The Influence the Language Learning Environment on Negative Transfer and its Connection to Status and Face-Saving Refusal Strategies

While there is some research into the role of the target language community in understanding negative transfer and refusals (e.g., Beebe, Takahashi, Uliss-Weltz,1990; Ikoma & Shimura, 1993; Takahashi & Beebe, 1987; Yamagashira, 2001), most research has focused on determining the extent to which L2 proficiency influences negative transfer without close attention to status of the interlocuter or face-saving strategies (e.g., Félix-Brasdefer, 2004; Chang, 2009, Jiang, 2015). In response to this gap in the research, this study reports on findings from an investigation into the negative transfer of refusals among 59 students studying in Japan and the USA. Findings demonstrated that learning environments play a role in determining the frequency of negative transfers and that negative transfer was sensitive to status. Content-related findings suggest that the production of face-saving refusals is sensitive to the learning environment as well. Implications for pragmatic research comparing L1 and L2 environments are offered.


INTRODUCTION
The question of how L2 learners move toward pragmatic competence, according to Félix-Brasdefer (2004), has been an ongoing discussion among researchers in the field of interlanguage pragmatics since the 1970s. For many, L2 proficiency has played an important role in predicting learners' levels of pragmatic proficiency. The debate is played out in research on examining the relationships driving L2 proficiency, pragmatic transfer and pragmatic proficiency. Much of the contemporary research has focused on the use of refusals. Refusals are a face-threatening act (Brown & Levinson, 1987), often made in response to an invitation, request, offer or suggestion and so may require not only a high level of pragmatic competence to execute skillfully (Chang, 2009) but also input from target language speakers (Beebe, Takahashi, Uliss-Weltz, 1990;Takahashi & Beebe, 1987;Morkus, 2018).
Less understood is what role the target language community plays in negative pragmatic transfer and refusals in terms of frequency, content and status. Seminal research by Takahashi and Beebe (1987) examined the refusal formulas, i.e., patterning of speech acts within a refusal, of EFL (English as a foreign language) and ESL (English as a second language) learners and found evidence of more pragmatic transfer in EFL than ESL at both beginning and advanced proficiency levels. The explanation for this finding was that EFL learners had fewer opportunities for pragmatic transfer than their ESL peers. Follow-up research into the influence of the EFL versus the ESL community did not focus on the transfer of refusals. Rather, researchers examined the use of refusals, metapragmatic instruction and the role of textbooks (Ishihara & Cohen, 2010, Shimizu, 2009, cross-cultural issues and L2 proficiency (Bu, 2012;Chang, 2009;Jiang, 2015), gains in pragmatic ability in relationship to social membership in the EFL environment (Barron, 2003) and gains in speed and accuracy as a result of instruction (Taguchi, 2008).
The lack of attention to the negative transfer of refusals and their connection to status and their content in contemporary pragmatic transfer research is an important oversight that has left unanswered questions about the role of the learning environment in negative transfer across cultures. As such, this study compares the occurrences and content 34 IJALEL 10(2): [33][34][35][36][37][38][39][40][41][42] of refusals across three status levels between Japanese adult English learners at equal levels of proficiency in two learning environments, Japan and the USA. The ESL group had resided and studied English at a university in the United States for one year. The EFL group in Japan had studied English at a college of engineering and sciences in Japan. Two reference groups, Japanese and American, were drawn from their peers at the same universities respectively. Data was collected in Japanese and English with the discourse completion task (DCT) developed by Beebe, Takahashi & Uliss-Weltz (1990). Questions examined the occurrences and the content of negative transfer among the learner groups in Japan and the United States.

LITERATURE REVIEW
To date, only a limited number of investigations have been conducted regarding the relationship between learning environments and negative pragmatic transfer in refusals (e.g., Beebe et. al., 1990;Ikoma & Shimura, 1993;Yamagashira, 2001). Still fewer, have examined the connection negative transfer has to content and status. Most research has instead focused on the influence of L2 proficiency on negative transfer (e.g., Chang, 2009, Félix-Brasdefer, 2004Jiang, 2015) and, as such, has left an incomplete picture of the role that learning environments play in negative transfer across status levels.
Early research by Takahashi and Beebe (1989) explored the influence of the EFL versus the ESL learning environment among Japanese English learners at two different proficiency levels. The study examined the frequency of negative transfer, the ordering of semantic formulations and the content of semantic formulations. Of the three, frequency is most relevant to the findings in this study. Negative transfer occurred in just three scenarios among the learners studying in the United States, the ESL environment, while negative transfer occurred in nine scenarios among learners in Japan, the EFL groups. The higher frequency of negative transfer was not attributed to higher L2 proficiency, as there was less negative transfer among the higher proficiency group than the lower proficiency group. Status was addressed implicitly as a dimension of the ordering of semantic formulations and not negative transfer. However, numeric results taken from the DCT do speak to the question of status. Findings reveal three instances of negative transfer among the ESL group include one higher-status scenario, one equal-status scenario and one lower-status scenario. Of the nine scenarios in which negative transfer occurred among the EFL group, five were equal status and four were lower status.
In later research, Ikoma and Shimura (1993) explored the pragmatic pragmatic transfer of refusals among 10 native speakers of Japanese and 10 American learners of Japanese. The taxonomy for analysis of refusals developed by Beebe et. al., (1990) was used to analyze refusals to invitations, offers, requests and suggestions. Findings revealed that Japanese students provided fewer alternatives than their American counterparts. They made distinctions by status levels. For instance, they tended to overuse 'Kekko desu' in interactions with friends (equal status), probably because of the similarity to 'No thank you' in English. Both were attributed to negative transfer. Yamagashira's (2001) examined use of refusals of Japanese English learners, drew on the Beebe et. al., (1990) taxonomy for analysis, and conducted follow-up interviews to analyze data. Findings demonstrated that the highest occurrences of negative transfer occurred with interlocuters in the higher-status and second in the equal-status situation. Unlike in Beebe et. al., (1990), there was only minor evidence of negative transfer in scenarios where participants held lower status in relationship to their interlocuter.
In studies linking proficiency levels to negative transfer (Chang, 2009;Jiang, 2015), less attention is paid to the influence of the learning environments. Of the limited number that have, most have drawn on the concept of face to explain differences in the selection and content of refusals. Chang (2009) examined negative transfer among 81 Chinese students majoring in English. Forty-one were freshmen at a lower level of proficiency and 40 were seniors at a more advanced level. Two reference groups were used: 35 American college student and 40 Chinese students majoring in Chinese. Regarding the selection of indirect versus direct refusals, Chang found that selection of indirect refusals was higher among the Chinese groups than the American group and attributed this difference to the Chinese cultural norm to preserve face in any interaction.
Similarly, when the content of the refusals was examined, Chinese learners were less direct than the Americans. In refusals of offers, Chinese students relied more heavily on direct refusals than indirect. With respect to requests and invitations, Chinese learners provided fewer specific reasons when they had to refuse someone of higher status. Presumably, the fewer specific reasons would invite fewer follow-up questions and thus save face for both speakers. Jiang (2015) recently argued that the content of refusals varies across cultures. Jiang drew on the same instrument as Chang (2009) and demonstrated that Chinese students relied more heavily on general excuses in refusals to requests and invitations than Americans. The findings were independent of L2 proficiency and so Chang (2009) andJiang (2015) relied on the concept of face to explain the difference. Chang drew on work by Ting-Toomey (1988) to explain that learners in collectivist cultures will try to avoid conflict so that both interlocuters can save face. Cultures that are more individualistic, such as American culture, will use more direct strategies to make a refusal while learners in more collectivist cultures would naturally draw on more indirect refusals. These findings are supported in work by Beebe et. al., (1990) and by Jiang (2015).
Finally, with regards specific research into refusals, Bella (2014) has similarly argued that an exposure to the target language environment represents a key variable in acquisition. This confirmed previous research which suggests that acquisition of refusals may be independent of level of L2 proficiency. A lack of pragma-linguistic competence is generally associated with limited exposure to the target language community and appears in learners' reliance on simplified expressions (Codina-Espurz, 2013) The Influence the Language Learning Environment on Negative Transfer and its Connection to Status and Face-Saving Refusal Strategies 35

METHODOLOGY Participants
The participants in this study were as follows: 1) 16 Japanese university students in Japan serving as a reference group (JPR), 2) 10 Japanese university students in America (JLA) serving as a learner group, 3) 17 American university students (AMR) serving as a reference group and 4) 16 Japanese, also university students, learners of English in Japan. All had scores of 400 and 500 on the Test of English for International Communication (TOEIC). Consistent with research practices in pragmatic transfer (e.g., Bu, 2012;Chang, 2009;Jiang, 2015), reference groups were selected to provide an intercultural baseline of data. Responses from the AMR group would serve as a reference for English refusals while the JPR group would serve as a reference for refusals in Japanese. The use of a reference group for the Japanese language rather than using the Japanese responses from the JLJ group is a technique for establishing a corpus of pragmatic data that is not subject to the problem of participants simply copying or translating the refusals in English from their L1.

Instrument and Procedure
Data for this study was drawn from a series of DCTs implemented by a number of researchers (Beebe et. al., 1990;Chang, 2009). The taxonomy used for analysis was used by Beebe et. al. (1990). While the use of a DCT has been criticized for the limits placed on participants' language use (Hinkel, 1997;Morrison & Holmes, 2003), it does allow researchers to gather accurate samples to use quickly and efficiently (Beebe & Cummings, 1996). The DCT had 12 different scenarios organized into four groups, termed situations. Each situation contained three scenarios. The four situations were grouped as follows: 1) refusals to requests, 2) refusals to invitations, 3) refusals to offers, and 4) refusals to suggestions. Situations were further structured at three levels of status in relationship to the interlocuter with one scenario at placing the participant at a higher level of status in relationship to the interlocuter, another placing participants at an equal level and the third at the lower level.
The JPR and JPJ groups took the DCT online using Qualtrics in Japan. In the United States, the study was conducted online and in-person. The AMR group took the English-only questionnaires online while the JPR took it in-person and in Japanese. The learner groups, JLA and JLJ, took the questionnaire in English and Japanese in-person and on paper. The authors administered the study.

Data Analysis
Analysis was guided by the classification scheme described by Beebe et. al., (1990). Drawing on Beebe et. al., the refusals were organized as semantic formulas. A semantic formula refers to "a word, phrase, or sentence that meets a particular semantic criterion or strategy; any one or more of these can be used to perform the act in question" (Cohen, 1996, p. 256). For instance, if a participant wrote, "I am sorry, but I cannot meet with you tomorrow", it would be coded as expression of regret/negative ability/excuse. A shorter example, might be "I refuse", direct refusal/ performative. Consistent with work by Beebe et. al., (1990) and Jiang (20015), the structure of semantic formulations is determined by its relationship to the use of semantic formulations by other groups. The following relationships were used to determine the frequency of negative transfer. A refusal that met the requirements for one of the patterns below was classified as negative transfer. The use of the ≈ symbol is used to indicate nearly equal.
• The number of semantic formulations used by JPR is greater than the learner group, and the number of semantic formulations produced by JLA or JLJ is higher than AMR. (i.e., JPR>JLA or JLJ>AMR) • The number of semantic formulations used by JPR is lower than the learner group, and the number of semantic formulations produced by the learner group is lower than AMR. (i.e., JPR<JLA or JLJ<AMR) • The number of semantic formulations used by JPR and the learner group are equal or about equal while the number of semantic formulations produced by AMR is higher than JPR or JLA and/or JLJ. (i.e., JPR≈JLA and/ or JLJ > AMR or JPR ≈JLA and/or JLJ) • Both JPR and JLA and/or JLJ groups use one formulation that the AMR group never uses. • The AMR group uses one strategy that neither JPR group nor JLA and/or the JLJ group uses.
Following the identification of semantic formulations, the frequency of negative pragmatic transfer was calculated for each scenario across the learner groups. Data were reported in terms of percentages to make relative proportions of semantic formulations clear. The percentages were calculated by dividing the total number of participants in a selected group into the total number of participants who used the semantic formulations. Occurrences of negative transfer were classified into one-two-and three-part formulations. The refusal, "I have another appointment", would be calculated as an excuse and represent a one-part formulation. A two-part formulation would be, "I am sorry. I have an appointment". Here, the formulation is written as regret/ excuse. A three-part formulation would add another phrase or sentence.
To ensure the validity of responses, data was analyzed by two authors. Each worked independently to code data. Discrepancies in analyses were discussed until an agreed-upon interpretation could be reached. No data were dismissed. Data for learner and reference groups were organized onto an Excel spreadsheet.

Questions
1. Does negative transfer of refusal formulations occur more frequently among JAL or JLJ groups? 2. When negative transfer does occur, are there differences in the frequency of status-related refusals between the JLJ and JLA groups?
IJALEL 10(2):33-42 3. Does the content of refusals vary for English learner groups in terms of face-saving strategies?

FINDINGS AND DISCUSSION
The findings are presented in the following order. Data which compares general findings across the reference groups' use of direct and indirect refusals as well as adjuncts is presented first, as it establishes a baseline for more detailed comparisons with the learner groups. Next, data from each of the four situations is presented. Comparisons are made across learner and reference groups in the frequency of semantic formulations. Patterning in the frequency of semantic formulations which met at least one of the conditions for negative transfer as described by Beebe et. al., (1990) are identified and discussed. Finally, at the most specific level of analysis, the content of the refusals is examined with the purpose of identifying the connections between face and the content of refusals. The analysis is given with these priorities in mind but with one caveat. Kasper (1992) described differences between groups in terms of statistical significance. The participant pool in this study prevented a large enough sample size to develop an argument around statistical significance. As such, a descriptive account is given.

The Use and Negative Transfer of Direct Refusals and Adjuncts Among Reference Groups
While no evidence of negative transfer of direct refusals was found, a number of observations related to the use of direct refusals with regards to status are in (see Table 1). First, across all status levels, the use of direct refusals was higher in equal-status scenarios than higher-status and lower-status scenarios. Second, the data indicate that the JLJ participants differentiate between higher-and lower-status interlocuters by a greater percentage than the AMR participants. The JLA group shows the smallest variance between higher and lower-status scenarios. Further disaggregation of the data suggests that the use of adjuncts may play an important role. The AMR and JLA groups used more adjuncts than the JPR and JLJ groups to soften their refusals. The highest use of adjuncts was among the AMR group which used a total of 50 followed by the JLA group which used a total of 43. The JPR group was the lowest at 16, and the JLJ group produced 19. A discussion of negative transfer across each of the situations follows. As there were no occurances of negative transfer in the use of direct refusals, the following discusion is restricted to an analysis of indirect refusals.

Negative Transfer of Refusals to Requests in Situation 1
Findings from this situation suggest evidence of status-related negative transfer among the JLJ group. Table 2 provides an account of the one-and two-part combinations and evidence of three instances of negative transfer among the JLJ group occurring in the lower-and equal-status scenarios. No evidence of negative transfer was found in the higher-status scenario. No evidence of negative transfer was found among the JLA group in any of the scenarios. Given that the JLA group did not demonstrate any evidence of negative transfer, this suggests that L1 environment may be one possible explanation for negative transfer in this particular situation.
In scenario 1, the participant owns a bookstore and must refuse a request for a raise from an employee. The JPR≈JLJ > AMR pattern was found, indicating evidence of negative transfer in the use of the two-part regret/excuse formulation among the JLJ participants. Only a small gap, just four percent, separated the JPR and JLJ groups in their preference for regret/excuse formulation, suggesting that the JLJ group relied heavily on native refusal norms.
Twenty-nine percent of the JLJ group relied on the regret/excuse formulation in comparison to just six percent of AMR group, showing strong separation in refusal norms between the two groups. The JLA group, however, showed a smaller statistical difference in the use of the regret/excuse formulation with the AMR group. A comparison of the JLA group revealed a15% difference in the use of regret/ excuse separating the JLA (10%) and the JPR (25%) groups, suggesting Scenario 2 is the equal-status scenario in which the participants must refuse a request from a classmate to borrow lecture notes. Table 2 shows two instances of negative transfer among the JLJ group. In contrast with scenario 1 which relied on the JPR≈JLJ > AMR pattern, the first instance of negative transfer of the regret/excuse formulation was expressed in the JPR<JLJ<AMR pattern. Forty-seven percent of the JLJ group selected the regret/excuse formulation, placing them just six percent below the AMR group but 41% above the JPR group.
In the second occurrence of negative transfer, JPR=JLJ > AMR pattern was found in the use of the excuse formulation. This was similar to the JPR≈JLJ > AMR pattern found in scenario 1. The use the excuse formulation is at 19% for both the JPR and the JLJ. Findings also revealed that equal percentages of the JLJ and JPR used criticism and non-performative (NP), suggesting a strong case of shared norms between the two groups in this scenario.
There were no instances of negative transfer in scenario 12. Unlike scenarios1 and 2, the JLJ group appeared to have aligned itself more closely with the AMR group than the JLA group. Selection of regret/excuse, the most common two-part formulation for both learner groups, was very close to the JPR group. As shown, 40% of the JLA group and 59% of the JLJ group chose the regret/excuse formulation. The JPR group similarly relied on regret/excuse more than any other refusal. Forty-three percent of the JPR chose regret/ excuse, placing it in between the JLA and JLJ groups.

Negative Transfer of Refusals to Requests in Situation 2
Table 3 reveals that negative transfer occurred at all three status levels in the use of the regret/excuse formulation for the JLJ. The JLA group, in contrast, split the use of negative transfer between scenarios three (higher status) and four (lower status), suggesting evidence of status-related differences in negative transfer. Use of the regret/excuse formulation was consistent throughout and just two different patterns for negative transfer were used. In scenario 3, the participant is the president of a printing company and must refuse the offer of a salesman for dinner. While both learner groups showed evidence of negative transfer of the regret/excuse formulations as expressed in the JPR>JLA or JLJ>AMR pattern, distribution of the refusal formulations differed between the learners groups. For the JLJ group, this suggested a closer approximation of L1 refusal norms when compared with the JLA group. Table 3 shows an 11% gap in the use of the regret/excuse formulation between the JLJ and the JPR but a 52% difference between the JLA and JPR groups. The opposite distribution was evident in comparison with the AMR group. Fifty-three percent more of the participants in the JLJ group drew on the regret/excuse formulation in comparison to JLA group, placing the JLA group much closer to the response rate of the AMR(6%) participants than the JLJ group. Scenario four is the contrasting higher-status scenario, but, unlike scenario three, scenario four shows more similarities than differences between the learner groups. Scenario four asks participants to refuse an offer from the boss to attend a dinner party with the additional problem being that the participant's spouse does not like the boss's wife. Table  three shows similar patterns of negative transfer between the learner groups. As such, there is a nearly equal occurrence of the regret/excuse formulations for the JLJ (47%) and JPR (43%) groups, placing their responses at parity with the JPR group. When scenarios four and three are compared, findings show that the JLA group increased their use of the regret/excuse formulation by 30% while the JLJ group showed just a 10% gain. The JPR group decreased their use of the regret/excuse formulation by 19%, placing both learner groups at near parity Scenario 10 is the equal-status scenario where the participant must refuse an invitation from a friend for dinner. The participant, however, does not like the friend's spouse. It is the only scenario of the three in which there is no evidence of transfer among the JLA group. Consistent with the JPR= JLJ > AMR pattern, participants in the JLJ group matched the JPR group in their use of the regret/excuse formulation. While the occurrences of request/excuse formulations in the JLJ group was quite close in scenarios 10 and four, the same was not true for the JLA group. The number participants in the JLA group increased by 20% in scenario four when compared with scenario 10, suggesting a status-related explanation for the change in the frequency of regret/excuse formulations.

Negative Transfer of Refusals to Offers in Situation 3
With no occurrences of negative transfer among the JLA group, this situation suggests that the L2 learning environment plays a stronger role in negative transfer than the target language community. Furthermore, Table 4 suggests evidence of status-related negative transfer for the JLJ group. Findings indicate three instances of the JLJ group relying on negative transfer, two in scenario 9 (equal-status scenario) and one in scenario 11 (the lower-status scenario), but not in scenario 7 (the higher-status scenario). No evidence of negative transfer was found for the JLA group.
Scenario 9 required two responses and so is divided into 9a and 9b. In scenario 9, the participant has to refuse an offer for a second piece of cake from a friend. Consistent with the JPR>JLJ>AMR pattern, the percentage of the JLJ participants drawing on the excuse formulation was between the AMR and JPR groups for both 9a and 9b. All refusals by JLA group were in the form of a one-part formulation and to the  point. This was very different from the AMR group in which 65% of the participants relied on two-part formulations and just 18% drew on one-part formulations. The one-part excuse formulation also functioned as negative transfer in 9b among the JLJ group. Again consistent with the JPR>JLJ>AMR pattern, the percentage of participants who selected the excuse formulation was between the JPR and AMR groups.
The final instance of negative transfer occurred in the scenario 11 with the use of the excuse formulation. In it, the participant must refuse an offer for a promotion that requires moving to a small town. The JPR group had the largest frequency of the excuse formulation and demonstrated negative transfer as expressed in the JPR<JLJ<AMR pattern. Both the AMR and JLA groups had a low selection rate for the excuse formulation, 10% for the JLA group and 12% for the AMR group.
In scenario seven, no evidence of negative transfer was found, but data shows strong evidence for the use of the hook formulation across all groups. Eighty percent of the JLA group and 63% of the JLJ group chose hook as a one-part formulation. Similar numbers were present with the reference groups. Eighty-eight percent of the JPR group and 82% of the AMR group chose hook as a one-part refusalformulation. The selection of the hook formulation is an unresolved issue in this research, a point that is consistent with work by Takahashi & Beebe (1987) and Yamagashira (2001).

Negative Transfer of Refusals to Suggestions in Situation 4
Situation four suggests a slightly stronger argument for negative transfer in the target language environment than the L1 environment, but less evidence of differentiation between status levels for the JLJ group than the JLA group. More specifically, negative transfer occurs among the JLA in all three scenarios. Negative transfer was expressed through the same pattern, JPR>JLA>AMR, for scenarios eight and five. Table 5 reveals that scenario eight casts the participant as a language teacher who must refuse the request of a student to add more conversation practice to class. The use of the excuse formulation for the JLA is between the JPR and the AMR, suggesting negative transfer. Notably, this is the only scenario of the three in which the JLJ group did not demonstrate evidence of negative transfer. While an increase of just two percent to 30% in the number of respondents would have made the responses of the JLJ group consistent with the JPR>JLJ>AMR pattern and thus indicated negative transfer, the JLJ group remains distinct from the JLA in their use of refusals. For instance, there were no uses of criticism by the JLA group but 36% for the JLJ used them. Conversely, guilt was not used at all among the JLJ, the AMR or the JPR groups, but it was present in 20 % of the JLA participants.
Scenario five reveals the final example of negative transfer. The refusals are in response to a suggestion from the boss to clean up the desk and write little reminders to finish various tasks. The JLJ transfers the use of excuse formulation, and negative transfer is consistent with JPR>JLJ>AMR pattern. Forty percent of the participants rely on the excuse formulation in in their refusal to a friend who suggests a new diet. The use of the excuse formulation is consistent with the JPR group, which was recorded at 55%, suggesting that JLJ relies closely on JPR norms. Probably since it was a diet, participants across groups made refusals which involved a lack of enthusiasm. As an example, AMR participant wrote, "I am not really into diets." Both learner groups demonstrated evidence of negative transfer in scenario six. The lower-status scenario asked the participants to refuse a suggestion from the boss to clean up the desk and write little reminders to finish various tasks. Evidence of negative transfer was found in the use of the excuse formulation. Also, the JPR<JLJ or JLA <AMR pattern was identified for both groups. A comparison on the learner groups revealed that 19% fewer of the participants in the JLA drew on the excuse formulation. Further, the JLA group restricted their use of one-part formulations to excuses while the AMR group includes excuse, promise and threat. Examples from the AMR are striking in their candor, "Boss, maybe you should try and organize yourself," (criticize) and "Maybe I will" (indefinite).

Content of Refusals
Content differences and similarities across learner groups were examined for the purposes of understanding the connections between face and status. Consistent with work by Chang (2009) andJiang (2015), the most common formulation, excuse, was selected for analysis. Two kinds of analyses were done. First, excuses were analyzed in terms of their specificity. Specific refusals included a specific event and/or a specific date or time within the excuse while non-specific refusals referred to general activities only. The distinction between specific and non-specific refusals is important, as it speaks to the workings of face in interpersonal exchanges. According to Merkin (1990), specific refusals, are common among American English-speakers who tend to draw more questions and can thus be viewed as face-threatening. General excuses are common to Japanese speakers and may not receive follow-up responses (Takahashi & Dufon, 1989). The second analysis concerned the content of direct versus indirect refusals. Direct refusals represent a more face-threatening act than indirect refusals and are found to occur more frequently among Americans when compared with Japanese (Beebe et. al., 1990), Indonesian (Widanta, Hudiananingsih, Sitawati & Ardika, 2019), Javanese (Wijayanto, 2016) and Egyptian-Arabic (Morkus, 2018).

Analysis of content of situation 2
The data suggest face-saving strategies are sensitive to place of residence among the learner groups. Evidence of this can be found in the distribution of responses in the use of unspecified excuses in equal and lower-status scenarios. The distinction between status levels was seen in situation two among the JLA group participants. With regards to the reference group data, the JPR group shows strong distinctions across all three status levels. In scenario four (the lower-status scenario), non-specific excuses total 29% while this tendency is reversed in scenario three (the higher-status scenario) with non-specific excuses at 37%. The AMR group shows strong contrast in responses to scenario four and 10 with non-specific responses at 57% and 69%, respectively. In scenario three, the distributions are similar between the two groups with both favoring the unspecified.
As for the selection of content, Table 6 reveals that most used excuses were "I am busy, sorry" or "I have already made other plans". Because they were used without reference to what the other plans might be or how long they might last, it was classified as a non-specific excuse. "I am busy, sorry" was used five times in scenario three and four by the JLJ group, but was most prevalent in scenario 10 where it was used four times by both the JLA and JLJ groups. It was used less in the reference groups. The JPR group used it twice in scenario three and four, respectively and once in scenario 10. The AMR group used it in scenario four, but did not use it in either scenario three or 10.
The use of "Sorry, I am busy" among the Japanese is important for its connection to Japanese culture. The direct translation from Japanese is "Gommennasai, ima insogashi desu", which literally means "Sorry, (I) am busy". All three Japanese groups used it when they responded in Japanese. The use of the literal translation at least suggests a reference to the face-saving strategy common to Japanese culture.

DISCUSSION
While there is some research into the role of learning environments in understanding negative transfer and refusals (e.g., Beebe et. al., 1990;Ikoma & Shimura, 1993;Takahashi & Beebe, 1987;Yamagashira, 2001), it has been done without adequate attention to the workings of content, status and face-saving strategies. Instead, most contemporary research has focused on determining the extent to which L2 proficiency influences negative transfer (e.g., Félix-Brasdefer, 2004;Chang, 2009, Jiang, 2015. The shift in research has left open important questions about the role of the learning environment in negative transfer. In response, this study compared the frequency of negative transfer in the use of refusals across two groups of Japanese participants, one studying in Japan and the other in the United States. Findings demonstrated that learning environment plays a role in determining  IJALEL 10(2):33-42 the frequency of negative transfers and that negative transfer was sensitive to status. Content-related findings suggest learning environments' effect on the production of face-saving refusals.
At the broadest level, the first finding suggests that the EFL environment has a stronger influence on the frequency of negative transfer than the ESL environment and is confirmed in research by Takahashi and Beebe (1989). The JLJ group drew on negative transfer in eight scenarios while the JLA group relied on negative transfer in just four. The total number of occurrences of negative transfer for the JLJ was 10. Two occurred in scenario two and two in scenario nine, both equal-status scenarios. Negative transfer occurred in the JLA group four times. In comparisons of direct and indirect refusals between the JLJ and JLA groups, findings demonstrated higher occurrences of negative transfer among the JLJ group.
When the findings are parsed in terms of status, they suggest the differential influence of the learning environments. This finding adds to the still limited amount of research which examines the role of the learning environment on the extent to which negative transfer is sensitive to status (e.g., Takahashi and Beebe, 1989). Of the 10 instances of negative transfer among the JLJ group, five were found in the equal-status scenarios. Three were located in the lower-status scenarios, and just two were found in the remaining higher-status scenarios. JLA group showed a stronger distinction. Two instances of negative transfer were present in the higher-status scenarios, and two were in the lower-status scenario. Just one instance of negative transfer occurred in the lower-status scenarios. Finally, just one occurrence of negative transfer was identified in the equal status scenarios.
Other findings which did not demonstrate evidence of negative transfer contribute to the research on status-related differences in the use of refusals. First, consistent with Beebe et. al., (1990), the use of direct refusals by the JLJ group demonstrated that native speakers of Japanese tend to differentiate more strongly between lower and higher-status interlocutors than native speakers of English. The JLA group, however, was the exception, showing very little distinction between status levels. Second, the higher use of adjuncts by the AMR and JLA group than the JPR and JLJ groups is consistent with findings by Chang (2009) who compared Chinese and Americans' refusal patterns. Adjunct are used to soften a request and are sensitive to status. Second, after approximately one year abroad, the JLA group produced significantly more adjuncts than their peers in Japan . This was in spite of the fact that both learner groups were at the same level of L2 proficiency and in contradiction to previous research which connects the use of adjunct to L2 proficiency (Morkus, 2018).
The final area of study concerns the connections between face-saving strategies and the content of status-related refusals. The content of the refusals was analyzed in terms of the connections between the concept of face and specific versus non-specific excuses for situation two. Findings from scenario two revealed differences in the use of unspecified excuses between the learner groups in scenarios three (higher status) and four (lower status). Scenario two also showed the same patterning of negative transfer for the JLA group. To date, researchers have argued that Japanese frequently make status-driven distinctions in the use of refusals (Beebe, et. al., 1990;Ikoma & Shimura, 1993;Yamagashira, 2001), but they have not linked these distinctions to place of residence and face-saving strategies. The finding is important because it introduces the possibility that negative transfer may not only be sensitive to place of residence but also to the use of face-saving strategies.

Conclusion and pedagogical implications
Findings from this research suggest that the influence of the negative transfer and time spent in the target language can inform classroom instruction. The analysis of how content and face saving strategies, a finding which demonstrated distinctions between the two learner groups, offers a starting point. In a class of adult learners, a teacher might begin with surveying the students on who and who has not studied abroad. This will give a thumbnail sketch of the kinds of pragmatic skills in the target language the students may already have. In order to begin instruction, Limberg (2015) suggests students explore differences between the native and target language use of the speech act. Students can interview classmates and friends about how refusals, in this case, change according to the level of status, e.g., a friend or a stranger. Textbooks are another source or pragmatic data, but as Ishihara and Cohen (2010) notes, it is important to understand that textbooks often lack examples of speech acts that are applicable. Creating their own scenarios from their findings can be an active and interesting next step. Students can work in groups and practice their scenarios. Follow-up discussions should be encouraged which encourage discussions about how status, culture and specific circumstances might influence the realization of a refusal.
In conclusion, this research has re-opened the question of to what extent learning environments figure into negative transfer. Findings linked place of residence to the frequency of negative transfer and showed that negative transfer is sensitive to status. More informative was that negative transfer was sensitive to face-saving strategies. While this was only found in one scenario, it is an area that deserves further attention. To date, research into this area has been limited to research on Chinese learners of English (Chang, 2009;Jiang, 2015) and is an area of study which has not received attention in research on the influence of the learning environment. As such, these findings are an early attempt to bring attention to the research on negative transfer and the learning environment. Future research could narrow investigations to specific settings within the target language environment. Much has been done in that regard, e.g., home-stay interactions (Cook, 2008); Kinginger, 2008), dorm room conversations (Diao, 2016), service encounters (Shively, 2011) and extra-curricular activities and part-time jobs (Taguchi, 2015), but, again, not in relationship to negative transfer and certainly not with eye towards understanding how they might be sensitive to face-saving strategies. Understanding how time spent in these settings interacts with the workings of negative transfer, status and face-saving strategies is a possible next step and one that will inform teachers and researcher alike.

ACKNOWLEDGMENT
I would like to thank the reviewers for their helpful critique during the preparation of this manuscript. Also, I would like to thank Nicholas Mitchell for his technical assistance in the analysis of the data.