i autistic 我自闭

There may be different autisms because different parts of our sensing, thinking, feeling and acting are not working, or are not working together. One person may struggle to filter noise but readily recognise faces and feel close to other people. Another may rely on logic because faces, emotions and relationships do not make sense in the same direct way. A third may understand a situation and care deeply, yet struggle to turn that understanding into action.

For me, logical thinking worked, so I relied heavily on it. Sensory filtering, knowing where my body was in three-dimensional space, recognising faces and experiencing emotions were missing or unreliable in different ways. Another autistic person could have a very different combination of things working and not working.

The question is not just “How much help does this person need?” It is “Which part is not working, what can they use instead, and what happens when something changes?” That could explain why the same advice helps one autistic person, does nothing for another, and makes life harder for a third.

This is an explanation I am exploring, not a proven account of autism. I have reasons to take it seriously, but not solid evidence that proves it. The research and scientific questions are at the bottom; the examples come first.

What do these systems do?

By “systems”, I mean the different functions that let us experience and participate in the world. Here are the questions I want to separate, with examples of what a gap might look like.

These examples are possibilities, not a diagnostic checklist. Each heading contains several functions: recognising a person is not the same as reading their expression; feeling an emotion is not the same as naming it. Nor does each row correspond to one separate brain component. Relational Meaning is my term for the felt significance described here.

Two people, different things working

Imagine two autistic adults, Alex and Beth. The following people and situations are invented to explain the idea, not reports of research results or assessments of particular people. Assume that these are their experiences:

Both have difficulty filtering noise. Both can think logically. Yet Beth can readily recognise distress that Alex needs explained, while Alex can begin a familiar task that leaves Beth stuck. Neither is simply the other with “more autism”. The difficulty lies in different places.

These are not the only two possibilities. Someone else might feel emotions strongly but not recognise them, or understand relationships but struggle to speak or act. This comparison covers selected functions, not everything involved in an autism diagnosis.

Put them in the same noisy cafe

A friend looks upset. Alex and Beth both sit silently. From outside, both might seem uninterested. But suppose Alex is struggling with the noise and cannot work out the friend’s expression, while Beth recognises the distress and feels concerned but is too overwhelmed to respond.

If those descriptions are right, here is what I would expect under different conditions:

Quiet can make an existing ability usable. Clear words can supply missing information. Practice can build a skill. These are different changes. Alex’s deliberate care can be genuine; it need not be identical to Beth’s experience to matter. But getting both to say the same sentence would not show that they arrived there through the same route.

The dishes are still in the sink

Now imagine three people who have agreed to wash up, but none has started. All three want the job done. The visible result is the same; the difficulty need not be.

A person could need more than one of these changes. Gloves do not teach the steps; a demonstration does not necessarily help someone start; starting together does not remove an intolerable sensation. The examples show why “more training” is not one solution to every unfinished task. They do not promise that any particular adjustment will work.

Before removing a workaround, find out what it does

An unfamiliar behaviour may be part of the person’s attempt to solve a problem, not only part of the problem we want to solve.

In When Adaptation Looks Like Laziness, Emily Paige Ballou distinguishes the useful function of an adaptation from its drawbacks. A television could ease the transition into homework while making concentration worse.

Imagine someone who begins sorting paperwork more readily with familiar audio playing, but then overlooks details. Removing the audio might reduce mistakes while making starting harder. Keeping it unchanged would leave the errors unaddressed.

An alternative worth trying could preserve the familiar sound during preparation, then reduce it for the part requiring careful checking. That is a proposal to test, not an assurance that this sequence will work.

Ask what changed in initiation, accuracy, effort and completion. An arrangement that improves one of these may still worsen another.

What function must the replacement preserve, and what cost should it reduce? Understanding the attempted solution gives us something to improve, rather than a choice between condemning it and protecting it from all questions.

Remembering a birthday is not the same as feeling why it matters

Three people fail to mark a partner’s birthday. Again, imagine different things happening underneath:

The third person could buy a thoughtful gift through deliberate consideration. That is not worthless or necessarily dishonest. But knowing the date, understanding the custom, feeling its personal meaning and carrying out a response are not the same thing. Equally, a forgotten birthday, dislike of celebrations or lack of attachment to a mug cannot by itself tell us which experience someone has.

What I mean by a system being “offline”

I mean that something the system would contribute is not available to the person’s experience or action. It might never have been readily available, work only sometimes, be overwhelmed, or fail to connect with another function. I do not mean that a brain scan has shown a component switched off.

The examples distinguish something missing, something blocked, and something not joining up. Beth’s response in the cafe is blocked by noise. The first person at the sink is missing knowledge. The third birthday account separates knowing a fact from feeling its significance. These need different questions, even when the behaviour looks similar.

A good result can come through a different route. Someone who does not reliably recognise faces could identify a colleague by their voice or introduction. A greeting then shows successful identification in that encounter, not necessarily improved face recognition. Conversely, failure to greet does not tell us whether the problem was recognition, noise, distraction or something else.

These distinctions do not mean we can inspect another person’s mind from outside. Their account and what happens across situations matter. Where we cannot tell, leave room for “not yet known” rather than assuming either an absent ability or a hidden one.

What it was like for me

I received an Asperger’s diagnosis in August 2001. I no longer identify with autism, including Asperger’s, but that diagnosis is part of my history.

I used logic to compensate: an internal “face-recognition program” to estimate familiarity, and representations to help direct my movements. The Mental Interfaces I Built describes these effortful workarounds.

Television and road noise disrupted concentration; quieter late-night conditions offered relief. Meanwhile, I experienced my body at a distance and the world as flat. Some input was difficult to keep out, while other experience was difficult to access or organise. A quieter daytime setting would have been a better aim than a routine that cost me sleep.

My account of learning to work with my senses describes changes in emotional experience, followed by a bodily and three-dimensional breakthrough around mid 2004, and further developments afterwards. These were changes in experience, not merely learning to hide unchanged difficulties. Access did not arrive with complete familiarity or skill: I still needed practice and help.

I have called my earlier experience “intellectually-disconnected”: I relied on reasoning while other experiences and connections were missing or fragmented. This does not mean superior intelligence or a complete absence of care or feeling.

A strength, a preference or the only dependable route?

A strong reliance on systemising can have different origins. Someone may find patterns and explicit explanations especially accessible or rewarding. Another may become a highly practised systemiser because other ways of understanding a situation are unreliable or unavailable.

These routes can reinforce each other. An initial strength receives more practice because it works; alternatives remain less developed or harder to access. What began as a useful solution may become an unnecessarily exclusive one.

In my case, logic worked while other experiences and connections were missing or fragmented. Reliance on it was therefore not simply a possible source of difficulty. It was also my response to difficulty. Research on compensatory strategies supplies related distinctions, but does not establish the origin of my experience.

Being good at systemising, preferring it and being unable to function without it are different circumstances. Do not weaken the most dependable resource while assuming another will appear. Ask what it makes possible, what it fails to represent and what additional information or capability could make another route usable.

A person missing information may need explanation; someone who understands but cannot act may need a different bridge. Someone with one effective but overused approach may benefit from practice with alternatives. The same outward behaviour need not call for the same intervention.

Knowing “this is my family” and feeling what that means

In Entering the NeuroTypical Parallel Universe, I describe reasoning about other people’s perspectives and caring about doing right, while lacking much of the felt personal significance that particular people, places, promises and relationships later acquired.

I call that dimension Relational Meaning. Earlier islands of connection became a much wider experience on 12 June 2018. The Heart, Eternal Truth and Eternal Flow are my words for the experience and its significance.

My development makes more sense to me as different functions and connections becoming available at different times than as everything improving together. However, another autistic person may already have the feelings and connections I was missing. Their difficulty might be protecting, recognising or expressing those experiences, not discovering them.

The descriptions “sensory-emotionally-overwhelmed” and “intellectually-disconnected” are shorthand for possible experiences, not two established types. Both can include strong reasoning and poor sensory filtering. Someone can recognise parts of both.

Why the same advice will not fit everyone

Imagine doctors who can recognise a fever but do not yet understand its causes. Knowing how hot each patient is would not tell them why each is ill. They could still offer useful care, but it would be a mistake to assume that the same temperature means the same underlying problem.

My suggestion about autism goes beyond noticing that different people need different help: some of these recurring differences may reflect different underlying autisms, involving different systems and connections. The everyday examples explain that possibility; they do not prove it. Difficulty starting the washing-up, for example, is not itself an explanation of what caused someone’s autism.

Different writers may therefore be describing different autisms, not competing to provide the one correct account. My sharing addresses gaps I experienced; another person’s sharing may explain how to protect an already vivid inner world. Readers may recognise my sensory difficulties without sharing my relational disconnection. They need not accept or reject either account as a package.

Do not mistake an experience someone cannot access for one they need only protect. Do not mistake an experience they already have for one they need to acquire. The task is to understand what needs accommodating, what might develop, and what conditions make either useful.

This differs from Many Lenses of Autism: lenses are ways of looking at autism; this explanation asks whether the same label covers different things happening within people.

A diagnosis can identify a useful pattern without supplying a sufficient explanation of the behaviour used to establish it. Reducing a measured difficulty, understanding the process behind it and making daily life more workable are separate achievements. Autism Beyond Behaviour: A Diagnosis Is Not an Explanation develops the fever comparison and my software-engineering analogy: different functions can work separately yet fail to coordinate, and passing a familiar test need not establish transfer or acceptable cost.

Complementary specialisations and a shared need to adapt

I also want to explore an evolutionary possibility: could some variation within Humanity reflect complementary specialisations, with different emphases on understanding dependable regularities and navigating relationships?

This is a hypothesis about variation within Humanity, not two established evolutionary branches or a division into two pure kinds of people. Empathising-systemising theory is relevant prior work; questionnaire differences do not establish separate lineages or validate this systems-and-connections account.

The practical question does not depend on settling the evolutionary one. When a preferred approach becomes insufficient, what enables a person to use another? One person may search for a stricter rule when a relationship changes; another may protect social agreement when the group’s explanation is wrong. Both can need adaptability, but the required movement differs.

My further hypothesis is that people who navigate social life readily receive more frequent, usable practice in interpersonal adjustment. Other people’s independent aims repeatedly change the situation. Someone else may encounter just as much uncertainty without understanding the feedback or learning how to respond differently.

System-oriented work can demand invention and revision too. A person may be flexible when debugging and inflexible with relationships; a socially fluent person may adapt between people but remain rigid about the group’s assumptions. The question is which forms of adaptation people practise and whether that competence transfers.

This adds preferred strategies, alternatives and switching conditions to the multiple-autisms inquiry. Longitudinal work would ask whether reliance on reasoning preceded other difficulties or developed in response to them. Different responses to support would not, by themselves, establish evolutionary origins.

Beyond the Script develops the personal argument. The value of these distinctions is whether they help predict a useful next step beyond what competent existing assessment already supplies.

Start with one real difficulty

Parents and caregivers do not need to decide which “type” a child is before helping. Choose one activity. What can the child do, where does it become difficult, and what changes when you make one part easier?

For example: keep the familiar activity but turn off the television; show one step instead of adding more spoken instructions; or offer a way to point or type rather than requiring speech. Keep the child’s preferences, comfort and response in view. These are manageable adjustments to ordinary activities, not ways to diagnose a hidden brain problem. Individual sensory and communication adjustments also feature in NICE guidance. [9]

Record the difference: “With the television off, my child followed two steps without another reminder.” That is more useful than “the sensory system is fixed”. Observe whether the benefit lasts, what effort it takes, and what remains difficult. Do not provoke distress to test the idea.

Combine the child’s own communication with observations over time and relevant professional assessment. New or worsening difficulties deserve attention to pain, health and circumstances rather than being dismissed as “just autism”. [9]

Development, responsibility and caution remain

Different difficulties require different help, not the abandonment of expectations. Support, practice, environmental changes and firmer boundaries may all be relevant. An explanation of difficulty does not make harmful conduct acceptable.

My warnings about imposters and exploitation remain: repeated inconsistency showing much greater relevant social capability than claimed, especially when linked to self-interest, is a reason to err on the side of caution. Check differences in conditions, effort and support, but do not ignore a concerning pattern. Protective precautions require less evidence than a public accusation.

My strategic focus on adults with low support needs also remains. Other things being equal, fewer or smaller support barriers leave less to overcome in reaching a self-sustaining life. That is a reason to invest in opportunities for breakthroughs, not to lower expectations. It does not guarantee an individual’s outcome or remove the obligation to provide essential care for people with higher needs.

My breakthrough is not everyone’s missing experience. Another person’s limitations are not everyone’s ceiling. The aim is to understand the particular person well enough to make more of Life available to them.

Research notes: evidence and open questions

The scientific proposal. I propose a functional-configuration account: different developmental pathways may affect different combinations of capacities, access to information and integration between functions, producing overlapping autistic presentations. “Integration” here concerns how information contributes to experience and action; it is not a claim about measured neural connectivity. The account must address repetitive activity, routines, flexibility, language and sensory experience, not only social interaction.

Related theoretical work. Happe, Ronald and Plomin (2006) argued against expecting one genetic or cognitive cause to explain all the features grouped under autism. This is prior theoretical work, not validation of my framework. [1]

Phenotypic and genetic heterogeneity. Litman et al. (2025) identified four phenotypic classes in 5,392 autistic children, with differing associated genetic patterns; phenotypic patterns were replicated in another cohort. This supports biologically informative distinctions without establishing four final natural kinds. [2]

Developmental and polygenic differences. Zhang et al. (2025) linked diagnosis timing with developmental trajectories and polygenic profiles. The authors describe gradients rather than discrete early-versus-late categories. Diagnosis age is not an individual test of cause or authenticity. [3]

Functional connectivity. Pagani et al. (2026) identified hypoconnectivity- and hyperconnectivity-dominant patterns across genetic mouse models and human imaging data. The corresponding human subtypes covered 25.1% of 940 autistic participants, not everyone. These signal-based measures do not identify which of my experiential functions is “offline”. [4]

Separating functions within broad labels. Cook et al. (2013) distinguished detecting expression differences from interpreting their emotion, and found that expression recognition was associated with alexithymia rather than autism severity in their sample. Livingston et al. (2019) found relatively good observed social behaviour alongside poor theory-of-mind task performance in some autistic adolescents. These support finer distinctions without making every improvement masking. [5] [6]

First-person evidence. MacLennan et al. documented mixed sensory responsiveness and variation with stress, tiredness and control. Buckle et al. documented difficulties starting, stopping or changing actions despite intention. These support investigating the distinctions illustrated above; neither study establishes my proposed types or a universally effective response. [7] [8]

Together, these findings make biological, developmental and functional differences worth investigating. Their classifications are not interchangeable, and none establishes my systems-and-connections account or measures Relational Meaning. My personal history supplies hypothesis-generating observations, not independent confirmation.

What else could explain the differences?

A spectrum of outward capabilities does not, by itself, explain how those differences arise. A range from short to tall describes height; it does not identify the processes producing each person’s growth. Examining unusually short or tall growth is a way to work back from the outcome towards the processes involved. We need to distinguish the outward result, the mechanism producing it, and the factors that brought that mechanism about. The same caution applies when investigating autistic differences.

Distinct-looking profiles may be combinations of continuous dimensions rather than separate underlying kinds. Different causes may converge on shared processes; circumstances and development may alter a profile over time. Support needs and severity remain useful, but do not by themselves identify the configuration or its cause.

Relational Meaning may overlap with emotional awareness, alexithymia, interoception, attachment, autobiographical meaning or compensation. These are competing or complementary explanations to test, not diagnoses assigned to me or other people. Its value depends on whether it captures a useful distinction beyond existing measures.

Improvement after a practice does not establish that the practice caused it, or how. Different responses can reflect starting skills, health, delivery or environment; different underlying causes may benefit from the same support. Whether an approach helps, why it helps and who benefits are separate questions.

How could we find out whether this is right?

Define and measure the proposed distinctions. Assess experience, comprehension, expression, performance, effort and context separately. Where self-report is difficult, an observer’s interpretation must not silently become evidence of an absent inner capacity.

Compare competing models. Test distinct classes, continuous dimensions and combinations of both, across independent samples. The configuration should add information beyond existing measures and competent assessment, not merely fit its originating accounts.

Predict before observing the outcome. Assess profiles before comparing methods. A treatment-by-profile interaction would show whether the relative benefit of Method A versus Method B varies as predicted. Improvement reaching statistical significance in one group but not another is insufficient to establish that difference.

Count costs and failures. Assess communication, autonomy, distress, fatigue, everyday transfer and continuing support needs. The hypothesis would be weakened by unreliable measurement, failed replication or no added predictive value. Calling every unsuccessful person “another type” afterwards would evade the test.

A useful dimensional profile could still be valuable even if discrete types were not supported. For now, this remains my working hypothesis, not a validated diagnosis, treatment-selection tool or basis for allocating services.

Sources cited

1. Happe, F., Ronald, A. and Plomin, R. (2006). Time to give up on a single explanation for autism. Nature Neuroscience, 9, 1218-1220. DOI: 10.1038/nn1770.
2. Litman, A. et al. (2025). Decomposition of phenotypic heterogeneity in autism reveals underlying genetic programs. Nature Genetics. DOI: 10.1038/s41588-025-02224-z.
3. Zhang, X. et al. (2025). Polygenic and developmental profiles of autism differ by age at diagnosis. Nature. DOI: 10.1038/s41586-025-09542-6.
4. Pagani, M. et al. (2026). Autism subtypes identified using cross-species functional connectivity analyses. Nature Neuroscience. DOI: 10.1038/s41593-026-02287-z.
5. Cook, R., Brewer, R., Shah, P. and Bird, G. (2013). Alexithymia, not autism, predicts poor recognition of emotional facial expressions. Psychological Science, 24, 723-732. DOI: 10.1177/0956797612463582.
6. Livingston, L. A. et al. (2019; online 2018). Good social skills despite poor theory of mind: exploring compensation in autism spectrum disorder. Journal of Child Psychology and Psychiatry, 60, 102-110. DOI: 10.1111/jcpp.12886.
7. MacLennan, K. et al. (2022; online 2021). In Our Own Words: The Complex Sensory Experiences of Autistic Adults. Journal of Autism and Developmental Disorders. DOI: 10.1007/s10803-021-05186-3.
8. Buckle, K. L. et al. (2021). “No Way Out Except From External Intervention”: First-Hand Accounts of Autistic Inertia. Frontiers in Psychology, 12, 631596. DOI: 10.3389/fpsyg.2021.631596.
9. National Institute for Health and Care Excellence (NICE). Autism spectrum disorder in under 19s: support and management. Guideline CG170, recommendations 1.1.9, 1.3.1 and 1.4.1.