I have been going on for a bit about how our psyches are constructed things. But I’ve been doing so in kind of general terms. So, let’s take a breather and see if we can’t assemble a simple belief engine from some of the pieces that have shown up on the blog so far. Or if engineering isn’t your thing, you can think of this as a sort of recipe.
Ingredient List
1 c. Visual perception of faces
1 c. The other race effect
1 Tbs. Generalization
Stir with a dash of cultural experience
Faces. Faces are a construct of our visual system, and a lot could be said out them. For example, here is an image from a well-known experiment seeking to determine if infants are born with an innate ability to recognize face-like images
The basic idea is to take a face and show it to infants of different ages. Of course, you then need other images so that you can make comparisons. Like, maybe make sure that the infant isn’t simply responding to the nose, mouth, etc., no matter how they are arranged (Linear). Or maybe check out if it’s a triangular arrangement of stimuli that the infant “likes” and compare that a more normal face, and the same image with the triangular arrangement flipped. This is some of what psychologists do!
Whether infants are innately predisposed to respond to face-like images, over time, most of us become better at distinguishing between faces. When our eyes encounter a face-like stimulus they tend to bounce around in a manner that picks out particular features.
Orginal work done by Alfred Yarbus (1967) and image found here.
Like any skill, this is something at which we become better and more efficient with repeated practice. We don’t just pick out faces as generic visual objects, but individual faces for our friends, parents, siblings and so on. And yes, we know that all of this construction is happening in a brain, and since that is the case, damage to the brain can interfere with our ability to construct faces. Prosopagnosia is a syndrome that refers to an inability to recognize faces brought on by strokes / damage. Individuals experiencing prosopagnosia might be able to perfectly describe the features of a face: the nose, the color of they eyes, the shape of a chin – but they cannot “see” the face.
The Other Race Effect. As we encounter faces, we become more and more adept at pulling out features that allow us to efficiently “recognize” the individual attached to any particular arrangement of these features. The general thinking is that this process results in the creation of a prototypical face. You can think of this as being the weighted average of all of the face-features that an individual has encountered in their day-to-day living. That prototype is then going to become your psyche’s starting point for recognizing new faces. Just like a few posts back, when we talked about how our visual system has beliefs about depth, and size, and relative brightness, our visual system constructs beliefs about faces. One of these beliefs goes by the name the other race effect.
Basically, the other race effect refers to the fact that your psyche is better at recognizing prototypical faces. If you’ve ever tried to use money in another country, you’ll have an appropriate analogy. Even after living in Germany for four year, I would sometimes get hung up on counting out change. I had trouble remembering which coins went with which amounts, and strangely this made it more difficult to simply add up and subtract amounts. The other race effect is something like that. It refers to the fact that individuals who regularly experience faces within a particular ethnic / racial category find it harder to distinguish between faces outside of that category.
A study from 2007 published in Psychological Science provides a nice conceptual demonstration of the other race effect.
In this particular study, the researchers asked about the development of the effect. In other words, they asked whether very young children showed the effect, and if not, when did it become pronounced. Their subject pool consisted of Caucasian infants from three age groups: 3-month, 6-month, and 9-month. In the experiment, these infants were presented with faces from four “ethnic categories”: Caucasian, African, Middle-Eastern, and Chinese. Please note that these are the designations from the article. All of the faces were taken from students who were 23 – 27 years old.
The way that these experiments typically work, is that an infant will be shown a stimulus until they become bored. “Bored” is measured by the amount of time that the infant spends looking at a stimulus vs. looking elsewhere (remember, in science we have to measure something!). This “boredom” is referred to as habituation. Once an infant is habituated to a stimulus, an experimenter can show that infant another stimulus and ask: “Does the infant remain bored?” If so, then that means that the infant doesn’t recognize a difference between the two stimuli. On the other hand, if the infant’s attention perks up, then we can conclude that the infant DOES recognize a difference between the stimuli.
That is the logic that was used to examine the development of the other race effect in the Daley et al. study. Infants were be shown a particular face until they habituated to it. Then there were shown the same face along with another novel face. That novel face either came from the same or another ethnic category. The question was whether the infants would recognize the difference of novel faces within an ethnic category. For example, if the infants were habituated to a Chinese face, would they perk up when presented with a new Chinese face?
Here is their data. I’ve changed it from what was provided in the actual paper, because the data there consisted of numbers in a table. I took those numbers and created a graph that makes the same points. Any bars above the dotted line indicate that the infants “recognized” the novel faces as novel. 3-mo old infants essentially treated all faces as unique. 6-mo olds, though, treated Caucasian and Chinese faces as unique, while tending to clump African and Middle-Eastern faces into generic categories. Finally, at 9 months, the infants showed the full on “Other Race Effect.” They treated faces from their own ethnic group, Caucasian, as unique, but tended to treat the faces from the other three ethnic categories as generic.
Generalization. Ok, so we know that faces are visually constructed things. We also know that we tend to learn to distinguish between commonly encountered faces, and that is going to tend to produce better recognition of faces within our ethnic group as opposed to outside of ethnic group. All of this is simple engineering. Let’s start to push it in a direction that we’d connect with more traditional notions of belief. We’ll do this by reminding ourselves of a psychological phenomenon known as generalization.
Like so many things in psychology, generalization is conceptually simple, but incredibly complex in its details. The basic phenomenon is this. If you learn something about a situation, you will tend to transfer that learning to “similar” situations. This transfer to “similar” situations, contexts and stimuli is termed generalization, and it is seen in so many animals that it is considered a basic feature of learning. Here is an example of generalization in pigeons. Pigeons that have learned that a vertical line either indicates food or the absence of food, will tend to generalize that learning to other line orientations.
However, notice that I put “similar” in quotes. Usually we think of “similar” as sharing features. If you are humiliated when you recite a poem in 4th grade, then maybe you will feel emotions of humiliation whenever you enter any classroom. Classrooms, after all, are “similar.” However, when we use the term “similar” what we really mean is that your brain is not recognizing a situation as completely unique. Someone else’s brain might very well. Also, this is not to say that at some point in the future you won’t recognize difference. You might very well learn that situation A is quite different from situation B.
So “similar” is a subjective thing. When we transfer learning, feelings, and assumptions to new situations, this says more about what we as individuals recognize as “similar” at that particular moment. As an example, when we learn a new language, one of the humps we have to overcome is hearing sounds as unique that we start off hearing as generic. In any event, generalization is the phenomenon of transferring learning to “similar” situations and stimuli. But remember, “similar” means, to some degree, “not recognizing difference.”
A Dash of Cultural Experience. So here is where we mix all of our ingredients together to construct a larger belief. First, we’ve pointed out that our visual system uses features in the environment to construct faces. Whether some of this construction is innate, or not, is beside the point. Faces are constructed things. Second, we’ve pointed out that how this happens creates categories of faces that we recognize as unique, and categories of faces that we treat as generic. Thirdly, we’ve reminded ourselves that if stimuli are generic, then this means that they are “similar,” and learning will tend to generalize across similar stimuli. What will happen, then, if culture pulls out a single face from an ethnic group and portrays it in a particular way? For example, let’s suppose that we are Caucasian child from the Kelly et al study. We see a show that portrays a Caucasian as a criminal. Well, since we tend to treat Caucasian faces as unique, then we won’t generalize criminality to a broad range of Caucasian faces. On the other hand, if we see a show that portrays an African, Middle-Eastern, or Chinese individual as a criminal, we will be more likely to generalize “criminality” to a broad range of African, Middle-Eastern, or Chinese individuals. That’s the thought experiment for this blog post.