Mixing Memory

Psychologists are now fairly confident that humans have an agency-detection mechanism, and an itchy agency-detection trigger finger. This is because the consequences of not detecting agency are often greater than those of agency-detection false positives. As Atran and Norenzayan put it:

From an evolutionary perspective, it’s better to be safe than sorry regarding the detection of agency under conditions of uncertainty.

When it looks like there’s a shape moving in the tall grass, it’s better to infer agency and run away than end up being eaten by a lion when it jumps out of the grass. Thus, when we’re not sure whether agency is present, we’re likely to infer that it is, regardless of what the object in question might be. This means that we’ll assign agency to just about anything, including dots moving on a screen2.

Subsequent studies used the same paradigm as in study 2, including using casual sex as the goal, and found that the goal persisted over time (e.g., when there was a delay between the activation of the goal and the helping task), and that participants were less likely to adopt the goal when they read about a person

Part of what it means to assign agency to an object is that we attribute intentions, or goals, to that object. Thus our hard-wired agency-detection mechanism is also a goal-inference mechanism, and we’re willing to assign goals to just about anything from a young age. As an example, take a look at this figure depicting the stimuli in an experiment by Fulvia Castelli3:

i-b9091efe9a487ffef0013619622e03fe-Goal1.JPG

In one condition of Castella’s experiment, participants (children with high-functioning Asperger’s syndrome, non-autistic children, and adults) watched the circle at the bottom of the “valley” move up the right slope three times, each time getting a little higher, and then finally coming to rest either at the bottom of the valley or midway up the right slope. Most of the children and the adults (including the autistic children) indicated that the circle’s goal was to get to the dark ball at the top of the right slope, even though the circle never actually got there. All it took was repeated “efforts,” getting closer each time, for children to infer that the circle had a goal. This result isn’t limited to individual objects, either. If groups of dots on the screen appear to be moving together, we’re perfectly willing to assign goals to a group of dots too. Silly humans.

Socially, inferring the goals that underlie people’s behaviors is very important (would you rather know only that someone’s being nice to you, or that they’re being nice to you because they want something from you?), so it’s not surprising that we’re quick to infer the goals of other humans as well as of dots. In fact, when describing people’s behavior, we’re more likely to do so in terms of the goals behind it than just about anything else4, and this goal-centered conceptualization of others’ actions occurs automatically and often unconsciously5.

In a recent paper, Aarts et al.6 hypothesized that the automatic conceptualization of goals may go further than merely inferring a goal’s presence. We may, in fact, adopt the goals we perceive. They wrote:

[W]e suggest that automatic processes that begin with the perception of another’s behavior may lead to what we call goal contagion: the automatic adoption and pursuit of goals that others are perceived to strive for. (p. 24)

In order to test this hypothesis, they conducted several experiments in which participants were exposed to behaviors that implied particular goals, and were then placed in a situation in which they themselves adopt the same goal. In the first study, participants first read a story in which the main character planned a vacation with friends. In one condition, the character went to work on a farm before going on vacation which was designed to imply the goal of making money for his vacation. In another condition (the control condition), he did volunteer work prior to going on vacation. After reading the story, participants were then told that they were going to participate in a second task. They were also told that if they completed this task in time, they would be able to complete in a third task in which they could earn money. The experimenters then recorded how quickly participants completed the second task, in order to determine whether they were motivated to make money. Finally, participants answered demographic questions designed to measure their financial status. Participants with lower financial status should have a higher “need for money,” and thus the “goal contagion” effect should be stronger for them.

If the goal contagion hypothesis is correct, participants who read the story in which the character worked on the farm should finish the second study faster than those in the control condition, and this effect should be strongest for high “need for money” participants (i.e., participants with lower financial status). This is what they found. Though overall, there wasn’t a statistically significant difference between the farm-working and control conditions, the difference was in the predicted direction. Furthermore, when they looked at high need for money participants alone, the difference was significant: they performed the second task much faster after reading the story about the person working on the farm than after reading the control story.

The second study was designed to test for the effect using a different goal: casual sex. In this study, all of the participants were heterosexual males. They first completed a simple task involving clicking a mouse, ostensibly to measure “computer skills.” Next, they read one of two stories. The first, the casual sex goal story, was about a guy named Bas (the study was conducted in Holland), who runs into an old friend at a bar and, well, wants to have sex with her. Here’s the story (from the appendix):

Bas is meeting a former college friend called Natasha while having a beer in his favorite pub. They are having a chat, and Bas tells her about his new job. The atmosphere in the pub is great, and a lot of people have been showing up. At the end of the evening Bas walks Natasha home. When they arrive at her home, he asks her, “May I come in?”

In the control condition, participants read a story about Bas meeting Natasha, chatting for a bit, and then watching people dance. Finally, participants were placed in a position in which they could help either a female or a male. Pilot studies showed that heterosexual Dutch men saw helping a woman as “a viable route to attaining the goal of having casual sex” (I know, I know, I didn’t design the study, I’m just telling you about it). Therefore, if the participants are motivated to attain casual sex, they should do more to help the female. In the helping task, they were told that either a male or female had designed the first “computer skills” task, and were asked to give that person feedback about the task. The participants’ motivation to help was then measured by looking at how long they took giving feedback, and how many words they wrote, the idea being that the more time they spent, and the more they wrote, the more motivated they were to help.

The prediction from the goal contagion hypothesis, in this case, is that participants who read the story about Bas attempting to attain casual sex from Natasha would be more motivated to attain casual sex themselves, and would therefore help the female more by writing a bunch and taking a long time to do it. The graph below presents the z-scores for both measures combined (taken from Table 2, p. 28):

i-62e82a7e4e2b8339686b1f352bf1e84c-casualsex.JPG

In case you’re having trouble interpreting that, this is what it shows. When the target was female, participants spent significantly more time and wrote more words after reading the story about Bas seeking casual sex than after reading the control story. When the target was male, the pattern was reversed: they actually helped the guy less after reading the casual sex story (I call this the “I’m not gay!” effect, or the homophobic effect).

Subsequent studies used the same paradigm as that used in study 2, including the use of casual sex as the contagious goal, and found that the when people adopted goals they perceived (in the actors in the stories), the resulting motivation persisted over time (e.g., when there was a delay between the activation of the goal and the helping task), and that participants were less likely to be motivated to seek casual sex when the story they read involved Bas attempting to have casual sex with Natasha under conditions that were inappropriate. In case you’re wondering what those conditions might be, here’s the story for that condition (from the appendix):

Bas is meeting a former college friend called Natasha while having a beer in his favorite pub. They are having a chat, and Bas tells her about the upcoming birth of his child. The atmosphere in the pub is great, and a lot of people have been showing up. At the end of the evening Bas walks Natasha home. When they arrive at her home, he asks her, “May I come in?”

It appears, then, that goals really are contagious, at least under the right conditions (e.g., if we perceive the goal in another under appropriate circumstances for acting on that goal). Notice that in both the money and sex studies, the goal-related behaviors that the participants performed were very different from the behaviors they read about, indicating that goal contagion is not simply a case of imitation. It is also interesting to note that during the debriefing sessions after each study, participants reported being unaware of the affect that the stories had on their own goals, indicating that the contagious goals were largely unconscious.

In their conclusion, the authors note that the “goal contagion” effect may, among other things, help to coordinate social behaviors. This may be generally be beneficial for the individual, but it might not always be a good thing overall. I suspect that, if the goal contagion effect is confirmed by further research, it may be used to explain all sorts of unseemly group behaviors. The effect immediately calls to mind things like negative behavior towards outgroups, for example, and makes me think of all of the social psychology research done in the 1950s in an attempt to explain the rise of Nazism, and the inhuman behaviors that produced. It might also help to explain many religious behaviors, as a large part of organized religion consists in the adoption of mutual behavioral goals. For now, of course, more research needs to be done to understand the effects that Aarts et al.’s studies produced, but it definitely seems like they’ve discovered something very interesting.


1Atran, S., & Norenzayan, A. (2004). Religion’s evolutionary landscape: Counterintuition, commitment, compassion, and communion. Behavioral and Brain Sciences, 27(6), 713-730.
2Heider, F., & Simmel, S. (1944). An experimental study of apparent behavior.
American Journal of Psychology, 57, 243-59.
3Castelli, F. (In Press). The Valley task: Understanding intention from goal-directed motion in typical development and autism. British Journal of Developmental Psychology.
4McClure, J. L. (2002). Goal-based explanations of actions and outcomes. European Review of Social Psychology, 12, 201-236.
5Hassin, R.R., Aarts, H., & Ferguson, M.J. (2005). Automatic goal inferences. Journal of Experimental Social Psychology, 41(2), 129-140.
6Aarts, H., Gollwitzer, P.M, & Hassin, R.R. (2004). Goal contagion: Perceiving is for pursuing. Journal of Personality and Social Psychology, 87(1), 23-37.

Comments

  1. #1 Sabina
    October 26, 2006

    Hi,
    I am wondering if part of what they were seeing in this study has to do with the Mirror Neurons? Has anybody done research on what different kinds of stimuli the mirror neurons ‘copy’? I know it works for visual, I am pretty sure they have ascertained they can also transmit or pick up on facial expressions, thus deducing emotions. Since the written word is encoded and processed, leading to mental imagery, I wonder if those mirror neurons are not heavily involved there too. Hence the popularity of romance novels and other more ‘explicit’ literature that can make people ‘visualize’ events? Just musing about this, while I should probably be doing some real work here…

    S.

  2. #2 Chris
    October 26, 2006

    Hi Sabina, you know, I have to admit that the idea of mirror neurons hadn’t really crossed my mind. I suppose I have a bit of an aversion to them. It may very well be that mirror neurons play a role in this, by say simulating the behavior of another individual. The behavior might then be associated with certain goals, and using context we can narrow down the appropriate one in a given instance. But that’s a pretty big story with very little empirical evidence.

  3. #3 Sabina
    October 26, 2006

    Thanks for your answer Chris.
    Maybe I ‘over-conceptualized’ it, but here was my thought:
    The latest research seems to show that the mirror neurons allow a person to view an action, and – if that person has performed that action in the past – to not only replicate it, but to also anticipate the outcome (Gallese, V. 2004). For that to occur the mirror neurons would have to create a representational concept of the action and the outcome in the brain. My point is that by reading about an action that implies a certain goal, a representation of that action is formed in the brain – including the projected outcome and all the emotions connected to it. This representation is now part of the neural mapping of the brain, part of the available memory (for however long), and if the outcome is a desired one, presumably the person will strive to replicate the action that produced the projected representation. I don’t think that there is any way to empirically test my theory at this point in time. Which is why it is just a theory, lol! :)

  4. #4 Sabina
    October 26, 2006

    Oh no! I just read your post from July about Mirror Neurons – sorry I brought it up, lol! :)