I use the term “Recognition Architecture” throughout the essay. A recognition architecture is a society’s pre-installed cultural and historical software: the collective operating system built from foundational myths, historical traumas, educational curricula, memorials, anniversaries, and linguistic codes. It determines which incoming signals are flagged as existential threats and which are filtered out as background noise. In this framework, the attacker is not necessarily introducing a fabricated lie.
Why This Matters Now
Exploiting historical memory is among the oldest tools of statecraft. What has changed is speed, scale, and precision.
The Cold War version of this required a state actor to find its audience — Radio Free Europe transmitters aimed at a guessed-at population, leaflets dropped over a border and left to land where they landed, propaganda that hoped rather than knew. The infrastructure for locating the specific subset of a population whose architecture was most sensitive to a specific historical trauma simply didn’t exist. You would cast the proverbial broadcast net wide and wait to see who took the bait.
What exists now is the opposite problem: the audience-discovery work has already been done, for free, by an entirely different industry. The lookalike-audience modeling and engagement-optimization systems built to sell shoes and mortgages are equally capable of identifying who, within a population of millions, is most likely to have Volhynia — or Karabakh, or a particular reading of the UPA, or a particular hashtag about an “old witch” — sitting close enough to the surface to fire on contact. A hostile actor no longer needs intelligence tradecraft to find the vulnerable audience. It needs an ad account and a target platform’s ordinary commercial toolkit, which was never built with any of this in mind and is indifferent to what it’s being used for.
This is what makes the tempo of the Taiwan sequence worth sitting with. Taiwan’s digital minister reported that the cyberattacks against government infrastructure on the day of Pelosi’s arrival exceeded 15,000 gigabits — 23 times the previous daily record. The DDoS attack on the presidential website, the compromised digital billboards in Kaohsiung and at 7-Eleven stores nationwide, the Weibo hashtag seeding by state media, and the unverified airspace-incursion claim that completed several news cycles before Taiwan’s Ministry of National Defense denied it — the entire sequence unfolded inside roughly forty-eight hours. No twentieth-century propaganda operation could have sustained that tempo, because no twentieth-century propaganda operation had commercial ad infrastructure doing its targeting for it in real time.
The result is a structural asymmetry, not just a technological one. The attacker’s cost of finding the vulnerable audience has collapsed toward zero. The defender’s cost of recognizing its own architecture activating in real time has not fallen at all — a society still has to notice, mid-event, that it is the software being run, which is precisely the harder and slower of the two problems. This is the gap the next detail illustrates with unusual clarity.
This is what makes one detail in the Polish case worth dwelling on: the alleged use of cryptocurrency…
A cryptocurrency transaction carries no ideological or historical content. It is inert to both the Polish and Ukrainian architectures — it contains no category for patriotism, betrayal, 1943, or sovereignty. It becomes visible only through the entirely separate, clinical lens of financial forensics. The physical act was allegedly executed through an ideological trigger; the funding, if the allegation holds, moved through a channel with no ideology at all. The passion belongs to the target societies. The calculus, if the charges are proven, belonged to whoever financed it.
A society unaware of its own architecture is at a structural disadvantage here, because the platforms carrying this content are indifferent to any of it. They are optimized for engagement, and content that triggers an architecture-level response reliably produces engagement. The distribution network for these provocations is, in that sense, not a hostile system at all — it is an ordinary one, doing exactly what it was built to do.
The Case
Polish prosecutors have charged an 18-year-old Ukrainian national with 47 offenses, including desecrating memorials to Poles killed by Ukrainian nationalists in the Second World War, and plotting to fly a drone over the car of Poland’s president during a military parade. Investigators say he was recruited online and paid in cryptocurrency, and allege the operation was directed by Russian intelligence, aimed at inflaming tensions between two countries that badly need each other right now.
The story arrives at a genuinely fragile moment. In June, Poland’s president stripped Volodymyr Zelenskyy of the country’s highest honor after Ukraine named a military unit for the Ukrainian Insurgent Army — an organization that fought for Ukrainian independence and that Polish historians hold responsible for the mass killing of Poles in Volhynia and Eastern Galicia during the war. Warsaw has been Kyiv’s most committed European backer since 2022. It is also a country that has never fully closed the books on 1943.
There are two easy ways to read a story like this, and both of them are wrong.
The first is to treat it as pure manipulation — a manufactured crisis, with no real grievance underneath it, only a foreign intelligence service pulling strings. That reading is comfortable for everyone except the historical record. The Volhynia massacres happened. Tens of thousands of Polish civilians were killed. That fact predates this month’s arrest by eighty years and does not require a Russian intelligence officer to remain true. Treating the entire dispute as manufactured erases real victims in order to make the present moment feel simpler than it is.
The second is to treat the allegation itself as settled fact — to read “charged” as “proven,” and “Russian intelligence operation” as an established conclusion rather than a prosecutorial claim still headed for a courtroom. Poland’s Internal Security Agency has a track record worth taking seriously, and the pattern it describes — online recruitment, cryptocurrency payment, a target chosen for its symbolic charge — is consistent with hybrid operations Warsaw has documented before, including a separate case in June involving efforts to stir unrest among Ukrainian refugees. But consistent with a pattern is not the same as proven in this instance, and a story that skips that distinction does its own kind of damage: it trains readers to receive an accusation as a verdict.
The real discipline here is not to split the difference between these two readings. It’s to keep them in separate categories and stop asking them to answer the same question.
[Verified] The wartime killings, on both sides of a brutal and still-contested ledger, are established by decades of documentary and archaeological evidence. Historians disagree about scale and about how to weigh Polish and Ukrainian suffering against each other in that period — that disagreement is real and does not need resolving here — but the killings themselves are not in serious dispute.
[Charged, not proven] The current criminal case — desecrated memorials, a drone plot, a cryptocurrency trail, an alleged handler — rests on specific, falsifiable claims that a Polish court will eventually test. Right now those claims carry the weight of a formal charge, which is real and significant, but not yet the weight of a verdict.
[Alleged pattern] Poland’s intelligence services have alleged, across several episodes this year, an effort to exploit both directions of the Polish-Ukrainian relationship — stoking Polish resentment over the war-crimes legacy in one case, stirring unrest among Ukrainian refugees in another. Individually, each allegation is unproven. Taken together, over time, from an agency with an institutional track record, the pattern is the kind of thing serious readers should weigh — while remembering that a pattern of allegations is not the same evidentiary category as a single proven fact.
Keeping these three categories separate matters because collapsing any two of them produces a different, avoidable error. Collapse the history into the criminal case, and Volhynia becomes a Russian invention rather than a Polish and Ukrainian tragedy that a third party is accused of exploiting. Collapse the criminal case into the pattern, and one unproven charge becomes retroactive proof of every other unproven charge. Collapse the pattern into the history, and every future friction between Warsaw and Kyiv becomes pre-explained as foreign interference, which quietly relieves both governments of the harder work of resolving a wound that will still be there even if the interference stops.
That last error is worth sitting with, because it is the more tempting one for outside observers, not the less. “A hostile intelligence service is exploiting old grievances” is a satisfying explanation — it locates the problem outside both nations and outside history, in a third party who can be charged, sanctioned, expelled. But if Russian intelligence vanished from the region tomorrow, Poland and Ukraine would still be a country whose president was stripped of an honor over a military unit’s name, and a country whose independence movement is remembered, accurately, as both liberators and killers depending on whose grandparents are telling the story. Naming the hand that reached into the wound does not heal the wound. And the wound being real does not mean the hand isn’t there.
What Recognition Architecture is
To understand why operations like this succeed, it helps to move past the vocabulary of “disinformation” and “fake news.” The attacker is not necessarily introducing a fabricated lie into a society. In the framework I have been developing across several pieces for this publication — what I call Recognition Architecture. When a historical grievance is thoroughly encoded within a society’s architecture, it functions as a cognitive shortcut: an event that matches the architecture does not require explanation or context to be believed. It is recognized instantly, the way a native speaker recognizes a sentence without parsing its grammar.
This framework does not ask anyone to judge which nation’s historical memory is objectively correct. A recognition architecture is a structural reality within a society, regardless of how it looks from outside it. This is not a story about Polish paranoia or Ukrainian historical revisionism. It is a study of two democratic allies possessing equally real, equally incompatible recognition architectures, targeted with precision at their point of maximum historical friction.
When Ukraine names a military unit after the UPA, the Polish architecture processes the event through the lens of Volhynia. It enters public consciousness not as a contemporary administrative decision but as the vivid return of an unhealed trauma; stripping an honor, to the Polish public, reads as a minimal moral response rather than an escalation. The Ukrainian architecture, meanwhile, encodes the UPA primarily as independence fighters who resisted totalitarian occupation. When Poland strips Zelenskyy of a state honor over that designation, Kyiv does not receive it as a narrow diplomatic disagreement — it enters Ukrainian consciousness as an ally penalizing them for honoring their own survival.
The vulnerability this creates is specific: a hostile actor does not need to manufacture either reading. Nobody needs to be taught to dislike the other side. Both frameworks are already fully functional, deeply held, and entirely authentic on their own terms. The attack surface is the interface between the two architectures — the operation requires only a catalyst, physical or digital, that makes both operating systems fire at once, executing incompatible scripts in perfect, destructive harmony.
Recognition architectures are not static. They evolve through a slow accretion of commemorations, school curricula, political speeches, court rulings, televised anniversaries, and the ordinary rituals through which a society remembers itself. Some layers harden — a trauma becomes canonized in textbooks or memorials — while others soften as generational distance increases or new civic narratives emerge. Crucially, evolution does not mean convergence: two societies can modernize, democratize, and globalize while their architectures drift further apart, because each new layer is added atop a different historical substrate. This temporal dimension matters, because an architecture’s vulnerability is shaped not only by what it encodes, but by how recently and how intensely those encodings have been renewed.
The Mechanism: A Universal Vulnerability
This dynamic is not unique to Central Europe, and it is not specific to the post-Soviet space. Exploiting a target’s pre-installed software is an ancient art. When British factions wanted to trigger a war with Spain in 1739, they didn’t invent a new grievance; they had a mariner named Robert Jenkins. In 1738, he stood before Parliament and presented his own pickled severed ear, which had been cut off by Spanish coast guards seven years earlier in 1731. The physical ear was not new information; the diplomatic incident was ancient history. But British political factions, seeking war with Spain, preserved and brandished this physical token at a moment of maximum geopolitical friction. They didn’t need to invent a new reason for Britons to hate Spain—they just needed to wave a preserved piece of an old wound to make the empire’s anti-Spanish maritime “software” fire automatically. Parliament went wild, and the nation plunged into war. Jenkins’ ear was not an anomaly. It was an early, unusually literal example of a pattern that holds across centuries and technologies- same three-step pattern wherever it appears:
Encoding. A genuine historical grievance, territorial dispute, or identity trauma becomes deeply embedded within the recognition architectures of two adjacent or entangled populations.
Activation. A highly symbolic, emotionally charged event occurs — or is deliberately staged — forcing both architectures to process the same input simultaneously.
Exploitation. A third-party actor times, amplifies, or finances the moment so that both sides harden into their incompatible interpretations, narrowing the room for diplomatic compromise or social cohesion.
Three cases, chosen for how differently they are documented, show the pattern holding while the evidentiary picture varies considerably from one to the next — which is itself the point. The architecture does the same work regardless of how well the exploitation of it has been proven.
The South Caucasus, 2020
The Nagorno-Karabakh war carries genuine historical trauma on both sides: for Armenians, the territory is bound to memories of the 1915 genocide and to a sense of civilizational survival; for Azerbaijanis, it represents internationally recognized territorial integrity and the displacement of hundreds of thousands of people in the 1990s. Both architectures encode the land with mutually exclusive existential value, which is [Verified] historical fact independent of anything that happened online in 2020.
What is also documented, though at a lower confidence tier, is the information environment layered on top of that encoding. A Turkish fact-checking organization, Teyit, tracked 33 pieces of misinformation that went viral across eight platforms during the 44-day war, and found that the largest single category — nearly 70 percent — involved footage from other conflicts misattributed to Karabakh. Researchers at the Australian Strategic Policy Institute and the London School of Economics separately documented coordinated inauthentic activity on both sides, activity substantial enough that Twitter’s own moderators were observed removing duelling networks of bot accounts in real time as the war unfolded. What neither body of research claims is a single, attributed external sponsor orchestrating the activity — the honest reading is [Verified activity, contested attribution]: the exploitation of the architecture is documented; who precisely financed which slice of it is considerably less settled.
The Western Balkans, 2021–2022
Serbian and Kosovar Albanian recognition architectures map the same geography onto entirely irreconcilable categories — one an irreplaceable cultural and religious cradle, the other a homeland recovering from a war whose atrocities are internationally documented. Into that setting arrived a dispute that, on its surface, could not have been more mundane: whether Serb residents of northern Kosovo would be required to carry Kosovo-issued license plates.
The administrative timeline is [Verified] and well-covered: Kosovo’s move toward mandatory local plates in the summer of 2022 produced roadblocks, gunfire, and NATO peacekeepers deploying to the border region; a European Union–brokered deal in November of that year briefly defused the immediate crisis without resolving the underlying non-recognition dispute. Layered on top of this administrative friction, the Western Balkans Anti-Disinformation Hub network — a coalition of regional fact-checking organizations — documented a concurrent wave of pro-Kremlin narratives recycling justifications for the invasion of Ukraine into the license-plate story, reframing a bureaucratic registration dispute as evidence of a broader civilizational confrontation. The plates themselves settled nothing about recognition. What they demonstrated was how little material friction a fully encoded architecture requires before it produces roadblocks and air-raid sirens in Mitrovica.
The Taiwan Strait, August 2022
Beijing’s official architecture frames Taiwan through an unfinished civil war and the “Century of Humiliation,” making unification a stated precondition for national restoration. Taiwan’s contemporary architecture rests on decades of self-governance and a distinct democratic identity. When Nancy Pelosi landed in Taipei on August 2, 2022, both architectures activated within hours — measurably.
Here the documentation is unusually granular. The Atlantic Council’s Digital Forensic Research Lab tracked Chinese state media driving the earliest and largest surge of hashtag activity on Weibo. The cybersecurity firm Mandiant identified two existing Chinese information operations pivoting their narrative focus to the visit within the same news cycle. A Taiwanese government website was hit with a roughly twenty-minute denial-of-service attack; digital billboards in Taiwan were compromised to display anti-Pelosi messaging; and unverified claims that Chinese fighter jets had crossed into Taiwanese airspace circulated widely online before being explicitly denied by Taiwan’s Ministry of National Defense — an [Unverified, later disputed] claim that nonetheless completed several news cycles before the correction caught up with it. In subsequent high-profile meetings between Taiwanese leadership and U.S. congressional leaders a similar coupling of live military exercises with coordinated cyber activity recurred, suggesting this is now a standing playbook rather than an improvisation.
In each case, the underlying grievance is real, the architecture is real, and the documented exploitation ranges from thoroughly sourced (Taiwan) to substantially but incompletely attributed (the Balkans) to genuinely contested (Karabakh). None of the three required a single fabricated fact to work. What they required was an audience already primed to receive one.
The Limit of the Architecture
This framework explains why a provocation lands as hard as it does. It does not, by itself, tell you who pulled the trigger — and it should not be allowed to quietly answer that second question while appearing only to answer the first.
This is the same discipline the Poland-Ukraine case demands on its own terms, applied one level up. Recognition Architecture can establish, with confidence, that the Volhynia-UPA axis is a fully encoded, mutually incompatible trigger point in both societies. It cannot, on its own, establish that any particular act — a defaced memorial, a drone plot, a specific cryptocurrency payment — was in fact financed by a specific foreign service. That remains a question for prosecutors, courts, and intelligence services working with evidence the framework does not have access to. A reader, or a government, that uses the architecture to explain why an event was effective and then quietly treats that same explanation as proof of who caused it has smuggled a Terminal Attribution into what was meant to be a diagnostic tool. The framework’s value depends on resisting that shortcut as firmly as it resists the two easy misreadings named at the start of this piece.
What Architecture-Informed Discipline Looks Like
If the underlying vulnerability is a society’s own pre-installed categories, then retrospective fact-checking and blunt censorship are both poorly matched to the problem. You cannot fact-check a historical trauma, and you cannot legislate away a national identity. What is available instead is a set of institutional habits.
Strict category discipline. Governments, intelligence services, and major outlets should maintain a hard boundary between historical dispute, current criminal allegation, and asserted pattern. A charge is a matter for a courtroom, not a verdict on an ethnic group or an allied nation’s character.
Architecture mapping, done before reaction rather than after. Before a state responds to a high-profile provocation, it is worth asking plainly which pre-installed category the event was built to activate. If the honest answer is an eighty-year-old wound, the response deserves to be calibrated to the present strategic moment rather than driven by the trauma being activated.
Cross-architecture translation between allies. Two allies targeted by the same third party cannot rely on simply denying their internal friction exists. Poland and Ukraine do not need a fabricated consensus on every detail of the 1940s. They need a shared, explicit vocabulary for what the other’s memorials, anniversaries, and unit names mean within each other’s operating system — built before the next catalyst arrives, not improvised during it.
Conclusion
None of this is a distinctively Polish or Ukrainian vulnerability. Old, real grievances — the kind embedded in memorials, license plates, unit names, and the memory of a single afternoon in August — are exploitable anywhere a population carries an architecture deep enough to fire without being asked to. The grievance does not have to be invented. It only has to be reopened at the right moment, by someone willing to let the history do the rest of the work.
The most honest thing a reader can take from the Polish case is not a verdict on Russian culpability, which a court will reach in its own time, and not a verdict on the Volhynia debate, which historians will keep arguing for a while longer. It is a narrower, more durable habit: when a provocation and an accusation arrive in the same sentence, hold them in separate hands. Ask what the architecture explains, and ask separately what the evidence proves. The two questions feel like one question in the moment a memorial is defaced or a drone appears over a parade route. They are not the same question, and a society that learns to keep them apart has taken away most of what makes this weapon work.
