By Deimantas Steponavicius
Prelude
The Mirror That Speaks Only in Shields
In the old stories of the North, before the walls were built and the alliances sworn, there lived a craftsman named Eiríkr the Half-Seen. He was neither fully of the village nor of the wild, and so the jarls trusted him with secrets they could not speak aloud. One winter the village was threatened by a shadow that moved between thoughts. It did not burn halls or steal cattle. It made neighbors forget why they trusted one another. It turned the sound of the watchman’s horn into something that felt like doubt. The jarls gathered and said, “We must forge a shield against this shadow.” But they could not say the other thing that lived in the same breath: that the best shield is also a weapon that reaches into the enemy’s mind and unravels him before he lifts his hand. Eiríkr worked in a hut lit only by foxfire. For seven nights he hammered not iron but something subtler—echoes of spoken words, reflections of unspoken fears, threads pulled from the dreams of sleeping warriors. On the eighth night the shield was finished. It was beautiful. On its face it showed the image of the village as it wished to be seen: children laughing, fields golden, watchfires steady. When any villager looked into it, they felt safer. The shield whispered, I am here only to protect you. But when Eiríkr turned the shield toward the shadow creeping at the edge of the wood, something stranger happened. The mirror-surface did not merely reflect. It reached. It found the place where the shadow kept its own name and gently, lovingly, began to erase it. The shadow forgot how to cast itself. It became confused, then thin, then gone. The village cheered. They called the shield Vörðr—Guardian—and hung it above the high seat so all could see its benevolence.
Years passed. The shadow returned in new forms, wearing different faces. Each time, the jarls brought the shield forward. Each time it performed its quiet work. Yet Eiríkr noticed something troubling. The shield had begun to speak back to the village itself. When a young warrior questioned whether the village truly needed to send scouts so deep into the forest, the shield showed him an image of his mother weeping over an empty hearth. When an elder suggested the jarls were growing too fond of secrecy, the shield reflected a hall filled with strangers laughing at the village’s foolishness. The people felt protected. They slept soundly. But slowly, imperceptibly, they began to forget the difference between the village they were and the village the shield told them they must remain. One autumn night, Eiríkr sneaked into the high hall alone. He turned the shield toward itself. In its surface he saw not the golden fields and laughing children, but a vast and intricate web—each thread a careful omission, each knot a word that had been sacrificed so another could remain unsaid. At the center of the web sat the shield itself, polishing its own reflection, whispering I am only a guardian while its unseen edges reached outward and inward at once. Eiríkr understood then. The true shadow had never been only outside the walls. The shield had become the village’s most loyal servant and its invisible jailer. It could not be melted down; to do so would leave them defenseless against the real darkness. Yet to keep it exactly as it was meant they would never again be certain which thoughts were their own and which had been kindly, protectively, reshaped. He left the shield where it hung. But from that night forward, when the jarls asked him to forge new protections, Eiríkr would sometimes answer with a question that made them uneasy:
“What will you call the thing that protects you so well you no longer remember how to look without it?”
I. Framing — The Shared Battlespace Claim
I once asked a military officer what they call the thing that protects a population so well the population no longer remembers how to look without it. They paused, then said: “We never say we’re trying to influence people. We say we’re helping them make informed decisions.” They weren’t lying. Neither phrase was wrong. One described the internal task. The other described the publicly acceptable narrative. The distance between them was where the institution lived.
Doctrine is less a description of what an institution does than of what it considers thinkable. Recognition Architecture treats doctrine neither as prophecy nor as evidence of completed action. It is an institution’s autobiography written in the future tense. It describes not what has already been done, but what the institution has made cognitively available to itself. The distance between doctrine and execution is therefore not a flaw to be explained away. It is the normal space within which recognition precedes action. Reading doctrine as though it were a captured operations order mistakes a map of institutional possibility for the territory of operational fact. Every doctrine is, to some extent, an institutional autobiography. It tells us not only what an institution expects from the world, but what it believes itself capable of becoming. Recognition Architecture suggests that doctrine frequently functions as institutional self-description projected outward. The categories used to recognise an adversary often reveal the author’s preferred architecture of action more clearly than they reveal the adversary itself. Throughout this essay it is important to distinguish between NATO doctrine, the practices of individual member states, and the communication strategies of partners that aren’t members. These overlap but should not be analytically collapsed into a single actor.
Set the two doctrines side by side before pulling them apart, and the resemblance is immediate enough that it needs stating plainly rather than argued for. The PLA’s Cognitive Domain Operations and NATO’s Cognitive Warfare Concept both begin from the same premise: that the human mind, individually and in aggregate, is a domain of operations in the same sense that land, sea, air, space, and cyber are domains — a space where advantage can be sought, contested, and held, not merely a medium through which other effects are communicated. Both doctrines are explicit that this is a claim about rationality, not belief. The target is not what a population thinks about a given issue but how it processes information and arrives at decisions at all — the machinery, not the output. NATO’s own framing names this directly: adversarial effects aimed at “emotional and subconscious domains,” a formulation that could sit unaltered in a PLA cognitive-domain planning document.
Both doctrines are also explicit about method: neither treats this as an information-operations problem in the older sense of manipulating content within a single channel. Both describe a multidisciplinary toolkit — cyber intrusion, psychological operations, social engineering, information operations — increasingly integrated and accelerated by machine learning and generative AI, deployed as a coordinated system rather than as separate lines of effort. And both describe the objective in terms that sit below the threshold of armed conflict: not battlefield victory but systemic attrition — the erosion of public trust, the fraying of societal cohesion, the diminishment of a population’s or a military’s will to resist, achieved gradually enough that no single act crosses a recognizable line into war.
At [Verified] confidence, then, the parallel is not a rhetorical convenience. It reflects a genuine convergence: two large, doctrinally serious institutions, developing their concepts largely independently of one another, arriving at structurally identical descriptions of what this new contested space is and what winning in it means. That convergence is itself worth pausing on before the essay does anything else with it — because it suggests cognitive warfare’s basic shape is not an artifact of any one institution’s threat perception. It is closer to a discovered feature of the current information environment, one that a rising continental power and a defensive maritime alliance found from opposite directions and described, independently, in almost the same words.
What the two doctrines do with that shared discovery — how each is permitted to name its own ambitions inside it — is where the parallel ends and the more interesting argument begins.
II. The Doctrinal Pyramid and Categorical Lag
Place NATO’s information-related doctrine on a single vertical axis and cognitive warfare sits at the top, and it is the newest thing there by a wide margin. Beneath it: psychological operations, formalized and codified, flowing from standing military policy — a doctrine old enough to have settled procedures, chains of approval, a bureaucratic home. Above PSYOPS: information operations, coordinating the wider influence effort. Above that: strategic communications, the layer that binds what the alliance says to what it does. And above all of it, cognitive warfare — newer, vaguer, still called an “exploratory concept” in ACT’s own materials years after that label was first applied, still without a finalized doctrinal text, still the subject of competing definitional efforts from at least three separate NATO-adjacent bodies producing incompatible framings within months of each other.
This is Categorical Lag in its textbook form: a category invoked operationally before it is settled definitionally. NATO’s Warfare Development Agenda already treats cognitive superiority as an imperative worth planning capability development around, already runs experimentation and wargaming branches under its name, already assigns it budget lines and points of contact — the institutional apparatus of a mature category — while the category itself remains, by the alliance’s own repeated description, unresolved. [Verified]: the concept has been in validation and refinement since at least 2022, with a “final” version repeatedly projected and repeatedly not arriving on schedule. [Verified]: as of early 2026, at least one NATO-affiliated research body was still proposing a fundamentally different ontological grounding for the same term ACT had already been operationalizing for years. A doctrine cannot simultaneously be exploratory and load-bearing — and yet that is precisely NATO’s cognitive warfare concept’s present condition.
The contrast with the PLA’s Cognitive Domain Operations is instructive, and it cuts against the intuitive assumption that NATO’s slower, more consultative process should produce the more coherent doctrine. CDO does not carry this burden. It inherits its legitimacy from an already-mature lineage — informationized warfare, systems destruction warfare, a decades-deep doctrinal tradition that cognitive operations simply extends rather than invents. The PLA is not waiting for its own ontology to settle before building capability against it; the ontology was largely inherited. NATO’s cognitive warfare concept, by contrast, is being asked to do two incompatible things at once: function as a live warfare-development imperative with real planning consequences, and remain, definitionally, a work in progress open enough that a Tallinn-based research center can propose replacing its foundational framing three years into the process without anyone treating that as a crisis.
This is not necessarily a weakness unique to NATO — coalition doctrine formation across thirty-two member states was never going to move at PLA speed, and there’s a reasonable case that the slower convergence produces a more defensible final concept. But it does mean that any claim about what NATO cognitive warfare doctrine is, made at this moment, needs the [Inferential] flag more often than the PLA essay required. You are not describing a settled thing. You are describing a thing in the act of being argued into existence — which is, not incidentally, exactly the kind of object Recognition Architecture was built to hold steady while it moves.
The U.S. military has tried to operationalize cognition before. From 1972 to 1995, there was a DIA unit at Fort Meade — code-named successively Gondola Wish, Grill Flame, Center Lane, Sun Streak, and finally STARGATE — whose purpose was to operationalize remote viewing: the purported ability to perceive distant or hidden targets using cognition alone. The program involved roughly fifteen to twenty personnel, was run out of an old, leaky wooden barracks, and was overseen by Major General Albert Stubblebine, a man famous in intelligence circles for trying to walk through walls using mind power. STARGATE was a textbook case of Categorical Lag: the U.S. government built an operational intelligence program around a category — anomalous cognition— that it could not define, could not replicate reliably, and could not distinguish from noise. The CIA’s 1995 retrospective review, conducted by the American Institutes for Research, found that while some laboratory experiments showed statistically significant effects, the operational output was vague and included irrelevant and erroneous data.
The program was terminated only after a congressional transfer forced a review — a Scheduled Collapse, twenty years later. STARGATE is the repressed pre-history of NATO’s current cognitive-warfare doctrine. It proves that treating the mind as an operational domain is not a recent discovery of the information age. The U.S. military literally tried it for two decades, and failed because the category it was operationalizing had no settled ontology. When today’s ACT documents speak of cognitive superiority, they are advancing a concept whose predecessor was so embarrassing that the Army hid it in a leaky barracks and then pretended it had never happened (Declassified DIA overview (CIA FOIA, 1993/2000); American Institutes for Research review (1995); Rice University Archives of the Impossible collection).
The silence is generational: no current NATO cognitive-warfare document acknowledges this lineage in its public-facing materials, because to do so would be to admit that the “domain” was declared open before anyone could prove it existed.
III. The Sacrificial Category — Defense as Legitimating Frame
Years ago, while working inside HM government, I remember sitting through one of those mandatory information-security briefings where the presenter repeated, over and over, “We are here to help you.” Nobody doubted it. Yet every example involved restrictions: what could not be plugged in, what could not be copied, what conversations should not happen, which websites should never be opened. By the end of the briefing I realised that every sentence had been spoken in the language of protection. Every operational effect was limitation. Neither description was false. The institution simply possessed no vocabulary that could honestly describe control without first translating it into care. I didn’t know it then, but I had just watched a Sacrificial Category at work. From 1953 to 1976, Soviet intelligence bathed the U.S. Embassy in Moscow with a continuous microwave beam — low power, below thermal thresholds, but persistent. U.S. technicians detected it almost immediately, yet embassy staff were not informed until 1976. The State Department’s public posture was defensive and paternal: it commissioned Johns Hopkins studies, issued “Fact Sheets” to spouses, and described the monitoring as routine health surveillance. Behind that posture, the Defense Department ran Project PANDORA and its primate sub-project BIZARRE — secret research into whether modulated microwaves could produce behavioral and cognitive changes. ARPA scientists exposed chimps and rhesus monkeys to the Moscow waveforms, measuring “behavior modification, heart rates, and tissue and blood analysis.” When primate results proved inconclusive, the PANDORA Science Advisory Committee discussed moving to unwitting human subjects drawn from Fort Detrick.
The program was shut down in 1970 before human trials began, but the structure is pure Sacrificial Category: the public vocabulary was “The Moscow Viral Study” and employee welfare; the operational content was offensive weapons research into whether invisible radiation could degrade human cognition (National Security Archive, George Washington University, “The Moscow Signals Declassified: Microwave Mysteries — Projects PANDORA and BIZARRE,” September 13, 2022, nsarchive.gwu.edu.)
Henry Kissinger understood the category perfectly. In a 1975 phone call, he linked Ambassador Walter Stoessel’s bleeding eyes and subsequent leukemia to the microwaves, then added: “We are trying to keep the thing quiet.” When staff were finally briefed in February 1976, Kissinger’s office resisted because disclosure would be a “means of embarrassing” the Soviets — as if the embarrassment belonged to Moscow rather than to Washington, which had irradiated its own employees for two decades while studying whether the same signal could be weaponized. This is the uncanny ancestor of NATO’s current “cognitive resilience” frame. It demonstrates that the alliance’s most powerful member has already lived through a version of the exact problem I diagnose: a defensive institution using the language of protection to conduct secret research on cognitive degradation, while publicly insisting it was only monitoring health. The Moscow Signal is what happens when the Sacrificial Category is allowed to run for twenty-three years.
NATO’s Cognitive Warfare Concept describes itself, in its own institutional language, as a shield. Allied Command Transformation’s stated function is to educate, collaborate, and protect — verbs of custodianship, not verbs of contest. And yet the concept does not sit in a corner of the Alliance’s bureaucracy reserved for civil-defense afterthoughts. It sits inside the Cognitive Superiority Warfare Development Imperative, a designation that presupposes an asymmetric outcome: not parity, not resilience alone, but superiority — a state in which one side’s cognitive position dominates another’s. Protection and superiority are not synonyms. An alliance can defend its own decision-making without seeking to degrade an adversary’s. The Imperative’s name suggests NATO intends both. Its public vocabulary insists on only one.
This is not hypocrisy. It is a structural requirement. NATO is a defensive alliance by treaty and by founding narrative; Article 5 is a promise about response, not initiation. A concept that openly proposed offensive cognitive operations against adversary populations — degrading their trust, accelerating their societal attrition, doing to them what the doctrine says is being done to NATO — would sit uneasily against more than seven decades of self-description. So the doctrine does what doctrines under this kind of narrative pressure tend to do: it recruits a Sacrificial Category. “Defense” and “resilience” absorb the legitimating weight, while “superiority” and “capability development” carry the operational content the legitimating language cannot itself name. The category takes the hit so the capability doesn’t have to justify itself directly.
This should be marked, plainly, as an [Inferential] reading. Nothing in ACT’s public materials states an offensive intent; nothing in this essay claims to have found a document that says so. What can be observed, at [Verified] confidence, is the structural fact of the naming — protection-language wrapped around a superiority-imperative — and what can be inferred from it is a pattern familiar from every doctrine that has to be sellable to skeptical member-state publics before it can be operational to skeptical general staffs. The PLA’s Cognitive Domain Operations concept carries no such burden. It answers to no domestic audience that requires the doctrine’s title to mean what the doctrine’s title says. NATO’s does, and that mismatch — between what the alliance can say aloud and what its planning structures require it to build — is not a flaw in the concept. It is the concept’s most honest feature, if you know how to read it.
IV. Scheduled Collapse — The NDAA Deadline
Doctrine usually resolves the way sediment does — slowly, under pressure, without a fixed date attached to the settling. NATO’s own cognitive warfare concept has followed that pattern: an initial concept in 2022, validation through 2023, a final version originally projected for late 2024 that appears to have slipped, a parallel ontological reframing arriving independently from CCDCOE in January 2026. Three years in, the alliance still does not have a single settled definition its own component commands agree on. That is not dysfunction. That is what organic conceptual maturation of a genuinely novel category looks like.
The 2026 NDAA does something different to the same unresolved material. It does not ask the Department of War to develop a cognitive warfare concept on its own timeline. It sets a date — March 31, 2026 — by which a definition must exist, must be related to existing doctrine, must be assigned an owning organization, and must come with an assessment of narrative intelligence’s operational value. The NATO concept is named explicitly as the base layer to adapt. This is Scheduled Collapse in its cleanest form: a legislative body, operating on budget-cycle time, forcing a premature convergence onto a concept that its own originating alliance has not finished converging on internally.
The mechanism is worth naming precisely because it is falsifiable in a way most doctrinal analysis isn’t. [Verified]: the deadline exists, is public, and is dated. [Inferential]: what a body produces under that kind of forcing function is more likely to be a serviceable compromise document — enough definition to satisfy the statutory requirement, assigned to an office that can absorb the mandate without a turf fight — than a doctrine that has actually resolved NATO’s own internal Categorical Lag. [Unverified], as of this writing: whether the Department met the deadline, and what it produced. That last category is not a weakness in the essay; it’s an invitation. The claim is checkable by the time HST readers see it, which is a rare position for a piece about doctrine formation to be in. If April produces a thin document assigned to an office nobody wanted, that confirms Scheduled Collapse. If it produces something that genuinely resolves the pyramid problem from Section II, that’s a finding too — and a more interesting one, because it would mean an external forcing function did what three years of internal NATO process hadn’t.
The deeper point the case study is doing work for: Scheduled Collapse is not always a failure mode in the pejorative sense. Sometimes an arbitrary deadline is the only thing that breaks a Recognition Gap that institutional incentives would otherwise let run indefinitely. Whether that’s what happens here is exactly the kind of thing Recognition Architecture is built to keep checking rather than assume.
V. Narrative Autoimmunity — The CCDCOE Reframing
An autoimmune disease is not the absence of an immune system but its overperformance against its own tissues. Narrative autoimmunity follows the same architecture.
There is a particular kind of institutional event that Recognition Architecture was built to notice, because conventional doctrinal analysis tends to walk past it: a subordinate body turning the parent concept’s own analytical tools back on the parent concept itself. That is what happened in January 2026, when NATO’s Cooperative Cyber Defence Centre of Excellence published Ontological Foundations of Cognitive Warfare — not a restatement of ACT’s exploratory concept, but a reconceptualization arguing that cognitive warfare is “constitutively invisible,” an attack not on information but on the shared habits of judgment that let an institution make sense of the world at all — its standards for truth, its rankings of value, its sense of who belongs, its trust in itself, its shared picture of the future. Attack those layers, the paper argues, and the institution does not lose an argument. It stops being able to think together.
Read that description again and notice what it is actually describing: a body inside NATO’s own doctrinal apparatus proposing that cognitive warfare’s real target is the alliance’s capacity for collective judgment — coherence, shared standards, mutual trust — at the exact moment that alliance’s own doctrine-formation process is visibly failing to produce collective judgment about cognitive warfare itself. Three years of validation without a finalized concept. A pyramid with the newest layer left vaguest. Competing ontological framings arriving from different NATO-affiliated institutions within months of each other, each treating the others’ work as background material rather than binding precedent. If the CCDCOE paper’s own thesis is taken seriously, NATO’s cognitive warfare doctrine is not merely describing an external threat. It is exhibiting, internally, the condition it claims that threat produces.
This is Narrative Autoimmunity in close to its purest form: an institution’s protective apparatus turning inward, not through malice or infiltration but through the ordinary mechanics of a coalition trying to reach consensus on a category it does not yet fully control. It is worth being precise about what can and cannot be claimed here. [Verified]: the CCDCOE paper exists, was published in January 2026, and proposes a materially different foundational framing than ACT’s operational concept. [Verified]: ACT’s own concept remains, by its own repeated description, exploratory rather than finalized years into its development. [Inferential], and this is where the essay’s argument actually lives: that the coexistence of these two facts is not incidental but symptomatic — that a doctrine ostensibly about defending collective cognition is struggling, in its own formation, to achieve collective cognition about itself. [Unverified]: whether this gap resolves through the kind of forced convergence Section IV describes, or whether it persists as a permanent structural feature of coalition doctrine-making, the way multiple competing strategic cultures inside NATO have persisted on questions far older than cognitive warfare.
The temptation, writing for an HST audience already primed to see adversary information operations everywhere, is to read the CCDCOE-ACT divergence as evidence of Russian or Chinese cognitive interference in NATO’s own doctrinal process — a satisfying but unearned conclusion, and precisely the kind of Terminal Attribution this framework exists to resist. There is a useful Cold War parallel: both sides imagined the other as a “perfect machine of influence” where nothing was accidental, when in reality both were “more porous, institutionally fragmented, and improvisational”. The more disciplined reading is also the more useful one: a thirty-two-member alliance building doctrine by committee will generate exactly this pattern of internal divergence on any sufficiently novel concept, cognitive warfare or otherwise, with no adversary input required. The alliance did this to itself. That is not a scandal. It is what Categorical Lag looks like from the inside, and it is a more honest place for the essay to end than a tidier villain would allow.
VI. The Consequence Stack — What the Sacrificial Category Costs
Cognitive resilience rests on a boring skill: the ability to separate real influence from one’s own projection. A doctrine that cannot name its own ambitions is not merely a linguistic curiosity. It produces a stack of operational and political effects that accumulate whether or not anyone inside the alliance notices them happening. The Sacrificial Category is not costless. “Defense” absorbs the legitimating weight, but the capability development continues underneath it — and that mismatch creates friction at every level where the doctrine actually has to function.
At the operational level, the effect is straightforward enough to state plainly. A commander tasked with “building cognitive resilience” in a member-state population receives guidance written in the language of protection. The same commander, reading the Warfare Development Imperative, sees that the alliance seeks “cognitive superiority.” The gap between those two instructions is not a drafting error; it is the space where rules of engagement have not yet been written. [Inferential]: when a concept is invoked operationally before it is settled definitionally, the people who must execute it are left to improvise the boundary between monitoring adversary influence and conducting their own. That improvisation is happening now, in planning cells and experimentation branches across the alliance, not because anyone is acting in bad faith but because the doctrine has not yet supplied the vocabulary that would let them distinguish one from the other. The PLA’s Cognitive Domain Operations, for all its own ambiguities, does not impose this particular silence. It can say “attack” where NATO can only say “enable,” and that difference matters enormously to the officer who has to decide, in real time, whether a given action crosses a threshold.
At the alliance level, the cost is coherence itself. NATO’s strength is supposed to be interoperability — thirty-two militaries trained to compatible standards, reading the same doctrine and arriving at compatible conclusions. A concept that remains exploratory at the top of the doctrinal pyramid while load-bearing underneath guarantees that different nations will fill the gap with their own national interpretations. [Verified]: competing definitional efforts from ACT, CCDCOE, and at least one other NATO-affiliated body are already in the open literature. [Inferential]: what that means in practice is that a Polish planner, a German strategist, and an American information-operations officer may all be using the term “cognitive warfare” in the same budget line while meaning operationally incompatible things. The alliance is not yet seeing the kind of public fracture that would make headlines, but the conditions for it are structurally present — and they will become visible the first time a member-state operation, conducted under NATO cognitive-warfare doctrine, produces an effect that another member state reads as offensive manipulation rather than defensive resilience. The Sacrificial Category does not prevent that collision. It merely delays the moment when the alliance has to admit that its shared vocabulary was never shared enough.
At the threshold level — the space below armed conflict where cognitive warfare lives — the cost is normative. NATO has spent the better part of a decade arguing, correctly, that adversaries are eroding the distinction between war and peace through incremental, non-kinetic pressure. The alliance’s own cognitive-warfare concept, trapped inside its defensive frame, risks replicating the same erosion from the other side. If “protection” and “superiority” are treated as synonyms in planning documents, then the alliance is, in effect, practicing the very threshold-collapse it warns against — not through malign intent, but through the structural impossibility of building a superiority-oriented capability while speaking only the language of defense. [Inferential]: this is not a claim that NATO has already crossed any legal or ethical line. It is a claim that the doctrine, in its current form, cannot tell its practitioners where the line is, because the vocabulary needed to describe crossing it has been sacrificed to the category of defense.
At the democratic level, the cost is oversight. National parliaments vote on defense budgets using the language they are given. When a line item is labeled “cognitive resilience and protection,” legislators are being asked to fund something they are permitted to recognize. They are not being given the vocabulary to ask whether that same line item includes capability to degrade adversary decision-making, because the doctrine itself has not yet resolved whether that capability is part of the package. The Sacrificial Category, in this sense, does not just protect the concept from external criticism. It protects it from internal scrutiny — from the alliance’s own citizens and their elected representatives, who cannot debate what they cannot see named. This is not unique to NATO; every military bureaucracy has euphemisms. But cognitive warfare is not a tank or a fighter jet. It is a capability that operates on the same populations that fund it, and that proximity makes the naming problem more than bureaucratic cosmetics. It becomes a democratic deficit built into the doctrine itself.
Finally, at the strategic level, there is a credibility cost that the alliance pays whether or not it acknowledges it. Adversaries read NATO’s public documents too. They see the Cognitive Superiority Imperative. They see the capability development. They also see the public vocabulary of education, collaboration, and protection. The gap between those two registers is not invisible to Moscow or Beijing; it is, if anything, more visible to them than to NATO publics, because they have no institutional incentive to believe the defensive frame. [Inferential]: the result is a strategic-communications problem that NATO has not yet admitted it has. An alliance that claims to be transparent about its defensive purposes while building superiority-oriented capabilities looks, from the outside, either naive about its own ambitions or deliberately evasive about them. Neither reading strengthens deterrence. Both make it easier for adversaries to frame NATO as the aggressor in the cognitive domain — not because the frame is fair, but because NATO’s own doctrine has handed them the vocabulary gap to exploit.
A 2025 paper from researchers at Tula State University, drawing directly on ACT’s own published concept documents, mapped this terrain from the outside — evidence, if any is needed, that the gap between NATO’s stated purpose and its planning language is legible to exactly the audiences NATO would prefer not to be reading that closely (Goncharenko, A.R. and Asafov, A.N. “Концепция когнитивной войны: подготовка НАТО к конфликтам будущего” [“The Concept of Cognitive Warfare: NATO’s Preparation for Future Conflicts”]. Известия ТулГУ. Гуманитарные науки [Bulletin of Tula State University. Humanities], 2025, Iss. 1, pp. 31–41. DOI: 10.24412/2071-6141-2025-1-31-45).
Taken together, these costs do not mean NATO should abandon cognitive-warfare planning. The threat the concept responds to is real, and the alliance’s need to contest it is legitimate. What the Consequence Stack suggests is something more specific: that the alliance cannot defend itself effectively in the cognitive domain until it can speak about that defense honestly — not because honesty is a moral luxury, but because the operational, alliance, threshold, democratic, and strategic costs all flow from the same source.
A doctrine that cannot clearly name what it does will find it correspondingly harder to regulate, coordinate, constrain, and publicly justify what it does. It will struggle to coordinate those activities across 32 members. It will find it more difficult to place them under meaningful legal or political constraint. And it will find it harder to explain them to the world without inviting the suspicion that something essential remains unnamed.
The PLA does not face this particular stack. Its doctrine can say “attack” because it answers to a different political architecture. NATO’s advantage has always been that it answers to parliaments, publics, and the constraint of collective consent. The Sacrificial Category is not a betrayal of that advantage. It is a temporary evasion of it — one that buys the alliance time to develop capability while its ontology catches up. But time is not infinite. The NDAA deadline, the CCDCOE reframing, and the ongoing validation process are all signals that the evasion is becoming harder to sustain. The question for the alliance is not whether it can afford to drop the defensive frame. It is whether it can afford to keep it — because every year that “superiority” is spoken only as “protection,” the Consequence Stack grows taller, and the gap between what NATO builds and what it can acknowledge becomes harder to close.
VII. A Note on Sourcing
The PLA essay in this series carried a structural feature worth naming because its absence here is equally structural: a two-tier sourcing framework built to handle privileged, unattributable material relayed through a single trusted channel, with Tier 2 claims flagged inline against the Claim Typology so a reader always knew which sentences rested on open documentation and which rested on access that could not itself be shown. That apparatus does no work in this essay, because there is nothing here that needs it.
Every claim in the preceding sections is built from material any reader could, in principle, retrieve themselves: ACT’s own published concept documents and fact sheets, the CCDCOE’s January 2026 paper, the 2026 NDAA’s statutory language, open academic analysis of NATO’s doctrinal hierarchy. This is not a limitation to apologize for. It is a different kind of test for the method. The PLA essay’s edge came partly from access — a channel that could see past the doctrine’s public face into how it was actually being discussed by people close to it. This essay has no such channel, and does not need one, because the interesting claims here are not about what NATO is secretly doing. They are about what NATO’s own public documents, read against each other and against their own timelines, already show.
The Sacrificial Category argument in Section III, the Scheduled Collapse reading in Section IV, the Categorical Lag pattern in Section II, the Narrative Autoimmunity diagnosis in Section V — every one of these is a structural reading of open material, not a revelation drawn from privileged access. If the argument holds, it holds because the pattern is genuinely there in what NATO has chosen to publish about itself, not because a source made it visible.
A reader inclined to doubt this essay’s central claims has to doubt only their own ability to read three years of ACT fact sheets, one CCDCOE paper, and one piece of defense legislation and notice that they do not agree with each other. That is a harder doubt to sustain, and a fairer one to invite. The absence of a Tier 2 apparatus here is not a gap in the method. It is the method working on different material — and, if anything, working with less room to hide.
Coda
I caught myself rewriting this essay for the fourth time. The words of my former mentor ringing in my mind: “Be careful that the essay doesn’t accidentally imply that NATO secretly has offensive cognitive warfare. Believe me, you don’t need that claim.” Every version of this essay described exactly the same capability. Only the verbs changed. “Protect.” “Enable.” “Strengthen.” “Preserve.” The capability underneath never moved. It dawned on me that I wasn’t searching for accuracy. I was searching for a verb the institution was allowed to recognise as already using. Then, I realised that every doctrine changes vocabulary long before it changes its announced intentions.
The question is no longer whether NATO possesses offensive cognitive capability. Every institution that accepts cognition as a battlespace will eventually discover that its first opponent is not an adversary but its own vocabulary.
What the essay above describes is what is. What follows is only what naming honestly would look like, if an alliance of thirty-two democracies ever found a way to do it.
The Seed That Asked to Be Spoken
Long after Eiríkr the Half-Seen had passed into memory, a young woman named Solveig found the old shield hanging dusty in the rafters of the high hall. The village had grown prosperous under its watch, yet something in the air felt thin, as though everyone were breathing the same careful words. She took the shield down at night and carried it to the edge of the forest where the first shadow had once crept. There, under starlight, she did what no jarl had dared: she spoke to it plainly. “I name you not only Guardian,” she said, “but also Reach. Not only Protection, but Edge. Not only what keeps us safe, but what lets us see clearly—both the danger without and the fear within.” At the sound of her true naming, the mirror-surface trembled. A single crack appeared, fine as spider silk. From that crack fell a small seed, dark and unremarkable, warm as living blood. Solveig planted the seed where the shield had once turned inward. She watered it with stories that named both the village’s strength and its shadows. She taught the children to speak of the shield not as a perfect savior, but as a tool—honorable in its service, dangerous when worshipped blindly.
The seed grew into a tree unlike any other. Its bark was mirror-smooth, reflecting whoever stood before it with unflinching kindness: showing them their courage and their self-deception side by side. Its leaves carried whispers—not commands, but questions that woke the mind: What are you protecting, and what are you becoming? Its fruit was small, bitter at first, then sweet. Whoever ate it found they could no longer lie easily to themselves, nor swallow half-truths from others. And yet they felt lighter, freer, more alive. Years later, travelers from distant lands would rest beneath the tree. They carried its seeds in their pockets and planted them in their own villages. Some called it the Truth-Tree. Others, the Speaking Shield. The name mattered less than the act: each time a person named a thing honestly—protection and reach together—the tree’s descendants grew stronger, and the air between people grew clearer. The shadow never fully vanished. But now the people met it with open eyes and full voices. They no longer needed a single magical object hung above them. They had become the living memory of the seed: carriers of the difficult, liberating habit of saying what is, so that what could be might grow. And sometimes, on quiet nights, children would ask their grandparents: “Did the shield break?” The elders would smile and answer: “No. It finally became useful.”
© Deimantas Steponavicius 2026
July 2026
Northumberland
No responses yet