Thursday 24 January 2013

Is hefting to perceive the affordance for throwing a smart perceptual mechanism?

In the last post, I reviewed Geoff's first paper looking at whether people can perceive the affordance for throwing an object to a maximum distance and a first swing at identifying the information specifying the affordance. People can perceive the affordance. Bingham et al then identified an invariant relation between the timing of the motions of the wrist and elbow when people hefted the balls they chose as optimal for throwing, and showed that this kinematic pattern specified a peak in the function which determined how much kinetic energy was transferred to the ball. They suggested that this relation in the joint movements served as information for the dynamic property which led to a maximum distance throw, and that this is how hefting was able to provide information about throwing. They suggested that this was a smart perceptual mechanism for perceiving the affordance property.

That was where things stood until Zhu & Bingham (2008) ran an extensive replication and extension of the original study, to test the specific smart perceptual mechanism proposed by Bingham et al (1989). 

Experiment 1: Replicating Bingham et al, 1989
This experiment had ten people first heft then throw objects to a maximum distance. There were 6 object sizes covering the full range of graspable objects (1-6in). Within each size there were 8 weights. The exact weights differed (see Zhu & Bingham, 2008, Table 1) because of limitations in the materials used; but within each size each weight increase was 1.55 times the previous weight. This constant geometric progression preserved the relational structure of the weight distribution within each size.

Participants hefted weights in increasing order within each size and ranked their top three choices for throwing to a maximum distance. The weighted average of the selected weights provides an estimate of the actual preferred weight for that size, overcoming the fact that there were only a finite number of actual options, samples from a continuous relation. In a separate session, all the participants threw all of the objects as far as they could three times in a random order.

Participants all threw with varying ability and picked different weights within each size, mostly explained by variations in the size of the participant. However, they all showed the same pattern as Bingham et al 1989; they picked heavier weights for larger sizes, and threw their preferred objects the farthest. To compare people, Zhu and Bingham rescaled the selected weights as a function of the preferred weight (to get everyone on the same scale) and plotted the distance thrown (Figure 1)
Figure 1. The distance thrown as a function of ball size and weight (scaled by preferred weight))
For each size, follow the line along the rescaled weight dimension. There is a peak distance for each combination, and that peak is projected down onto the floor of the plot. 0 on the rescaled weight axis corresponds to the preferred object and the line is arranged around the preferred object weight. As in the previous study, people can heft objects and perceive the affordance for throwing to a maximum distance.

Experiment 2: What if you heft with something other than your hand?
Bingham et al 1989 suggested that people choose heavier weights as size increases to offset the size-induced increased in stiffness about the wrist, which would in turn preserve the timing between the elbow and wrist movement that optimised the flow of kinetic energy along the arm to the projectile. This, they suggested, was the smart perceptual mechanism people were employing to perceive a throwing affordance via hefting. If this is the case, then hefting without involving the wrist would interfere with this perception, even though both size and weight were being perceived.

Two groups of participants therefore hefted the balls with either their elbow or their feet. They could see the size and feel the weight; but could they perceive the affordance? The answer is not really; their judgements were much more variable and they systematically selected objects that were too heavy

Figure 2. Left: Mean selected weights using the hand (diamonds), elbow (squares) and foot (triangles). Right: Standard deviations of these judgments

Experiment 3: Hefting with your foot to throw with your foot
The smart perceptual mechanism being investigated suggests that people can use hefting-by-hand to perceive throwing affordances because the dynamics (specifically, the optimal way to transmit force) of the two tasks overlap. This isn't the case for the elbow or the foot; but can people select objects they can throw with their foot the farthest if they heft with their feet? 

The  answer is no, not really. People's judgments were fairly random, and people did not typically throw their preferred objects the farthest. This presumably would get better with practice, but in general Experiments 2 & 3 together suggest that the perception of the throwing affordance relies on the overlap between hefting with the hand and throwing with the hand, because that's how we typically interact with the dynamics of projectile motion.

Experiment 4: Do size and weight both matter for achieving maximum distance?
The specific smart perceptual mechanism proposed by Bingham et al (1989) is that both size and weight affect the dynamics of both hefting and throwing in the same way. Weight is known to affect the dynamics of throwing; as weight increases, release velocities come down. What about size?

4 participants threw objects of 4 sizes, 8 weights within each size. The authors measured the release angle produced (which was, on average, a quite low 24°; the optimal angle for maximum distance given air resistance is about 36°). Zhu & Bingham then used these angles and the thrown distances for these objects from Experiment 1 in a simulation of projectile motion to estimate the release velocities (this reflects limited access to the necessary equipment, I think; our more recent work measures these release parameters directly from the high speed video footage). 

Once the weights reached .05kg, the release velocities decreased with increasing weight. However, neither release angle nor release velocity showed any consistent effect of size, suggesting that only weight affects the dynamics of throwing. Size has it's effect via the dynamics of projectile motion; as size (specifically, the cross sectional area) increases, drag increases, and distance goes down. Zhu & Bingham conclude that the specific smart perceptual mechanism proposed in Bingham et al (1989) is not correct.

This paper investigated whether hefting an object provides information about that object's throwability to a maximum distance via the smart perceptual mechanism proposed in Bingham et al (1989). That paper suggested that both hefting and throwing share certain dynamical features. Throwing entails transmitting force along the arm and the efficiency of this depends on achieving a particular timing pattern between the limb segments. This force is stored briefly in the tendons about the wrist and then delivered to the projectile in a fast snap. This snap and the limb segment timing are preserved if object weight increases as object size does, because the increase in size affects the wrist dynamics. So Bingham et al suggested that both object size and weight affect the dynamics of hefting and throwing identically, and hence the former can produce information about the affordances for the latter.

Experiments 1-3 of the current study support this hypothesis; people can perceive the throwing affordance but only when hefting by hand. However, Experiment 4 suggests that this specific smart mechanism might not work, because object size does not affect the dynamics of throwing (no effect on release angle or velocity). It does affect distance, but it does so by affecting the dynamics of projectile motion. In order to learn to perceive throwing affordances via hefting, therefore, people may need to have experience seeing how far balls of different sizes travel when thrown (i.e. experience with the dynamics of projectile motion). The next paper I'll review examines this hypothesis in more detail. 

ResearchBlogging.orgBingham, G., Schmidt, R., & Rosenblum, L. (1989). Hefting for a maximum distance throw: A smart perceptual mechanism. Journal of Experimental Psychology: Human Perception and Performance, 15 (3), 507-528 DOI: 10.1037//0096-1523.15.3.507 Download

Zhu, Q., & Bingham, G. (2008). Is hefting to perceive the affordance for throwing a smart perceptual mechanism? Journal of Experimental Psychology: Human Perception and Performance, 34 (4), 929-943 DOI: 10.1037/0096-1523.34.4.929 Download

No comments:

Post a Comment