Uniform multilingualism: A media genealogy of Google Translate

Ido Ramati*, Amit Pinchevski

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

10 Scopus citations

Abstract

This article applies a media geneaology perspective to examine the operative logic of Google Translate. Tracing machine translation from post–World War II (WWII) rule-based methods to contemporary algorithmic statistical methods, we analyze the underlying power structure of algorithmic and human collaboration that Translate encompasses. Focusing on the relationship between technology, language, and speakers, we argue that the operative logic of Translate represents a new model of translation, which we call uniform multilingualism. In this model, the manifest lingual plurality on the user side is mediated by lingual uniformity on the system side in the form of an English language algorithm, which has recently given way to an artificial neural network interlingual algorithm. We conclude by considering the significance of this recent shift in Translate’s algorithm.

Original languageAmerican English
Pages (from-to)2550-2565
Number of pages16
JournalNew Media and Society
Volume20
Issue number7
Early online date22 Aug 2017
DOIs
StatePublished - 1 Jul 2018

Bibliographical note

Funding Information:
Weaver’s (1955b) allegory suggests the mathematization of languages as a way to reach a universal common ground. Extricating a mathematical logic would allow numerical-based computers to work with humans’ natural languages, and getting “down to the common base of human communication––the real but yet undiscovered universal language” (p. 23). Revisiting this idea in the field’s first collection of essays in 1955, Weaver suggests seeing machine translation as a “Tower of Anti-Babel.” While this new tower is not intended to reach to the sky, it does hark back to “that mythical situation of simplicity and power when men could communicate freely together” (Weaver, 1955a: vii). Weaver’s Anti-Babel promises reversing the old curse of translation through mechanical means, but as was the case with the mythical Babel, this inverted model similarly entails eradication of divergence in favor of “simplicity and power.” The impulse toward uniformity within universality is also articulated in Weaver’s reference to Russian in his letter to Wiener: seeing Russian as English in disguise is a commitment to seeing the familiar within the foreign, thereby providing the means for rendering the language of the other as “our” language. This version of universality is also present in Weaver’s involvement in establishing international research collaborations, which he deemed as a “world-wide fraternity of scientists” which might advocate for world peace. However, as the Cold War became a geopolitical reality, executive officers at the Rockefeller Foundation sought to make this vision “a pro-American and anti-Soviet mold” (Geoghegan, 2011: 109). Machine translation projects were increasingly supported by government, military, and intelligence agencies, which collectively saw them as a part of the race for scientific superiority between the United States and the USSR (Gordin, 2015). The first public demonstration of machine translation from Russian to English in 1954 drew the attention not only of government power brokers, but also of the Soviets, who immediately initiated their own machine translation projects (Hutchins, 2000).

Publisher Copyright:
© The Author(s) 2017.

Keywords

  • Actor–Network Theory
  • Google Translate
  • algorithmic culture
  • media genealogy
  • neoliberalism
  • participatory culture
  • translation

Fingerprint

Dive into the research topics of 'Uniform multilingualism: A media genealogy of Google Translate'. Together they form a unique fingerprint.

Cite this