Politeness strategies when native Japanese speakers and Uzbek Japanese learners give warnings about prohibited acts: a cross-sectional study

1.

Introduction

This study began with the question of how the relationship between signs indicating prohibition and linguistic expressions is. We are interested not only in what linguistic expressions are used in prohibition signs, but also in how people warn others after seeing a prohibition sign. The way people warn others may differ depending on the language. For example, in a no-smoking situation, speakers of a language will clearly order people not to smoke, based on the idea that smoking goes against social rules. On the other hand, speakers of another language will issue vague warnings so as not to offend the hearer. The authors, who are Japanese language teachers, became interested in warnings based on prohibition signs in Japanese and began to investigate them. Some people may have the stereotype that Japanese people give vague warnings out of consideration for the hearer. It is necessary to clarify whether this idea is correct and, if it is correct, what form vague warnings take. We are interested not only in the speech acts of native speakers of Japanese, but also in the speech acts of learners of Japanese. The purpose of this study is to clarify pragmatic features when native speakers of Japanese and Uzbek learners of Japanese give warnings about prohibited acts. The authors investigated the classification criteria for speech acts such as prohibition and warning that previous studies have shown, and verified their validity. We did not create the classification criteria for warnings based on prohibition signs from scratch, but rather improved on the classification criteria proposed by previous studies. The classification proposed in this study was developed with the aim of analyzing Japanese data. However, due to its versatility, it can also be applied to languages other than Japanese, and thus this classification can be seen as a framework that forms the basis for more comprehensive classification criteria in warning research. Japanese and Uzbek, both of which are Asian languages, are agglutinative and have a subject-object-verb structure and similar usage of auxiliary verbs: e.g., both -te miru in Japanese and -b koʻrmoq in Uzbek mean “to try to Verb” using verbs derived from the word “to see” (Yamazaki 2017). For this reason, one might expect that Uzbeks could learn Japanese grammar relatively easily. However, even if learners are familiar with Japanese grammar and can produce grammatically correct sentences, they may not be able to successfully implement Japanese speech acts that warn others based on a prohibited sign.

There are two reasons for focusing on Uzbek learners of Japanese. First, as Japanese language education has flourished in Uzbekistan in recent years (Iwasaki and Umarova 2019: 231), we expect research on Uzbek learners of Japanese will continue to develop. In order to avoid communicative misunderstandings between Uzbek Japanese learners and native Japanese speakers, it is necessary to clarify the pragmatic characteristics of learners. Second, our 2021 survey¹ examined warning of prohibitions by native speakers of Japanese, as well as by Chinese, Korean, and Uzbek learners of Japanese, but failed to elucidate the characteristics appertaining to Uzbek learners. To address this oversight, this study raises the following research questions: 1) Which politeness strategies do native Japanese speakers and Uzbek Japanese learners use more frequently when warning of prohibitions? 2) What what pragmatic features other than the politeness strategies do the two language groups display?

2.

Methods

Data analysis

2.3

We created a classification scheme, based on request and prohibition studies comprising Searle (1976), Blum-Kulka et al. (1989), Trosborg (1995), Kishie (2008), and Bella (2012). First, we presented two head acts and one supportive move, with reference to the frameworks of Blum-Kulka et al. (1989) and Trosborg (1995). A head act is the minimal unit which can realize a warning against prohibition and is the core of the warning sequence. A supportive move is a unit external to the warning, which modifies its impact by mitigating its force. While the head act is essential, the supportive move is not obligatory. We divided the two head acts into directive and representative based on Searle’s (1976) classification of warnings. The framework of Kishie (2008) as well as Searle (1976) influenced the classification proposed in this study. The category of directive by Searle can be equivalent to the request-type of Kishie (2008). When previous research on requests has focused on politeness, indirectness has often been the subject of analysis. In typical studies such as Blum-Kulka et al. (1989) and Trosborg (1995), request strategies were classified into a scale of nine or eight. Kishie (2008) established the categories of Request I and Request II, as confirmed in the previous studies section, but these categories were not named based on indirectness. We established the following eight categories in terms of indirectness based on the research of Blum-Kulka et al. (1989) and Trosborg (1995): Level 1 imperative “Don’t take photos!”; Level 2 explicit performative “I ask you not to take photos”; Level 3 hedged performative⁵ “I would like to ask you not to take photos”; Level 4 obligation statement “You should not take photos”; Level 5 want statement “I want you not to take photos”; Level 6 suggestory formula “How about not taking photos?/Let’s not take photos”; Level 7 query preparatory “Could you not take photos?”; Level 8 hint⁶ “You can take photos in another place.” We called these categories Directive-Warning (DW) because the request-type categories can only be interpreted as directive.

As the second head act, we proposed Representative-Directive-Warning (RDW). Searle’s (1976) point that warnings have representative features as well as directive features led to the name RDW in this study. For example, the sentence “Photography is prohibited” functions as a directive in that the speaker stops the hearer from being about to take photos, but superficially it seems to be representative. Since there are sentences that have both representative and directive features, we made the RDW-type. As with the DW-type, when considering the sub-classification of the RDW-type, we wondered how many levels of indirectness should be assumed in the RDW-type. First, we referred to the explicit and hedged performatives in the DW-type. The difference between these strategies is based on whether or not a hedge marker is used. We also focused on hedge markers in the RDW-type and created the category of explicit and hedged infringement-indications⁷: hedged infringement-indication is more indirect than explicit infringement-indication. Second, when we observed the RDW-type data, we found expressions that confirm facts, such as “Isn’t it prohibited?” In the DW-type, the expression of questions is highly indirect, except for hints: Level 6 suggestory formula and Level 7 query preparatory. For this reason, we identified fact-checking as a category with a higher degree of indirectness than hedged infringement-indications. Based on the above, we created the following categories in the RDW-type based on indirectness: Level 1 explicit infringement-indication “Photography is prohibited”; Level 2 hedged infringement-indication “Looks like photography is prohibited”; Level 3 fact-checking “Isn’t it wrong to take photos?”

After examining examples of the RDW-type from a different point of view than indirectness, we found that there are characteristic differences between Jap. and UJap. in the form of the predicates. We set up four predicate forms, “Prohibited to Verb”-form, “Not possible to Verb”-form, “Wrong to Verb”-form, and “Right to Verb”-form (Verb is hereinafter abbreviated as V), and found a difference in their utilization between the Jap. and UJap. groups. Furthermore, when observing the data on hedged infringement-indications of the RDW-type, we found various hedge markers. We classified them and created the following three categories: inference from evidence, citation of a fact, sentence-final ellipsis.

With reference to Blum-Kulka et al. (1989) and Bella (2012), who conducted the research on request, we made the following classification scheme for supportive moves: (i) preparator “I have a favor to ask. (Can you stop calling here?)”; (ii) getting a precommitment “Can I have a word with you? (Looks like drinking isn’t allowed)”; (iii) grounder “There is a sign that says, “Staff only!” (You can’t enter there)”; (iv) alternative “(You can’t enter there) I’ll show you to a restroom”; (v) apology “Sorry. (Photography is prohibited here).” One can place the first and second categories not only in the context of a request, but also in the context of a warning. The third category, grounder involves external mitigation where the speaker gives reasons, explanations, or justifications for his/her warning. The fourth alternative corresponds to Blum-Kulka et al.’s (1989) “promise of reward.” In the case of a request, the speaker may give the hearer a reward, but in the case of a warning, the speaker may suggest to the hearer an alternative plan to replace the prohibited act instead of giving a reward. Regarding the fifth category, Bella (2012), who investigated linguistic mitigation in Greek requests, set this category as one of the supportive moves.

Based on the above, this study created two head acts, the DW-type (eight subcategories) and the RDW-type (three subcategories), and one supportive move (five subcategories). The RDW-type was also classified into four categories according to predicate form. Furthermore, hedged infringement-indication in the RDW-type was classified into three types based on the type of hedge marker. In the utterance of a warning against prohibited acts, either a head act of the DW or RDW-type is mandatory, but a supportive move is not mandatory. While observing the data, we found situations in which nothing was said, that is, no warning was given. Such no warning situations were labeled by the participants as “Don’t do FTA.” The three authors used this coding manual to label the data: Figure 1 is a summary of the coding procedure.

Figure 1.

Summary of the warning categories against prohibited acts in Japanese and Uzbek.

Following the above procedure, we labeled all the utterances pertaining to warning against prohibited acts as follows.

(2) Koko shashin totcha dame mitai yo.              Asoko ni hyooji ga aru.

  Hedged infr-ind (Wrong to V, Inference from evidence)      Grounder

  You’re not supposed to take pictures here.            There is a sign over there.

The example (2) above is from Situation 1-i “No photography” in Japanese. In the example, there was one hedged infringement-indication for the RDW-type, and one grounder for the supportive moves. In the RDW-type, one “Wrong to V”-form was counted. Furthermore, in the case of the hedged infringement-indication, the type of hedge marker was noted: this example showed that there was one inference from evidence.

We presented the eight scenarios ( Table 1) to three groups (36 participants each) and obtained the following data sets: Jap. 288, Uzb. 288, and UJap. 288 (total 864). These data were analyzed quantitatively. First, we counted the number of main categories and subcategories of the DW-type, RDW-type, and supportive moves. The number of data labeled as “No warning” (i.e., “Don’t do FTA”) was also recorded. The percentage of each subcategory of DW and RDW-types was divided by the total number of head acts. To calculate the utilization rate of the four predicate forms of the RDW-type, the number of each predicate form was divided by the total number of RDW-type. Furthermore, to calculate the utilization rate of the hedge markers of hedged infringement-indication, the number of each hedge marker was divided by the total number of hedged infringement-indication. To calculate the utilization of the three subcategories of supportive moves, the number of each subcategory was divided by the total number of supportive moves. In addition, we analyzed each social parameter. Since there were many apologies among the supportive moves, we counted the number of apologies in each social parameter (i.e., social distance and age). Also, we counted the number of “Don’t do FTA” for each social parameter. To obtain the utilization rate of the DW-type, RDW-type, supportive moves, and “Don’t do FTA,” the number of those four semantic formulae was divided by the total number of semantic formulae. Calculations were made using Microsoft Excel (see extended data). We conducted a statistical analysis to show the differences between the two languages by comparing data from native Japanese speakers and Uzbek learners of Japanese (i.e., t-test).

One point to note about classification was that the same expression form may be labeled differently depending on the context. For example, in a situation where “No entry for unauthorized personnel,” the following example was found in the Japanese data.

(3) a. Sochira wa kankee-sha nomi ga haireru tokorodesu. (That place is for authorized personnel only.)

   b. Soko wa kankee-sha iriguchinanode, hoka no toire o sagashimashoo!

   (That is the entrance for the authorized personnel, so let’s find another restroom!)

The above (3a) was classified as hint of the DW-type. If this is the only response, it can be interpreted as hint to prevent the hearer from entering the place. On the other hand, the underlined part of (3b) means that the place is the entrance for unauthorized personnel, and it has the same meaning as (3a). However, the underlined part is a reason for looking for another toilet, in other words, it functions as a grounder. We coded the data while considering the context.

3.

Results

Table 2 indicates the DW-type categories and examples found in the Jap. and UJap. data. Contrary to our expectations, we did not find any examples of categories other than the imperative, want statement, and hint. Also, we did not find any examples of categories other than imperative, obligation statement, suggestory formula and hint in the Uzb. data.⁸ Negative imperative forms that would stop the hearer’s action were classified as imperative. The desiderative form in which the speaker wishes the hearer to stop an action, were classified as want statements.⁹ Vague warning without a specific form was classified as hint. We can say that these three languages lack diversity in relation to the DW-type.

Table 3 shows the categories of the RDW-type. The most direct category in this RDW-type is explicit infringement-indication, which is an expression with a bare predicate form of prohibition, such as “kinshi da” and “dame.” Hedged infringement-indication is a category that is one step more indirect than explicit infringement-indication. This is an expression in which a hedge is added to explicit infringement-indication (For details of hedge markers, see Table 5). The most indirect category in RDW-type is fact-checking, which asks the hearer whether his or her action is prohibited. The sentences in which explicit infringement-indication was negativized and interrogativized were classified into this category. This study shows that Jap. had examples of this but UJap. did not.

As shown in Table 4, we divided the three RDW-type categories into four in terms of predicate form. The first “Prohibited to V”-form means that the hearer’s action is not permitted according to a rule. The second “Not possible to V”-form has the meaning that the hearer cannot perform the action. Using the third “Wrong to V”-form and the fourth “Right not to V”-form, the speaker communicates that the hearer’s act is bad and not good based on the speaker’s value judgment.¹⁰ If the speaker warns by using the second, third and fourth forms without adding an expression like “It’s a rule” or “There’s a no photography sign over there,” the basis for their warning is not due to a rule, but a value judgment. However, if the speaker uses the first form, they can warn the hearer on the basis of the rule, even without giving a reason for the warning. Concerning the third “Wrong to V”-form, we set two forms: dame and ikenai. As described below, we found differences in the utilization of these two forms between the Jap. and UJap. groups.

Table 5 shows the three categories of hedge markers in hedged infringement-indication: Inference from evidence, citation of a fact, and sentence-final ellipsis. The category of inferences from evidence is a way of expressing inferences based on a situation, such as “it looks like”, “it’s like”, and “I think”, and in both Japanese and Uzbek, this hedged marker is placed after the predicate. Detailed observation revealed that there were six forms in Japanese and four forms in UJap. in this category. The category of citation of a fact is an expression that cites a prohibited fact. Like inference from evidence, this category is also placed after the predicate in both Japanese and Uzbek. The last category is sentence-final ellipsis, which is unique to Japanese grammar. According to Iwasaki (2006: 262), “kedo” and “ga” are used to express a contrary conjunction such as although, but, and were originally only used as conjunctions that connect to the predicate on the left sentence. In modern Japanese, he pointed out that these conjunctions appear even when there is no main clause and function as pragmatic particles that weaken the force of the assertion. In this study, not only the conjunctions “kedo” and “ga,” which indicate a contrary conjunction, but also the conjunction “node,” which expresses a reason such as because were observed as sentence-final ellipsis in Jap. and UJap.

Table 6 shows the numbers and percentages for the DW and RDW-types. Figure 2 shows the percentages in graphical form.

Figure 2.

Percentages of head acts (Directive-Warning and Representative-Directive-Warning types)’ categories.

The following shows the results of the t-test for the DW and RDW-types in the Jap. and UJap. groups.

The results of the statistical analysis showed that there were statistically significant differences in the use of imperative, explicit infringement-indication, and hedged infringement-indication. From the above, it was clear that the UJap. group used imperative and explicit infringement-indication more frequently than the Jap. group, while the Jap. group used hedged infringement-indication more frequently than the UJap. group.

Table 7 shows the numbers and percentages of the four predicate forms in RDW. Figure 3 shows the percentages in graphical form.

Figure 3.

Percentages of Representative-Directive-Warning type’s predicate forms.

The following shows the results of the t-test for the predicate forms in the Jap. and UJap. groups.

The Jap. group used the “Prohibited to V”-form the most (about 70%), the Uzb. group used the “Not possible to V”-form the most (about 50%), and the UJap. group used the “Prohibited to V” and “Wrong to V”-forms to the same degree (30-40%). Statistical analysis showed significant differences between the Jap. and UJap. groups in the use of the “Prohibited to V,” “Wrong to V,” and “Right not to V”-forms. The Jap. group used the “Prohibited to V”-form more than the UJap. group, but the UJap. group used the “Wrong to V” and “Right not to V”-forms more than the Jap. group.

When the “Wrong to V”-form was subdivided, the forms “dame” and “ikenai” were identified. The number and percentage of these two forms are shown in Table 8.

Table 9 shows the numbers and percentages of hedge markers in hedged infringement-indication. Figure 4 shows the percentages in graphical form.

Figure 4.

Percentages of hedge markers in Hedged infringement-indication.

The following shows the results of the t-test for hedge markers in the Jap. and UJap. groups. A statistically significant difference was found only in the category of inference from evidence.

From the statistical analysis, a significant difference was found between the Jap. and UJap. groups in the category of inference from evidence. The Jap. group used inference from evidence more than the UJap. group.

When subdividing the hedge markers for inference from evidence based on the categories in Table 5, we obtained the following results ( Table 10).

The table above shows that the Jap. group used several hedge markers, centering on “mitai” and “rashii”, but that the UJap. group used the two categories of “mitai” and “to omou” exclusively.

As Table 11 shows, the Jap. and UJap. groups did not produce any examples of “preparator” and “getting a pre-commitment” in the supportive moves. The Uzb. participants also gave no examples of these categories. Table 12 shows the numbers and percentage of supportive moves used, and Figure 5 shows the percentages.

Figure 5.

Percentages of supportive moves’ categories.

The results of the statistical analysis of the supportive moves showed that the only significant difference between the Jap. and UJap. groups was in the use of apologies. The UJap. group used apologies more than the Jap. group. As in the examples of apology in Table 11, the DCT data revealed that apology¹¹ almost always preceded a head act.

Furthermore, Figure 6 shows the number of apologies by social parameter. According to the results, it was clear that while Jap. participants apologized only to people they did not know (+SD), whereas UJap. participants apologized not only to people they did not know (+SD), but also to people they knew well (-SD).

Figure 6.

Number of apologies by social parameter.

Participants noted that they do not warn the hearer in some situations. We categorized the case as “Don’t do FTA.” Table 13 indicates the number and percentage of DW, RDW, supportive moves, and Don’t do FTA.

The results of the t-test for the four categories above in Jap. and UJap. are shown below. Of these, only supportive moves showed a significant difference. As a result, the UJap. group used supportive moves more than the native groups, about twice as often.

Figure 7 shows the sub-classification of “Don’t do FTA” by social parameter. This figure revealed that all the groups implemented the “Don’t do FTA” strategy for +SD, namely unknowns.

Figure 7.

Number of Don’t do “face-threatening act (FTA)” by social parameter.

4.

Discussion

One of the significant findings from this study was that Uzbek learners of Japanese tended to be verbose. Analysis of the data in Figure 4 and Table 13 revealed that the UJap. group used more supportive moves, especially apology, than the Jap. and Uzb. groups, although the native groups avoided verbose utterances by providing only the necessary information. The verbosity of the learner group would not have been influenced by their native language, but by the characteristics of the learner language. If the result showed that the Uzb. group frequently used supportive moves, or if there were previous studies showing that Uzbeks tend to speak in a redundant way, there would have been a possibility of a cultural influence of the native language. However, the data from this study showed that the influence of the learners’ language was stronger than the influence of Uzbek and its culture. The influence of the learner’s language can be derived from the similarity of the data in Blum-Kulka and Olshtain (1986), which investigated the redundancy of the learners’ language. Blum-Kulka and Olshtain (1986: 172) showed the ratio of supportive moves with and without native speakers and non-native speakers as follows ( Table 14).

Meanwhile, the results of the ratio of supportive moves in this study were as follows ( Table 15). Cases where no warning was issued were labeled as “Don’t do FTA,” and the number and ratio of these cases were shown.

The following is a comparison between Blum-Kulka and Olshtain (1986) and the present study. According to Blum-Kulka and Olshtain (1986), non-native speakers used supportive moves nearly 15% more often than native speakers (native speakers 40.0%; non-native speakers 55.5%). On the other hand, the non-native UJap. group in this study used the supportive moves about 20.0% more frequently than the native Jap. and Uzb. groups (Jap. 18.4%; Uzb. 17.0%; UJap. 38.1%). This study clearly showed a tendency for more learners to use supportive moves than Blum-Kulka and Olshtain (1986). Blum-Kulka and Olshtain (1986: 177) asserted that lengthening of speech act patterns and addition of supportive moves in languages spoken by learners, especially advanced learners, are due to the learners’ lack of confidence. They want to get their message across but are not confident in their speaking ability. Blum-Kulka and Olshtain (1986: 173) said that grounder, which provides reasons and explanations to justify the need to make a request, was the most common in the supportive moves that learners used in request. Ito (2002) reported that lengthening of speech act patterns (especially reasons for refusal) occurred in refusal expressions in Japanese made by groups of native Malay speakers. UJap. subjects tended to be at the intermediate level of Japanese language proficiency (N3-N2 levels), but as Blum-Kulka and Olshtain (1986: 177) previously noted, the subjects exhibited “lengthening of speech act patterns” found in advanced level learners.

Unlike the previous studies of request and refusal, this study showed that apology was more common than grounder. In the scenario where the speaker warned about a prohibition, many Uzbek learners of Japanese may have assumed that the speaker does not need to actively explain the reason for the warning because the situation pertaining to the warning is clear and that the speaker expects the hearer to know that one is not allowed to take pictures in certain areas of a museum or use their cell phone in certain areas of a hospital. For this reason, UJap. participants probably implemented the negative politeness strategy of apology to express their feeling of being sorry for having to issue a warning rather than to give reasons for the warning. Also, Figure 5 shows, UJap. participants apologized to both strangers and those who they knew, while Jap. participants apologized only to strangers. The more extensive use of apology in UJap. than in Jap. means that Uzbek learners of Japanese may apologize more readily than Japanese. There may be a stereotype that learners speak bluntly to their hearers, but they used a lot of the politeness strategy called apology, which is not often used by Japanese people. Then, we consider why this learners’ group used apology so much. The possibility of the influence of Uzbek culture is low. Surveying the results of cultural studies on Uzbeks, it is possible that Uzbeks apologize frequently (Turdonova 2024; Abduvahobovna and Ibragimovna 2024). However, Figure 4 shows that there were few instances of apology in the Uzb. group. If apology was used frequently in Uzb. as well as in UJap., it would be possible to say that it was influenced by Uzbek culture. This phenomenon of frequent apologies can be explained by the characteristics of the learner language. It was reported that in refusal situations, Japanese learners whose native language is Thai tended to apologize more than native Japanese speakers (Roungtheera 2005). From the above, it can be said that Japanese learners generally tend to apologize. Because the expression of apology is convenient, learners may use it frequently. The UJap. group’s apology is a politeness strategy that can be easily done by simply placing the word “Sumimasen” before the warning. Unlike reasons, this strategy does not need to be explained in accordance with the situation. Furthermore, by making use of the stereotype that Japanese people apologize frequently, UJap. may have frequently used the apology strategy. According to Takahashi (2012), this characteristic of Japanese people to apologize often is in contrast to the characteristic of Chinese people who do not admit their own mistakes, and is a cause of cultural friction between Japan and China. In situations where Uzbek learners of Japanese warn Japanese people based on a prohibition sign, it may be better not to apologize too much.

An analysis of the head acts, the core of warning, revealed that Jap. warning was indirect. Their warning was based on the rules of prohibition, and it also showed their own speculation that the existence of the rules is uncertain. On the other hand, UJap. warning was direct and that it was based on the speaker’s value judgment. First, regarding indirectness, see the RDW-type’s categories in Figure 2. The Jap. group used hedged infringement-indication frequently, while the UJap. group used a lot of explicit infringement-indication, which was more direct. While the Jap. group in Figure 3 used “Prohibited to V”-form frequently, the UJap. group used both “Prohibited to V” and “Wrong to V”-forms frequently. From this fact, it can be seen that while Jap. tended to give warnings based on rules, UJap. tended to give warnings based on the speaker’s own judgment. While Jap. transferred the responsibility for warning to the museum or hospital, about half of the UJap. participants themselves took on the responsibility that originally belonged to those institutions and conveyed the opinion that the hearer’s action was wrong. Furthermore, from the results of the frequent use of hedge markers such as “inference from evidence” in the Jap. group in Figure 4, it was found that Jap. thought that the existence of the prohibition rule is uncertain.

From the results of the analysis of the head acts above, we clarified the characteristics of the Jap. and UJap. groups. We wondered why the UJap. group used “Wrong to V”-form so frequently. Since the Uzb. group used the “Not Possible to V”-form most frequently and hardly used the “Wrong to V”-form at all, it is likely that the reason why the UJap. participants used the “Wrong to V”-form so frequently was due to the influence of the learner language, not the influence of their native language. If native language or culture had an effect, we would have seen the “Wrong to V”-form frequently in the data of Uzb. group’s data as well. The UJap. group may have used the “Wrong to V”-form frequently due to the influence of the learner language, particularly the textbook. The influence of textbooks on learners has been studied in the past. Certainly, there are studies that do not emphasize the influence of textbooks, such as Yamada (2004: 150), but some studies pointed out that the influence of textbooks is strongly apparent in the initial stages of learning, such as Park (2017) and Honda (2019). Also, Honda (2019: 125) stated that “the first 1000 words that Japanese learners learn are often influenced by the textbooks they use.” Park (2017: 101) also pointed out that the actual usage of “totemo” (very) by Japanese learners differs from that of native Japanese speakers, and is closer to the usage of “totemo” in textbooks, especially beginner-level textbooks. For the “Wrong to V”-form, two subcategories were set up: “dame” and “ikenai.” As Table 8 shows, there are differences between the Jap. and UJap. groups, with Jap. using “dame” frequently and UJap. using “ikenai” frequently. The “ikenai” form is practiced in the early stages of Japanese language learning. This form appeared in Lesson 15 of Minnano Nihongo, Shokyuu (Beginner) I, a textbook that is representative of Japanese language textbooks worldwide (Three A Network 2012a: 126-129), and is within the first 1000 words that Japanese language learners learn. On the other hand, “dame” appeared in Lesson 9 of Minnano Nihongo, Shokyuu (Beginner) I, but it only appeared once and was not used as much as “ikenai” for practice. It is likely that UJap. participants used the word “ikenai” frequently because they had practiced it a lot in their textbooks.

As Figure 3 shows, UJap. used “Right not to V”-form as well as “Wrong to V”-form more than Jap. frequently. According to Takanashi (2011: 3-4), the Japanese phrase “shinai hooga ii,” which belongs to the “Right not to V”-form, is an expression that “gives the hearer a benefit or favor and gives the hearer the option of not carrying out the action.” In other words, such expressions are not just a way of advising the hearer not to suffer any disadvantage. They are also an indirect way of showing respect for the hearer’s face and conveying that the hearer does not have to follow the speaker’s advice (Brown and Levinson 1987: 172). The result that UJap. used this form frequently may suggest that in Uzbek society, when giving advice, there is a culture that respects the hearer’s face and gives the hearer options. As Figure 3 shows, the Uzb. group also had a higher usage rate of the “Right not to V”-form than the Jap. group. The high usage of the “Right not to V”-form in the UJap. group may be due to the influence of the culture of their native language.

Regarding the indirectness of Jap. and UJap., we noted above that Jap. made frequent use of hedged infringement-indication, while UJap. made frequent use of the more direct explicit infringement-indication. The fact that Jap. used hedges to speak ambiguously is supported from the perspective of cultural studies. Yang and Cao (2005: 46-47) stated that “Japanese people consciously change their speaking style from direct to indirect, and from indirect to euphemistic, and deliberately use ambiguous expressions to avoid direct expressions,” indicating that Japanese people tend to prefer ambiguous expressions. In second language acquisition research, there were reports of research showing that language learners are more direct than native speakers (Fukushima 1990; Cenoz 1995). Fukushima (1990) pointed out that offer and request by Japanese learners of English were more direct than the forms used by native English speakers. Many UJap. participants did not use hedged forms that Jap. participants used frequently. The tendency for Japanese language learners not to use hedges was also pointed out in Hotta and Horie (2012), who investigated hedges in the speech of refusals. The results of this study showed the same tendency as their research. Japanese hedges can be difficult for learners in general.

The UJap. and Jap. groups showed differences not only in the usage rate of hedges, but also in the type of hedges. As Table 10 shows, UJap. tended to use two types of hedge, “to omou” and “mitai,” within the category of inference from evidence. This result made us wonder why there were so few types of inference from evidence in UJap. One possible reason for this scarcity is the influence of the native language, Uzbek. From Table 10, we found that Uzb. only showed the word “ekan” as inference from evidence. The word “ekan” is used to express what the speaker learned from a third party or what the speaker judged or guessed based on circumstantial evidence, and is equivalent to verbal suffixes such as “mitai,” “rashii,” and “yooda” in Japanese (Shimada 2019: 182). As inference from hearsay, Uzbek only has “ekan,” but Japanese has more types. The small number of varieties of inference from evidence in Uzbek may have led to the small number of forms in the category of inference from evidence in UJap.

It is imperative to pay attention to the aforementioned statement, which pointed out that “UJap. tended to use two types of hedges, “to omou” and “mitai,” within the category of inference from evidence.” The Japanese “to omou” is translated as “I think that,” and “mitai” is translated as “looks like, it’s like, it seems.” The fact that the UJap. group used the verb “omou” (about 48% of the total) so frequently in the category of inference from evidence may be due to the influence of learner language. This theory is also supported by the statement in Hotta and Horie (2012: 15) that “learners tended to use verbs like “omou” (I think), and Nikula’s (1997) observation that learners tended to overuse “I think” as a hedge. On the other hand, the reason for the high frequency of “mitai” (48% of the total) in UJap. was unknown. As in the case of “ikenai,” we considered the possibility of the influence of the textbook, but this is unlikely. While the form “mitai” is not taught in the beginner’s level of “Minnano Nihongo,” it is taught in the second lesson of Intermediate I (Three A Network 2008). The lessons on the other forms of inference from evidence were as follows: “yooda” is taught in Lesson 47 of Beginner II, “soo” is taught in Lesson 3 of Intermediate I, “rashii” is taught in Lesson 11 of Intermediate I, and “ppoi” is taught in the third section of the Additional Section of Intermediate II (Three A Network 2008, 2012b, 2013). If we base the theory that learners tend to frequently use the items they learn at the beginning of the textbook, it was expected that because UJap. learn “yooda” at an earlier stage of beginner level, they would use it frequently in this survey, but this was not the case. We also considered the possibility of influence from their native language, Uzbek, but this is also unlikely. As we pointed out above, the Uzbek word “ekan” indicates a guess based on hearsay, and it does not correspond only to the Japanese form “mitai.”

The above is a discussion of the RDW-type of head acts. As Figure 2 shows, the DW data was less than the RDW data. Although participants of the three language groups could show indirectness in the DW-type as well as in the RDW-type, they did not use any indirect strategies of the DW-type except for hint. While request categories and DW-type categories of warning were the same, the frequency of use was different. Harting (2008: 125) showed that in request situations, more than 50% of the Japanese participants used the highly indirect query preparatory and about 20% of them used the moderately indirect want statement. However, in the warning situations, none of the groups used such indirect DW strategies. The participants may have avoided using such indirect strategies because the strategies were too euphemistic or because directive speech acts were not appropriate for the warning scenarios.

None of the three language groups used a preparator or attempted to get a precommitment in the supportive moves. In the warning scenarios, although the speaker does not have the authority to prohibit, they can serve as a spokesperson for the authority. Because of this role, the participants may have thought that there was no need to use a preparator to inform the hearer that a warning was about to be given, nor does the speaker feel the need to try to secure a pre-commitment before the warning. The participants in all the language groups would not have used such circuitous strategies because the speaker does not need to adopt a humble attitude toward the hearer in warning situations.

We found that the three language groups did not warn strangers. All the groups implemented the “Don’t do FTA” strategy towards the +SD attribute, that is, unknowns (see Figure 6). Warning was given to people with other attributes. Participants in all the groups issued warnings without hesitation not only to the hearer who were the same age as the speaker and whom the speaker knew well (-SD, x=y), but also to the hearer who were older than the speaker and whom the speaker knew well (-SD, x<y). In all the language groups, whether or not the speaker knows the hearer well was an important indicator of whether or not a warning would be given. It is possible that they did not want to threaten the face of a stranger and receive a hostile reaction from the stranger.

We summarize the characteristics of Jap. and UJap. from the perspective of politeness strategies. As mentioned in previous studies, Brown and Levinson (1987: 98) stated that the speaker will not provide compensation if it is in the hearer’s benefit to perform an FTA. They pointed out that the warning itself shows that the speaker is concerned about the hearer, so the politeness strategy does not work, and the warning may be carried out baldly. However, the following politeness strategies were actually observed in the data from this study. The Jap. groups used the strategy of “State the FTA as a general rule” by using the “Prohibited to V”-form. This is a strategy that conveys that the speaker does not want to interfere, but is only compelled to do so by rules, regulations, or obligations (Brown and Levinson 1987, negative politeness strategy 8). With the “Prohibited to V”-form, many Jap. participants did not want to warn, but may have wanted to imply that they had to warn due to the rules of the museum or hospital. Although the use of the “Prohibited to V”-form has the advantage of allowing the speaker to transfer the responsibility for the warning to an authority, there is a danger that the expression alone may lead the hearer to feel that the speaker is an overbearing person who enjoys wielding authority. To avoid this risk, the Jap. group also frequently implemented the second politeness strategy of hedge (Brown and Levinson 1987, negative politeness strategy 2). Of the hedges, the Japanese used inferences from evidence frequently. This is the hedge that Brown and Levinson (1987: 164-165) explained as “the speaker does not take full responsibility for the truth of his utterance.” By using this hedge, the Jap. participants tried to escape from responsibility by conveying that the fact of the prohibition was only their own prediction. In contrast, the UJap. participants used the “Wrong to V”-form to take on the responsibility that the institution should have originally taken on, and to express his or her opinion that the hearer’s action was wrong. The UJap. probably thought that the hearer would be offended by the situation and added the apology (Brown and Levinson 1987, negative politeness strategy 6). The UJap. group also used the “Right not to V”-form, a strategy of indirectness that gives the hearer the option of whether or not to do the action (Brown and Levinson 1987, negative politeness strategy 1). From the above, we could see that there were differences in the politeness strategies used in warning between Jap. and UJap. If Uzbek learners of Japanese were to find employment in Japan and have to communicate with Japanese people, they would probably need to be careful about the use of apologies. They may be apologizing unnecessarily in situations where Japanese people would not.

Above, we compared data from native speaker and non-native speaker groups. Finally, as a supplement, we contrast the native languages of Japanese and Uzbek. The results of this contrast supplement the explanation of the characteristics of UJap. shown above. Firstly, the main similarities between Japanese and Uzbek are that RDW-type was used more than DW-type, and that there was a certain number of people who did not give strangers any warning. On the other hand, the main difference occurred in the RDW-type. The Jap. group used “Prohibited to Verb”-form frequently, while the Uzb. group used “Not possible to Verb” frequently. In other words, while Jap. focused on the fact of prohibition, Uzb. focused on the possibility of prohibition. Furthermore, when observing the hedge markers of Jap. and Uzb. in the RDW-type, it was found that the number of types of inference from evidence differed between the two languages. The Jap. group showed various forms of the hedge markers in the category of inference from evidence. In Jap., six types were used: “mitai,” “rashii,” “ppoi,” “soo,” “yooda,” and “to omou,” but in Uzb., only one type was used: “ekan” ( Table 5). Based on the results of this contrastive analysis, we can predict that it would be difficult for Uzbeks to understand and learn the characteristics of Japanese hedge markers that they cannot use in Uzbek. In fact, the hedge markers used in UJap. were mainly two types, so it can be seen that UJap. does not fully master Japanese hedge markers. This contrastive analysis supplemented the feature of UJap.

5.

Conclusions

Referring to prohibition and request studies, this study created a taxonomy of warnings against prohibited acts and compared the speech act used by native Japanese speakers and Uzbek Japanese learners. The two language groups used the RDW-type more than the DW-type. Among the RDW-type, the Japanese participants used indirect expressions (i.e., hedged infringement-indication) more frequently than the Uzbek learners of Japanese. By using the “Prohibited to V”-form, the Japanese participants conducted a politeness strategy of stating warning as a rule. However, simply using the “Prohibited to V”-form runs the risk that the hearer feels that the speaker is an overbearing person who enjoys wielding authority. To avoid this risk, the Japanese participants used the hedge’s politeness strategy, especially the inference based on the fact of prohibition (e.g., “Looks like photography is wrong”), to make it clear that the right to prohibit rests with the public authority, not with themselves. On the other hand, the Uzbek learners of Japanese used more direct expressions (i.e., explicit infringement-indication) than the Japanese. Previous studies on second language acquisition indicated that learners tend to speak more directly than native speakers. The Uzbek Japanese learners tended to apologize more than the native Japanese speakers. Because of frequently using explicit infringement-indication, it seems as if they gave a blunt warning. To compensate for doing the FTA, the learners implemented the politeness strategy of apology. The Japanese participants also made apologies, but only when the hearer was a stranger. In contrast, Uzbek learners of Japanese made apologies regardless of whether the hearer was a stranger or not. From the above, we can see that the native Japanese speakers and Uzbek Japanese learners used different politeness strategies. In this study, there were participants who not only did not use politeness strategies, but also did not give warnings. Whether or not the speaker gives warnings depends on the social parameters of the hearer. The native Japanese speakers and Uzbek Japanese learners only used the “Don’t do FTA” strategy only with strangers. Whether or not the hearer was an acquaintance was an important indicator of whether or not to give warnings, whether the hearer was a native Japanese speaker or a learner of Japanese.

Based on the results of this research, we can make the following suggestions to Uzbek learners of Japanese. Don’t do FTA to strangers means that there are cases where it is better not to warn strangers. For learners with a high level of Japanese ability, we recommend that they try to give warnings using general rules such as “kinshi desu” (It’s prohibited) and using hedges that Japanese people frequently use. If they say “kinshi desu” without using hedges, there is a possibility of FTA and interpersonal friction. In the case of learners whose Japanese language skills are insufficient, it would be best not to give any warnings. If such inexperienced learners carelessly warn a stranger, they may end up threatening the face of the stranger and getting themselves into trouble. Uzbek learners’ overuse of apology means that they may apologize excessively in situations where the Japanese do not. Uzbek learners of Japanese do not need to apologize when warning Japanese people. Although the learners may be a stereotype that Japanese people apologize, this is a different situation to a warning.

We will point out the possibilities and issues for future research. The present study created a classification scheme for warning against prohibitions, based on the Japanese and Uzbek Japanese data, and referring to the Uzbek data. In the future, we will use this framework to analyze this speech act of learners of Japanese other than Uzbeks and native speakers of Asian languages other than Japanese. By examining data from a large number of languages, we can develop a more general classification scheme. As mentioned above, this study found cases where participants of all the groups did not warn. Each of the DCT’s questionnaire statements did not ask whether the warning was given in the first place. In the DCT, if we ask not only “What do you say to the hearer to warn him/her?” but also “Do you warn the hearer in this situation?,” the majority of the participants may not warn unknowns. Some participants gave reasons for not warning in certain situations. Because the number of such comments was small, we did not analyze them qualitatively in this study. In the future, we would like to create a questionnaire that asks why participants choose not to give warnings when they do not, and perform a qualitative analysis to clarify the cultural backgrounds of Japanese and Uzbek people. Since we did not collect data with a view to conducting a qualitative analysis in this study, qualitative analysis is a task for the future. In response to the need to expand the study of Uzbek Japanese, we conducted a pragmatic study focusing on Uzbek learners of Japanese as well as native Japanese speakers. In order to survey pragmatic competence, it is important to understand not only the speech act but also the social and cultural context (Bachman 1990: 42). In the past, when it was difficult for Uzbeks to go to Japan, their motivation for learning Japanese was to learn about Japanese culture and to get good marks in the Japanese Language Proficiency Test. In such situations, the learners may not have thought about being able to use the language appropriately in Japanese society. However, since 2020, there has been an increase in the number of Uzbeks studying Japanese in order to get a job in Japan (Japan Foundation 2022). Obtaining a work visa for Japan and working for a Japanese company has become a motivation for their studies. There will probably be more learners and teachers who think about being able to use the appropriate language in work situations in Japan, such as avoiding friction with Japanese businesspeople. It will also become important for Japanese language teachers in Uzbekistan who teach Japanese to their students to provide not only Japanese language knowledge to help them pass the Japanese Language Proficiency Test, but also lessons that convey appropriate language use for Japanese people. Research that supports Japanese language teachers in Uzbekistan is needed. In addition, it is also important to study the acquisition of Japanese by Uzbek learners of Japanese from the perspective of acquisition research. The acquisition of speech acts involves going through several stages in order for speech acts to be performed appropriately. In order to clarify the developmental process of learners’ acquisition of warning, it would be effective to divide learners into levels. In this study, we gathered a group of intermediate learners with an average level of N3-N2. In the future, we will need to increase the number of participants from Uzbek learners of Japanese and investigate the warning according to the level of Japanese ability. Furthermore, in order to observe the differences in the acquisition situation of Uzbek learners of Japanese and the acquisition situation of other foreign learners of Japanese, it would be effective to use a large-scale database called the “International Corpus of Japanese as a Second Language.” This corpus can show results by controlling several conditions such as native language, learning environment, and developmental level. We will be able to compare the Japanese of Uzbeks with the Japanese of other foreigners at each stage of development. One possible direction for research would be a contrastive study of Japanese and Uzbek. By contrasting not only warnings but also other speech acts, it would be possible to clarify the cultural backgrounds of Japanese and Uzbek people. Furthermore, the results of this research can also be applied to the creation of Japanese language textbooks. In Japanese language education, the creation of textbooks based on the “Can do” approach, such as the “Marugoto” series published by the Japan Foundation, has already begun. Using the classification scheme and survey results from this research, we would like to contribute to the creation of Japanese language teaching materials based on the “Can do” approach for warning. In addition, there are Japanese language textbooks that aim to teach speech acts such as requests and apologies, such as Shimizu (2013). With the aim of creating this textbook, we plan to write Japanese language teaching materials related to speech acts that include warnings.

Ethics and consent information

Retrospective ethical approval for this study on the 01/05/2023.

All participants provided implicit consent by completing the online questionnaire.