character voice in anime subtitles peter howell,...

CHARACTER VOICE IN ANIME SUbTITLES

Peter Howell, Hiroshima Prefectural University, [email protected]

AbstractOne of the main functions of dialogue in film narrative is the construction of character voice, which is partly realized through the skillful use of the sociolinguistic and pragmatic resources of language. In the English subtitling of Japanese animation, there are however two barriers to recreating this function. The first barrier is the major pragmatic and sociolinguistic difference between Japanese and English, and the second consists of the technical constraints of subtitling as a mode of translation. The article describes from a textual point of view how subtitlers have addressed the stylistic problem of character voice, focusing on the work of Neil Nadelman in the film Grave of the Fireflies (“Hotaru no Haka”) and in single episodes from two TV series, Revolutionary Girl Utena (“Shōjo Kakumei Utena”) and Slayers Try (“Sureiyāzu Torai”), and on the work of David Fleming in the film Akira (“Akira”). For all four works, another commercially available subtitled version, either in English or French, is considered for purposes of comparison. Comparison suggests that strategies vary with regard to character voice, but that the decimation of stylistic function is not inevitable in subtitling as a mode of translation.

Key words: Anime subtitles; character voice; compensatory procedures.

IntroductionSarah Kozloff’s study, Overhearing Film Dialogue (Kozloff 2000), indicates in its title the dual-layered pragmatic context of dialogue in film. Within the inner pragmatic layer of the film, dialogue is addressed primarily to other characters, but in the context of the outer layer it is addressed to an “overhearing” audience to fulfill narrative functions within the overall semiotic design of film as a work of art. Kozloff outlines the functional nature of dialogue in film, and highlights the importance of these functions in view of the fact that film is often regarded as a predominantly visual medium. Among the chief functions outlined is characterization:

Each time a character opens his mouth, filmgoers learn more about him – is his accent “upper class” or “hillbilly”? Is he or she polite? brusque? thoughtful? quick? lazy? (Kozloff 2000: 43).

Manfred Pfister explains that characterization can also be non-verbal (facial expression and costume) and may be explicit (e.g. in commentary by other characters) (Pfister 1988). But writers, whether in film, drama or novels, typically seek to differentiate between characters by giving them a particular character voice as a kind of implicit badge of identity:

…the recurrence of certain stylistic features delineates the contours of the figure’s identity and distinguishes him from the other figures. (Pfister 1988: 148)

The stylistic features often involve the writer making use of the diverse varieties of language found within speech communities, including idiolect, dialect, sociolect and register. In Japanese anime and manga, writers make use of the connotations of language varieties to create a ‘virtual’ yakuwari-go (“role-language”) and bring to life stereotypical characters such as the kindly old

292

0907-676X/06/04/292-14 $20.00Perspectives: Studies in Translatology

© 2006 Peter HowellVol. 14, No. 4, 2006

scientist, the cosseted young rich girl, the wise-cracking Osakan etc. (Kinsui 2003).

As language varieties are closely linked to particular cultures and societies, they present special problems for translators of dialogue in verbal art. Japanese language and society, in particular, diverge widely from language and society in America, Britain and other English-speaking countries, and this creates a significant barrier to attempts to achieve functional correspondence of character voice in translation. The challenge is compounded in the case of subtitling, which is a mode of translation constrained by technical factors: in particular the need to reduce the number of words in the target text in comparison to the original.

Linguistic barrier: sociolinguistic and pragmatic differencesOne of the most difficult problems for translators is geographical dialect. In the target language it is not possible to reproduce dialect’s function of implicitly indicating the regional origins of a character (Levý 1969: 101). The Japanese local colour of, for example, Kōbe dialect cannot be directly reproduced in another linguistic culture. But in functional translation some procedure should ideally be sought to recreate the general connotations conveyed by the dialect in the source culture. Although in prior eras the substitution of source-language dialect by target-language dialect was not uncommon (Czennia 1992: 113), contemporary norms in literary translation firmly reject that procedure. Japanese-English literary translators have described it as “ludicrous” (Keene 2001: 328) and “ridiculous” (Seidensticker quoted in Richie 2000: 77). However, this interdiction on dialect substitution seems to be less absolute in the popular culture world of anime dubbing and comic-book translation. In its homepage liner notes to the TV series Macross (episode 12), the anime distribution company Animeigo (2006) explains that it “has traditionally translated the Osaka-ben dialect into Southern English”. In the translation of the SF manga Neon Genesis Evangelion, a stereotypical Brooklyn dialect is used for a character speaking Osaka dialect. An alternative translation procedure is to use standard, non-regional language in the target text, but a style level that is marked for spoken language. Herbst (1994: 108-109) reports that this is a common procedure in dubbing into German and can recreate the connotations of social intimacy, directness and emotion conveyed by regional dialect. If, however, spoken style is extended to include non-standard and slang words, this risks producing an effect demeaning to a fictional character, quite different from that of the dialect in the original text (Schogt 1988: 119; Schreiber 1993: 211).

The problems of translating regional dialect would seem essentially the same for all language pairs. But pragmatic aspects of Japanese constitute a particular difficulty for the English translator because of the fact that linguistic resources used in Japanese to express pragmatic meanings often have no directly corresponding counterparts in English, and that indeed Japanese socio-cultural structures – social hierarchies, in-group out-group membership, gender construction etc. – are significantly different from those in English-speaking countries. One such pragmatic difficulty is presented by sentence-final particles, some of which are used to “express the speaker’s emotion or attitude toward the hearer” (Makino & Tsutsui 1986: 45), and which may – for instance in the case

Howell. Character Voice in Anime Subtitles 293

of the particles ze and zo – be strongly associated with informal male speech. Such particles do not exist in English, a fact which can cause them to be labelled by language teachers as untranslatable (e.g. Kawashima 1999). Japanese terms of self-reference and address are also very different from those used in English. When addressing someone outside one’s intimate group or of higher status, it would be normal to use the addressee’s family name + the address suffix-san, while the use of name + the address suffix -sama represents an instance of honorific language elevating the addressee to a yet higher status. Conversely, the term of address omae (“you”) is socially acceptable among males of the same social status and in-group, but is rude in other contexts where it may convey aggression and contempt. Furthermore, the non-standard pronunciation omē is considered still rougher and more aggressive. Younger Japanese people, such as schoolchildren, students and junior company employees, may often simply use the term senpai (“senior”) to address those older than them in a particular social hierarchy.

The recreation of these distinct pragmatic aspects of character voice presents a major challenge when translating represented speech from Japanese into English. Three types of overall strategic approach may be adopted in the face of this challenge. A foreignizing strategy could conceivably seek to convey something of the nature of the Japanese social system by simply transferring address terms such as senpai and suffixes such as -san directly from Japanese. Although rare in commercial translations, this is a strategy not infrequently found in non-commercial fan translations. It risks however creating an exotic stylistic effect totally absent from the original dialogue. A neutralizing strategy would simply omit the social meaning encoded in the Japanese dialogue by the use of an unmarked style in English. It results in stylistically flat, bland dialogue. A domesticating strategy, on the other hand, would have recourse to compensatory procedures using a variety of sociolinguistic and pragmatic resources of the target language which, although they are not necessarily equivalent, recreate a corresponding stylistic effect. A heavily domesticating strategy can make the characters sound distinctly American, which may potentially be jarring for more purist fans of anime as a specifically Japanese genre, or for non-American audiences, particularly if the setting of the anime is Japan itself.

Modal barrier: the constraints of subtitling Academic literature on subtitling has stressed the technical constraints of subtitling as a mode of translation. Titford (1982), noting the physical limitations of space and time imposed by technical factors, coined the term “constrained translation” for subtitling. De Linde (1995: 9) explains that subtitlers have to “minimize the loss of image” and also ensure that titles can be read by the viewers. If the title is too long and appears on the screen for too short a time, it may not be able to be read by the average viewer. For the case of intralingual subtitling in Britain, de Linde & Kay (1999: 11) report the Independent Tele-vision Commission’s conclusions that average adult reading speed is 66% of speaking speed and so subtitles should be reduced by one-third in comparison to audio dialogue. Dollerup (1974: 202) outlines three procedures used to shorten subtitle texts in comparison to original dialogue: omission (of words

2006. Perspectives: Studies in Translatology. Volume 14: 4294

and dialog sequences unnecessary for viewers’ comprehension); concentration (of longer phrases into a few words); and fusion of several utterances into one. De Linde (1995) reports Irena Kovačič’s findings that linguistic items carrying interpersonal and textual functions are omitted more readily than carriers of the representative function, and goes on to describe further the types of reductions in subtitles in comparison with the original dialogue. Her description is similar to that made by Dollerup, distinguishing between total reductions (deletion) and partial reductions (condensations or paraphrases). However, de Linde breaks new ground by suggesting that patterns of reduction may depend on the type of film, many more reductions occurring in the English subtitles for the fantastical French comedy Delicatessen than in the more written-style dialogue of La Lectrice.

Another constraining aspect of subtitling as a type of translation relates to the shift from the oral-auditory channel of communication to that of the written channel. This shift has been highlighted by Gottlieb, who points to what he calls the diagonal, diasemiotic nature of subtitling, in that it “crosses over” from one channel to another (Gottlieb 1995, 1998). As Fawcett points out, using the example of 1980s BBC subtitling, there may be different tolerance norms for taboo language depending on whether the taboo item is delivered in the oral-auditory channel or in written (subtitled) form (Fawcett 1983: 187). Nedergaard-Larsen (1993: 213) is another commentator who suggests subtitles should in some cases be “toned down” because of the stonger effect of written forms.

In view of the constraints outlined above, it is perhaps not surprising that subtitling is often seen strategically in terms of certain functional deficits. With regard to indices of character, Marleau suggests that subtitles may fail to fulfil the same emotive function as the original dialogue:

Le texte écrit véhicule l’affectivité de celui qui parle: ce qu’il manifeste ordinairement grâce au débit, à l’intonation et au rythme. Le fait que ce dialogue subisse, dans bien des cas, une compression, pourrait être un obstacle pour éveiller ces émotions. [The written text is a vehicle for the emotions of the person speaking: usually manifested by speed, intonation and rhythm. The fact that this dialogue undergoes a compression in many cases may be an obstacle to conveying these emotions.] (Marleau 1982: 274, translation added)

Becquemont (1996: 153) writes that cinema subtitles sometimes only provide a skeleton of information to accompany the visual elements of films, and stresses their “aspect sec et schématique, centré sur la fonction référentielle” [“dry and schematic aspect, centred on the referential function”]. Mason (1989) and Hatim & Mason (1997) emphasize that one of the most important aspects of reductions in subtitling has to do with interpersonal pragmatics, in particular the omission of features of politeness such as the use of boosters and down-toners. The opinion is expressed that reductions are likely to have a significant effect on the narrative function of conveying the personality and interrelationships of the characters:

Cumulatively, the absence of the politeness features which we have discussed cannot fail to convey a different idea of the personality of the characters on screen and of their attitudes towards each other. Items which, taken in isolation, appear to be of relatively little significance turn out to be powerful indicators of rhetorical purpose and the dynamics of interaction. (Mason 1989: 24)


Lorin Card’s study (Card 1998) focuses on the capacity of subtitles to attain a “balanced equivalence” in respect to the register of the original dialogue and its connotations. Investigating English subtitles for two French films, Au revoir les enfants and 37º2 le matin (“Betty Blue”), Card, like other commentators, notes the tendency of subtitles to attenuate the orality of the original dialogue, particularly in the case of taboo elements. However, it is contended that equivalence is possible if the subtitler uses an English register that corresponds to the French original. Card also points out the possibility of graphical compensation for the switch to the written channel in subtitles: both by devices of spelling, as in the orthographic representation of a German accent (e.g. “Your jam ‘ist konfizkated’!”) and by typographic devices as in the use of capitalized words or bolder typeface. Finally, Card also presents some examples in which, far from being diluted, the connotational impact of the English subtitles is stronger than the French dialogue.

Neil Nadelman: Grave of the Fireflies Grave of the Fireflies (“Hotaru no Haka”) is an animated feature film directed in 1988 by Takahata Isao, based on a story by novelist Nosaka Akiyuki. Set in the Western Japanese city of Kōbe during the closing stages of the Pacific War, it tells the sad story of two children, Seita and his younger sister Setsuko, whose mother is killed in an American bombing raid and who go to stay with an unsympathetic aunt in the countryside. They leave their aunt’s house to live by themselves in an unused bomb-shelter, but eventually die from the accumulated effects of malnutrition just after the Japanese surrender in 1945. It gives a sentimental yet powerful reminder of the terrible suffering that the war brought to many Japanese people, including children. Two sets of subtitles were investigated for the purposes of this article. The main focus of interest are the subtitles produced for the American market on the Central Park Media DVD released in 1998, and translated by Neil Nadelman (labelled #1ES in the examples). These differ quite substantially from the English subtitles on the Warner Home Video DVD released for the Japanese market in 2000 (labelled #2ES in the examples). The translator of the latter set of subtitles is uncredited. It should be borne in mind that English subtitles on DVDs sold in Japan are often intended for language learners and that therefore a simpler, ‘neutralised’ translation is by no means inappropriate for the target audience in this case.

In Grave of the Fireflies, almost all the characters use dialectal Japanese. The narrative function of the dialect is not necessarily to develop character stereotypes or character relationships, but to provide local colour. Nadelman cannot reproduce the local colour effect of the dialectal Japanese. He does, however, often translate using a spoken-style level that retains some of the social intimacy suggested by dialect. By contrast, the unnamed translator for the Warner Home Video DVD (on sale in Japan) renders Japanese dialect into stylistically flat, simple English dialogue. One translation procedure employed by Nadelman is the orthographic device of pronunciation spelling. Pronunciation spelling can be defined as “a spelling that is supposed to represent a pronunciation more closely than a traditional spelling, as lite for light, wanna for want to” (American Heritage Dictionary 4th edition, 2000). In the following examples, Setsuko’s speech is marked as local dialect by, amongst other things, her use of the non-


standard question particle nen rather than standard no in the first line, and her use of the non-standard copula ya rather than standard da: [bold print in all examples added by the author for highlighting purposes]

(1) Setsuko Nani shiton nen? (J) What are you doing? (gloss) Whatcha doing? (#1ES) What is it? (#2ES)

(2) Setsuko Chūsha iya ya (J) I don’t like shots. (gloss) Don’t wanna shot! (#1ES) I hate shots! (#2ES)

The choice of lexical register typical of informal style is another way in which Nadelman’s subtitles contrast with those in the Warner Home Video DVD, as is evident in examples of lines spoken by a variety of incidental characters in the film. In example (3), from the pre-credits opening scene, the station attendant’s use of local dialect (itemai yoru de compared to standard itte shimau darō: “will soon die”) helps locate the action in Kōbe for Japanese viewers and adds local colour. Similarly, in the air raid scene immediately after the credits (example 4), dialect is used in incidental comments by a neighbour (den na compared to standard desu ne). Nadelman again cannot reproduce the local colour of the dialect, but can indicate the informal tone of the dialect by choosing English lexis belonging to spoken register (a goner, in for a big one). By contrast, the subtitles provided for the Japanese market have a stylistically neutral tone (dying, hundreds of bombers):

(3) Station attendant Kocchi no yatsu mo mō jiki itemai yoru de. (J) This guy also will soon die. (gloss) This one’s a goner, too. (#1ES) This one’s dying too. (#2ES)

(4) Neighbour Teki-san daihentai rashii den na. (J) Enemy (seems) like (a) big squadron. (gloss) We’re in for a big one! (#1ES) Hundreds of bombers. (#2ES)

This contrast in lexical register is found in the lines of other characters appearing throughout the film. In example (5), where a kindly naval officer uses the dialectal yarō as opposed to standard darō, Nadelman’s knocked around contrasts with beat, and in (6), where the dialectal copula ya is used rather than standard da, gobble down stands out in contrast to eat:


(5) Naval Officer Kondake nagurya ki ga sunda yarō. (J) If (you) beat (him) this much (you) should be satisfied. (gloss) You’ve knocked this poor kid around enough. (#1ES) You beat him enough. (#2ES)

(6) Aunt Sett-chan shiroi gohan ya to yō taberu nā. (J) Sett-chan if (it) is white rice eats a lot. (gloss) You’re just gobbling it down tonight, aren’t you Setsuko? (#1ES) You eat a lot tonight. (#2ES)

Neil Nadelman: revolutionary Girl Utena and Slayers TryTwo other sets of subtitles by Neil Nadelman were compared with alternative subtitles, although in this case alternative English subtitles were not available and so commercially produced French subtitles were considered. The works of anime were the first three episodes respectively of the animated TV series Revolutionary Girl Utena (“Shōjo Kakumei Utena”) (henceforth Utena) and Slayers Try (henceforth Slayers). Utena, produced in 1997 by TV Tokyo, tells the story of Tenjō Utena, a tomboy student at a fantastical private boarding school, and her relationship to the mysterious Himemiya Anthy, the “Rose Bride”. The show can be categorized in terms of genre as shōjo (“girls”) anime. A VHS-cassette with English subtitles by Neil Nadelman was released by the distributor Software Sculptors in 1998. A French subtitled cassette was released in 2000 by the Belgian-based distributor Dynamic Visions, with translation credited to “S.M” and adaptation to Charles Lewis. Produced in 1997 by TV Tokyo/SOFTX, Slayers Try is the third TV series of the Slayers franchise. Categorized by its American distributor Software Sculptors as comedy/adventure and by its distributor for the French-speaking world, IDP, as heroic fantasy, it continues the adventures of the teenage sorceress Lina Inverse and her friends Gourry, Zelgadis and Amelia, as they encounter a mysterious companion, Filia, and set sail beyond the Kingdom of Seyruun. A cassette subtitled by Neil Nadelman was released in 2000 and a cassette with French subtitles by an uncredited translator was released in the same year. Although French versions of anime and manga are sometimes translated using English as an intermediary language, in this case the French subtitles for both Utena and Slayers are based directly on the original Japanese.

Salient differences in translation procedures emerge from a comparative juxtapositioning of the subtitles in English (ES) and French (FS), specifically with regard to typography and spelling, interjections and terms of address. In Utena capitalization is used for emphatic effect on 14 occasions in English, but not at all in the French subtitles. And in Slayers it is found 21 times in English, but not in French. Using the typographic device of capitalization enables the American subtitler to convey the forceful personalities of Utena and Lina, indicated in examples (7) and (8) by their use of the emphatic sentence-final particle yo. The capitalization of HAVE and THAT compensates somewhat for the loss of sound in the diasemiotic translation from oral to written language. In the French translation of Utena’s line, the emphatic adverb vraiment (“really”)


may be considered a stylistic compensation using lexical means:

(7) Utena Iya kimi wa itta hō ga ii. Zehi iku beki da yo. (J) No, you should go. (You) definitely ought to go. (gloss) You HAVE to go to it! (ES) Tu devrais y aller! Tu as vraiment besoin d’amis. (FS)

(8) Lina Sonna koto de machi o osotta no yo!! (J) For that (you) attacked a town. (gloss) So you attacked a town to do THAT?! (ES) Et la ville, alors? (FS)

Nadelman, as in Grave of the Fireflies, also employs pronunciation spelling in Utena’s and Lina’s lines, helping to recreate their informal colloquial style, exemplified in the following two lines by sentence-final particles (wa and yo), contractions (ja nai rather than de wa nai), and lexical choice (yatsu [“guy”]). The French subtitles do not use spelling as a compensatory procedure in this way, instead making use of conversational syntax and lexis:

(9) LinaJōdan ja nai wa yo. (J)(It’s) no joke. (gloss)You’ve gotta be kidding! (ES)Tu rigoles? C’est quoi,cette histoire de candidate? (FS)

(10) Utena Yurusenai wa Saionji no yatsu. (J) (I) cannot accept that guy Saionji (‘s behaviour). (gloss) That Saionji guy’s gonna pay for this! (ES) Ce Saionji ne s’en tirera pas comme ҫa. (FS)

Whereas interjections are rarely used in either set of French subtitles, Nadelman uses them five times in Utena and more than twenty times in Slayers. In example (11) the American interjection man is used in the English subtitle to convey the emotional emphasis suggested by the sentence-final particle na, whereas the French subtitler uses the adverb vraiment (“really”):

(11) UtenaNandaka henteko na me ni awasareta na. (J)Somehow (I) had a weird experience. (gloss) Man, this has beenone truly weird day. (ES)Vraiment étrange,ce qui s’est passé là-bas. (FS)

In his translations of Lina’s lines, Nadelman sometimes combines interjections with the typographic and orthographic devices mentioned earlier. In examples (12) and (13) the combination of translation procedures helps reproduce the orality of the Japanese lines which again feature a conversational sentence-final


particle: nē (elongated for emphasis):

(12) LinaSasuga minatomachi nē. Osakana oishii!! (J)As one would expect in a port town. The fish is delicious. (gloss)Man, you gotta love seaport food! (ES)Le poisson a la criée,il n’y a rien de tel! (FS)

(13) LinaKekkō jibun ni shōjiki na renchū nē. (J)(They are) a bunch quite honest with themselves. (gloss)Man, now THOSE guys arehonest with themselves! (ES)Finalement, il n’y est passi attaché. (FS)

Nadelman is also able to recreate the pragmatic meanings associated with the use of address-suffixes in Japanese. Address plays an important role in character relations in Utena. Anthy addresses Utena honorifically as Utena-sama when the two girls become ‘betrothed’ after Utena defeats the villainous Saionji in a duel. In the English subtitles Anthy addresses Utena as “Miss Utena”, while in the French subtitles, she uses “vous” verb forms. In the story, Utena is displeased by this unusual honorific and the topic of address is raised in the dialogue itself:

(14) UtenaSono Utena-sama tte iu iikata wa yamete yo ne. (J) Stop that “Utena-sama” style of address. (gloss)And another thing. Would youstop calling me “Miss Utena”? (ES)Et arrête de me vouvoyer, d’accord? (FS)

Address forms are also drawn attention to in the Japanese script when Anthy pointedly changes her style of address to Saionji after he is defeated by Utena in a duel, and loses his right to her as the Rose Bride. Before the duel she addresses him as Saionji-sama, using an honorific address-suffix, but afterwards she deliberately addresses him as a normal senior student with Saionji-senpai. Nadelman translates this switch in politeness level by having Anthy address Saionji as Saionji, my master before the duel and as Saionji my classmate after he has lost. In the French subtitles Anthy uses the tu form of address; however there is no immediately contiguous use of vous forms with which to form a contrast:

(15) Anthy (Before the duel) Sumimasen Saionji-sama (J) Sorry Lord Saionji. (gloss) Forgive me, Saionji, my master. (ES) Je suis désolée, Saionji. (FS)


(16) Anthy (After the duel) Gokigenyō Saionji-senpai. (J) Farewell upper-classmate Saionji. (gloss) Cheer up, Saionji…

…my classmate. (ES) Ne t’inquiète pas, Saionji. (FS)

In Slayers, although address terms are not highlighted in this way in the dialogue itself, there is also differentiation between the characters in the way they address each other. Amelia, a polite and somewhat stuffy ‘princess’, always addresses other characters using the -san suffix, normally associated in Japanese with polite, respectful, but distant usage. In the English subtitles, the -san form is translated by using the old-fashioned combination of the title Mr. or Miss with the given name of the character being addressed. The French subtitler uses the tu/vous distinction as is the case in Utena: while the tomboyish Lina addresses Amelia with tu, Amelia uses the vous-form with Lina:

(17) Amelia Dōshita n desu ka Lina-san. (J) What happened Lina? (gloss) What happened, Miss Lina? (ES) Pourquoi êtes-vous partie, Lina? (FS)

(18) Amelia Lina-san sore tte moshikashite… Lina, possibly that (means) …(gloss) Miss Lina, you mean… Are you saying…? (ES) Lina, vous voulez dire… (FS)

The same procedures are used, respectively, by the American and French subtitlers to translate priestess Filia’s polite and reserved use of -san.

David Fleming: AkiraAkira originated as a science-fiction comic-strip by Ōtomo Katsuhiro. It began publication in Young Magazine in 1982, reaching its conclusion ten years later after a total of more than 2,000 pages had been drawn. An animated film version of Akira was written and directed by Ōtomo, and became the highest grossing film of 1988 in Japan. The film was also given theatrical release in the US and Europe, and is credited with sparking the major increase in interest in Japanese manga and anime which occurred in western countries in the 1990s. It was released on the UK home video market by Manga Entertainment Ltd., using dubbed and subtitled translations from the initial American release by Streamline Pictures. It was re-released in DVD format by Pioneer in 2001, with a newly translated dubbing script and subtitles. Akira thus provides another opportunity to make an intralingual comparison between two commercial subtitled translations of the same anime film.

The story of Akira, set in 2019, is centred around two teenage delinquents, Tetsuo and Kaneda, in the quasi-totalitarian society of neo-Tokyo. Tetsuo, a picked-upon junior member of a biker gang, starts to develop ‘psionic’ powers after crashing into a strange adult figure with the body of a child, one of a


number of survivors of a government-financed experimental project, the most powerful and mysterious of whom is Akira. As Tetsuo’s powers increase to a terrifying and cataclysmic extent, his gang-leader Kaneda seeks to find and kill him, allied with a group of revolutionary resistance fighters. In an apocalyptic ending whose meaning is not altogether clear, Kaneda and his friends are left to ride in freedom through the streets of the destroyed city, while Tetsuo and Akira are absorbed into what appears to be the creation of a new parallel universe.

As in Nadelman’s subtitles, in the new set of subtitles for Akira, translated by David Fleming, orthographic and typographic devices are used to recreate character voice. In example (19) whereas the older subtitles neutralize the spoken register of Kaneda’s sentence-final particle sa and lexical choice (koitsu), the new subtitles use pronunciation spelling (gonna) to convey a sense of orality:

(19) Kaneda Koitsu o todokete yaru no sa. (J) (I’ll) send this thing. (gloss) I’ll give this to him. (#1ES) I’m gonna send Yamagata his bike. (#2ES)

Whereas Nadelman’s subtitles used capitalization for emphasis, the later subtitles in Akira more regularly use italicization to suggest emphatic pronunciation in the dialogue. Italics are used in over 25 of Kaneda’s lines in the second set of subtitles, but only once in the first. This is an interesting translation procedure because italics are not used in this way in the earlier English subtitles or indeed in the French subtitles for Akira, and discussions of subtitling, whether written in French (Becquemont 1996: 148, Cornu 1996: 160) or English (Hervey & Higgins 1992: 160), tend to suggest that italics should be restricted to off-screen voices.

The difference in the translation of pragmatic meaning between the two sets of subtitles can be clearly seen in example (20), in which a lexical intensifier (the hell) and phonological intensification represented by italics help convey the emotional tone of the exclamation (nan da!) and the contraction of pronoun and topic particle (are wa > arya):

(20) Kaneda Nan da arya! (J) What is that (thing)? (gloss) What’s that? (#1ES) What the hell is that? (#2ES)

And in the next example (21), the rudeness of the imperative form (dokē) is again conveyed more strongly in the later set of subtitles, by means of an interjection and use of italics:

(21) Kaneda Dokē. (J) Move out of the way. (gloss) Get out of my way! (#1ES) Damn it, move! (#2ES)

In example (22), the combination of italics and operator ellipsis in the later subtitle convey the aggressive attitude of Kaneda’s exclamation using the rough address-term, (omē [“you”]), more strongly than the standard question form


used in the earlier title:

(22) Kaneda Nan da yo omē wa. Sōshikigaeri ka. (J) What are you? Coming back from a funeral? (gloss) Are you on your way from a funeral? (#1ES) What’s with you? You just come back from a funeral? (#2ES)

A final example from Kaneda’s lines (23) shows a similar contrast between the two sets of subtitles in their treatment of pragmatic aspects (interjection, non-standard pronunciation, sentence-final particle) of the original Japanese. Italics (damn) and non-standard language are used in Fleming’s translation only:

(23) Kaneda Chikushō kitanē zo Tetsuo. (J) Shit, that’s dirty Tetsuo. (gloss) None of your dirty tricks! Let’s fight fair! (#1ES) God damn it! This ain’t fair, Tetsuo! (#2ES)

When we also examine the translation of the dialogue of the other main character, Tetsuo, we find that, as is the case with Kaneda, Fleming’s subtitles regularly make use of typographic and orthographic devices – specifically: italics and pronunciation spelling, and on a single occasion capitalization (26). These help to recreate the rough character voice of Tetsuo, the juvenile delinquent gang member. In the following examples, he uses the informal masculine n da form rather than polite no desu, an aggressive, contemptuous form of address (temēra), and the direct imperative form kure (“give”) with the emphatic particle yo (27):

(24) Tetsuo Na…nan da aitsura… (J) What are those guys? (gloss) What are they? (#1ES) W-Who the hell are they? (#2ES)

(25) Tetsuo Nan da temēra wa. (J) What are you? (gloss) Who are you? (#1ES) Who the hell are you?! (#2ES)

(26) Tetsuo Doko ni iru n da. (J) Where is (he)? (gloss) Where is he? (#1ES) WHERE IS HE?! (#2ES)

(27) Tetsuo Kapseru kure yo. (J) Give capsules (gloss). Give me some capsules. (#1ES) Gimme some…capsules. (#2ES)


ConclusionsIn comparing two sets of subtitles for Takahata Isao’s well-known anime film, Grave of the Fireflies, it was found that in the shift from the spoken to the written channel neither subtitler reproduced the local colour suggested by Kōbe dialect. However, the American subtitler Neil Nadelman used pronunciation spellings such as whatcha and gonna, and choice of spoken register to recreate some of the social intimacy associated with dialect. This contrasted with the neutral, flat, and simple style of English subtitles used on the DVD marketed in Japan. And in investigating the subtitles of two television anime series, Revolutionary Girl Utena and Slayers Try, we found that Nadelman was again able to recreate indices of character from the Japanese dialogue. Japanese sentence-final particles which conveyed an emphatic voice for strong female characters in the original dialogue could be translated using compensatory procedures such as typographic devices (capitalization), orthographic devices (pronunciation spelling), and interjections. Nadelman was also able skillfully to recreate aspects of interpersonal politeness arising from modes of address in Japanese. The French subtitlers for Revolutionary Girl Utena and Slayers Try did not use typographic and orthographic devices, although they were able to make use of other resources in the respective target language, such as the tu/vous distinction. A considerable difference was noted between two English subtitled translations of the SF film Akira. In the earlier subtitles, many indices of character were neutralized, but in the more recent version, the subtitler, David Fleming, used compensatory procedures similar to those used by Nadelman to recreate character voice.

This would suggest that the influence of the technical constraints of subtitling on stylistic functions has sometimes been overstated. It seems true to say that subtitling may involve such a reduction in text volume that much stylistic information is lost. But, depending on the type of subtitling adopted, the narrative function of characterization in the Japanese dialogue can be reproduced by the use of a variety of compensatory procedures.

Technical considerations are not the subject of this article, but it may be noted incidentally that there appears to be a relationship between quantitative aspects of the subtitles and their ability to reproduce narrative function of character voice. When a simple quantitative comparison is made of the two translations of Grave of the Fireflies, there is a significant difference between the Central Park Media and the Warner Home Video subtitles. Although both translations have approximately the same number of subtitles, the subtitles written specifically for American viewers are much denser, comprising over 40% more words and many more double-lined subtitles. Similarly, there is a significant quantitative difference between the two sets of subtitles for Akira. The later Pioneer translation on DVD contains about 200 more titles than the earlier VHS release, and almost 50% more words. It is thus clear that there are, and have been, differences in subtitling practice with regard to text quantity (and hence expected reading speed). These differences have been highlighted in the exchange of views between professionals in the pages of the film journal Sight and Sound (Minchinton 1988, Packham 1988). An interesting task for future study would be to investigate the relationship in subtitling between such quantitative factors and recreation of character voice.


Denser subtitles are indeed able to reproduce corresponding effects with regard to characterization, but at the same time it should be noted that in no subtitles do we find the more domesticating procedures which are sometimes used in dubbing and comic-book translation. Specifically, procedures such as dialect substitution, and autonomous insertion of swearing, humour, or foreign language are not found in subtitles. Subtitling is thus ‘foreignizing’ in that it does not replace the original text as is usual in general translation, and it incorporates fewer adaptations than is the case in dubbing and comic-book translation. But it is also a mode of linguistic transfer which can potentially recreate interpersonal meanings and successfully ‘domesticate’ the stylistic function of character voice.

NoteThis article is adapted from parts of my Ph.D. dissertation completed at the University of Edinburgh.


character voice in anime subtitles peter howell,...

Documents