The Evolutionary Default: Why Racial Disparities Are Expected (And Not Racism)
50,000 years of divergent selection created the trait distributions that drive modern gaps.
I never wanted to write about race or anything “racial.” I never went looking for “race science” or “racial genetic research.”
What pushed me here was watching “disparities prove racism” become the mandatory narrative from 2016 onward — accelerating through COVID and George Floyd. Every gap in outcomes was attributed to discrimination. Every institution was “systemically racist.” Every skeptical question was itself evidence of racism.
But the environmental explanations didn’t survive contact with basic questions.
Why do the same approximate population outcomes appear across radically different countries, decades, and policy environments even after multiple generations?
Why do gaps persist or widen after massive intervention?
Why do higher-SES members of some groups underperform lower-SES members of others on standardized measures?
Why do higher-SES members of some groups commit more violent crime per capita than lower-SES members of others?
The answers I got weren’t answers — they were accusations. Ask why the model keeps failing and you’re told the question itself is racist. That’s when I started digging.
I didn’t arrive at hereditarian ideas because I wanted them to be true. I arrived here because every alternative explanation collapsed under scrutiny, and because the people insisting loudest that the science was settled were the same ones blocking the science that could settle it.
Here’s what I actually believe, laid out plainly.
The Core Logic
Human populations were geographically separated for 50,000+ years under different selection pressures. What we call ‘race’ is simply a crude but functional proxy for these divergent genetic clusters (ancestry groups). Cognitive and behavioral traits are heritable and polygenic.
We acknowledge population differences in height, disease susceptibility, skin pigmentation, lactose tolerance, and athletic aptitude without controversy.
There is no principled reason the brain was exempt from divergent selection.
Therefore, we should expect differences in trait distributions across populations — means, variances, tails. These would produce different outcome distributions even in identical environments.
This isn’t a hypothesis I’m tentatively proposing. It’s the logical default.
The burden of proof falls on anyone claiming tens of thousands of years of separation under different conditions produced zero cognitive or behavioral divergence while producing divergence in everything else.
That’s the extraordinary claim.
The Evidence Lines Up
Diaspora patterns track origin populations worldwide. Immigrants from different regions show outcome distributions that correlate with their ancestral populations, not their destination countries. While elite migration policies can select for the "smart fraction" of a specific group (e.g., Indian engineers or Nigerian physicians representing the top 0.1% of their home nations), unselected waves (refugees, lottery winners, birthright cohorts) consistently revert to the rank-order of their origin populations. This holds across continents, across decades, across radically different policy environments.
Gaps survive and sometimes invert SES controls. Higher-SES members of some groups still underperform lower-SES members of other groups on standardized measures. The “unmeasured environment” escape hatch requires ever-more-elaborate auxiliary hypotheses.
Environment is largely downstream of genetics. Treating the environment as an independent variable is the “Sociologist’s Fallacy.” Parental genes shape family environments, and aggregate genetics shape institutions. While individuals can experience absolute gains by moving to better environments (better nutrition, safety), the relative rank-orders persist. When researchers “control for SES,” they are often controlling for a downstream effect of the very trait they’re trying to isolate.
Environmental interventions show diminishing returns. Decades of massive educational investment haven’t closed gaps. At some point you have to ask whether the model is wrong.
Athletics demonstrates the principle without controversy. West African ancestry dominates sprinting. East African ancestry dominates distance running. The NBA is ~70% Black. Nobody calls this a “hierarchy” — it’s just trait distributions shaped by evolution. Apply the same logic to cognition and suddenly it’s forbidden.
The global pattern is remarkably consistent. Whether you look at national outcomes, regional outcomes within countries, or diaspora outcomes across different host nations, the same population rank-orders appear. Environmental theories require these patterns to emerge from completely different causal structures in each context while somehow producing identical rankings.
The hereditarian explanation is parsimonious: one upstream cause, many downstream effects.
The Genetic Evidence Is Arriving
Critics have long said: “Show us the genes.” Fair enough. The tools weren’t available. Now they are.
Whole-genome sequencing is closing the “missing heritability” gap. The November 2025 Nature paper by Wainschtein et al. using WGS on 347,630 individuals found that sequencing captures ~88% of pedigree-based heritability. Rare variants (previously invisible to GWAS) explain ~20% of the genetic variance.
For educational attainment (EA), h²_WGS = 0.347 (h²_PED = 0.409)
For fluid intelligence (FI), h²_WGS = 0.328 (h²_PED = 0.405)
This matters because rare variants have larger effect sizes and disproportionately impact distribution tails. If populations differ in rare variant burdens due to bottlenecks, drift, or historical selection, the tails diverge more than the means — exactly where high-stakes outcomes (elite performance, severe dysfunction) are determined.
Polygenic scores predict across populations despite claims they can’t. The standard objection is that European-derived PGS don’t “port” to other ancestries. But Piffer & Kirkegaard (2024) tested European-trained polygenic scores on variation within 28 Chinese provinces using 137,000 Han Chinese genomes.
Results:
European EA polygenic scores predicted provincial IQ differences at r = 0.52
European height PGS correlated with measured height at r = 0.71–0.82
Within-family European height PGS (controlling for parental environment confounds) tracked the north-south gradient at r = 0.82
A score trained entirely outside Asia captures non-trivial regional variation inside China, consistent with shared causal genetic architecture. PGS portability may work far better than critics claim — especially as sample sizes grow.
The “no selection signal” objection is becoming dated. Critics claimed no robust evidence of divergent selection on cognitive traits between populations. But detecting selection on highly polygenic traits is methodologically difficult with older tools. Jensen’s method applied to Chinese provincial data shows small but significant evidence of directional selection (r = 0.25–0.27 for trait-relevant SNPs).
Portability debates are often constrained by data: the weakest performance typically appears where large ancestry-matched GWAS/WGS training sets are still scarce, which limits PRS construction and validation. That limitation doesn’t imply genetic influence is absent; it mostly implies we’re still underpowered outside Europe (and increasingly East Asia).
As WGS data accumulates and rare variants are mapped, expect this to clarify.
The Mainstream Position Is Incoherent
Watch how the goalposts move:
“Race is a social construct.” Used to dismiss genetic research. But somehow race is real enough to measure arrest rates, test score gaps, income disparities, and representation in professions. If the categories are too fuzzy for genetics, they’re too fuzzy to diagnose discrimination. Pick one.
“Disparities prove discrimination.” This assumes the blank slate as the baseline — that equal outcomes are expected absent bias. But if populations differ in trait distributions for any reason (genetic, cultural, or otherwise), disparities are expected even with zero discrimination. “Disparity = discrimination” is a non sequitur.
“Within-group heritability doesn’t imply between-group genetic differences.” Technically true. But this objection implies too much. It doesn’t establish that between-group differences aren’t genetic — it just says within-group heritability alone can’t prove they are. The question is empirical, and the emerging evidence (PGS cross-population prediction, consistent global patterns, WGS rare variant mapping) increasingly points toward some genetic contribution.
“We need more research before drawing conclusions.” Funny how this caution only applies in one direction. Environmental causation is assumed by default despite relying on unfalsifiable constructs (”structural racism,” “historical trauma,” “stereotype threat”). Hereditarian hypotheses require impossible proof while environmental hypotheses require none.
Gatekeeping Is Real and Documented
This isn’t paranoia. It’s on the record.
James Lee, principal author of the largest GWAS studies on educational attainment, documented in City Journal (October 2022) that NIH was denying researchers access to the Database of Genotypes and Phenotypes (dbGaP) for intelligence research. The stated justification? Studying the genetic basis of cognitive traits is “stigmatizing.” NIH demanded updates on ongoing research with implied threats to withdraw access if answers weren’t satisfactory. Lee searched NIH regulations and found no legal basis for these restrictions—they appear to be invented by “anonymous bureaucrats with ideological motivations.”
Stuart Ritchie (King’s College London) reported encountering the same barriers. The NIH Alzheimer’s database explicitly prohibits research “into the genetics of intelligence” — even when the goal is studying Alzheimer’s risk factors. When he inquired, NIH confirmed: “the association of genetic data with any of these parameters can be stigmatizing.”
Nature Human Behaviour made the policy explicit in an editorial: the journal will not publish studies showing “the wrong kind of differences between human groups.”
The 2023 Hastings Center consensus report (”Wrestling with Social and Behavioral Genomics”) defined racial comparisons for cognitive traits as research of “greatest concern” requiring “compelling justification” and heightened methodological scrutiny that isn’t applied to environmental research.
Panofsky et al. (2024) in the Hastings Center Report went further, explicitly calling for a “more-closed evidentiary culture” in behavioral genetics. They praised the 2023 consensus report as “a salutary development” precisely because it moves toward restricting inquiry.
This enforcement isn't new; it has just become more bureaucratic. Decades ago, J. Philippe Rushton was ostracized not because his data was fabricated, but because he committed the moral transgression of observing differences. His evolutionary frameworks (like differential K theory) were debatable, but his primary 'sin' was simply documenting that populations differ in trait distributions. Critics conflated description with moral hierarchy, accusing him of 'building a hierarchy' simply for reporting data. Today, that same moral panic has been formalized into NIH funding restrictions and journal submission guidelines.
This isn’t inference about unstated bias. This is stated policy, in major journals and federal funding agencies, documented by the researchers being blocked.
The result: Hereditarian work migrates to preprints and independent platforms while mainstream journals gatekeep. Then critics point to the lack of mainstream publication as evidence the work isn’t credible. It’s circular.
Meanwhile, the tools that could resolve the question (large-scale multi-ancestry WGS studies on cognitive traits) don’t get funded. We could test this directly. We choose not to. And then we pretend the absence of evidence is evidence of absence.
Related: The Anti-Hereditarian Shell Game
The Policy Costs Are Real
If the blank-slate premise is wrong, policies built on it cause harm.
Affirmative Action
The math is simple. If you lower selection thresholds for one group to achieve demographic representation, you mathematically guarantee lower average qualifications among selected members of that group. This follows directly from truncating different distributions at different points.
The SFFA v. Harvard data showed Black applicants in lower academic deciles had higher admission probabilities than Asian applicants in higher deciles. This creates qualification gaps among admits that persist through graduation because elite schools graduate nearly everyone they admit.
If qualification metrics predict performance—and they do (MCAT predicts board passage, LSAT predicts bar passage)—then preferences create real competence gaps in the professional pipeline.
Cremieux’s analysis shows this clearly: affirmative action makes it rational to statistically discriminate against beneficiary groups, because their credentials signal less than the same credentials from non-preferred groups. The policy designed to help them undermines trust in their qualifications.
The Compounding Effect
This has been operating for decades. Lower standards → less qualified practitioners → degraded training for the next generation → institutional quality decline → the people setting future standards are themselves products of the degraded system.
Medical data is consistent with this: Blacks are ~5% of medical residents but ~20% of those dismissed from programs. Black physicians received 4.4% of California Medical Board complaints while being 2.7% of physicians.
Killer King
The Martin Luther King Jr./Drew Medical Center in Los Angeles — ”Killer King” — operated from 1972 to 2007 with catastrophic outcomes:
$20.1 million in malpractice payouts (1999–2003)—worst per-patient in California
Patients with minor injuries died from incompetent care
Police wouldn’t let wounded colleagues be taken there
Locals fled from ambulances to avoid the hospital
Documented cases:
A 9-year-old with scrapes and broken teeth—sedative overdose, wrong ventilator settings, dead.
A sheriff’s deputy with gunshot wounds—joking with nurses on arrival, dead from a lethal drug combination two days later.
A woman receiving AIDS-positive blood during a hysterectomy because nobody checked.
Why wasn’t it fixed? Anyone who tried was called racist. Maxine Waters organized demonstrations when LA County discussed reforms. Politicians were cowed for decades. The hospital lasted 30 years after the problems were publicly known.
The victims were disproportionately the minority patients it was supposed to serve.
The Policing Misdirection
The 2020 discourse assumed what it needed to prove.
Observed: Black Americans arrested and shot at higher rates per capita.
Assumed: This proves police bias.
Ignored: If crime commission rates differ, arrest rates should differ proportionally even with zero bias.
Victim surveys (where victims report offender race) roughly match arrest demographics for violent crime. The disparity in arrests largely tracks offense rates, not police bias.
The entire “systemic racism in policing” narrative depends on assuming equal underlying crime rates—the blank slate applied to behavior. If that assumption is wrong, the inference collapses.
This is the same logic applied everywhere: achievement gaps prove racist schools, income gaps prove racist employers, arrest gaps prove racist police. All of it assumes equal distributions as the expected baseline. All of it collapses if the baseline is wrong.
What I’m Claiming
Let me be precise, because people will misrepresent this:
I’m claiming:
The hereditarian hypothesis follows from basic evolutionary logic
The evidence is consistent with it and increasingly direct
The blank-slate alternative requires implausible assumptions
Policies premised on blank-slate causation have measurable costs
The asymmetric treatment of these hypotheses is unscientific
Gatekeeping prevents the research that could resolve the question
I’m not claiming:
That magnitudes are precisely known
That any group is “superior” or “inferior” (data does not determine “worth”)
That individuals should be treated based on group averages
That environment doesn’t matter at all
That all hereditarian research is methodologically perfect
That racial differences are permanent (future selection can radically reshape these distributions again — perhaps fully bridging gaps or creating more divergence)
The scientifically honest position: populations separated for tens of thousands of years under different selection pressures almost certainly differ somewhat in cognitive and behavioral trait distributions.
The magnitude is an empirical question, but “some difference” is the logical default — not something requiring extraordinary proof.
Why This Matters
If population differences in trait distributions are real (even partially) then:
“Disparity = discrimination” collapses as a framework. Unequal outcomes don’t prove unequal treatment. Policy built on that assumption is built on sand.
Affirmative action is counterproductive. It selects for lower performance, degrades institutions, and harms the people it claims to help by devaluing their credentials.
Integration without acknowledgment of differences creates permanent grievance. If gaps can’t close because they partly reflect genetics, then the framework that attributes persistent gaps to persistent racism guarantees endless accusations of discrimination that can never be resolved.
We’re lying to ourselves about human biodiversity. We accept it for height, disease, athletics, and physical appearance. We deny it for the brain. There’s no principled reason for this—only political convenience.
The Demand for Honest Inquiry
I’m not asking anyone to accept hereditarian conclusions uncritically.
I’m asking for symmetric rigor:
End the evidentiary double standard. Currently, environmental theories are accepted by default based on loose correlations (e.g., “poverty correlates with outcomes”), while hereditarian hypotheses are rejected unless they provide perfect molecular causality. This asymmetry makes honest inquiry impossible.
Fund the research that could actually resolve the question. We need large-scale, multi-ancestry Whole Genome Sequencing (WGS) on cognitive traits. The tools exist; the funding is being withheld.
Confront the opportunity cost. We cannot solve problems we misdiagnose. Continuing to pour resources into environmental interventions that yield diminishing or negative returns isn’t virtuous; it’s wasteful. We need to know the true magnitude of the genetic contribution so we can stop demanding impossible outcomes from our institutions.
Stop treating “this could be misused” as grounds for suppressing inquiry. The cost of ignorance (indefinite waste and misdiagnosis) is higher than the risk of knowledge.
Acknowledge that the current “consensus” is enforced by gatekeeping, not by evidence.
The truth doesn’t care about our preferences. Neither should science.