LAST summer, around the anniversary of Martin Luther King Jr.’s “I Have a Dream” speech, I wrote a column making . gingerly . a case for optimism about race and politics in America.

My argument was basically this: As much as racial controversy has marked the presidency of Barack Obama, our race-related policy cleavages are still less dramatic than at any previous point in America’s history. It isn’t just that there’s no contemporary equivalent of the conflict over Jim Crow, in which one side had to be defeated utterly for racial progress to be possible. It’s that on many racially entangled issues, from education to criminal justice to various socioeconomic challenges, the key policy debates are less polarized than in the 1970s and 1980s, and the impact of potential reforms on whites and blacks seems much less zero sum.

But after watching as Ferguson, Mo., seethed and smoldered, it’s worth offering a case for greater pessimism. Not because the optimistic arguments are no longer credible, but because we’ve just had an object lesson in why they might be proved wrong.

This lesson isn’t exactly new; indeed, it’s been offered by both parties throughout this presidency. Ultimately, being optimistic about race requires being optimistic about the ability of our political coalitions to offer colorblind visions of the American dream . the left’s vision stressing economics more heavily, the right leaning more on family and community, but both promising gains and goods and benefits that can be shared by Americans of every racial background.

In the Obama era, though, neither coalition has done a very good job selling such a vision, because neither knows how to deliver on it. (The left doesn’t know how to get wages rising again; the right doesn’t know how to shore up the two-parent family, etc.) Which has left both parties increasingly dependent on identity-politics appeals, with the left mobilizing along lines of race, ethnicity and gender and the right mobilizing around white-Christian-heartland cultural anxieties.

For a while the media has assumed that this kind of identity-based politics inevitably favors the left, because 21st-century America is getting less white every day.

But that’s too simplistic, in part because the definitions of “white” and “minority” are historically elastic. If a “white party” seems sufficiently clueless and reactionary, it will lose ground to a multicultural coalition. But as African-Americans know from bitter experience, “whiteness” has sustained itself by the inclusion of immigrants as well as by the exclusion and oppression of blacks. That history suggests that a “multicultural party” may always be at risk of being redefined as a grievance-based “party of minorities” that many minorities would prefer to leave behind. (And leadership matters, too: A protean figure like Barack Obama can put together a genuine rainbow coalition, but it’s not clear how many other politicians can do the same.)The key point here, though, is that whichever coalition is ascendant in this scenario, a politics divided primarily by identity is likely to be more poisonous than one in which both parties are offering more-color-blind appeals.

Unfortunately, identity is also the most primal, reliable form of political division. And Ferguson has provided a case study in exactly how powerfully it works.

There was a moment, early in the debate over the death of Michael Brown, when it felt as if this story might vindicate the case for optimism about racial politics . that the original tragedy might be sufficiently transparent, the subsequent police misconduct in quelling protests sufficiently clear-cut, for Ferguson to become a more powerful exhibit in the increasingly bipartisan case for various criminal justice reforms.

But then it became clear that the situation was murkier . that the cop had witnesses and physical evidence supporting his side of the story, that police had to deal with looters as well as peaceful protesters. As John McWhorter wrote in Time magazine, by the time the grand jury handed down its non-indictment the original narrative about Ferguson could only survive with “a degree of elision” and “adjustment.” Which meant, predictably, that the potential for consensus receded, and how people felt about the story became primarily a matter of identification instead.

Do you identify more with a black teenager or with a cop? With protesters menaced by playing-soldier cops or with business owners menaced by the protest’s violent fringe? With various government spokesmen or with, say, Al Sharpton?Again, this is not unusual; this is how political division and racial division often interact.

And there’s still nothing inevitable about this interaction. Rand Paul, the Republican who’s pushed hardest to change the old paradigm on race and crime, is still talking about criminal justice reform in the wake of Ferguson. The path to a less identity-driven kind of politics is still open.

But it’s clearer today how easy, how human, it will be to leave that path untaken.