8 Minutes To Read

Shooting the Facebook Messenger (Part III)

8 Minutes To Read

Thomas Dowling presents an argument for how Facebook could potentially (and inadvertently) advance Burmanisation in Myanmar.

This post is Part Three of a four-part series. Part I introduces and provides pertinent background for the topic. Part II considers and evaluates some of Facebook’s mitigation strategies and outlines some issues that remain problematic. Part III argues that Facebook’s exclusive preference for Burmese at the expense of all other ethnic languages runs the risk of increasing Burmanisation. Part IV offers some reflections and concluding remarks, ultimately arguing for a more nuanced understanding of Facebook’s presence in Myanmar

Blocking (Ethnic) People: The Burmese Linguistic Monopoly

Modern Myanmar currently comprises of 135 recognised ethnic groups. Debates over the validity of this number and whether the Rohingya should be included are left for other papers, what concerns us here is language. Many of these groups speak their own—Chris Myers states that there are over 100 different languages. Not all, however, possess a written equivalent of their spoken words.

Unsurprisingly, this ethnic and linguistic proliferation makes Myanmar one of the most diverse countries in Asia. Non Burmese speaking peoples, however, are not equally represented. The Bamar population dominate not only the Defence Services but also politics, minimising the political role of the other 134 minority groups.

But these ethnic groups do at least, to varying degrees, have a stake (however small) in the political apparatus of the State. This is not the case on Facebook which is exclusively available in Burmese: not Shan, nor Kachin, and of course, one can forget about the disenfranchised, stateless Rohingya and their Bengali dialect. The Bamars have a linguistic monopoly. When considered from this perspective, Facebook looks less like an important democratic driver but rather a megaphone for the dominant, privileged Bamar majority. To contextualise this, Burmese is just one of the over 100 languages Facebook is available in; coincidentally, roughly equating to Myanmar’s own linguistic diversity.

At first sight, this linguistic issue might not seem like much, especially to outsiders with little knowledge of Myanmar – a label Facebook and Zuckerberg evidently wear. However, language matters a great deal. Language can be equated to power and dominance of one group over another: Latin in Imperial Rome; Greek in the Byzantine world; English within the British Empire, and Japanese during their occupation of Korea offer a few broad, simplistic, reductionist examples. Advancement in these cultures often rested on adopting the language (and sometimes religion) of the rulers in the realms of administration and the army. Whether Facebook was—or is now—conscious that Burmese is not the only available language option for this pluralistic society should be discussed more in Menlo Park. Failure to do so, I contend, will ultimately do more to harm Myanmar’s rich ethnic tapestry as surely as other forms of State-sponsored Burmanisation have done in the past, including, but not limited to conflict, displacement (and replacement with Bamars in so-called ‘model villages),’ supplanting Buddhism as the State religion, and significantly reducing the teaching of ethnic languages in primarily non-Bamar states and regions.

Again, in fairness to Facebook, and with Warofka’s update in mind, the social media giant asserted the difficulty of encoding Burmese language characters on Unicode—the international text encoding standard—when other tech giants bucked at the problem. Warofka wrote ‘[t]his lack of a single standard poses some very real technical challenges for us and others. It makes automation and proactive detection of bad content harder, it can weaken account security, it means less support for languages in Myanmar beyond Burmese, and it makes reporting potentially harmful content on Facebook more difficult. Resolving compatibility issues is also important for Myanmar’s technological development and its economic growth.’ These problems can explain some of the issues moderators have hitherto encountered in detecting hate-speech. Steve Stecklow offers insight with some examples: ‘The Reuters article […] showed that the translation feature was flawed. It cited an anti-Rohingya post that said in Burmese, “Kill all the kalars that you see in Myanmar; none of them should be left alive.” […] Facebook had translated the post into English as “I shouldn’t have a rainbow in Myanmar.”’ Evidently these sorts of challenges were missed by Facebook’s grossly understaffed team at the time. With such translational issues with the dominant language in Myanmar, it is not surprising that Facebook would not want to embrace the maelstrom of providing their services to other ethnic languages.

This feels like an appropriate juncture to note the original genesis of Shooting the (Facebook) Messenger. The work herein was inspired almost exclusively by Olivia Solon’s opinion piece for the The Guardian some months ago. Solon wrote:

“[Facebook] also said that one of the reasons it couldn’t moderate content effectively was because users weren’t using its reporting tools. This might have something to do with the fact that those reporting tools – including the text in drop-down menus attached to objectionable posts – were only translated into Burmese in late April/early May this year.”

It was these few sentences that caught my attention in what was a solid, though ground-well-trodden piece. Personally, I—and perhaps others did also—had always assumed that Facebook in Myanmar, warts and all, was in Burmese (prior to Solon’s article I never even thought to question the willful jettison of all those other ethnic languages). With this in mind, it was startling to discover that it was only in the early part of the year that Facebook’s reporting tools and drop-down menus became available in Burmese.

Why does this matter? It matters because Burmese is not the only language in Myanmar. While language changes regarding reporting tools and drop-down menus are further positive steps to improve things in Myanmar by Facebook, this move, yet again, favours Burmese-speakers, most obviously the Bamars themselves (though it’s fair to say those non-Bamar who speak Burmese would also be privy to the same advantages). Reporting tools and drop-down menus favour the Bamars precisely because it means that they can more easily report abuse or hate-speech online, but it remains as difficult as ever for speakers of other ethnic languages who either do not speak Burmese or would otherwise prefer to use their own language when uprising the same services on the platform as the Bamar majority. Thinking about this, we must ask whether a Kachin jade-miner in Hpakant, or a Pastor in Shan State, or a Rakhine farmer in Sittwe has the same access to Facebook, its services, and abuse-report mechanisms as a Bamar soldier on deployment in Kayin State, civil servant in Naypyidaw, or rickshaw driver in Mandalay? Patently no. Facebook’s platform and mitigation strategies are distinctively Bamar-friendly.

While Burmese is undoubtedly the main language in Myanmar, as Facebook itself correctly identifies, it is not a lingua franca. The ethnic regions continue to speak their own words and are officially recognised by the State (indeed, an aptitude to speak one of Myanmar major languages is a crucial requirement of citizenship). As articulated above, there are more than 100 languages now available on Facebook, some of which have relatively few speakers. For instance, Maltese and Corsican have 400,000 and 200,000 speakers respectively, that they are catered for. In contrast, the 3 million or so Shan speakers in Myanmar’s Northeast are not, nor are the 900,000 Kachin and their language Jingpho. Nevertheless, Facebook does offer the means to request a language that is not currently offered, and can be accessed via this link. As a first step, I have requested that the Shan language be considered owing to their sizeable speaker-pool and as a partial antidote to the further dominance of Burmese in the country. (Given that many problems with Burmese script have been overcome, Facebook has less of a technical and engineering case to argue; they’ve already set a precedent for tackling Myanmar’s primary language.)

Facebook and Zuckerberg refer to the country as ‘Myanmar’ in their correspondence and public addresses, which in some respects may well figure into some of the issues. Leaving name-related debates over ‘Myanmar’ and ‘Burma’ to one side, I would argue that Facebook have failed to grasp one of the most important aspects of the former – it’s full name: The Republic of the Union of Myanmar. I emphasis Union here precisely because it is not a homogenous or mono-linguistic  nation in the way South Korea is (where my family and I reside). Modern Myanmar comprises of many peoples and many tongues, and it is in this arena where Facebook should concentrate its next round of efforts to make their platform not only safer and freer from hate-speech, but also more linguistically representative, inclusive, accessible, and ultimately, democratic.

My main concern—as I’ve attempted to articulate in this Part III, is that by adding those report mechanisms, drop-town tabs, and whatever other future innovations they may develop in the Burmese language at the expense or ignorance of Myanmar’s other ethnic peoples—is that Facebook continues to declare, in effect, that Burmese is the language of the Myanmar nation. In this way, Bamars and Burmese are posed to be (more) synonymous with the country in the way that Facebook is synonymous with the Internet. Indeed, it is accurate to say that both Bamars/Burmese and Facebook are dominant in their respective spheres, but Myanmar is not (yet)  entirely encapsulated by either. One could equally make the case that Myanmar is defined by its ethnic diversity; meanwhile, other telecommunications and social media companies are slowly perforating the market to increase competition. In a country of 51-55 million people, as many as 30-35% are not ethnic Bamars – that’s at least 12 million people, or roughly the population of Ohio  State in the US. Those 134 other ethnic groups have their own cultures, traditions, and languages that enshrine their own unique identities. Consequently, the exclusivity of Burmese as the language option, inclusive of the new drop-down menus both favours the Bamar majority, place further social disadvantages upon Myanmar’s non-Bamars.

Solon’s piece highlighted a dangerous reality in Myanmar concerning report mechanisms and drop-down menus Burmese, but also shown light on a grander problem: the dominance of Burmese on the platform at the expensive of all others. Language in this case does not simply reflect political dominance of one over many, it is a key cultural marker and signifier of identity. One need only consider the Rohingya to appreciate how language marks out one group from another.

Arguably, and I concede this might be highly contentious to say, Facebook’s linguistic omissions of Myanmar’s other main languages (to say little of the dozens of smaller ones) is potentially more damaging than the hate-speech that Facebook has been charged with facilitating. Given the speed of communication via social media, hate-speech can—and evidently has—led to outbreaks of violence and greatly exacerbated conflict, particularly in Rakhine State. And while the social media platform was used to pollute Myanmar’s pluralistic air with a fervent, chauvinistic Buddhist nationalism that will likely linger for some time, the short history of Facebook in Myanmar illustrates that hate-speech ebbs and flows, shown in part by Raymond Serrato’s study; it is potent and deadly as the events in 2012, 2013, and after August 25th 2017 make plain, but is not permanent.

I contend that Facebook’s simplistic viewing of Burmese as the de facto language of Myanmar, whether because of ease, cost, or technical hurdles, makes the company partly responsible for furthering the primarily Bamar-administered state’s long-held ambition to Burmanise the country. The cost of this could be the eventual loss of Myanmar’s ethnic diversity and cultural uniqueness: Burmese is not Myanmar. I am not qualified to appropriately judge whether the casualties inflicted from hate-speech are more or less preferable than the death of entire cultures, but I suspect that many of the people belonging to these ethnic groups would prioritise the longevity of their own distinctiveness. Further or increased Burmanisation via Facebook would put that in jeopardy for minorities more so, in my opinion, than even the most-virulent strain of hate-speech. Moreover, I would also suggest that such a global force like Facebook showing such preference for Burmese inflates the narrative of Bamar superiority precisely at a time when the Tatmadaw are fighting insurgent groups in Rakhine, Shan, Karen, Chin, and Kachin states.

Myanmar’s many ethnic groups are not blocked from Facebook’s services nor report mechanisms, but without the appropriate language skills they are at a significant disadvantage. Facebook and its services distinctly favour the majority Bamars—just as most other aspects of life do.

Given all that we have discussed in the previous three Parts, is it time to shoot the messenger?

Thomas Dowling is a Ph.D student with the University of Leicester. His primary research interests revolve around environmental security in Myanmar (particularly in the context of human security), viewed through the prism of securitisation theory. Thomas is also well-travelled in Myanmar, and lived in Taunggyi, Shan State for a short while. Previously, Thomas earned degrees in Ancient History (BA, MA; Bristol University) and International Security Studies (MA; Leicester University). Presently, Thomas lives in Daegu, South Korea, with his wife, baby, and Jack Russell.