At three months old, infants already show a preference for faces of their own race. They have not learned to walk, talk, or recognize themselves in a mirror, but the brain has already begun sorting the world into familiar and foreign. This is where racism begins, and that is the most important conversation we are not having.
Racism is not simply the product of evil minds. It is the product of ordinary human psychology shaped by evolution, activated by culture, and sustained by the choice not to deeply examine it. That distinction matters enormously, because while the origins of bias may be human, its consequences are not abstract. Racism kills. It impoverishes. It tells people their lives are worth less. Every year in hiring decisions, in hospital rooms, in interactions with police, in the quiet mathematics of who gets opportunity and who does not, bias becomes damage. Understanding why racism exists is not an invitation to excuse it; it is a starting point to dismantle it.
For the vast majority of human history, people lived in small, homogenous tribal groups where the ability to rapidly distinguish “us” from “them” was not a social preference but a life or death calculation. The brain adapted accordingly, developing fast, automatic categorization systems that sort people into groups before conscious thought kicks in. Psychologist Henri Tajfel tested just how little it takes to trigger this machinery. In his Minimal Group experiments, subjects were divided into groups based on something as meaningless as a coin flip, and immediately began favoring their own side. The bias came first. The justification came later.
The wiring is universal. What gets loaded into it is not. Children begin absorbing more complex racial attitudes between the ages of three and five, long before they possess the critical framework to evaluate what they are learning. Psychologist Rebecca Bigler demonstrated that children do not need to be explicitly taught prejudice; they need only to observe which groups their environment treats as different, and their categorization systems do the rest. This research is supported by John Jost's system justification theory, which reveals how the brain is motivated to perceive existing social hierarchies as legitimate and natural, even when those hierarchies are unjust, because a predictable world, even an unfair one, feels safer than an unpredictable one. By the time most people reach adulthood, bias is not a belief they have chosen. It is the architecture they have inherited.
And yet, inheritance is not destiny. The same brain that automates bias is also capable of interrupting it. Research by Jennifer Eberhardt shows that simply slowing down decision-making can reduce racial bias in high-stakes environments like policing. Studies on implicit bias, including the work of Mahzarin Banaji, reveal that awareness alone does not erase prejudice, but it creates a crucial pause, a moment where instinct can be questioned rather than obeyed. That pause is where responsibility begins.
This is why the conversation we avoid matters so much. It is easier to talk about racism as an anomaly, something extreme and distant, embodied by individuals we can point to and condemn. It is far harder to confront it as something ordinary, something quietly operating in everyday decisions, including our own. But if racism were only the product of extreme hatred, it would be far easier to solve.
None of this is an excuse. It is a call to precision. If we misunderstand the origins of bias, we will misdiagnose the solution. Shaming individuals without addressing the systems that reinforce bias changes very little. At the same time, focusing only on systems while ignoring personal accountability leaves those systems intact. The work is both internal and external: to question our reflexes, and to redesign the environments that shape them.
There is something deeply hopeful in this, even if it does not feel like it at first. If racism were purely innate, it would be unchangeable. If it were purely cultural, it would be easier to shift. The reality sits in between. Although bias may be automatic, it can be unlearned, interrupted, and reshaped, but only if it is first acknowledged.
