This study theoretically explains how Japanese learners of English (JLEs) develop their knowledge of (short-distance) subject wh-questions in English. Although first language (L1) acquisition research on wh-questions in English shows that the subject wh-question is easier to acquire than the object wh-question (Stromswold, 1988; Tyack & Ingram, 1977), some second language (L2) acquisition studies suggest that it is more difficult for JLEs to acquire subject wh-questions than any other wh-questions (Shirahata & S. Ogawa, 2017; Shirahata et al., 2017). This study adopts the following two theoretical perspectives: i) a syntactic approach: the Cartographic approach (Rizzi, 1997), and ii) a semantic approach: the animate vs. inanimate contrast in the subject noun, and attempts to explain why subject wh-questions (subject who- and what-questions) are difficult for JLEs.

Participants were 45 first year university students in Japan, and they were divided into three proficiency groups (Low, Intermediate and Advanced levels). The test sentences included two types of subject wh-questions; Type 1: animate subject who-questions (e.g., Who bought this bag?); Type 2: inanimate subject what-questions (e.g., What changed Mary so much?). The data were collected through a multiple-choice task. The results indicate that JLEs at the initial stage are syntactically and semantically influenced by L1 Japanese properties. Then, at the intermediate stage, they overuse DO (YOU), and this may be caused by the use of a Focus head as a single probe for all wh-questions after receiving a certain amount of do-support from L2 input. With the progress of JLEs' acquisition of subject wh-questions, they gradually come to realize that Focus and Force Phrases are used in disjunctive environments, resulting in the correct use of subject wh-questions. Thus, we propose the JLEs' developmental stages for the acquisition of subject wh-questions.

This paper presents part of the results of a learner corpus study of the acquisition of English as a second language (L2) based on two cross-sectional corpora constructed independently in Japan and Australia, namely Japanese Learner Corpus of English Narratives (JaLCEN) and WSU-Xi'an Jiaotong ESL Corpus. Different data elicitation tasks were used for these corpora, that is, oral and written narratives for JaLCEN and various spoken and written tasks for WSU-Xi'an Jiaotong ESL Corpus. The present study examines the spoken narratives by 88 Japanese native (L1) speakers from JaLCEN and the picture description spoken data by 51 learners in Australia with Japanese L1 and Chinese L1 backgrounds from WSU-Xi'an Jiaotong ESL Corpus. The study investigates acquisition of English grammar, focusing on the lexical mapping, based on the developmental stages defined by the Lexical Mapping Hypothesis (LMH) within Processability Theory (PT; Pienemann, 1998; Bettoni & Di Biase, 2015a). Analyses show that learners from both corpora exhibited an implicational relationship following the PT developmental stages when the emergence criterion was used and that different data-collection tasks did not affect the developmental sequence. This study contributes to the further understanding of systematic aspects of L2 acquisition. It also suggests that different learner corpora may be used to investigate the development of L2 grammar.

This study examined whether phonological neutralization in L1 Japanese, where the vowel sequences /ei/ and /ee/ are claimed to be incompletely neutralized to [eː] while /ou/ and /oo/ to be more completely neutralized to [oː], would transfer to the production of L2 English diphthongal vowels /eɪ/ and /oʊ/. Data were obtained from 100 Japanese learners of English in the J-AESOP corpus, which includes their oral proficiency judgments by trained phoneticians. Acoustic and statistical analysis revealed that low-proficient Japanese learners of English produced both /eɪ/ and /oʊ/ with significantly smaller formant changes than native English speakers. More proficient learners produced /eɪ/ with near-native formant changes, but their production of /oʊ/ was still significantly different from native production. Based on the results, it is argued that: (1) Japanese speakers tend to produce English /eɪ/ and /oʊ/ as monophthongal [eː] and [oː] due to the transfer of phonological neutralization, and (2) the acquisition of diphthongal /oʊ/ is more difficult than that of /eɪ/ because the transferred /ou/-/oo/ neutralization is more complete than the /ei/-/ee/ neutralization. The nearly complete neutralization of /ou/ and /oo/ may also explain why Japanese speakers often confuse English /oʊ/ and /ɔː/ (e.g., boat vs. bought).