Professional Documents
Culture Documents
Article
Hannah E. Dahlberg-Dodd
Abstract
This paper examines diachronic changes in the pronoun use and phonetic per-
formance of protagonists in shōnen anime, a genre of animated work that is
aimed predominately at adolescent boys. Utilising nearly forty years of shōnen
anime, this study constructs a diachronic analysis of first-person pronoun
usage, the primary pragmatic index of gender performance in Japanese, as well
as average pitch and pitch range, which are frequently cited as salient phonetic
markers of gender performance but are understudied in this area with regard to
language in media. By analysing the way that changes in masculinity structures
are reflected in performance of fictional protagonists, this paper demonstrates
the necessity for further research on language use, particularly by protagonists,
in fictional media, as well as on the way that dominant language ideologies are
reproduced and consumed in the popular culture market.
Introduction
Despite its constant presence in modern culture, language use in fictional
media and its acoustic properties are understudied as subjects of linguistic
inquiry, largely due to the fact that early sociolinguistic research consid-
ered language produced outside of a ‘natural’ context to be inauthentic
Affiliation
Ohio State University, USA.
email: dahlberg-dodd.1@osu.edu
In the early 1990s, the Japanese economy suddenly declined, causing the
previously hegemonic sarariiman (lit. ‘salaryman’) model of masculinity
brought on by the last several decades of success in the financial quarter
to fall in popularity as job security and hiring rates also fell (Dasgupta
2000:199). While there has been significant work done on recent mas-
culinities, particularly as it concerns the salaryman model as a gendered
construct (e.g. Dasgupta 2000, 2005, 2013; Hidaka 2010; Deacon 2013;
Smitsmans 2015), there has been little published work on language ide-
ologies in this area as they interact with media, particularly fiction. The
bulk of research on language and gender has been dominated by studies
of female, Standard speakers, and many of these studies have been based
on self-reflection rather than data-collecting, but in the last fifteen years,
more studies of men’s speech have emerged (e.g. Miyazaki 2004; SturtzS-
reetharan 2004; Saito 2012).
In analysing nearly 40 years of shōnen protagonists, we can see a change
in speech patterns that corresponds with this late-twentieth-century shift
in masculinity, resulting in a greater characterological and linguistic range
for the portrayal of teen and young adult male characters in media. To
put it more generally, language use in Japanese popular media reacted to,
rather than directly affected, the societal construction of masculinities, a
phenomenon that is similarly reflected in the language use of female char-
acters in manga in Unser-Schutz (2015).
Through a phonetic analysis of the average pitch (f0) and pitch range
of protagonists from forty years of shōnen anime, as well as a text-based
analysis of their first-person pronoun choices, this study aims to illustrate
not only the ways in which main character figures are designed for their
audiences, but also bring to the light the nature of the ideological con-
versation between a work’s author and its audience through those main
character figures. Additionally, by approaching the first-person pronouns,
one of the most salient markers of one’s gendered persona in Japanese, in
conjunction with the f0 behaviour of protagonists, I hope to contribute to
a more thorough understanding of masculine ideology and its phonetic
manifestations in contemporary Japanese media.
of the novel of Gone with the Wind by Margaret Mitchell (1957) and the
Japanese translation of the English movie script by Sidney Howard (1994),
Nakamura (2007) illustrates how the linguistic ideologies as related to
hegemonic masculinity have changed during the twentieth century. Below
in Examples 1 and 2 are Japanese translations of the same sentence, spoken
by the main male character Rhett Butler to Scarlett O’Hara, the protago-
nist. The accompanying gloss is not a reflection of the original sentence
in Gone with the Wind, but rather than attempt to capture the tonal dif-
ference between the two in Japanese. The annotation was done using the
standard put forth by Shibatani (1990), while all bolding and underlining
are my own.
Example 1
(Mitchell 1957, as cited in Nakamura 2007)
Example 2
(Howard 1994, as cited in Nakamura 2007)
only the innocuous odoru (‘to dance’). Finally, when speaking of women,
Example 1 utilises the word josei, which is neutral in terms of politeness,
while Example 2 uses onna, which when used on its own can carry more
sexualised, often derogatory connotations due to its usage as a euphemism
to refer to sex workers as evidenced in onna’s dictionary entry.
Of particular note to this study, however, are the first-person pro-
nouns boku and ore in Examples 1 and 2, respectively. According to Kinsui
(2003:92), in a given fictional setting, the protagonist1 typically has the least
marked speech style (i.e., most faithful to ‘Standard’) unless there exists
a narrative reason dictating otherwise. In the first half of the twentieth
century, male protagonists typically used the first-person pronoun boku,
which according to Kinsui, gives the impression of an educated, upwardly
mobile student (Kinsui 2003:124). This usage was particularly common
in shōnen works, where it remained until the 1950s, suggesting that the
expected protagonist of a shōnen work has an interest in education and
social mobility. In the mid-twentieth century, probably beginning with the
shōnen manga Kyojin no hoshi (1966) and Ashita no Joe (1968), it became
increasingly common for protagonists to use ore as not only audience
expectations for the preferred, idealised ‘hero’ changed, but more impor-
tantly, the hegemonic masculine ideal had begun to shift (Kinsui 2003:124).
During this period, there was an increase in desire for the ‘hot-blooded
hero’, an aggressive, no-nonsense character (Nakamura 2007:64–5). At the
same time, though characters that used boku characters did not necessar-
ily disappear altogether, at least not all at once, the personality attribute
‘hot-blooded’ was not available to its users. Boku and ore exist in relation
to what Eckert (2008:454) has dubbed the ‘indexical field’, a ‘constella-
tion of ideologically related meanings, any one of which can be activated
in the situated use of the variable’. In other words, boku and ore exist in
different, but overlapping, indexical fields. However, though boku and ore
could index some overlapping qualities, boku lacked the ability to index the
loose-cannon, ore emulated. To put it differently, simply being educated
and upwardly mobile was not sufficient to embody the kind of idealised
masculinity that had captured the cultural imagination of male viewers and
fiction-oriented content producers during the 1950s and 1960s.
While salaryman masculinity is by no means the only masculinity that
could conceivably utilise the first-person pronoun ore, the ideological
connection between ore and salaryman image can still be seen in media
aimed at salarymen. For example, though the average salaryman on the
job utilises either watashi (a first-person pronoun that can be utilised in
formal situations regardless of gender) or boku (Kinsui 2003:127–8; see
also Miyazaki 2004), the inner-thoughts of salarymen can often be seen
352 Hannah E. Dahlberg-Dodd
portrayed using ore. One example is from the yearly Sarariiman Senryū
competition hosted by life insurance company Dai-Ichi Life. The competi-
tion, which has run every year since 1987, is a way for salarymen to submit
comical haiku-style poems (known as senryū) about their daily lives. Not
only do the submissions available on the website use exclusively ore, but the
advertisements on the website portray caricatured salarymen also using
ore.2 This is due to the fact that while salarymen are beholden to speaking
according to certain politeness registers that may demand either watashi
or boku, ideologically there exists a connection between the personality
traits associated with ore and the societal image of the salaryman. In other
words, ore is in an ‘indirect indexical’ (see Ochs 1992) relationship with
the salaryman image, but its situated use in an advertisement (such as for
Sarariiman Senryū) is enough to call upon this image if used in conjunction
with any other features of what could be called the ‘salaryman indexical
field’, such as reference to the workplace, company hierarchical structures,
and so forth through certain lexical choices.
The mid-1990s was met with a less than ideal economic situation. Known
as the ‘Lost Decade’, the Japanese economy underwent a downturn driven
by a bursting of the speculation-driven ‘bubble’ of the early 1990s. This
resulted in slowed economic growth and economic downsizing, effectively
throwing a wrench in the idealised school-to-white-collar-employment
path (Dasgupta 2013; Deacon 2013). The economic downturn was particu-
larly evident among younger generations; in the 1980s, companies recruited
heavily from high schools and universities, and with less positions avail-
able for new employees due to company restructuring and lay-offs in the
90s, resulting in particularly high employment among younger generations
(Hidaka 2010:105–8). The breakdown in this pipeline, according to Deacon
(2013:159), ultimately caused those who had subscribed to the salaryman
model of masculinity to question ‘their very identities as fathers, as provid-
ers, and as men’ (italics in original). That being said, salaryman masculin-
ity did not cease to be the hegemonic model. Regardless of any losses in
legitimacy and authority caused by the bubble burst (Roberson and Suzuki
2003:9), as stated in both Dasgupta (2013) and Smitsmans (2015), the sala-
ryman model still ultimately remains the ideal despite emergence of ‘sub-
cultural masculinities‘ such as ‘herbivore men’ (sōshokukei danshi), the
antithesis the salaryman’s ‘carnivore man’ (nikushokukei danshi), into the
mainstream (Smitsmans 2015:19). This is consistent with Connell (2005)’s
description of hegemonic masculinity as not necessarily the most wide-
spread in terms of practice, but as a unity of the most ‘favourable’ traits
(Connell 2005:79; Smitsmans 2015:9).
Voices of the hero 353
are still scarce (but see Hiramoto 2010, 2013; Unser-Schutz 2015; Maynard
2016), probably due to the same perception mentioned earlier that fiction-
alised language is unworthy of linguistic study (Djenar 2015). Manga, in
particular, and by extension anime, hold positions as particularly low-brow
media that do not require ‘real’ reading, a stance that may have contrib-
uted to its relatively low amount of serious scholarly consideration com-
pared to other fictional genres in the Japanese diaspora (Ito 2000). But, as
pointed out by Agha (2007:151), language as it appears in fiction is at its
core ‘metadiscursive’, giving linguists insights into the processes of, as is
relevant for this paper, the ideological processes involved in the creation of
the characterological figure of the protagonist, particular with an audience
of adolescent boys in mind. This study intends to (i) help address the gap in
literature concerning language and gender in fictional media, while simul-
taneously (ii) moving towards an understanding of the linguistic qualities
that make one a ‘hero’ (or in some cases, an ‘anti-hero’) in an adolescent
male-targeted fictional work.
Anime began to be adapted from Shōnen Jump shortly after the maga-
zine’s inception, with the first anime (Otoko ippiki gaki daishō) airing
in 1969. Shueisha, the publishing company that owns Shōnen Jump, has
little to do with its animated adaptations at a production level, however.
Works that are adapted from manga that run Shōnen Jump are managed by
numerous different companies, but because these works are likely to retain
the same target audience due to their roots in Shōnen Jump and the fact
that few major changes are made in their animated adaptations, it is most
appropriate to think of these as Shōnen Jump anime rather than necessarily
as distinct products of their various animation companies.
Variables
The variables under investigation in this study are (i) first-person pronouns,
(ii) average f0, and (iii) average range of f0. In Japanese, first-person pro-
nouns, along with sentence-final particles, are typically cited as particularly
salient with regard to gender indexing in a given utterance (SturtzSreeth-
aran 2004; Unser-Schutz 2015). Sentence-final particles, however, regularly
change from utterance to utterance (especially in non-fictional discourse),
depending on the speaker, as well as context, thus an investigation of their
role in masculinity would be better suited by either a corpus-based study,
discourse analysis study, or some combination of the two, putting this
particular facet of identity construction outside the scope of this study
(see Unser-Schutz 2015; SturtzSreetharan 2004; Ueno 2006). First-person
pronouns, however, are much slower to change diachronically, due to
the fact that personal pronouns are generally considered to be the most
salient index of gender (or gender-based performance) in Japanese (Abe
2004; Unser-Schutz 2015). Additionally, even more so than sentence-final
particles, first-person pronouns carry salient gender indexing informa-
tion, due in large part to the traditional prescriptive approach to pronoun
usage. According to Ide (1997) Standard Japanese pronouns vary primarily
among two axes: gender and formality register, which can be evidenced in
Table 1 below.
Masculine Feminine
Gendered ideologies ore/boku watashi/atashi
Students’ patterns ore boku uchi atashi
wara (2003, 2007) is a notable exception in her study of the voice quality of
protagonists versus antagonists across a number of different anime. While
her study focused more on voice quality than f0, she did note that the
average f0 of the fictional characters in her study was considerably higher
than averages reported by studies on non-fictional speakers. In her study,
adult male heroes had an average f0 of 191.8 Hz, while male child heroes
had an average of 342.5 Hz (Teshigawara 2003:50). Instead of range of f0,
Teshigawara (2003) reported on standard deviation of average f0, which
was 47.7 Hz and 84.5 Hz, respectively.
In light of the observations by Nakamura (2007) and Kinsui (2003)
regarding the boku to ore shift across the early and mid-twentieth century
period with the shift in hegemonic masculinity, this study aims to explore
whether or not a similar shift in first-person pronoun usage can be seen in
the late 1990s with the economic downturn. Additionally, because recent
media gives us access to a vocal performance of these masculinities rather
than only textual evidence of their usage, the average f0 and f0 range are
also under consideration due to their perception as salient indicators of
gender performance.
Methods
A total of 78 different anime series have been adapted from Shōnen Jump
manga between 1968, the year of Shōnen Jump’s first issue, to 2015, when
these data were collected. The protagonist of each work was determined
based on the work’s reported protagonist (designated by the word ‘shujinkō’).
However, because this data set combines an analysis of pronoun usage as
well as an acoustic analysis of the protagonists’ speech, the works analysed
had to be limited based on a number of factors. The principles for exclusion
are listed below with the number of works that were excluded for that reason.
First-person pronouns
For first-person pronouns, the pronoun used by the character was recorded.
Any deviation was noted, though within this data set, characters used only
a single pronoun and did not deviate. This is contrary to data taken from
speakers outside of fiction, where a Standard Japanese speaker may change
five times during a conversation (Miyazaki 2004), but more likely changes
pronoun usage according to context of formality.
Results
First-person pronouns
Across the thirty-five total years of analysis, ore was by far the most preva-
lent first-person pronoun used by protagonists, and it was used across age
groups. The second most common, boku, was used almost predominately
by teens, which is in opposition to the ideological association of boku
in casual speech with children (Nakamura 2007). Other pronouns that
appeared were ora (n = 1) and oira (n = 1), both of which are associated
with a ‘country’ archetype, sessha (n = 1), which is an out-of-use pronoun
used by warrior characters, watashi (n = 2), and washi (n = 1), which is
associated with old men.
f0 measurements
A total of 1060 intonation phrases were analysed for average f0, minimum
f0, maximum f0, and f0 range. The average length of the intonation phrases
in this data set is 1.35 seconds.
Because this is an analysis of f0 qualities, it is worth noting before enter-
ing the analysis that the protagonists themselves were voiced by both male
and female voice actors. This particular point is also brought up Teshiga-
wara (2003)’s analysis of the voice qualities and f0 of both male and female
heroes and villains in anime, and as with Teshigawara (2003), the gender of
the voice actor did not factor in to the final analysis. This is because though
male and female speakers tend to utilize different average f0 and average
pitch ranges (e.g. Yuasa 2008; Teshigawara 2003) the actors performing
these characters are not speaking in a day-to-day manner. Rather, they are
performing the role of a male character of a specified age. As pointed out by
Unser-Schutz (2010:406) with regard to manga, ‘a certain level of realism’ is
necessary for the audience to enjoy it. In other words, a vocal performance
that creates a disconnect between the purported age and gender of the
protagonist and the audience’s expectations of that vocal performance can
disrupt immersion into the work on the part of the audience.
Each of these voice actors (n = 47) are performing male protagonists that
are suited to animated works that have been adapted from Shōnen Jump
Voices of the hero 361
F p
1990
mean f0 37.164 5.585 × 10–09
max f0 42.298 6.256 × 10–10
min f0 26.665 5.872 × 10–07
range 26.295 6.953 × 10–07
2000
mean f0 19.719 1.216 × 10–05
max f0 14.363 1.784 × 10–04
min f0 22.328 3.377 × 10–06
range 5.5594 0.01895
2010
mean f0 4.3568 0.03793
max f0 9.0548 2.902 × 10–03
min f0 2.3377 0.1276 (n.s.)
range 6.2286 0.01325
Discussion
Based on the statistically significant difference between average f0, minimum
f0, maximum f0, and f0 range, we can see that f0 indeed plays a role in the
framing of an ore-using protagonist versus a boku-using protagonist. That
being said, exactly what ‘role’ this is seems to change depending on the
decade, as boku’s f0 measurements in relation to ore’s vary, and the difference
between them decreases, yet retains its significance, in the most recent data.
First, however, it is illustrated that during the 1980s, there was not a par-
ticular range or pitch that was necessarily associated with exclusively ore,
as indicated in Figures 2 and 3. As seen in Figure 1, ore was overwhelmingly
the most frequent pronoun, and it was utilised by protagonists with both
Voices of the hero 363
Figure 2: Combination violin and box-and-whisker plot of average f0 for each pro-
noun by decade.
Figure 3: Combination violin and box-and-whisker plot of f0 range for each pronoun
by decade
364 Hannah E. Dahlberg-Dodd
high voices and low voices, as well as large ranges and small ranges, sug-
gesting that the first-person pronoun was not an immediately salient part of
the linguistic (and physically) ‘bricolage’, or stylistic practice, that made up
the shōnen hero (Hebdige 1984; Eckert 2008). Rather, ore worked together
with numerous factors, such as that particular character’s f0 range, average
f0, as well as sentence-final particles, to structure the linguistic semiotic
presentation of that particular shōnen protagonist. To reiterate the earlier
quote from Kinsui (2003:66–7), the protagonist of a given work uses the
least-marked speech style in its fictional environment unless a narrative
reason justifies otherwise, but what constitutes as ‘least-marked’ varies
with not only the demographic of the portrayed protagonist, but also the
writers of that work as well as the intended audience. For the writers and
audience of the 1980s, the ‘hot-blooded hero’, in whatever manifestation it
appeared, was the least-marked protagonist in works created by a nearly
exclusively male staff for a young male audience.
With the reintroduction of boku in the 1990s, we can see that expectations
for protagonists in shōnen works changed as the power of masculinity struc-
tures that were dominant during the 1980s began to weaken. This suggests
that not all of the qualities that were desired in protagonists as of the mid–
1990s were able to be portrayed with only ore as an available first-person
pronoun. Boku, however, which was still in regular use in natural speech
but not used in Weekly Shōnen Jump for either protagonists or side charac-
ters within this data set, seemed to serve as a viable masculine alternative.
Though boku had been the pronoun of choice for male protagonists shōnen
and otherwise in the early twentieth century, it was in the 1960s that audi-
ences and creators moved away from the upwardly mobile, educated image
of boku to the ‘hot-blooded hero’ image of ore (Kinsui 2003:124). But, with
the reintroduction of this pronoun, not only can we see a movement away
from exclusively hot-blooded heroes, but we can understand boku as a kind of
Bakhtinian ‘chronotope’, invoking in its usage the qualities and traits of early
twentieth century protagonists that existed prior to the economic downturn
(see Bakhtin 1981; Blommaert 2015). This chronotopic quality can be seen
in the vast difference in acoustic qualities between ore-users and boku-users
when boku was first reincorporated as illustrated in Table 6 above (p < 0.000
for all acoustic measurements). Over time, however, as boku continued to
be utilised in a consistent 30 per cent of appearing shōnen protagonists, the
acoustic differences between ore-users and boku-users declined, suggesting
that through its continual use, boku gradually began to lose its chronotopic
qualities as it became acoustically renaturalised into the shōnen genre. Thus,
as its acoustic performance merges with that of protagonists using ore, both
boku protagonists and ore protagonists are allowed to coexist as unmarked
fictional male voices.
Voices of the hero 365
Conclusion
Through this diachronic analysis of f0 and first-person pronoun usage in
Weekly Shōnen Jump, we are able to see a relationship between mainstream
masculine ideologies and the linguistic construction of male protagonists
in a male-targeted media genre. During the 1980s, ore was the first-person
pronoun of choice for protagonists in the shōnen genre, but in the 1990s
the previously hegemonic variety of masculinity began to weaken. This
resulted in a loss of dominance for the ore-using ‘hot-blooded hero’, a shift
that is evidenced by the reintroduction of the first-person pronoun boku,
which had fallen out of use in shōnen works in the 1960s and denotes an
upwardly mobile, educated protagonist (Kinsui 2003:124).
At its reintroduction, boku protagonists inhabited an overlapping, but
different acoustic space than ones using ore, with highly significant differ-
ences in terms of average f0, maximum f0, minimum f0, and f0 range. This
difference suggests that though boku occupied 30 per cent of the protago-
nists that appeared following 1996, through its use during an unsteady time
for hegemonic masculinity structures, it functioned as a kind of Bakhtin-
ian chronotope, harkening back to a period before the destabilisation of
current hegemonic masculinity through economic turmoil (Bakhtin 1981;
Blommaert 2015). However, the decrease over time in acoustic difference
between protagonists using boku versus those using ore suggests that it is
renaturalising as a viable masculine alternative to the ‘hot-blooded hero’
style.
As it stands, language use in media, particularly its phonetic proper-
ties, remains understudied in the area of sociocultural linguistics, though
as of the twenty-first century, this has begun to change (Stamou 2014).
Its necessity, however, is demonstrated in the fact that media, particularly
fictional media, has become an increasingly prominent figure in our social
environments. Further study of linguistic and metalinguistic practices with
regard to the construction of characterological figures can only enrich our
understanding of the way that speakers interact with variously meaningful
ideologies, as well as the way we choose to construct and consume our
linguistic landscapes.
Notes
1 Although Rhett Butler was not the protagonist of Gone with the Wind, as pointed
out by Nakamura (2013), dominant gender ideologies are actually more salient in
translated works than works originally created in Japanese.
2 The website for the Minna no Sarariiman Senryū project is available at http://event.
dai-ichi-life.co.jp/company/senryu (accessed 20 November 2016).
3 One exception was made to this in the case of a show which had a poorly reviewed
original run, and it was remade within two years with a different voice actor with a
run of several hundred episodes. In this case, the remake was considered.
References
Abe, H. (2004) Lesbian bar talk in Shinjuku, Tokyo. In S. Okamoto and J. S. Shibamoto-
Smith (eds) Japanese Language, Gender, and Ideology: Cultural Models and Real
People 205–21. Oxford: Oxford University Press. https://doi.org/10.1525/jlin.2005.
15.1.38
Agha, A. (2005) Voice, footing, enregisterment. Journal of Linguistic Anthropology
15(1): 38–59.
Agha, A. (2007) Language and Social Relations. Cambridge: Cambridge University
Press.
Androutsopoulos, J. (2012) Repertoires, characters and scenes: sociolinguistic differ-
ence in Turkish–German comedy. Multilingua 31(2/3): 301–26. https://doi.org/
10.1515/multi-2012-0014
Bakhtin, M. M. (1981) Forms of time and chronotope in the novel. In M. Holquist (ed.)
and C. Emerson and M. Holquist (trans.) The Dialogic Imagination: Four Essays
84–258. Austin, TX: University of Texas Press.
Bell, A. (2001) Back in style: reworking audience design. In P. Eckert and J. R. Rickford
(eds) Style and Sociolinguistic Variation 139–69. New York: Cambridge University
Press.
Blommaert, J. (2015) Chronotopes, scales, and complexity in the study of language in
society. Annual Review of Anthropology 44: 105–16. https://doi.org/10.1146/annu
rev-anthro-102214-014035
Boersma, P. and Weenink, D. (2016) Praat: a system for doing phonetics by computer
(version 6.0.16). Retrieved from www.praat.org.
Camp, M. (2009) Japanese lesbian speech: sexuality, gender identity, and language.
Doctoral dissertation, University of Arizona, USA. Retrieved from http://arizona.
openrepository.com/arizona/handle/10150/195371.
Connell, R. W. (2005) Masculinities. Cambridge: Polity Press.
Dasgupta, R. (2000) Performing masculinities? The ‘salaryman’ at work and play. Jour-
nal of Japanese Studies 20(2): 189–200. https://doi.org/10.1080/713683779
Dasgputa, R. (2005) Salarymen doing straight: heterosexual men and the dynamics of
gender conformity. In M. McLelland and R. Dasgputa (eds) Genders, Transgenders,
and Sexualities in Japan 168–82. Oxford: Routledge.
Dasgupta, R. (2013) Re-reading the Salaryman in Japan: Crafting Masculinities.
London: Routledge.
Deacon, C. (2013) All the world’s a stage: herbivore boys and the performance of mas-
culinity in contemporary Japan. In B. Steger and A. Koch (eds) Manga Girl Seeks
Voices of the hero 369
Yomiuri Shimbun Morning Edition (2003) Shōnenshi sakuhin: joshi nimo ninki. Yomi-
uri Shimbun Morning Edition (16 June).
Yuasa, I. P. (2008) Culture and Gender of Voice Pitch: A Sociophonetic Comparison of the
Japanese and Americans. London: Equinox Publishing.