I

Luther began his translation of the New Testament in 1521, having gone into hiding at Wartburg Castle after facing trial for heresy following his excommunication – a result of his publication of the 95 Theses, in which he criticized the church’s practice of selling indulgences (i.e., forgiveness of sins) and questioned the religious authority of the Pope. The Reformation was thus already well underway when Luther the exegete became Luther the translator. And yet Luther’s work as a translator was so tightly bound up with his work as a theologian that his challenge to the church might not have been possible without also challenging the canonical status of the Latin Vulgate – the official version of the Bible used by the church throughout the medieval period, translated in the 4th century on the orders of the Pope. But if Luther broadly sought to return Christianity to its early message by putting the individual relationship to Scripture above the mediative authority of the church and its traditions, he in many ways also followed in the footsteps of two church fathers, whose work in the late 4th and early 5th centuries exerted a profound influence on Christianity throughout the medieval period. These figures were Augustine, the great exegete, philosopher, and mystic, and Jerome, the translator of the Vulgate.

Albrecht Dürer, “Der heilige Hieronymus im Gehäus,” 1514

Discussing Luther’s exegesis and its relationship to his translation, Bluhm describes Luther as combining “the genius of two of his most distinguished predecessors in these two fields,” with “the depth of Augustine” and “a literary skill greater than Jerome.” [3] While each stood in a certain metonymic relation to the church, Luther can be described as having built on their legacies at the same time as challenging them. Luther developed a theology based on Augustine’s doctrine of grace – essentially a rejection of the idea that man is free to choose between good and evil (aka Pelagianism), with Augustine arguing that the good works needed for salvation can only come from faith, delivered by God’s grace. Luther radicalized this idea by arguing that faith alone (sola fide) brings redemption, making any attempt at earning justification through virtuous deeds spiritually dangerous. This tenet marked Luther’s theological break with the church and remains the central principle of Protestantism to this day. But it is also at the heart of his legacy as a translator, and on multiple levels.

In 1530, a few years before publishing his translation of the entire Bible, Luther wrote a letter in which he vehemently defended himself against critics of his translation of the New Testament (first published seven years earlier) and explained his approach. [4] At the core of his argument was a passage in Romans, in which Paul (its author) states that man is justified not by good works but “by faith alone” (“allein durch den Glauben”). This led to accusations that Luther added the word “alone” to make Paul’s words fit his own theological framework, since no equivalent existed in either the Vulgate or the original Greek. In his letter, after railing against the “donkeys” who criticized his translations without themselves knowing how to translate, Luther states that he had “wanted to speak German, not Latin or Greek” when translating, arguing that his choice of words reflects how Paul’s message would naturally be expressed in German as it was spoken by “the mother at home, the children on the street, the man at the market.” [5] In defending his translation, then, Luther did not so much refute the idea that he had inscribed his theology into his translation as affirm it, by invoking his own understanding of the Bible’s message and the idea that Scripture should speak to the faithful in their own voice. And with this, he might be described as having picked up and radicalized an argument prefigured in another letter, every bit as combative as Luther’s, written more than eleven centuries earlier by Jerome.

II

Around 395, Jerome wrote a letter to his friend Pammachius in which he, like Luther, defended himself against criticisms of his supposedly overly free style of translation; in this case, the dispute centered not on Jerome’s Vulgate but on a piece of correspondence he translated from Greek into Latin privately for a friend, which was subsequently made public by an adversary. [6] Considered the “founding document of Christian translation theory,” [7] Jerome’s letter was also the first text to expound upon the distinction between word-for-word and sense-for-sense translation, which is first attributed to Cicero and forms a tension that expresses itself in different ways throughout the subsequent Western tradition. But if in some ways Jerome’s letter directly prefigures Luther’s in its core argument, then at the same time it contradicts it. For after stating that “I not only admit but freely announce that in translating […] I render, not word for word, but sense for sense” and evoking a line of classical Roman scholars in defending this approach, Jerome then immediately adds an exception for Scripture, in which “even the syntax contains a mystery.” [8] While the distinction Jerome draws here between sacred and non-sacred text is not contradictory in itself, his letter does subsequently contradict this argument by citing examples from the New Testament in which the Apostles themselves seem to take a freer approach when translating passages from the Hebrew Old Testament into Greek. Like Luther, then, Jerome might also be described as facing in two directions, torn between a classical Greco-Roman culture that emphasized translation’s value as a pedagogical or rhetorical tool and an ecclesiastical culture in which word-for-word translation had normative authority in matters of Scripture – and where being seen to adulterate the word of God was consequently liable to bring charges of heresy. But the ambiguity in Jerome’s argument is also enlightening in a broader sense, insofar as it points to the way that a fundamentally reductive binary, inherited from classical culture, has often served to structure other models and arguments that complicate or even contradict it.

Benozzo Gozzoli, “Agostino legge la lettera di San Paolo ai Romani e si converte,” 1465

One such example comes from the translator and translation scholar Douglas Robinson, who describes Augustine and Jerome as having initiated an ongoing ascetic tradition within translation, in which the translator is expected to “empty him- or herself of personal desires in order to achieve the neutral, impersonal transfer of the author’s meaning to a reader in another language.” [9] Each of them typifies a distinct strain of this ascetism, whose structure mirrors the two modes of monastic life: Where Augustine’s cenobitism is “impersonal, perfectionist, and systematic,” Jerome’s eremitism is inward-looking, pragmatic, and “riddled with internal contradictions.” [10] Translation, here, is viewed in terms of temptations to be resisted, from the desire to cling “‘too’ closely to the source-language text” to the urge to “strike off ‘too’ boldly in a new direction.” [11] But while the cenobite is primarily concerned with resisting the latter and the eremite the former, the primary distinction between the two ascetic modes is actually one of orientation: Where the cenobite turns to the community and its traditions in interpreting the word, the eremite seeks an “unmediated relation” to the truth beyond it. [12] The difference might also be described as one between how translation is theorized and practiced – between an idealized view of language and one that knows from experience what it means to struggle and doubt and still find a solution. If Jerome marked the beginning of the eremitic tradition in translation, then, Luther surely brought it into the modern era. In emptying himself of desire and seeking an immediate relationship with the meaning beyond the text, he could not help but write himself into his translation. And in doing so, he created not just a new Bible but a new version of the German language, infused with his theological perspective and “the vitality of his own sense of self.” [13]

For all of Luther’s skill as a translator, the fact that his Bible exerted such a strong influence on the development of the German language as a whole was also and above all a result of the forces unleashed and accelerated by the invention of the printing press. Latin’s supreme sacral status within Western Christianity was already under threat by the time Luther entered the scene, as humanist Bible scholarship increasingly turned to Greek and Hebrew as the languages of exegesis, and new markets emerged for religious texts written in vernacular languages. [14] But if print capitalism provided the necessary conditions for Luther’s Bible, he perhaps more than anyone fully harnessed its “revolutionary vernacularizing thrust,” which played a central role in the rise of national consciousness in Germany as elsewhere. [15] In bringing the language of Scripture closer to language as it was spoken, then, Luther helped spawn a new era in which the ultimate power of the aristocracy and the church began to give way to that of the nation-state – an era that now seems to be coming to an end, driven by a technology whose power to reshape human society, culture, and thinking seems just as great as that of the printing press.

III

If the momentous technological shift brought about by the rapid advancement of artificial intelligence threatens to reverse the democratizing effects of the printing press by making states and their citizens increasingly dependent on the digital infrastructures owned by a tiny class of tech barons, [16] then the most immediate question for translation as a concrete field of activity is what role human translators might play in an order in which the effort and expense of their work is likely to become increasingly unjustifiable in the majority of contexts. For those interested in translation as a linguistic phenomenon, this also raises the question of what normative frameworks will emerge in this new context, and what influence they will exert on language more widely. While the ability of LLMs to translate so effectively would be impossible without the fundamentally subjective experiences and decisions of the authors and translators whose work served as training data for their models (or without the efforts of those tasked with manually fine-tuning them), it is impossible to predict how these algorithms will adapt to and shape dynamic norms of equivalence between languages in the long run – norms that remain fundamentally grounded in individual human interpretation, even when they become embedded in technological systems such as those of the book, the internet, or an LLM.

That these norms are often largely invisible to us in our own time is in their nature. And yet now, as ever, a tension remains. While reader-oriented or familiarizing approaches now dominate in the majority of contexts, within translation theory, fluency in the target language has often been seen as coming at the cost of the identity of the original work and language. This idea is most prominently associated with the early Romantic theologian and philosopher Friedrich Schleiermacher, whose theory of translation has come to stand at the center of a modern hermeneutic tradition. But if Luther is in many ways paradigmatic of the “domesticating” method that is now routinely (if somewhat reductively [17] ) contrasted with the “foreignizing” one retrospectively attributed to Schleiermacher, [18] then at the same time he also embodies many of modern translational hermeneutics’ key principles: that translation is dependent on the subjective understanding and experience of the translator; that the translator’s interpretation of the source text is dependent on individual, cultural, and historical factors, just as the reader’s interpretation of the translation is; that translators require creativity in overcoming both the differences between languages and the gaps in their own understanding; and that in overcoming these differences and gaps, translators establish temporary norms that then have the potential to be adopted more widely. [19]

This modern hermeneutical perspective on how equivalence is constructed between languages emerged in the second half of the 20th century in line with the rise of machine translation and natural language processing, as successive attempts at expanding and diversifying the translation unit beyond the word level led to a recognition that translation must be viewed holistically if it is to be understood properly, as a complex and fundamentally contingent phenomenon that exceeds any single theoretical framework. [20] But it takes on new significance in the context of a technology whose inner workings are partly opaque by nature, and whose ability to map linguistic form and meaning across a dizzying number of dimensions (in the case of Chat GPT, 12,228 precisely [21] ) raises profound ethical and epistemological questions for translation, many of which ultimately boil down to “Who decides?” Normativity, here, remains as tightly bound up with questions of power as ever, but with a new sense of epistemic urgency. Because for all of LLMs’ seeming ability to synthesize or emulate an endless number of different modes, styles, and voices, the evidence so far suggests their overall effect is actually one of homogenization. [22] There is also the crucial question of what linguistic variation means when difference becomes a matter of probability and stochastics, [23] rather than of individual experience, interpretation, and expression. The issues AI poses to translation are thus bound up with a wider threat to the diversity of human thinking in general. If translators are to play any part in resisting this flattening of difference, it will not involve the neutral, impersonal transfer of meaning from one language to another. If two millennia of theorizing and practicing translation have proven this to be an impossible ideal, then this impossibility has also seemingly often provided the basis for translation’s power to create new meaning. At this moment in history, then, the big question might be what happens if translators stop trying.

Ben Caton is a translator, editor, and author based in Berlin.

Image Credit: Public Domain.

Notes