The Map Is Not the Territory
Why the public image of autism fails almost everyone it's meant to describe.
Photo by A. C. for Unsplash+
Content note: This piece briefly discusses Nazi-era child murder in the context of the Asperger's diagnostic history.
A note on context: I write on Substack primarily to restore my own coherence after a late-in-life autism diagnosis that coincided with severe burnout and a cognitive, nervous system, and coherence collapse. My following is small, and I typically write long-form. Most articles reach 100 to 200 reads over the course of months. My recently published autotheory piece, over 20,000 words, has been opened nearly 700 times in its first two weeks. Real human beings like me have reached out privately to say how well it describes their inner world. They feel seen, heard, legitimized. Real. This is not a small thing. The system has entirely missed or misread us.
Most people hold a fairly consistent image of autism. It has emerged over decades of public narrative—research, media, advocacy—and it centers on a particular presentation: early, visible, and significant in its support needs. That presentation is real. The people it describes are real. The research built around them is real.
But it was never the whole picture. The gap between that image and the actual territory is wide enough that people have been falling into it for decades: misdiagnosed, undiagnosed, or diagnosed and then told they don’t quite count.
Here’s what the image misses: autism is not a single dial running from mild to severe. It’s several independent dimensions—perceptual style, processing architecture, pattern recognition, communication style, executive function profile, interoceptive and somatic awareness, nervous system regulation, and support needs—that combine in different ways for each person.
Change one, and the others don’t automatically follow. A person can be nonspeaking and possess cognitive capacities that no standardized instrument was designed to assess. A person can hold a graduate degree yet still require substantial environmental support to function. A person can have spent forty years successfully masking their autism at enormous internal cost and only now be finding language for what that cost was.
When the image becomes the definition, everyone who doesn’t match it becomes incomprehensible to clinicians and systems, and sometimes even to themselves.
The muddiness runs deeper than the autism category alone. Until DSM-5, ADHD and autism could not be co-diagnosed. DSM-4 explicitly excluded ADHD when autism was present. This was a categorical decision that assumed mutual exclusivity and offered little supporting evidence. DSM-5 removed that exclusion, and what followed was the formal recognition of something clinicians already suspected: the co-occurrence rate is between 50 and 70 percent.
For decades before that, autistic people with ADHD were either diagnosed as only one or the other, or diagnosed with autism and had their ADHD symptoms attributed to it and left untreated. Half the picture, missing for the duration of a human life. Both autism and ADHD independently produce spiky cognitive profiles: high capacity in some domains, significant difficulty in others, with executive function a consistent weak point in both.
When they co-occur, the profile becomes spikier, the gap between peak capacity and sustainable functioning widens, and the cost of sustaining the mask rises. Add twice exceptionality, the co-occurrence of high cognitive capacity in some domains alongside significant difficulty in others, and the profile becomes spikier still. The person may be capable of genuinely extraordinary things in the right conditions and unable to reliably manage ordinary demands in the wrong ones. Both are often true at once. The categories have no framework for that. They were built to sort, and twice exceptionality resists sorting by definition. The categories were kept separate by diagnostic convention. The people were never separate at all.
How the Map Got Drawn
This isn’t a cultural accident. It has a specific institutional cause, and beneath that cause lies a specific economic logic.
Diagnostic categories are not neutral scientific descriptions. They are administrative instruments. They determine who receives resources, who qualifies for accommodation, and who is sortable to insurance, school, workplace, and legal systems. I would argue that categories exist, in large part, because capitalism requires sortable units. A person who cannot be placed in a recognized category cannot be efficiently processed, and systems built for throughput do not handle the unclassifiable gracefully. They either force a fit or deny the need.
This is the context in which the DSM operates. It is written by psychiatrists, yes—but it is written largely for insurers and institutions, and the pressure to standardize, to make presentations legible, to create boundaries clean enough to administer, shapes every edition. The result, I would argue, is a document that serves the system’s need for sortable categories more reliably than it serves the people those categories are meant to describe.
Until 2013, the DSM—the diagnostic manual that governs how mental health conditions are identified and treated in the United States—maintained separate categories for different presentations of autism, including Autistic Disorder, Asperger’s Disorder, and PDD-NOS. These categories were imperfect and inconsistently applied, but they gave people named identity pegs. Asperger’s, in particular, became a community, a self-understanding, and a way of locating oneself that felt specific rather than approximate.
It is worth pausing on that name. Hans Asperger was not simply a pioneering researcher whose work was later applied to a diagnostic category. Historical investigation—most rigorously by Edith Sheffer in her 2018 book Asperger’s Children, corroborated independently by historian Herwig Czech through medical records and referral letters—has documented his active complicity in the Nazi regime’s child-killing program. Asperger referred children directly to Am Spiegelgrund, a killing center where children the Reich deemed unfit for social conformity were murdered, as part of what became known as Aktion T4. These referrals are documented in letters signed by Asperger himself. The category that became a community, that gave thousands of people their first coherent self-understanding, was named after a man whose career was entangled with eugenics. Many people who built their identity around that label learned this history only after the fact, if at all. This, too, is part of what it means to have frameworks built without you. You may not know what they were built on.
DSM-5 collapsed it all into a single category: Autism Spectrum Disorder. The stated rationale was that the boundaries between categories weren't clinically reliable, which wasn't exactly wrong. The eugenics history that made the name itself problematic would only be fully documented years later, after the consolidation had already happened. But the consolidation solved neither problem cleanly. It removed a compromised label without replacing it with anything that served the same orienting function, and it grouped wildly different presentations under one umbrella, handing that umbrella a public image built around the most visible, historically studied presentation. The nuance didn’t disappear from reality. It just lost its diagnostic vocabulary, such as it was.
The one exception was Rett Syndrome, which had been included under DSM-4's pervasive developmental disorders alongside the other autism categories and was removed entirely from the autism section in DSM-5 when its underlying genetic cause, a mutation in the MECP2 gene, was identified. It was removed from the category based on biological evidence, without the people it described having any say in the matter. It is a preview of what may happen again as research advances: categories shifting, lines moving, and people left to renegotiate their understanding of themselves in the wake of decisions made elsewhere.
This is the primary reason I advocate for self-understanding beyond diagnostic categories.
The DSM is written by psychiatrists, largely for insurers and clinicians. Autistic community organizations advocated from outside the process during DSM-5's development and achieved partial influence, but there was no formal autistic representation within the workgroup itself, and the lived experience of the people being diagnosed has never been structurally included in shaping the document that determines whether that experience is recognized.
This matters because we are now living with the consequences, which are unfolding at a particular historical moment. One in which larger systems are failing faster than they can be repaired, and the people whose nervous systems are most attuned to incoherence are registering that failure first.
The Fights We Are Having
Outside the autistic community, a conflict is unfolding over who gets to define the territory.
On one side, researchers and clinicians who built their understanding of autism from a specific population, using specific methods, and are now defending that understanding against evidence that doesn’t fit. In March 2026, Dame Uta Frith, one of the field’s most influential figures, gave an interview to TES magazine, arguing that the autism spectrum has “widened to the point of collapse.” She characterized the later-diagnosed group, mainly adolescents and women, as people who “might feel highly anxious in social situations” and are “perhaps characterised mainly by a sort of hypersensitivity.” On masking, she was equally dismissive: “We are all masking, all the time, trying to adapt to our society’s norms,” she said, with no particular problem except exhaustion, and that exhaustion “could arise from lots of other causes.”
Notably, her characterization of the later-diagnosed group does not appear to be grounded in research on late-identified autistic adults. Her framework rests on externally observable traits, early identification, and childhood presentation. The epidemiological concern driving her 2026 comments appears to be rising diagnosis rates and strain on educational systems, rather than clinical evidence about the population she characterizes. I have seen no direct statement from her citing research on late-identified autistic adults as the basis for these claims. She is applying a model built without this population to a population it was never designed to describe—and, predictably, finding that they don’t fit.
On one point, she is unambiguously right: “The current situation is dire, and something must be done.” The disagreement is not whether the system is failing. It is about who it is failing, why, and what doing something should actually mean.
These are not fringe positions. They come from a pioneering researcher with decades of influence on how autism is understood and resourced. And they are worth examining precisely because they sound reasonable until you look at what they’re actually doing.
It is worth noting that Frith’s own theoretical framework—weak central coherence, the tendency toward detail-focused rather than gestalt processing—is itself a description of a cognitive style that appears across the broader autism presentation her research helped legitimize, including among many of the late-identified adults she is now questioning. The framework and the population arrived together.
It is also worth noting, and the critiques of her position often miss this, that Frith’s language is carefully hedged. She uses might and perhaps. She is not making definitive claims about this population. She is expressing uncertainty. That intellectual caution deserves acknowledgment. But it also raises a question her argument doesn’t answer: if you are uncertain, why does that uncertainty point toward narrowing the diagnostic doorway rather than investing in better research on the population you’re uncertain about? Uncertainty should produce more investigation, not less inclusion. The hedged language and the gatekeeping conclusion don’t actually fit together.
And the specific uncertainties she's expressing are being actively resolved by new research that runs counter to her conclusions. Studies increasingly point to ecological mismatch—chronic sensory overload, the sustained demand of masking, and the hypervigilance required to navigate environments not designed for autistic nervous systems—as the primary driver of anxiety in autistic people, rather than anxiety as a co-occurring condition in its own right. The treatment implication follows directly: accommodation and environmental redesign address the cause; standard anxiety interventions often target the symptom while leaving the conditions intact. I've written about what this looks like from the inside in Mapping the Territory of Coherence.
Calling it anxiety and stopping there mistakes the symptom for the origin, and it’s just another attempt at categorization that doesn’t help anyone. The claim that masking is something everyone does entirely misses the structural difference: neurotypical people mask for various social reasons, while autistic people mask because the alternative is being unintelligible to the world they have to live in. The cost is categorically different. As the writer behind Adult With Autism observed in a recent piece, the implication beneath these arguments is troubling: if the old model didn’t see you, perhaps you aren’t really there.
The attributes don’t disappear when the label is removed. The sensory overload, social fatigue, pattern recognition, burnout, and lifelong sense that something operates differently remain exactly where they were. Removing the diagnosis doesn’t explain them. It just leaves them unexplained.
On the other side, people are building alternative frameworks because the existing ones don’t fit. Clinicians, researchers, and community members are documenting what late-identified autism actually looks like—the masking, the compensation, and the collapse that arrives years after the performance—and arguing that diagnostic tools need to catch up to the population, not the other way around.
This is unfolding in real time. In April 2026, a doctoral researcher known as The Cognitive Ecologist, whose work focuses on late-identified autistic adults and transformative social change, published a model she calls DSM-6 Concept Draft: Autism Recognition After Adolescence. The model proposes what the DSM has never formally encoded: that autism—as it’s currently defined in the DSM-5—can be traced across a lifespan, even when it was hidden, misread, compensated for, or metabolized through survival. That masking is evidence, not absence. That the internal cost of functioning is data, not drama.
There is something else worth naming, though the research hasn’t caught up to it yet. Among the late-identified autistic adults I’ve encountered, particularly those clustering around the gestalt processing profile, I’ve noticed consistent descriptions and demonstrations of speech patterns and processing delays that suggest a shared underlying architecture. Language that arrives in whole-meaning chunks rather than in analytically assembled pieces. A processing style that takes in the full gestalt before producing output.
The formal theory of gestalt language processing remains contested and underspecified in the literature. But the phenomenon itself—the community noticing it in each other before the researchers do—forms part of the argument this piece makes. The lived pattern precedes the category. It always does. And when the researchers do arrive, the first thing they will do is try to make it measurable. Which means they will operationalize it, standardize it, and sort it. Something will be lost in the translation. It always is.
This is the community doing the diagnostic work that it was excluded from. Building the map that the institutions didn’t build. Doing so, notably, while living inside the gap the map failed to account for.
Within the autistic community and among families of autistic people, the conflict looks different but stems from the same source. People who found stability, identity, and hard-won self-understanding within a particular framework for understanding autism—and parents who built their advocacy, their community, and their children's care around that same framework—experience its expansion as a threat to the coherence they built around it. This isn't bad faith. Coherence is not a luxury.
When your nervous system has spent decades finding its footing in a world that didn't accommodate it, or when you've spent years fighting to have your child seen and supported within a system that resists both, the framework that finally names that experience becomes load-bearing. Someone else's different-but-also-valid experience can feel like it undermines the ground you're standing on.
The result is a familiar social media landscape: late-identified autistic adults being told they don’t deserve their diagnosis because they “made it this far.” The implicit logic is that survival is evidence against need. If the support wasn’t there and you managed anyway, the support was never necessary.
This argument has it exactly backward. The person who made it this far often did so at enormous cost—burnout, collapse, and self-accommodation that consumed decades of bandwidth. The absence of a diagnosis didn’t protect them from any of that. It only meant they bore it alone, without a framework, without accommodation, and without access to the support that might have changed the trajectory. Surviving without support is not evidence that support wasn’t needed. It is evidence of what happens when support isn’t available.
I want to be clear about something. I am not interested in fighting any of these fights. Not with researchers defending old models, not with autistic people who have found their footing in a framework that doesn't include me, and not with clinicians who were trained on incomplete information and are doing their best with it. That stance doesn't minimize what any of us are carrying, or what's at stake in these arguments. It's a recognition that the conflict itself is a symptom, not the disease. Continuing to fight inside it means the institutions that produced it never have to answer for what they built.
What interests me is what becomes possible if we stop. The people on every side of these arguments have something real at stake—accurate recognition, adequate support, a coherent identity, and resources that actually reach those who need them. None of those things is in competition. A late-identified autistic adult receiving a correct diagnosis doesn’t take anything from a high support needs autistic child who needs residential support. A better diagnostic framework that recognizes masked presentations doesn’t erase the people the old framework was built around. A workplace accommodation policy that accommodates sensory and communication differences benefits more people than it costs.
The barrier is not disagreement. It is architecture. These groups rarely occupy the same systems, waiting rooms, diagnostic pathways, or research conversations. The silo is not incidental; it is how institutions manage populations they were never designed to serve as a whole. Divide by category, allocate resources by category, and the people within those categories never have to be seen in relation to one another.
The coalition that doesn't yet exist—uniting people with high support needs and their families, late-identified adults, autistic researchers and clinicians, and community members who build alternative frameworks in their spare time—has real leverage. Divided, each group asks institutions to see them. Together, they provide the evidence that institutions and criteria need to be rebuilt. Infighting nearly guarantees that it won't happen. It keeps the people with the most at stake from building the power to change what's failing them all.
The Resource Argument
There is a related claim worth addressing directly: that late-identified autistic adults are diverting resources from autistic people with higher support needs.
This claim is largely unfounded. Late-identified autistic adults, in practice, do not compete for the same resource pool as people with high support needs. They do not compete for residential placements, intensive interventions, supported employment programs, or the Medicaid waivers that fund most of that infrastructure. The accommodations this population actually needs—workplace flexibility, sensory-informed environments, competent mental and medical health care—come from almost entirely different systems.
What there is is a legitimacy economy. Autism has gained cultural and political recognition, and some advocates fear that late-identified adults dilute that recognition, making the category seem less serious. This fear is understandable, but the conclusion is wrong. Diagnostic legitimacy is not a finite resource.
Public confusion about what autism means is real and has proven exploitable. When the category appears to stretch in ways the public can't follow, it creates openings for bad-faith actors to weaponize that confusion, question the validity of the diagnosis itself, undermine funding for research, and cast doubt on the reality of neurodevelopmental difference altogether.
The legitimacy anxiety within the community is partly a rational response to witnessing that happen. The answer, however, is not to narrow the category to protect it from misuse. It is to build an accurate public understanding of why the category is genuinely complex, and to stop ceding that complexity to people with no stake in getting it right.
And the supports that do exist are less available than people believe, including for the populations they were designed to serve. Support infrastructure was built to serve the categories that the system could see. That means it was built around the presentations that were legible to the instruments built to serve the system. The circularity is not accidental. It is how institutions sustain themselves, by defining need in terms of what they already know how to provide.
For late-identified autistic adults, particularly those who are high-masking, the gap is not merely inadequate support. It is categorically misaligned support. Frameworks designed for childhood presentations, delivered by practitioners untrained to recognize adult-onset identification, and offered in environments that do not accommodate the very nervous systems they purport to serve. The category doesn’t fit. The tools don’t reach. And the system, having sorted you into the wrong bin or no bin at all, has no remaining mechanism to correct itself.
One further category error compounds all of this. Autism is not a mental health condition. It is a neurodevelopmental difference, a difference in how the brain and nervous system are structured and process information from birth. Mental health conditions are generally acquired, episodic, or primarily responsive to psychological intervention.
Autism is none of those things. Its dimensions are neurological, sensory, physiological, and ecological. Categorizing it alongside mental health conditions is an administrative convenience—it has to live somewhere in the DSM, and the DSM is organized around psychiatric categories—but the category actively misdirects care. It suggests that the intervention target is mood or behavior rather than the environment or accommodations. It sends autistic people into mental health systems that aren’t built for them. It psychologizes what is physiological and individualizes what is systemic.
What would actually help: training for occupational therapists, speech pathologists, mental health providers, and physicians on late-identified presentations, grounded in the premise that autism is a neurodevelopmental and physiological reality rather than a psychiatric one. Workplace accommodation frameworks that take the sensory, communication, and executive-function needs of autistic employees seriously. Protected flexibility in environment, schedule, and communication modality. Not more of the same, delivered faster. Something differently designed.
There is a broader implication worth naming. Late-identified autistic adults face unemployment and underemployment at significantly higher rates than the general population, and many reach the conclusion—often after repeated burnout—that standard workplace environments will only continue to harm them. This is not a personal failing or a limitation of the individual. It is an environmental design problem.
Workplaces built around neurotypical processing styles, communication norms, and sensory conditions are genuinely hostile to autistic nervous systems, especially those already eroded by decades of unaccommodated demands. The diagnostic label matters because it provides a framework for understanding that hostility and, theoretically, a legal basis for requiring accommodation. But the deeper point is this: if environments had been designed with neurological variation in mind from the beginning, many of the conditions that make diagnosis feel urgent—the burnout, the collapse, the desperate need for a framework that explains what keeps going wrong—might never have reached crisis. The diagnosis names the mismatch. A better design would have prevented it.
This is not abstract. I was diagnosed at 51. My daughter was diagnosed at 4. Between us, we represent opposite ends of the presentation spectrum the system claims to serve, and neither of us has been adequately supported within it.
For me, the system’s failure predates diagnosis by decades. My autistic nervous system was engaging with medical settings from pre-adolescence onward. ER visits, symptoms that didn’t cohere into a recognizable story, and care that treated each presentation in isolation because no one had the framework to see them as part of the same story. The cumulative cost of that unrecognized mismatch is physiological.
Chronic HPA axis activation doesn’t resolve on its own. It produces chronic physical pain. It requires constant, careful management, and it is, in part, what decades of an unaccommodated nervous system looks like from the inside.
It also cost me my ability to work. After my collapse, I was unemployable for eight to nine months while my nervous system recovered enough to function. I am only now rebuilding, tentatively and on my own terms, because standard workplace environments are not ones I can safely return to. That is not a personal limitation. It is an accurate read of what those environments often do to a nervous system like mine.
I wrote about what post-diagnosis disclosure looks like in practice, and how variable the clinical response remains even when you name it directly in What Happens After You Say, “I’m Autistic”. Disclosure doesn’t close the gap. It just makes the gap visible.
The Measurement Problem
Underneath all of this is a question that rarely gets asked directly: what are we actually measuring, who built the instruments, and what did they need them to do?
Standardized assessment tools serve standardization. They were designed to produce scores that could be compared, filed, and acted upon at scale. A number is administratively useful in ways that a nuanced cognitive portrait is not. It can determine classroom placement, support allocation, and insurance reimbursement. The sorting function requires the score. The score requires the test. The test requires that cognition be legible in a particular, narrow way.
Cognitive assessment tools were designed around a particular definition of intelligence, one that privileges certain kinds of analytic processing, neatly reduces to standardized task formats, and was normed against populations that did not include the full range of human neurological variation. The result is not a measure of intelligence. It is a measure of performance on those specific tasks, in that specific format, on that specific day. It is a measure of how well a mind can make itself legible to the instrument, which is itself a kind of cognitive task, not a neutral one.
I have been labeled intellectually gifted by these instruments at both ages five and fifty-one. I don’t experience that label as inaccurate exactly, but I also don’t experience it as complete. The way my mind actually works—the pattern recognition, the systems thinking, the integrative leaps that happen faster than language, the things I know before I can explain how I know them—none of that is what the test was measuring. The label is a partial read of a much larger signal.
I have also watched someone I love be labeled severely intellectually disabled by a system that had no tools to reach her. The label said nothing about her mind. It said everything about the measurement's limits.
The instrument fails in both directions, and the systems built downstream of those instruments—the diagnoses, the support allocations, the accommodation decisions—inherit every one of those failures. When the map is wrong at the foundation, everything built on it is misaligned.
This is not a peripheral issue. It is the issue. The public image of autism, the diagnostic categories, the support frameworks, and the community conflicts—all of it rests on measurement tools that were never adequate to the territory they purported to describe.
There is another problem that the instruments cannot address. Interoceptive access—the ability to perceive one’s internal states, to know what the body is registering before it becomes a crisis—is not always a stable trait among autistic people. It is load-dependent. The more the environment demands, the less available it becomes. This means the clinical assessment, conducted in a quiet room under low demand, may capture the best case. The disability lives in the worst case. A system built on static, sortable categories has no mechanism for measuring something that degrades precisely when it matters most.
What This Moment Is
It would be a mistake to look at all this conflict and conclude that autism is simply a contested category undergoing normal scientific revision. Something larger is at play.
Systems that were already straining are now failing visibly and quickly. The institutions that were supposed to provide diagnosis, support, and accommodation were built slowly, around narrow assumptions, by people who were not the ones they were meant to serve. They were also built to be efficient, which, in practice, meant being standardized and designed for the person who could be sorted most cleanly. The edges—the presentations that didn’t fit, the needs that didn’t map to existing categories, the people who were too complex to process quickly—were not prioritized. They were deferred. And deferred problems compound.
What is happening now is compounding. The scale of unmet need that better recognition has revealed is not new. It is an old need, finally named.
The people whose nervous systems are most attuned to incoherence—who process patterns, sense rupture, and feel the wrongness of misaligned systems before they can articulate it—are registering this failure first and most acutely. Some of what looks like autistic burnout right now is that. Not just autism in an unaccommodating world, but an accurate perception of a world that is genuinely becoming harder to navigate, hitting hardest those with the least margin and the highest sensitivity to incoherence.
The researchers defending rigid structures are also often operating within institutions under pressure, defaulting to frameworks they know because the alternative is open-ended uncertainty in a moment that already has too much of it. That doesn’t make the defense right. It makes it legible. A systemic failure, not just individual bad faith.
What Better Looks Like
The Cognitive Ecologist’s model does something quietly radical: it shifts the diagnostic question from what this person looked like in childhood to whether we can trace an autistic neurotype across this person’s development, including where it was hidden. Masking becomes evidence rather than a disqualifier. Internal cost becomes data. Burnout becomes a diagnostic signal rather than a separate complaint.
It also includes an exclusion safeguard, a formal prohibition against ruling out autism because a person made eye contact, has relationships, succeeded professionally, was not disruptive in childhood, or can describe their emotions. In other words, a prohibition against using evidence of successful adaptation as evidence against a diagnosis.
This is the logical correction the current system needs. It was built by someone outside the institution, shared through the only available channel here on Substack, while also living the experience the framework describes.
That is worth sitting with. The people who were excluded from shaping the systems are now, at their own expense and on top of everything else they carry, building the replacements.
Categories are administrative tools. They were built to sort populations, not to describe people. Every human being is as unique as a fingerprint, and that is true within the container of autism as anywhere else. We forget this when categories start doing the work that only attention to actual individuals can do. The map stops being a tool and becomes the territory. And then we spend our energy defending the map instead of seeing what’s actually there.
The image most people have of autism is not wrong. It’s incomplete in ways that have consequences, especially for the people it leaves out, for the community it pits against itself, and for the systems built around it that are now failing people at both ends of a distribution they were never designed to see clearly.
Maybe you came into this piece thinking you had a reasonable grasp of what autism is. A spectrum. A range. A set of observable traits identifiable in childhood.
Maybe the territory is larger and stranger than that. Maybe the map has always been the issue.
This piece references work by The Cognitive Ecologist, a doctoral researcher specializing in late-identified autistic adults and transformative social change, whose Substack article DSM-6 Concept Draft: Autism Recognition After Adolescence proposes a lifespan model for autism recognition, and Adult With Autism, whose essay “The Curious Case of Being Told You Don’t Actually Exist” addresses the lived consequences of diagnostic gatekeeping.
The problem is not that people are over-identifying.
The problem is that the map is so badly drawn that cognition, distress, adaptation, trauma, personality, culture, communication and access needs all get collapsed into pathology.
The benchmark is too narrow — and then it calls itself normal.