Authentic Communication: Proceedings of the 5th Annual JALT Pan-SIG Conference.
May 13-14, 2006. Shizuoka, Japan: Tokai University College of Marine Science. (p. 32 - 47).

Preferred complaint strategies in Japanese and English

by Carol RINNERT (Hiroshima City University)
Yoko NOGAMI (Graduate student, Hiroshima City University)
Chiaki IWAI (Hiroshima City University)


This study, based on two data collection stages, aims to determine what English complaint strategies are preferred by Japanese university EFL (JEFL) learners. The first stage compared Japanese complaint formulations with previously collected English responses by JEFL learners and native English speakers in two complaint situations. The second stage elicited judgments of appropriateness and effectiveness of various complaint formulations in the same two situations. The findings from this study indicate which aspects of complaints may cause difficulties for JEFL learners. This study suggests the need to raise their pragmatic awareness regarding the use of complaint strategies in particular contexts.

A complaint is defined as an expression of "displeasure or annoyance" in response to an action that is seen by the speaker as unfavorable (Olshtain & Weinbach, 1993, p. 108). Complaining about a perceived problem can be extremely face-threatening for both speaker and hearer, whether in one's native language or not (Brown & Levinson, 1987; JACET SLA-SIG, 2005). Most seriously, non-native speakers run the risk of unintentionally offending others by complaining too directly, which can be seen as criticism and lead to unpleasant consequences (Murphy & Neu, 1996).
In a previous cross-cultural survey of speech act productions (Iwai & Rinnert, 2001), English Discourse Completion Tests (DCTs) were used to elicit complaints in two fairly serious situations in which the relationship between speaker and hearer differed (see Appendix 1 for exact wording):
Situation 1: A professor made an apparent grading error
Situation 2: A roommate frequently made noise late at night

The aim of the present study is twofold. The first goal is to determine which factor (grammatical, pragmalinguistic, or sociopragmatic) may have caused Japanese students to choose a specific English complaint strategy by comparing their English complaints with complaints in Japanese produced by similar Japanese university students. The second goal of this study is to determine what English complaint strategies are judged to be most socially appropriate and potentially effective by English NSs, fluent NNSs, and JEFL learners in these same two situations, and by extension which strategies are seen as potentially problematic across cultures.

[ p. 32 ]


Data collection

In the first stage of this study, a questionnaire survey about complaint speech acts in Japanese (see Appendix 1 for an English translation) was conducted online in the fall of 2005. A total of 196 Japanese university students answered the questionnaire. A majority of them were first to third year students majoring in international studies, arts, economies, humanities, and information technology. Additionally, in order to compare the the complaints by Japanese university students with English complaints by native English speakers (US) and Japanese EFL learners (JE), the previously collected DCT responses by American and Japanese respondents (n = 100 for each group, Rinnert & Iwai, 2003) were also used.
For the second stage, a questionnaire was constructed based on prototypical and potentially problematic English complaint strategies from the Rinnert and Iwai (2003) study (see Appendix 2 for sample questionnaire items). The English questionnaire was administered online to elicit judgments of appropriateness and effectiveness from native (NS) and fluent non-native (NNS) English speakers (n = 31) in the summer of 2005, and a translated Japanese version was given to less fluent JEFL students (n = 40) in the spring of 2006. The native speakers (n = 20) and fluent non-native speakers (n = 11) had an average age of 44.8 years (range 22 - 70). They were mainly teachers with high levels of education (8 held BA, 15 MA, and 7 PhD degrees) from eight countries (the U.S., Japan, U.K., Canada, Denmark, Australia, Germany, and Italy). In contrast, the less fluent JEFLs (n = 40) were all Japanese undergraduate students (second and third year International Studies majors) with little or no overseas experience.


The analysis of the data collected in the first stage of the present study focused on three aspects of the complaints: (1) main components, (2) level of directness, and (3) amount of mitigation. The main components consisted of Initiators (I: greetings, address terms, and other opening formulas), Complaints (C: expressions of negative evaluation, including justification) and Requests (R: direct or indirect attempts to get the hearer to redress the situation), or combinations of these components. Three levels of directness of the complaints were identified:
  1. Indirect (no explicit mention of offense, implied offense only);
  2. Somewhat direct (mention of offense, but no mention of the hearer's responsibility);
  3. Very direct (explicit mention of offense and hearer's responsibility for it).
For the determination of the amount of mitigation, softening expressions were identified and counted. Examples of softeners from the previously analyzed English complaints included a little, sort of, you know, would/could, and I think/wonder; Japanese softeners found in the Japanese complaints included chotto [a little, a bit], toka [or something], warui kedo [sorry, but], dekireba [if possible], and ki ga suru [feel like] (see Appendixes 3 and 4 for complete lists of all softeners).
The second-stage analysis looked at the relative effects of the same three aspects (components, directness, mitigation) on the judgments of appropriateness and effectiveness of twelve systematically constructed complaints and four distracters, using a 5-point scale ranging from very appropriate/effective (5 points) to very inappropriate/ineffective (1 point), along with an opt-out choice of "I can't determine" (see Appendix 2). The effects of the three factors were tested statistically by means of a 3-way analysis of variance [ANOVA: components (3 levels: I + C, I + R, I + C + R) x directness (2 levels: Direct, Indirect) x mitigation (2 levels: minimal softening, multiple softeners)]. In addition, the judgment scores were compared across the two groups: NSs/fluent NNSs (hereafter referred to as NFSs) 2 vs. less fluent EFL learners (hereafter referred to as JEFLs) through a 4-way ANOVA (components x directness x mitigation x group).

[ p. 33 ]


Stage 1

Component analysis

The first analysis concerns overall response patterns. As explained above, each response in the DCT data was classified into one of seven response patterns based on which speech act combinations (Initiator, Request, Complaint) it contained.

Figure 1
Figure 1. Complaint response patterns in the Professor situation

The most common response patterns among all groups wereI + C and I + C + R in this situation. The frequency of the C + R and I-only responses among the less fluent EFL group varied considerably from the other two groups. I-only usage by the this same group was one-third less than in the other two groups. The C + R response, however, was three times more frequent in this group. Although none of the results were statistically significant, a statistical difference that was significant was observed among all three groups for combinations in which no initiator appeared. In other words, this means that if ****** STATE CONDITION ****** then ***** STATE RESULT **********. Table 1 indicates the frequency of complaints without initations for the three groups.

Table 1. Frequency of no initiator component patterns (C, R, & C+R)
Group: JJ JE US
Sample Size: 196 100 100
Percentage of No initiator use: 13.3% 30% 16%
χ2 = 8.44, p = 0.015

[ p. 34 ]

Figure 2 compares the same parameters when a roommate is involved rather than a professor.

Figure 2
Figure 2. Complaint response patterns in the roommate situation

In this case, the χ2 was 15.25 and level of significance 0.001. Almost half of the US group used I + C + R combination responses. However, JE respondents offered this combination at only about a 10% level, while 27% of JJ respondents employed this pattern. The production of the JE respondents demonstrates the fact that they tend to use a different strategy in the English language, which differs from their native language and the target language norms.
Other features which stand out include the fact that C + R and I + C combinations are analogous among the three groups at around 20% and 15%, respectively. Furthermore, variations can be seen in the C-only segment. Making complaints (including justifications) without Initiators and Requests is significantly more frequent in the JJ and JE groups than the US group (χ2 = 8.59, p = 0.014). It is likely that whether response patterns include an Initiator component or not is one of the important differences across the three groups (Table 2). The differences among the groups for the two factors were significant (χ2 = 7.17, p = .027 and χ2 = 11.79, p = .003 levels, respectively).

Table 2. Initiator non-containing component patterns; F situation (frequency)
Group: JJ JE US
Sample Size: 196 100 100
% Initiator contained:1 44.4% 29% 59%
% No initiator:2 52.5% 56.0% 32.0%

N.B. 1 Including I, I + C, I + R, I + C + R reponse pattern
2 Including C, R, and C + R response patterns

[ p. 35 ]


The second analysis concerns the level of directness of complaint components. The numbers given in Figure 3 are the means degrees of directness, based on three levels.
A one-way ANOVA was carried out for each situation, and the results indicate that the group difference was significant in both situations (with the professor, F = 18.29, p = .000 and with the roommate F = 5.29, p = .005). The JE average directness was notably higher than the other two groups in the situation with the professor. The differences between the JJ and JE groups and between the JE and US groups were significant in accordance with post hoc Scheffé tests (at p = .000 levels for both). The roommate situation results also indicated that the group differences for JJ vs. US as well as JE vs. US were significant. The average levels of directness of the JJ and JE groups were similar, and higher than that of the US group (between JJ and US groups, p = .034, and between JE and US groups, p = .01).

Figure 3
Figure 3. Direct complaint components


Thirdly, we investigated the use of softeners with both professors and roommates quantitatively. Figure 4 displays a summary of the complaint patterns when dealing with the professor. The results for the JJ and JE groups appear similar to each other, but not to those of the US group (χ2 = 89.34, p = 0.000).

Figure 4
Figure 4. Use of softeners with complaining to a professor

[ p. 36 ]

Almost 65% of the JE respondents (n=**) and over half of the JJ respondents (n=**) did not use any softeners. In contrast, more than 56% of the US group (n=**) used multiple softeners. This illustrates that native Japanese speakers and JEFLs use considerably smaller numbers of softeners than English NSs.
In the F-situation, more than 55% of the US respondents employed manifold softeners. Most notably, more than two-thirds of the JE respondents did not apply any softeners, and none used more than two softeners. The JJ group softener usage is similar to the US group usage. No-softener use was only 14%; however, the percentage of multiple softeners reached 38%.
Examining the JJ group softener use in the two situations, large differences (χ2 = 98.42, p = .000 level) can be observed, as seen in Figures 4 and 5.

Figure 5
Figure 5. Use of softeners with complaining to a roommate

Stage 2

Complaint Evaluations

The second analysis focuses on evaluations of selected English complaint formulations. The separate judgments of appropriateness and effectiveness for each complaint item on the evaluation questionnaire were compared across the two groups: NFSs and JEFLs.

Professor situation

The mean appropriateness and effectiveness judgment scores in the P-situation are displayed in Figure 6. The left third of the graph (containing four judgment scores for each line) represents the Initiator + Complaint component combination; the middle section, the Initiator + Request combination; and the right section, the three components combined (Initiator + Complaint + Request). Within each of the component sections, the two judgments of the direct versions are shown on the left (the unmitigated version first, then the mitigated one), and the indirect versions are on the right (again unmitigated, followed by mitigated).

Figure 6
Figure 6. Group means of acceptability/effectiveness when complaining to a professor

[ p. 37 ]

As can be seen in Figure 6, the NFS judgments of appropriateness for each of the three component combinations form a clear progression from the lowest scores on the left (direct, unmitigated), second lowest next (direct, mitigated), to second highest (indirect, unmitigated) after that, and the highest (indirect, mitigated) on the right. Thus for the NFS group, the indirect complaint versions were perceived as more acceptable than the direct ones, and within each of these categories the mitigated version was judged more acceptable than the unmitigated one. As can be seen by comparing the two lines for each group, in most cases the judgments of appropriateness and effectiveness were fairly similar, especially for the NFS group.
Statistical analysis of the judgment scores showed that although component was not a significant factor, indirectness and mitigation both strongly influenced judgments of acceptability and effectiveness of complaints in the P-situation. In particular, indirect strategies were rated as much more acceptable (F = 252.56, p = .000) and effective (F = 14.80, p = .000) than direct ones, and mitigated versions were much more acceptable (F = 90.34, p = .000) and somewhat more effective than unmitigated ones (F = 4.51, p = .011) for the two groups combined.
Although the groups did not significantly differ overall, their assessments of indirectness and components were somewhat different. Most notably, within the Initiator + Request component, JEFLs judged direct and indirect requests as equally appropriate, and direct requests as more effective than indirect ones. The two groups also differed in their component preferences: Initiator + Complaint alone by the NFS group, Initiator + Request alone by the JEFL group.
The mean judgment scores of the appropriateness and effectiveness of the complaints a roommate for the two groups are presented in Figure 7.

Figure 7
Figure 7. Group means of acceptability/effectiveness when complaining to a roommate

Opposite to the P- situation, in the F-situation directness was perceived to be significantly more appropriate (F = 17.25, p = .000) and effective (F = 73.43, p = .000) than indirectness for the two groups combined. Thus, both groups preferred more direct strategies when complaining to a friend.

[ p. 38 ]

On the other hand, although mitigation significantly raised the level of both appropriateness (F = 8.32, p = .004) and effectiveness (F = 4.20, p = .041), for the NFS group, it had no effect on the JEFL group. In other words, the NFS group preferred softened complaints over non-softened ones to the roommate as well as the professor, whereas the JEFL group only saw the same necessity when speaking to a higher status interlocutor.
Third, the component factor exerted a significant effect on the appropriateness (F = 6.12, p = .002) and effectiveness (F = 12.12, p = .000) judgments of the two groups combined. In particular, complaint strategies that included Requests were judged more appropriate and effective than those without Requests, and for both groups, the combination of Initiator + Complaint were considered the least appropriate and effective in this situation.
Finally, the group factor was found to be statistically significant. Overall, the EFL effectiveness judgments were significantly higher than those of NFSs. In other words, the learner group showed more optimism about the success of complaining to a friend, regardless of the strategies employed.


The first research aim concerned which factors (grammatical, pragmalinguistic or sociopragmatic) led to Japanese EFL learner's dissonant complaint production as compared to other speakers of English as a first or second language. This question was addressed mainly from three aspects using the cross-cultural and cross-linguistic DCT data.
First, regarding components the use of Initiators showed a significant difference among the three groups as attested by Rinnert and Iwai's (2003) cross-regional study. JEFL students did not use Initiators as much as respondents of the JJ and US groups in either situation. This leads us to infer that pragmalinguistic problems arise for Japanese producing English complaints, especially when speaking to the professor, where the difference appeared clearly. When addressing the roommate, the frequency of Initiators decreased from the US, to JJ, to JE groups, which may indicate that there were slightly different sociopragmatic norms in Japanese and English in this kind of situation between friends. Moreover, in the roommate situation, the JE's infrequent use of the I + C + R response pattern appears to reflect JEFLs' inadequate knowledge of that pattern. This manifests potentially complex problems related to pragmalinguistic and sociopragmatic awareness.
As for the directness of the analysis, the results appeared to be dissimilar in the two situations. In the situation with the professor, the JJ and US groups' degrees of directness seem to reflect similar socio-cultural norms. The JE group showed a significantly higher level of directness than the other two groups. Therefore, we can say that Japanese socio-cultural norms when complaining to a professor are not reflected in JEFL learners' production in the target language, meaning that there is an underlying pragmalinguistic problem. This interpretation is supported by Nakabachi's (1996) investigation of Japanese and English complaints by Japanese EFL learners, elicited in a situation similar to the complaints to the professor. In that case 33% of respondents also used mitigated and implicit expressions in Japanese, as opposed to unmitigated and explicit ones in English.
In contrast when complaining to a roomate, the problematic aspect regarding directness is likely to reflect differences in cross-linguistic norms. The correspondence of higher complaint directness levels across Japanese and English by Japanese students as opposed to American respondents can be explained in terms of problematic sociopragmatic conflicts (Blum-Kulka, 1982; JACET SLA SIG., 2005; Nakabachi, 1996; Olshtain, & Weinbach, 1993).

[ p. 39 ]

The last aspect of the Stage 1 analysis concerns the amount of softener use. The previous study (Rinnert & Iwai, 2003) reported less use of softeners by JEFLs. The present study showed that for the situation with the professor, the JJ group used fewer softeners than both the other groups. On the other hand, for the situation with the roommate, the usage of softeners was similar to the US group quantitatively. However, this is because Japanese tends to use post-positional particles such as -yone, and -desyo for mitigation, not unlike the English modals such as could and would. However, theese Japanese mitagators appear in friendly and casual talk rather than formal settings (Nakabachi, 1996). On such occasions, honorific expressions (often at sentence end) are used for mitigation. Japanese honorific language has a more systematic character than softeners; thus, honorifics were not counted as softeners in this study. Consequently, this led to identification of less softener use by the JJ group in the situation with the professor. By taking into consideration the characteristics of Japanese mitigation described above, JEFLs' reduced use of English softeners can be explained as a pragmalinguistic problem because Japanese softeners, especially with respect to particles, are difficult to translate into English. Thus, it is highly likely that softener usage is not easily transferred from Japanese socio-cultural practice to English pragmatic use, even though many research studies on pragmatics have shown the high possibility of a pragmatic transfer (e.g. Blum-Kulka, 1982; Takahashi, 1996).
The rather abrupt English language complaint strategies adopted by many JEFL learners may defy simple interpretations, and require a multifaceted analysis touching upon sociopragmatic and pragmalinguistic concerns.
The second research goal was to determine which complaint strategies had been acquired by the JEFLs in the second stage of this study. To this end, the same three aspects analyzed in the first stage were tested, using an experimental design, and the degree of acquisition of awareness of these strategies by the learners in this study was inferred by comparing their answers with those of the NFSs.
With regard to component choice, the resultss appear to match what was found in the earlier study (Rinnert & Iwai, 2003) and Stage 1 of this study.3 In particular, when complaining to a professor, the JEFLs, like the NFSs and many of the Stage 1 participants, preferred a three-part (I + C + R) component formulation. Similarly, the frequent use of the I + C pattern by the US students in Stage 1 was reflected in the higher preference for complaints alone among the NFS group in Stage 2, whereas the learners showed a greater preference for requests alone. However, the learners' high regard for requests (I + R pattern) in this stage does not match the performance of any of the groups in Stage 1. Uncovering the reasons for such a positive assessment of requests would seem worthwhile, particularly because the low assessment of requests alone by the NFS group suggests that it could lead to potential problems for the learners who might use them. In contrast, for the situation with the roomate, the learners' clearly prefered the I + C + R pattern (the preferred one for both groups of native speakers, but not the learners in Stage 1) suggests that these intermediate learners have acquired an awareness of the merits of using this pattern when complaining to a friend.
Concerning directness, the second aspect tested in Stage 2 of this study, the results from the complaints to the professor suggest that these JEFL learners had acquired pragmalinguistic knowledge about the expression of indirectness in English complaints to a higher status interlocutor. That is, their positive judgments of indirect complaints in this situation corresponded to the frequent use of indirectness in the Japanese and US English complaints, and contrasted with the more frequent use of directness by the JEFLs in the preceding stage. At the same time, the learners' relatively positive assessment of direct requests contrasted strikingly with those of the native/fluent group, indicating that problems could arise from the use of directness in request components as part of complaints to a person of higher status.

[ p. 40 ]

Turning to the F-situation, the results of this experimental study appear to contradict the earlier results, in that both groups chose directness over indirectness when complaining to a friend, whereas the Americans in Stage 1 used significantly more indirect complaints. However, this apparent lack of correspondence between production and perceptions among native English speakers probably stems from the fact that Stage 1 measured the directness of only complaint components, as opposed to a combination of complaint and request components in Stage 2. Also, 3 levels of directness were identified in the first stage, whereas only 2 levels were included in the second stage. Nevertheless, further investigation is necessary to determine the most appropriate/effective levels of directness for English complaints to a friend.
Finally, the third complaint strategy examined was mitigation. The findings from the P-situation indicate recognition by these learners of the appropriateness and effectiveness of mitigating complaints to a professor. Thus, even though the JEFLs in Stage 1 were unable to produce much mitigation in their complaints in this situation, the learners in Stage 2 were clearly aware of the benefits of using multiple softeners to someone of higher status. On the other hand, the F-situation results show no recognition by the learners of the importance of softening an English complaint to a friend, even though multiple softening was found in the Japanese complaints in this situation in Stage 1. This result matches that in the previous stage, where the learners produced little or no softening in their English complaints. However, a question remains as to why the learners in the second stage preferred softening in the P-situation but not in the F-situation. It is likely that the answer relates to Japanese socio-cultural norms of politeness that require deference to be shown to higher status interlocutors (Niyekawa, 1999), which would make it easier to transfer awareness of the need for mitigation in this situation, as opposed to a situation with a same-status interlocutor.
Overall, the findings from Stage 2 suggest that receptive pragmatic competence precedes productive competence. That is, an awareness of the appropriateness and effectiveness of such strategies as using particular component choices, indirectness and mitigation appears to come earlier than the ability to use these strategies in producing complaints, even when learners are given the time to reflect on their strategic choices, as they were in Stage 1.


This two-stage study has investigated the production and evaluation of complaints by Japanese EFL learners in their L1 and L2 and those by native/fluent English speakers in order to discover what English complaint strategies are preferred by these learners, and possibly why. The findings suggest that a complex combination of linguistic, pragmatic and sociopragmatic factors affect learners' knowledge of appropriate and effective ways to complain. Thus, in order to teach appropriate ways to perform intricate face-threatening acts such as complaints, English teachers need to raise their own awareness of the complexity of the factors involved. Once teachers and researchers have established a relatively clear picture of acceptable norms across languages, as well as the current level of their students' pragmatic knowledge, they should be able to develop effective teaching methods (cf. Occhi 2006) to empower their EFL students to complain appropriately and effectively.

1. This research was undertaken as part of the Prag-PEACE Project (Pragmatic Paradise of Easy Access for Communicative English) at Hiroshima City University. Prag-PEACE sites can be accessed at or

2. The two groups were combined due to the small respondent numbers.

3. Unfortunately, the component 'Initiator' was not tested in this design, as it was deemed necessary to limit the number of questionnaire items to avoid the fatigue factor. It was assumed that learners at an intermediate level of English proficiency would recognize that an Initiator would raise the levels of appropriateness and effectiveness of most English complaints, but this assumption requires empirical confirmation.

[ p. 41 ]


Blum-Kulka, S. (1982). Learning to say what you mean in a second language: A study of the speech act performance of learners of Hebrew as a second language. Applied Linguistics, 3, 29-59.

Brown, P., & Levinson, S. C. (1987). Politeness: Some universals in language usage. Cambridge: Cambridge University Press.

Iwai, C., & Rinnert, C. (2001). Cross-cultural comparison of strategic realization of sociopragmatic competence: Implications for learning world Englishes. Hiroshima Journal of International Studies, 7, 155-179.

JACET SLA SIG. (2005). Japanese university students' receptive pragmatic competence. JACET Bulletin, 40, 117-133.

Murphy, B., & Neu, J. (1996). My grade's too low: The speech act set of complaining. In S. M Gass & J. New (Eds.), Speech acts across cultures: Challenges to communication in a second language (pp. 191-216). New York: Mouton de Gruyter.

Nakabachi, K. (1996). Pragmatic transfer in complaints: Strategies of complaining in English and Japanese by Japanese EFL Speakers. JACET Bulletin, 27, 127-142.

Niyekawa, A. (1999). Minimal essential politeness. Tokyo: Kodansha International.

Occhi, D. (2006). Teaching pragmatics: From example to theory. Pragmatic Matters: JALT Pragmatics SIG Newsletter 7 (2), 8-9.

Olshtain, E., & Weinbach, L. (1993). Interlanguage features of the speech act of complaining. In G. Kasper & S. Blum-Kulka (Eds.), Interlanguage pragmatics (pp. 108-122). Oxford: Oxford University Press.

Rinnert C., & Iwai, C. (2003). Variation in complaint strategies in three regions. In M. Swanson, D. McMurray, K. Lane (Eds.) The proceedings of JALT 2002 in Shizuoka (pp. 372-378). Tokyo: The Japan Association for Language Teaching (CD-Rom).

Takahashi, S. (1996). Pragmatic transferability. Studies in Second Language Acquisition, 18, 189-223.

Thomas, J. (1983). Cross-cultural pragmatic failure. Applied Linguistics 4, 91-112.

Main Article Appendix 1 Appendix 2 Appendix 3 Appendix 4

2006 Pan SIG-Proceedings: Topic Index Author Index Page Index Title Index Main Index
Complete Pan SIG-Proceedings: Topic Index Author Index Page Index Title Index Main Index

[ p. 42 ]
Last Next