Freedom of expression is not an absolute of modernity. It is a conquest that lives within a context, not a good suspended in a vacuum. It changes when the means by which we speak to each other, the public spaces, and the way we imagine ourselves as a collective change. We are still rightly terrified of formal censorship, which is the status quo in much of the non-Western world. But it is no longer the only or the main one in democracies. Alongside attempts to limit what we can say, another front has emerged. An information ecosystem that decides whether those words circulate or vanish.
The history of freedom of expression is the history of the conditions that give power to speech. To understand this, one must go back, before social media and television, to when speech was a concession, not a right. In the modern sense, it was born with the Enlightenment. For centuries, speech was a territory monitored by the sovereign, the Church, or the community. Saying something unwelcome meant risking reputation and life. Galileo saved himself by abjuring. Bruno did not. They were not exceptions: speech was not a right, it was a revocable favor.
The turning point comes when the individual becomes a bearer of rights, not a recipient of permissions. It is the French Revolution that translates that principle into a political act. Speaking without asking for authorization means feeling like an active part of the world. Freedom implies responsibility and action. But when a principle enters reality, it deforms. The principle is universal, the terrain is not. In 1793, the Revolution guillotined Olympe de Gouges, author of the Declaration of the Rights of Woman, for opposing the Jacobins. Freedom was universal in words, selective in deeds.
With the nineteenth century, the mental landscape of societies changes. Industrial press, nation-states, and literacy transform speech into a political infrastructure. The press does not just inform. It creates communities among people who do not know each other but share news and imagery. Publics that exist primarily in shared perception. Opinion becomes measurable, directable, mobilizable. It is a new actor of power, which no sovereign had ever had to face. Newspapers arrive designed to orient the world, not just to tell it. Before, one convinced a king, now one can act on the pressure of the population. Papers like Il Risorgimento in Italy or Le Petit Journal in France show how information is already a battlefield. The first newsrooms do not just collect news: they build narratives that switch consensus on or off. Meanwhile, the conditions for the twentieth-century leap mature: diffusion, literacy, politicization.
The leap comes with the totalitarian century. Radio, cinema, and industrial press create a continuous information flow. For the first time, the population lives immersed in flows that count more than direct experience. In 1922, Walter Lippmann brings it into focus. The pseudo-environment, the mental image constructed by the media, becomes more influential than direct experience. Totalitarianisms do not limit themselves to censoring, they construct the perceptual context itself. Here one sees environmental power: shaping the world in which citizens believe they live. Mass societies become sensitive to communicative oscillations; consensus ceases to be an accessory and becomes a condition of power. Bernays in 1928 formalizes it by defining propaganda as “engineering of consent”. Totalitarianisms apply it on an industrial scale. Goebbels builds a media ecosystem where the German population sees no alternatives to the Nazi narrative: he does not just censor, he reorganizes the perception of the real and shapes the symbolic frame with which society interprets itself.
The postwar period brings television, the last common Western space before digital fragmentation. It combines video, simultaneity, and capillary diffusion, but remains a controlled space. Few channels, schedules decided from above, access limited to those with resources or power. The audience is common but passive: it receives, does not produce. Television returns a unitary image, built by the verticality of the medium. The 8 PM news as a collective ritual, the moon landings followed by hundreds of millions of people simultaneously, great sporting events as moments of national identity. Television created the illusion—real in its effects—that the nation was watching itself at the same moment.
With the Internet, that common world dissolves. Digital reverses everything: from centralization, we move to dispersion. That unity had costs—verticality, passivity—and should not be regretted. However, it had a shared horizon. The dream of the Internet as a new public space seems shattered, faced with the constellation of micro-spaces we have today. Each with its own memory and its own rhythm.
Fragmentation is only the first step. Humans no longer inhabit these spaces alone. There are automated networks: bots that simulate emotions and conflicts, AI that generates infinite streams of contents. Public opinion becomes a hybrid entity, where human and artificial confuse. Speech does not disappear because it is censored but because it loses its foothold. It no longer finds a common ground on which to settle.
During the 2016 American elections, the Russian Internet Research Agency managed thousands of automated accounts that generated political content on Facebook and Twitter. There was no need to eliminate critical voices. It was enough to submerge them under an artificial flow that simulated consensus or dissent. On TikTok, the algorithm decides in fractions of a second whether content will live or die. A video can reach ten million people or stop at one hundred views, not due to intrinsic quality but due to opaque engagement criteria. The difference between visibility and oblivion does not depend on a censor who blocks, but on a system that selects according to its own logics. It is not classic censorship but environmental selection. Content formally exists, but without a context to support it, it vanishes as if it did not exist.
Social media reorganizes discourse according to its own logics such as speed, conflict, emotion, attention capture. There, with a growing part of the “public” composed of automatic entities, the individual loses not only visibility but also political weight and orientation. The environment reacts artificially, amplifying what is convenient and ignoring what disturbs. Speaking in this context means losing the link with the real because synthetic networks build illusions of consensus and dissent. The debate appears formally intact but is emptied inside.
In democracies, it is no longer so much the State that limits speech, and not mainly through classic censorship. Control passes through the configuration of the private context—because social media are private—in which speech circulates. Censorship does not disappear but coexists with a power that acts before the content, shaping the place where a word can make a mark or dissolve. It is a less visible power, not less incisive.
Freedom, where it exists, thus enters its most ambiguous phase. Formally everyone can speak, concretely the algorithm decides. And paradoxically, the appeal to freedom of speech is used to obscure the new mechanisms that render it ineffective even without forbidding it. In an environment where narratives can be replicated, amplified, or distorted infinitely, there is no need to impose a truth nor to censor, nor to do propaganda in the classic sense. It is enough to alter the context in which reality takes shape. If every version of the facts circulates with the same force, the multiplication of narratives makes reality irrelevant. It is no longer a question of imposing a lie, but of dissolving the very possibility of a shared truth.
The paradox is evident. The individual freedom to publish is enormous, but the collective capacity to build a shared reality is more fragile. Not because someone forbids, but because the common ground crumbles and everything competes with blows of stimulus. The front passes between those who defend a readable space and those who live off its chaos.
How is an ecosystem under pressure defended? Regulating platforms risks reintroducing forms of vertical control that resemble censorship. Who decides what to regulate, with what criteria, and who watches over the regulators? Digital literacy is necessary but structurally insufficient when systems are designed to exploit cognitive biases. One cannot ask individuals to resist mechanisms optimized by engineers to maximize addiction. Non-commercial digital public spaces, Mastodon, federated platforms, independent forums, exist but remain marginal, unable to compete with the resources and network effect of large platforms. The answer is not a single solution: freedom of expression is defended by recognizing that the context in which it lives is never neutral. Those who control algorithms, infrastructures, and amplification rules exercise political power even without censoring a single word. The stake is guaranteeing that speaking still makes sense. Defending freedom of expression today means defending the conditions that make it possible.