AMODERN 8: TRANSLATION–MACHINATION

Christine Mitchell, Rita Raley

Amodern 8 explores the contexts and implications of translation as mechanism, media, technique, and transmission. Our tethering of “translation” to “machination” marks our intention to move beyond the habit of situating MT and computer-generated language in the familiar crisis poses of fakery, treason, and inauthenticity. Rather than regarding the machine as marking the limits of translation – an assumption that risks walling off translation practice from media and communication studies concerns, while still absorbing its products – our aim is to continue to investigate the possibilities and configurations of translation as machined, and translation as machining meaning, historically and in the contemporary moment.

Because “translation” can so easily be used metaphorically as a way to theorize all manner of signal recordings and transmissions, the term forsakes some of its practical solidity as attention shifts away from translation-as-it-happens to other media, other technologies, and other concerns. The metaphorical borrowings apply not only to translation, but to language as well. While it can certainly be generative to expand the translational scope to other media transformations, for the purposes of this special issue we have chosen to retain the more precise usage. In sum: we simply still care about language as a particular kind of medium – and in our contemporary computational environments, questions about what language is, how it moves, and how it is used, manipulated, maintained, transformed, and understood seem more pressing than ever before.

Our contributors grapple with such machinations by revisiting, uncovering, and extending machine translation’s historical record; by close reading translations, not primarily or only as hermeneutic inquiries, but as analyses of process and procedure; and, in a similar vein, by (re)activating machine translation development and operation through prototyping, re-creating, and experimenting with its devices and programs.

LOVE.ABZ AND THE TASK OF THE MACHINE TRANSLATOR

An Interview with Otso Huopaniemi

Rita Raley, Otso Huopaniemi

love.abz (2011-2015) was a series of performances that began when Finnish playwright and performance scholar Otso Huopaniemi ran his play, An ABZ of Love, through Google Translate, repeatedly translating from Finnish to English in a recursive loop. This poetic experiment in machine translation then evolved into a Babelian performance work that used live acting and live machine translation to explore the fragility of language and our ongoing relationship with machines, as well as questions of textual stability, semiosis, and authorship. love.abz was performed in New York, Berlin, Cardiff, Karlsruhe, and Helsinki, with several different casts and using a variety of languages. In this extended dialogue, Huopaniemi and Rita Raley together consider "the task of the machine translator."

THE TRANSLATION OF PROCESS

John Cayley

There are practices of literary art distinguished by the compositional use of software. As a constitutive aspect their writing, computation may generate text from a store of linguistic elements or reconfigure previously composed texts. When asked to translate a work of this kind, we are confronted by actual ‘prior inscriptions,’ programs, that played some part in its writing. What critical or interpretative priority should we give to these inscriptions? Should we translate these programs? If so, how and when? "The translation of process," however, addresses liminal cases of works that are accepted as literary, possess existing translations, and which also instantiate program-like processes.

WHETHER SOMETHING WORKS

Finding the Human in the Margins of Machine Translation

Christine Mitchell

The success of statistical methods for computerized translation is often recalled as the victory of empiricist data analysis over rationalist rules-based approaches. The new methods just “worked.” This paper adds nuance to historical claims by revisiting events that dramatized the methodological opposition. What’s uncovered is the human in the machine loop of translation; while new assessment practices paved the road to technological success, they were justified by the paradoxical expectation that human language learning would underwrite machine translation functionality.

OPTOPHONIC READING, PROTOTYPING OPTOPHONES

Tiffany Chan, Mara Mills, Jentery Sayers

This article details the contributions of blind readers to the development, design, and marketing of the optophone, a text-to-tone transcription machine introduced in the early twentieth century. We combine archival research with prototyping to investigate the dimensions involved in past coding and decoding practices. If archives provide testimonial fragments about individual use, 2D to 3D translation helps scholars to more broadly characterize optophone reading and understand technical affordances.

CRYPTO-MONOLINGUALISM

Machine Translation and the Poetics of Automation

Avery Slater

This article tracks the rise of two separate Cold War endeavors: machine translation and computer-generated literature. Technocratic imperatives to circumvent the unhomogenizable nature of linguistic plurality influenced machine translation’s theory and design from the outset. Thus, the machinic ingenuity that makes all languages in the world accessible to every other actually works to intensify polylingualism’s outmoding. Couched in terms of universal intelligibility, this latently monolingual fantasy constitutes an ideology of “crypto-monolingualism.” Placing the machine that writes alongside the machine that translates, this comparative historical analysis of machine translation and computer-generated literature also meditates on how neural-networked machine translation aligns with powerful global incentives to accelerate technologically-assisted monolingualism.

THE CRYPTOLOGICAL ORIGINS OF MACHINE TRANSLATION

From al-Kindi to Weaver

Quinn DuPont

I argue that the origins of machine translation are to be found in the archeology of cryptanalysis (or “code-breaking”). I trace this archeology through the cryptanalysis and translation practices of medieval Arabic scholars, the machine translation efforts of Renaissance and Early Modern universal language planners, the cryptanalysis techniques developed by NSA chief cryptanalyst William Friedman, and through Warren Weaver’s “Translation” memorandum. Together, these histories reveal some of the ways that various machineries were developed to “crack” unknown text, be it a foreign language or ciphertext. As such, this history also points to a deeper, ontological relationship between language and cryptology.

TRANSLATING THE FUTURE

Transpilers and the New Temporalities of Programming in JavaScript

Andrew Pilsch

This essay explores the intersection between machine translation and contemporary software development. Specifically, the article considers ES6 transpilers, JavaScript utilities that translate the as-yet-unimplemented future standard for the JavaScript language (ES6) into executable code. These programs translate from a language that might exist in the future into one that exists in the present, mirroring several conversations about contemporary translation (including work by Rita Raley and Michael Cronin) and writing practices (most prominently Vilém Flusser). Flusser in particular sees digital writing as a process of translating the future into the present in dialogue with machines, and this essay argues that ES6 transpilation is an early example of this new temporal rhythm.

THE COMPLEXITY OF POETIC PATTERN

Recreating Early Work in Machine Translation

Nick Montfort

A 1961 system by Victor Yngve, developed to be part of a machine translation project, produced unusual and provocative sentences. Although this generator was made for a utilitarian translation pipeline, it was eventually essentially claimed (although not by its developer) as an avant-garde poetry generator. In 2016, as an engagement with the history of computing, digital art and literature, and machine translation, I re-implemented this system in three different forms, including as a Commodore 64 BASIC program. I printed this program’s output on a dot matrix printer and presented it at a poetry reading; the three programs I developed were also made available as free software.

TRANSLATE AND DESCRIBE

Archive-Based Image Description as an Intermedial Translation Technique

Jane Birkin

The archive-based translation of image to text is explored in relation to the direct and literal translation of texts between languages. A visual content-based description applies institutional language and record-keeping methodologies, cataloguing visual elements and spatial relationships. It operates, like interlingual translation, within a system of shared knowledge and linguistic structures. Creativity and intuition are perceived as problematic to both translation processes and analysis of meaning is not a requirement. My set-piece description of a news image is provided as an example of translation in practice. It is also the vehicle for a metatranslational discussion, through which a case is made for raising the status of the “secondary” text.

TRANSLATION AND OPTICAL MEDIA

Spirit-Channeling in Edward FitzGerald’s Rubáiyát of Omar Khayyám

Karin Littau

Drawing on Victorian visual studies, this essay addresses the ways in which Edward FitzGerald co-opted optical media into his translational practice and theory for his numerous renditions of the Rubáiyát of Omar Khayyám. The aim is to show that his stance on translation – like that of his contemporary Dante Gabriel Rossetti – while imbued with older cultural beliefs and practices, such as the transmigration of spirits and souls, is also strikingly contemporary when it comes to one of the key media technologies of his age: the magic lantern. The focus on the little noticed connections between translation and optical media is intended as a step towards a more comprehensive media history of translation that pays attention to translation not only in the familiar contexts of oral and written cultures but also in visual and screen cultures.

TRANSLATIONAL GRANULARITY

Joe Milutis

What is the relation of translation to transmission? How does mechanical reproduction introduce the possibility of a translation that goes beyond alphabetism, into the realms of image and noise? As we see in W. G. Sebald’s use of Xeroxes, the granularity of a text, a form of utter specificity, is a sign of its existential untranslatability and an indication of the totality of mutation. This translational materiality neither requires dogmatic postmodernism, nor a return to presumably “humanist” bromides about fidelity to a text’s meaning. Rather, an embrace of the imperfection inherent in a text’s transmission and reproducibility allows for a kind of superfidelity to the constantly shifting ground of translation.