Capturing and representing asymmetries in Japanese EFL learners

Capturing and representing asymmetries in Japanese EFL learners’ mental lexicon SUGINO Naoki (Ritsumeikan University, Japan) Simon FRASER (Hiroshima University, Japan) AOTANI Noriko (Tokai Gakuen University, Japan) SHOJIMA Kojiro (The National Center for University Entrance Examinations, Japan), and KOGA Yuya (Graduate School of Education, Waseda University, Japan) Paper presented at the Vocab@Vic 2013 Conference RHLT 2, Rutherford House, Victoria University of Wellington, New Zealand 11: 45 -12: 05, Wednesday 18 December, 2013 This study is supported by Grant-in-Aid for Scientific Research (B) (25284110) from Japan Society for the Promotion of Science. 1

Background • Background: A part of a research project • To provide individualised (visual) feedback • on how learners have learned the words • focusing on “co-occurrence” 2

Background • Background: A part of a research project Aotani, Kameyama, Sugino, & Amaya (2009) Aotani, Kameyama, & Sugino (2010 a) • To provide individualised (visual) feedback • on how learners have learned the words acceptance of metaphorically expanded uses • focusing on “co-occurrence” (polysemous) Adjective – Noun run – Noun 3

Background • Background: A part of a research project Aotani, Kameyama, & Sugino (2010 b); Aotani, Kameyama, Sugino, & Koga (2011); Aotani & Sugino (2012) • To provide individualised (visual) feedback learner traits (strategy, motivation, self-esteem…) SOM (self-organising map) • on how learners have learned the words acceptance of metaphorically expanded uses • focusing on “co-occurrence” (polysemous) Adjective – Noun 4

Background • semantic relatedness through paradigmatic associations • + syntagmatic associations 5

Asymmetry / Directionality • Association measures • Pecina (2009): 82 lexical association measures • Asymmetry • Evert (2007/2009) • Schmitt (2010), Durrant (2008) • Gries (2013), Ellis (2006) • Corpus to psycholinguistic studies • Durrant & Doherty (2010) • Michelbacher, Evert, & Schütze (2011) • Yamashita & Jiang (2010); Wolter & Gyllstad (2011, 2013) 6

Kusanagi (2013) • focuses phrases on the asymmetry within the “N and N” • association measure based on the ratio of transitional probabilities • reaction times and acceptability judgment scores of the Japanese EFL learners • general insensitiveness to asymmetry observed among the learners 7

Background • Background: A part of a research project (a)symmetry within syntagmatic relations • To provide individualised (visual) feedback AMISESCAL • on how learners have learned the words acceptability judgment on the attributive/predicative uses • focusing on “co-occurrence” Adjective–Noun / Noun–copula–Adjective • cf. Karasawa (2003) 8

Asymmetry in this study • A pair of an adjective and a noun is symmetric if • Adj. →N and N→Adj. are both (un)acceptable, • to the same degree. • A pair of an adjective and a noun is asymmetric when • either Adj. →N or N→Adj. is acceptable. 9

The present study • Three types of adjectives main, only, own • predictive-only: afraid, alive, alone • both uses: beautiful, happy, small • attributive-only: • Three types of nouns teacher, friend • [inanimate; concrete]: car, house • [inanimate; abstract]: idea, love • [human]: • All words are with the familiarity scores of above 5. 50 (of 7. 00) (Yokokawa, 2006) • 108 target sentences (between 7 to 9 words each) 10

The present study main love of rock music. • At the meeting, George’s idea seemed to be only. • Finally, Albert learned that his idea was very own. • The idea was so afraid that I couldn’t tell. • The car looked alive on the streets of Tokyo. • An alone car was parked on the street. • I found the idea so beautiful that I smiled. • They lived in a happy house by the sea. • A small car would be easier to drive. • They shared the 11

The present study • 51 Japanese EFL learners at the university level • those who did not complete the test were excluded → 45 learners’ data were used • Five native speakers of English • Instructed to give acceptability judgement on the 5 - point Likert scale to each of the 108 sentences; to be completed in 20 minutes. • The acceptability scores (1=totally unacceptable to 5=totally acceptable) were then converted into the proximity scores (1→ 5, 2→ 4, 3→ 3, 4→ 2, 5→ 1). 12

AMISESCAL (Shojima, 2011, 2012) • http: //www. rd. dnc. ac. jp/~shojima/ams/index. htm 13

AMISESCAL (Shojima, 2011, 2012) • Asymmetric von Mises Scaling • A kind of asymmetric multidimensional scaling, based on the von Mises distribution in directional statistics • Normal • μ: distribution in directional statistics mean direction κ: concentration 14

AMISESCAL (Shojima, 2011, 2012) • Asymmetric von Mises Scaling 15

AMISESCAL (Shojima, 2011, 2012) 16

(A)symmetry in AMISESCAL • Symmetry is represented by the proximity on a two dimensional map. • both acceptable: two words will be placed close to each other. • both unacceptable: the two words will be placed far from each other. • Asymmetry is represented by • its v. Md’s degree (µ), concentration(κ), and size (ω). 17

Data structure • 1 -mode 2 -way, asymmetric double bipartite proximity data • Nouns are fixed at the radius of 0. 400 18

Results: Native speakers of English 19

Results: Japanese EFL learners 20

Results: JEL 1 & JEL 2 21

Conclusion • By analysing acceptability judgement scores using AMISESCAL, (a)symmetry within syntagmatic associations consisting of an adjective and a noun is visually represented in terms of their positions and the v. Mds. • NSEs’ representation is in accordance with the normative grammar; JELs representation is characterised by the lack of v. Mds and the similar distance from the nouns. • Seen individually, • No apparent patterns/similarities, but (a)symmetries are present. • The presence of ‘spiky’ v. Mds, implying the associations are represented as fixed chunks. 22

Issues for further study • Visualising AMISESCAL • ∆P asymmetries observed in corpora by (Gries, 2013; Ellis, 2006) • Establishing protocols for interpretation • accumulation mappings and classification of the 23

References Aotani, N. , Kameyama, T. , Sugino, N. , & Amaya, Y. (2009). A study of the effectiveness of the CALL program, ‘Adjective Sommelier’, as a learning tool to improve learners’ analytical approach to the polysemous senses of TL adjectives. Proceedings of the 14 th Conference of Pan-Pacific Association of Applied Linguistics, 415 -418. Aotani, N. , Kameyama, T. , & Sugino, N. (2010 a). An analysis on how Japanese learners of English perceive polysemous senses of words that are peculiar to English: Based on the acceptability patterns of the senses of ‘run’. Proceedings of the 15 th Conference of Pan-Pacific Association of Applied Linguistics, 267 -272. Aotani, N. , Kameyama, T. , & Sugino, N. (2010 b). On the significance of improving learners’ metaphorical thinking abilities for language acquisition. Proceedings of the 4 th CLS International Conference , 45 -54. Aotani, N. , Sugino, N. , Kameyama, T. , & Koga, Y. (2011). An analysis on how derivation inferences instruction affects acquisition of polysemous senses of words of Japanese learners of English. Proceedings of the 16 th Conference of Pan-Pacific Association of Applied Linguistics , 385 -388. Aotani, N. , & Sugino, N. (2012). Profiles of Japanese EFL learners and their acceptance of transfer in word meaning expansion. Proceedings of the 5 th CLS International Conference. Durrant, P. (2008). High frequency collocations and second language learning (Ph. D dissertation. University of Nottingham). Retrieved from http: //etheses. nottingham. ac. uk/622/1/final_thesis. pdf Durrant, P. & Doherty, A. (2010). Are high-frequency collocations psychologically real? Investigating thesis of collocational priming. Corpus Linguistics and Linguistic Theory, 6, 125– 155. doi: 10. 1515/CLLT. 2010. 006 Ellis, N. C. (2006). Language acquisition as rational contingency learning. Applied Linguistics, 27, 1– 24. doi: 10. 1093/applin/ami 038 Evert, S. (2007). Extended manuscript for Evert, S. (2009). Corpora and collocations. In A. Lüdeling and M. Kytö (Eds. ), Corpus Linguistics. An International Handbook, article 58. Mouton de Gruyter, Berlin. Retrieved from http: //cogsci. uniosnabrueck. de/~severt/PUB/Evert 2007 HSK_extended_manuscript. pdf Gries, St. Th. (2013). 50 -something years of work on collocations: What is or should be next … In S. Hoffmann, B. Fischer- Starcke, & A. Sand (Eds. ), Current issues in phraseology. International Journal of Corpus Linguistics [Special issue], 18, 137– 166. doi: 10. 1075/ijcl. 18. 1. 09 gri Karasawa, S. (2003). Patterns of elaboration and inter-language development: An exploratory corpus analysis of college student essays. In D. Archer, P. Rayson, A. Wilson, & T. Mc. Enery (Eds. ), Proceedings of the Corpus Linguistics 2003 Conference [Special issue of UCREL Technical Paper], 16, 394– 401. Retrieved from http: //ucrel. lancs. au. uk/publications/CL 2003/papers/karasawa. pdf Kusanagi, K. (2013). Second language learners’ processing of symmetric formulaic sequences: A preliminary study focusing on English N and N phrases. LET Journal of Central Japan, 24, 15– 24. Michelbacher, L. , Evert, S. & Schütze, H. (2011). Asymmetry in corpus-derived and human word associations. Corpus Linguistics and Linguistic Theory , 7, 245– 276. doi: 10. 1515/CLLT. 2011. 012. Pecina, P. (2009). Lexical Association Measures: Collocation Extraction. Institute of Formal and Applied Linguistics. Schmitt, N. (2010). Researching vocabulary: A vocabulary research manual. Basingstoke, UK: Palgrave Macmillan. Shojima, K. (2011). Local dependence model in latent rank theory. Japanese Journal of Applied Statistics, 40, 141– 156. Shojima, K. (2012). On the stress function of Asymmetric von Mises Scaling. Proceedings of the 4 th Japanese-German Symposium on Classification, 18. Wolter, B. & Gyllstad, H. (2011). Collocational links in the L 2 mental lexicon and the influence of L 1 intralexical knowledge. Applied Linguistics, 32, 430– 449. doi: 10. 1093/applin/amr 011 Wolter, B. & Gyllstad, H. (2013). Frequency of input and L 2 collocational processing: A comparison of congruent and incongruent collocations. Studies in Second Language Acquisition, doi: 10. 1017/S 0272263113000107 Yamashita, J. & Jiang, N. (2010). L 1 influence on the acquisition of L 2 collocations: Japanese ESL users and EFL learners acquiring English collocations. TESOL Quarterly, 44, 647– 668. doi: 10. 5054/tq. 2010. 235998 Yokokawa, H. (Ed. ). (2006). Nihon-jin eigo gakushusha no eitango shinmitsudo (Moji hen): Kyoiku, kenkyu no tameno dai ni gengo detabesu [Database of familiarity of English words for Japanese EFL learners: Orthography]. Tokyo: Kuroshio. 24

gwisno@is. ritsumei. ac. jp 25

Asymmetry in collocation studies • Asymmetry/directionality sequences in formulaic • “whether word 1 is more predictive of word 2 or the other way round” • “bidirectional/symmetric association measures conflate two probabilities that are in fact very different: p(word 1|word 2) is not the same as p(word 2|word 1), just compare p (of|in spite) to p(in spite|of). ” (Gries, 2013, p. 141) 26

Asymmetry in collocation studies • ∆P= p (outcome | cue=present) – p (outcome | cue=absent) • “∆P is the probability of the outcome given the cue (P(O|C)) minus the probability of the outcome in the absence of the cue (P(O|-C)). When these are the same, when the outcome is just as likely when the cue is present as when it is not, there is no covariation between the two events and ∆P= 0. ∆P approaches 1. 0 as the presence of the cue increases the likelihood of the outcome and approaches -1. 0 as the cue decreases the chance of the outcome - a negative association. ” (Ellis, 2006, p. 11; cited in Gries, 2013, p. 143) • ∆P 2|1= p (word 2 | word 1=present) – p (word 2 | word 1=absent) • ∆P 1|2= p (word 1 | word 2=present) – p (word 1 | word 2=absent) 27

AMISESCAL (Shojima, 2011, 2012) 28

Individual differences among NSEs 29

Individual differences among NSEs: NSE 1 30

Individual differences among NSEs: NSE 2 31

Individual differences among NSEs: NSE 3 32

Individual differences among NSEs: NSE 4 33

Individual differences among NSEs: NSE 5 34

Results: JEL 3 & JEL 4 35

Results: JEL 5 36

Aim of the Study • To demonstrate how learners’ understanding of different uses of English adjectives can be visualised by employing AMISESCAL 37