Concentric: Studies in Linguistics 42.1 (May 2016): 73-101 DOI: 10.6241/concentric.ling.42.1.04 Apologizing in Mandarin Chinese: A Study on Developmental Patterns* Yuh-Fang Chang National Chung Hsing University This study of the pragmatic development of the speech act of apology compares perception and production data from Mandarin-speaking participants in different age groups. Using a cross-sectional approach, subjects from 4 age groups (3rd grade, 6th grade, 10th grade and college freshmen) were selected to represent children, early adolescents, adolescents, and young adults groups. Data were collected using the perception assessment task and the discourse completion task. Differences across age groups were statistically significant in the perception of the obligation to apologize and the likelihood of apology acceptance but not statistically significant in the perception of the severity of the offense indicating the ability to judge offense severity develops earlier than the ability to assess what action to take after the offense and how people will respond to the apology. Developmental changes in apology production were also found through quantitative analysis of production data. Key words: pragmatic development, speech act of apology, Mandarin 1. Introduction The research literature on language development now boasts a large number of studies providing a rich description of children’s development in phonetic and morphemic, semantic, and syntactic knowledge. Nevertheless, research on growth in their capacity to “do things with words” (i.e., pragmatic development), though fruitful and continuing to flourish, has progressed at a much slower pace. While there has been substantial study of L1 pragmatic development, many aspects of pragmatic development remain under-researched. The development of pragmatic competence in the speech act of apology is one such area. Apologies have the power to restore a damaged relationship, mitigate loss of face, and stimulate forgiveness (Lazare 2004). To apologize appropriately requires that one first recognize having done something wrong, something that calls for an apology. Then, it is necessary to assess the severity of the offense and the weight of contextual variables, such as power and distance, in order to select an appropriate apology strategy. Appropriate performance of the speech act of apology requires the development of cognitive ability, social understanding, and linguistic skills. Due to the complexity of performing the speech act of apology, one would expect children of * This research was funded by the Ministry of Science and Technology (MOST 101-2410-H-005-044). The author would like to thank the anonymous reviewers for their valuable comments on an earlier version of this article. 73 42.1 (May 2016) different ages to understand and use apologies differently. However, very little is known about how development in this area proceeds. Most apology studies to date have examined the apologetic behavior of adults, focusing on cross-cultural comparisons or second-language learning issues (e.g., Blum-Kulka & Olshtain 1984; Olshtain 1989; Holmes 1989, 1990, Robin 1992, Sugimoto 1995, 1997, Hussein & Hammouri 1998, Suszczynska 1999, Tsai 2000, Chang 2004, Bataineh & Bataineh 2008, Chen 2008, Kim 2008, Liu 2008, Kamph & Blum-Kulka 2009). There have been a few studies exploring the development of the speech act of L1 apology, with the majority observing preschoolers or elementary school children. However, ways in which apologies by adolescents and young adults differ from apologies by children has remained unexplored. In addition, a review of the literature on pragmatic development reveals that Mandarin-speaking participants are a relatively under-researched speaker group. The limited age range covered in the existing literature and the paucity of developmental apology studies involving Mandarin-speaking participants leaves an incomplete picture of the developmental pattern of apology. This study attempts to fill gaps in current knowledge of the development of the speech act of apology by comparing percep- tion and production data from Mandarin-speaking participants across different age groups. 2. Literature review To apologize appropriately requires that one recognize having done something wrong that calls for an apology. Researchers have pointed out that the concept of right and wrong changes as children develop. Piaget (1932) distinguished two stages of moral development, the first stage being heteronomous morality (seen from 4 to 7 years of age) in which one judges behavior by considering its consequences rather than the intention of the actor. Children at this stage believe that rules are set by authorities and cannot be altered. The second stage is autonomous morality (from 10 years of age and older) and, according to Piaget, children at this stage are able to judge an action by considering both its consequences and the actor’s intention. They are aware that rules are created by people and subject to change by consensus. Kohlberg (1976, 1986) distinguished six stages of moral development, grouped into three levels: pre-conventional (under the age of 9), conventional (age 10-20), and post-conventional (over 20). The first stage is punishment and obedience, in which children’s moral reasoning is based on punishment. Authorities/adults define right and wrong. The second stage is individualism and purpose, in which moral reasoning is 74 Chang: Apologizing in Mandarin Chinese based on rewards and self-interest. The third stage, beginning the conventional level in Kohlberg’s theory, is interpersonal norms. At this stage, the individual often adopts his/her parents’ standards, emphasizing good-person stereotypes and concerns for approval. The fourth stage is social system morality, in which moral reasoning is based on an understanding of social order and justice. The fifth stage, beginning the post-conventional level, is community rights versus individual rights. At this stage, the individual understands that standards may differ from one person to another. The sixth and highest stage is universal ethical principles, in which the individual develops a moral standard based on universal human rights. In addition to changes that children experience in thoughts about right and wrong, their cognition also develops rapidly. Piaget identified four stages of cognitive development: sensorimotor stage (birth to 2), preoperational stage (ages 2 to 7), concrete operational stage (ages 7 to 11), and formal operational stage (ages 11 to 16). In the sensorimotor stage, infants construct an understanding of the world through the coordination of sensations with physical movement. In the preoperational stage, egocentric thinking is a salient feature. Children at this stage are able to use symbols (words, images, and drawings) to represent the world. In the concrete operational stage, individuals are capable of logical reasoning and can perform manual tasks but lack the capacity for abstraction. In the formal operational stage, individuals become capable of thinking in the abstract. The development of cognitive ability, social understanding, and linguistic skills influences the ability to perform the speech act of apology. It would be reasonable to expect that individuals at different ages understand and use apologies differently. Darby & Schlenker (1982) examined the responses of children of different ages (K/1st graders, 4th graders, and 7th graders) to four variations of apology: 1) no apology, 2) a perfunctory apology (“Excuse me”), 3) a standard apology (“I’m sorry, I feel badly about this”), 4) an elaborated apology (“I’m sorry, I feel badly about this. Please let me help you”). They found that all three groups gave the elaborated apology more favorable ratings for forgiveness, liking, blame, and punishment. However, the older age groups were more sensitive than the youngest age group (K/1st graders) to the different levels of apology. Ohbuchi & Sato (1994) investigated the perceptions of Japanese children (2nd graders and 5th graders) regarding three types of responses made by transgressors in hypothetical situations. The three accounts that transgressors offered for their actions included: 1) apology condition: saying “Sure, I did it. I am very sorry. I feel bad now. Please, forgive me!” 2) excuse condition: saying “Sure, I did it, but I was very upset at that time because of the defeat,” 3) no account condition: saying “I did it.” The results showed that the older children (5th graders) evaluated the transgressors who apologized as less negative, less intentional, and more 75 42.1 (May 2016) remorseful than those who made excuses or gave no accounts for their behavior. The second graders were insensitive to the differences between the accounts. Smith (2009) examined the understanding of apology and emotion in children (age 3-9). The findings revealed that even preschool-age children understood that apologies can make a victim feel better. In addition, the children preferred a genuine apology to a non-genuine apology. Kochanska, Casey & Fukumoto (1995) found that toddlers were able to use apology to mitigate offense when they were led to believe they had caused damage to toys and other objects, whereas they did not feel responsible when a damaged object was simply presented to them. Ljungberg et al. (2005) examined post-conflict behavior among preschool-age boys in Sweden and found that apologizing was rare compared to other types of affiliative post-conflict behavior, such as an invitation to play. In contrast, examining post-conflict behavior among preschool-age children in Japan, Fujisawa, Kutsukake & Hasegawa (2005) found apologizing was the most common form of post-conflict behavior. Ely & Gleason (2006) examined the developmental pattern of apology across the age period 1;2 to 6;1 by analyzing the corpora from the Child Language Data Exchange System (CHILDES). The results showed that the average age at which girls produced apologies (2;2;0) was 3 months earlier than for boys (2;5;18). The average age of onset of apology production was 2;4. The authors also found an unbalanced U-shaped curve in the development of apology, with the lowest number of apologies being produced by three year olds. In addition, children’s apologies became more elaborate as they grew older. Lin (2009) investigated the developmen- tal pattern of apology of Chinese children aged 4-8. The findings showed that younger children (under 7 years old) employed more direct apology strategies than older children. Older children used more combined apology strategies than the younger children. Through natural observation of Israeli preschool children and young adolescents, Kamph & Blum-Kulka (2009) explored how children of different ages apologize. They found that the range of apology strategies increased with age. While both the young and older groups employed the formulaic Illocutionary Force Indicating Device (IFID) forms slixa ‘forgive’ and ani micta’er ‘I’m sorry’, only the adult group used mitnacel ‘apologize’, a more formal IFID form. In addition, “promise of forbearance”, “repair” and “minimization” and “intensification” occurred only in adult apologies. The findings also revealed that the range in types of violations increased with age. Additionally, younger children produced apologies mainly with regard to “breach of expectations.” Review of these studies revealed that research on the development of L1 76 Chang: Apologizing in Mandarin Chinese apology, compared with the other strands of apology studies, points to a need for further exploration. The relatively few studies on the development of L1 apology have shown that age is an important factor. The majority of the studies on pragmatic development in L1 apology have examined young children. The paucity of research involving adolescents and young adults has resulted in a less than full understanding of how the production of apologies differ in different age groups. In addition, little attention has been paid to Mandarin-speaking children’s develop- ment of apologetic behavior. Given that aspects such as politeness values and the linguistic forms used in the speech act of apology are culture-specific, it is reasonable to expect that development of apologetic behavior also varies across language and culture. Hence, it is important to examine the developmental patterns of the speech act of apology in many languages in order to determine not only differences but commonalities. In addition, examination of the development of L1 apology from different L1 speakers will provide baseline data for researchers in the field of interlanguage pragmatics to better our understanding of the development of L2 pragmatic competence. In addition, most research on the role of gender in apology has shown that males and females differ in their choices of apology strategies (e.g., Fraser 1981, Holmes 1989, Bataineh & Bataineh 2006, Chen 2008). However, these findings were from studies of adults. Whether gender differences in apology strategies occur in young children and adolescents remained unexplored. Finally, a review of the literature revealed that the majority of the research findings in pragmatic development of apology depend on either production data or perception data. In exploring developmental patterns, collection of production data allows the researcher to establish participants’ pragmalinguistic knowledge of the strategies and linguistic forms that can be used to realize the speech act of apology. Including perception data can help in specifying participants’ sociopragmatic norms and in understanding their sociopragmatic knowledge of contextual factors. This study employed both data collection techniques, thus providing additional insight into the developmental pattern of the speech act of apology. The research questions that guided the study are as follows: 1) Are there developmental differences in perception as to the severity of offense, the obligation to apologize, and the likelihood of apology acceptance across different age groups and across gender among native Mandarin speakers? 2) How does the use of apology strategies differ across different age groups and across gender among native Mandarin speakers? 3) How does the repertoire of apology strategies expand as the age increases? 77 42.1 (May 2016) 3. Methods 3.1 Subjects A cross-sectional approach was adopted to investigate the pragmatic development of the speech act of apology. To explore how apologies by adolescents and young adults differ from apologies by children, subjects from four age groups (3rd grade, 6th grade, 10th grade and college freshmen) were selected to represent children, early adolescents, adolescents, and young adults. Their ages were 9, 12, 16, and 19 years old. Each age group was composed of 60 people: 30 males and 30 females. A total of 240 subjects participated in the study. All the subjects were native speakers of Mandarin Chinese. Data of the 3rd graders and 6th graders were collected from the same elementary school in central Taiwan. The 10th graders were all from the same high school in northern Taiwan. The college freshmen participating in this study were from a research-oriented university in central Taiwan. 3.2 Data collection This study employed a written discourse completion task (DCT) for data collection with an acknowledgement of its limitation in representing the naturally occurring interactions. The written DCT was considered an appropriate instrument for this study due to its strength in reflecting speakers’ pragmalinguistic knowledge and sociopragmatic knowledge of the strategies and linguistic forms selected (Kasper 2000) and in allowing the researchers to manipulate the variables of interest (Beebe & Cummings 1996). To make a comparison between the development of L1 and L2 apology possible, this study adopted scenarios devised by Chang (2010), including bumping into people, losing a borrowed book, being late and speaking ill of someone (See Appendix I for a sample of the questions in the questionnaire). Each scenario was repeated once with either a classmate or a teacher as the hearer. Respondents were presented with a description of the context and the social status of the interlocutors for each scenario. A total of 1,920 responses were collected and analyzed. As mentioned earlier, this study collected both perception and production data. The perception data were collected using a perception assessment questionnaire which elicited subjects’ perceptions of three factors in each scenario: 1) severity of offense, 2) offender’s obligation to apologize, and 3) likelihood of the apology being accepted. This format follows the assessment questionnaire in Bergman & Kasper (1993). 78 Chang: Apologizing in Mandarin Chinese 3.3 Procedures The DCT questionnaire was distributed to the subjects in a classroom environment with a research assistant present to explain the procedures to the participants. The perception assessment task was administered first. In this task, the subjects were presented with the scenarios, each of which was followed by three items: (1) how serious is the offense? (2) Do you have the obligation to apologize? (3) Is your classmate/teacher likely to accept your apology? After reading the scenarios, they were asked to rate each offense on a 9-point scale for these three context-internal factors: 1) severity of offense, 2) offender’s obligation to apologize, and 3) likelihood of the apology being accepted. After the participants finished the perception assessment task, they were asked to write down what they would actually say in the situation. The participants were informed that the classmate in each scenario was of the same age as themselves, whereas the teacher was around age 40. Both the classmate and the teacher were characterized as acquaintances rather than strangers and were of the same gender as the participant. 3.4 Data analysis With regard to the perception data, the means of the quantitative ratings and standard deviation of contextual factors were calculated and compared for each age group. To answer the first research question (Are there developmental differences in perception as to the severity of offense, the obligation to apologize and the likelihood of apology acceptance across different age groups and across gender among native Mandarin speakers?), two-way analysis of variance (ANOVA) was performed for age and gender to assess the statistical significance of differences across age and gender groups. To analyze the production data, the researcher developed a coding scheme based on classifications of previous studies (e.g., Olshtain & Cohen 1983, Trosborg 1987, Blum-Kulka, House & Kasper 1989, Bergman & Kasper 1993). Table 1 presents the coding scheme used for the present study. The apology strategies produced by the participants were analyzed as consisting of a sequence of semantic formulas. For example, if a respondent apologized for losing the borrowed book, saying, “I’m sorry. I lost the book. I did not mean it. I will buy you a new one,” this was coded as: [IFID]-[admission of fact]-[lack of intent]- [repair]. The data were coded by the researcher and a trained research assistant. Intercoder reliability was 90 percent. 79 42.1 (May 2016) Table 1. The coding scheme of apology strategies of the present study I. Illocutionary force indicating device (IFID): a. Expression of regret or offer of apology, e.g., “duibuqi 對不起” or “baoqian 抱 歉” “I’m sorry.” “I apologize.” b. Request for forgiveness, e.g., “qing yuanliang wo請原諒我” “Please forgive me.” II. Adjuncts 1. Explanation or account of the cause which brought about the violation 2. Expression of the speaker’s responsibility for the offense: a. Explicit self-blame, e.g., “shi wode cuo是我的錯” “It’s my fault.” b. Expressing lack of intent, e.g., “wo bu shi guyi de 我不是故意的” “I didn’t mean it.” c. Acknowledgement, e.g., “wo bu yinggai zheme zuo我不應該這麼做” “I shouldn’t have done it.” d. Admission of fact, e.g., “wo chidao le 我遲到了” “I’m late.” 3. Offer of repair, e.g., “wo zai mai xinde gei ni我再買新的給你” “I’ll buy you a new one.” 4. Promise of forbearance, e.g., “xia ci bu gan le 下次不敢了” “It won’t happen again.” 5. Minimization of the degree of offense, e.g., “na mei sheme ya那沒什麼呀!” 6. Speaker showing concern for offended party, e.g., “ni haihao ma 你還好嗎？” “Are you all right?” 7. Intensification, e.g., “feichang非常” “very” 8. Alerter, e.g., “laoshi 老師…” “teacher” 9. Justification, e.g., “nide gushi zhende hen wuliao 你的故事真的很無聊” “Your story is really boring.” To answer the second research question (How does the use of apology strategies differ across different age groups and across gender among native Mandarin speakers?), the apology production data were analyzed in terms of: 1) the complexity of the participants’ apology strategy pattern, and 2) the developmental pattern in the choice of strategy. The complexity of the participants’ apology strategy pattern was examined in terms of the combination pattern of the apology strategies and the total number of strategies employed by each group. The developmental pattern in the choice of strategy was examined in terms of the difference in the frequency and the content of the each type of apology strategy employed. Analysis of the frequency of the strategies used across groups provided information concerning whether one strategy emerged before another. Chi-square analysis was performed to examine 80 Chang: Apologizing in Mandarin Chinese whether the difference was statistically significant or not. The analysis of the content of the strategies allowed us to discover the difference in the linguistic forms used for each type of strategy across the different groups. The analysis of content/linguistic forms of the apology strategies involved qualitative analysis. To answer the third research question (How does the repertoire of apology strategies expand as the age increases?), the analysis method that Chang (2010) proposed was adopted. Disregarding the contextual variations, Chang (2010) ranked the use of each type of strategy, based on the highest percentage that had occurred among scenarios, to investigate whether a certain apology strategy emerges earlier than others. For the rationale and the details of this procedures, please see Chang (2010). 4. Results 4.1 Differences in the perception of the severity of the offense, the obligation to apologize and the likelihood of apology acceptance Tables 2 and 3 present the means and standard deviation for the perceived severity of the offense, the obligation to apologize, and the likelihood of apology acceptance across age and gender groups. A two-way analysis of variance measure was performed to test the effects of age and gender. As illustrated in Table 4, a significant main effect was found for age regarding perception of the obligation to apologize and perception of the likelihood of apology acceptance (F = 18.898, p = .000; F = 40.062, p = .000). There was no significant main effect found regarding perception of the severity of the offense (F = 1.484, p = .217). The standard deviation was high for the third graders, which may indicate that variability in the ratings of the 3rd graders was higher than that of the other groups. Table 2. Means and standard deviation for the perception of the degree of severity, the obligation to apologize, and the likelihood of apology acceptance across age groups Grade 3 Grade 6 Grade 10 College Children Young adolescents Adolescents Young adults Severity of the 6.74 7.00 6.98 6.85 offense (2.60) (2.22) (1.91) (1.94) Obligation to 7.24 7.75 8.16 7.87 apologize (2.54) (1.88) (1.53) (1.53) Likelihood of 5.17 5.66 6.69 6.42 apology acceptance (2.80) (2.64) (2.15) (1.98) 81 42.1 (May 2016) Concerning perception of the obligation to apologize, the results of a Scheffe F test revealed that the 3rd graders perceived the obligation to apologize as significantly lower than did the 6th graders, 10th graders and the college group. The results also disclosed that differences were statistically significant between the 6th graders and the two other age groups (10th graders and 3rd graders) in the mean ratings of obligation to apologize. The differences between the college freshmen and 10th graders were not statistically significant. Table 3. Means and standard deviation for the perception of the degree of severity, the obligation to apologize, and the likelihood of apology acceptance across gender groups Grade 3 Grade 6 t t Gender M F M F p p 6.13 7.34 -5.193 7.03 7.00 .289 Severity of the offense (2.98) (1.99) .000 (2.20) (2.24) .773 6.59 7.88 -5.749 7.83 7.67 .926 Obligation to apologize (3.06) (1.68) .000 (1.89) (1.88) .335 Likelihood of apology 4.81 5.52 -2.809 5.91 5.42 2.027 acceptance (3.02) (2.52) .005 (2.69) (2.57) .043 Grade 10 College t M t Gender M F F p p 6.99 6.97 .097 6.50 7.20 -4.045 Severity of the offense (1.96) (1.91) .923 (2.12) (1.66) .000 8.18 8.14 .238 7.68 8.06 -2.756 Obligation to apologize (1.71) (1.44) .812 (1.70) (1.33) .006 Likelihood of apology 6.78 6.61 .864 6.36 6.49 -.714 acceptance (2.33) (1.96) .338 (2.12) (1.83) .475 Table 4. ANOVA Table for the perception of the degree of severity, the obligation to apologize, and the likelihood of the apology acceptance across age and gender groups SS df MS F p age 21.350 3 7.117 1.484 .217 Severity of theoffense gender 100.072 1 100.072 21.369 .000 age 210.612 3 70.204 18.898 .000 Obligation toapologize gender 65.701 1 65.701 18.219 .000 Likelihood ofapology age 700.880 3 233.627 40.062 .000 acceptance gender 60.949 1 60.949 7.891 .005 82