The Geopolitics of Context: Mordor, Russia and Google Translate

By Pip Thornton

“One does not simply walk into

the Russian Federation”

Over the last week several media outlets, and many more Twitter feeds, have been spreading news of a series of ‘glitches’ in Google Translate which saw the word Russia being synonymised with Mordor when translated from Ukrainian to Russian. Furthermore, Russians became occupiers and for a short time the name of the Russian Foreign Minister Sergey Lavrov returned the result sad little horse. 


Twitter / Vadim Nakhankov (

Noting that ‘the terms mirror language used by some Ukrainians following Moscow’s annexation of Crimea in 2014′, some suggested the possibility of foul play; that the algorithm was ‘hacked by spies‘, ‘jokers’ or ‘mischievous pro-Kiev activists’,  or that the words had been inserted manually by users as alternative translations presumably in order to ridicule Russia and humiliate Lavrov. Other sources referred to a ‘bug’ or an ‘automated error’ in the algorithm, an explanation seemingly substantiated by the way Google quickly issued an ‘embarrassed apology’, stepping in to ‘fix’ their wayward algorithms as soon as the matter came to light.

However, although these anomalies could in theory have been caused by a programming error, by ‘hacking’, user manipulation of data, or indeed from ‘inside’ the company, the statements which Google issued to the press seemed simply to reiterate what we already know about how Google Translate works; that it does not actually translate from language to language in any traditional way, but relies on the corpus of ‘big data’ available on the web in a particular language to provide most likely matches based on the frequency and proximity of words and phrases. The algorithm has no way of knowing whether individual words are idiom, metaphor, or a literal translation; as Google stated, it simply ‘looks for patterns in hundreds of millions of documents…  the meaning of words depends on the context in which they’re used’.

“The algorithm has no way of knowing whether words are

idiom, metaphor or a literal translation”

Whether or not the data has been manipulated, the ‘glitches’ which emerge from Google Translate are therefore not down to a ‘bug’ or an ‘error’, in fact the algorithms are doing exactly what they are programmed to do. They are merely reflecting the current usage of language based on the available data and as such – for all its clunkiness and drawbacks – Google Translate might be a better indicator of local politics and trends than other Google products such as Search and Analytics, which have the added burden of optimisation and market forces to contend with. If we are talking about the extraction of useful and real-time trends from Big Data, then the relatively low-tech Google Translate might even be a better place to look than platforms such as Twitter or Facebook, precisely because of their heavy curation and perhaps incomplete user datasets.

So why does the data available to the Google Translate algorithms contain so many Lord of the Rings references? It turns out there are in fact several reasons why a Ukrainian text might use the word Mordor in reference to Russia. It is a term which has, over the last few years ‘crept into official and semi-official pronouncements of Ukraine’s political elite: ministers, spokespersons for the government, and the like’, and the Ukrainian president Petro Poroshenko himself has compared the Eastern Ukraine area of ‘New Russia’ to Mordor. In his 2014 article Putin’s Russia, Tolkien’s Mordor: What’s the Difference?, Leonid Bershidsky notes how Russia was popularly likened to Mordor as far back as 1961, a translator of Tolkien’s text suggesting that Sauron was based on Josef Stalin – a suggestion refuted by Tolkien himself. Bershisky also notes Putin’s apparent resemblance to Sauron, and mentions a rumour that ‘a design company was planning to light up an enormous Eye of Sauron over a Moscow skyscraper’ in 2012. Such is the cultural conflation, that if the words Russia and Mordor become interchangeable in the limited amount of Ukrainian text which Google Translate has to work on, then ‘glitches’ like this are bound to happen.


Sergey Lavrov (

And what, then, of the sad little horse? One explanation comes from a Russia Today report which points to ‘an offensive monument to the Russian Foreign Minister which was erected in Ukraine’s southeastern city of Zaporozhye where the diplomat was caricatured as a horse’. The monument can be seen here from a photograph which was shared widely on Ukrainian social media and on Twitter, a platform whose content is privileged in search rankings and is often reproduced exponentially as pictures, tags and phrases ‘go viral’.

lavrov real 2

Sad little horse (

Of course there is the possibility that the anomalous glitches could have been deliberate, and similar ones could easily be replicated with the appropriate amount of effort; it is after all relatively easy to ‘play around’ with Google, as some Google Map hackers and Googlebombers have shown, although presumably it is becoming harder to do so anonymously as more and more Google services require logging in. According to some it is not just for the purposes of embarrassing public figures or making jokes about places that Google services can potentially be manipulated. In 2013 security experts were so convinced they had uncovered a secret communication channel in Google Translate that they contacted the US intelligence industry. In his article Lorem Ipsum: Of Good & Evil, Google & China, Brian Krebs explained how for a short time in August 2013, the cod-latin placeholder text Lorem Ipsum returned some ‘apparently geopolitical and startlingly modern phrases’ when translated from Latin to English, churning out words like China, NATO and internet. Was this a ‘hidden in plain sight’ spy network? Or maybe an ingenious way to bypass the censorship of China’s firewalls? According to Krebs, some experts remain convinced there was a clandestine system in place; their convictions strengthened by the abrupt manner in which the translations stopped soon after their discovery.

However, just as with the Russia/Mordor incident, there is a far more prosaic explanation for why this might happen, and it reveals far more about the types of website which use Lorem Ipsum than it does about covert operations. The web only holds a finite amount of real Latin text and corresponding translations, so when the Lorem Ipsum text is used as a placeholder which corresponds with the content of a website which can be viewed in different languages, in effect diluting the word-stock of that language, then the algorithm knows no better than to match the ‘Latin’ word to the English one. As Lorem Ipsum tends to be used as placeholding text on the multi-lingual websites of official organisations, governments, diplomatic pages or multinational businesses, so the ‘translations’ will reflect the language used on those sites.

Once again, it is all down to context; the way words move through and are ordered within digital spaces, which is what I am trying to explore in my PhD. But what is crucial to my thesis, and what ultimately explains both the Russia/Mordor and the Lorem Ipsum incidents, is that while the proximal context of digitised words is imperative to how the web functions; the literal, or semantic context is largely irrelevant. Platforms and tools such as Google Translate and Google Search (which is where the focus of my research lies) use words simply as their raw materials; without them the algorithms have nothing to work on and no means of generating profit. My argument is that this reliance on decontextualised words (or the code or tags which represent them) is having a subtle yet profound effect not only on the literary integrity of language, but on the wider discourse too. Some of the ‘glitches’ generated by algorithmically mediated language seem harmless enough, for example the funny phrases which autocomplete often throws up, or the intriguing, yet ultimately innocuous Lorem Ipsum example. But others, for example ones which reflect dominant content which may itself be racist or sexist, can not only offend, but also compound and disproportionately magnify certain stereotypes or agendas.


screenshot author’s own

The example I have used before to illustrate this point in Google Search is a set of autopredictions for the phrase David Cameron is… the top one of which was generated for many months after a photograph of a house in Leicester went viral on social networks in 2013.

Despite the humour and misspelling (which has perhaps already been autocorrected), a political statement has nevertheless made it onto a platform (i.e. Google) which over 90% of the UK population use on a daily basis and on which many people rely, if not as a source of truth or knowledge, as a source of confirmation and solidarity. Going back to Mordor, however, although both the Sergey Lavrov and the David Cameron ‘glitches’ have been generated by similar (de)contextualised algorithmic processes,  the difference between these two examples is the geopolitical context in which they occur.

“Google has escalated a local war of words

into a global diplomatic incident”

While Google calling the Russian Foreign Minister a sad little horse might make for an amusing news story here (and seemingly in parts of the Ukraine), what the automated translation tool has inadvertently, but effectively done is to magnify and escalate a localised war of words into a international diplomatic incident. It seems, therefore, that context matters at both ends of the scale; at the granular level of encoded and sorted linguistic data, right up to a global geopolitics of context.  I hope that with these examples I have been able to show how the instability of decontextualised and digitised language, untethered by semantic meaning, yet structurally fundamental to modern communication and capital production, has potentially significant and dangerous consequences to the stability of an entire region.




One thought on “The Geopolitics of Context: Mordor, Russia and Google Translate

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s