Dictionary Theory
Russian information confrontation with the west
My DPhil examined the co-construction of Russian post-Soviet identity during Putin’s third presidential term, less a study of propaganda than of why it resonated, in both supply and demand terms. Why do certain narratives land? What needs did they meet? What interpretive infrastructure made them feel true rather than imposed?
Since 2022 it has been difficult to avoid the pull of an inverse question: if you understood how the Kremlin’s dictionary was built and why it worked, could you use the same understanding to work against it? That question has found little appetite in the West, sometimes for ideological reasons, more often simply because the field has been overwhelmingly focused on defence. How do we protect ourselves from Russian disinformation? How do we identify it, label it, counter it?
I think we are getting even that defensive fight wrong. And I want to try to explain why, albeit with the caveat that Western-facing Russian information operations are not primarily my domain. My work has focussed on Russian domestic propaganda and, in more recent years, on information ecosystems in occupied Ukrainian territories. What follows is a first attempt to apply a framework developed in that context to a problem I have watched from the side.
That framework, which I have been developing under the working name of dictionary theory, starts from a simple and far from original premise: that Russia’s information confrontation is not primarily about the content of what is communicated, but about the reference frame through which all content is interpreted. The shared definitions, connotations, and narrative rules that determine what reality means.
For over a decade, Western governments, tech platforms, and think tanks have waged war on Russian disinformation. They built fact-checking units, trained citizens in media literacy, pressured social media companies to remove false content, and monitored Kremlin-linked accounts and mapped troll farms.
None of it has been harmful but none of it has really solved the problem at hand, because we are fighting in the rear and ignoring the real battle. We have been focussed on the false words Russia introduces while Russia is busy rewriting the dictionary.
Graffiti in Kharkiv by Hamlet. ‘Reality reminds me of a film I haven’t watched’
Look it up
When you don’t know what a word means, you look it up in a dictionary. You trust what it tells you. Russia wants to rewrite that dictionary, so that when you look up “invasion” it says “liberation,” when you look up “sovereignty” it says “the right to defend your civilisation against Western destruction,” when you look up “Nazi” it says “anyone who resists Russian power.” This is about lies but it is also about something much larger, more expansive, than lying. Ultimately, you are being given definitions and definitions often feel like facts.
But a dictionary alone is not enough to construct a version of reality. You also need grammar, or the rules for how words combine into meaning. Grammar tells you who acts and who is acted upon. Who can be a victim and who cannot. What causes what. What is natural and what is aberrant. Russian ‘information confrontation’ is thus not just about rewriting individual definitions but also about rewriting the rules for how those definitions fit together into a picture of the world.
This is why our counter-measures fail. Those who work to identify Russian disinformation and label it false are using their own established dictionary and grammar to do so, their own categories of truth and falsehood, their own narrative frameworks for what counts as legitimate and what counts as propaganda. They are checking Russia’s words against their definitions and finding them wrong.
But Russia is not trying to win an argument using our definitions. It is trying to replace the terms. While we are busy demonstrating that specific Russian claims fail our tests, Russia is quietly rewriting the tests. It is like correcting the typos in a nonsense verse: fluent, careful, and entirely beside the point.
Russia understands information as an environment, rather than a product that is true and false. I think the Russian understanding is superior to our own. As such, in the Russian approach, the lies are not the primary element, changing the dictionary is. That is how to alter the shared stock of meanings, connotations, and narrative rules through which we interpret everything we read, see, and hear. Once the dictionary is changed, true information produces false conclusions. The target’s own mind does the work. You cannot fact-check a dictionary.
How We Process Information
We can map this onto what happens when information enters your mind.
First, perception. Something happens. Your grandfather dies in front of you. That is a fact. You register it. Russia does sometimes work at this level, including seeding doubt about whether events occurred at all, or whether what was directly witnessed can be trusted. MH17. Bucha. The maternity hospital. This is real, and it matters. But it is also the most fragile part of the operation and the part our existing defences handle best. Flat denial requires constant maintenance and tends to fail when evidence accumulates. The more important and more durable work happens on what comes after perception.
Second, categorisation. Instantly, automatically, without thinking. Is this a threat or an opportunity? Is this person a friend or an enemy? Is this event justice or aggression? Is your grandad’s death suspicious or expected? A tragedy or a relief? You are not consciously deciding. You are using categories you already have, built up over years of experience, culture, education, and the stories you were told about how the world works.
Third, narrative. Isolated facts are unstable. The mind wants a story, something that explains where this event came from, what it means, and what comes next. You reach for the nearest available story that makes sense of what you have seen.
Fourth, moral judgement. Good or bad. Deserved or outrageous. Worth fighting or worth ignoring. This follows automatically from the category and the narrative.
Your grandfather dies. That is the fact; you have perceived it, registered it, the death is not in dispute. But how he died, why he died, what his death means, what death itself means, all of that depends on the interpretive layers that follow. On culture, values, religion, the stories your community tells about suffering and loss and who is responsible. Russia does not need to tell you your grandfather is still alive. It needs to control what his death means.
Russia is not primarily operating on the facts you receive. It is working on the categories and the narrative rules, the dictionary and the grammar through which all facts are processed. If it can shape those, it does not need to lie to you. It feeds you real (albeit perhaps partial) information, and your own mind, using a corrupted interpretive structure, reaches the conclusion Russia wants or at least a conclusion amenable to Kremlin interests.
This Is Not New.
What I am describing is not a novel discovery, nor is it improvisation. It is doctrine with a lineage stretching back decades.
The Soviet mathematician and psychologist Vladimir Lefebvre developed the foundations of what would become known as reflexive control between 1963 and 1967. Put simply, reflexive control is the art of causing an adversary to make the decision you want using their own rational faculties. You do not override their reasoning. You shape the inputs to their reasoning (so the premises, the categories, the available narratives) so that their own cognition produces your preferred output. The most powerful operation is invisible to its target, because it works through their own mind rather than against it. Lefebvre’s formal mathematical model had its critics. For a start, it assumed rational actors and required detailed knowledge of the opponent. For years the concept was too unorthodox to be named openly in Soviet military publications, with officers writing around it using phrases like “control of the enemy.” It was not formally adopted by the Soviet military until the late 1970s. But the underlying insight survived the critique of the model and outlasted the Soviet Union itself. Contemporary Russian information doctrine is recognisably built on the same foundation: shape the inputs, and the target’s own mind delivers the output you want.
The Soviet active measures tradition shared the same underlying logic, though it developed more through practice than theory. It explicitly distinguished between disinformation (tactical falsehoods for specific purposes) and the deeper strategic operation: shaping the ideological correlation of forces, the conceptual environment in which all information lands. The forgeries and planted stories were almost secondary. The real work was cultivating what felt natural, inevitable, and common-sense across foreign audiences, working on their dictionary from the outside, over time.
As Jānis Bērziņš and other contemporary analysts of Russian military doctrine have documented, information confrontation occupies a central place in Russian thinking about modern warfare; there is a (messy) thread running from Slipchenko’s theorisation of sixth generation and no-contact warfare through to current doctrine. The objective is to shape how reality is perceived and interpreted, and through that, to shape what actions feel possible, legitimate, or necessary.
Nataliya Bugayova and Kateryna Stepanenko at the Institute for the Study of War put it directly: the ultimate target of Russian cognitive warfare is not belief but will. Russia does not need you to believe it. It needs you to doubt yourself, doubt your institutions, doubt whether action is possible or worthwhile. Once will is degraded, physical capability becomes irrelevant. The West can outspend Russia ten to one and still be paralysed.
But reading this through the lens of my research into the Russian domestic information space, I suspect that paralysis is not Russia’s desired end state but perhaps rather the precondition for something more ambitious. You cannot install a new dictionary into a space that already has a functioning one. First you have to discredit the existing dictionary, making it feel contradictory, exhausted, or corrupt. Once the shared vocabulary of liberal democracy feels hollow, the space opens for alternatives. Not necessarily Russia’s own dictionary wholesale, but dictionaries more amenable to Russian interests, more sceptical of Western institutions, more sympathetic to sovereignty as a trump card, more willing to see Western power as the primary aggressor in the world.
Surkov’s domestic operation illustrates what this clearing process looks like before the alternative is offered. His management of Russian political space in the 2000s was not primarily about propaganda; it was about managed ontological confusion. By simultaneously funding and undermining every political tendency, by making irony and cynicism the dominant register of public life, by ensuring that no stable shared reality could form, he guaranteed that the Kremlin remained the only available source of coherence. Paralysis was not the end state so much as it was the clearing operation that made the Kremlin the only reliable dictionary, the only actor with a grammar coherent enough to be adopted wholesale.
Three Similar Wars
In my reading, Russia is running the same approach in multiple environments simultaneously but at very different stages of completion, and with very different degrees of freedom. The method is the same: discredit or exhaust the existing dictionary, create the conditions in which an alternative feels necessary, and promote dictionaries more amenable to Kremlin interests into that space. What differs is how much of that work is already done, and how freely Russia can operate.
At home, the operation is about maintenance. The Putin regime has controlled dictionary formation for decades, through curriculum, commemorative culture, state media, and the architecture of public memory. The dictionary does not need to be installed. It needs to be kept in good repair, reinforced by periodic crisis moments, like wars, Western sanctions, or terrorist attacks, that make it feel validated by events. The operation is remarkably light-touch, or it could be, if the Kremlin were not so very paranoid. You do not need to constantly lie to a population whose interpretive infrastructure reliably produces the conclusions you want.
In Europe, the operation is different. Russia cannot control dictionary formation here. It does not usually try. What it tries to do instead is corrupt the existing shared dictionary just enough that it stops functioning. The key targets are the foundational words of liberal democratic vocabulary: sovereignty, legitimacy, freedom, rules-based order. Not to replace them with Russian alternatives, but to make them feel hollow, contested, and hypocritical. To produce not belief in Russia, but disbelief in the shared framework. The goal is paralysis, again not as an end state but as a clearing operation. A population whose shared vocabulary has been sufficiently degraded cannot organise coherent resistance. It is in that condition that alternative frameworks begin to feel coherent and attractive.
Elsewhere, across the Global South, in post-colonial states with long memories of Western intervention, or among sympathetic communities in the West, the operation is different again. Here Russia offers a genuinely competitive alternative dictionary: sovereignty as the supreme value, multipolarity as liberation, the West (or elites) as exploitative or aggressors. This finds real purchase not because Russia created the vulnerability, but because Western conduct did. Russia is working with raw material it did not manufacture and the populations it is targeting often already possess grammars and meanings that partially align with its own.
Russia Is Not the Only Author
Russia did not invent the fractures it exploits in Western societies. The distrust of institutions, the economic grievances, the cultural dislocations are all real, and they were produced by Western conditions, Western choices, Western failures.
This matters for two reasons. First, it means we cannot attribute every dictionary disruption to Russia. To do so is analytically wrong and politically counterproductive; it renders every social grievance as enemy action and every domestic critic a Kremlin asset.
Second, and more important: it means Western societies have agency here. The conditions that make populations vulnerable to dictionary corruption are substantially within our power to address. If institutional trust is the immune system, then rebuilding it is a defence policy as much as a social one. Janis Bērziņš has argued precisely this: that addressing the social conditions which create openings for Russian information operations is not a soft optional extra but a strategic requirement.
Russia is also not the only actor producing alternative dictionaries. Across Europe and North America, domestic far-right movements have independently developed interpretive frameworks that converge substantially with Russia’s preferred vocabulary: anti-liberal, sovereigntist, civilisationally anxious. Russia did not build these movements but where their dictionary overlaps with its own, amplification is sufficient. No installation is required and the operation can run on convergence as much as coordination.
This means the disinformation paradigm has no answer for a large portion of the actual problem. Where there are no Russian fingerprints, there is no Russian operation to counter. But the dictionary corruption proceeds regardless.
The Acceleration
Everything above describes a slow operation. Dictionary corruption, historically, is resource-intensive. Shaping the interpretive infrastructure of a society takes decades. The Soviet active measures apparatus was enormous, involving thousands of officers and decades of patient cultivation, precisely because this work is hard and slow.
Artificial intelligence is ending that constraint. Content that works on this metaphorical dictionary and grammar, or that introduces and normalises particular connotations, narrative frames, and category structures, can now be produced at industrial scale across languages, platforms, and communities simultaneously.
Perhaps the deeper problem is not the volume so much as what volume does to the shared conditions under which dictionaries are maintained and corrected. Dictionaries are not fixed. They are continuously negotiated by communities, through journalism, conversation, cultural production, public argument. When something happens, we talk about what it means, and through that talking, shared meaning is either confirmed or adjusted, or maybe even neologisms emerge. This is how healthy interpretive communities function.
AI-generated content at scale introduces corrupted vocabulary and degrades the process of interpretation itself. When the information environment is sufficiently polluted, and Russian operations like the Pravda network work on exactly that, or when volume alone overwhelms processing capacity, when synthetic and authentic voices become indistinguishable, when the sheer density of competing interpretations makes negotiating shared meaning feel impossible, communities lose the ability to maintain and correct their dictionaries through normal social processes. Our shared vocabulary fragments. The spell-check stops working not because the dictionary was changed, but because the dictionary can no longer be updated at all. And people will start looking for and co-creating their own dictionaries, with connotations offered by Russian-inflected if not outright Russian tools.
This is the clearing operation Russia has historically needed to conduct slowly and expensively. AI does it automatically, continuously, and on the cheap.
We are entering a period in which the pace of dictionary corruption will outrun any counter-measure operating at the content level. Fact-checking, platform moderation, media literacy programmes were all inadequate before. Against AI-accelerated dictionary pollution they may well become essentially irrelevant. Not wrong, just operating at the wrong level, too slowly, on too small a surface area.
What Might Actually Help
The framework implies its own counter-strategy, even if the details are complex.
You cannot address dictionary corruption by working on content. You have to work on the same level as the attack, namely on the interpretive infrastructure itself.
That means investing in the conditions under which healthy shared dictionaries are maintained. Not just media literacy (the ability to identify false claims) but something harder: the cultivation of what we might call frame awareness, the capacity to notice that you are interpreting rather than simply perceiving. Communities with high frame awareness will not become immune to dictionary corruption, but they may become more resilient to it.
It means taking seriously the domestic conditions (inequality, institutional failure, cultural displacement) that make populations search for alternative dictionaries. These are not soft problems separate from security.
And perhaps most controversially, I think it means developing what I will euphemistically call ‘counter-dictionary operations’: not fact correction, but active work on the shared vocabulary of our own countries and of other countries, including Russia. Understanding which words have been most successfully corrupted. Building interventions that work at the level of connotation and narrative frame, not claim and counter-claim. This is what genuine strategic communication would look like if it were designed for the actual threat rather than the threat we find it easier to measure.
What this looks like in practice is still being developed. In my own (closed) work on Russian information space I have used a concept I describe as the ‘mythscape’, or the distinctive architecture of myths, historical grievances, and narrative frames through which various Russian audiences interpret reality. I have been testing a methodology built on a simple premise: that messages designed to reach Russian audiences must first be fluent in the Russian dictionary and grammar for interpreting the world. Not to endorse that grammar, but to speak from within it. Messages that arrive in foreign grammar are blocked at the door, via a phenomenon psychologists call psychological reactance, the automatic resistance triggered when a message feels alien or threatening to one’s interpretive framework. Messages that are fluent in the existing grammar get through. Once inside, they can do work that externally framed counter-messaging cannot. The dictionary does not have to be replaced all at once. It can be worked on from within.
Finally, we need a frank reckoning with AI. Not primarily as a tool for generating disinformation (that framing is too narrow) but as an accelerant for the structural degradation of shared epistemic conditions, of shared reality itself. The policy question is not just how do we label AI-generated content but how do we preserve the social processes through which communities maintain shared meaning in a high-volume synthetic environment. We do not yet have a serious answer to that question.
The Lesson From Orwell
George Orwell is responsible for me teaching myself Russian at the age of twelve, and so I return to his essays and his fiction in the hope that he will help me work my way out of this mess (studying Russia). Thinking about his writing on the Soviet Union, on language, on propaganda, and above all about Nineteen Eighty-Four, what he is writing about is not primarily lies. It is the destruction of language. The systematic reduction of vocabulary until certain thoughts become literally unthinkable. Newspeak removes the words you would need to believe something different.
Russia’s operation is not Newspeak, certainly not in the West. It arguably already is in Russia but in the West it does not reduce the dictionary, it just corrupts it through the selective cultivation of certain connotations, certain narrative frames, certain categories and the emotions attached to them. The dictionary still looks full, the spell-check still runs. Everything comes back correct.
That is why we have struggled to say precisely what is happening, even as we see something bigger than just disinformation is happening. It is why we cannot find the right words. And why, until we start working on the dictionary rather than the words, we will keep shooting in the dark while more and more of our own side join the enemy.