The Inversion of Babel
On monoculture, polarization, and binary systems.
Diversity in language use has always been a passion of mine. In my home region, I still speak the dialect of my youth. At work, I speak the standard Flemish of the media and education. In my hometown just across the border with the Netherlands, the language sounds a bit more Dutch. I also love the linguistic differences in the bilingual newsroom of our press agency in Brussels. I enjoy the broken English of international meetings. I savour my wife's Nigerian English and the weird street slang my children sometimes speak. In that multiplicity, I felt at home. Every layer has its own pace, its own double meanings, its own untranslatable words. In recent years, however, a new layer has been added: the smooth, efficient language of the machine.
That language is seductive. It is always available, never faltering, never searching for the right word. But as I read those texts, I notice something else: the sentences look alike, whether they are written in English, Dutch, French, or a local Nigerian language. The words are there, but the soul is gone. It is as if, in a single decade, we are rebuilding the failed Tower of Babel, and this time not with bricks, but with clicks, tokens and bytes.
In this essay, I trace the echo of Babel in three movements. First, I return to antiquity, where the story of the Tower of Babel can be read as a kind of border control against human hubris.
Next, I look at our digital Inverse Babel: the promise of the singularity, the rise of linguistic entropy, and the binary trap of platform logic.
Finally, I search for a counterforce in the present: independent journalism as a defender of the space in-between, where human faltering and doubt are still allowed to exist.
I. The Myth You Think You Know
The Tower of Babel is a story many people have heard. Humanity, united by a single language, builds a tower reaching to the heavens. God takes revenge on their ambition, creating a confusion of tongues by making everyone speak a different language, and scatters humanity to all corners of the world. A story of hubris, of multilingualism, and of diversity as a punishment for mankind.
But Babel can also be read as a warning against centralization, against the dangers of absolute uniformity. This is what the rulers of Babel say in Genesis: “Come, let us build ourselves a city, with a tower that reaches to the heavens, so that we may make a name for ourselves; otherwise we will be scattered over the face of the whole earth.” The builders are not naive villagers; they are the first urban planners, engineers, and strategists. Their dream is not just a tower, but a monoculture: one language, one mankind, one name.
Theologians and political thinkers see this as an early critique of centralization. British theologian Dan Strange speaks of “God’s problem with centralized power”: Babel is a project in which religious and political power converge in an infrastructure designed to neutralize any deviation. In a Lutheran reading by Michael Laffin, Genesis 11 is read as a warning against totality, against systems that no longer recognize diversity. The intervention, the confusion of tongues and the dispersion, then acts as a safety mechanism. A protection against totalitarianism, a pluralistic correction.
Comparative mythology shows variations worldwide. Sumerian tales like “Enmerkar and the Lord of Aratta” recall a time when “all people spoke one language,” until a god confounds their tongues.
The Quran relocates the tower to Egypt: Pharaoh orders Haman to build a high tower to climb up to the God of Moses, a story in which hubris and divine correction are central, and which is adopted into Islamic narrative traditions featuring dimensions of the confusion of tongues. In Greek culture, the god Hermes confuses languages. Among the Mexican Toltecs, there is a story of people building a high zacuali, a tower, to protect themselves against a new flood. The deity confused their languages and scattered them across the earth. The Aztecs have similar stories of people building high pyramids, after which the gods destroy the top, scatter the builders, and confuse their language. In the Polynesian Maori tradition, men build a great structure into the heavens; the god grows angry, tears it down, and changes their language so they speak in different tongues. Among the Lozi in Zambia, they build a tower to reach Nyambe, the sky god, but they fail and are scattered.
The Tower of Babel is thus not just a story about language, but an archetype of what happens when a culture believes that total unity and total control are the path to salvation.
II. The Digital Inverse Babel
Fast forward to the 21st century. We are no longer building a tower of bricks, but an immaterial structure of code, data, and models. Where ancient myths warn humans not to “want to become like gods” in knowledge, power, and immortality, the singularity vision sells precisely that package as progress: superintelligent AI, radical life extension, and data immortality.
In this context, large language models emerge as the machines undoing Babel. They bridge language barriers, offer real-time translation, and speak neatly in all the languages and dialects of the world. At first glance, they seem to fulfill the dream that Babel started with: a single communicative layer for all of humanity.
But under the hood, they work differently. LLMs optimize for probability: the most likely next word, in the most likely order. They are “machines of the average,” as stated in a recent study titled “Generative Monoculture in Large Language Models” by Fan Wu, Emily Black, and Varun Chandrasekaran.
This has two effects.
The first effect is linguistic entropy. Studies on AI-mediated English show how generative systems reinforce standard varieties of English and marginalize regional or minority languages. A study by Lana O'Sullivan published on Reset.org speaks of a technological monoculture in which large models perform particularly well in English and a few dominant languages, while other varieties flatten out or disappear. In algorithms, most languages become a museum version of themselves: correct, recognizable, but lifeless. You hear the machine speak with your mother's tongue, but the soul, the rhythm, the layering, the untranslatable is gone.
The second effect is the binary trap. Digital systems are not only probabilistic on top, they are also binary at the bottom. Interfaces force decisions into yes/no, like/dislike, follow/unfollow. Various corpus studies of online comments show that algorithmic environments amplify divisive rhetoric: statements that divide the world into friend and foe generate more engagement, and thus more visibility. Linguistic and rhetorical analyses point to a shift toward antithesis forms in digital communication: a structure that suggests clarity but reduces nuance to two poles. I have indeed begun to notice how abnormally often the antithesis pops up as a stylistic device everywhere, appropriate or not, in recent digital writings.
LLMs mirror this logic, which was initiated years ago by social media. They are trained on corpora where these binary patterns are already dominant, and they learn to reflect that style back.
In many AI texts, you see the same rhythm: first a safe antithesis, then a statement. In these texts, you rarely see genuine doubt, the faltering, the detour. The rhetorical richness of tricolons, epiphoras, polyptotons, unexpected metaphors, or openly expressed dubitatio is replaced by a smooth middle style that applies everywhere, and therefore truly belongs nowhere.
However, while language is pulled toward the center, the political and social structure shifts toward the extremist edges. Complexity research by U.S. Political Science professor Jenna Bednar shows that democracies are more robust when there are many different fault lines, diverse scopes and multiple checks and balances.
Once the information sphere is compressed into a single dominant axis, reduced to one binary dimension, the system becomes more vulnerable to autocratic capture.
Hannah Arendt described totalitarianism as a project that destroys plurality and installs an all-encompassing narrative, where citizens experience “organized loneliness”: apparent unity on the surface, deep isolation, and strong hidden resentment leading to explosive polarization. A digital Tower of Babel creates no hive mind, no unity. The tower creates flattening and opposition.
Digital infrastructure can reproduce exactly that pattern. When generative monoculture and platform dominance ensure that we speak roughly the same middle style in all languages, while the underlying interaction is forced onto binary tracks, we approach a digital Babel: not a cacophony of languages, but a monotonous choir dividing the world into two camps.
III. The Human in the Space In-Between: Journalism as Counter Architecture
In such a world, language becomes infrastructure. The question “is this sentence fluent?” loses meaning, because every actor, every state, every company, every troll farm, every hobbyist, has the same generative toolbox.
The distinction between human and synthetic prose blurs.
The crucial question then becomes: where does this signal come from, who created it, and with what intent?
With these kinds of questions, we return to the space in-between, the place where doubt, friction, and dialogue reside.
Journalism fulfills exactly this role. To use that contagious antithesis as a stylistic device: not as a content factory to produce even more efficiently, but as an unmasker, an awareness-raiser, a signaller of what is really going on.
Emily Kubin, a German social and political psychologist, writes in “Human Communication Research” that a simple form of “polarization literacy” in news consumption, an explicit contextualization of polarizing content and situations, can reduce affective polarization.
An American study conducted by political scientist Curtis Bram, specializing in polarization and political psychology, discusses an experiment with “underreported news” and shows that when readers are consciously confronted with relevant stories that their own information bubble pays little attention to, issue polarization decreases and viewpoints become more nuanced.
Engagement and solutions journalism projects show that newsrooms consciously steering away from horse-race politics toward substantive, community-oriented reporting contribute to fewer “us/them” frames in public perception.
At the systems level, UNESCO and the Council of Europe emphasize that media pluralism and editorial independence are a necessary component of our democratic infrastructure. Due to the omnipresence of binary algorithms on a limited number of global media platforms, the spectrum of observable voices shrinks, paving the way for a narrative monoculture. Studies on authoritarian-populist contexts show that this effect is further amplified by media concentration in the hands of players directly or indirectly controlled by the state.
The "Democracy Shield" of the European Commission, the first comprehensive, flagship strategy for strengthening democratic resilience within the EU, explicitly starts from the assumption that independent media increase the resilience of democracies against disinformation, platform monopolies, and political interventions.
The journalist is, and remains, primarily a delegated witness. Someone who says: “I, a human of flesh and blood, traveled to this place, spoke to these people, saw these documents, checked this, and asked these questions, and I put my name and reputation on the line to verify this account.”
In a world where language itself is cheap and reproducible, value shifts to origin and responsibility. Journalism thus becomes the defense of the vital space in-between: the delay between signal and acceptance, the space where questions can be asked: who is speaking, why, with what power, and with what agenda?
Perhaps that is the deepest irony of Babel. The confusion of tongues forced people to meet each other, to translate, to misunderstand, and to try again. Our digital Inverse Babel threatens to liquidate that confusion in favor of a smooth, global middle style with binary tracks underneath.
The rulers of Babel wanted to make a name for themselves, similar to today's corporate and political rulers. The rulers of Babel succeeded in conquering a place in our collective memory, but not as winners. The winner was diversity: a chaotic mix of languages, cultures, and traditions. So let's embrace the space in between us. It's what makes us human.
Synthetic image/AI-generated
This blog is written by Patrick Lacroix in a personal capacity. AI tools are used for research, structuring, drafting and language support. All content is selected, verified, and edited by the author, who retains full editorial responsibility.

