have an influence on the learning of meaning of words though only in the subtitles condition. • Aptitude does not seem to have an effect on the supposed benefits of being exposed to subtitles in the intervention group. • Different scenario for the control group, where aptitude (LLAMA total) affects learning of TWs’ meaning and TEs’ form and meaning. Explicit learning aptitude 34 Results RQ3: Vocabulary Size & Proficiency Intervention Vocab. size Gains TWs L2 OPT Listening OPT Grammar OPT Total .278* .461** .473** .510** .045 .002 .001 .000 .309* .331* .392* .421** .030 .020 .007 .004 Gains TEs L2 .123 .175 .120 .158 Gains TEs L1 .361* .337* .598** .560* .018 .013 .000 .000 Gains TWs L1 *p 0.05 level – 2-tailed **p 0.01 level – 2-tailed 35 Results RQ3: Vocabulary Size & Proficiency Control Vocab. size OPT Listening OPT Grammar OPT Total Gains TWs L2 .206 .058 .257 .110 Gains TWs L1 .365* .493** .540** .509** .043 .008 .004 .007 Gains TEs L2 .121 .116 .228 .138 Gains TEs L1 .522* .423* .635* .622** .005 .022 .001 .001 *p 0.05 level – 2-tailed **p 0.01 level – 2-tailed 36 Results RQ3 Intervention High (N=21) vs. Low (N=18) Proficiency + Vocabulary size OPT Listening Vocab. size OPT Grammar OPT Total Gains TWs L2 .017 .021 .007 .003 Gains TWs L1 .013 .043 - .016 Gains TEs L2 - - - - Gains TEs L1 .009 .000 .002 .037 Control • High (N=13) > Low (N=10) Proficiency + Vocab. size In Meaning of TEs for proficiency (OPT grammar p=.009; OPT total p=.009) and vocabulary size (p=.044) 37 Discussion RQ3 Do proficiency level and vocabulary size have an effect on vocabulary learning from subtitled TV series? • In the intervention condition, they clearly play a role in learning the form of new words and meaning of both new words and expressions, as opposed to aptitude. Higher proficiency relevant to learning form and meaning of TWs and meaning of TEs. • In the control group, proficiency is only relevant to learning the meaning of new words and expressions in one’s L1, not to learning the form of new words and expressions in L2. • Therefore, extra exposure (and proficiency) relevant to learning of TWs (meaning + form) and TEs (meaning), but not to TEs (form). Number of occurrences? Cognitive load for multiword expressions? • [Intentional learning + learning strategies + proficiency / voc. size] > cognitive aptitude(s) or extra exposure, as shown in the results for RQ1 & RQ2. 38 Conclusion • Intentional learning • Learning strategies • Cognitive aptitude(s) • Proficiency • Extra exposure • Vocabulary size 39 Limitations & current research • No comparison subtitling / non-subtitling conditions • Only one term • Training effects towards session 3 of the intervention • Lack of motivation in the control group • In-depth study on vocabulary learning: § Word Features – Frequency, saliency, cognateness, PoS § Retention effects – Delayed post-test • Other language skills: § Content comprehension § Speech segmentation § Spelling 40 THANK YOU! OBRIGADOS! 41 Results RQ2 All participants LLAMA B Gains TWs L2 .156 Gains TWs L1 .290* Gains TEs L2 .140 Gains TEs L1 .146 LLAMA D .067 LLAMA E .274* LLAMA F LLAMA TOTAL .135 .199 .063 .274* .031 .105 .224 .022 .031 .097 .031 -.050 .121 .020 .176 .125 .082 .217 *p 0.05 level – 2-tailed **p 0.01 level – 2-tailed