Analysis according to the Autismology framework

Executive summary

This part explicitly changes the frame of reference. It uses the conceptual framework developed by Autistan and Autismology to examine the controversy at a level upstream from diagnosis. Its starting point is the distinction between autism as an autistic nature and ASD as a diagnostic category. A classification may be too broad, imperfect, or poorly applied without the reality it is trying to identify being itself incoherent.

The analysis reconstructs a causal chain: autistic nature → characteristics → interaction with the environment → adaptations, compensations, or harms → observable manifestations → clinical interpretation → possible diagnosis. In this architecture, visible difficulties do not automatically constitute the nature of autism. They may depend on the intensity of certain characteristics, co-occurring conditions, the environment, sociogenic harms, and possibilities for adaptation.

This framework makes it possible to interpret late diagnosis differently. A person may have been autistic throughout development while only being recognized several decades later, especially if the manifestations were less immediately visible or if the person learned the majority social codes. The second-language analogy distinguishes adaptation from assimilation: learning to function according to non-autistic conventions does not mean losing or replacing one’s own nature.

Autismology also proposes looking further upstream for a common core through the hypothesis of the Natural Reference Framework: a relation to coherence, authenticity, precision, fine distinction between pieces of information, and non-arbitrariness. These concepts are presented as working hypotheses rather than established findings in the scientific literature. This version sets out conditions that could make them testable, notably by separating what remains stable across environments from what varies with accessibility and sociogenic harms.

From this perspective, differences in needs, visibility, genetic trajectories, or timing of diagnosis may be real without necessarily implying several distinct autistic natures. False diagnoses should be corrected, co-occurring conditions distinguished, and social phenomena studied separately. Caution is directed mainly at the risk of turning detection categories into ontological boundaries and socially recreating “real” and “false” autistic people.

Key points

• Autistic nature and the diagnosis of ASD are two distinct levels.

• Visible manifestations lie downstream from interactions among characteristics, environment, adaptation, and co-occurring conditions.

• Age at diagnosis first describes a history of detection and recognition.

• Adaptation is not assimilation; a person can learn majority codes without ceasing to be autistic.

• The Natural Reference Framework and other concepts specific to Autismology are hypotheses to be confronted with data and explicit tests.

Status of this part. This part explicitly changes the frame of reference. It takes up and reorganizes concepts developed in the work of Autistan and Autismology. These concepts are not presented as conclusions of the external scientific literature: they form an analytical framework and working hypotheses to be confronted with evidence.

C.1 — Autism and “autism spectrum disorder” are not the same thing

In the conception developed by Autistan, a very clear distinction must be made between:

autism as an autistic nature

and

autism spectrum disorder as a diagnostic category.

This distinction is not the definition currently used by the DSM or ICD. It is a conceptual proposal by Autistan intended precisely to prevent the medical definition of a disorder from being taken as a definition of autism itself.

Autistic nature refers to what fundamentally makes a person autistic: a particular, deep, and relatively stable organization of the human being, present throughout development.

ASD, by contrast, is a human diagnostic construction. It rests on observation of certain behaviors, particularities, difficulties, and manifestations considered sufficiently characteristic to warrant a diagnosis.

This distinction is essential because the two objects do not have the same status.

Autistic nature exists in the person independently of whether a psychiatrist, psychologist, or administrative system recognizes it.

Diagnosis, by contrast, necessarily depends on:

historically variable scientific and medical knowledge, conventional criteria, assessment tools, professional practices, the culture of a period, and sometimes human error.

Diagnostic criteria can therefore evolve without the nature of autistic people changing.

C.2 — An imperfect classification does not necessarily make its object incoherent

This is a central point in the current controversy.

Suppose the ASD diagnosis has now become too broad.

Suppose even that many people are incorrectly diagnosed.

That would constitute a serious problem with the diagnosis of ASD.

But it would not demonstrate that autism itself had become an incoherent reality.

The confusion can be illustrated very simply.

If a measuring instrument is badly calibrated and begins producing many erroneous measurements, those errors do not thereby become new properties of the phenomenon being measured.

A repeated measurement error remains a measurement error.

Similarly, if some contemporary diagnostic practices identify as autistic people who are not autistic, the logical conclusion is not:

there is now a new kind of autism.

The first conclusion should be:

our diagnostic instrument may be producing false positives; we therefore need to improve our ability to distinguish autism from non-autism.

A diagnostic error should not be allowed to retroactively reshape the reality it was supposed to identify.

C.3 — The fundamental problem: looking for autism in its consequences

Medicine and a large part of research have historically begun with what is observable.

They see behaviors, social difficulties, sensory particularities, reactions considered unusual, repetitions, intense interests, communication difficulties, or crises.

The whole is then grouped under the name “autism.”

For Autistan, this reasoning begins too far downstream.

We propose instead distinguishing a chain of levels:

autistic nature → autistic characteristics → interaction with the environment → adaptations and compensations → difficulties or advantages depending on the situation → possible harms and suffering → visible reactions and manifestations → interpretation by relatives and professionals → possible diagnosis of ASD.

A visible manifestation therefore lies at the end of a long chain.

It does not necessarily constitute the definition of what is at the beginning.

This is probably one of the main reasons why autism appears so mysteriously “heterogeneous” when one begins with its manifestations.

C.4 — Autistic nature is not defined by its “disorders”

The very vocabulary of “autism spectrum disorder” encourages another confusion.

It easily suggests that autism consists of a set of disorders and that the more of these disorders a person presents, the more autistic that person is.

We reject this equivalence.

In our conception, difficulties possibly associated with autism must be studied separately from autistic nature itself.

Some difficulties may be intrinsic to certain individual configurations.

Others may arise from co-occurring conditions that are not autism itself.

Still others may arise mainly in the encounter between an autistic person and an environment profoundly unsuited to that person.

And a characteristic considered a “deficit” in one context may constitute a quality in another.

High precision may be disabling when an environment constantly demands rapid social approximations.

That same precision may become valuable in a field requiring the detection of very fine differences.

Low tolerance for inconsistency may become extremely painful in a contradictory administrative system.

It may also make it possible to identify contradictions that others no longer notice.

Resistance to certain social conventions may be interpreted as a failure to adapt.

It may also constitute useful resistance to conformity.

In other words:

the value of a characteristic cannot be determined independently of the frame of reference and environment in which it is observed.

C.5 — The Natural Reference Framework: looking for the common core upstream

The theoretical work developed by Autistan proposes precisely to search for the common core of autism much further upstream than diagnostic criteria.

For this, we use the notion of the Natural Reference Framework.

This hypothesis considers that autistic nature may be characterized in particular by a particularly direct relationship with certain properties of reality: coherence, authenticity, precision, distinction between things that are close but different, fidelity of information, and non-arbitrariness.

This approach leads to a different interpretation of a number of characteristics frequently observed in autistic people.

What is described as “rigidity” may sometimes be difficulty accepting a contradiction arbitrarily.

What is described as “excessive attention to detail” may sometimes be a greater preservation of distinctions actually present in the information.

What is called “social difficulty” may sometimes arise from communication in which one person transmits something fairly directly while the other rapidly reconstructs meaning from categories, implicit assumptions, and conventions.

This hypothesis is not presented here as a scientifically demonstrated theory.

It nevertheless indicates an essential methodological difference:

instead of first asking what disorders autistic people have in common, we look for what may be common in their fundamental organization before that organization interacts with extremely different environments.

C.6 — Sociogenic harms can profoundly transform the observable picture

Another fundamental distinction concerns what Autistan calls sociogenic harms.

An autistic person never exists independently of the environment.

The person continually encounters sensory, human, institutional, and communicational environments.

These environments may impose:

noise, lights, or other sensory demands that are difficult to tolerate; unpredictable changes; contradictions; implicit expectations; arbitrary obligations; repeated misunderstandings; mistaken interpretations; accusations; pressure to mask one’s nature; incoherent administrative procedures; incomprehensible social constraints.

A visible reaction may then be the endpoint of a causal chain:

environmental harm → disturbance → accumulation → suffering → visible autistic reaction.

If the observer begins the analysis only at the last stage, what is seen is:

“problematic autistic behavior.”

If the chain is traced backward, a very different reality may be discovered:

“an understandable reaction by an autistic person to a series of aggressions or environmental incompatibilities.”

This distinction is crucial when comparing different autistic people.

Two people with relatively similar autistic characteristics may develop extremely different observable pictures because their environments, histories, and possibilities for adaptation are not the same.

Heterogeneity of consequences therefore does not necessarily prove fundamental heterogeneity of the nature situated upstream.

C.7 — Why age at diagnosis is a particularly misleading variable

The proposal to divide people according to the age at which they were diagnosed seems particularly problematic to us.

Age at diagnosis is not an intrinsic property of autism.

It is first of all the moment when a human system finally recognizes something.

This distinction may seem obvious, but it changes almost the entire interpretation.

A person diagnosed at age 45 did not become autistic at age 45.

At minimum, we must distinguish:

the existence of autistic nature;

the appearance of certain perceptible manifestations;

the moment when people around the person notice them;

the moment when someone thinks of autism;

and the moment when a diagnosis is finally made.

These moments may be separated by several decades.

C.8 — Early and late diagnoses should logically produce different populations

There is even a very simple reason why populations diagnosed early and late should not be expected to look perfectly alike.

Imagine two autistic children.

In the first, manifestations are extremely visible from the earliest years: very strong reactions, major communication problems, significant regulation difficulties, behaviors immediately noticed by adults.

It is extremely likely that someone will quickly consider a developmental disorder.

In the second, the particularities are much less spectacular. The child speaks, learns, works around some difficulties, gradually builds strategies, and more or less manages to follow what others expect.

That person may go twenty, thirty, or forty years without a diagnosis.

This mechanism almost automatically produces:

particularly visible manifestations or major difficulties → earlier diagnosis

and

less immediately visible manifestations or greater possibility of adaptation → later diagnosis.

The two groups will therefore necessarily have different statistical characteristics.

Because some of their differences are precisely what contributed to their entering groups diagnosed at different ages.

This is a classic population-selection problem.

We are not randomly choosing two groups of autistic people and then discovering that they differ.

We classify them according to the result of a detection process that itself depends on their differences.

C.9 — A very concrete example: being diagnosed at age 48

The Secretary-General of the Autistan Diplomatic Organization himself received an autism diagnosis at around age 48.

This obviously does not mean that a particular form of autism appeared at age 48.

It existed beforehand.

But for several decades, its manifestations and difficulties had not led to a correct identification of their nature.

He had been able to function, travel, work, learn, adapt to many situations, and develop numerous strategies.

That does not make his autism later.

It makes his diagnosis later.

This experience illustrates the danger of directly transforming “age at diagnosis” into “type of autism.”

C.10 — “Mild autism”: an imperfect expression but an understandable intuition

In ordinary language, one might say that some people are “more mildly autistic.”

This expression poses several problems if used as a rigid clinical classification.

But it expresses an intuition that should not simply be discarded: some characteristics constituting autistic nature can probably be present with greater or lesser intensity.

Many human characteristics exist in degrees.

It is therefore not absurd to consider different intensities of autistic nature without necessarily assuming several distinct species of autism.

The difficulty is that the intensity of autistic nature and the intensity of observed difficulties cannot be mechanically equated.

Between the two intervene:

the environment, accessibility, personal history, acquired abilities, support received, co-occurring conditions, sociogenic harms, and many other variables.

A person whose autism is barely visible may experience considerable suffering.

Another person with very strongly autistic characteristics may function remarkably well in a particularly suitable environment.

We must therefore distinguish:

possible intensity of autistic nature

from

level of disability or suffering in a given situation.

C.11 — What the 2025 genetic study actually shows

The main new scientific argument behind Frith’s proposal comes in particular from a study published in Nature in 2025 by Zhang and colleagues.

This study is important and should not be minimized.

It shows that age at diagnosis is associated with different developmental trajectories and identifies two polygenic factors that are only moderately correlated with each other. One is more strongly associated with early diagnoses and lower social and communication abilities during childhood; the other is more strongly associated with later diagnoses and shows stronger genetic correlations with ADHD and several mental-health problems.

The authors conclude that age at diagnosis may partly reflect different developmental and polygenic trajectories rather than depending only on diagnostic or environmental factors. They also indicate that the broad term “autism” may encompass several phenomena with different etiologies.

This is a genuine scientific finding.

But it does not resolve the conceptual question.

C.12 — Genetic differences do not necessarily mean differences of nature

An especially important line of reasoning is needed here.

The existence of different genetic pathways leading to different expressions of a characteristic does not automatically imply the existence of several unrelated fundamental realities.

The same complex property can result from many genetic combinations.

Conversely, the same genetic factors can contribute to several different characteristics.

Human genetics is rarely a system in which:

one gene or one set of genes = one perfectly separated natural category.

The Zhang study therefore shows something important:

people diagnosed early and later are not genetically interchangeable on all the dimensions studied.

But the question that concerns Autistan remains:

Do these genetic differences produce two fundamentally different natures, or do they modulate different dimensions and expressions of a common nature?

The study alone cannot answer this ontological question.

It should also be recalled that the authors report that, in some cohorts, diagnoses relied on community diagnoses or reports rather than standardized clinical assessments, and that a variable delay may exist between the appearance of characteristics and formal diagnosis.

We must therefore avoid turning “genetic correlation with age at diagnosis” into “the genetics of two different autisms.”

C.13 — The risk of confusing etiology, intensity, and visibility

Another possibility must be considered.

Suppose certain genetic combinations favor manifestations that are more visible early in life.

Those people will be more likely to be diagnosed early.

Suppose other combinations produce much subtler manifestations in childhood that become problematic when social and relational demands increase during adolescence.

Those people will be more likely to be diagnosed later.

We would then indeed find two polygenic profiles associated with age at diagnosis.

But this would still not be sufficient to demonstrate that the fundamental phenomenon we call autism is absent from one of the two populations.

Genetic differences could influence, among other things:

intensity, form of expression, when certain difficulties become visible, particular co-occurring conditions, or adaptive capacities.

This is precisely why we believe we first need to know what we are looking for genetically.

Looking for the genetics of people who currently meet ASD criteria is not exactly the same as looking for the biological bases of autistic nature.

C.14 — The case of girls and women makes the problem particularly clear

A large Swedish study published in the BMJ in February 2026 followed nearly 2.8 million people.

It shows that the ratio of male to female diagnoses falls sharply as age at diagnosis increases. By adulthood, cumulative rates are much closer to equality than traditional representations of autism suggested.

Frith uses this difference between early and late diagnoses as another indication favoring separation of the groups.

But the authors of the study reach a very different interpretation.

They describe a substantial catch-up in female diagnoses and conclude above all that we need to understand why girls and women receive their diagnosis later than boys and men.

This difference in interpretation is revealing.

What is observed is:

more women among late diagnoses.

One possible inference is:

autistic women have historically been less readily recognized and are therefore diagnosed later.

Or:

the late-diagnosed group may correspond to another category.

The data alone do not choose between these two readings.

The conceptual framework in which they are interpreted is therefore decisive.

C.15 — Misdiagnoses should be removed from the problem, not turned into a subgroup

Frith also raises a question that should not be avoided: the massive increase in interest in autism may also produce misdiagnosis.

There are now thousands of videos, testimonies, lists of signs, and other content about autism.

Some people discover through them an explanation they have been seeking for decades.

In some cases, that discovery may be extraordinarily relevant.

In others, some people may mistakenly recognize themselves in very general descriptions.

Some may be looking for an identity.

Some professionals may apply criteria too loosely.

Fashion effects or imitation probably also exist.

Frith explicitly mentions self-diagnosis, identity-seeking, social media, masking, and what she calls “social contagion” among the factors that may contribute to rising diagnoses.

Autistan does not consider that these phenomena must be denied in order to defend autistic people.

On the contrary.

Correctly identifying autism also requires being able to say that a person is not autistic when that person is not.

But this leads to a very different conclusion from creating a new subtype.

If a person has been incorrectly diagnosed as autistic:

the person does not represent a new form of autism.

The person represents a diagnostic error.

C.16 — Mixing genuinely late-recognized autistic people with false positives artificially creates heterogeneity

This is probably one of the most concerning confusions.

The statistical group “people diagnosed late” may contain:

people who really are autistic and had never been identified; women or men whose manifestations were less well recognized; people who developed many adaptive strategies; people whose difficulties became much more visible as social demands increased; but also, potentially, people who were diagnosed incorrectly.

If all these situations are mixed into a single category called “late diagnosis,” that category necessarily becomes extremely heterogeneous.

Finding afterwards that this category differs from people diagnosed very young should therefore surprise no one.

And inferring that there may be a second autism risks turning a sample-construction problem into a conclusion about the nature of autism.

C.17 — Neurodiversity: essential contributions, but also contemporary confusions

Autistan has long had a complex relationship with the notion of neurodiversity.

There is a fundamental contribution in this approach: it has helped challenge the idea that every neurological or developmental difference should automatically be thought of as an intrinsic deficit to be corrected.

It has promoted listening to the people concerned.

It has made problems of environment, accessibility, discrimination, and normalization more visible.

All of this matters.

But the contemporary development of the extremely broad category of “neurodivergence” also creates confusion.

It can bring very different phenomena together under the same identity banner.

In some settings, it may encourage an excessively subjective conception of belonging.

It can sometimes transform a developmental reality into an identity that might almost be chosen.

Autistan has no reason to regard every claim made in the name of neurodiversity as scientifically or conceptually correct.

But the existence of these excesses still does not change the nature of autism.

C.18 — Social movements around autism should not define autism

This is another level that Frith’s article sometimes seems to bring too close to diagnostic questions.

Social movements evolve rapidly.

Fashions change.

TikTok did not exist when Kanner wrote his first descriptions.

The term “neurodivergent” did not have the contemporary meaning it has today.

Activist movements appear, evolve, split, or disappear.

Autistic nature, if it corresponds to a real developmental and biological organization, obviously cannot change at the same pace.

We must therefore distinguish:

what autism is

from

what a given society, at a given time, says about autism.

Social phenomena can greatly influence who receives a diagnosis.

They should not determine what the diagnosed object is.

C.19 — Masking is a good example of possible confusion

Frith is also critical of the importance acquired by the notion of masking or camouflage.

She raises a real methodological problem: if every absence of an observable manifestation can be explained by invisible camouflage, a diagnosis may become difficult to falsify.

This objection deserves to be taken seriously.

But it does not justify concluding that masking is simply an invention used to bring people artificially into autism.

Within Autistan’s theoretical framework, masking arises almost naturally from a situation of adaptation.

An autistic person gradually learns something like a second social language.

The person discovers how to speak, look, react, hold the body, answer, or conceal certain reactions in order to avoid negative consequences.

This can make external adaptation more successful.

But it does not necessarily change the person’s nature.

A French person living in Brazil can learn excellent Portuguese and master many local codes without thereby ceasing to be French.

Likewise:

adaptation is not assimilation.

The existence of compensatory strategies must therefore be evaluated rigorously, neither accepted automatically as diagnostic proof nor automatically rejected as fiction.

C.20 — The paradox of the autistic person who adapts “too well”

This line of thought creates an important paradox.

The more capacities a person has for learning non-autistic social rules, the less visible that person’s autism becomes.

The less visible the autism is, the less likely the environment is to consider that the person may be autistic.

And the more likely the person is then to be interpreted according to non-autistic norms that do not entirely fit.

Successful adaptation can therefore delay recognition.

A late diagnosis is then not evidence of autism that appeared late.

On the contrary, it may be the consequence of an exceptional ability to conceal or compensate for certain incompatibilities over a very long period.

C.21 — Different difficulties do not mean different natures

Frith rightly emphasizes the enormous difference between, for example:

a non-speaking autistic child with intellectual disability and a permanent need for assistance;

and

a university-educated, autonomous, working adult diagnosed at age forty.

Their needs are obviously not identical.

But this obvious fact still does not allow us to infer that their fundamental autistic nature must be different.

The first picture may correspond to:

autistic nature + high intensity of certain characteristics + intellectual disability + language difficulties + other co-occurring conditions + a particular environment.

The second may correspond to:

autistic nature + another intensity or configuration + high intellectual abilities + substantial social learning + adaptive strategies + a different environment.

The visible results may be almost opposite.

That does not necessarily prevent the existence of a common core situated much further upstream.

C.22 — Autism must be distinguished from co-occurring conditions

This distinction also seems insufficiently present in many discussions of autism “severity.”

An autistic person may also have:

intellectual disability, epilepsy, sleep disorders, ADHD, anxiety, depression, trauma, motor disorders, language difficulties, or many other particularities.

These elements can radically transform the person’s situation.

But they should not automatically be incorporated into a mysterious substance called “severe autism.”

Otherwise, once again, we risk asking autism to explain everything that happens to an autistic person.

Clinical heterogeneity may thereby be artificially amplified because several independent dimensions are amalgamated.

C.23 — Is there really a “spectrum”?

Autistan also questions the necessity of the word “spectrum” itself.

Historically, this notion helped recognize that autistic manifestations did not fit neatly into a few rigid categories.

That was probably an advance over earlier classifications.

But the term itself can become misleading when we imagine a kind of mysterious object called “the spectrum” that now needs to be cut into pieces.

Why not consider more simply that there is an autistic nature with variations of configuration and intensity, as with so many other human characteristics?

We do not generally speak of a “spectrum of human height,” although some people are very short, some very tall, and there are all intermediate heights.

We do not necessarily speak of a “spectrum of human pigmentation.”

Characteristics vary.

Sometimes that is enough.

C.24 — Continuity does not imply a natural boundary in the middle

The comparison with skin pigmentation illustrates another risk, without of course equating autism with the social and political history of racism.

Some people have very light pigmentation, others very dark pigmentation, and there is an immense range in between.

The social categories “white” and “black” can give the impression of two naturally separated groups.

But human biology does not necessarily contain a universal objective line saying:

white ends here, black begins here.

The boundary is to a large extent constructed by classification.

We should ask whether we risk doing something analogous when we take gradual variation and decide:

autistic group 1 | autistic group 2.

Then perhaps:

1A | 1B | 2A | 2B.

A classification may be useful.

But its practical usefulness does not prove that the boundaries it draws exist in nature.

C.25 — The political risk: recreating “good” and “bad” autistic people

This question is not only theoretical.

New classifications can have social and political consequences.

The history of autism has already known divisions between:

“high-functioning” and “low-functioning” autistic people;

“Asperger” and “classic autistic” people;

autistic people considered able to speak for themselves and autistic people whose speech would be ignored;

people considered disabled enough to receive support and people assumed not to be.

A new division based on age at diagnosis could gradually produce:

the “real” autistic people diagnosed as children

versus

late-diagnosed autistic people treated as more suspect.

Or:

autistic people falling under the medical model

versus

autistic people falling under neurodiversity.

Such polarization would be particularly dangerous.

The British National Autistic Society has in fact responded to Frith by saying that contemporary subtyping proposals remain theoretical, without established clinical application, and that a new division could lead to stigma, misinterpretation of labels, and difficulties accessing support.

C.26 — This does not mean that every existing diagnosis should be protected

Rejecting a premature subdivision does not mean defending every existing diagnosis without criticism.

If some people are not autistic, it must be possible to say so.

If some criteria have become too broad, they should be re-examined.

If some concepts are applied circularly, they should be corrected.

If social media really produce phenomena of mistaken identification, this should be studied.

If some professionals diagnose too easily, practices should be improved.

Autistan has no interest in artificially inflating the autistic population.

On the contrary.

A correct understanding of autistic nature requires distinguishing as precisely as possible:

autism and non-autism.

But precision does not necessarily mean multiplying subtypes.

It begins by defining the object better.

C.27 — A fundamental difference from deficit-based logic

One obstacle to this definition is the way autism was historically discovered.

It was observed mainly when something was “wrong.”

People see a doctor because a child is suffering, is not communicating as expected, seems unable to adapt, or presents concerning behavior.

Science therefore logically accumulated a great deal of knowledge about difficulties associated with autism.

But this method creates a profound bias:

we may then define the person’s nature solely from the reasons the person arrived in the doctor’s office.

It is as if one tried to define the fundamental nature of a plant exclusively from the symptoms it displays when growing in unsuitable soil.

The plant may wither.

Its leaves may become distorted.

Its growth may be disrupted.

These phenomena are real.

But they do not necessarily constitute the definition of the plant.

C.28 — Autism as a nature: neither good nor bad in itself

In our conception, autistic nature should therefore not be described as intrinsically good or bad.

It is first of all a nature.

We consider that it has many potential qualities, and probably important functions.

But these do not automatically come to fruition.

Like any human configuration, they interact with circumstances.

The same characteristic can become:

a strength in one context;

a difficulty in another;

or something neutral in a third.

This reasoning allows us to move beyond two equally simplistic extremes:

“autism is a disease made up of deficits”

and

“autism never creates any difficulty.”

Neither correctly describes the real complexity.

C.29 — The hypothesis of self-protection of Naturalness

At a more theoretical and still speculative level, Autistan has long explored an additional hypothesis.

Autistic nature might constitute a form of preservation or self-protection of human Naturalness.

This formulation does not mean that “nature decided to create autistic people” for a particular purpose.

It should not be understood as simplistic teleology.

The hypothesis is rather that, within human evolution, forms of organization may persist that are less readily absorbed by certain artificial social constructions and more strongly attached to coherence, authenticity, or directly perceived reality.

In a society strongly based on conformity, conventions, rapid interpretations, and symbolic transformation of reality, these characteristics can become very costly.

But they may also have fundamental human value.

This hypothesis remains to be studied and seriously confronted with scientific knowledge.

It nevertheless illustrates why Autistan does not first look for the meaning of autism in its symptoms.

C.30 — The real scientific question should be located further upstream

Frith’s article asks, in substance:

has the spectrum become so broad that it should be divided?

For us, a prior question must be asked:

what makes a person autistic?

Only then:

what variations can this nature display?

what characteristics follow from it?

how do those characteristics interact with different forms of environment?

which difficulties are intrinsic, which are situational, and which arise from co-occurring conditions?

what consequences are produced by sociogenic harms?

what criteria actually make it possible to identify this nature?

and how can people who were diagnosed incorrectly be detected?

As long as these levels remain mixed together, moving diagnostic boundaries risks moving the confusion rather than resolving it.

C.31 — What Autistan nevertheless retains from Uta Frith’s warning

We therefore do not regard Frith’s editorial as uninteresting.

On the contrary, her discomfort with contemporary diagnostic heterogeneity probably reveals a real problem.

The ASD category may have become insufficiently precise.

Late diagnoses need to be better understood.

The effects of cultural changes need to be studied.

False positives need to be identifiable.

The contributions and excesses of some neurodiversity discourses need to be discussable without taboo.

And the genetic findings of Zhang and colleagues require us to take seriously the existence of different developmental and genetic trajectories among people currently diagnosed as autistic.

But none of these observations yet compels the conclusion:

“autism should be divided in two according to age at diagnosis.”

That conclusion seems premature to us.

C.32 — Autistan’s provisional position

At this stage, our position can be summarized as follows.

The increase and broadening of ASD diagnoses are legitimate subjects of study.

Different subpopulations may exist among people currently diagnosed.

Diagnostic errors may exist.

Different genetic and developmental trajectories may exist.

Autistic nature may have very different intensities and configurations.

There may also be co-occurring conditions that radically change a person’s life and needs.

But none of these elements yet demonstrates that autistic nature itself should be divided into several distinct categories.

And the age at which a doctor finally makes a diagnosis seems to us a particularly fragile criterion for establishing such a division.

C.33 — Adaptation without assimilation: the second-language analogy

An important addition to the Autismology framework is the distinction between adaptation and assimilation. Adaptation means learning to function in an environment organized according to other codes without having to give up one’s own way of functioning. Assimilation, by contrast, assumes that the person must durably erase, inhibit, or deny their particularities in order to be considered acceptable.

The analogy used is that of a second language. A person who lives for a long time in another country may learn the local language, understand its codes, and function effectively without losing their mother tongue or having to become internally identical to people born in that country. Similarly, an autistic person can learn non-autistic conventions as a second system of expression and reception while retaining autistic functioning as the primary reference.

This distinction is directly relevant to the debate on late diagnoses. Socially effective adaptation can reduce the external visibility of characteristics without demonstrating their absence. Conversely, the ability to learn majority conventions does not prove that a person belongs to a different nature. It may simply show that the person has developed, sometimes over decades, an adaptive competence.

This also leads to distinguishing adaptation from masking. Not every adaptation is necessarily pathological concealment; it may be voluntary and functional learning. But when the environment demands permanent assimilation, prohibits the person’s own modes of communication or regulation, and sanctions their manifestations, the effort of adaptation may become costly and contribute to sociogenic harms.

C.34 — From conceptual hypothesis to a testable research program

If Autismology is to propose a level of explanation further upstream than diagnostic criteria, it must also formulate hypotheses that can be tested. The Natural Reference Framework, the distinction between autistic nature and environmental consequences, or the idea of adaptation without assimilation cannot serve only as alternative vocabulary: they must make it possible to generate predictions distinguishable from those of other models.

A first avenue is to distinguish what remains relatively stable when the environment changes from what varies greatly with accessibility. If some characteristics belong more to autistic nature while others are primarily sociogenic consequences, improving the environment should reduce some distress, crises, or exhaustion without necessarily making the more fundamental characteristics disappear. This distinction could be studied longitudinally.

A second avenue would be to compare autistic people with very different visible needs while searching for upstream invariants: modes of information discrimination, relation to coherence and arbitrariness, forms of perceptual or conceptual precision, relation to the implicit, and modes of social learning. The challenge would precisely be to avoid defining these invariants through the difficulties they are supposed to explain.

A third avenue concerns age at diagnosis. Instead of treating “early” and “late” as two essences, one could separately model the intensity of certain characteristics, co-occurring conditions, social visibility, environmental accessibility, adaptive strategies, and detection mechanisms. It would then be possible to test how much of the difference attributed to the diagnostic group persists once these dimensions are taken into account.

A fourth avenue concerns false positives. Before using the heterogeneity of late diagnoses as an argument for several autisms, one should, as far as possible, distinguish genuinely autistic people recognized late from people whose diagnosis is erroneous. Otherwise, contamination of the sample by false positives risks artificially producing part of the heterogeneity that is then to be explained.

Finally, if several genetic architectures lead to different autistic expressions, the relevant question is not only whether they differ, but what they may have in common at the functional or developmental level. A plurality of etiologies does not necessarily exclude a phenomenological or functional common core; conversely, an apparent common core does not guarantee a single etiology. Autismology must therefore leave this question open and confront it with data rather than resolve it by definition.

This requirement of testability is important in order to maintain a clear distinction between a fruitful working hypothesis and a doctrine. The Autismology framework gains strength if it can indicate which observations could support it, refine it, limit it, or refute it.

Note on sources. URLs for external sources are placed directly in the body of the text at the point where they document a claim. There is deliberately no system of numbered references to a final list.

C.35 — Conclusion — Do not divide what we have not yet sufficiently defined

The controversy created by Uta Frith has at least one merit: it forces us to ask again a question that the multiplication of criteria, diagnoses, genetic studies, and social movements may have pushed into the background:

what is autism?

For Autistan, the fundamental mistake has long been to look for the answer mainly in disorders.

We propose the reverse movement.

Begin by looking for autistic nature.

Then distinguish its characteristics.

Then their different intensities and configurations.

Then the effects of their encounter with the environment.

Then sociogenic harms, adaptations, compensations, and co-occurring conditions.

Only then the manifestations that become visible.

And finally the human diagnosis that attempts to recognize all of this.

This architecture explains why two autistic people can appear extraordinarily different without having to assume two unrelated natures.

It also explains why a diagnosis at age three and a diagnosis at age forty-eight do not necessarily describe two different autisms: they may simply correspond to two very different histories of visibility, adaptation, and recognition of the same fundamental category of human nature.

If false diagnoses exist today, correct the diagnoses.

If fashion effects exist, study the fashion effects.

If there are excesses in some neurodiversity discourses, criticize those excesses.

If different genetic trajectories exist, study them seriously.

But do not ask these peripheral phenomena to define autism in our place.

Before deciding into how many pieces the “spectrum” should be cut, perhaps it is time to look seriously for what connects the people we call autistic.

That is probably where the real subject lies, much further upstream from disorders and diagnoses.