Acta Linguistica Asiatica, 10(2), 2020. ISSN: 2232-3317, http://revije.ff.uni-lj.si/ala/ DOI: 10.4312/ala.10.2.105-125 NORMATIVE FORMS AND SYNTHETIC STRUCTURE OF JAPANESE IN THE INCUBATION PERIOD OF L2: SUBJECT TO SENTENCE-FINAL FORMS IN LONGITUDINAL DISCOURSE DATA OF KOREAN RETURNEE SISTERS’ JAPANESE HWANG Young Hee Hanyang Cyber University, Korea hwang@hycu.ac.kr Abstract In this paper, I examine the change mechanism of Japanese sentence-final forms (SFF) maintained by two Korean returnee sisters for over 10 years after the cessation of L2 contact, and focus on the negative formal style of verb sentences and its deviation from the actual use of norms (analysis form) and non-norms (synthetic form). Findings are based on a comparison of two Korean sisters’ Japanese with that of thirteen Korean adults’ colonial Japanese maintained for over 60 years, which is also in the incubation phase. In the sisters’ Japanese sentence-final forms that were incubating as their L2, they rarely used the non-norms, while the norms were stably maintained, and the retention of the synthetic structure of their returnee Japanese correlated with the duration of the language acquisition period and with the elapsed time of contact cessation. That is, the sisters used more of the norms in the Korean colonial Japanese SFF than they did in their Japanese; I attribute this to the sisters’ 10-year incubation period. Specifically, the Korean returnee sisters’ speech included interventions of the explanatory [N] in the past affirmation of -ta desu, heavy use of non-norms in adjective and noun sentences, and connecting sentence-final particles to further grammatical structures. However, there were fixed and conventional Norms in the Korean returnee sisters’ Japanese, and once these are acquired, masu forms are retained for long periods in mirror image, especially on elder A. To summarize, in terms of the format of the returnee Japanese SFF, the two Korean returnee sisters were slower to shift from norms (masu) to non-norms (desu) than were adult Korean speakers of colonial Japanese. The same shift is observed with synthetic structure even after cessation of the language contact. Keywords: normative forms; incubation period; sentence-final forms of Japanese; synthetic structure; analysis; Korean returnee sisters Povzetek V tej študiji preučujem mehanizem sprememb v končniških oblikah povedi v japonščini, ki sta jih dve korejski sestri povratnici uporabljali v obdobju 10 let po prenehanju neposrednega stika z L2. Predvsem se osredotočam na negativni formalni slog glagolskih stavkov ter odstopanj v dejanski uporabi norme (analitična oblika) in nenorme (sintetična oblika). Ugotovitve temeljijo 106 HWANG Young Hee na primerjavi japonščine omenjenih sester ter japonščine trinajstih odraslih Korejcev kolonialne japonščine, ki se je ohranila več kot 60 let. Obe vrsti japonščine sta ocenjeni kot jezika v fazi inkubacije. V oblikah korejskih sester so le redko opazne nenorme, medtem ko so bile norme stabilno vzdrževane. Njuno ohranjanje sintetičnih oblik je sorazmerno tako s trajanjem usvajanja japonskega jezika kot tudi s pretečenim časom od prekinitve stika z jezikom. Poleg tega je razvidno, da sta po povratku v domovino uporabljali več analitičnih oblik kot v času bivanja na Japonskem, kar pripisujem desetletnemu obdobju inkubacije. Njun govor je na primer vključeval razlagalni [N], ki se je pogosto vmešal v pozitivno preteklo obliko ta desu, prekomerno rabo sintetičnih oblik v pridevniških in samostalniških stavkih ter nadaljno navezavo na končniški povedne členke. Hkrati je bilo v njunem govoru opaziti precej ustaljenih in običajnih analitičnih oblik, ki so se, potem ko so bile usvojene, v zrcalnih slikah obdržale zelo dolgo. Tako se je -masu oblika še posebno dolgo obdržala pri starejši od sester. V splošnem lahko rečem, da se je japonski govor korejskih sester povratnic počasneje preusmeril iz norme (-masu) v nenormo (desu) v primerjavi z govorom korejskih govorcev kolonialne Japonske. Podoben premik je opazen tudi pri sintetični strukturi navkljub prenehanju jezikovnega stika. Ključne besede: normativne oblike; inkubacijska doba; končniške oblike povedi v japonščini; sintetična struktura; analiza; korejski sestri povratnici 1 Introduction The field of applied linguistics has had a recent focus on “L2 maintenance” or “L2 attrition”. In that focus, the mechanism of language change is either maintained, or else declines after L2 is acquired, and it is important to provide feedback and savings effect to the relearning of L2 learner (Tomiyama 2004). This includes cases of Japanese learners who learned in high school, expatriates who had returned to Korea after staying in Japan, and international students who wanted to maintain a certain level of Japanese. I refer to the language of such Japanese speakers as the “Japanese of the incubation period.” In addition to studies on L2 acquisition, L2 attrition researchers such as Weltens and Cohen (1989), Ono (1994), Shibuya (1995), Hansen (1999), Tomiyama (2000), Kiguchi (2004), Kim (2010), and Matsumoto (2013) have conducted studies on Japanese learners to examine subtractive changes over time. However, very few researchers have explored the changes in sentence-final forms (desu and masu style) that occur over time among learners of Japanese as L2. In this study, I investigate the latent Japanese that remains in two Korean children who returned to Korea after six years in Japan. By examining characteristics of sentence patterns in different stages of L2 acquisition, namely maintenance, attrition, and decline, I aim to disprove the hypothesis of “analysis” that Dorian (1977), Shibuya (1995), Kiguchi (2004), and Hwang (2013) have asserted to be normal in language retention. Normative forms and synthetic structure of Japanese in the incubation … 107 Furthermore, some researchers, such as Noda et al. (2001), have proposed the unstable interlanguage states of the sentence-final forms (SFF) (desu and masu style) with Japanese learners. However, because learners involved in earlier studies had ceased learning Japanese during their stays in Japan, including the current Japanese learner, their Japanese had been latent for more than 10 years. Further research is needed on the maintenance status of the Verb, i-adjective, Noun and na-adjective SFF to elaborate on results by Kiguchi (2009), Sanada (2009), and Hwang (2010). I aim to clarify the processes of maintaining and attrition, which relate to the processes of learning SFF and mirror relationships, in order to elucidate the holistic changes in learners’ Japanese. 2 Previous research 2.1 The L2 life cycle Among L2 researches on language changes after active learning stops that were conducted around 1980s, Weltens and Cohen (1989) on L2 attrition identified the early stage of experimental verification as “order of attrition, relations with literacy, age, and supplemental strategies” from a symmetric perspective on language acquisition. Education for returnee from overseas began to be recognized as an issue in Japan already in the late 1960s (Kubo 1994). In the 1990s, language problems of Japanese returnees attracted attention, with the main theme being the changes in young people’s English ability. For example, Ono (1994) pointed out that Japanese returnees’ English vocabulary skills had rapidly declined regardless of how long they had stayed abroad, their ages at return, or their English vocabulary skills immediately after their return. Tomiyama (2000) surveyed Japanese returnees (all eight years old) from the United States for four years after they had returned to Japan, and found that signs of decline appeared after six months though the speakers had maintained the capacity to accept and produce English after four years. Tomiyama (2008) investigated the differences in age and proficiency in L2 attrition between two siblings and suggested maturational factors. Meanwhile, Kim (2010) discussed attrition-related changes in fluency, vocabulary, and postpositional particles in Japanese as L2 according to the passage of time in Korean child returnees. The author also found out that fluency degraded over time. In the present study, I compare children’s L2 Japanese in the incubation period with that of Korean colonial adults who had had little contact with Japanese after liberation in 1945. Korean colonial Japanese has been maintained for over 60 years since its acquisition and can therefore be defined as an attrition language similar to Korean returnee Japanese for over 10 years after the cessation of L2 contact. There are 108 HWANG Young Hee very few longitudinal studies on the maintenance status of Koreans’ L2 Japanese after language contact. For example, “Japanese Learners of Chinese and Korean Native Language Longitudinal Utterance Corpus (C-JAS [http://c-jas.ninjal.ac.jp/index.py]),” which was released in Japan, includes a three-year longitudinal discourse data of six Japanese learners, in which authors focused mainly on language acquisition, and less on Japanese maintenance. Therefore, I consider findings of Korean colonial adults’ Japanese to be optimal as a contrast data. Researchers have recently investigated Japanese attrition with Japanese L2 learners who speak English (Hansen ed. 1999, Hayashi 1999, Tomiyama 2000), and with colonial Japanese speakers who remained in Taiwan and the South Sea Islands (Kan 2004, Shibuya 2001, Matsumoto 2013). Their works are cited examples of colonial Japanese in which L2 has been restructured without suffer decline for over 50 years. Kiguchi (2004) and Hwang (2009) have also examined colonial Japanese among Koreans who maintain Japanese as their L2. On the other hand, Kiguchi (2004) worked on the context of language attrition, and Hwang (2009) elucidated the relevance of variations between speakers and their language acquisition environment at the point of language maintenance. There has also been little research on individual circumstances after L2 active learning stops and L2 cessation because of the difficulties associated with continuing the longitudinal research. 2.2 General characteristics of L2 in the L2 life cycle Dorian (1981) examined the dying language of a Scottish Gaelic dialect and pointed out the following linguistic features (instances of Japanese usage being regressed are excerpts from Kiguchi 2004). (a) No possibility of selecting formal versus informal style (b) Replacing a synthetic structure with an analytic structure (e.g., mashi-ta [FML- Past] and ta-desu [Past-FML]) (c) Occurrence of analogical leveling (e.g., shabereru [POT] → shaberareru [POT]) In addition, Shibuya (1995) reported on the declining status of Japanese remaining in Palau in the South Sea Islands in terms of the following five features of speech structure. (1) simplified formal style usage, (2) simplified discourse management with reduced use of sentence-final particles, (3) analysis of using nai-desu [NEG-COP] rather than masen [FML-NEG] or using kitto ‘surely’ to indicate likelihood, (4) use of chunks such as sentence-final particles yone [Compound SFP] and negatives like masen and wakara-nai-ne ‘understand’ [NEG-SFP], Normative forms and synthetic structure of Japanese in the incubation … 109 (5) avoiding complex phrases such as passive and causative sentences. In the above examples, characteristics observed in the L2 in its incubation state are considered as “simplification, analysis, or transfer of the mother tongue”, and are the factors driving major language change. 2.3 Characteristics of Japanese sentence-final forms (SFF) appearing in the L2 life cycle Noda et al. (2001) found that interlanguage speakers construct formal sentences with past-tense verbs using nonstandard Non-norms series (past + formal) much more often than using standard Norms series (formal + past). Table 1 presents the two series of formal SFF. Table 1: The two series of Japanese SFF [FML] Sentence Past Non-past Norms Non-norms Norms Non-norms Verb Positive V-masita V-tadesu* V-masu V-desu* Negative V-masendesita V-nakattadesu V-masen V-naidesu i-adj. Positive i-desita* i-tadesu i-desu Negative i-arimasendesita i-nakattadesu i-arimasen i-naidesu Noun Positive N-desita N-dattadesu N-desu Negative N-dewaarimasendesita N-dewanakattadesu N-dewaarimasen N-dewanaidesu na-adj. Positive na-desita na-dattadesu na-desu Negative na-dewaarimasendesita na-dewanakattadesu na-dewaarimasen na-dewanaidesu * non-standard type This table is a modified version of the SFF in Noda et al. (2001), with additional non- past forms. Norms conjugation indicates synthetic forms as chunks or units (following synthetic structure by Dorian 1981), and Non-norms conjugation indicates analytic forms (following analytic structure by Dorian 1981). For example, if Japanese learners recognize and use expressions such as -tarayokatta and -naitodame as chunks when making conditional statements then they are applying the synthetic forms, whereas if these chunks are recognized and used separately, such as -tara+yokkatta or - nai+to+dame(ikenai), they are applying the analytic forms. Actually, in colonial Japanese, an analytic form is a case where a pause (,) is inserted into the +part such as Netara, yokkatta ‘I wish, I had gotten to sleep.’, Sinakereba, dameda ‘You have to, do it.’. On the other hand, in returnees’ Japanese there is no pause in the +part and it is 110 HWANG Young Hee spoken as a chunk, like Okitarayokkatta ‘I wish I woke up.’, Ittaradameda ‘Don't go’, which corresponds to the synthetic form. Kiguchi (2004), who described the colonial Japanese of Korean elders, observed that Non-norms such as ta-desu [Past-COP], nai-desu [NEG-COP], and Verb+desu [COP] tend to be productive, and also further detailed the declining tendency at the stage of language attrition. In addition, Hwang (2013) examined Chinese-Korean colonial Japanese and found universal simplification and analysis in the Japanese learners’ interlanguage acquisition process in case of non-norms. Besides, Hwang also presented various potential inter- language factors, such as the possibility of influence from Korean, the correlation with verbal internal factors, predominantly using the norms of speakers with high academic ability, and discrimination based on discourse functions in sentences. In this study, I will discuss similarities and differences in the usage of verb SFF with a particular focus on speaker attributes, age of contact cessation, types of verbs, collocation relationships with sentence-final particles, and insertions of [N]. In particular, in relation to the age of contact cessation, I aim to present how both the synthetic and the analytic predicate structures have been maintained in both the Japanese of Korean returnee children and colonial Japanese of Korean adults. 3 Survey overview I have collected natural discourses of two Korean-Japanese bilingual Korean sisters for over 4 years, 11 months in order to examine the mechanism of acquisition, maintenance, attrition, and regression of Japanese as L2 that was acquired approximately 15 years earlier. The discourse data I use in this study consist of language data that were collected roughly 10 years before and after Japanese contact cessation. Details of the two sisters Japanese L2 acquisition life cycle will be given shortly in Table 2 below. 3.1 Informants The informants I studied are two Korean returnee sisters: A, born in 1999, and B, born in 2001. They lived in Tokyo and Osaka for six years from 2001 to 2008, with two short returns to Korea, and received just over 4 years of local kindergarten and elementary education (1st and 2nd stage; acquisition period). Ten years after they had returned permanently to Korea in 2008, they were experiencing regression of their Japanese proficiency. As of July 2008 the sisters (elder sister A was in the 3rd grade and younger sister B was in the 2nd grade) were gradually deteriorating in their Japanese because they had Normative forms and synthetic structure of Japanese in the incubation … 111 no contact with the language. At that time they attended the “returned children’s class” established in Korea's elementary school until February 2009 (3rd stage; retention period). After that, until August 2011, they gradually became monolingual speakers; during that time, they had no contact with Japanese (4th stage; incubation and attrition period). A entered high school in March 2015 and B in March 2017, and at that point they both began studying Japanese again (5th stage; re-contact period). In short, A and B learned Japanese in the 1st and 2nd stages of the L2 acquisition cycle and often used to speak Japanese on a regular basis instead of Korean. In the 3rd stage, Japanese was maintained as their basic (default) language, but in the first half of the 4th stage, their basic language changed from Japanese to Korean. In the second half of the 4th stage, the sisters ceased using Japanese. As noted above, I classify the data on the two informants across the stages of the L2 acquisition life cycle: 1st and 2nd stage (acquisition period), 3rd stage (retention period), 4th stage (incubation and attrition period), and 5th stage (re-contact period). Table 2: Information of informants and Japanese contact Stage Residential area and period Education Information 1st stage (acquisition 1) ① Seoul (A, 0:0-2:0; B, 0:0-0:7)* ② Tokyo & Yokohama (A, 2:1-4:6; B, 0:8-3:1) ③ Daegu (A, 4:7-4:10; B, 3:2-3:5) ④ Osaka (A, 4:11-5:1; B, 3:6-3:8) ⑤ Daegu (A, 5:2-5:6; B, 3:9-4:1) ① Resident in Korea ② Resident in Japan ③ Kindergarten in Korea ④ Kindergarten in Japan ⑤ Child-care institutions in Korea 2nd stage (acquisition 2) ⑥ Osaka (A, 5:7-8:9; B, 4:2-7:4) ⑥ Kindergarten and Elementary school in Japan 3rd stage (retention) ⑦ Seoul (A, 8:10-9:4; B, 7:5-7:11) ⑦ Elementary school in Korea 4th stage (incubation and attrition) ⑧ Seoul (A, 9:5-15:3; B, 7:12-13:10) ⑧ Elementary and junior high school in Korea 5th stage (re-contact)** ⑨ Seoul (A, 15:4-present; B, 13:11-present) ⑨ High school in Korea * In column 2, (year: month) indicate the elapsed time since the births of informants A and B. ** Results are taken from the surveys conducted in November 2016, March 2017, and March 2020. 3.2 Discourse data information I began recording the linguistic data in September 2006 (A, 7:11; B, 5:6) and recorded informants’ conversations until August 2011 (A, 11:10; B, 10:5). This includes having the girls look at a picture book in Japanese (Kim 2010) four times, from just before they returned to Korea in July 2008 to December 2009. I recorded the discourse materials 112 HWANG Young Hee at the sisters’ home in stable scenes. I recorded conversations when the sisters or their family members were playing together or talking to each other before sleeping. At that time, the parents did not use Japanese, as they communicated mostly in Korean. Six months after the girls returned to Korea, after the January 2009 survey, most of their natural conversation occurred in Korean, and in the March 2009 survey that was conducted eight months after their return, I observed examples of misuse of Japanese and mixed usage of Korean and Japanese. Then, in the January 2010 survey, which was conducted 17 months after the return to their home country, they were no longer using Japanese. After the 5th period when Japanese was no longer used, a translation survey was conducted on the use of words. I expected some variants to be retained as latent knowledge, however, they were absent from the discourse data. To clarify the latent linguistic ability, an additional survey was conducted in November 2016, March 2017 and March 2020 when voluntary use of Japanese disappeared completely to obtain data on translations of Korean sentences. Previous researchers mainly gathered data for their studies through artificially manipulated speech, reading, and survey responses. However, the discourse in this study contains natural Japanese used in conversations involving only two sisters and their family members. The girls mostly spoke casually with each other, but role plays in formal style conversation have also been detected. The proportions of formal style were 12.4% for A and 15.5% for B (see Table 3). Moreover, the recorded data I have collected for five years are presented in Table 2 [④~⑧ ] account for 1,149 minutes (~19 hours) in total. Considering the entire duration of the investigation, this total is an equivalent to one-hour conversations recorded every three months. In addition, considering the recording time, the number of characters, and the number of times, the ratio of the volume of discourse from the 1st to the 5th stages is 4:4:3:3:4 with rounding to the nearest integer. This does not include the time I presented Kim’s (2010) materials and collected those data. 3.3 Contrast data information To clarify the characteristics of the SFF in the Japanese spoken by the two returnee- sisters to Korea, I contrast their data with the discourse data of the colonial Japanese spoken by Korean and Chinese adults that have undergone similar regressions. Although the two samples are very different in size, they are appropriate for comparison because the conversations with the Japanese native speakers and the aspects of discourse development are similar. I furthermore include two sets of colonial Japanese data for Korean elders, the first of which was recorded in Daegu and Seoul in South Korea in 2006. The informants were eight Korean speakers over 80 years old (3 male and 5 female) who had been screened Normative forms and synthetic structure of Japanese in the incubation … 113 for having little experience with relearning Japanese after liberation in 1945. Their 61 years since Japanese language contact cessation was significantly longer than the equivalent period for the Korean returnee sisters. The total recording time was 530 minutes at an average of 44 minutes per person. The second data set, discourse among five Korean elders (1 male and 4 female) and two Japanese native speakers, was recorded in 2010 in the Yanbian area (Yanji city and Tumen city) of northeastern China, an average of 42 minutes per person. These speakers were over age 80 and had had little contact with Japanese since 1945. For more information on the survey subjects, refer to Hwang (2013). Further, these thirteen elderly were about the same age as elementary school students, and were in contact with Japanese students and Japanese people around them. As such, in terms of the degree of Japanese contact, the duration of Japanese acquisition period, and Japanese learning method similar to current immersion programs, these thirteen elderly had a similar background of Japanese acquisition that was similar to that of the Korean returnee sisters. 4 Results and discussion 4.1 Informants’ actual use of sentence-final forms (SFF) Before proceeding to the main text, I examine the overall outline and characteristics of SFF through the frequency of use of formal and informal verb sentences in Japanese in incubation. Table 3 shows the informants’ rates of using the whole SFF, including both noun and adjective as well as verb sentences. Table 3: Informants’ usage of SFF SFF Returnee A Returnee B Koreans Chinese Koreans ➊ Total 2397 1789 1626 1112 ➋ Informal style 2100 1512 603 225 ➌ Formal style 297 277 1023 887 ➍ Percentage of formal style 12.4% 15.5% 62.9% 79.8% Note: ➍=➌/➊ As shown in Table 3, the two Korean returnee sisters used substantially less formal style Japanese in the incubation period than did the Korean and Chinese-Korean elders. This is attributed to the differences in the attributes of the conversation partners, i.e., Korean returnee children versus Japanese native adults. 114 HWANG Young Hee 4.2 Use of formal verb sentences Table 4 shows the number of appearances in the L2 life cycle by period, and the data is divided into positive, negative, informants (vertical axis) and tense, series, and period (horizontal axis). Furthermore, regarding L2 incubation, I focus on the time after acquisition, namely the 3rd or the retention period. Table 4: Usage of formal verb sentences Sentence Past Non-Past Total Norms (masi-ta) Non-norms (ta-desu) Norms (masu) Non-norms (ru-desu) I II III IV V I II III IV V I II III IV V I II III IV V Positive A 15 12 - - - 8 2 - - - 44 45 7 1 18 3 5 1 - - 160 27 - 10(4) - 89 26 8(5)* 1 B 2 2 - 112 - 4 - - - 14 33 2 - 13 4 2 - - - 188 4 112 4(2) - 47 15 6(3) - Sentence Past Non-Past Total Norms (masen-desi-ta) Non-norms (nakat-ta-desu) Norms (masen) Non-norms (nai-desu) I II III IV V I II III IV V I II III IV V I II III IV V Negative A - - - - 4 1 - - - 1 8 5 1 - 14 1 - - - - 35 - 4 1 1 13 15 1 - B 1 - - 1 2 - - - - 7 - 5 4 1 6 3 - 1 - 1 32 1 3 - 7 5 11 3(2) 2 * Examples with ta-desu, nakatta-desu, ru-desu and nai-desu followed by [N] are indicated in parentheses. Table 4 shows the following: (a) In Japanese SFF in the incubation period, speakers use Norms more than Non- norms in all areas except for the past negative (nakat-ta-desu) used by B. (b) A shows use of more appropriate SFF than B does (Examples 1-5), indicating that A better distinguished the use of the two series. In addition, fixed and idiomatic Norms were prominent in the speech of B, who had less contact with Japanese (Example 6). Normative forms and synthetic structure of Japanese in the incubation … 115 (1) A: Meeru-janakute koko-ni atebi-o kaku-n-desu-yo. (II 200711)1 e-mail-COP-NEG here-LOC date-OBJ write-[N]-COP-SFP ‘You must write the date here, not on the e-mail.’ (2) A: Mousikomi anata-ga si-ta-n-desu-ka. (II 200807) Application you-SUB do-Past-[N]-COP-INT ‘Did you apply for it?’ (3) A: Minna wakari-masi-ta-ka. (III 200810) everyone understand-FML-Past-INT ‘Did everyone understand?’ (4) A: Kochosensei-ga kimeru-tte okaasan-ga ii-masi-ta. (III 200810) principal-SUB decide-Quotation mother-SUB say-FML-Past ‘The mother said that the principal decides.’ (5) A: Nani-degozai-mashoo-ka. (IV 200903) what-COP-Extraordinary polite-FML-SPEC-INT ‘What is going on?’ (6) B: CD-wa mot-teori-masu-ka. Mot-teori-masen-ka. (III 200812) CD-TOP have-ASP-FML-INT have-ASP-FML-NEG-INT ‘Do you have the CD? Or don’t you?’ (c) There are characteristic differences by period, for instance that B’s use of Non- norms in the negative past is more prominent toward the 5th period, the period of attrition. Specifically, A generally used Norms whereas B used both Norms and Non- norms and transitioned to Non-norms. B’s usage was superior to A’s method in terms of her ability to convert the sentence styles in proportion to Japanese contact. (7) B: Isogasiku-te tegami-o kak-e-nakat-ta-desu. (V 201611) busy-Cause letter-OBJ write-POT-NEG-Past-COP ‘I was busy and could not write a letter.’ 1 The parentheses indicate the period of the utterance as the year and month. 116 HWANG Young Hee (8) B: Moo [gamansuru]2 gaman-deki-nakat-ta-desu. (V 201611) anymore [endure] endurance-can-NEG-Past-COP ‘I could not endure it anymore.’ Table 5 summarizes the percentage of use per person of colonial Korean Japanese and colonial Chinese-Korean Japanese of the incubation period in Table 4 (Decimal point, rounding off). The use of A and B were only for the 3rd to 5th periods, the incubation period of L2, and in the same way, I regarded colonial Japanese as Japanese of the incubation period from the 3rd to 5th period. Table 5: Usage of formal verb sentences in the L2 incubation period Sentence Informants Past Non-past Total Norms Non-norms Norms Non-norms Positive Returnee A - - 26 1 27 Returnee B 112 - 2 - 114 Koreans 22 39 69 32 161 Chinese-Koreans 52 79 90 61 282 Negative Returnee A 4 1 15 - 20 Returnee B 3 7 11 2 23 Koreans - 2 4 30 36 Chinese-Koreans 4 10 13 17 43 The distribution of usage of formal verb sentences in Table 5 can be generally understood in terms of the ability to use sentence forms according to linguistic functions in sentences, as well as the differences in the different informants’ levels of Japanese contact. In other words, there are many Norms in the order of A > B, but B uses Non-norms prominently in the Japanese negative past tense; this is also a common phenomenon with the colonial Japanese speakers in this study (Hwang 2010, 2013). Hwang’s conclusions on Japanese acquisition, maintenance, attrition, and regression for Korean learners were tentatively a mirror image as follows. (a) Non-past negative forms of formal verb sentences (i) masen > (ii) nai[N]desu > (iii) masen > (iv) nai[N]desu > (v) naidesu 2 [ ] is the interviewer's utterance. Normative forms and synthetic structure of Japanese in the incubation … 117 (b) Past negative forms of formal verb sentences (i) masendesita > (ii) nakatta[N]desu > (iii) masendesita > (iv) nakatta[N]desu > (v) nakattadesu In the incubation period, the normative and synthetic masu conjugation forms in formal verb sentences are in unmarked positions in terms of the numbers used (Examples 10 and 11), but in colonial Japanese, non-normative and analytical desu conjugation forms are in the unmarked position. (9) K: (Chuugakkoo-wa) iki-masen-desi-ta. junior high school-TOP go-FML-NEG-COP-Past ‘I did not go to (Junior High school).’ (10) A: Ki-te-masen-desi-ta-ka. (II 200807) come-ASP-FML-NEG-COP-Past-INT ‘Did not you come?’ (11) B: Urusaku-te nemure-masen-desi-ta. (II 200701) noisy-Cause sleep-FML-NEG-COP-Past ‘I could not sleep because it was noisy.’ In addition, Non-norms (ta-desu) of positive and past only appears in the acquisition period (1st and 2nd stage), which is characterized by being used in the typical form ta-deshoo with the confirmation and requirement functions (Examples 12-16). (12) A: Oneechan-ga mot-teki-ta-deshoo. (II 200607) sister-SUB bring-AUX-Past-COP-SPEC ‘Sister brought it, did not you?’ (13) A: Jeojjok-e hwajangsil. Mie-nakat-ta-deshoo. (II 200612) there-LOC toilet. see-NEG-Past-COP-SPEC ‘There's a toilet over there, did not you see?’ (14) A: Oneechan-ga tot-teta-deshoo. Kime-ta-deshoo. Hutari-de. (II 200701) sister-SUB take-ASP-Past-COP-SPEC. decide-Past-COP-SPEC. two people-with ‘Did sister take it? We decided with two people?’ 118 HWANG Young Hee (15) A: Kore junbisuru-kara doko-de it-ta-deshoo. Osoku dasi-te-nai-tte. (II 200712) This prepare-Cause where-LOC say-Past-COP-SPEC. late send-ASP-NEG-Quotation ‘Because I'll prepare this. Somewhere did I say that ‘Do not send late’?’ (16) B: Koo ioo-to omo-tta-deshoo. (II 200807) like.this say-Will-Quotation think-Past-COP-SPEC ‘You thought to say like this?’ In the above examples, the returnee sisters use non-norms in the verb SFF of Japanese in the incubation period substantially less than do the colonial speakers of Japanese. This indicates that the girls have well maintained norms inputted as chunks in the background rather than following the strategy of “rationalization (analysis) to desu” as a universal strategy that appears mainly in adult Koreans learning Japanese. 4.3 Use of i-adjective and na-adjective (noun) sentences Next, I examined the interrelationships between the SFF used in i-adj. and na-adj. (noun) sentences, as well as the forms used in the verb sentences described in previous sections. Table 6 summarizes each informants’ actual use of the formal forms of i-adj. sentences. After the acquisition period, the use of formal i-adj., na-adj. and Noun sentences decreased dramatically for the returnee sisters. The use of informal styles is indicated with brackets. Even though few in number, the translation surveys clearly show the tendency of decreasing Norms. I examined the results of the translational survey described in §3.2, and its use in a discourse. Table 6: Usage of the formal forms in i-adjective sentences Sentence style Returnee A Returnee B Koreans Chi.-Kor. Total Positive Past - (4) - (-) 10 39 49 Non-Past - (25) 1 (13) 97 59 157 Negative Past 5 (2) 4 (2) 4 3 16 Non-Past - (2) - (-) 37 4 41 Table 6 shows that A used masu forms such as atsuku arimasendesita ‘It was not hot’ (V 202003) in i-adj., whereas B only used desu forms: atsuku nakattadesu ‘It was not hot’ (V 202003). This suggests that in the i-adj., the learner's analysis strategy tends to commence after the learner reaches the 5th period of re-contact and attempts to unify SFF into the desu forms. Normative forms and synthetic structure of Japanese in the incubation … 119 Table 7: Usage of formal na-adjective (Noun) sentences Sentence style Returnee A Returnee B Koreans Chi.-Kor. Total Positive Past 1 (3) - (-) 7 24 32 Non-Past 2 (87) 2 (46) 282 140 426 Negative Past 6 (4) 5 (4) - 3 14 Non-Past - (1) - (3) 17 13 30 Table 7 summarizes the actual usage of formal noun sentences from the survey in the 5th period; the table shows that although A used both Norms and Non-norms (Examples 17ab, 18ab), B only used Non-norms (Examples 19 and 20). (17a) A: Kankokujin-janaka-tta-desu. (V 201611 and V 202003) Korean-NEG-AUX-Past-COP ‘He was not a Korean.’ (17b) A: Chuugokujin-dewaarimasen-desi-ta. (V 201703) Chinese-AUX-NEG-FML-COP-Past ‘He was not a Chinese.’ (18a) A: Sizuka-janaka-tta-desu. (V 201611 and V 202003) quiet-AUX-NEG-Past-COP ‘It was not quiet.’ (18b) A: Sizuka-dewaarimasen-desi-ta. (V 201703 and V 202003) quiet-AUX-NEG-FML-COP-Past ‘It was not quiet.’ (19) B: Chuugokujin-janaka-tta-desu. (V 201611, V 201703 and V 202003) Chinese-AUX-NEG-Past-COP ‘He was not a Chinese.’ (20) B: Sizuka-janaka-tta-desu. (V 201611, V 201703 and V 202003) quiet-AUX-NEG-Past-COP ‘It was not quiet.’ Thus, considering that A had remembered Norms as chunks in the early learning stage and that the sentences I investigated were formal rather than from discourse 120 HWANG Young Hee data, I consider the phenomenon to represent a style shift for A that reflected her higher Japanese ability. These results were the same for verb sentences (Examples 21, 22). (21) A: Isogasiku-te terebi-o mi-masen-desi-ta. (V 201703) busy-Cause TV-OBJ watch-FML-NEG-COP-Past ‘I did not watch TV because I was busy.’ (22) B: Sono-hito-wa kekkonsi-naka-tta-desu. (V 201703) That.person-TOP marry-NEG-Past-COP ‘He did not get married.’ 4.4 Linguistic factors related to the use of commentary predicate [N] Next, I examined the characteristics of the informants’ [N] through their discourse functions. As was the case in Examples 23-25, A distinguished [N] according to its discourse function in the sentence. (23) A: Ayamat-ta-n-da-ne. (III 200812) apology-Past-[N]-COP-SFP ‘I apologized.’ (24) A: Si-teru-n-da-yo. (IV 200903) do-ASP-[N]-COP-SFP ‘I'm doing it.’ (25) A: Watara-nakyaikenai-n-dat-ta. (IV 200908) cross.over-Obligation AUX-[N]-COP-Past ‘I had to cross over it.’ In the context of politely conveying simple facts, A was distinguishing the Norms and the proper ru-[N] and ta-[N] as she explained information that was unknown to her speaking partner. One case of using [N]-da to connect the sentence-final particle yo with the information presentation function and one case of connecting other sentence-final particles behind can be understood as examples of proper use according to the discourse function. Normative forms and synthetic structure of Japanese in the incubation … 121 However, B, who did show the use of Norms, also showed non-normative variations; her mixture of variations containing [N] is remarkable (Examples 26-28). There were five cases with sentence-final particle yo, two cases with other sentence- final particles, and eight cases without these particles. That is, in the discourse in the L2 attrition period, A had only two cases of no discourse function [N] of explanation, confirmation, or assertion, which was much fewer than B’s eight cases. I attribute this to A’s lower usage of Non-norms. (26) B: Suru-n-da-yo. (III 200812) do-[N]-COP-SFP ‘I will do it.’ (27) B: Nac-chat-ta-n-da-yone. (IV 200905) happen-ASP-Past-[N]-COP-SFP ‘That's happened.’ (28) B: Nat-ta-n-daroo. (IV 200908) happen-Past-[N]-COP-SPEC ‘Is it happened?’ 4.5 Collocation relationships with sentence-final particles (SFP) Finally, as shown in Table 8, I observed a continuous pattern of usage classification in the collocation relationships with sentence-final particles in the speech of A and B. Table 8: Korean returnee sisters’ usage of sentence-final particles Sentence styles A B I II III IV V I II III IV V Informal styles + Formal styles 1955 251 92 - 92 1419 144 143 1 82 yo 276 8 4 - 2 189 15 4 - - ne 132 2 - - - 76 2 5 - - na 54 6 1 - - 34 4 2 - - The ratio of sentence-final particles 24% 6% 5% 0% 2% 21% 15% 8% 0% 0% Notes: Only verb sentences are included. In addition, A had 140 cases (no, 72 cases) and B had 44 of sentence-final particles kai, ka, sa, ze, zo, no, mono, ya, and wa. This is considered a feature of child L2 in contrast with adult L2. 122 HWANG Young Hee For both A and B, the total usage of sentence-final particles decreased drastically beginning in the 3rd period. However, A showed more relevant sentence-final particles based on the degree of Japanese contact. A distinguished between Norms and Non- norms according to the discourse function in sentences. By contrast, B misused Non- norms that were inappropriate for the scene. She then began to use Norms incorrectly, and there was no use of [N] with explanation, confirmation, and assertion function where there should have been (Example 29). (29) B: Kak-e-nai, dat-ta-desu. (V 201611) write-POT-NEG, COP-Past-COP ‘I can write, it was.’ However, A and B simultaneously used fixed Norms more than colonial Japanese. From this reason I conclude that the verb SFF in L2 in its incubation period maintained sequentially according to the elapsed time after Japanese contact cessation along with the linguistic factors such as the usage according to the discourse function in the sentence. 5 Conclusions In this paper, I have described actual usages of maintenance, attrition, and regression of SFF in the L2 Japanese of two Korean returnee children compared with that in the colonial Japanese of thirteen elderly. Based on these observations which focused on the verb sentences of Japanese in the incubation period, I present the following study conclusions. (1) The lower the level of Japanese contact in the L2 incubation period, the more likely that the conjugation forms that predominantly use sentence Non-norms will mainly appear in analytical forms in the 5th regression period. [N] begins to intervene irrelevantly for a discourse function in the past and affirmative ta-desu. However, the usage of fixed and idiomatic Norms is maintained. These tendencies are stronger in the Japanese of two Korean child returnees than in colonial Japanese. There is also a higher ratio of formal Norms observed. (2) The informants all tended to use Non-norms in not only verb, but also adjective and noun, sentences. (3) The use of sentence-final particles and the use of Non-norms varied according to the sentence discourse function, which affected the analysis of SFF, correlated with the Japanese contact level as well as with the elapsed time (=age) of contact cessation. Normative forms and synthetic structure of Japanese in the incubation … 123 (4) In terms of the forms of Japanese sentences at the learning stage and at the cessation of contact with Japanese, it is possible to deduce the process of creating verb sentences across the life cycle by partly modifying Hwang’s (2013) conclusions that Japanese acquisition, maintenance, attrition, and regression for Korean learners is tentatively a mirror image. To conclude, I propose that the SFFs of Japanese in the incubation period are recursively reconstructed and treated as the interlanguage in which the ongoing analysis, which is often observed in adult Korean learners of Japanese, takes place in terms of conjugation forms. Acknowledgments This study was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2019S1A5A2A01035182). References Anderson, R. W. (1982). Determining the linguistic attributes of language attrition. Lambert, R.D. & Freed, B.F. (eds.) The Loss of Language Skills, 83-118, Rowley MA: Newbury House. Dorian, N. C. (1981). Language death: The life cycle of a Scottish Gaelic dialect. Philadelphia: University of Pennsylvania Press. Han, K. (2000). A Study on the English Loss of Returnees; focused on elementary school students (귀국학생의 영어손실에 관한 연구: 초등학생을 중심으로). Yonsei University Graduate School of Education Master's Thesis, Seoul: Yonsei University Press. Hansen, L. (1999). Investigating second language attrition: An Introduction, In Hansen, L. (ed.) Second language attrition in Japanese contexts; Tokyo, 3-18. Oxford University Press Hansen, L. (ed.) (1999). Second language attrition in Japanese contexts. New York; Tokyo: Oxford University Press. Hayashi, B. (1999). Testing the regression hypotheses, the remains of the Japanese negation system in Micronesia. In Hansen, L. (ed.) Second language attrition in Japanese contexts. 154-168. Oxford University Press. Hwang, Y. (2009). Pragmatic approach to interlanguage in the second language retention period; focused on the sentence-final particle Yo of Korean elder Japanese (第二言語保持 期の中間言語への誤用論的アプローチ: 韓国高年層日本語の文末詞ヨを対象に), Japanese Language Studies, 24, 225-245. The Japanese Language Association Of Korea. Hwang, Y. (2010). Preservation of sentence forms in Japanese learners during Japanese colony; focused on verb formal sentences (일제강점기 일본어학습자에 있어서 문말형식의 보존: 동사문 정중체를 중심으로). Japanese Language Studies. 28, 286-300. The Japanese Language Association of Korea. 124 HWANG Young Hee Hwang, Y. (2013). Preservation of desu and masu in Japanese of incubation period in Korean and China (韓国と中国の潜伏期日本語におけるデス形とマス形の保持), Japanese language education research, 25, 95-110. Korean Academic Society of Japanese Education. Hwang, Y. (2015). Second Language Retention Mechanism of Korean Japanese Learners; Variations according to Japanese contact type (韓国人日本語学習者の第二言語保持メ カニズム: 日本語接触の類型による変異形), Japanese language education research, 31, 273-291. Korean Academic Society of Japanese Education. Kan, K. (2004). Remaining Japanese in Taiwan (台湾に残存する日本語の実態 ). Osaka University Graduate School of literature Doctoral Thesis, Osaka: Osaka University Press. Kanazawa, H. (2008). The Japanese of foreign students is Japanese in the future; Dynamism of Japanese change (留学生の日本語は、未来の日本語: 日本語の変化のダイナミズム), Hitsuji Shobo. Kiguchi, M. (2004). Studies on the attrition of residual Japanese in Korea; Through the analysis of the Korean elder corpus (韓国における残存日本語の摩滅に関する研究: 韓国人高 齢者コーパスの分析を通して), Chung-ang University Doctoral thesis, Seoul: Chung-ang University Press. Kiguchi, M. (2009). A Study on the Change of Remaining Japanese in Korea (韓国における残 存日本語の変容をめぐって). Japanese Studies, 24, Seoul: Japan Research Institute of Chung-ang University. Kim, C. (2002). The need for bilingual education for returnees (귀국학생을 위한 이중언어교육의 필요성). Bilingualism, 21, 122-140. Institution of Bilingualism. Kim, M. (2010). Maintenance and attrition of the Japanese in child returnees (帰国子女の日本 語の維持と摩滅), Japanese Studies, 29(14), 183-195. Meiji Shoin. Kubo, M. (1994). Current Status of Child Returnees Education (帰国子女教育の現状 ), Japanese language, 13(3), 4-12. Meiji Shoin. Matsumoto, K. (2013). Pragmatic Variations and Changes in Palau Japanese (パラオ日本語の 語用論的変異と変化 ), Okamura, T. (ed.) The linguistic world of Oceania, 220-262. Keisuisha. Noda, S., Sakoda, K., Shibuya, K., Kobayashi, N. (2001). Grammar acquisition of Japanese learners (日本語学習者の文法習得), Taishukan Shoten. Noda, H. (2004). Factors related to the use of negative and non-negative statements; Based on the case study and the younger year anchor survey (否定ていねい形「ません」と「な いです」の使用に関わる要因 : 用例調査と若年層アンケート調査に基づいて ), Quantitative Japanese Linguistics, 24(5), 228-244. Ono, H. (1994). Retention of the ability of the returning child (帰国子女のバイリンガルの能 力の保持). Japanese language, 13(3), 4-12. Meiji Shoin. Park, A. (2014). Qualitative Research on the Reminiscence of Middle and High School Students in Korea; Focusing on the case of living in Japan with parents (귀국학생의 중·고등학교생활 회고담에 관한 질적 연구: 부모와 동행하여 일본 생활을 한 경우를 중심으로). Educational Anthropology Research, 17(2), 231-286 Sanada, S. (2009). Dynamics of Eastern Asia Residual Japanese (東アジア残留日本語のダイ ナミックス), Monthly language, Vol.38. No.1, 82-89. Normative forms and synthetic structure of Japanese in the incubation … 125 Sibuya, K. (1995). Possible expressions of Japanese remaining in the South Sea Islands (旧南洋 群島に残存する日本語の可能表現), Musha, 2, 81-96. Sibuya, K. (2001). The Status of Japanese Remaining in the Palau; Report: Preface (パラオにお ける日本語残存の実態 報告書: 序章), Sanada, S. (eds.) Investigations on dialects that have been disappearing in Japanese, Grant-in-Aid for Scientific Research (A), Emergency Surveys on the ‘Endangered Languages’ of the Pacific Rim, Achievement report, 285-302. Tanomura, T. (1994). A Quantitative Survey on the Choice of Negative Forms of Politeness; masen and naidesu (丁寧体の述語否定形の選択に関する計量的調査: 「～ません」 と「～ないです」), Osaka University of Foreign Studies, 11, 51-66. Tomiyama, M. (2000). Child second language attrition; A longitudinal case study. Applied Linguistics 21(3), 304-332. Tomiyama, M. (2004). Loss and maintenance of L2 (第二言語の喪失と維持), Terauchi, M. Kinoshita, K & Narita, M (eds.) The present of second language acquisition research, 239- 256. Taishukan Shoten. Tomiyama, M. (2008). Age and proficiency in L2 attrition: Data from two siblings, Applied Linguistics 30(2), 253-275. Weltens, B. & Cohen, A. D. (1989). Language attrition research; An introduction. Studies in Second Language Acquisition, 11, 127-133. Winford, D. (2003). An introduction to contact linguistics. Oxford: Blackwell. Abbreviations and symbols A. = Elder Korean returnee ASP. = Aspect AUX. = Auxiliary B. = Younger Korean returnee COP. = Copula, Be verb (e.g. da, desu) FML. = Formal i-adj. = i-adjective sentences (e.g. takai, yasui) INF. = Informal INT. = Interrogative LOC. = Locative [N]. = Nominalizer, Commentary predicate n(o)da na-adj. = na-adjective sentences (e.g. genkida, sizukada) Norms. = Normative sentences-final forms Non-norms. = Non-normative sentences-final forms NEG. = Negative OBJ. = Object POT. = Potential SFF. = Sentence-final forms SPEC. = Speculation SUB. = Subject TOP. = Topic