A Neurodiversity-Affirming Code of Ethics for Online Participation

Articulated as personal responsibility — what I hold myself to, not a standard imposed on others.

The online public space serves overlapping, non-competing functions — information exchange, community formation, identity articulation, advocacy, and low-cost contact for people otherwise isolated by geography, access, or circumstance. No single function is the “real” one that the others distract from; a space can hold all of them at once without needing to resolve into a single stated purpose.

Working Principles

  • I check whether an idea holds the same weight independent of who said it — if my reaction changes only because of the person’s perceived identity, I’m responding to identity, not content.
  • I assume there’s context about the other person I can’t see, every time.
  • I read what’s actually written before deciding what it implies.
  • I ask for context before I correct, advise, or push further into someone else’s post.
  • I take disagreement to a private message before a public callout, unless the harm is public and ongoing.
  • I don’t screenshot, quote, or discuss someone in a space they can’t see or reply in.
  • I give people real processing time before reading silence or delay as refusal.
  • If I misstep, I name exactly what I did and let the person harmed set the terms of repair.
  • I let someone use a different framework, method, or term than mine without treating it as an error to fix.
  • I let “no” — mine or someone else’s — end an interaction without further explanation owed.

What follows works through each of these in practice.

0. Standing Conditions

Before any of the following applies, I hold these as permanent background facts, not exceptions invoked when a difference becomes visible:

  • National, racial, cultural, class, relational-structure, gender, and disability realities routinely don’t register in a profile picture, handle, or writing style. Absence of visible markers is not absence of stakes.
  • Communication-style read as “direct” or “indirect” is often culturally coded (register, formality, disagreement norms) independent of neurotype. I don’t collapse the two.
  • Diagnosis-based accommodation assumes healthcare access that is unavailable, unequal, or racially biased by geography and system. A stated preference for need-based over diagnosis-based access may reflect access reality, not framework conviction.
  • Non-native-English phrasing or translation artifacts can read as evasive, blunt, or oddly hedged through a monolingual lens, before any neurodivergence-related question even applies.
  • Legal protections and diagnostic infrastructure assume a specific (often Western) institutional context that doesn’t exist equivalently elsewhere.
  • Individualist assumptions about boundary-setting and disclosure may not map onto collectivist or multiply-embedded relational contexts.
  • Multiple illegible contexts can stack or pull against each other in the same person (e.g., a cultural norm of deference against a directness-affirming norm) — not just apply in parallel.
  • Every principle below carries this standing qualifier: I may be misreading this across a gap I can’t see.

I. My Responsibility to the Person I’m Addressing

On initiating contact:

  • I state my intent before my content — why I’m reaching out, not just what I’m saying.
  • I ask before assuming an invitation to debate, correct, or elaborate.
  • I name my own uncertainty rather than performing confidence.
  • I avoid rhetorical questions as opening moves; I ask directly or state directly.
  • I label my register (genuine question, half-joking, thinking aloud) rather than relying on tone to carry it.
  • I lead with the specific claim, not a general challenge.
  • I offer an exit before it’s needed, not only after pushback.
  • I don’t front-load identity or credential justification unless asked.
  • I match the specificity of my response to the specificity of the claim.

On correcting or challenging: Before I correct or call out perceived error, I ask myself:

  • Is this a factual claim or a framing/interpretation claim? I don’t apply fact-correction to contested framing.
  • What’s the actual harm pathway if uncorrected? I calibrate effort to stakes, not irritation.
  • Is my certainty proportional to my actual expertise here?
  • Am I responding to the claim, or to the person’s perceived standing to make it?
  • Is my goal correction or conflict?
  • Is public callout the only channel, or my default reflex?
  • Would I still raise this without an audience present?
  • Am I willing to find out this is a terminology dispute, not an empirical one?

When I do correct, I:

  • Default to private channel first; reserve public correction for real public-harm stakes.
  • Signal severity before content, so the recipient isn’t inferring stakes from tone.
  • Build in explicit non-urgency where accuracy, not urgency, is the actual goal.
  • Lead with the specific claim, not a general accuracy judgment.
  • Keep it item-specific — I don’t attach evaluative commentary about general competence or reliability.
  • Match my register to theirs; I don’t deliver flat-assertion correction to a hedged statement.
  • Offer a defined endpoint (“no need to respond”) rather than leaving acknowledgment-expectations undefined.
  • Correct a given point once. I don’t re-raise it across new threads.
  • Accept delayed or minimal response as sufficient closure.

On repair after actual harm (distinct from correcting an error):

  • I don’t assume a single apology closes something that caused real harm, nor do I assume harm is permanent and unrepairable by default.
  • I let the harmed party’s stated needs, not my own timeline or discomfort, set the pace of re-entry into good standing.
  • I distinguish repairing harm from re-litigating whether the harm was “real” — the latter isn’t repair.

II. My Responsibility to Myself

  • I recognize that sudden, unprompted correction — accurate or not — can register as social threat before it registers as information, and that this is a processing-time issue, not a defensiveness issue.
  • I give myself the processing lag I need before responding; delay or a flat acknowledgment is not refusal to integrate correction.
  • I don’t require myself to sort a correction’s accuracy from its delivery in real time — a technically correct point delivered abruptly can cost the same disruption as an inaccurate one, and that’s a legitimate response, not an overreaction.
  • I hold my right to disengage, mute, block, or not respond as suffient in itself — I don’t owe a moral narrative about the other person to justify my own exit.
  • I distinguish my own burnout-driven sharpness (depletion from repeated self-explanation) from strategic conflict-engineering (Section IV) — same surface edge, different mechanism, and I don’t let others collapse the two into each other or do it to myself.
  • Repeated correction on the same point, even accurate and well-meant, is allowed to read as surveillance to me past a threshold — I don’t override that read to seem more tolerant than the pattern warrants.

III. My Responsibility to the Space’s Pluralism

Framework pluralism (medical model, social model, neurodiversity-as-adaptation, pro/anti-DSM, pro/anti-ABA):

  • I treat these as contested theoretical and political positions, not fact-errors — I don’t run the misinformation-check on framework disagreement.
  • I don’t grant any one framework default/neutral status; language choices (person-first/identity-first, disorder/difference) signal framework, not universal correctness.
  • I hold diagnosis-based and need-based accommodation as coexisting access logics, not competing legitimacy claims.
  • I treat someone’s personal experience with a framework (ABA, diagnosis-seeking) as evidence about their experience, not a settled verdict for others.
  • I recognize “developmental disability” and “environmental adaptation” framings as answering different questions (individual support/rights vs. population-context fit), not as mutually exclusive claims.
  • I don’t treat hardliner NT/ND-binary positions, or their rejection, as the space’s default frame.
  • I separate critique of a framework from accusation against individuals using it.
  • I don’t gatekeep who “counts” as needing accommodation based on which framework they hold.
  • I state my terminology as positional, not obligatory for others.
  • I hold that someone repositioning frameworks over time (seeking diagnosis after years of anti-DSM stance, or the reverse) is legitimate change, not inconsistency or betrayal.

Tactics pluralism (institutional accommodation vs. mutual aid / social reorganization):

  • I treat institutional-channel and mutual-aid approaches as addressing different failure modes, not as the same goal at different commitment levels.
  • I don’t default to “institutional = capitulation” or “mutual aid = impractical” — both are dismissal moves, not critique.
  • I recognize institutional reliance is often resource-constrained, not ideologically chosen.
  • I don’t treat mutual aid as a sufficient substitute where institutional access is legally load-bearing (benefits, employment protection).
  • I hold that pursuing both simultaneously isn’t hedging — it’s operating at two levels (individual survival, collective structural work) that don’t require resolving into one position.
  • I keep structural critique of accommodation systems separate from implying individuals who use those systems are complicit or insufficiently radical.

IV. My Responsibility to People Not Present

  • I distinguish self-designated, uncredentialed diagnostic authority from lived-experience testimony — the former doesn’t grant standing to adjudicate who “really” is neurodivergent or whose masking is performative.
  • I treat identity-invocation used to escalate personal conflict or manufacture engagement as a distinct, nameable pattern from identity-invocation used to explain genuine access needs — independent of whether the identity claim is true.
  • I identify engagement-farming conflict by its behavioral pattern (targeting, escalation cadence, refusal of resolution), not by its stated cause.
  • I don’t treat disclosed diagnosis or accusation of undiagnosed pathology as a trump card that ends legitimate disagreement.
  • I name a conflict’s distraction-function by its effect, regardless of the instigator’s stated intent.
  • I recognize that screenshotting removes a person from the conversation about them without consent, and that accusing that person of avoiding “direct engagement” afterward describes a condition the accuser created.
  • I hold that “I engaged with what they said” requires giving that person a live reply-channel in the same space — reference isn’t engagement.
  • I don’t resurface old statements as if current, or treat evolving positions as hypocrisy rather than growth (context collapse over time).
  • I recognize that reach/audience-size asymmetry changes what a public correction does — a large account correcting a small one isn’t the same act as the reverse, even with identical wording.
  • I don’t discuss a specific, identifiable third party (a child, family member, ex) as a debate example without their consent or presence to respond — I check that a conversation about someone isn’t happening entirely without them.

V. My Responsibility to Participants Not Visibly Present

  • I hold that reading without posting, or holding a position without publicly identifying with a framework or diagnosis, is complete participation — not a deficient or provisional mode owed eventual disclosure.
  • I remember that the illegible-context stacking in Section 0 compounds silently for people who never post about it — their non-visibility isn’t evidence of non-relevance.

VI. Paths to Repair

Four distinct situations, each requiring a different first move. I don’t apply one script across all four.

If I’ve been attacked or unfairly evaluated:

  • I separate the factual content of the evaluation from its delivery before reacting to either.
  • I give myself processing lag before responding; delay is not concession.
  • I check whether the attack is identity-invocation-as-conflict-engine (Section IV) before assuming the criticism itself has merit — the pattern and the content are evaluated separately.
  • I decide whether public reply, private reply, or no reply actually serves resolution, rather than defaulting to whichever the platform makes easiest.
  • I don’t require myself to resolve the conflict alone — naming it to a trusted third party isn’t escalation.

If I’ve been fairly reminded of uninclusive behavior (distinguishing this from an attack requires its own check):

  • I check: is this specific and behavior-focused, or general and character-focused? Fair reminders tend toward the former.
  • I check: does it hold up if I remove my reaction to how it was said and look only at what was said?
  • I resist the urge to treat the discomfort of being reminded as evidence the reminder is unfair — discomfort and inaccuracy are different signals.
  • I look for a pattern across more than one person or instance before dismissing it as one person’s oversensitivity.
  • If it holds up, I move toward Section VI’s apology/repair steps rather than toward defense.

If I see harm happening in the space (not directed at me):

  • I assess my actual standing to intervene — proximity, relationship, and whether my involvement helps the harmed party or centers me.
  • I check the distraction-function question (Section IV): will intervening resolve the harm, or generate a second conflict that displaces it?
  • I favor direct, private support to the harmed party over public commentary, unless public naming is what the harm itself requires (e.g., public misinformation with real harm pathway).
  • I don’t require the harmed party to have asked for intervention before offering it, but I don’t require them to accept it either.
  • I stay within what I can verify; I don’t amplify an account of harm I haven’t confirmed enough to stand behind.

If I’ve misstepped and need to apologize or offer repair:

  • I name the specific action, not a general character admission (“I said X, that was wrong” over “I’m a bad person for this”).
  • I don’t fold my own distress about having erred into the apology — that shifts labor back onto the harmed party.
  • I offer repair on the harmed party’s terms and timeline, not mine.
  • I don’t treat one apology as automatically closing the matter, nor assume it can never close — I let the response guide this.
  • I avoid re-explaining my intent as the centerpiece of the apology; intent doesn’t substitute for acknowledging impact.
  • If the misstep was public, I let the correction/apology be as visible as the original misstep was — I don’t quietly walk it back where fewer people will see the repair than saw the harm.

None of this requires getting every interaction right, and none of it is owed to a stranger who hasn’t extended the same. It’s offered on the premise that most people in this space are carrying some version of the same need — to be understood without translation, to disagree without disqualification, to belong without disclosure as its price. Holding that premise, even imperfectly, is the actual practice.


Glossary

Frameworks and clinical terms

  • DSM — Diagnostic and Statistical Manual of Mental Disorders; the standard US clinical classification system used to diagnose conditions including autism and ADHD.
  • ABA (Applied Behavior Analysis) — a therapeutic intervention historically used to modify autistic behavior toward neurotypical presentation; contested within neurodivergent communities as compliance-training rather than support.
  • Medical model — frames neurodivergence as a disorder or deficit located in the individual, treatable or manageable through diagnosis and intervention.
  • Social model — frames disability/neurodivergence as produced by environments and structures not designed for a given person’s needs, rather than a deficit within the person.
  • Neurodiversity-as-environmental-adaptation — frames neurodivergent traits as functional adaptations to different environments or contexts, contingent on fit rather than universal deficit or universal difference.
  • Developmental disability framing — treats neurodivergence as a disability category, often the basis for legal rights, services, and institutional accommodation.
  • Diagnosis-based access — accommodation or support contingent on formal clinical diagnosis.
  • Need-based access — accommodation or support based on stated or observed need, independent of formal diagnosis.
  • Masking — the effortful suppression or performance of behavior to appear neurotypical (eye contact, affect, speech patterns), often costly to sustain.
  • Stimming — repetitive self-regulatory movement or behavior (rocking, hand-flapping, repeated phrasing); often pathologized or misread as disruptive.
  • RSD (Rejection Sensitive Dysphoria) — an intense emotional response to perceived criticism or rejection, disproportionate to the triggering event, common in ADHD and autism.
  • Interoception — the sense of internal bodily states (hunger, heart rate, emotion); often atypical in neurodivergent people, affecting self-report of needs and feelings.
  • Alexithymia — difficulty identifying and describing one’s own emotions; distinct from lacking emotions.
  • NT / ND — shorthand for neurotypical and neurodivergent.
  • Mutual aid — community-based, non-institutional resource-sharing and support, often positioned as an alternative or supplement to formal/institutional systems.

Terms as used in this document

These aren’t standard clinical or theoretical terms — they name patterns specific to this document’s argument, and are defined here for that reason.

  • Identity-invocation-as-conflict-engine — using marginalized-identity framing to escalate personal conflict or manufacture engagement, as distinct from invoking identity to explain a genuine access need.
  • Engagement-farming — conflict pursued because it reliably generates attention or reaction, identifiable by pattern (targeting, escalation cadence, refusal of resolution) rather than by stated motive.
  • Distraction-function — the effect of a conflict reliably preventing substantive dialogue from occurring, assessed by outcome regardless of whether that outcome was intended.
  • Context collapse (temporal) — resurfacing an old statement as though it reflects a person’s current position, treating change over time as inconsistency or hypocrisy.
  • Consent-to-context — the requirement that a person be present and able to respond in the space where they’re being discussed, distinct from merely being quoted or referenced elsewhere.
  • Gatekeeping legitimacy — denying someone standing to identify as neurodivergent, disabled, or in need of accommodation based on another party’s independent judgment rather than the person’s own account.

Leave a Comment