{"id":251,"date":"2013-10-04T14:48:08","date_gmt":"2013-10-04T18:48:08","guid":{"rendered":"http:\/\/blogs.law.harvard.edu\/andresmh\/?p=251"},"modified":"2013-12-08T02:58:42","modified_gmt":"2013-12-08T07:58:42","slug":"multilingual-interactions-through-machine-translation-numbers-from-socl","status":"publish","type":"post","link":"https:\/\/archive.blogs.harvard.edu\/andresmh\/2013\/10\/multilingual-interactions-through-machine-translation-numbers-from-socl\/","title":{"rendered":"Multilingual Interactions through Machine Translation\u2014Numbers from Socl"},"content":{"rendered":"<p>For the past two years, social media platforms have been rolling out\u00a0<a href=\"http:\/\/en.wikipedia.org\/wiki\/Machine_translation\">machine translation<\/a>\u00a0in the hopes of\u00a0enabling\u00a0multilingual interactions. However, the people interacting in these platforms often know each other already, and have a language in common (i.e., friends). But what happens when machine translation is used to facilitate interactions\u00a0<em>among strangers<\/em>, who perhaps have common interests but\u00a0<em>not a common language<\/em>?<\/p>\n<p>The earliest social media platform to enable machine translation was probably Facebook, which began\u00a0<a href=\"http:\/\/mashable.com\/2011\/10\/06\/facebook-translation-tool\/\">autotranslating conversations in Facebook pages<\/a>\u00a0(a good place to start\u00a0given that Pages are more likely to bring together heterogeneous languages). Likewise,\u00a0<a href=\"http:\/\/mashable.com\/2013\/08\/20\/google-translate-plus\/\">Google+<\/a>\u00a0and\u00a0<a href=\"http:\/\/allthingsd.com\/20130701\/twitter-testing-bing-powered-translation-feature\/\">Twitter<\/a>\u00a0later released similar features, enabling, for example, Spanish-speaking Twitter users to read the\u00a0<a href=\"https:\/\/twitter.com\/MuhammadMorsi\/status\/352164641352327168\">tweets from the now toppled Egyptian president<\/a>\u00a0Muhammad Morsi, translated from Arabic to Spanish:<\/p>\n<p><a href=\"http:\/\/metaremixing.files.wordpress.com\/2013\/10\/morsi.png\"><img decoding=\"async\" id=\"i-81\" src=\"http:\/\/metaremixing.files.wordpress.com\/2013\/10\/morsi.png?w=554\" alt=\"Image\" \/><\/a><\/p>\n<p>&nbsp;<\/p>\n<p>How often do these types of multilingual interactions occur, though?\u00a0<a href=\"http:\/\/civic.mit.edu\/blog\/erhardt\/ethans-five-questions-about-mapping-attention-at-links-2013\">Ethan Zuckerman posed<\/a>\u00a0a similar question when wondering how often people use their browsers&#8217; machine translation to pay attention to content outside their immediate reach.<\/p>\n<p><!--more--><\/p>\n<p>With that in mind, we decided to look into some numbers using data from our own social media platform:\u00a0<a href=\"http:\/\/so.cl\">Socl<\/a>,\u00a0which started offering machine translation\u00a0<a href=\"http:\/\/blog.fuselabs.org\/post\/16769656635\/introducing-the-new-so-cl-translator\">since last year<\/a>.\u00a0Socl, like Twitter, often brings strangers together who might not speak the same language, example:<\/p>\n<p><a href=\"http:\/\/metaremixing.files.wordpress.com\/2013\/10\/socl.png\"><img decoding=\"async\" id=\"i-82\" src=\"http:\/\/metaremixing.files.wordpress.com\/2013\/10\/socl.png?w=447\" alt=\"Image\" \/><\/a><\/p>\n<p><img decoding=\"async\" title=\"More...\" src=\"http:\/\/wordpress.com\/wp-includes\/js\/tinymce\/plugins\/wordpress\/img\/trans.gif\" alt=\"\" \/><\/p>\n<div><strong>Multilingualism in Socl<\/strong><\/div>\n<div><\/div>\n<div>In the\u00a0<em>3 months of Socl data\u00a0<\/em>we looked at,\u00a0we found more than 6,000 multilingual posts: threads like the one above, where the language of one or more of the comments, or the thread-starter, were different\u2014presumably representing people being able to communicate with people in other languages through machine translation.<\/div>\n<div><\/div>\n<div>We found that most multilingual threads (85%) are contain two languages, and the remaining 15% have 3 or more languages, up to a handful of threads with 5 languages in one single thread.<\/div>\n<div><\/div>\n<div><a href=\"http:\/\/metaremixing.files.wordpress.com\/2013\/10\/langnum.png\"><img decoding=\"async\" id=\"i-84\" src=\"http:\/\/metaremixing.files.wordpress.com\/2013\/10\/langnum.png?w=650\" alt=\"Image\" \/><\/a><\/div>\n<div><\/div>\n<div>Furthermore, the majority of multilingual posts involved English and some other language, with\u00a0<em>English-<\/em><em>Portuguese\u00a0and English-Spanish<\/em><em>\u00a0being the most common pairings\u00a0<\/em>among bilingual threads:<\/div>\n<div>\u00a0<a href=\"http:\/\/metaremixing.files.wordpress.com\/2013\/10\/langpairs1.png\"><img decoding=\"async\" id=\"i-87\" src=\"http:\/\/metaremixing.files.wordpress.com\/2013\/10\/langpairs1.png?w=650\" alt=\"Image\" \/><\/a><\/div>\n<div><\/div>\n<div><\/div>\n<div><\/div>\n<div>These \u00a0numbers reflect the demographics of Socl itself, as almost half of the visitors come form outside the US, mainly from Brazil, India, and Germany.<\/div>\n<div><\/div>\n<div>It is important to note that these numbers are produced using automatic language detection, which, while it has improved a lot in the past few years, still fails when dealing with emoticons and other unusual Internet lingo.<\/div>\n<div><\/div>\n<div>More work is needed to understand the degree to which machine learning can support deep cross-language communication, but providing seamless automatic translations appears to be working across different platforms. That said, language is just one barrier, cultural is a much more difficult one to address, especially through algorithmic methods.<\/div>\n<div>\n<p><em>Many thanks to\u00a0<a href=\"http:\/\/elenaagapie.com\/\">Elena Agapie<\/a>,\u00a0James van Eaton, and Bruce Haly, for helping with this post.<\/em><\/p>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<p>For the past two years, social media platforms have been rolling out\u00a0machine translation\u00a0in the hopes of\u00a0enabling\u00a0multilingual interactions. However, the people interacting in these platforms often know each other already, and have a language in common (i.e., friends). But what happens when machine translation is used to facilitate interactions\u00a0among strangers, who perhaps have common interests but\u00a0not [&hellip;]<\/p>\n","protected":false},"author":3887,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3112],"tags":[],"class_list":["post-251","post","type-post","status-publish","format-standard","hentry","category-internet-culture"],"jetpack_featured_media_url":"","_links":{"self":[{"href":"https:\/\/archive.blogs.harvard.edu\/andresmh\/wp-json\/wp\/v2\/posts\/251","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/archive.blogs.harvard.edu\/andresmh\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/archive.blogs.harvard.edu\/andresmh\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/archive.blogs.harvard.edu\/andresmh\/wp-json\/wp\/v2\/users\/3887"}],"replies":[{"embeddable":true,"href":"https:\/\/archive.blogs.harvard.edu\/andresmh\/wp-json\/wp\/v2\/comments?post=251"}],"version-history":[{"count":2,"href":"https:\/\/archive.blogs.harvard.edu\/andresmh\/wp-json\/wp\/v2\/posts\/251\/revisions"}],"predecessor-version":[{"id":264,"href":"https:\/\/archive.blogs.harvard.edu\/andresmh\/wp-json\/wp\/v2\/posts\/251\/revisions\/264"}],"wp:attachment":[{"href":"https:\/\/archive.blogs.harvard.edu\/andresmh\/wp-json\/wp\/v2\/media?parent=251"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/archive.blogs.harvard.edu\/andresmh\/wp-json\/wp\/v2\/categories?post=251"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/archive.blogs.harvard.edu\/andresmh\/wp-json\/wp\/v2\/tags?post=251"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}