Introduction: Mediatization, perceptual studies, and japanese language varieties
This article investigates the perception of a mediatized Japanese language variety, Osaka dialect, as it is spoken by the male protagonist in the popular film Soshite chichi ni naru/Like father, like son, released in 2013 (Kore'eda Reference Kore'eda2013). Particular attention is given to the frequency with which the protagonist is judged as using the Osaka dialect or some other variety of Japanese including Standard Japanese. We are motivated by Agha's framework of mediatization (Agha Reference Agha2007, Reference Agha2011) as well as by Campbell-Kibler's (Reference Campbell-Kibler2007, Reference Campbell-Kibler2010) perceptual work on sociolinguistic variants as carriers of social meaning. We report our findings from a perceptual sociolinguistic experiment of a mediatized variety of the Japanese language. We end by discussing the indexical field of Osaka dialect as perceived by lay Japanese speakers.
The semiotic value of language, especially marked language varieties or styles, cannot be understated. Regional varieties or dialects of languages can work as semiotic cues to index region but can also index meanings far beyond region (e.g. Johnstone Reference Johnstone2009). Dialects do not need to be tied to regional origins; rather than demanding that regionality be one index of a dialect, Ball (Reference Ball2004:356) suggests that we understand dialect as ‘a variety in a repertoire whose primary function is microcontextually socioindexical’. Indeed, much scholarship demonstrates the ways in which language varieties serve a creative indexical function by contributing to affective or personality traits of speakers in real life situations (e.g. Johnstone Reference Johnstone2009; Remlinger Reference Remlinger2009). Over time people intuitively assign highly salient metalinguistic labels to the speech they hear, tapping into language ideologies of ‘those kinds of speakers’ within a community (e.g. Johnstone, Andrus, & Danielson Reference Johnstone, Andrus and Danielson2006). This same process occurs in media as well (e.g. Stamou Reference Stamou2014).
In media contexts, for example, Lopez & Hinrichs (Reference Lopez and Hinrichs2017) demonstrated how a mediatized Jamaican Creole used in a widely seen car commercial functioned to display one car owner as someone who is happy in contrast to his coworkers who are not happy with their own cars. Pua & Hiramoto (Reference Pua and Hiramoto2018), in an analysis of James Bond films set in East Asia, show how a general ‘Asian dialect/accent’ is placed onto villains and allies in an uneven fashion. In both of these cases, a close examination of the languages used by the actors proved to be complex amalgamations of actual language forms—Rastafarian speech in the case of the Jamaican Creole—plus segmental variation salient to the English-speaking audiences such that the languages became ‘recognizable’ as Jamaican Creole or an ‘Asian’ language such as Chinese or Japanese. Agha (Reference Agha2011) refers to this phenomenon as mediatization of an enregistered language form (Agha Reference Agha2007). Mediatization is when the repeated association and dissemination of recognizable features of language and their social meanings come to be readily associated by individuals with particular personas, identities, activities, and other culturally relevant categories of people or contexts. Particular kinds of characters are developed and voiced with semiotic clues, including recognizable features of language(s), allowing the viewing audience to make assumptions about the actor/character that specifically support the intended plot lines and narrative arcs (e.g. Hiramoto Reference Hiramoto2009, Reference Hiramoto2013; Occhi, SturtzSreetharan, & Shibamoto-Smith Reference Occhi, SturtzSreetharan; and Shibamoto-Smith2010; Lippi-Green Reference Lippi-Green2011; Bleichenbacher Reference Bleichenbacher2012). These mediated forms of language are important sites of sociolinguistic inquiry given that they carry embedded social meanings to the viewing audience (e.g. Stamou Reference Stamou2014; Androutsopoulos Reference Androutsopoulos and Coupland2016; M. Inoue Reference Inoue, Coates, Fraser and Pendleton2020). However, the social meanings that are perceived by the audience remain a question.
Most relevant to this current examination of the use of marked varieties of a language is previous analyses of the film, Like father, like son (SturtzSreetharan Reference SturtzSreetharan2017a, Reference SturtzSreetharanb). Looking both quantitatively and qualitatively at specific grammatical features including first- and second-person pronouns, sentence final particles, and lexical items, SturtzSreetharan suggested that the film moved Osaka Dialect (OD) beyond its regional index (with the Osaka region of Japan) to include a caring affective father located anywhere. The voice of the caring father was combined with specific behaviors that included spending hands-on time with his children. As Table 1 describes, the caring OD-speaking father protagonist was explicitly juxtaposed to the other father protagonist who spoke Standard Japanese (SJ). The SJ-speaking father's language was found to be more formal and less masculine, which, when combined with his behavior of absentee fathering practices, gave an impression of a cool, distant father. While the SJ-speaking father and his family live in the Tokyo metropolitan area, the OD-speaking father and his family live in the rural area of Maebashi, about sixty miles (100 km) from Tokyo and more than 300 miles (480 km) away from where OD is spoken. No explanation is given in the film for the OD-speaking father's linguistic style. In short, the film supports and reinforces the stereotypical associations of SJ and OD by unhooking OD from its regional tethers and mapping its indexical values onto behaviors that resonate with common social assumptions about SJ versus OD speakers and attitudes. However, this previous work only concerned itself with lexical and morphological analyses of OD as produced by one of the protagonists, ignoring how lay audiences perceive this mediatized form of the Japanese variety. Our current endeavor advances this work by conducting a perceptual investigation of the choice of mediatized language varieties in the film. Such a perceptual investigation has bearing on how mediatized language varities are interpreted by individuals which in turn affect the social meanings claimed (see Campbell-Kibler Reference Campbell-Kibler2010; Pharao, Maegaard, Møller, & Kristiansen Reference Pharao, Maegaard, Janus, Møller; and Kristiansen2014; Levon & Fox Reference Levon and Fox2014). Our dialect recognition experiment investigates the degree to which the OD-speaking father's speech is recognized as OD by lay listeners, and the focus-group interviews examine the affective meanings that are invoked by OD in listeners’ minds.
In the past decade, a number of perceptual studies have investigated the social meanings that listeners attribute to specific productions of language (Campbell-Kibler Reference Campbell-Kibler2007, Reference Campbell-Kibler2010, Reference Campbell-Kibler2011; Yuasa Reference Yuasa2010; Levon Reference Levon2014; Pharao et al. Reference Pharao, Maegaard, Janus, Møller; and Kristiansen2014; Podesva, Jamsu, Callier, & Heitman Reference Podesva, Jamsu;, Callier; and Heitman2015). For example, Campbell-Kibler (Reference Campbell-Kibler2007) found that listeners perceived speakers who pronounced the verb ending -ing with the alveolar -in as more likely Southern and less likely homosexual and/or urban compared to speakers who used the velar -ing. Yuasa (Reference Yuasa2010) investigated listeners’ perceptions of creaky voice and reported that it was more often attributed to upwardly mobile educated women; but, she found that it was also perceived as informal and hesitant indicating contrasting social meanings indexed by the same linguistic feature. Similarly, Pharao and colleagues (2014) examined perceptions of Danish [s] and a fronted [+s] variant in ‘modern’ and ‘street’ Copenhagen registers finding that the fronted [+s] variant was more often labelled as homosexual and feminine in only the ‘modern’ register revealing the linguistic features’ potential meanings in different contexts. Finally, Podesva and colleagues’ (Reference Podesva, Jamsu;, Callier; and Heitman2015) investigation of released [t] by political figures revealed that social meanings of this sound are constrained by social as well as linguistic factors. These studies elegantly demonstrate Eckert's (Reference Eckert2008) idea of indexical field and serve to account for the unequal success of speakers to use the forms effectively (e.g. Podesva et al. Reference Podesva, Jamsu;, Callier; and Heitman2015).
The Japanese language, representation, and indices
The Japanese language is well-known as one that is marked for honorifics and gender (e.g. Kindaichi Reference Kindaichi1957; Ide Reference Ide1982). Much earlier scholarship, in fact, has focused precisely on this intersection of polite language forms and gender of speaker (e.g. Reynolds Reference Reynolds1985; Shibamoto Reference Shibamoto1987; Takahara Reference Takahara1991; Endo Reference Endo1997). Additionally, scholarship on Japanese language and gender has also focused on the use of regional varieties by women (mainly) (e.g. Okamoto Reference Okamoto1998, Reference Okamoto, Jones and Ono2008; Sunaoshi Reference Sunaoshi, Okamoto and Smith2004; Didi-Ogren Reference Didi-Ogren2011) with some investigations of their use by men (SturtzSreetharan Reference SturtzSreetharan2004, Reference SturtzSreetharan2009)
OD is a western regional variety of the Japanese language. Among the many regional varieties of Japanese, OD has been widely audible through radio broadcasting since the early twentieth century due to the popularity of manzai, a two-person comedic form that is said to have originated in the Osaka region (Koyano Reference Koyano2004). More recently, OD has been recognizied as an enregistered mediatized language form through various audio-visually programming much of it comedic in nature including reality TV (see SturtzSreetharan Reference SturtzSreetharan2015, Reference SturtzSreetharan2017c). Mediatized varities of any language are not representative of any real speakers, or to put it another way, mediatized varieites are not an accurate form of any spoken language. This is the case for mediatized OD as well. Nonetheless, it is easily recognized and identified when used by televisual characters or when written due to its distinct lexicon and morphology; consequently, OD is commodified on T-shirts, coffee mugs, and other souvenirs in much the same way that Johnstone describes for ‘Pittsburghese’ (Reference Johnstone2009; Johnstone, Andrus, & Danielson Reference Johnstone, Andrus and Danielson2006). In addition, OD forms are used frequently in manga and anime as well as by young people throughout Japan in their speaking and texting ( Kinsui Reference Kinsui2003; Tanaka Reference Tanaka2011; Tanaka, Hayashi, Meada, & Aizawa Reference Tanaka, Hayashi, Maeda and Aizawa2016).
Analyses of mediated representations of regional varieties of Japanese have also been undertaken. These studies have demonstrated that young modern heroines avoid dialect, preferring to use SJ instead; heroes, by contrast, are represented as speaking dialect, especially western dialects such as OD (Shibamoto Smith & Occhi Reference Shibamoto and Occhi2009; Occhi et al. Reference Occhi, SturtzSreetharan; and Shibamoto-Smith2010). In contrast, northeastern Japanese regional varieties were selected to represent slave speech and poor white speech in the Japanese translation of Gone with the wind (Hiramoto Reference Hiramoto2009). As Miyake (Reference Miyake1995) has noted, SJ has come to index modernity while many of the regional varieties have come to be viewed as the language of lower, uneducated, and rural classes of people with OD and other western regional variants being an exception.
The indexical field of OD has shifted over time. Prior to the 1980s, outsiders viewed OD as garish, gaudy, and harsh (F. Inoue Reference Inoue2009). Then, in the 1980s, OD gained an image of fun and laughter (owarai kotoba) which Inoue (Reference Inoue2009) suggests is due to a resurgence of interest in manzai (two-person comedy). Finally, in the 1990s, a ‘dialect boom’ brought national appeal to OD (Tomosada Reference Tomosada1995) with people reporting that OD is cool and fun/funny (Tomosada & Jin'nouchi Reference Tomosada and Jin'nouchi2004). The continued destandardization of the Japanese language finds OD showing up in many places, especially instant messaging and social media chat platforms (see F. Inoue Reference Inoue, Heinrich and Galan2011). Indeed, a language attitude survey of over 250 Japanese college students undertaken in 2007 (Tanaka Reference Tanaka2011) demonstrated that the people in Tokyo described various Japanese language varities, including OD, as ‘interesting’, ‘cool’, and ‘masculine’ (Table 2).
Drawing on previous perceptual studies of the social meanings that people attribute to particular variabilities in language (e.g. Campbell-Kibler Reference Campbell-Kibler2007, Reference Campbell-Kibler2010, Reference Campbell-Kibler2011; Pharao et al. Reference Pharao, Maegaard, Janus, Møller; and Kristiansen2014; Podesva et al. Reference Podesva, Jamsu;, Callier; and Heitman2015), we conducted two perceptual studies on the mediatized OD in Like father, like son. First, the dialect recognition experiment investigated the extent to which the variety spoken by the OD speaking protagonist (Yūdai) is recognized as OD by audiences. Second, the qualitative ethnographic focus-group interviews investigated the extent with which the social meanings reported in the literature for OD are, in fact, perceived as such by audiences. We begin with the dialect recognitions experiment.
Dialect recognition experiment
In order to address the recognizability of the OD used in the film, we conducted a dialect recognition experiment using the online tool Qualtrics to recruit listeners widely. The link to the task was distributed through personal networks and social media.
A total of 135 self-identified native Japanese speakers participated in the experiment. Participants who did not complete at least half of the trials (n = 40) and those who had seen the film (n = 26) were removed from analysis. The responses from the remaining participants (n = 69) were analyzed. Of these sixty-nine participants, sixty-eight of them completed 100% of the trials while one participant completed over 90%.
Participants’ ages ranged from nineteen to sixty-two years old (average age = 24.8 years old). Of the participants, forty-one (60%) were female and twenty-eight (40%) were male. At the time of the study, the majority (65%) were residing in the western region of Japan (Kansai), including Osaka, and fewer in the eastern region (Kantō), including Tokyo (32%). One participant was living in the southwestern region (Chūgoku) and one did not specify current residence. Participants were also asked to indicate where they were reared. About 51% of participants grew up in the western region, and 28% in the eastern region. About 13% were originally from the southern region of the main island and 4% were from the southern island of Kyūshū. The remaining participants were from other areas with one from each of the following regions: the northeastern region (Tōhoku), the southwestern region (Chūgoku), and the southwestern island (Shikoku). Just one participant declined to identify their birthplace.
Participants were also asked to provide information about the variety of Japanese they regularly spoke. The response choices were (i) SJ only, (ii) mainly SJ and sometimes dialect, (iii) dialect only, (iv) mainly dialect and sometimes SJ, (v) other, and (vi) decline to state. As shown in Table 3, responses were fairly evenly spread out; each choice except for ‘other’ applied to about 20–30% of respondents but most (28%) indicated that they used dialect only. The five participants who selected ‘other’ indicated that they were all speakers of SJ and some dialects. Those who indicated that they speak a dialect (n = 49) were asked to report what dialects they spoke by typing their answer in a response box (Table 4). The results show that the participants were fairly experienced with Japanese dialects and most with the dialect(s) of the western region, including Osaka.
a Kansai dialect is often used to refer to varieties of dialects spoken in the western region of Japan including Osaka, Kyoto, Nara, and Kobe.
A total of eighty-four short sound clips were chosen as stimuli from various scenes throughout the film for this experiment. We identified eighty scenes in the film, separated by changes in the physical location, events depicted, characters, and topics featured. A total of ten scenes were represented in the stimuli. Note that stimuli were sounds only, and did not include images, as the study focuses on the effects of OD rather than a constellation of auditory and visual semiotic cues as used in the film.
Of the eighty-four sound clips, fifty-four (64%) came from OD-speaking Yūdai's lines as it was his dialect that was the focus of our investigation. We also selected thirty from other characters (see Table 1): fourteen from SJ-speaking father Ryōta, ten from Ryūsei (Yūdai's son), and six from Keita (Ryōta's son). While the focus of this study was listeners’ responses to Yūdai's stimuli, Ryōta's clips were included to verify that his language is SJ, and children's clips were added to include a variety of voices to the task. The number of clips from each character could not be completely balanced as each character had different amounts of screen time and background sounds of the film limited the number of clips that could be used for an experiment like this. Nonetheless the clips were chosen from scenes throughout the film that featured lexical and morphological features of OD, some originally identified as emblematic of Yūdai's OD in SturtzSreetharan (Reference SturtzSreetharan2017a). We also aimed to choose lines of speech from the same scenes across the two fathers in order to match content of speech across the stimuli.
Each stimulus contained a unit of speech forming an utterance, marked by an intonational boundary tone at the right edge, typically followed by a pause or a turn. The stimuli were thus fairly short, ranging from 0.5 and eleven seconds in length. (1) and (2) are two examples.Footnote 1
We divided the eighty-four stimuli into four sets of twenty-one, each containing thirteen or fourteen Yūdai stimuli, three or four Ryōta stimuli, and four child stimuli. We created four versions of the dialect recognition task with each set of twenty-one stimuli. These four versions were created so that the recognition task could be completed without exhausting participants while we collected responses to a wide range of samples taken from the film. A pilot participant responded to trials in one sitting without taking breaks, and the task was completed within ten minutes. Each version of the task (i.e. each stimulus) received eleven to twenty-one responses.
Task and procedure
Participants were randomly assigned to one of the four versions of dialect recognition task. After completing demographic questions, participants listened to each of the twenty-one audio stimuli, and identified the dialect of the speech. Each trial presented one auditory stimulus at a time, with the simultaneous display of seven response choices on the same online page. The participants were instructed to click a sound icon to listen to the stimulus, and then click their response choice. They could listen to each stimulus as many times as they liked before submitting their response. Possible response choices were: the person spoke (i) clearly SJ, (ii) probably SJ, (iii) clearly OD, (iv) probably OD, (v) clearly a dialect, although not OD, (vi) probably a dialect, although not OD, or (vii) cannot decide. When participants submitted their response choice, the next trial was presented. After making a judgement for all twenty-one stimuli, participants were asked whether they had seen the film or not and their responses were discarded if they had.
Analysis and results
‘Clearly’ and ‘probably’ response choices were provided to assist with the task in case participants’ confidence varied across stimuli. However, since the level of confidence was not critical to our research question, we combined ‘clearly’ and ‘probably’ responses into single categories so that there were four response categories of ‘standard’, ‘Osaka’, ‘other’, or ‘no response’.
We first examined the overall pattern of responses. Responses to stimuli spoken by each of the four characters (i.e. Yūdai, Ryōta, Ryūsei, and Keita) were tallied, and the results in percentages are illustrated in Figure 1. Percentages are used here since the number of stimuli varied across the characters, and the number of responses also varied across stimuli. The total raw number of responses was 921, 239, 175, and 101 for Yūdai, Ryōta, Ryūsei, and Keita respectively. Figure 1 shows that Ryōta and his son Keita were almost always judged as SJ speakers (95% and 93% respectively). This is not surprising as they are presented as unambiguously metropolitan Tokyo people throughout the film and their perceived dialect follows suit. By contrast, the responses to the stimuli for Yūdai (the focus of our investigation) and his son Ryūsei were mixed. While Yūdai was most often heard as a SJ speaker (55%), he was also heard as an OD speaker (24%) or as a speaker of some other dialect (15%). Ryūsei, his son, was more often heard as an OD speaker (54%).Footnote 2 In order to examine if the use of dialect was indeed distinct across the two fathers, Yūdai and Ryōta, we conducted a Pearson's Chi-Square test to compare the percentage data for the two characters. The results confirmed that the pattern of perceived dialect was different between the two father characters (χ2(3) = 131.53, p < .001). Interestingly, the comparison between responses to Yūdai's speech and his son Ryūsei's speech also showed statistical significance (χ2 (3) = 91.86, p < .001). This makes Yūdai's speech pattern very distinct: it is different from Ryōta's ‘clearly SJ speech’ pattern, and it is different from Ryūsei's ‘clearly OD’ pattern. However, we report the comparison between Yūdai and Ryūsei with some caution, since the sample size was much smaller for Ryūsei (101) compared to Yūdai (921) and Ryōta (239). Nevertheless, these results establish that linguistic behaviors of the two fathers were different as judged by listeners: whereas Ryōta's speech was consistently heard as the standard variety, Yūdai's speech did not show a consistent pattern. While listeners clearly perceived OD or another regional dialect in Yūdai's speech, considerable instances of SJ were also identified.
In order to further probe the mixed results of Yūdai's speech, we examined the pattern of dialect identification across scenes of the film. As noted earlier, ten out of eighty scenes from the film were represented in the stimuli, and Yūdai's speech appeared in seven of them. As Table 5 shows, some of the seven scenes feature the two fathers talking with each other while others feature Yūdai talking with his friend or family members. Figure 2 illustrates Yūdai's perceived dialect across the seven scenes.
Figure 2 suggests that Yūdai was heard as an OD speaker with different frequencies across scenes. He was heard more often as an OD speaker in scenes such as 42 (51%), 39 (37%), 58 (31%), and 22 (29%). A notable characteristic of these scenes is that they include interactions between Yūdai and Ryōta or Yūdai and Ryōta's family. An example conversation from Scene 42 in which the two families meet at an indoor playground is shown in (3). While the children are playing, Yūdai and Ryōta argue about what to do about their respective sons. The speaking lines of Yūdai that were tested in the dialect recognition task indicate the percentage of dialect selected. Note that listeners ‘recognized’ OD in Yūdai's speech frequently (51%) in this conversation. NR in the parenthesis indicates ‘no response’.
By contrast, Yūdai is heard as speaking SJ dominantly in Scenes 32 (73%) and 29 (74%). Unlike the four scenes discussed above, Yūdai is speaking with people in his inner circle in these scenes, at home with his family (Scene 32) and with a good friend at his shop (Scene 29). Example (4) is a conversation taken from Scene 32, where Yūdai is at home with his family, and Keita is visiting them. Yūdai takes everyone to the late grandmother's altar to introduce Keita to her; the whole family then has breakfast together. Notice the listeners’ judgements of Yūdai's dialect are mostly SJ.
As these excerpts and Figure 2 suggest, Yūdai was perceived to sound more like an OD-speaker in scenes in which both Ryōta and Yūdai interacted (Scene 42) or Yūdai interacted with Ryōta's family. However, Yūdai was perceived to sound more like that of a SJ speaker in scenes where Ryōta was absent (Scene 32).
To contrast the difference between Yūdai and Ryōta's speech patterns more clearly, we identified three scenes (15, 22, and 42), which were part of the dialect recognition experiment and where both Yūdai and Ryōta were speaking. In these three scenes, Ryōta was judged as speaking SJ 84%, 98%, and 99% of the times respectively; Yūdai was judged as speaking OD 27%, 29%, and 51% of the times respectively. Yūdai and Ryōta's attitudes are starkly different in Scene 42, and their language use was also perceived as significantly different, (χ2(3) = 97.13, p < .001). Whereas Ryōta is consistently and reliably judged to be a SJ speaker, Yūdai's language is perceived inconsistently and unreliably as an OD speaker. Critically, however, Yūdai is consistently judged to be a non-SJ speaker. In this way, as was pointed out by SturtzSreetharan (Reference SturtzSreetharan2017a, Reference SturtzSreetharanb), the contrast in the two fathers’ attitudes, socioeconomic class statuses, and affect amplifies the juxtapositioning of the languages used by each, especially the nonstandard variety used by Yūdai. It is the social meaning of these perceived differences that we turn to next.
Ethnographic focus-group interviews
The results of the dialect recognition experiment found that listeners did recognize some of Yūdai's speech as OD. By contrast, there was very little ambiguity in the recognition of Ryōta's speech as SJ. The findings also show that Yūdai's OD production changed across scenes. Those scenes in which he and Ryōta were at odds in their parenting styles brought about more frequent judgements of Yūdai's speech as OD. The dialect recognition experiment, however, could not answer our second question; that is, it could not tell us precisely what social characteristics, including his portrayal as an affective, hands-on father as we argue, were being conveyed to audiences by the deployment of OD in Yūdai's speech. In order to answer this question, we follow Campbell-Kibler (Reference Campbell-Kibler2010) and Podesva et al. (Reference Podesva, Jamsu;, Callier; and Heitman2015) in their use of semi-structured interviews to elicit possible social meanings associated with OD.
Participants and interview groups
Twelve native Japanese speaking participants (three men and three women from the Tokyo area and three men and three women from the Kansai area) participated in focus-group interviews conducted by the first and second authors. Previous work suggests that two focus groups allow identification of at least 80% of the most prevalent themes relevant to the inquiry (Guest, Namey, & McKenna Reference Guest, Namey and McKenna2016). All participants were between the ages of eighteen and thirty (average age is twenty) and were enrolled in a university located in the pacific northwest; none had been studying in the United States for more than six months prior to the interviews. The interview groups were separated by gender and region so that the three women from Tokyo were in a group together and the three men from Tokyo were in another group together. This was mimicked for the two groups of Osaka participants. The three members of each group were friends. Initial contact was made with one of the members, and they were asked to bring two friends from the specified region. This was done so that participants felt comfortable sharing their thoughts regarding gender, region, social meanings, and language varieties in Japan. The interviewees were not familiar with the movie, and therefore, did not have any expectation or bias toward the language usage in the movie. The demographic information of each interviewee was collected before the interviews.
The interview participants listened to audio stimuli and answered questions asked by moderators (the first and the second authors). Note that images of the film were never shown to participants. For these interviews, we developed six longer stimuli each lasting thirty to forty seconds, containing five to fourteen stimuli used in the dialect recognition experiment so that the results of the interviews shed light on the aspect of meaning imbued by the pattern of dialect choice revealed in the recognition experiment. The strength of OD for the stimuli involving Yūdai's speech was determined as strong, moderate to strong, moderate, or weak based on the results of the recognition experiment (Table 6). Stimulus 1 through 3 contained a conversation taken directly from a given scene of the film. Stimulus 4 through 6 contained compiled lines taken from multiple scenes. The lines for compiled stimuli 5 (of Yūdai) and 6 (of Ryōta) were selected carefully so that the lines included similar topics and spoken in similar moods across the two fathers. For instance, both stimuli included lines from an arguing scene (Scene 42), where both of the fathers expressed anger and aggression, so the perception of the two characters was not biased.
a Stimuli from Scenes 15 and 22 were not included as these featured too many other voices along with a frequent usage of first names.
The interviews, involving three participants and two moderators at once and conducted in Japanese, were audio-recorded. Each interview was about thirty minutes in duration. The interviews were semi-structured allowing participants to initiate discussions on their own. Therefore, the flow was guided but not fixed by the questions. Each interview was conducted in three steps to address different aims. Step 1 intended to provide the general context for the participants and draw attention to Yūdai as a man, father, and a husband. Thus, the three stimuli containing complete dialogues between Yūdai and another speaker were played in this step (stimulus 1, 2, and 3 in Table 6). The interviewees were then asked to share their thoughts on who the speaker might be as a person, a man, and a father. Participants were instructed to ‘focus on the first man in the audio clips’ to draw their attention to Yūdai while not making reference to the film.
In step 2, we aimed to determine the participants’ perceptions of Yūdai's language variety specifically. Participants heard two stimuli of Yūdai's lines from which all other characters’ lines were removed (clips 4 and 5 in Table 6). Critical to this step, the two stimuli respectively contained Yūdai's utterances that had received lower OD responses (clip 4) and those that received higher OD responses (clip 5) in the dialect recognition experiment. To probe listeners’ responses to this critical difference, we played these stimuli of weak OD and moderate to strong OD back to back in this step. After hearing these two stimuli, participants were asked what dialect they thought ‘the man’ was speaking. We aimed to establish the interviewee's recognition of Yūdai's dialect and whether it was consistent with that of the dialect recognition experiment respondents’ judgements. In addition, we asked the participants to compare the two men speaking in stimulus 4 and stimulus 5 (although they were both Yūdai) in terms of social meanings such as who is the speaker as a person, a man, and a father. They were also asked to discuss what kind of people speak OD. This last question was meant to draw out interviewees’ opinions toward OD more explicitly rather than implicitly as had been done in step one.
In step 3, we presented OD and SJ in a more direct way by playing the compiled clip of Yūdai's lines from the high OD scenes (stimulus 5) and another compiled clip of Ryōta's lines from scenes in which he was perceived, in the dialect recognition experiment, to be unambiguously a SJ speaker (stimulus 6). While the second step sought to elicit the participants’ reactions toward the style shifting of Yūdai between weak and moderate/strong OD, the third step advanced to investigate how the participants perceived the contrast between OD-speaking father and SJ-speaking father. The participants were asked to discuss their attitudes about who Ryōta might be as a person, man, and father. Finally, the participants were asked to judge the traits of masculinity and fatherhood for the two men.
Analysis and results
All recordings (a total of eighty-two minutes in duration) were transcribed and coded by the first and second authors. The transcripts were iteratively read and emerging themes were identified (Bernard, Wutich, & Ryan Reference Bernard, Wutich and Ryan2016). The themes that emerged from the directed questions in the focus groups included dialect, masculinity, and fatherhood. All focus-group interviews were coded for these themes.
The results of the interviews provided insight into the social meanings ascribed to both OD and SJ. In particular, the ways in which these speaking styles connect to fatherhood were revealed. Firstly, while all of the participants identified Ryōta as a SJ speaker and Yūdai to be an OD speaker, the group of Osaka men expressed confusion about which part of the Kansai region Yūdai came from; moreover, the Tokyo men could not decide whether Yūdai sounded like an OD speaker or a Tohoku-dialect speaker (Northeastern Japanese regional variety). When asked why they thought Yūdai was an OD speaker, the group of Tokyo women vaguely referenced lexical or morphological aspects of his speech that they called his hanashikata ‘way of speaking’. From this we could see that the focus-group participants were judging the dialect of the speakers in ways similar to those in the dialect recognition task. That is, they seemed to think of the speaking style as OD but the confidence level was not very high.
When asked about the possible occupational status of Yūdai, all interviewees guessed that he was someone who was self-employed or works with his hands, possibly as a carpenter; in short, he was not heard as a white collar worker (salaryman). Many also commented that he might not have graduated high school or perhaps only attended a technical college suggesting that Yūdai's speech indexed a lower socioeconomic status in the minds of the interviewees. Participants judged Yūdai to be friendly and approachable, which aligns with stereotypical indices of OD as noted in the extant literature discussed above.
When specifically asked about Yūdai's masculinity, all participants except for the Tokyo men described Yūdai as otokorashii ‘masculine’ and related his masculinity to physical strength (chikaramochi). The Tokyo men, however, were somewhat more critical. They described Yūdai as sounding ganko ‘stubborn’ and ibatteiru ‘arrogant’, explaining that these traits align with a mukashi-no ‘old-fashioned’ kind of masculinity. This might be due to the fact that they were all asked this question after describing Yūdai as someone who engages in manual labor or does work that requires strength. Another possibility is a regional difference that just these Tokyo men felt towards an OD speaking man because none of the Tokyo women participants or Osaka participants (women and men) made such a comment.
When asked about their perceptions of Yūdai as a possible father figure, participants gave responses that aligned with a caring and affective father. When asked about the fatherhood of Yūdai, a woman from the Tokyo group stated that he seems like a father who would “protect (his) children”. A man from the Tokyo group noted that in contrast to Ryōta, the “one speaking in [OD]” (Yūdai) would be a “good father” who has a “good relationship with his kids”; he demonstrates metalinguistic awareness when he goes on to say that his opinion may be influenced by the speaker's use of OD which is “more open and sociable” than SJ.
When asked about how Yūdai might be as a possible husband, many participants expressed that Yūdai might not be an ideal partner. They described him as stubborn, too traditional in his thinking of the roles of men and women, and possibly rough. A woman from the Tokyo group described the OD speaker as teishu kanpaku sō ‘seems like a domineering husband’. Another woman from this group described the OD speaker as someone who “holds the impression that a father is strong but women are weak”.
When asked about Yūdai's weak OD and moderate to strong OD (step 2 of the interviews), all groups agreed that the stronger OD version was Kansai dialect. None of the groups said it was specifically OD. Recall that Kansai is a larger area that includes Osaka. Men from Osaka said that the speaker could be from Osaka, Hyogo, Kyoto, or Shiga, prefectures adjacent to Osaka, and they could not determine which specific dialect it was. Similarly, men from Tokyo said the speaker was from “Kansai”, but it could be a variety from a region more southwest than Osaka. This highlights the fact that Yūdai's dialect represents a mediatized version of Kansai dialect, although such a dialect actually does not exist in reality. When asked to directly comment on the impressions of the two styles, men from both Osaka and Tokyo described the weaker OD version as fōmaru ‘formal’ or kyorikan ga aru ‘distant’. While women from Osaka said the stronger OD version was shitashimi yasui ‘familiar’, women from Tokyo described the weaker OD version as yasashii ‘kind’, kashikomatta ‘formal’, and chikarazuyosa ga nakunatta ‘lost strength’.
Finally, we had participants listen to audio-clips of Ryōta speaking and asked them to answer the same questions regarding who the speaker was as a person, man, and father (step 3 of the interviews). The majority of all participants responded that Ryōta sounds like a salaryman who graduated from university, an elite, modern man. They also described him as majime ‘serious’ and gōriteki ‘rational’, not stubborn like the other father. Compared to the perception of OD-speaking Yūdai, SJ-speaking Ryōta was judged as someone who belonged to a higher socioeconomic status. With regard to the perception of Ryōta's masculinity and fatherhood, women from Tokyo judged him to be less masculine than “the other man”. Men in Osaka judged Ryōta as otokorashikunai ‘not masculine’. They expressed that he did not sound like an “honest artisan/craftsman” but more like an “average guy” (lit. common person); but, they expressed that he would “look after his children properly”.
An important aim of step 3 of the interview was to compare OD and SJ directly. In general, speakers thought that Ryōta would be a good (and better) husband but that Yūdai would be a more pleasant father. A woman from Tokyo was the minority voice in her judgement of preferring Yūdai as a husband. And, one woman from Osaka expressed that Ryōta would be the better husband because his ‘speaking style is calmer’ hanashikata ga odayaka. This could be due to the content in the clips in which there is an argument between the two fathers. Whereas Yūdai's energy and emotional openness seems to be judged by some as friendliness and approachable, others judged it to be irrational and aggressive. Lastly, a man from Osaka described Yūdai as the “better father” because chichioya rashii koto o suru ‘he would do father-like things’, while Ryōta was obliquely described as the “better husband” as he would provide seikatsu ga antei suru ‘a stable lifestyle’ due to his proper job at a company.
The results of the interviews illustrated that the OD and SJ as used by Yūdai and Ryōta indeed gave listeners the impressions of a hands-on, affective father and a cold, distant father respectively as they were meant to be portrayed in the film. Yūdai was also perceived as lower in social class, and more masculine by all but the Tokyo men. Ryōta, by contrast, was perceived as a college graduate, elite, salaryman; almost exactly as his character was framed in the film. While Ryōta was not perceived to be as masculine as Yūdai by most, the Osaka men were clear in their perception that Ryōta was not masculine, indicating some differences in the perception of masculinity between Tokyo and Osaka men towards other men. Finally, of note, all participants thought Yūdai would be a more pleasant father while Ryōta would offer more stability and be a better husband due to his good career and more rational, calm personality. Perception of the two men can be summarized as seen in Table 7.
Previous work on the film Like father, like son had shown the ways that lexical and morphological features of OD were deployed by Yūdai (SturtzSreetharan Reference SturtzSreetharan2017a, Reference SturtzSreetharanb); however, the perception of Yūdai's use of non-SJ had not been investigated. Moreover, the social meanings that audiences perceive when hearing men use OD remained unexplored.
Our results from the dialect recognition experiment, presented above, clearly show that Yūdai's dialect was more strongly recognized as OD (or, at least a nonstandard variety) when the two fathers are interacting, and it was less recognized as OD in scenes where he is interacting with his own family and friends. These findings provide evidence to support previous arguments that the choice to have Yūdai use OD was strategic and instrumental in contrasting the two men's styles of fatherhood in the film rather than just being an aspect of Yūdai's character and back story (SturtzSreetharan Reference SturtzSreetharan2017a, Reference SturtzSreetharanb). In the meantime, Ryōta is overwhelmingly and consistently recognized (95% of the time) as being a SJ speaker. Intrigued as to what might underlie the lack of recognition of Yūdai's speaking style as OD, a comparative acoustic analysis was conducted with some of Yūdai's stimuli and the production of the same materials by two Japanese men who were born and raised in Osaka and self-identify as OD speakers. Although we do not have the space here to go into detail, we found two points of interest: (i) the verb morphology of some of Yūdai's utterances did not always adhere to typical OD features; and (ii) Yūdai's pitch production did not consistently conform to OD pitch patterns. We speculate that these non-OD patterns of pitch and morphology resulted in a lack of constant and uniform recognition by our respondents with regard to Yūdai's use of OD.
Our results from the ethnographic focus-group interviews confirmed that many of the social meanings (indices) previously identified for OD are, indeed, perceived by audiences who hear a man speaking OD. That is, when men speak OD it has been suggested that they sound ‘cool’, ‘informal’, and ‘masculine’ (Occhi et al. Reference Occhi, SturtzSreetharan; and Shibamoto-Smith2010; Tanaka Reference Tanaka2011). Moreover, the focus-group interviews also confirmed that OD can enhance the affective perception of a father as suggested by previous work on this film (SturtzSreetharan Reference SturtzSreetharan2017a, Reference SturtzSreetharanb). Women and men from both the Osaka and Tokyo areas of Japan agreed that Yūdai sounded like he would be the kind of father who “protects his children”. Speakers described Yūdai as someone who would daiji-ni shite kureru ‘take care of someone’, which connotes a more loving or cherishing attitude towards those for whom he cares, for example, children. He was also described as “available” and “easy to approach”—two characteristics that could be interpreted as positive qualities in a father. While both groups of women and the men from Osaka described Yūdai as masculine, the men from Tokyo did not. Likewise, both groups of women and the men from Tokyo described Ryōta as less masculine than Yūdai; the men from Osaka specifically described Ryōta as not masculine. This is attested in the popular literature with statements by OD speaking men that SJ-speaking men sound ‘effeminate’ and ‘soft’ (e.g. Okamoto & Shibamoto-Smith Reference Okamoto and Shibamoto-Smith2016).
Taken together, these two studies demonstrate the ways in which social meanings are created through an OD-speaking subject as well as through an SJ-speaking subject. To perceive OD as warm and masculine must be understood in juxtaposition to the perception of SJ as cool/cold and not-so-masculine. Moreover, in accordance with perceptual studies of sociolinguistic variants such as Campbell-Kibler (Reference Campbell-Kibler2007, Reference Campbell-Kibler2010), we find that the perceived social meanings are not the same across all audiences. In point of fact, Tokyo men and Osaka men do not perceive the others’ language variety as masculine even though women do. Additionally, our findings demonstrate that it is the voice of the SJ-speaking protagonist that is judged to be the more desired husband; this was due to the perception that he (Ryōta) would be able to provide a “stable lifestyle”. This judgement is echoed in previous findings by Darling-Wolf (Reference Darling-Wolf2004) who notes that women report the desire to marry a salaryman due to his ability to provide a stable lifestyle for his spouse/family. In contrast, however, the OD-speaking protagonist (Yūdai) is judged to be the more caring father with the assumption that he would be available, intimate, and close to his children. These characteristics go together with contemporary notions of the ikumen ‘child rearing men’ who are valorized for their abilities to undertake hands-on care responsibilities of their children (Mizukoshi, Kohlbacher, & Schimkowsky Reference Mizukoshi, Kohlbacher and Schimkowsky2016).
In this article, we investigated the perceptions of a mediatized variety of Japanese, OD, when used by an OD-speaking male protagonist in the film Like father, like son. We undertook a dialect recognition experiment to ascertain the ways that native lay speakers of Japanese would recognize and identify the language used by the OD-speaking protagonist, Yūdai. In addition, we conducted ethnographic focus-group interviews in order to understand the kinds of social meanings that native lay speakers of Japanese ascribe to the mediatized OD. Studies of linguistic variation and social meaning focus mainly on speakers and the ways in which specific features point to salient social meanings. Fewer studies have focused on how these variables and features are perceived, checking to what extent the features are taken up by an audience as cues to the social meanings rather than incidental correlates. Here we demonstrated that listeners not only recognized Yūdai's use of mediatized OD but they also reported a link between OD and certain affective meanings about masculinity, fatherhood, and husband fitness in this mediatized context. Our study found that Yūdai was most recognized as speaking OD when he was interacting with the other male protagonist, Ryōta, who speaks SJ. In scenes where Yūdai interacted with his own family or with neighbors, he was recognized to be an SJ speaker. We suggest it is his frequent non-OD pitch production along with incorrect OD verb morphology that works to create an unevenness in his OD production.
Our findings with regard to the social meanings that are created by OD demonstrate that OD-speaking men are judged positively to be cool, available, and masculine as well as to be good fathers who will most likely care for and protect children. Likewise, OD-speaking men are judged negatively to be aggressive, stubborn, traditional, and (lower) working class. This mixture of social meanings supports arguments made elsewhere about OD but had not yet been empirically substantiated.
Finally, our dialect recognition experiment along with the ethnographic focus-group interviews are significant given that this is one of only a few studies in Japanese that specifically examines the role of perception in the construction and uptake of indexical meanings of regional varieties. We are contributing directly to building knowledge with regard to what shapes and impacts a language variety's indexical field. The mediated version of the Japanese language variety, OD, has clearly gone beyond its regional boundaries, successfully creating social meanings that index masculinity, informality, and intrinsic qualities, including what sounds like a ‘good father’ and ‘good husband’.