Hack40.Blind to Change

Hack 40. Blind to Change

We don't memorize every detail of a visual scene. Instead, we use the world as its own best representationcontinually revisiting any bits we want to think about. This saves the brain time and resources, but can make us blind to changes.

Both our vision [Hack 14] and attention [Hack #34] have far coarser resolutions than we'd have thought. What's more, there are gaps in our vision across time [Hack #17] and in space [Hack #16], but our brains compensate for these gaps and knit together a rather seamless impression of the world.

And this gapless impression is utterly convincing. Most of the time we don't even realize that there are holes in the information we're getting. And so we believe we experience more of the world than we actually do. There are two possibilities as to what's going on here. The first is that we build a model inside our heads of the world we can see. You can test to see whether this is the case.

Imagine you are looking at a picture. There's a flicker as the picture disappears and appears again. What's different? If we made and kept a full internal representation of the visual world inside our heads, it would be easy to spot the difference. In theorybefore memory decay set init should be as easy as comparing two pictures (before and after) side by side on a page. But it isn't.

So that puts paid to the first possibility. The other is that you don't build a full internal model of what you're seeing at allyou just think you do. The illusion is maintained by constant sampling as you move your eyes around, a part of what is called active vision [Hack #15] . After all, why bother to store information about the world in your head when the information is freely available right in front of your very eyes?

The proof of the pudding for active vision is testing the consequence that, if true, you should find it very difficult to spot changes between two scenes, even with just a short flicker in between. Since most of the two separated images aren't stored in memory, there's no way to compare them. And, true enough, spotting any difference is very difficultso hard, in fact, that the phenomenon's been labeled change blindness.

3.8.1. In Action

You can try an animated GIF demo, which we made, at http://www.mindhacks.com/book/44/changeblindness.gif, both frames of which are shown in Figure 3-8. Shown side by side, the difference between the two versions of this picture is obvious.

Figure 3-8. The difference is easy to spot when you're allowed to look at both versions of the "same" picture at once¹

But if you don't know what you're looking for, it can be impossible to spot. Load the images in the following URLs and have a look. If you're finding the first one hard, have a look at the man's noseyou can be looking right at the change in the image and still not spot it for a frustratingly long time.

• http://nivea.psycho.univ-paris5.fr/ASSChtml/couple.gif (an animated GIF)

• http://www.usd.edu/psyc301/Rensink.htm (a Java applet)

3.8.2. How It Works

You need the momentary blink between the pictures so you are actually forced to compare the two pictures in memory rather than noticing the change as it happens. Interestingly enough, the blink doesn't actually even need to cover the feature that's changing, as another demonstration at http://nivea.psycho.univ-paris5.fr/ASSChtml/dottedline.gif shows. Rather than blanking out the entire image, distracting patterns momentarily appear overlaid on it to divert your attention from the change.

You're just as blind to the altering feature when patterns flash up, even though the picture as a whole remains present the entire time. It's enough that your attention is momentarily distracted from picking up on the change, forcing you to rely on your memory for what the scene looked like half a second agowe're not talking long-term memory here.

3.8.3. In Real Life

This isn't just lab theory. Change blindness can help you pull some great tricks outside of the lab and without the aid of a computer. A classic experiment by Daniel Simons and Daniel Levin2 is a perfect example. One of the pair would stop a passerby to ask for directions. In the midst of the kindly passerby's attempt at giving directions, two men would carry a door between the experimenter and passerby. During this distraction, the experimenter switched places with his colleague, who was a different height and build, sounded different, and was wearing different clothes. Despite these blatant differences, a full half of the people they tried this on didn't notice any difference between the man who started asking for directions and the man who finished listening to them.

3.8.4. End Notes

1. The road markings on the right of the picture change location.

2. Simons, D. J., & Levin, D. T. (1998). Failure to detect changes to people during a real-world interaction. Psychonomic Bulletin and Review, 5, 644-649.

3.8.5. See Also

• Daniel Simons' lab provides a nice collection of movies they've used to demonstrate change blindness (http://viscog.beckman.uiuc.edu/djs_lab/demos.html).

• J. Kevin O'Regan has a great talk entitled "Experience Is Not Something We Feel but Something We Do: a Principled Way of Explaining Sensory Phenomenology, with Change Blindness and Other Empirical Consequences" (http://nivea.psycho.univ-paris5.fr/ASSChtml/ASSC.html).