Using kinematics to perceive expectations and deception

Experiment 1 of Runeson & Frykholm (1983) established that an invisible source of influence could be perceived even when only the movement kinematics of an actor was present. Whether this perceived invisible source referred to the intention of or outcome achieved by the actor remains an empirical question. The subsequent experiments in this paper explicitly studied whether it was possible to perceive internal sources of influence via kinematics.

Experiment 2: Perception of weight expectations

Postural adjustments don't come only after an object is being lifted. Instead, they take effect even before one has come into contact with the object. These postural preadjustments typically depend on the actor's expectations of the object's weight. We see this everywhere in our daily lives. For instance, when we know an object is heavy, we tend to take precautionary measures prior to lifting so as to put ourselves in a biomechanically safe and advantageous position to effectively lift it. Note the emphasis on postural preadjustments that take place before the object is lifted. This was essentially what the second experiment explored.

In experiment 2, actors were filmed walking up to an object before lifting it. Here, they were made aware of the weight to be lifted before lifting, and hence could be assumed to be approaching each lift with a certain expectation of object weight. Like in previous experiments, observers only saw patch-light displays of the actors (made possible using retroreflective tape). The key difference here was that the participants only viewed these patch-light videos up to the point just before the objects were being lifted. That is, the observers only had access to the actors' kinematics prior to the actual lifting motion. After each clip, the observers were then asked to report what they thought was the actual weight of the object. 

Based on the above analysis, actors' kinematics before lifting should be informative of their expectations of the object's weight. Hence, if observers are sensitive to this information, their judgments of the object's actual weight should correspond to the weight expected by the actor. And this was exactly what was found! In fact, three of the observers managed to perceive all the expected weights correctly, with an additional two observers only making one error each. These results suggest that the KSD principle applies to dynamic influences like expectations -- there is visual information in the ambient optic array that kinematically specifies expectations of object weight. 

Experiment 3: Perception of deceptive intention

The third experiment aimed to determine if people could use kinematics to perceive the intent to deceive. This entails perceiving both the actual state of affairs and the deceptive intention. In this experiment, actors were first asked to lift objects of 3 different weights (6.5, 11.5, and 19 kg). Then, they were given a lighter object (4 kg) and asked to lift it, on different occasions, as if it had the same weight as those previously mentioned 3 weights. Patch-light displays of these conditions were then randomly presented to a group of observers. The observers had two tasks. Firstly, they reported whether they thought the actor was engaging in deceptive behaviour. Secondly, if they thought the actor was not being deceptive, they then reported the actual weight of the object. On the other hand, if they thought the actor was being deceptive, they were asked to report both the actual object weight and the weight deceptively intended by the actor. 

Results showed that deceptive lifts were identified with increasing frequency as the intended object weight went up. This makes sense, since it is usually easier to fake lifting weights that are closer to the actual weight than farther. As expected, among true (non-deceptive) lifts, observers' judgments of real weight corresponded closely with those of intended weight (see Fig. 1). On the flip side, among faked (deceptive) lifts, while judgments of intended weight increased with actual intended weight, judgments of real weight (while overestimated) remained relatively constant across all actual intended weights. This demonstrates that observers were able to perceive both the actual weight of the object, as well as the intended object weight as conveyed by the deceptive actions of the actor. Hence, this suggests that both these dynamic influences (actual object weight and deceptive intent) are specified in the kinematics of the activity. 

Fig. 1 Average judgements of real and intended weight

Like expectations, deceptive intentions can be clearly perceived. And this does not have to be through a mechanism of statistical inference as proposed by many information-processing theories -- experiment 3 suggests that the deception is uniquely specified in the kinematics of the task. It makes intuitive sense too that we can perceive deceptive intent. I'd argue that this is why we can appreciate the art of miming -- we can perceive both what the mime is trying to convey (deceptive intent) and the actual state of affairs! And their art is all about taking advantage of movements and kinematics to convey a certain dynamic that is specified by those kinematics. 

At this point, it is helpful to be reminded that for all this talk about KSD, we still are largely ignorant of the exact informational variables specifying different dynamics (e.g., intentions, deception, expectations etc). Runeson & Frykholm (1983) have provided some candidate areas to explore (e.g., postural adjustments relative to object acceleration), but these remain potential sources until we actually find the exact informational variable. What makes this more problematic is that it is also difficult to quantify the dynamic influences described by the authors. How do you mathematically relate information to intentions, expectations, and deception? It might be more straightforward to search for information specifying physical properties like weight, but what will this look like for more psychical properties? And are these relevant questions, or have I inadvertently introduced a form of Cartesian dualism into my analysis? These will be important questions and challenges to grapple with as I delve deeper into the ecological psych literature.

References

Runeson, S., & Frykholm, G. (1983). Kinematic specification of dynamics as an informational basis for person-and-action perception: Expectation, gender recognition, and deceptive intention. Journal of Experimental Psychology General, 112(4), 585–615. https://doi.org/10.1037/0096-3445.112.4.585