 |
| Where is Waldo? |
You found Waldo? (squint your eyes and it's the darkest spot on the beach) But did you spot the cactus? did you notice there was not a single shadow? did you notice there was no lifeguard? did you spot the guy with polka dots sunburn? Chances are you didn't, because given a limited amount of time, you probably tried to skim through this picture containing a flurry of details with only one thing in mind, the red-stripped shirt pattern : Waldo (the implicit mission of this picture).
The invisible gorilla experiment is very funny, but it's a "gotcha" trick and it only surprises you once (with 50% success! no bragging rights there). Like when you point away to turn someone's head for an instant, and take advantage of their distraction to slap them in the face... It's a joke requiring the willing trust of the person.
- Hey, look away! ... Ah-ha, gotcha!
- But you told me to look away...
- Exactly.
That's basically what this "scientific" test comes down to. They test our attention, we kindly obey to the directives... only to learn the directives themselves were lies. This breaks the pupil-professor social contract. How cheap is it to catch someone off-guard through a deceptive premise? Anything goes in the realm of jokes, but as a scientific procedure it's pretty sloppy.
They don't ask you to spot a gorilla, they divert your attention to something else (count the passes of the team in white) and then blaming you for not noticing something else (which they implicitly ordered you not to pay attention to). That's what illusionists do : misdirection. Granted, the gorilla is a big detail to miss, but it was concealed carefully... Is it a coincidence if the gorilla is black (just like the team you are specifically asked to ignore), the same size as all the players, and moving across the screen at a comparable speed? Of course not. It's part of the strategy to avoid attracting your attention. The gorilla suit is only there to ridicule you, but it's merely the irruption of a random intruder within a scene we have no business keeping track of, except for the ball.
The center of gaze (corresponding to an area of the retina called the fovea), sees only a cone of vision of 2 degrees wide. This is what we use to decipher letters and read for instance. The rest of the field of vision (peripheral vision) is more blurry, unfocused and less acute. What this "awareness test" does is monopolize the attention of the fovea exclusively onto the white team players (and more precisely on the white players holding the ball), leaving the black team players in the peripheral vision. This is why, in the peripheral vision (less competent for subtle details, more competent in tracking motion), the gorilla can easily pass as a shape and behaviour resembling a black-team player without raising enough attention to the brains to require a fovea double-check (which would identify the difference in a fraction of a second). To make this trick work, the fovea must stay away from the gorilla at all time. Which is why you are asked to concentrate on a moving ball.
If a real gorilla was to pop in the scene, several details would have alerted our peripheral vision : abnormal posture and displacement of one of the "black-team player", animal sounds one would assume, different ground vibrations, and first of all, the panic of all players and the end of the passes. The delimited frame of the video also removes important aspects of our environment. The fact a gorilla moves in up-close, amongst a group of people without causing mayhem, keeps our senses quiet. Noticing and reacting to the panic in somebody's eyes, face or unusual gestures is part of our defense system. Even if these scientists like to make fun of our blind spots, we still have a pretty good instinct for survival, notably to spot danger in a familiar setting before it arrives, with all our senses, even if it is not under the direct scrutiny of the fovea. I guarantee you that if someone rushed in on screen, bouncing up and down, in a gorilla suit or not, even in the corner of the screen, 100% of spectators would be able to say they noticed something odd during the counting, even without looking at it.
- Hey, look away! ... Ah-ha, gotcha!
- But you told me to look away...
- Exactly.
That's basically what this "scientific" test comes down to. They test our attention, we kindly obey to the directives... only to learn the directives themselves were lies. This breaks the pupil-professor social contract. How cheap is it to catch someone off-guard through a deceptive premise? Anything goes in the realm of jokes, but as a scientific procedure it's pretty sloppy.
They don't ask you to spot a gorilla, they divert your attention to something else (count the passes of the team in white) and then blaming you for not noticing something else (which they implicitly ordered you not to pay attention to). That's what illusionists do : misdirection. Granted, the gorilla is a big detail to miss, but it was concealed carefully... Is it a coincidence if the gorilla is black (just like the team you are specifically asked to ignore), the same size as all the players, and moving across the screen at a comparable speed? Of course not. It's part of the strategy to avoid attracting your attention. The gorilla suit is only there to ridicule you, but it's merely the irruption of a random intruder within a scene we have no business keeping track of, except for the ball.
The center of gaze (corresponding to an area of the retina called the fovea), sees only a cone of vision of 2 degrees wide. This is what we use to decipher letters and read for instance. The rest of the field of vision (peripheral vision) is more blurry, unfocused and less acute. What this "awareness test" does is monopolize the attention of the fovea exclusively onto the white team players (and more precisely on the white players holding the ball), leaving the black team players in the peripheral vision. This is why, in the peripheral vision (less competent for subtle details, more competent in tracking motion), the gorilla can easily pass as a shape and behaviour resembling a black-team player without raising enough attention to the brains to require a fovea double-check (which would identify the difference in a fraction of a second). To make this trick work, the fovea must stay away from the gorilla at all time. Which is why you are asked to concentrate on a moving ball.
If a real gorilla was to pop in the scene, several details would have alerted our peripheral vision : abnormal posture and displacement of one of the "black-team player", animal sounds one would assume, different ground vibrations, and first of all, the panic of all players and the end of the passes. The delimited frame of the video also removes important aspects of our environment. The fact a gorilla moves in up-close, amongst a group of people without causing mayhem, keeps our senses quiet. Noticing and reacting to the panic in somebody's eyes, face or unusual gestures is part of our defense system. Even if these scientists like to make fun of our blind spots, we still have a pretty good instinct for survival, notably to spot danger in a familiar setting before it arrives, with all our senses, even if it is not under the direct scrutiny of the fovea. I guarantee you that if someone rushed in on screen, bouncing up and down, in a gorilla suit or not, even in the corner of the screen, 100% of spectators would be able to say they noticed something odd during the counting, even without looking at it.
In fact, Derren Brown makes use of this gorilla awareness test in one of his shows (Evening of Wonders, 2007).
Daniel Simons created another version (The Monkey business Illusion) destined to catch off guard the people who already heard about the gorilla. While viewers will be content spotting the gorilla, other subtle changes have been added to trick them. The fake curtain (added on green screen) changes from a red hue to an orange hue (without altering the general luminosity of the scene, which would otherwise alert our peripheral vision), and they use the entrance of the gorilla to exit one of the black-team player by the same side of the screen at the exact same moment (which is also the moment the curtain colour changes, for maximizing confusion and stealth), which easily fools the peripheral vision in thinking the black shape moving offscreen is the same as the lookalike black shape coming back on screen from the same spot.
In fact the gorilla didn't go unnoticed. We just didn't register its presence at the top level of our waking consciousness. If we are not told about the gorilla at the end of the test, we could easily retrieve this information under hypnosis. Under a state of trance (or focused concentration) we could revisit our memories of the video, even the details that we are not capable to access/verbalize consciously. This demonstrates that the center of gaze is not the only area that sees and remembers. (see Derren Brown's Trick or Treat Season 2 Episode 1; May 2008) We see more than what we are able to recall/spell out, because most details are relegated to low-level priorities and removed from verbalizable memory.
Again this is a trick only possible because of the offscreen suspension of disbelief. They use the off-screen space to sneak in/out, thus creating a false environment. Theatre uses the curtains to facilitate a live performance on stage, but if they started to abuse this convention to sneak in gorillas, spectators would give up disbelief and suspect everything going on, which would kill the necessary immersion in the narrative universe. The camera frame is likewise a convention we accept as long as it is not turned against us, to expose the blindspots created by our suspension of disbelief. For the benefit of the narrative plausibility, filmmakers don't need spectators to become suspicious of the non-diegetic space, behind the camera, or in the off-screen wings.
This is a valid experiment in the field of psychology, to test the compliance to a procedure, regardless for its irrelevance. I'm not sure it is as pertinent to the field of neurology, since the attention deficit is caused by the suppression of information in the fabric of the test itself. Seriously, it doesn't prove anything, and certainly not what it says it does. The two guys, Christopher Chabris and Daniel Simons, interviewed on a radio show (CJAD radio) are more ambivalent/measured about the conclusions, when given a chance to expand on it.
In fact the gorilla didn't go unnoticed. We just didn't register its presence at the top level of our waking consciousness. If we are not told about the gorilla at the end of the test, we could easily retrieve this information under hypnosis. Under a state of trance (or focused concentration) we could revisit our memories of the video, even the details that we are not capable to access/verbalize consciously. This demonstrates that the center of gaze is not the only area that sees and remembers. (see Derren Brown's Trick or Treat Season 2 Episode 1; May 2008) We see more than what we are able to recall/spell out, because most details are relegated to low-level priorities and removed from verbalizable memory.
Again this is a trick only possible because of the offscreen suspension of disbelief. They use the off-screen space to sneak in/out, thus creating a false environment. Theatre uses the curtains to facilitate a live performance on stage, but if they started to abuse this convention to sneak in gorillas, spectators would give up disbelief and suspect everything going on, which would kill the necessary immersion in the narrative universe. The camera frame is likewise a convention we accept as long as it is not turned against us, to expose the blindspots created by our suspension of disbelief. For the benefit of the narrative plausibility, filmmakers don't need spectators to become suspicious of the non-diegetic space, behind the camera, or in the off-screen wings.
This is a valid experiment in the field of psychology, to test the compliance to a procedure, regardless for its irrelevance. I'm not sure it is as pertinent to the field of neurology, since the attention deficit is caused by the suppression of information in the fabric of the test itself. Seriously, it doesn't prove anything, and certainly not what it says it does. The two guys, Christopher Chabris and Daniel Simons, interviewed on a radio show (CJAD radio) are more ambivalent/measured about the conclusions, when given a chance to expand on it.
On a short term prospect, the brains intentionally ignores certain elements (that are kept in check by the familiar permanence of the context) in order to become more efficient with a very specific task. But if there really was something fishy going on, the correlation of other signals (ear, feeling of vibrations, disorder of the environment...) would abort the silly task at hand (counting passes because we were asked to) and assess the general situation with all our senses, even for a second, until the suspicion is cleared up. At the cinema, we are dependent on what the filmmaker decides to reveal, visually on the screen, and aurally in the speakers.
That's why jump scenes are the easiest tricks in movies, and the cheapest (artistically speaking) effects. Like shouting "boo" in the back of an unsuspecting friend. Everyone falls for it. If we are in a safe environment, we never double check on everyone passing behind our back. We let ourselves betrayed by an excess of confidence with familiarity. In an unfamiliar place, we are on our guard, and would less likely let something fishy happen behind our back without double checking. The illusion of such tests is to pretend that the delimited diegetic world contained in a brief video sequence (to which we willingly grant the proverbial "suspension of disbelief") makes any claim on our real world perception. There are a lot of elements we take for granted in reality, that could fool us eventually, without checking them time and again, especially not more than once.
The Movie Perception Tests (Levin/Simons); Whodunnit? (WCRS&Co); Evolution of Games (MysteryGuitarMan)
Did you notice all the changes?
They play a brief clip from an unfamiliar scene where everything is new to us. We have to make sense of the context, while listening to a dialogue. And they ask us to pick up on gotcha changes that happen against the very logic of the scene and unbeknownst to us...
In a real world environment we know what elements are unknown and which ones are known, because we've tested them before, we also assume that things stay the way they are. We also have a good sense of which elements are most likely to change in certain circumstances. A sudden change from the corner of our eyes or an atypical noise from the ambiance might make us reassess the permanence of things and check things again as if new. But when the edit ellipsis of a shot-countershot convention abuses our trust to slip in prop changes, it is not something we would naturally pay attention to, nor would such trivial details (fork switch, arm positions...) prove to be interesting at all. The Whodunnit clip abuses the offscreen space left by the close-up panoramic, to alter props that the main character (a detective of all characters) can see in plain sight. At the movies, we suspend disbelief for the editing shorthands and other camerawork tricks, provided the ellipsis maintains the coherence of the whole. We also rely heavily on the knowledge of the onscreen characters (even if some of them may lie purposefuly for the benefit of the story). But if ALL characters play pretend changes that occur in the diegetic world didn't happen (however we do expect them to ignore the technical staff from the non-diegetic world of course), then the narrative has no foundation to build on. And that's this loophole (the willingness of spectators to give credit to stage artifices) that these gotcha test exploit, which is a low blow to the suspension of disbelief.
Applying this observation directly the movie narrative is rather abrupt and pointless. Comparing a chess master memory of checker conformations to reading the content of a movie is ludicrous. Of course, we let a lot of details pass by unchecked, especially when the talking and editing is accelerated. Just like in reality we could be fooled by something or someone we didn't go check out and assess firsthand, because details are too many around us, and because only a few of them will have any importance/influence in our life. When someone with an uniform on introduces herself as a nurse, we assume the hospital didn't let her enter and work there without qualifications. We don't verify her resume ourselves. Maybe we should. When someone walking in the park with a dog in leash says "this is my dog", we believe it without running a thorough administrative check up. Yet it could be another occasion to fool us. Another loophole of the "inattentional blindness" (an half-empty glass way of looking at the admirable evolution of one of our vital sense). Just like we assume that in a scientific video that unfolds very calmly, without scream or laughter everyone on screen was put there on purpose by the scientists, and we have no reason to suspect anything tricky in this department, nor the directives ordered onto us.
Source: The Invisible Gorilla (Chabris/Simons)
What You See Is What You Get (Manohla Dargis; NYT; 8 July 2011)
What You See Is What You Get (Manohla Dargis; NYT; 8 July 2011)
Good and good for you (David Bordwell; 10 July 2011)
To be continued...
Related:
Aucun commentaire:
Enregistrer un commentaire