Mixing Memory

One of the more sophisticated theories in embodied cognition is Lawrence Barsalou’s perceptual symbol systems theory (which I’ve talked about before, here). Starting from the premise that cognition is for action, it argues that concepts are represented in the same brain systems that directly serve perception and action. In the realm of memory, Arthur Glenberg put it this way1:

Suppose that memory and conceptualization work in the service of perception and action. In this case, conceptualization is the encoding of patterns of possible physical interaction with a three- dimensional world. These patterns are constrained by the structure of the environment, the structure of our bodies, and memory. Thus, how we perceive and conceive of the environment is determined by the types of bodies we have. Such a memory would not have associations. Instead, how concepts become related (and what it means to be related) is determined by how separate patterns of actions can be combined given the constraints of our bodies. (p. 1)

.

One position of this view is that both the encoding and retrieval of memories should be affected by our current bodily and emotional states. Early evidence for this position comes from a classic study by John Riskind, in which he found that participants showed better recall for emotional experience (positive or negative) if their facial experiences or body postures were consistent with the emotions of the experience2. For example, if they were told to smile, their recall of positive experiences was better. But this explanation could be written off as a simple case of mood congruency. Since recall is easier when your emotional state is similar to the one you were in when you stored the memory, and since facial expressions cue emotions, it could just be that people recalled happy memories better when they were smiling because they were happier when smiling than when their facial expressions were mood incongruent (if they were frowning, say).

To overcome this problem, Dijkstra, Kaschak, and Zwaan produced an even more radical demonstration of what we might call “body congruence.” They selected participants from two age groups (younger, 21.5 years old on average, and older, 69.7 years old on average), and asked them to retrieve six different kinds of typical autobiographical memories: going to the dentist, playing sports, opening a door for a visitor, clapping their hands at a concert, waving at someone, placing their hand on their heart. These were the target memories. They also asked them to recall something that happened the day before the experiment, and an “imaginary event.” While they were recalling the events, the participants were instructed to position themselves in a posture that was either consistent or inconsistent with the event in the memory. For example, in the consistent condition, participants recalling a visit to the dentist would lie in a recliner, while in the inconsistent condition, they would stand with their hands on their hips. Each participant recalled three of the target memories in a consistent posture, and three in an inconsistent posture. The prediction, of course, is that recall would be better in consistent postures than in inconsistent ones.

As a first test of the prediction, Dijkstra et al. measured the time it took the participants to recall each type of memory. Here are the reaction time results in seconds (taken from their Table 1, p. 144, colors chosen for the Oregon Ducks fans among us):

i-e3855a23b3f74fcbbcab386c2e89fb86-Dijkstraetalgraph.JPG

As you can see, recall was faster for younger adults than for older ones. No surprise there. For both younger and older adults, recall was faster when body posture was consistent than when it was inconsistent. The effect was stronger for the younger participants, and Dijkstra et al. explain this by referring to answers to one of the questions they asked participants after each event was recalled: “Are you reexperiencing the event?” The younger participants reported reexperiencing the event more than the older participants. Perhaps this “reexperiencing” facilitates bodily congruence effects.

Two weeks after the first session, the participants called the experimenter, and were then asked to recall the autobiographical memories they’d reported in the first session. Participants were better able to remember the autobiographical events they’d previously recalled if, during the first session, they’d recalled them in a consistent posture. Here’s the percentage recall data (taken from their Figure 2, p. 145; don’t ask me why I picked these colors):

i-f3ca55414a7fe950f439d19b7ac8dcfa-Dijkstraetalgraph2.JPG

As you can see, the effect was again stronger for younger participants. Again, this may have to do with “reexperiencing.”

So, consistent with perceptual symbol systems theory and related theories of embodied cognition, remembering an experience while one is in a bodily position that is consistent with the experience facilitates recall. As I’ve said before when talking about perceptual symbol systems theory, it inspires really cool experiments, and this is certainly one. The argument that Barsalou and others have made is that the results of these experience can only be predicted post hoc, by theories of cognition that don’t place an emphasis on embodiment. Unfortunately for proponents of embodied cognition, this obviously is not the case with this experiment. Despite Dijkstra et al.’s claim that their experiment provides support for embodied theories, “amodal theories” (theories that claim concepts are represented outside of particular sensorimotor systems) have been predicting results like these for a few decades (e.g., Riskind’s experiment). It seems to me that the “body congruence” effect is simply another case of encoding-specificity (which Dijkstra et al. actually mention in the introduction to their paper). If you’ve taken a course in cognitive psychology, you may remember that memories are better recalled in contexts similar to those in which they were stored. Since posture is an important part of the context in which certain events occur (like, say, a visit to the dentist), it’s not surprising that posture facilitates recall for such events. Of course, just because this experiment doesn’t actually differentiate between amodal and embodied theories doesn’t mean it’s not really cool anyway.

1Glenberg, A.M. (1997). What memory is for. Behavioral and Brain Sciences, 20(1): 1-55.
2Riskind, J. H. (1983). Nonverbal expressions and the accessibility of life experience memories: A congruence hypothesis. Social Cognition, 2, 62-86.
3Dijkstra, K., Kaschak, M.P., & Zwaan, R.A. (2007). Body posture facilitates retrieval of autobiographical memories. Cognition, 102(1), 139-149.

Comments

  1. #1 Ryan Fox
    December 19, 2006

    Hi Chris,

    What do these graphs represent? The numbers aren’t given any meaning. I’m guessing that the first graph is recall time, but I have no clue as to what the other one is.

    By the way, you probably went with orange and blue because you picked one of the colours, and then picked its complement unconsciously.

  2. #2 andrew
    December 19, 2006

    i’ll bet the second graph shows proportion of memories recalled.

  3. #3 Chris
    December 19, 2006

    Hey guys, thanks for pointing out the inadequacy of the graphs. That’s what I get for making graphs quickly, based on someone else’s tables. I added text in the post to explain what they are.

  4. #4 Fido the Yak
    December 20, 2006

    What are the possibilities here? You posit that the body is a context in which things are remembered, but in Glenberg’s view the body is a separate contingency from the environment. Can this be sorted out experimentally? What about the view that the body is itself a mnemonic instrument, or a mnemonic agency? Is there any observable relationship between the body and memory that could not be ascribed to environmental context?

  5. #5 Chris
    December 20, 2006

    Fido, “can that be sported out experimentally?” That’s exactly the problem! It can’t. I mean, obviously the body has a privileged place, to some extent, in that we have senses that just provide information about it, and its always a part of our “context,” but it’s hard to sort out the “modal” effects from mere context effects, or in this case, encoding specificity.

The site is currently under maintenance and will be back shortly. New comments have been disabled during this time, please check back soon.