It’s both soothing and sad to recall that there was a time, not so long ago, when America’s attention could be commanded by a controversy as harmless as “the dress.” Remember that photo? Some people saw a white and gold dress; others insisted it was blue and black. Neither could understand how anyone could see anything different. But after some head-shaking, the two camps walked away friends.

The Internet is full of such optical illusions, which simultaneously expose our brain’s hidden subsystems and their mistakes. Some, like the dress, highlight hiccups in our visual cortex. But others reveal more troubling flaws in the way we integrate the information we acquire — flaws that are becoming more pressing as the nation absorbs video of protests happening in Portland, Ore., and elsewhere.

For example, there’s a “selective attention test” available on YouTube that illustrates perceptual distortion by asking you to count the number of times certain players pass a basketball to each other. Go ahead and do it if you’re interested, because further down, I’ll spoil the punchline.

You might imagine that the biases exposed by exercises such as these get worse when politics gets involved. But no need to use your imagination. Researchers at the Yale Cultural Cognition Project showed people a video of protesters clustered around the entrance to a building and asked them to decide whether the protesters were illegally blocking the entrance.

The first twist is that they showed two protest videos — one involving a military recruiter, the other an abortion clinic — and found that it didn’t matter which video you saw as much as whether your politics agreed with the protesters. The second twist is that these were actually the same underlying video. Only the subtitles of what was happening had been altered.

Even without the press of politics, we’re prone to pretty serious processing errors. If you watched the selective attention test, did you see the gorilla? For those who didn’t, there’s a guy in a gorilla suit — except people are often so intent on counting passes that they don’t notice. This “inattentional blindness” is so profound that I once showed that video to someone whose brother worked for the professor who made it, and not only did he fail to see the gorilla, he didn’t even spot his own brother passing basketballs.

Our blind spots are already, in other words, pretty significant. Add politics in, and they can make it hard to see anything straight.

This weekend, a short clip circulated on Twitter, showing a female protester in a street as a line of cops advances. One officer shoves her, hard enough to make her stumble and drop her cellphone; seconds later, a man flies in from outside the frame, smashing into the cop and knocking him to the ground. Other officers wade in; other protesters do the same; the chaos resolves, after a lot of shoving, with protesters on the ground being handcuffed.

I saw it via GQ’s Julia Ioffe, who tweeted, “This isn’t the police keeping the peace. This is them treating their fellow citizens as enemy combatants.” Many replies echoed the sentiment. Others saw, with equal conviction, police responding with restraint after being physically attacked.

Neither was wrong about what was in the video: A police officer was attacked, American citizens were manhandled. But all anyone saw was the element that had commanded their attention — and that was whatever fit the story they were already telling about violent protests or police brutality.

Video is inherently a heavily edited medium. Someone chooses what you see, and unscrupulous editors can use those choices to force viewers into an optical illusion. But even studied neutrality involves choosing where to point the camera and how long to keep it on.

But because video contains so much rich visual information, we tend to feel as if we’re there instead of receiving a highly selective retelling. That makes video seem more authoritative than other mediums. That authority can be a force for good: Cellphone cameras made debates over police tactics less a matter of whether to believe accusers than whether to believe your own eyes. But we still need to remember that what we’re seeing is in some sense an illusion, stripped of vital context by the narrow funnel of a camera lens — and that there can be giant holes in how we integrate what we do see into the rest of what we know.

What happened before and after? How representative of the whole scene, or American society, are the events shown? What is it actually like to be there, breathing tear gas, dodging fireworks? Americans have forgotten to ask themselves those questions. Instead, they’re wondering why the other side can’t see the plain truth.

America can easily survive seeing two dresses. But we’re in pretty bad shape when the visual distortion field is large enough to encompass the most important issues facing our country.

Watch Opinions videos:

Read more: