Thanks to CuriosityStream for supporting this episode!

Go to CuriosityStream.com/Psych to learn more. [♪ INTRO]. You've probably heard of the uncanny valley before.

It's the sense of unease you feel when you encounter a slightly-too-lifelike robot or the feeling of yikes when you stumble across a very creepy doll. The sense of nope, no thank you, do not want that you might get if you, say, very hypothetically, you kept seeing unsettling CGI cat-humans cavorting around your social media feeds. The uncanny valley is the steep drop-off in how likable we find a character or robot to be when it gets… slightly too close to appearing human.

These days, it's everywhere in pop culture, so it might surprise you to know that there's still a lot we don't understand about what causes it -- and whether it even really exists. The idea of the uncanny valley was first proposed in an essay written by a Japanese roboticist named Masahiro Mori in 1970. He hypothesized that a person's affinity for a robot would increase as the robot became more and more human-like.

And then, very abruptly, that affinity would drop off and be replaced by revulsion. He provided examples by plotting human likeness against likability for things like robots and puppets, and came up with a graph with a sharp dip: that is the uncanny valley. This idea wasn't totally new.

Fifty years earlier, Freud had written about “the uncanny,” and the terror caused by things that reminded us of the familiar. But the obsession with the uncanny valley in pop culture didn't really take off until the early 2000s. In particular, it's been attributed to -- go ahead and guess -- yes the 2004 Polar Express film starring Tom Hanks.

Plenty of early CGI movies flopped in large part because they freaked people out. And scientists really do cite that freakiness as part of the history of the uncanny valley. Then, in 2005, Mori's essay was translated into English, and psychologists started to actually study the phenomenon in earnest.

Most of the research since then has looked at whether or not it's possible to recreate his uncanny valley-shaped plot with real data. And the results have been mixed. An early study of 45 people in 2006 asked participants to rate the likeability of a series of morphed images, ranging from a mechanical robot through an android to a human.

And they did see an uncanny valley-type effect. But even then, the researchers questioned whether a spectrum of morphed images would always produce an uncanny valley. Basically, was the uncanny valley inevitable?

It might not be, because the effect doesn't always show up. In general, studies that use artificially manipulated images morphing between one thing and another have been more likely to support the valley's existence than those using images of real robots. But at least one study, published in 2016, where 342 participants looked at 80 real-life robot faces, did find an uncanny valley curve.

So… it's kind of hard to say. Many of us feel something deep down when we see CGI Taylor Swift, but the research simply hasn't pinned it down yet. Neither is there a clear idea of what might cause it.

There are tons of hypotheses! Some of the less likely ones invoke our perceptions. Like the pathogen avoidance hypothesis, which suggests that disgust towards uncanny faces might have helped us avoid someone who might be carrying a transmissible disease.

Or the mortality salience hypothesis, which suggests that uncanny faces, like clowns, dolls, wax figures, corpses, and zombies, might literally remind us of death. But none of these really have enough evidence to support them. Other hypotheses invoke what's going on in our brains, rather than what we perceive.

The violation of expectation hypothesis suggests that uncanny faces might lead us to think they're going to behave in a way that is human-like and then violate that expectation. The strongest evidence for this comes from research showing that robots and characters with a mismatch between their appearance and movements, or their appearance and voice, freak us out more than ones that just look creepy. But researchers still believe this one doesn't answer all the questions.

The categorical uncertainty hypothesis also tries to make sense of how our brains are processing uncanny faces. This one argues that there's uncertainty at any category boundary. So it's not just humans vs. robots that freaks us out, it's anything that our brains can't put into a nice neat box.

Like a cat person. But there's a problem:. A lot of this research relies on morphed images, but it's possible that the very process of morphing images creates visual artifacts our brains simply Do.

Not. Like. That might confound a lot of these findings.

And then, there are the theories that look at what we think it means to be human. The mind perception hypothesis suggests that we find robots uncanny when we think these non-human things might be capable of the human ability to think, plan, and feel. One 2012 study surveyed 165 participants and found that if they believed that a robot could sense and experience things, that played a role in making a human-like robot seem creepier than a mechanical one.

The study also found that a non-humanoid robot could be made creepy -- if participants were told that it was capable of human-like thinking, feeling, and planning. The dehumanization hypothesis looks at the same idea from a different angle: robots aren't creepy because they look and act more like humans. Rather, it's because they look like us, but don't act enough like us.

Our brains are super good at recognizing faces. The idea here is that when we see one and it doesn't behave the way we expect. Actual Human People to act, that triggers uncanny feelings.

There are arguments to be made, and at least some evidence, in favor of basically all of these hypotheses. So we don't have a lot of answers when it comes to the uncanny valley. It's a phenomenon many of us are pretty darn convinced is real.

We feel it in our guts, even if psychologists can't really tell us why. When Mori wrote his essay in 1970, he was interested in figuring out how roboticists and animators could overcome it -- and that's clearly something we're still struggling with. Some researchers, though, argue that we may never overcome it -- and, in fact, that it's worth keeping around.

It lets us know that we can tell the difference between robots and humans, and that's actually a pretty cool thing for our brains to be able to do. And it might come in handy in the future. Our perceptions definitely filter the way we see the world -- not just androids.

The CuriosityStream original What Is Reality is about that idea, looking at how our brains construct the world around us. And that's just one of the over 2,400 documentaries and non­fiction titles available on CuriosityStream, including exclusive originals, so there's something there for anyone who enjoys learning. And you're here, so you probably do.

CuriosityStream is available worldwide, and their titles cover topics ranging from science to lifestyle, and nature to tech. Right now, SciShow viewers can get unlimited access starting at just $2.99 a month, and the first 31 days are completely free if you sign up at curiositystream.com/Psych and use the promo code “psych” during the sign-up process. So thanks for checking them out -- and for supporting SciShow. [♪ OUTRO].