Daniel Simons: It's what we know that just ain't so

The world isn't as it seems, experimental psychologist Daniel Simons says, and demonstrates through repeated short clips the problem isn't what we know, but what we know just ain't so.*

Beginning with an illustration drawn by Julian Beever - an IF alum, by the way - he makes the point about the power of illusion and about how we should call into question our own beliefs about what our eyes are telling us. 

The problem is compounded when our attention is drawn to a particular feature of the world.

He demonstrates this using the well known "gorilla" videos for which he is known, but after a number of people correctly guess that a gorilla has invaded the scene, Simons replays it to point out the other changes - blindingly obvious in retrospect - that have been missed by the smartypants in the audience. Updated "gorilla" videos can be found on the Web.

"Satisfaction of search" accounts, he adds, for the number of suits filed against radiologists. You're only aware of things you have noticed, not what you have missed. It's a simple and profound thought. And while a missed cancer diagnosis is tragic, noticing everything would be far more problematic. In fact, it's impossible, except, perhaps, for Dunning–Kruger graduates.

Beside being humbling, these limits on our cognition have real world implications. Juries tend to be very trusting about confident eyewitnesses, and confident eyewitnesses are often wrong. He proceeds to illustrate this from his own experience of being completely and hopelessly wrong about a vivid memory related to his interaction with graduate students. This is a human problem, but when they occur with public figures and are amplified by partisan point-making, for example, the tendency is to assume that the teller is either crazy or lying. Not helpful.

One consequence of trusting our intuition about what we've seen can be illustrated in the video posted here. About half the people approached in this test recognize, according to Simons and the test subjects' accounts, that they're giving direction to yet a second stranger, and still continue as if nothing has happened. What's up with that?


* shout out to Will Rogers.