Where autism comes from
Last reviewed November 16, 2026. OwieClear editorial team. Waiting for clinician review. Not medical advice. Talk with your doctor about your own care.
Autism is often discussed as though it arrived recently, which makes the search for a modern cause feel urgent.
The evidence points the other way: the traits are ancient, strongly genetic, present across every population studied, and most likely persistent for reasons that have nothing to do with anything modern.
Related pages: Autism
The bottom line
- Autistic traits are distributed continuously in the general population.
- The genetic variants involved are common and ancient.
- Autism occurs in every population and culture studied.
- The rise in diagnoses is explained mainly by changes in definition and recognition.
- Whether autistic traits were actively maintained by selection or are simply tolerated variation.
- How autistic people were understood in societies before the diagnosis existed.
The story, step by step
- 1. A continuum ()
- 2. Ancient variants ()
- 3. Why it might persist ()
- 4. Not an epidemic ()
- 5. Before the name ()
What the evidence labels mean
Not a line in nature
What happens
The genetics imply something important about what autism is.
What the research shows
- The common variants contributing to autism are present throughout the population, and everyone carries some.
- Autistic traits, measured with questionnaires, are distributed continuously rather than falling into two groups.
- Relatives of autistic people often show milder versions of the same traits, sometimes called the broader autism phenotype.
- A diagnosis marks a point on that continuum where traits are numerous enough and cause enough difficulty to meet criteria.
- This is not unusual in medicine. Blood pressure and cholesterol work the same way, with thresholds drawn across a gradient for practical reasons.
Why it matters
Where the line falls is a practical decision about who needs support, not a discovery about a natural boundary.
What's still uncertain
- Whether the continuum view holds equally for autism with intellectual disability is debated.
Read more
Next: Ancient variants
Carried for a very long time
What happens
If autism-associated variants were simply harmful, natural selection would be expected to have removed them.
What the research shows
- Many of the common variants associated with autism are ancient and are found across human populations.
- Some have been associated with traits that are advantageous in other contexts, including aspects of cognitive performance.
- Rare de novo variants, which arise fresh in each generation, behave differently and are not maintained in the same way.
- The persistence of the common variants suggests they were either useful, or neutral enough not to be selected against.
- This is population genetics rather than a claim about any individual.
Why it matters
That the variants persisted at all is evidence against a simple model in which they only ever cause harm.
What's still uncertain
- Inferring past selection from present genetics is indirect and contested.
Read more
- NIMH: Autism spectrum disorderGovernment
Next: Why it might persist
Hypotheses, honestly labelled
What happens
Several explanations have been proposed for why autistic traits remain common, and none is established.
What the research shows
- Specialization: traits such as sustained focus, pattern detection, tolerance for repetition and attention to detail are useful in a group, and a group containing people with different cognitive styles may do better than a uniform one.
- Systemizing: an inclination to understand rule-governed systems is valuable in many domains, including those that produced technology.
- Honest signalling and reduced social deception: not following social convention has costs and may also have had value.
- All of these are speculative. They are reasonable hypotheses with limited means of testing, and they should not be presented as findings.
- What can be said more firmly is that the variants persisted, which is itself a fact requiring explanation.
Why it matters
These ideas are often repeated as established. They are not, and treating them as such does autistic people no favours either.
What's still uncertain
- Evolutionary explanations for behavioural traits are difficult to test and easy to construct after the fact.
Read more
- NIMH: Autism spectrum disorderGovernment
Next: Not an epidemic
What the rising numbers mean
What happens
The strongest evidence that this is not new comes from how the numbers behave.
What the research shows
- Prevalence estimates vary enormously between places that cannot differ biologically. In recent US surveillance, rates differed roughly fivefold between monitoring sites.
- When adults are screened in the general population rather than relying on existing diagnoses, autistic adults are found at rates similar to those in children, implying they were always there and uncounted.
- Each broadening of the criteria has been followed by an increase in diagnoses, which is what you would expect if definition were the driver.
- Children who would once have been diagnosed with intellectual disability or language disorder are now diagnosed autistic, a shift that has been measured directly.
- Groups historically missed, girls, adults, people of colour, are now being identified, which raises counted prevalence without changing anything underlying.
Why it matters
An epidemic produces consistent increases everywhere. A change in counting produces exactly the pattern observed: wide variation by place and by who is doing the counting.
What's still uncertain
- A small real increase cannot be excluded, and this remains an open question.
Read more
- CDC: Autism spectrum disorderGovernment
- OwieClear: The diagnostic evolutionOwieClear
Next: Before the name
Autistic people in history
What happens
If autism is ancient, autistic people lived long before anyone had the word.
What the research shows
- Historical accounts describe individuals whose behaviour is recognizable, though retrospective diagnosis of historical figures is speculation and should be treated as such.
- Descriptions of children with these characteristics appear in medical and folkloric literature well before 1943.
- Societies with different structures may have accommodated such people more easily, or far less easily, and the romantic version of this is not supported by evidence.
- Many were very likely institutionalized, labelled with intellectual disability or mental illness, or simply lived unremarked in roles that suited them.
- The diagnosis did not create autistic people. It gave a name to something that had no name.
Why it matters
The useful question is not whether autistic people existed before 1943, but what happened to them, and the honest answer is that we largely do not know.
What's still uncertain
- Retrospective diagnosis from historical records is unreliable and often says more about the diagnoser than the subject.
Read more
- NIMH: Autism spectrum disorderGovernment
- OwieClear: When autism was first describedOwieClear