In defending some comments she made regarding white liberals and their support for President Obama, Melissa Harris-Perry looks at three common objections to conversations about race in the United States: “prove it,” “I have black friends,” and “who made you an expert?” While these are all familiar responses, the first one is a particularly sociological point that raises questions about how we view society and how this plays out in court:

The first is a common strategy of asking any person of color who identifies a racist practice or pattern to “prove” that racism is indeed the causal factor. This is typically demanded by those who are certain of their own purity of racial motivation. The implication is if one cannot produce irrefutable evidence of clear, blatant and intentional bias, then racism must be banned as a possibility. But this is both silly as an intellectual claim and dangerous as a policy standard.

In a nation with the racial history of the United States I am baffled by the idea that non-racism would be the presumption and that it is racial bias which must be proved beyond reasonable doubt. More than 100 years of philosophical, psychological and sociological research that begins, at least, with the work of W.E.B. Du Bois has mapped the deeply entrenched realities of racial bias on the American consciousness. If anything, racial bias, not racial innocence is the better presumption when approaching American political decision-making. Just fifty years ago, nearly all white Democrats in the US South shifted parties rather than continuing to affiliate with the party of civil rights. No one can prove that this decision was made on the basis of racial bias, but the historical trend is so clear as to require mental gymnastics to imagine this was a choice not motivated by race.

Progressives and liberals should be particularly careful when they demand proof of intentionality rather than evidence of disparate impact in conversations about racism. Recall that initially the 1964 Civil Rights Act made “disparate impact” a sufficient evidentiary claim for racial bias. In other words, a plaintiff did not need to prove that anyone was harboring racial animus in their hearts, they just needed to show that the effects of a supposedly race neutral policy actually had a discernible, disparate impact on people of color. The doctrine of disparate impact helped to clear many discriminatory housing and employment policies off the books.

Michelle Alexander brilliantly demonstrates in The New Jim Crow, the pernicious effect of the Supreme Court moving away from disparate impact as a standard to forcing plaintiffs to demonstrate racist intention. This new standard has encouraged the explosive growth of incarceration of African-Americans, turning a blind eye to disparate impact while it demands “proof” of racial bias.

I believe we must be careful and judicious in our conversations about racism. But I also believe that those who demand proof of interpersonal intention to create a racist outcome are missing the point about how racism works. Racism is not exclusively about hooded Klansmen; it is also about the structures of bias and culture of privilege that infect the left as well.

I like how Harris-Perry flips this objection: looking at the broad sweep of American history, from its days of more overt racism to more covert racism today, why don’t we assume that racism plays a role in everyday life in this society? Can we really assume, as many seem to do, that the issues with race ended at some point, either in the Civil Rights legislation of the 1960s or in the election of minority politicians or the ending of segregationist society in the South? With plenty of indicators of racial disparity today, from online comments from young adults to incarceration rates to homeownership to wealth to residential segregation, perhaps we should we see racism as a default feature of American society until proven otherwise.

Harris-Perry hints at one reason why it is difficult for Americans to see the effects of racism: the court system moving to the burden of proof shifting to “proving” “racist intention.” Without the proverbial smoking gun, it then becomes more difficult to develop arguments just from data and patterns, even if they are overwhelming. While the recent court case involving gender discrimination at Walmart and sociologists siding with the prosecution isn’t about race, it illustrates some of these principles. The data suggests discrimination may have taken place as more women did not receive promotions or pay raises. But without “proof” that this was a deliberate Walmart policy meant to harm women, the numbers may not be enough. The same holds true with race: “statistical discrimination,” stereotypes about large groups of people, may be okay because no individual or corporation can be held directly responsible for the outcome.