As I said yesterday, I love research that challenges the common sense view that perception, especially visual perception, represents the world as it "is." The paper I talked about there showed cognitive influences (memory) on relatively low-level visual processes (color perception occurs pretty early). Minutes (like 2!) after I finished writing that post, ScienceDirect sent me an email alert with the contents of the October issue of the Journal of Personality and Social Psychology, in which there is a paper by Emily Balcetis and David Dunning1 describing research that demonstrates an influence of motivation on fairly high-level visual perception -- object recognition.
I've talked a lot about motivated reasoning in the past (here, here, and here), but this paper presents what you might call motivated vision. In motivated reasoning (sometimes called hot cognition, because it isn't coldly rational), our judgments are biased by self-serving motivations. Thus we are biased towards conclusions that confirm our existing beliefs, or that enhance our self-esteem. Motivated reasoning is likely the norm -- unless we're wholly un invested in what it is we're reasoning about, motivational biases are likely to play a role in determining what conclusions we arrive at, be they conclusions that confirm pre-existing beliefs, that enhance our self-esteem, or whathaveyou. Balcetis and Dunning, in a set of five studies, set out to see just how far the influence of motivation can go. Can it actually affect the way we see?
The idea that motivation might influence perception is not new. If you've taken an introductory psychology course, you might remember hearing about a study from the 1940s by Bruner and Goodman2 in which they had "rich children" and "poor children" (their categories, not mine; see the figure below) estimate the sizes of coins. They found that all children overestimated the sizes of the coins (kids like money!), but that poor kids' overestimations were significantly greater than those rich kids. Here's a graph of the data from the paper (which you can read here):
At first, this might seem like a case of motivation (poor kids wanting money) affecting perception, but there was a problem with this study, and the line of research that it spawned. Since the kids were estimating the size of coins without real coins sitting in front of them, what motivation was affecting was their memory of the sizes of coins, and not necessarily their perception of coins.
So the door's open for Balcetis and Dunning. In their first study, they had participants sit down in front of a computer screen, and told them that the computer would randomly select which of two liquids they would drink for an experiment on taste. The first was orange juice, and the second was, well, I'll let Balcetis and Dunning describe it for you:
[A] gelatinous, chunky, green, foul-smelling, somewhat viscous concoction labeled as an organic veggie smoothie.
Naturally, it was expected that participants would want to drink the orange juice, and not the green goo. Participants were told that the computer would indicate which liquid they were to drink by giving them either a random letter or a random number. For half of the participants, a letter meant orange juice, and for the other half, a number meant orange juice. After spending three minutes describing what they expected the two liquids to taste like (making them all the more eager to get the orange juice, and not the goo), a figure that can either be interpreted as a 13 or a B flashed on the screen.
After the figure had been on the screen for 400 ms, the computer seemed to crash. The experimenter came over, and asked if anything had been on the screen before it had crashed. Participants then reported whether they'd seen a B or a 13. The prediction, of course, is that participants will want to see either a number or letter, depending on which one means they get to drink orange juice, and therefore when they're presented with the ambiguous figure, they'll be more likely to interpret it as a B or a 13 depending on which means orange juice. Consistent with this prediction, 72% of the participants said they saw a B (none said they saw a 13, and the rest said they didn't see anything -- I bet they saw a 13) when a letter got them orange juice, and 60.5 reported seeing a 13 when a number got them orange juice.
The second study was similar to the first. This time, participants could get a neutral-tasting liquid (bottled water), candy, or "a bag of gelatinous and partially liquefied canned beans." This time, which of the substances they got to taste was determined by their score in a game. The game presented participants with pictures of animals ("farm animals" and "sea creatures"). One type of animal got them positive points, and the other type got them negative points. They were told that a positive score got them candy, a negative score got them the liquefied beans, and zero got them water. As participants played the game, they could watch their score, and as they reached the end of the game, their score became more and more negative. Finally, with only one animal to go, their score was so negative that only one animal (either a horse, if farm animals gave positive scores, or a seal, if sea creatures did) could give them a positive score. The final image of an animal was this one:
As you can probably see, this can either be a horse's head or a seal. Participants were then asked what they'd seen on the screen when the last image was presented. As in the first study, participants tended to see the figure that got them the candy, and avoided the beans. In this case, 66.7% saw the horse when farm animals gave positive points, and 72.7 saw a seal when sea creatures did.
In a third and fourth study, this result was confirmed by showing that participants' first eye movements (movements that are likely not consciously controlled) after being presented with the ambiguous figure was to the category (farm animal or sea creature) that meant positive points, and that reaction times were consistent with seeing only one animal (i.e., they hadn't seen both, and chosen the "good" one). The fifth study also showed that they'd seen only one animal, because they were told that due to a computer error, they would get to drink orange juice (instead of green goo) if an animal from the category opposite to what they'd been told previously had been shown last. So, if they'd originally been told that farm animals yielded positive scores, they were now told that if the last image shown was of a farm animal, they would drink goo. If participants saw both the horse and the seal, then they could report seeing the one that would now yield the desired outcome, but instead they tended to report seeing the animal that, prior to the game, they had been told would yield the desired outcome.
These studies provide pretty strong evidence, then, that participants' object perception was affected by their motivation to eat or drink something that tastes good, or at least by their motivation not eat or something that looked like it would taste really bad. How does this work? Balcetis and Dunning suggest that motivation biased the participants' "perceptual sets." In other words, participants want to see something, so they're expecting to see it, and when presented with an ambiguous figure, this means they're more likely to see it. However, there are other potential explanations, and hopefully further research will attempt to test these. For now, though, I'm just going to revel in one more study showing that the world we see is largely a creation of our own -- in this case, of our own desires.
1Balcetis, E., & Dunning, D. (2006). See what you want to see: MOtivational influences on visual perception. Journal of Personality and Social Psychology, 91(4), 612-625.
2Bruner, J. S. & Goodman, C. C. (1947). Value and need as organizing factors in
perception. Journal of Abnormal Social Psychology, 42, 33-44.
But why would you revel in something that is obviously untrue? I see the police sirens in my rearview mirror even though I don't want to, and fail to see an extra zero on my paycheck although I do want to, and, obviously, I would get along disastrously in life if that weren't the case. Why should you consider results like these to be anything more than a cute trick? (Unlike phenomena like the Kanisza triangle illusion, which illustrate how the visual system has evolved cheats in order to approximately solve the impossible task of mapping from light reflectance back to its true sources- and which has nothing to do our personal desires or expectations.)
Other Chris, I think you might be missing the point. Your examples aren't really the same sort of thing as what they describe in the paper. Their claim is that motivation can influence perception under the right conditions. In the experiments, there was a high degree of uncertainty (ambiguous figures), for example. One could probably induce a similar result by presenting images subliminally, at the periphery of the visual field, or for a very, very short period of time. Because the system is looking for something, it will be biased to see that in anything that looks familiar. To be honest, that's not really a new or surprising finding, except that in this case, motivation was responsible for the expectation.
Chris, what struck me wrong was not the paper per se, which seems neat but of unclear significance, but your statement that you "revel in one more study showing that the world we see is largely a creation of our own -- in this case, of our own desires." Call me a naive realist, but I don't see how any of the evidence you mention supports the notion that the world we see is largely a creation of our own desires, beliefs, etc; at best this paper might support the notion that our perceptual world is shaded at the margins by our desires. I guess from there it would be interesting to know the mechanism and purpose of such a phenomenon, and whether it is always operative or only in particular circumstances, like food selection or social interaction (assuming people haven't already looked at these, and better, questions).
Other Chris, I didn't mean to imply that this study, alone, demonstrates that, but that it's another piece of evidence in a long line of such evidence.
I think the clearest way to say what you are getting at is "perception is flawed". That is, a perceptive event, even in a rational being, may be quite skewed. The B/13 is, of course, a wonderful example. it if, really, BOTH. A B can have vertical separators, and a 13 can be drawn very close. Only a very few people will be able to hold the B/13 dichotomy, especially after seeing it for only 4/10th of a second.
Optical illusions in general are interesting - one can make an optical illusion playing to any of the visual subsystems. Similar illusions can be played on the other senses, although visual is the least dynamic, and thus most easily able to be demonstrated. i've heard illusions of "inserted notes" - that is, a peice of music where a repeating note is left out along the way, but the litener fills it in from the pattern, for instance. And most people are familiar with the water/cold illusion.
Now, what does fallacious perception mean to the person who is not ideatic and has not been trained to be hyper-alert (i.e., 99.9% of the world)? It means that you should not trust someone who says something was what they saw, especially if them seeing it would have a positive effect on their life. It is likely a misperception at some level. One reason that phtographs made newspapers much more believable was that the photograph was not influenced by the perceptive - and language- tree of the individual.
If a farm animal would save you from the green goo, you might be thinking "please be a farm animal, please be a farm animal" and thus prime recognition of a farm animal.
But what if the task was framed as "anything BUT a farm animal will save you from the green goo"? then you might be thinking to yourself "please DONT be a farm animal, please DONT be a farm animal" and still be priming recognition of a farm animal.
Opposite motivation (wanting a farm animal vs not wanting a farm animal) but the same perceptual effect.
What do you think?