A patient describes long-term post-viral exhaustion, with symptoms running across several body systems. They leave with a prescription for anti-anxiety drugs, and their account goes into the record as somatic symptom disorder.
Every institution with authority over people’s lives keeps a boundary between legitimate evidence and noise, and it rarely presents that boundary as a political choice. It presents the boundary as scientific rigor, an objective standard or neutral procedure.
Whoever defines what counts as knowledge decides which harms need an institutional response and which can be dismissed as subjective complaints.
Maps and categories were drawn to serve power
Knowledge systems don’t float outside the history of power. The categories an institution uses to measure reality were built by specific people with specific material incentives.
European colonial mapmaking didn’t merely record terrain. It drew boundaries that dissolved Indigenous land tenure, turned common grazing grounds into state property and erased existing settlements from legal records. In nineteenth-century medicine, psychiatric categories pathologized resistance to enslavement as drapetomania and female autonomy as hysteria. In mid-twentieth-century urban planning, redlining maps used actuarial criteria to classify integrated neighborhoods as hazardous, cutting off credit while presenting the decision as objective risk management.
These categories weren’t innocent errors made with primitive instruments. They worked as instruments designed to turn extraction and domination into technical necessity, and they let bureaucrats carry out violence while seeing themselves as disinterested administrators of fact.
Thresholds still keep inconvenient evidence out
Modern institutions still use thresholds of evidence to shield their operations from contradiction.
When patients describe chronic physical symptoms (environmental toxicities, connective tissue dysfunction, long-term post-viral exhaustion), clinical workflows often file the testimony under anxiety or somatic symptom disorder. The dismissal protects the institution’s model of moving patients through quickly: investigating complex complaints across several body systems is expensive, while prescribing anti-anxiety drugs keeps the schedule moving.
Regulatory agencies often discount health observations from communities near chemical plants or industrial ports unless the residents can supply long-term epidemiological data. Marginalized communities lack laboratory funding, so their physical suffering registers as anecdotal noise while the factory keeps running.
Benefits decisions keep evidence out with opaque algorithms. When a system uses a risk-scoring model to deny healthcare coverage or cut off disability payments, claimants are told that the model’s proprietary weights represent objective statistical optimization, and what a claimant can’t do day to day is treated as an unreliable claim against the authority of the model.
In each case, calling vernacular knowledge “illegitimate” serves an administrative goal: it insulates the institution from having to update its records or pay for its errors.
Keep the rigor, change who gets heard
Challenging this doesn’t require abandoning empirical rigor or embracing relativism. Reality exists independently of institutional models, but access to it is politically organized: evidence has to be generated, indexed, heard and linked to authority before it can trigger a course correction.
Making institutions accountable for what they count as knowledge takes four shifts in how evidence is treated.
First, separate what can be known from what an institution has certified. Facts exist before agencies create billing codes for them. A community living with groundwater contamination knows its water is toxic before state hydrologists spend three years publishing a verification study.
Disability shows what happens when certification also becomes the price of help. Mike Oliver, who named the social model of disability, argued that people are disabled less by an impairment than by the barriers built around it, like a missing ramp or an absent interpreter.1 The help that removes a barrier still waits on a certificate: under the Americans with Disabilities Act, an employer can ask for documentation from a health care professional before granting an accommodation for a disability that isn’t obvious, and a worker who doesn’t provide it isn’t entitled to one.2 So the institution builds the environment that disables, demands you prove you’re broken before it’ll help you through it, and charges you for the certificate.
Second, treat vernacular testimony as a primary empirical signal. When workers, patients and residents repeatedly report that their lived reality doesn’t match official performance dashboards, institutions should handle their accounts as sensor readings rather than cognitive distortion.
Third, demand standards of evidence that people can meet. When an institution demands proof of harm, it shouldn’t place the burden of collecting data entirely on the injured party. Making claimants fund their own forensic research is governance through exhaustion.
Fourth, preserve channels for revising the institution’s categories. A system of knowledge is rigorous only if contradictory evidence can force it to revise its core assumptions.
Believing the dismissed has a structural limit. Living through an injury gives a person direct knowledge others lack, but personal experience doesn’t make anyone an infallible judge of systemic truth. What people at the margin know counts as vital empirical evidence. The remedy is an open evidentiary hearing whose findings institutions have to act on, with every conclusion still subject to scrutiny and appeal.
Notes
Mike Oliver, The Politics of Disablement: A Sociological Approach (Macmillan, 1990). Oliver coined the term “social model” for the distinction between impairment and disability that the Union of the Physically Impaired Against Segregation had drawn. For a short account of the model and the barriers it names, see Ravi Malhotra, “The New Politics of Disablement: The Contribution of Mike Oliver,” Socialist Project, March 2019.
U.S. Equal Employment Opportunity Commission, “Enforcement Guidance on Reasonable Accommodation and Undue Hardship under the ADA,” question 6. The documentation can be limited to what’s needed to establish an ADA disability, and it can’t be required when the disability and the need for the accommodation are both obvious.