Thinking about translation as the engine rather than the caboose of literary history has led me to thinking about literary works as collective objects, made up of past, present, and future editions of the work in multiple languages and formats (Born Translated). I have come to think of a work's first edition as comprising not only the first imprint of the work in one language but also the aggregate of imprints across media, languages, and versions of languages. The first edition is one and many. It exists in multiple formats as well as in multiple languages—and, indeed, format alters, greatly determines, and diffuses language. For the novel, digital and audiobooks alter through sight, sound, and touch our experience of the page and the words on the page. Format alters the use of idiom because the meaning of words depends on variations to orthography, font, and illustration. As we consider translation across media as well as across lexicons, we see that the multilingualism of the book precedes, compounds, and extends the multilingualism of the text.

My vision of translation intersects with the aqueous borders, dynamic multilingualism, and proleptic groupings that contemporary scholars of the archipelago have associated with archipelagic thought. A paradigm for understanding island clusters such as the Philippines and the Antilles, archipelagic thought draws its concepts from objects at once individual and collective, one and many, whose boundaries and substance such as shorelines, land mass, and estuaries are unfixed and changeable (Roberts; Roberts and Stephens). The archipelago serves as both metaphor and agent of translation. Because inhabitants of archipelagos move frequently across overlapping spaces, archipelagos create new, blended languages while hindering efforts to count and distinguish languages. The archipelago, like the practice of translation, establishes spatial, cultural, and linguistic networks that converge and diverge over time. Throughout history, islands have been cauldrons of language making, language borrowing, and language negotiation, where contact and mobility transform the content and shape of local communication. What we call islands, archipelagic thinkers have emphasized, may be paradigmatically solitary, but they begin as dynamic collectives. Their multilingualism drives translation.

Engaging with multilingual communities that operate within what have appeared to be monolingual geographies—the nation, the city, the island—contemporary artists, writers, and filmmakers are designing works that require translation. Yet unlike the enriched multilingualism that rewards linguistic scope (knowing many languages) or linguistic depth (knowing the patois of the neighborhood or region), the new multilingualism cultivates linguistic deficit: the experience of not knowing or knowing less than one language (Walkowitz, “Less than One Language”). Because they require translation, these works impede fluency, but they also welcome relative fluency. They force audiences to engage with tongues that are intentionally and explicitly derivative, extrinsic, and foreign—not original, intrinsic, and native—and they suggest that this mode of engagement is necessary and universal rather than compensatory or second-rate.

I have described the phenomenon of not knowing in works of literature such as Jhumpa Lahiri's recent fiction, the novels of Yoko Tawada, and essays and novels by Valeria Luiselli (Walkowitz, “On Not Knowing” and “English”). In these instances, we encounter visual images, unspoken words, withheld names, nonsense games, and visceral gestures that are irreducible to language or remain inaccessible to speakers or readers of any single language, whether German, English, Italian, Japanese, Spanish, or another language in which these works have appeared. Lahiri, Tawada, and Luiselli are designing literary works whose languages are strategically uncountable (Sakai). Normally, we expect multilingual works to involve several national languages on the page. However, these writers approach the concept of multilingualism not by adding new languages but by withdrawing and detouring access to even one language. By forcing readers of dominant languages to encounter at least some of the work in translation, these writers establish a more inclusive, more hospitable multilingualism.

In the remainder of this essay, I turn to prominent works of multilingual film. These works are notable because they must be seen and heard in translation, even when they are consumed by their primary audiences. They operate in the original and in translation from the start. In fact, the films go further than this: the translation is part of the original because the dialogue features multiple languages, often in a single scene. Characters understand at least one language but not all of them. This is generally true for audiences too. Subtitles are part of the work rather than simply external or supplementary to the work. Deficit multilingualism is crucial to the films’ stories about migrant labor, colonialism, globalization, transnational empathy, and civic hospitality. When the films travel out to global audiences, translators encounter a novel problem. How do you provide subtitles in one language that can convey the experience of dialogue in many languages? How do you show that characters are speaking a language that is not a dominant language or that is not a language that other characters, who speak minor or regional languages, can understand?

Perhaps the most prominent example of a multilingual film for which there are no native speakers is Drive My Car (2021), which won the 2022 Academy Award for best international feature. Drive My Car begins in at least ten languages. A review in The New York Times lists the languages as “Japanese, Korean, English, Cantonese, Mandarin, Tagalog, Bahasa Indonesia, German, and Malaysian,” whereas the Criterion edition of the film lists “Japanese, English, Korean, Korean Sign Language, German, Mandarin, Tagalog, Indonesian, and Malay.” It is difficult to count the number of languages because there are so many and because there is no general agreement across the institutions of filmmaking, film distribution, film marketing and reviewing, and film teaching about what a distinct language is. Are Cantonese and Mandarin two languages in the same way that Korean and Japanese are? What about Korean and Korean Sign Language (KSL), which appear as distinct units in the notes to the Criterion edition but not in the New York Times review, where only Korean appears (Dargis)? As we'll see, the film treats KSL and Korean not as one but as two, or at least as more than one, in important ways.

The protagonist of Drive My Car is Yusuke Kafuku, a theater director and actor who specializes in the staging of plays in translation. However, here is the catch: he does not just translate from one language to another language—for example, a Russian play into French, or a French play into Japanese. Instead, he translates one play into several languages at the same time, assigning a language or a version of a language to each of the speaking parts. The characters in the play do not notice. The diegetic play is monolingual, whereas the extradiegetic play, as it is experienced by audiences within the film (fig. 1) and by audiences of the film (fig. 2), is multilingual. The film makes us hyperaware of watching, hearing, seeing, reading, as well as listening to languages.

Fig. 1.

Yoon-a and Kafuku performing Uncle Vanya in Drive My Car (perspective from stage).

Fig. 2.

Yoon-a and Kafuku performing Uncle Vanya in Drive My Car, English subtitles for Korean Sign Language (perspective from audience).

The plays we see in Drive My Car are Samuel Beckett's Waiting for Godot and Anton Chekhov's Uncle Vanya. Both are modernist plays that began in French and Russian, respectively, and have had a vast impact in translation and adaptation across many additional languages. As many will know, Beckett wrote Waiting for Godot in French and then rewrote the play in English; in this sense, the play began twice, once in each language. The film Drive My Car is based on a short story by the Japanese novelist Haruki Murakami, whose early career in fiction overlapped with his career as a translator and who is widely known for inventing a version of Japanese that sounds like it is translated from another language and is ready to be translated into additional languages (Walkowitz, Born Translated 14–15). Drive My Car is a film that incorporates other arts, including printed words, theater, and audio tape, and is based on a short story whose original language was engineered for translation. The film uses the multigeneric, multimedia, and multilingual legacy of twentieth-century modernism to explore twenty-first-century experiences of social, political, linguistic, and psychic isolation.

In Drive My Car's production of Waiting for Godot, Didi and Estragon are speaking Japanese and Indonesian, respectively. Above the stage, we see supertitles, which appear in Japanese and English when Estragon speaks and in Indonesian and English when Didi speaks (fig. 3). Kafuku seems to imagine an audience of spectators who know English, Indonesian, or Japanese but not all three. It is significant to his project that Japanese and Indonesian are not neighboring languages. Because he has chosen tongues from relatively separate geopolitical orbits, the supertitles are necessary. Without them, the dialogue would be partly inaccessible to everyone. Put another way, the collective needs translation, but the individuals within that collective do not need translation in the same way. There is no monolingual experience of the play, or of the film: they cannot be experienced in one language.

Fig. 3.

Performance of Waiting for Godot in Drive My Car, Japanese and English supertitles for Indonesian and English subtitles for Indonesian.

We watch as well as hear languages in every version of the film, so listeners are aware of multiple idioms even if they cannot identify or count them. We can see as well as read the supertitles: the scripts change, and we can observe the shift between Japanese and Indonesian visually as well as verbally. The subtitles signal graphically: Estragon's Indonesian is translated with parentheses in the English version, whereas Didi's Japanese appears in subtitles without parentheses or, in the Japanese version, is not subtitled at all. This use of parentheses or sometimes brackets has become a common visual strategy for translating the relationship between minor and dominant languages for audiences who operate in neither language as well as for audiences who might not realize that two different languages are being spoken. Imagine, for example, viewers who cannot hear the difference between Japanese and Korean, Hebrew and Arabic, or English and Spanish.

Drive My Car was produced by the Japanese company Bitters End and the Japanese director Ryusuke Hamaguchi, but it resists classification as a Japanese-language film. As Kevin Dettmar rightly explains, Drive My Car is “inconceivable without subtitles.” This is true because there are so many languages spoken, because even the Japanese version requires translation for Japanese-speaking viewers, and because the film presents relative fluency as a prerequisite for compassion as well as collectivity. Relative fluency, for our purposes, might be understood as the experience of not knowing as well as knowing languages. Hospitality, the film suggests, depends on estrangement and on communication across estrangement.Footnote ¹ As the film's speaker of KSL explains, “People not understanding my words is normal for me.” Drive My Car seeks to make “not understanding” normal for everyone.

Later scenes in Drive My Car show us the staging of Uncle Vanya in Hiroshima, a symbolic location for a play expressing the difficulty and the necessity of empathy. Here, we encounter Yoon-a, the KSL-speaking actor, and her husband, Yoon-su, one of the theater's organizers and a translator among Japanese, Korean, and KSL. As for Waiting for Godot, there are supertitles for the play and subtitles for the film, but this time it is more complicated. With more characters, there are more languages, and the role of Sonya is assigned to Yoon-a, whose subtitles need to be read even by speakers of Korean. KSL appears in italics in the English subtitles, adding a second visual cue, along with the use (or absence) of parentheses, to distinguish among the languages and the kinds of language we encounter (fig. 4). The presence of KSL reminds audiences that sometimes one language is several languages and that a language can be major and minor at the same time: in this case, Korean is historically minor in relation to Japanese but major in relation to KSL. Drive My Car asks audiences to calibrate and recalibrate the relationships among languages and speakers of languages. By signposting political and social relations as linguistic relations, the film allows us to consider how the experience of linguistic deficit can drive the experience of collective feeling.

Fig. 4.

Performance of Uncle Vanya in Drive My Car, supertitles in several languages for Korean Sign Language and English subtitles for Korean Sign Language.

The discussion and presentation of KSL, both on the stage and in the couple's home, illuminate the embodiment of language and the richness of silence. Yoon-a and Yoon-su had concealed their marriage from Kafuku so that her audition would not be compromised by their relationship. Speaking to that decision, but also speaking to the impact of his wife's performance on stage, Yoon-su reflects, both seriously and ironically: “Silence is golden.” And as Yoon-a points out, through her husband's translation, no one knows the volubility of silence as well as she does. In the subtitles assigned to Yoon-su's rendering of Yoon-a's speech, quotation marks indicate that he is translating across medium as well as lexicon (fig. 5).

Fig. 5.

Yoon-a talking with Yoon-su, Kafuke, and Misaki in Drive My Car, English subtitles for Japanese translation of Korean Sign Language.

We are asked to notice, throughout Drive My Car, that Yoon-a's performance is communicating a general truth about the embodiment of words: gesture modifies speech, gesture conveys speech, and gesture takes the place of speech. There is no lexicon without medium. We see that languages, even those we think we understand fluently, are modified by silence: what we do not hear, what is not said, or what we cannot understand. We are asked to watch, read, hear, and listen, but we are sometimes watching words and reading images, or listening to words that we cannot understand and that thus function as sound instead of sense. The film makes language into a palpable substance that is very often out of reach, or out of any complete or fully graspable reach. The experience of not knowing languages commands our attentiveness as well as our empathy, especially for characters from communities whose languages are less known, infrequently learned, or undervalued. The command of attentiveness is crucial to the ethical posture of the film: we must listen, watch, read, and hear intently. Explaining his decision to bring Yoon-a with him to Hiroshima from Korea, where she had family and friends who speak and understand KSL, Yoon-su explains, “I figured I could listen to her like a hundred people.” Yoon-su has learned Yoon-a's language so that he can communicate with her. We are meant to understand in his use of the verb “listen” a sense of irony about what it means to be absorbed and attentive to the language of others. We are meant to register his insistence, and the insistence of the film, that those who have to listen intently may be more compassionate—better listeners—than those who have learned a language automatically or natively.

Before the film begins, Yoon-a has come from South Korea to Japan. In Japan, as we have seen, her language requires two layers of translation: translation from KSL into Japanese, and translation from gestural speech into oral speech. In the English subtitles, we see translation from KSL into English and from roman text and into italicized text; italics signal the translation from hands into voice. At the end of the film, the driver, Misaki, a young Japanese woman, travels in the opposite direction, from Japan to South Korea. We can see that it is South Korea because the language on her license plate has changed from Japanese to Korean and the signs on storefronts are in Korean and English. The story has advanced, and the omnipresence of masks signals the presence of a global pandemic: we are in COVID time (fig. 6). People are living separately together in new ways. Misaki has learned or is learning Korean. Parentheses used in the subtitles for her speech designate the minor tongue. Yet she appears much more at ease and much more at home than she did in her native Japan. It is interesting that Korean is placed in parentheses here: it is a minor tongue for Misaki, a Japanese immigrant, but a major tongue for the Koreans in this space. Learning a minor language, the film suggests, allows Misaki to live and to speak more freely. The final scene is subtitled in several ways. And by subtitled, I mean it is full of images of language and conditions of expression that obstruct as well as facilitate communication. The scene is subtitled for Misaki, who is speaking in an acquired tongue. It is subtitled for English-speaking viewers, who are asked to see English and Korean on the screen. And it is subtitled for the Korean characters who have to listen, watch, and understand the masked, plexiglassed speech of their neighbors.

Fig. 6.

Misaki in South Korea in Drive My Car, English subtitles for Korean.

Drive My Car is not the first or the only recent visual narrative to require subtitles for all audiences. The thematic relationship among dominant and minor languages is especially visible in the popular television serial Pachinko (2022), which was produced originally in Korean and Japanese audio, and occasionally in English, and focuses on three generations of Korean migrants in imperial and near-contemporary Japan. There are subtitles in every version, and episodes are preceded by title cards providing location and year in the serial's three languages. Subtitled dialogue assigns each language to a different color so that audiences who do not speak or cannot recognize one or more of the three retain the experience of contrast. Japanese appears in blue, Korean in yellow, and English in white. Again, visual, graphic, and verbal strategies of multilingualism ensure that the episodes are experienced in translation even by viewers who speak the three principal tongues.

We can also consider the Mexican film Roma from 2018, which was produced in Mixteco and Spanish and provides subtitles for Mixteco in the original version of the film. In the English subtitles, Mixteco is bracketed, so that we can see the difference from Spanish, even if we are unable to hear it (fig. 7), and Spanish appears in subtitles without brackets. The Criterion edition of the film brackets as well as subtitles Mixteco in the original version (fig. 8), but the version streaming on HBO only subtitles Mixteco (fig. 9). The Criterion strategy has the effect of highlighting visually the social, economic, and cultural differences between the Mixteco-speaking characters and the characters who speak only Spanish. Mixteco is emphatically minor. The use of bracketing translates the political register of language difference for a global audience, including viewers in Spanish-dominant regions outside Mexico, where the context of Mixteco as an indigenous language may not be well known.

Fig. 7.

Roma, Criterion edition, English subtitles for Mixteco, English translation.

Fig. 8.

Roma, Criterion edition, Spanish subtitles for Mixteco, original version.

Fig. 9.

Roma, HBO edition, Spanish subtitles for Mixteco, original version.

Another notable example is Fauda, an Israeli-made serial from 2015, which brackets Arabic in the English subtitles while leaving Hebrew, the dominant language, unbracketed. Steven Spielberg's 2021 film version of West Side Story is an interesting exception to the emphasis on multilingual subtitling. The production operates in Spanish and English with no subtitles in the original, implying that English and Spanish are one language, even if Spanish is unknown or partly unknown to many viewers based in the United States and wholly unknown to many English-speaking viewers outside the United States. The idea was to make English-only audiences “more attentive,” either because they can figure out the language they do not know well or because they cannot, explains Rita Moreno, who performs in the film and is one of the executive producers (qtd. in Acuna). However, when the film is translated and even when it is captioned in the original, the words on the screen lack any graphic markers (italics, brackets, or parentheses) that would tell viewers that there is more than one language being spoken (fig. 10). In this way, West Side Story operates multilingually for some English-speaking viewers but monolingually for viewers who are speakers of other languages.

Fig. 10.

West Side Story, HBO edition, English captions.

In these films and visual narratives, we are in the thematic space of the archipelago: a space of internal borders, contested territories, rival languages, and human movement caused by violence, famine, flows of labor, and flows of capital. We can see and hear idioms in the process of being made. Language is rendered uncountable. These narratives generate linguistic hospitality in the service of multilateralism: by making audiences less fluent and thus more attentive to words, they hope to make audiences act with greater empathy toward the foreignness of others.

Thinking about multilingual counting, we can revisit Jacques Derrida's claims about what it means to speak one language. In his classic formulation, Derrida avers that “we only ever speak one language” and we “never speak only one language” (7). He recalculates the idea of “one language” in three principal ways. First, he suggests, there is no pure idiom: there is no one way to speak or write a language, since everyone speaks a distinct idiom, and languages are the aggregate of those idioms. Second, he insists that we do not own any language—his one language does not belong to him (1)—since we are always speaking to and for someone else. The minute we speak or write, we are engaging with someone else's understanding of that language. It is not ours alone. All languages are in some way borrowed, taken, or imposed. They do not begin with us. Finally, he argues, it is possible to use, value, and even preserve languages without, as he says in my translation, “archiving” them. Languages are not meaningful and useful because they are coherent and fixed. On the contrary: they are alive because they overlap with other languages, because they are used, and because they change.

From the perspective of new migration and the resurgence of ethnonationalism, including ethnonationalism rooted in the isolation of languages, contemporary writers and filmmakers add to Derrida's argument. First, they emphasize the media and format of transmission: the orthography and font of multiple languages, and the visual bracketing of minor tongues. Second, they emphasize the embodiment of languages: the idea that communication always involves gesture, touch, color, and sound, including the absence of sound. Third, they emphasize the ethics of unknowing languages, especially dominant languages: the idea that all audiences should have to be language learners as well as language knowers. And, fourth, they ask us to engage generously with other languages: we become readers, interpreters, listeners, watchers, and handlers of words we do not know or do not know well. Approaching one language as an additional language—additional to itself, additional to the languages around us, and additional to the languages it has been and will become—is central to their vision of civic hospitality, human rights, and multilateral coexistence. They call for solidarity without fluency.

Global citizenship depends on our capacity to experience the aliveness of languages and to be alive in languages, to share the languages we know and to engage with languages that others know. Recognizing the history and future of many languages in any one language involves changing the way we understand the original idiom of novels, films, and visual narratives; the way we organize the institutions that collect, study, teach, and interpret these works; and the way we calculate the audiences who belong to them.