Language in Tech and The Creep of Conformity
When AI, social media, and coding are all dominated by English, the rest of the world must question whether conformity is at the door
In the midst of the COVID-19 pandemic in May 2020, as the entire world was dealing with the realities of being under lockdown and worrying about their physical as well as financial well-being, one institution was engaged with a very different concern. Just like everyone else, the Académie Française was discussing COVID…specifically which article was the correct gender for the term. Beyond the fact that it declared that it is, in fact, la Covid even though everyone uses the masculine article, this linguistic decision was also easy to mock given the seriousness of the context in which it was made.
While this example appears as insignificant and perhaps ridiculous, it points to a broader questioning that is taking place, challenging the utility of such academic traditional institutions. Do we still need these institutions that appear to entrench old rules of language at a time where language is under increasing flux? Do they still matter? Given the inflexibility and pomp of some of these institutions as well as the hierarchies of power that they have entrenched, it is easy to argue that they are fading in value. It is also important to point to the fact that the objectives that actors like the Académie promote, notably ‘protecting’ the French language, are regularly co-opted by extremist parties like le Rassemblement National to construct racist and discriminatory narratives about the natural evolution of language (more on this later). But at a time where two seemingly separate trends are taking place - the shrinking study of Language and Literature across the Anglo-Saxon academic sector as well as in countries like France and Germany, and the rise of artificial intelligence – an argument exists to defend the role of these institutions and the ‘protection’ of language, specifically the protection of linguistic diversity. As this piece explores, language is inherent to the dynamics of power and these dynamics are evolving even more in favour of English through artificial intelligence and the digital space.
But in order to study the power of this ‘digital lingua franca’, this piece will not treat the issue from an Anglo-Saxon perspective. Rather, it will explore what amounts to an English domination from the perspective of the language that used to dominate the world: French. For all the questions around the Académie Française, the seemingly useless stubbornness against ‘anglicisms’ and the language’s own oppressive past with the ‘Francophonie’, the clear articulation of French’s attempt at avoiding decline makes it a useful lens through which to understand both how English is becoming the sole language of tech as well as the effects of a digital lingua franca.
The Power of Language
Understood as a language used for communication between groups that speak different languages, the term lingua franca was first used to describe a French- and Italian-based jargon developed by Crusaders and traders in the eastern Mediterranean during the Middle Ages, although the concept stretches back to examples like Latin in continental Europe or Armaic for the Neo-Assyrian, the Neo-Babylonian, and the Achaemenid empires.
While the power of a lingua franca can be associated to diplomacy and commerce, recent history means that it cannot be separated from colonial intents. To a significant extent, it is the British colonial empire that established the foundation for English to become today’s lingua franca. Sue Wright explains how indigenous language was erased either by “ousting the original inhabitants”, having the survivors “corralled into reservations” as in North America and Australasia, or within the oppressor-oppressed dynamics found in plantations and mines across Africa and Asia where English became “the language of contact between rulers and ruled”. Although the older French lingua franca emerged as a result of both the Enlightenment [1] and the country’s military conquest of Europe which resulted in French becoming the language of nobility and of European courts, it too followed the same patterns as the British imposition of language, using not only violence but also exploiting processes of translation and codification. The implementation of French colonial law alongside pre-existing indigenous legal structures necessitated a codification of indigenous customs which called for the translation of codes and texts – translation can rarely be perfect but in the ‘interpretative’ space that existed between French and colonized languages, France was able to implement its control. For example, because indigenous interpretations about the right to a territory where different from how French law understood territorial rights, France was able to deliberately exploit the imperfection of translation to justify its appropriation of indigenous lands.
Education was another crucial tool in using language to implement control and domination, France opening schools across its colonies which allowed for potential opposition to be eliminated. In Sénégal, for example, the opening of secular schools (‘écoles laïques’) was a chance to regulate and challenge the indigenous religious schools to which local leaders sent their children. In one interesting linguistic case, French general and colonial administrator Louis Faidherbe opened a school destined to form future local collaborators which was called “École des fils d’otages” (School for the sons of hostages) before the name was changed to “École des fils de chefs” (School for the sons of chieftains). Whether under British or French control, the result of the imposition of language was the same: as Frantz Fanon describes, to speak the language of the oppressor is to oppress oneself as it presents the colonised culture as inferior or irrelevant to the culture of the coloniser.
Fanon’s words, interestingly, not only ring true to the populations that existed in the periphery of the empire, but also for those within the core. Here, once again, France provides clear examples as to the internal oppression that can occur through language, with Breton being actively stifled in favour of French until the mid-20th century. In this linguistic oppression, the role of schools was once again central, constructing structures that played on social dynamics to shame and punish. These structures were defined by ‘la vache’ (the cow) or ar vuoc’h which was a symbol carried by students to indicate they had been caught speaking Breton (or any other regional language that the French government was attempting to stamp out) – these symbols ranged from stones and pieces of wood to signs that had to be worn around the neck or wooden clogs. Beyond being a blatant tool for shaming, the ar vuoc’h held the more complex role of being able to shatter linguistic solidarity through the fear of punishment. For, if the shamed student happened to catch one of their classmates talking in Breton, not only was the ar vuoc’h handed over to the new offender but so was its punishment: any student unfortunate enough to still be holding the ar vuoc’h by the end of the school day faced punishment that ranged from chores, to writing lines, to seeing points docked off their homework. As Fañch Broudig explains, this would perturb friendships in the schoolyard, where the maternal language was most likely to be used.
Unfortunately, such domination is far from gone, even in the particular case of Breton. Here is where the discourse of the Rassemblement National comes back. As recently as 2022, Marine Le Pen stated that the learning of regional languages was only acceptable outside of school which not only stands in opposition to French law aimed at protecting regional languages but is also an interesting stance to take by a woman whose roots directly stretch to Brittany with her own extremist father auto-baptising himself as ‘le Petit Breton’. Undeniably, the Rassemblement National’s approach to the French language also finds its roots in more ‘traditional’ forms of oppression and colonialism, as they clamour for the ‘saving of the Francophonie’ and demand for French to be the sole official language in all of France, including its overseas territories. In the latter case, a Rassemblement National deputy in la Martinique even states that “it is out of the question” to attribute a co-official status to créole and, therefore, raise it to the same level as French. The echoes of Fanon remain ever-present over seventy years after his publication.
A Digital Lingua Franca
Having established how control over language provides power and exercises domination of populations, the question remains: what does this have to do with artificial intelligence?
The maintenance of English as the lingua franca is an important piece of context to understand because it is this act of maintenance that has a direct throughline to Silicon Valley. There is, technically, no concrete date as to when English began to be treated as a lingua franca albeit generally picking up steam around the 18th century with the aforementioned colonialism and Industrial Revolution. There is, however, a clear date for when English took over French as the lingua franca: 28 June 1919, as the signing of the Treaty of Versailles took place. After years of French being the language used for treaties and diplomacy, the Treaty of Versailles was signed in French as well as English; a result of the simultaneous effects of France having been weakened by conflict, British colonial strength but also, crucially, the United States. Indeed, the reason why English is the lingua franca today is not because of Britain, it is because of the United States.As the country that intervened in both world wars, the global economic powerhouse, the technological as well as academic hub that gave rise to inventions such as the nuclear bomb, and the soft power leader of the world, it is more so because of the United States that we all learn English today than because of Britain. The latter may have established the foundation but its spread of the language pales in comparison to the influence of American communications. Hence, with modern technology finding its launchpad in the American garages of Silicon Valley (see my previous post) – in part, thanks to the United States’ superpower status – the fact that English became the language of tech finds a throughline back to the events of 1919.
English as the digital lingua franca is upheld through a multitude of innovations that were established in English-speaking contexts (e.g., the Internet, the smartphone, social media platforms), through a vast breadth of scientific papers and their associated research/academic organizations, and through the physical localization of start-ups and companies in Silicon Valley. The most important facet of this digital language is, however, the use of English in coding. In January 2023, famed computer scientist Andrej Karpathy tweeted “the hottest new programming language is English”. Thinking in a similar vein, Nvidia CEO Jensen Huang announced in 2024 that “everybody in the world is now a programmer” thanks to AI although a perhaps more accurate statement would be that everyone capable of speaking English is now a programmer.
In a 2019 article, Gretchen McCulloch points to the English default around coding even though, in theory, a programming language can be composed out any kind of language or symbol. She highlights how, in spite of the increase in multilingual software programs and social media platforms, there remained an important lack of multilingual programming languages. Only four were widely available in multiple languages and she estimated “maybe a hundred or so” programming languages that were available in one or two languages other than English. Additionally, McCulloch comments how newer languages like Python, Ruba, and Lua may have originated in non-English speaking countries but they all still rely on English-based keywords.
What McCulloch points to is that, while the outputs may be multilingual and allow as many consumers to engage with a given product, the basic building blocks behind these technologies often remain bound to the self-imposed constraints of using English. So, although the dominance of English over online social media content has decreased with the expansion of platforms to non-English speaking countries, the true power – that is associated to designing and building these technological tools – remains in the same hands it always has. The rise of China has undeniably presented as a challenge to the English dominance over tech but when considering the complexity of Mandarin characters as well as the language currently being limited to one country, even if it is one of over 1 billion people, the argument that Chinese may replace the reach of English in coding feels somewhat weak at this time.
The relationship of coding to English has, in turn, become a stepping stone to another area of linguistic domination: artificial intelligence. The colonial dynamics of artificial intelligence as a whole, be it in the appropriation of land and drinkable water, in the exploitation of cheap labour in the Global South, or the possession of as much data as possible, is increasingly discussed across the field. In addition to the more classical areas of colonialism, the question of language re-emerges just as it did with the empires of past. In her book, Empire of AI, Karen Hao argues that “large language models accelerate language loss” with only fifteen languages being supported by OpenAI’s GPT-4 above an 80% accuracy. She continues that, as these models become foundational pieces of digital infrastructure, it means that accessibility for communities outside of these fifteen most developed languages is reduced which shrinks their access to economic opportunities while incentivizing them to prioritize learning and speaking English. In doing so, the language’s grip on the lingua franca status only continues to strengthen.. But this then brings on another issue: linguistic conformity threatens conformity of thought as it promotes a singular way of expressing oneself – boundaries form around how the world is seen, interpreted and shared.
Dealing With The Conformity of It All
This question of conformity brings the discussion full circle, back to the Académie Française. Such an institution cannot be separated from the dynamics of power and domination that are exercised through language – be it in France, Britain, Portugal or Spain, these academic bodies are, in one way or another, participants in the colonial projects that continue to have repercussions today. But these institutions, albeit more generally reactive than proactive, are also those that codify positive change into society. Be it in the (sometimes begrudging) acceptance of slang such as ‘bled’ that entered French vernacular through the immigration of Algerian families into the banlieues or the research conducted as to the feminisation of the language, the Académie Française can also be seen as a benchmark of societal progress and the inevitability with which such progress seeps even in the most entrenched mindsets. So, through this ability to codify change, linguistic institutions do have a responsibility nowadays to protect language.
But language must not be protected in the way that extremist parties claim; to protect a language as an excuse for protecting historical power relations will serve only the very people promoting discriminatory rhetoric while erasing markers of history. Such protection of language means the erasure of others and once that process has started, it is difficult to turn it around. For some, a language like Breton, which was commonplace only two generations ago, now faces a ‘daily struggle’ in ensuring its survival. Language disappears with greater ease than one might expect and the continued grip that English has on increasingly crucial digital services will only facilitate that further.
Linguistic institutions are right in believing that they have a responsibility towards not just a language but a culture, a way of thinking, and a way of being. They are right in worrying as to the impacts about the rise of English-dominated technology. But they must adapt to this new era of linguistic change and understand that what they are ultimately protecting is not language supremacy but language diversity. Hao concludes her book by discussion Te Hiku Media as an example of how artificial intelligence can be developed without a colonization of data, space and resources. Focused on “Māori language revitalisation” as well as archiving and training, Te Hiku is an excellent example of how technological progress can protect language diversity. There are ways to continue moving forward without following the pre-determined path; it’s just a matter of choosing to step off.
[1] It is worth noting that the expansion of English as a lingua franca is also tied by some to the fact that the origins of the Industrial Revolution took place in Britain but the reliance of this revolution upon colonial structures further cements that fact that English linguistic expansion was, in one way or another, tied to aspirations of empire