27 February 2009

The Uncanny Valley

This weekend the University of Pittsburgh's theatre dept will host a Theatre & Cognitive Science Symposium. E & I--Theatre & Cognitive Science majors, respectively--feel the universe aligning ever so slightly.

E will be presenting a paper on facial expressions--a subject of enduring interest & study for Artists, Actors & Cognitive Psychologists. Among other things, this provides an opportunity to employ a term we've been eager to use in public:
The Uncanny Valley.

In brief, the Uncanny Valley is the slight but disturbing gap between a human representation that looks almost real, but not quite. We're talking about 3D animation, robots or sculpture. The hypothesis is that the 'not quite' element--this uncanny valley-- is upsetting to people, and much moreso than a facsimile that looks totally unnatural or completely convincing. Think of the creepiness of a B-grade wax museum.

The most fun thing I've seen on this recently is this graph below. This earnestly charts Familiarity across Human Likeness, comparing Moving vs. Still examples. The range runs from An Industrial Robot to A Healthy Person, and helpfully includes Humanoid Robots, Stuffed Animals, Bunraku Puppets, Corpses (still), Zombies (moving), and a Prosthetic Hand.