Take a look at this video. Your job: decide which person speaks first:

Click to view Video 1 (QuickTime required)

Let’s make this a poll:

Now, try another one. Once again, decide which person speaks first:

Click to view Video 2 (QuickTime required)

Make your response here:

David Rose and Tanya Clark showed videos like this to 17 student volunteers, and asked the same question: who spoke first in each video? The videos are point-light displays, which show movements of particular points on an actor’s body, but not the nuances of facial movements and expressions.

The videos were used in an earlier experiment about identifying emotion (which we’ve discussed on Cognitive Daily). That study showed that people are accurate in identifying emotions in point-light displays. Now they wanted to find out if people can determine who’s talking in the same videos. Here are the results:

As you can see, accuracy depends on the emotion being portrayed. The videos depicted actors reading scripts while expressing one of several possible emotions. While in the previous study viewers were able to accurately identify the emotions, this experiment shows that people can’t always accurately figure out who’s speaking. While angry and joyful conversations were easy to parse, for conversations expressing fear and disgust, respondents didn’t do better than chance.

If we duplicated these results in our example, accuracy should be much better for Video 1 (anger) than Video 2 (fear).

Rose, D., & Clarke, T. (2009). Look who’s talking: Visual detection of speech from whole-body biological motion cues during emotive interpersonal conversation Perception, 38 (1), 153-156 DOI: 10.1068/p6279