1. Introduction

Prismatic Jane Eyre began as an experiment in collaborative close reading across many languages.1 We were provoked by Franco Moretti’s manifesto for distant reading, according to which close reading can have no place in the study of world literature since it “necessarily depends on an extremely small canon”: instead, “literary history will […] become ‘second-hand’: a patchwork of other people’s research, without a single direct textual reading […] If we want to understand the system in its entirety, we must accept losing something” (48–49). This famous assertion has animated much eye-opening research (Moretti, Graphs; Bode; Erlin and Tatlock; Alfano and Stauffer), and generated much debate (Kristal; Arac; Armstrong et al.). Like others, we regretted the loss that Moretti so airily welcomes: we wanted to hold onto the experience of reading and the granularity of linguistic and cultural difference—everything, in fact, that resists the idea that world literature can be reduced to ‘the system’ which can be understood ‘in its entirety’. But, rather than reaffirming the value of close reading as traditionally practised (Trumpener, Tally), we wanted to see if it could be made to work in new ways in a world-literary context, maintaining its heuristic power while looking across a large collection of texts from different times and languages.

This meant turning close reading into a collaborative enterprise: what practices of sharing, what modes of explanation, what norms of turn-taking would enable our group of thirty researchers working on twenty-four languages to have illuminating intellectual interactions? It meant selecting elements for us all to focus on: which passages, words, phrases, metaphorical patterns, and grammatical forms would nourish our combined attention? And it meant working on translations, since it seemed likely that the ground which translations have in common with their source text and one another would enable revealing differences to emerge through comparison. Jane Eyre (1847) was chosen because it has been much translated in many parts of the world (at least 619 times into at least 62 languages, by our latest count); and also because its polemical charge as an early feminist text, its ideological conflicts (for instance in the representation of Bertha Rochester), and generic instability (between romance and realism) seemed likely to generate interesting variances in translation, not only across languages and cultures but between different historical moments.

The project builds on the work done in Translation Studies since the 1970s to show that the inevitable differences between translations and source texts should be seen, not as errors to be denounced, but as indexes of linguistic, cultural, and historical difference and of translators’ individual imaginative energies (Toury; Bermann and Porter); and it extends the approach developed in Reynolds (ed.), Prismatic Translation (2019), viewing translation as a process that inevitably generates multiple texts which ask to be analysed and interpreted comparatively. In this, it offers a different understanding both of close reading and of canonicity from Moretti’s. Of course, Jane Eyre is a ‘canonical’ novel; but as soon as you look at it in a global perspective the significance of that statement changes. We are faced, not with a single text, but with at least 619, that is, something in the region of a hundred million words. This enormous and varied cloud of textuality has been produced not by one renowned Author but by hundreds of writers, almost all of whom are not in the least celebrated and many of whom are unknown. The novel is not held within the canonizing structures typical of national cultures (university courses, etc.) but is remade, sometimes (to be sure) as a literary text but also as a romance, adventure, or young adult fiction: the translations take their place among a multitude of variously popular and irreverent remakings, including melodramas, abridgements, film versions, literary responses, erotic versions, fan fiction, and merchandise including the ‘Jane Eyre’ pencil cases and tote bags that are popular in South Korea (Choi). As soon as you set about close reading translations, you are no longer faced with an “extremely small canon”, and in fact not really with a ‘canon’ at all, since translations, though they emerge from and contribute to the canonization of their source text, are not in themselves canonical. Correspondingly, the aim of close reading will change: no longer establishing the distinctiveness of a single text, but rather tracing interconnections and variances among the many co-created versions which constitute the global Jane Eyre.

We soon realized, however, that close reading of such a phenomenon cannot proceed alone. If we were to make a justifiable selection from this plethora of textuality, we needed first to be able to grasp it, which meant mapping it (Figure 1). So what began as a reaction against distant reading turned into a collaboration with it. In this article, which is really an interim report on the project, we offer some reflections on the interaction between digital techniques and traditional practices of scholarship and close reading; we outline the process of mapping many translations of Jane Eyre, the new perspectives on translation that it created, and some discoveries that arose therefrom; and we present a technique for the comparative textual analysis of translations in different languages, together with some early results.

The ‘Basic Map’, zoomed out to provide a snapshot of the global distribution of Jane Eyre translations
Figure 1 

The ‘Basic Map’, zoomed out to provide a snapshot of the global distribution of Jane Eyre translations. © OpenStreetMap contributors.

2. Constructing the Maps

Publication data about the translations was initially sourced from WorldCat and UNESCO’s Index Translationum, but it soon became apparent that these databases, though large, provide only a very partial account of the field. So the whole research group collaborated in improving the data, searching national and local library catalogues both online and on paper, and also using bookshops: for instance the two (abridged) Gujarati translations known to us were found in bookshops by Abhishek Jain. The work is ongoing, as the digital humanities (DH), archival and literary-critical strands of the project are running simultaneously, with a dialogic relationship between them. This is an untidy, and in some ways frustrating method: where the DH researcher prefers a complete dataset to work with, archival scholars and readers are always excited to uncover new things. Nevertheless, our practice has the benefit that a DH intervention can prompt a new archival or literary-critical discovery, and vice versa—a process of interaction that will be described further below.

The provisional dataset that underpinned the maps we have so far created (and currently—March 2021—available on our website) consisted of 593 acts of translation into 57 languages. We defined an ‘act of translation’ as including both the first publication of a new translation and its republication in a different place, for reasons we will explain below; and we did not attempt to distinguish between strict translations and abridgements or freer versions, so long as they remained within the genre of prose fiction. This is both because such categories are historically and culturally contingent (Toury, 62) and because we would not have been able to establish our own definition of them without close study of every text in the dataset—an impossibility given the quantities involved. The number of acts of translation per language in our provisional dataset varied from fifty-six for Russian to fifteen languages boasting only one translation each. However, we now know that those figures were in many respects wide of the mark, and we are currently aware of 668 acts of translation into 62 languages. The biggest disparity concerns translations into Chinese: when we made the provisional dataset we knew of forty-two Chinese translations; but as we were writing this article Hongtao Wang and a team of researchers at Beijing Foreign Studies University completed work that uncovered a further eighty-eight translations. With a total of 130 translations, we now know Chinese to be, by some distance, the language that has most welcomed and remade Jane Eyre. The Russian data has shifted in the opposite direction, as it transpired that many texts where the translator’s name was not given were in fact reprints of earlier versions. Thanks to work by Eugenia Kelbert, helped by Karolina Gurevich, we now count six translations, propagated by an interesting succession of acts of translation which we describe further below.2

An updated list of translations—‘final’ so far as the project is concerned, though it will no doubt still have imperfections, and will rapidly be superseded as new translations appear—will be published in 2022, and the maps revised accordingly. The embrace of open and interactive publication is a key aspect of our project: materials are made available progressively on the website (https://prismaticjaneeyre.org/), and the final electronic book publication, linked to the website, will be open access. This commitment embodies the recognition that our work is implicated in what it describes: it is not only a study of the global translation and reception of Jane Eyre, but is itself also an instance of that phenomenon. In this, we aim to align the project with Andrew Piper’s argument that DH avoids the meme of the critic “as heroic individual” and reveals scholarship’s embeddedness in “social practices” (8), while equally embracing Francis B. Nyamnjoh’s recommendations, in decolonial studies, for “convivial scholarship” which understands itself to be engaged in an “ontology of incompleteness” (5, 2). Our practice also has the advantage of opening our research to improvement by the public: we have already benefitted from the expertise of several Brontë fans and book collectors.

The bibliographic information collected by our colleagues presented some problems of data uniformity. Sometimes further work was needed to identify the place and/or date of publication. In roughly one third of cases the translator is unknown. In other instances, the name listed in catalogues is not that of the actual translator; and on occasion it seems possible that the translator’s name provided in the book itself is a pseudonym designed to conceal a plagiaristic act of publication. Sometimes, much is known about a translator’s life; often, nothing at all. It is therefore impossible to engage in any systematic visualization based on translators’ lives. This is a particular shame as a translator’s location, life-story, or regional identity may be interesting in themselves, and may also leave their marks on the use of language in the translations.

The dataset employed for the maps included eighteen possible attributes, of which eight were necessary: place of publication named in the book, city (which sometimes differs from the named place of publication), country, language, date, translator(s) when known, publisher, and geographical coordinates. The dataset was restructured to compose GeoJson files and shapefiles necessary for the visualization of point maps, and polygonal elements corresponding to the countries where Jane Eyre translations have been published. The data were therefore adapted to different types of digital visualization. We worked with the online JavaScript-based applications StoryMaps JS (storymap.knightlab.com/) and Carto (carto.com/) to create the ‘Covers Maps’ and ‘Time Map’ (see description below) and, in doing so, to correct the location data in intermediate drafts of the project. The main, synchronous global maps were then developed in the R programming language through the Leaflet package (rstudio.github.io/leaflet/) which allows the use of homonymous JavaScript libraries dedicated to spatial analysis (leafletjs.com/).

3. Locating the Translations

In order to build the maps, each translation of course had to be given a location, and this raises a fundamental intellectual question: where does a translation belong? What area or entity is a translation done ‘into’? A standard response would be that a translation is done into a ‘target language’, which would seem to entail that, on a map, each translation should be pegged to the geographical distribution of its language. But there are several problems with this idea. The distribution of a language is too complex and variable to be mapped: even the nuanced (and expensive) language maps offered by ethnologue.com cannot account for migration and travel, lone speakers or small clusters of them, or (even more importantly in the case of a text) the difference between speaking a language and reading it. And then of course both languages themselves and their distributions change and shift over time. Sowon S. Park points out that Yu JongHo’s translation into Korean (1970) was substantially revised by him in 2004 to take account of language change: the “ornate and literary Chinese vocabulary” of the first version was replaced by “more up-to-date modern Korean” in the second (Park). Another stark instance (researched by Sasha Mile Rudan) is Giga Gračan & Andrijana Hjuit’s 1974 Croatian translation, rewritten by Gračan alone in 2008 using post-independence linguistic forms. On a larger timescale, the point holds true of all languages and all translations: a common reason why new translations are commissioned is that old ones are felt to be out of date. So, the idea of mapping translations onto languages leads to a recognition of the variability and changeability of the language-scape in which translation is done, and the impossibility of using language-distribution as the basis of our maps.

What about nation-states? They stake claims to literary history, so that the target culture for the translation of a novel is often taken to be a national literature; and they also often limit the market for a translation. But here again we faced the problem of shifting borders (one need only think of Russia and the USSR, but there are very many other instances); and it is also the case that translations are sometimes reprinted—and therefore obviously read—outside the state where they first appeared. So our solution, in our global and chronological maps, is to follow what, for a literary scholar, is the traditional way of locating books, and peg the translations to the places where they were published, as given (usually) on their title pages. This is generally a city, except in rare instances such as the 1983 Italian translation by R. Cenni, published by Malipiero in ‘Ozzano Emilia’, just outside Bologna. Such places—and cities—may change names, but they do not usually move. This means that our maps show the location from which each translation went out into the world, and do not engage in a necessarily doomed attempt to indicate the boundaries of its distribution, let alone of its readership. Languages have very variable diasporic existences; and, in the end, any translation can be read anywhere.

This choice came about because of practical considerations: how best could we construct the maps to present our data visually? Yet it also had a fundamental effect on how translation is conceived throughout the project. This is an example of the ongoing dialogue between DH on the one hand and archival and literary critical work on the other. Our choice binds the act of translation closely to the act of publication: as soon as you start making a map, what strikes you most about a translation is that it regenerates the text in a new location. This gives translation a distinctive spatial dynamic: rather than being ‘into’ a language or ‘into’ a national culture, translation comes to look like a text emerging ‘from’ place after place, city after city. In line with this view, for the purposes of our maps, we decided that if a translation is republished in a different city it should be represented as a new ‘act of translation’. This draws our approach into harmony with the idea of ‘bibliomigrancy’ developed by B. Venkat Mani, and means that each text figures in two ways in the project’s representation of data. In the maps, each act of translation—as we are calling them—figures separately, because cartographic representation necessarily foregrounds the fact that translations are located. But in the list that will appear in the book of the project, new translations will be understood in the traditional way, with republication in new locations being included in the appended bibliographic information. This double perspective required our colleagues in the project to do further archival work. It turns out that such migratory republication does not happen very often; but two examples will show the interest of this perspective when it does.

The first is a classic translation of Jane Eyre into Russian by V.O. Stanevich, which has been researched by Eugenia Kelbert. The Stanevich translation was first published simultaneously in Moscow and St Petersburg in 1950. It was much reprinted in both those cities, but its geographical publication-life also extended a great deal further. On our maps, you can watch it appearing in Alma-Ata, in what is now Kazakhstan (1956), Barnaul, in Altai Krai (1958), Gorkji, now Nizhny Novgorod (1958), Tashkent, in what is now Uzbekistan (1959); and then Makhachkala on the Caspian Sea (1986), Minsk in what is now Belarus (1988), Saransk, in Mordovia (1989), Voronezh (1990), Izhevsk, in the Urals (1991), Krasnoyarsk, in Siberia (1992—see Figure 2), Omsk, again in Siberia (1992), Kaliningrad on the Baltic (1993), Kazan, on the Volga (1993), Krasnodar (1993), Ulan-Ude, in the Russian Far East (1994), and Nal’chik, in the Caucasus Mountains (1997). This distribution gives us a window onto the conditions of publication, first of all during the 1950s and then during the period of Glasnost and the dissolution of the USSR. The very fact that these successive publications in different locations were necessary shows that the first acts of translation in Moscow and St Petersburg did not translate the novel ‘into Russian’, in the sense of making it available to all Russian speakers, nor ‘into the USSR’ or ‘into Soviet culture’, in the sense of conveying it to all inhabitants of that state or participants in that culture. The first acts of translation brought the novel to smaller linguistic, geographical, and cultural areas; and more acts of translation were needed to carry it further.

The ‘World Map’, zoomed in to show the 1992 publication of Stanevich’s Russian translation in Krasnoyarsk, Siberia
Figure 2 

The ‘World Map’, zoomed in to show the 1992 publication of Stanevich’s Russian translation in Krasnoyarsk, Siberia. Maps © Thunderforest, Data © OpenStreetMap contributors.

The second instance is the vivid and compelling translation of Jane Eyre into Spanish by Juan González-Blanco de Luaces, first published in Barcelona in 1943. As Andrés Claro recounts, Luaces, a republican, had had a career as a novelist before the establishment of Franco’s regime and the imposition of censorship (Claro; Ortega Sáez). Like others in his position, Luaces turned to translation both to make a living and as a way of continuing the imaginative life that was no longer available to him as an author. His version of Jane Eyre was much reprinted in Barcelona; it then crossed the Atlantic to Argentina, where it appeared in Buenos Aires in 1955, joining four other Jane Eyre translations that had been published there during the 1940s. Claro explains how the significance of the act of translation changes in the new context:

the translation boom that began in Spain in the early 1940s, in the face of censorship and abrupt stifling of local creative expression by the Franco dictatorship (1939–1975), can be understood as a literary compensation and even an implicit alibi serving to avoid oppression and write between the lines, not least in relation to the Francoist National-Catholic programme and the domestic role for women that it promoted.

But when transplanted to Buenos Aires, the Luaces translation became

part of an explicit programme of opening up to and interacting with foreign languages and literatures as a way of creating a local ethos and literature emancipated from Spanish colonialism, especially in the young republics of the Southern Cone, where the novel became widely known at the very time the female vote and other civil rights for women were being secured. (Claro)

It was only after this that Luaces’s translation came out in Madrid, where the literary scene was more Francoist than in Barcelona (Claro): it was first published there in 1967, by the same transnational firm, Espasa-Calpe, which had brought it to Buenos Aires. Reprints continued (and still do) in all these cities; and in 1985 Luaces’s translation appeared in a second South American location, Bogotá, from a different publisher, Oveja Negra. Tracking this text shows us an instance of the transnational dynamics of Spanish-language publishing, and the varying pressures that have encouraged translation in different locations and times.

Mapping the translations by cities reveals a world in which they—in line with the publishing industry in general—are variably concentrated or dispersed. The complex travels of the Stanevich and Luaces translations are made possible by the variegated publishing environments which they inhabit. However, when you zoom out and look globally, you find that it is more common for translation into a language, and in a state, to be highly concentrated in a single city. In Korea, all the twenty-seven translations into Korean are published in Seoul; in Japan, all twenty-two translations into Japanese are published in Tokyo except for one (Kyoto, 2002); in Turkey, thirty-two translations are published in Istanbul, three in Ankara, and one in Izmir (see Figure 3); in Iran, thirty-three translations into Persian are published in Tehran, three in Qom, one in Tabriz, and one in Mashhad. In Europe, you sometimes find a similar pattern: Greece, twenty-two translations in Athens and one in Thessaloniki (1979); France, seventeen in Paris and one in Poitiers (1948). But there are also more dispersed environments, more along the lines of Russia. Of Italy’s forty-one translations, twenty-two are published in Milan, but the rest are shared out among twelve other cities. German has a yet flatter distribution which spreads across Germany and beyond: six in Berlin; two each for Stuttgart, Leipzig and Frankfurt; one each for six other German cities; and then five in Zurich and two in Vienna.3 The markets, the distribution networks, and the socio-political dynamics are different in each case; and so therefore is the significance of the act of translation.

The ‘World Map’ zoomed in to show the distribution of Jane Eyre translations in Turkey, Greece, Albania, North Macedonia, and Bulgaria
Figure 3 

The ‘World Map’ zoomed in to show the distribution of Jane Eyre translations in Turkey, Greece, Albania, North Macedonia, and Bulgaria. Maps © Thunderforest, Data © OpenStreetMap contributors.

The maps also reveal the large stretches of the world where translations of Jane Eyre have not occurred. In Africa, eight translations into Arabic have appeared in Egypt, one of which—the canonical version by Munir Baalbaki—has been republished in Casablanca.4 But, aside from a version for children in Afrikaans done by Antoinette Stimie and published in Cape Town in 2005, there is only one translation into an African language other than Arabic: Amharic, by an unknown translator, published in Addis Ababa, Ethiopia, in 1981. The intersecting reasons for this absence of translation across the continent have been explored by Annmarie Drury: among them, that Jane Eyre was not one of the novels selected for translation by British imperial authorities; that Anglophone Africans might encounter the book at school or just read it in English; and that it is not the sort of text that has appealed to translators in post-independence African literary cultures (Drury). This picture in some ways resembles and in others contrasts with that in the Indian subcontinent where the first, abridged, translation into an Indian language happened only in 1953, done by K. Appātturai into Tamil, and published in the South Indian city of Tirunelvēli, Tamil Nadu (it was perhaps helped into existence by the successful Hindi film version of Jane Eyre, Sangdil, in 1952). A second, even more abridged translation, done by Suraiya Akhtar Begum into Bengali followed in 1977, since when interest in translating the novel, and opportunities for doing so, seem to have grown, and sixteen further translations have appeared, into Punjabi, Malayalam, Gujarati, Assamese, Nepali, Hindi, Kannaḍa, and again Tamil and Bengali (Kragh and Jain).

This very uneven distribution of translations complicates both the globality and the canonicity of what we have hitherto been calling a ‘global’ and ‘canonical’ novel. Though Jane Eyre is very widely translated, it is not universally so (indeed, there are many more languages into which it has not been translated than into which it has). What we have been mapping, then, is a particular “significant geography” (Laachir, Marzagora, and Orsini), in which many factors—including Jane Eyre’s position in the canon of English literature—intersect to open some pathways to translation and close off others.

4. Translations through Time

In Atlas of the European Novel (1998), Moretti proposed a schema for understanding the distribution of translations of English and European texts. It happens—he says—in three waves: first, nearby literary cultures (which he calls ‘core’), then a pause; then somewhat further afield (‘semi-periphery’); and after that more distant cultures (the ‘periphery’) (1). The approach builds on Itamar Even-Zohar’s polysystem theory, adopting its tripartite structure from Immanuel Wallerstein’s theorization of ‘The Modern World-System’, in which ‘core’ economies accomplish tasks that require “a high level of skill”, the periphery provides “‘raw’ labour power”, while the semi-periphery contributes “vital skills that are often politically unpopular” (Wallerstein, 350). There are obvious difficulties with the application of this economic model to literary writing: for instance, as Johan Heilbron has pointed out, “the world-system of translation […] does not quite correspond to the predominant view in world-systems theory […] Cultural exchanges have a dynamic of their own” (432). A further problem is the implicitly low valuation given to the skill and creativity inherent in translation, and to the imaginative energies of local cultures: we will return to this below. But first, does the translation history of Jane Eyre conform to the template of the three waves?

Yes—to an extent. The novel was translated into cultures well connected to English in the decade after its first publication in 1847: Germany, Belgium, Holland, Russia, Denmark, Sweden, Spain, and France. During the 1860s there was only one new translation, into Polish in 1865.5 This might be taken as the first indication of a second wave, which then began to surge more vigorously with the first Hungarian translation in 1873, followed by Czech in 1875, Portuguese in 1877 (an incomplete version serialized in a magazine), Italian in 1904, and Brazilian Portuguese in 1916. There is then what might be called a third wave during and soon after the Second World War: Argentina, 1941; Chile, 1944; Brazil and Turkey, 1945; Mandatory Palestine, 1946; Greece, 1949; Iran 1950; and Burma and India, 1953.

Yet the model does not entirely fit. It is not only in Europe that Jane Eyre is translated in the first decade of its publication: in 1850–51 a rendition of the novel appeared in Havana, Cuba, first of all serialized in the newspaper Diario de la Marina (in a form that allowed the pages to be cut out and sewed into a book) and then published as a standalone volume; the same text appears to have been reprinted in Matanzas, Cuba, and La Paz, Bolivia soon after.6 And it is not only in ‘semi-peripheral’ places that the second wave happens, but also in Japan (1896—albeit an incomplete version) and Armenia (1908). These distinctive instances (or ‘anomalies’ if we take the model as our law) suggest the importance of particular connections, interventions, and channels. For example, the translation published in Cuba came about not directly from the English text that Charlotte Brontë wrote, but from a version written by the French literary journalist Paul-Émile Daurand-Forgues, under the pseudonym ‘Old Nick’, and serialized in the Brussels-based Revue de Paris during 1849; the Spanish translation was first published in Paris, as Jay Dillon has discovered.7 The Revue announced Old Nick’s version as, not a “traduction” [translation] but a “réduction” [reduction]: the subtitle was changed from “An Autobiography” to “Mémoires d’une gouvernante”; and Charlotte Brontë’s authorial pseudonym, Currer Bell, did not appear at the beginning of the serialization, though it is given at the end of the volume, in small type, in the “Table des matières” (Forgues, vol. 4, 119, 283). In the version published in Cuba, this new subtitle was faithfully translated: Juana Eyre: Memorias de un Aya (as indeed was the rest of the text); and the author’s name was given as ‘Old-Nick’ (title page). Now, the name Currer Bell has wholly disappeared—and the name of the Spanish translator is not mentioned either.

What we can see from these details is that it is the French text by ‘Old Nick’, not Brontë’s English novel, which has been translated into Spanish and published in Havana; and that it is a connection from the Revue de Paris, via the Paris-based Administración del Correo de Ultramar to the Diario de la Marina that has enabled this to happen. Nevertheless, large chunks of Brontë’s text have in fact also been translated via this relay, for Old Nick’s text is a ‘reduction’ of a particular kind. It reconfigures the startlingly confessional first-person voice that Brontë created for Jane as an epistolary narrative, and exploits the decorum of that form to summarize the more agonised and intimately suggestive psychological passages, such as Jane’s incarceration in the ‘red-room’. But dialogue, and dramatic sections like the abandonment of the wedding and the revelation of Bertha Rochester, are translated pretty closely. Not everything is shrunk in the Old Nick version—it is not the sort of ‘reduction’ one tries to achieve for a sauce. Rather, it cuts some elements of the book’s complex generic weave, while leaving others. The result is less a novel of character and more a straightforward tale of action and romance. A lighter book, then, and one that is perhaps easier to transport from place to place. It is certainly quicker to translate and less onerous to print in a newspaper (it is about half the length of Brontë’s text). We might lament the role of Daurand-Forgues in so reducing the novel—reducing it in complexity and depth as well as length. But, by doing so, it seems likely that he enabled at least something of Jane Eyre to travel further and more swiftly than it otherwise would have done.

This example begins to reveal the intricacies and cross-currents of transmission through translation, showing the impossibility of enforcing distinctions between translations, abridgements, and freer versions, and complicating the idea that translations radiate out from the source text at the centre to ever more peripheral contexts of reception. The role of relay translation via French in the nineteenth century is well known (Cardwell, 4), but the potential for creativity in that process is less fully appreciated. It was creative work by Daurand-Forgues that enabled Jane Eyre to travel to Cuba via Paris. Here again we have an instance of the productive interaction between DH techniques and traditional bibliographic scholarship and close reading. The map made the question of how Jane Eyre got to Cuba visible and urgent, but a person had to locate and read the texts to find the answer. This answer then prompted a further question: how many of the early translations of Jane Eyre were in fact taken from Old Nick’s Jane Eyre: Mémoires d’une gouvernante. If you look at the eleven translations that appeared in the first three years of the novel’s life, i.e. by the end of 1850, you notice that three of them—beyond the Spanish one—take their subtitles from Old Nick: Jane Eyre, of Het leven eener gouvernante (Dutch, translator unknown, 1849); Jane Eyre, eller en Gouvernantes Memoirer (Danish, translator unknown, 1850); Jane Eyre: Memoiren einer Gouvernante (German, tr. Ludwig Fort, 1850). So we can see the influence of Old Nick’s choice, or at least of a shared idea that memoirs of a governess were likely to sell. Yet, when you study the books themselves, you find that the first two are in fact translated from Brontë’s English. It is only Fort’s German text that is (faithfully) translated from Old Nick’s reduced version.8 Of the remaining seven translations done in those first three years, all take their titles from the English book—either in the form of plain Jane Eyre or with Brontë’s own subtitle added: An Autobiography. Yet not all of them turn out to be translated from the English: one, Jane Eyre: en själf-biographie (Swedish, 1850), is from the Old Nick version. So we have discovered a crowded field of generic and creative impulses. It was possible to jump on the bandwagon of the Old Nick title and yet translate directly from Brontë’s text; and it was also possible to do the reverse—to keep faithfully to the evocative reticence of Brontë’s title and subtitle, and then serve up a translation of Old Nick’s reduced version. Innovation and agency intrude unpredictably, and push in different directions. Here again, the work of making the map has prompted us to find this out, even though this level of detail is too fine-grained to represent visually in the maps as we have made them (though, in the final version, it will at least be discoverable by clicking through to the bibliographic notes).

Once you recognize the agency and creativity that go into each act of translation, and the new energies that flow from it, you can see the distribution of points on the map in a different way. Not the successive delivery of goods made and packaged in the UK but, in each case, a new interaction and a remaking for fresh purposes.

Scrolling forward through the decades, we find that in some periods of the twentieth century the most intense translation activity provoked by the novel has been located not in western Europe but in Turkey, Iran, China, or Korea (Figure 4). For instance, in the 1980s there were eight new translations into Korean but none into French or German; between 2010 and 2020 there were twenty-four into Persian and again none into French (though there was one into German); across the three decades 1990–2020, by far the most translations globally were done into Chinese, with 113. In none of these cases is this merely a belated catching-up with Europe, but rather an adoption and redirection of the materials of Jane Eyre to distinctive cultural purposes.9 Though the world of translation may be traversed by three broad chronological undulations, these are choppy waters, with squalls and counter-currents pushing in varied directions. Each translation of Jane Eyre is not a passive act of reception but an active engagement; and each culture where this happens is not merely peripheral to a centre that lies elsewhere but remains in many respects a centre to itself.

The ‘Time Map’, zoomed in to show intense Jane Eyre translation activity
Figure 4 

The ‘Time Map’, zoomed in to show intense Jane Eyre translation activity in East Asia between 1980 and 2020.

5. Experiencing the Maps

The maps offer knowledge in themselves; and they prompt further research. They also, in a more experiential way, change one’s sense of what the novel is and where it belongs. No longer anchored in Yorkshire, where it was written, or London, where it was first published; and no longer held within the borders of the tradition of Eng. Lit., Jane Eyre is made visible as a trans-temporal and geographically multiple text, with many writer-translators, publishers, and readers collaborating to bring ever-new energy to its plural existence (Figure 5).

A visualization that nicely suggests the complex multi-locational, trans-temporal, and pluri-scriptorial energies of Jane Eyre in translation
Figure 5 

A visualization that nicely suggests the complex multi-locational, trans-temporal, and pluri-scriptorial energies of Jane Eyre in translation.10

The third kind of map we have created—the set of Covers Maps—plays a particular role in this reimagining. As detailed above, they are constructed using Storymap JS and allow you to follow the chronological sequence of the book covers of translations into a given language (there is also an international selection), pegged to the street addresses of the publishers who produced them. These maps are necessarily partial: often we were not able to source cover images; and the covers of most of the earlier translations were of course unrevealingly plain. Nevertheless, working with these maps, we can discern international trends: favourite images for Jane Eyre covers are a solitary woman (often the 1850 George Richmond pencil portrait of Brontë); a solitary woman reading or writing (Figure 6); a romantic couple; a young woman with a big house in the distance; the house burning; and many variants of Jane’s first encounter with Rochester when he falls from his horse. We can discern national trends (Germany has a long-standing preference for the solitary woman), and observe the global influence of film versions and BBC adaptations: a cover inspired by the 1943 Robert Stevenson film with Orson Welles appears in Buenos Aires in 1944; the 1996 Zeffirelli version prompts covers in Paris (1996), Sâo Paolo (1996), and Istanbul (2007); and the 2006 BBC series with Ruth Stevens and Toby Wilson leaves its mark again in Istanbul (2009), as well as in Colombo (2015) and Giza (2016). More than this, though, the Covers Maps help one to shift perspective from the global view of our other maps, and zoom into the particular grid of streets where a given publisher was located, prompting one to remember and try to imagine the individuals—with their lunch-breaks and commutes—who were involved in the design and production process. This combination of the global with the granular is one of our key aims for the project, and it shows through also in our approach to textual analysis.

The Arabic ‘Covers Map’ showing an interesting variant of the solitary woman theme, as published in Casablanca, 2005
Figure 6 

The Arabic ‘Covers Map’ showing an interesting variant of the solitary woman theme, as published in Casablanca, 2005.11

6. Analysing and Reading across Languages

Here again, the core of the project is formed by human beings reading books. The participants in the project got together and, through discussion of each other’s findings and ideas, identified various aspects of the source text and translations that it seemed interesting to explore collaboratively. These include: the title; addresses to the reader; handling of pronouns; key words such as ‘passion’, ‘elf’, ‘conscience’, ‘touch’, ‘strange’, ‘walk’, ‘wander’; and particular passages such as Jane’s incarceration in the ‘red-room’, the incursion of Bertha Rochester into Jane’s bedroom, and the telepathic moment at which she hears Mr Rochester call “Jane! Jane!” across the moors. These aspects of the text were not decided a priori according to some objective criterion: for instance, we did not set out to look at the most frequent words or the most-cited passages. Rather, we allowed them to emerge from our experience as readers and our judgments as literary scholars, through discussion around a table. Furthermore, we did not derive our sense of what was worth investigating from the source text alone. It was Mary Frank’s study of the distinction between ‘Sie’ and ‘du’ in the German translations that led us to look at pronouns; and Yunte Huang’s idea that the descriptions of walking were likely to be interesting in Chinese that prompted us to focus on how the novel and the translations reiterate the distinction between ‘walk’ and ‘wander’ that appears in its opening sentences: “there was no possibility of taking a walk that day. We had wandered, indeed …”. We therefore give equal dignity to the translations as acts of writing. In tracing the metamorphoses of these aspects of the text, we are interested in uncovering the variety of modes of imagining and expression that are enjoined by linguistic and cultural difference, chronological change, and the preferences of individual translators.

All this work will be presented in traditional (though open access) form in the book of the project—although there is what might be called a supplementary DH aspect to it, in that the intricate comparisons of texts across languages seem better expressed in animations online than in the static page of a book (examples are at https://prismaticjaneeyre.org/close-reading/.) The more substantial DH involvement, however, consists in the beginnings of dialogue between human reading and digital textual analysis, comparable to that between digital mapping and human scholarship described above. In this case, there is a particular challenge since, though computers have impressive powers in comparing and analysing texts in what is defined as being the same language, they are—at present—less able to look across languages. Correspondingly, while there is a substantial body of work applying DH techniques to the study of translations into a single language (Cheesman et al.; Bizzoni, Reboul, and del Grosso), there is much less that has a multilingual range (Civiliene is a distinguished exception). How best might DH be drawn into a study of translations that sees them as taking place not only into separate languages but across a continuum of language difference?

This work is at an early stage, and other scholars are exploring various possibilities: for instance, Sasha Mile Rudan and Eugenia Kelbert are pursuing the analysis of cross-language stylistic features using their Bukvik and LitTerra platforms, while Yuri Bizzoni has been working on automatic alignment of the texts and on sentiment analysis. In what follows, we show how a comparatively simple measure, lexical density, can connect with human reading and yield results. We take as our corpus a small group of texts that our work on the maps had led us to select: the first translation into French (leaving aside the Old-Nick version) by Noémi Lesbazeilles-Souvestre (1854), the first translations into Italian (anonymous, 1904) and Brazilian Portuguese (anonymous, c. 1916), and one of the two earliest translations into Spanish in Spain (anonymous, 1928).12 The lemmatizer adopted for this first phase of the investigation was TreeTagger, which was determined to have a very high accuracy (always over 99.4%) due to the fact that the printed texts were correct and well transposed into digital format (Schmid).13

First, we calculated the Type/Token Ratio (TTR), which gives an indicative measurement of the lexical variety of a given text (Youmans).14 This value varies between zero and one: as it approaches one it indicates greater lexical variety. TTR is a fairly rough measure: for example, it does not take into account all the particularities of the lexicon since it treats morphological variants as unique types. Nevertheless, its utility in making general comparisons between source texts and translated texts is well established (Fokin). We found that it could also be deployed in a more focused way, guiding a reader to zoom in on moments of particular interpretive interest across a selection of translations.

Figure 7 shows the TTR of Charlotte Brontë’s text segmented by chapter.15 It is important to note that this graph does not enable like-for-like comparison between chapters since TTR is affected by chapter-length. As a stretch of writing gets longer, it is likely that the number of tokens will increase proportionately more than the number of types (unless a writer is pursuing variety with extreme determination!). If you look at the three chapters with the greatest lexical variety in our analysis—1, 36, and 38—you find that chapters 1 and 38 are much the shortest in the book, and their high lexical variety will owe something to this fact. Chapter 36, on the other hand, is a roughly average-length chapter; and when you turn to it as a reader you can see why its lexical variety is so high. It brings together different kinds of writing, each of which has its own distinct and varied vocabulary: landscape description, a wrought-up language of emotional response, and lengthy representation of demotic speech. It is this sense of a chapter as being a unit of style within the novel that underpins the technique of looking at TTR as calculated by chapter. The alternative is to segment the book into sections of equal length for analysis: that approach gives a more accurate picture of TTR variation, but at the cost of losing the connection to the novel’s own internal divisions.

TTR of Charlotte Brontë’s English text represented chapter by chapter
Figure 7 

TTR of Charlotte Brontë’s English text represented chapter by chapter.

The real interest of the chapter-by-chapter approach, however, lies in the ready means it gives us for comparing the text across languages. Figure 8 presents the chapter-by-chapter TTR measure of the source text—let’s call it the TTR signature—alongside the TTR signatures of the translations into French, Italian, Spanish, and Brazilian Portuguese.

Comparative TTR by chapter of the English text and the translations
Figure 8 

Comparative TTR by chapter of the English text and the translations into French (1854), Italian (1904), Brazilian Portuguese (c. 1916), and Spanish (1928).

It is evident that each line roughly follows the signature of Brontë’s English, with occasional divergences. The distance between the Italian, Spanish, and Brazilian Portuguese signatures, on the one hand, and the source text signature, on the other, is most likely owing to the different TTR norms for those languages, rather than to any stylistic choice by the translators: Sergiy Fokin has presented evidence to suggest that “when translating from an analytic language into a more synthetic one, the TTR rises; translating in the opposite direction, it decreases”. By contrast, it is surprising how tightly the French line sticks to the English one, given the differences between French and English: this may be an indication of a very close translation strategy. However, what interests us for the purposes of this study are the moments where the lines diverge. In chapter 36, the sharp increase in lexical variety in Brontë’s text is only partially matched by all the translations; and there are two occasions where the 1904 Italian translation is markedly out of line: in chapter 5 its TTR increases much further than the source text, and in chapter 24 it decreases much further.

First, let us investigate the case of chapter 36. As we have seen, in Brontë’s English this chapter has a high degree of lexical variety; and our TTR graphs suggest that this is not reproduced in the translations. Is this suggestion borne out by a reading? In other words, do the numbers connect to anything that has interpretive significance at the level of style? Some examples may reveal the answer.

Once Jane has reached Thornfield, she is nervous about what she might discover when she looks at the house. Brontë conveys this anxiety with unusual vocabulary and syntax:

I advanced my head with precaution, desirous to ascertain if any bedroom window-blinds were yet drawn up.

The phrases “advanced my head with precaution” and “desirous to ascertain” are likely to strike an attuned reader as strangely elaborate, and attention to the word-choices substantiates this impression: this is the only appearance of ‘precaution’ in the chapter; ‘desirous’ and ‘ascertain’ appear on only one other occasion each; and, as it happens, the appearances of the verb ‘draw’ and the noun ‘blind’ in this sentence are unique in the chapter as well. Now, look at the French of Lesbazeilles-Souvestre:

J’avançai ma tête avec précaution pour voir si aucun des volets des chambres à coucher n’était ouvert.

The translator begins by following the English phrasing; and ‘précaution’ appears only here in the chapter, just like ‘precaution’ in English. But as the sentence advances, Lesbazeilles-Souvestre shies away from the distinctiveness of Brontë’s choices: not ‘desirous to ascertain’, but ‘pour voir’; not ‘drawn up’, but ‘ouvert’—and all those French words appear on other occasions in the chapter. So we can clearly see both how Brontë’s style here generates a high TTR, and how Lesbazeilles-Souvestre’s choices produce a lower one.

Other examples point in the same direction:

I looked with timorous joy towards a stately house.

Je dirigeais mes regards joyeux vers une belle maison.

‘Timorous’ and ‘stately’ appear only here in the chapter; whereas all of the words in the French translation are reused at other points. The representation of the innkeeper’s speech is a special case of the same trend:

Such an immense quantity of valuable property destroyed.

Des valeurs énormes ont été détruites.

‘Immense’, ‘quantity’, ‘valuable’, and ‘destroy’ are not in themselves very unusual words; but they are not common in Jane’s narrative voice, and in fact appear only here in the chapter. The same is true of ‘valeur’ and ‘énorme’; but it is ‘détruites’ that marks the stylistic difference between Brontë and Lesbazeilles-Souvestre. In the French, ‘détruites’ appears on one other occasion in the chapter and other forms of the same verb twice more, at points where Brontë reaches for a series of different verbs: ‘crashed in’, ‘fallen’, ‘wrecked’.

In this case, just as with the maps, there is a productive interaction between DH techniques and traditional practices of scholarship and interpretation. The TTR measure pointed us to a chapter that seems anomalous in both source text and translations. On studying it, we discovered that, in Brontë’s English, its vocabulary is unusually varied; and that Lesbazeilles-Souvestre’s translation does not match that variety. This conclusion is in line with many studies (building on pioneering suggestions by Gideon Toury and Mona Baker) which show that, overall, the language of translations is less varied than that of source-texts, since translators are often under pressure to conform to linguistic norms (Malmkjær and Windle, 84–86). What is new in our approach is that it allows us to identify, via DH, particular chapters within texts where the disparity is especially marked and then to understand, via human reading, why that should be. Generally, Lesbazeilles-Souvestre is good at following the shifts in Brontë’s handling of lexis: it is only in this most intensely diverse chapter that she balks.

But what about the 1904 Italian translation which—strangely—diverges in both directions, with both less variety, and more? What do we discover in this case when we look from the graph to the text?

This anonymous translation is an interesting piece of writing. It cuts throughout, reducing the original’s approx. 187,000 words to around 145,000 (another instance of the impossibility of separating ‘translations’ from ‘abridgements’); but it is nevertheless quite stylistically responsive and interpretively alert. Moreover, many word-choices make clear that this translation is derived both from Brontë’s English and from the French of Lesbazeilles-Souvestre: here, as with the Cuban text discussed earlier, we see the mediating role of French, and also the connection between the chronological and geographical story shown in the maps and the linguistic character of the translations concerned. The case of chapter 24 in the Italian translation seems continuous with that of chapter 36 in the French: it has less lexical variety than the English, and no doubt the pressure to conform which we have just mentioned has played a part in that outcome. But the graph poses a further question: why this chapter?—and why did the Italian translator respond to it differently from Lesbazeilles-Souvestre (who on this occasion, as with most of the novel, seems closely aligned with Brontë)? Again, an example can give us some answers:

Your eyebrows have become as thick as my finger, and your forehead resembles, what, in some very astonishing poetry, I once saw styled, ‘a blue-piled thunder-loft’.

Vos paupières sont devenues aussi épaisses que mon doigt, et votre front ressemble à celui d’un Jupiter tonnant.

Le vostre palpebre sono gonfie come il mio dito e la vostra fronte è densa di nubi.

Here—as elsewhere in the chapter—Jane and Rochester are sparring flirtatiously, and reaching for outré vocabulary with which to do so: you will not be surprised to learn that ‘blue-piled’ and ‘thunder-loft’ do not reappear elsewhere in the chapter. The choices made by Lesbazeilles-Souvestre are not so extreme, yet they do not diminish the TTR because ‘Jupiter’ and ‘tonnant’ likewise appear only here. Turning to the Italian, we can note some resemblances to the French, but also see how the word choices become more ordinary: for instance, ‘nubi’ appears elsewhere in the chapter. So the Italian translator moves the text towards plainness and, on occasions like this where the source text’s vocabulary is baroque, the result is a marked lowering of the TTR.

How then can we explain the opposite shift which—to judge from the graph—occurs in chapter 5? Again, let us look at the texts (the French is not necessary to the point here so we will leave it out for the sake of clarity):

‘Good-bye to Gateshead!’ cried I, as we passed through the hall and went out at the front door.

Addio, Gateshead, – dissi passando sotto il portone.

It wanted but a few minutes of six, and shortly after that hour had struck, the distant roll of wheels announced the coming coach;

Erano le sei meno qualche minuto quando un rumore di ruote annunziò l’arrivo della diligenza.

I was puzzling to make out the subject of a picture on the wall, when the door opened, and an individual carrying a light entered.

Mi studiavo di capire che cosa rappresentasse un quadro appeso al muro, quando qualcuno entrò con un lume.

In all these instances—as throughout the book—the Italian translator pursues their work of shortening, achieving a kind of Hemingwayization of Brontë’s style. But look at the kinds of words that are cut on this occasion: ‘as we passed through the hall and’; ‘shortly after that hour had struck, the distant’; and ‘when the door opened, and’: how different they are from ‘blue-piled thunder-loft’! These words are being put to subtle stylistic use in the English, registering the young Jane’s impressions; but all of them are common words, and all but one of them (‘opened’) are reiterated elsewhere in the chapter, so not translating them will generate, not a decrease, but an increase in the TTR. The Italian translator’s commitment to shortening the text is consistent throughout their work on the novel, and it is in line with the aim stated in the preface to their translation, to make it available to all—“tutti e tutte”—including mothers and girls, “madri e ragazze”. Here is a particular purpose which contributes to setting this translation apart from the others that we have considered. Yet to characterize the 1904 Italian translation as simply an ‘abridgement’ would be too abrupt for, as we have discovered, its procedures can have contrasting stylistic effects. They can lead to a flatter style, as in chapter 36; but they can also produce a more tensely concise one, as in chapter 5.

7. Conclusion

Throughout this article we have observed the productive interaction between DH techniques and longer-established processes of scholarship and close reading. Together, they have created fresh ways of perceiving translations and new tactics for analysing them. The scholarly work of cataloguing Jane Eyre translations generated the need for maps; the process of mapping turned out to require a new conception of an ‘act of translation’ (to include republication in a new place), running in tandem with translation in its more usual sense; this led to further scholarly discoveries illustrating the complicated workings of creativity and agency in the production of translations; and that, in its turn, affected our understanding of what the maps portrayed. Now we saw in them the image of an enormous, plural, trans-lingual, trans-temporal, and multi-locational text, in which the signifying potential first manifest in Charlotte Brontë’s English plays out in ever-varying ways. How, then, to set about analysing this plural text—to read the translations, not only as being done ‘into’ separate languages or national traditions, but as coexisting in a landscape of language variety? Here again, DH and human reading work well in dialogue, with the simple measure of TTR enabling comparison of texts across languages, and pointing us to moments of interpretive interest, with a corresponding literary-historical and theoretical yield. Together, digital mapping and distant reading guided close reading to form a more nuanced picture of the transformation of Jane Eyre across language difference, identifying the particular stylistic challenge of chapter 36 for Lesbazeilles-Souvestre, while also helping us so see that, though the 1904 Italian translation may be mediated by Lesbazeilles-Souvestre’s French, it is also quite different stylistically, creating its own brisk version of Jane.

These are only first steps, though we hope they may be of interest. In the next phase of the project, we will bring into the conversation more techniques, and a larger corpus, so as further to investigate the world of translations as exemplified in the case of Jane Eyre.