Freyd, J. J., Pantzer, T. M., & Cheng, J. L. (1988). Representing Statics as Forces in Equilibrium. Journal Of Experimental Psychology: General, 117, 395–407. doi:dx.doi.org/10.1037/0096-3445.117.4.395

Summary

In this paper, Freyd et al. hypothesize that the mind represents dynamic forces, even for static scenes. In particular, they point out that in static scenes, it is not truly the case that there are no forces: it is just that forces are in equilibrium. They run a series of four experiments to test this hypothesis, finding that people’s memory for static objects is distorted in the direction that those objects would move if they were unsupported.

Methods

In Experiment 1, Freyd et al. showed participants a scene with a plant either on a table or hanging from a hook for 250ms. The scene was removed for 250ms at which point an identical scene was showed for 250ms, except that the table or hook was removed. The scene was again removed for 250ms, and then a new scene was shown which was either (1) the same, (2) different, with the plant moved slightly upwards, or (3) different, with the plant moved slightly downwards. Participants had to determine whether the last scene was the same as the original one or not. They found that people were quite accurate at determining “same” trials (8-13% error) and worse at detecting “down” trials (57% error) than “up” trials (37% error).

Experiment 2 was a control in which Freyd et al. never showed any form of support. The results showed basically no difference between the “down” trials (33%) vs. the “up” trials (34%). Interestingly, though, the error rate for the “same” trials was higher than in Experiment 1 (26%).

Experiment 3 was similar to Experiments 1 and 2, except that they used a different display (a lock hanging from a hook) and tested different distances from the true position. This was done in order to get a more fine-grained estimate of the distortion in people’s memories. As in Experiments 1 and 2, they found distortions when there was support, and that the shift was positive (i.e. “down” trials had more error) when there was support and that it was close to zero when there was no support.

In Experiment 4, Freyd et al. tested another new display, this time with a block resting on a spring. This allowed them to test for distortions in memory both in the upward and downward directions. There were two conditions: one in which there was initially no block (which would predict a downward shift, as the spring should compress), and one in which there was initially a block (which would predict an upward shift, as the spring should decompress). The results supported the predicted direction of memory distortions.

Algorithm

n/a

Takeaways

These are a cool set of results, related to representational momentum, that indicate that people have strong perceptual expectations about the way that objects should move. In particular, I think the results of Experiment 4 are really striking: it’s not just that people expect things to move down due to gravity, but that they expect them to move in the way they actually would. I expect you would find similar results with objects that typically move up (e.g. balloons). I also wonder if you would find these effects if they depended on higher-level knowledge about object properties (such as mass). For example, in the spring experiment, if you knew the block was extremely light (and thus would not compress the spring), would the memory distortion still take place?