Explanation Roundup (February/March Edition)
Fascinating finds on the cognitive science of explanation for February/March.
Posted Mar 26, 2012
Explanation has made some interesting appearances across the web in the last few weeks. So rather than introducing a new paper or study in this post, I encourage you to explore these fascinating finds:
* In this blog post, Psychology Today blogger Art Markman beats me to presenting a cool new paper by Professor Cristine Legare. Legare taught 2- to 6-year-old children about a device that turned on when some blocks (but not others) were placed on top of it. She then showed children some new blocks that either did or didn't generate the expected effect (for example, a block that looked like one that had previously made the machine turn on, but that didn't have any effect). She found that children were more likely to offer explanations for blocks that generated (or failed to generate) the expected effect, and that - here's the really neat part – the kind of explanation that a child offered could be used to predict how that child would play with the blocks and device when later allowed to explore them. In other words, children's explanations appeared to guide their exploration and therefore what they learned about the world.
* In this post from Scientific American's blog Literally Psyched, Maria Konnikova discusses compelling evidence that telling stories – and in particular offering explanations for our own behavior – is so basic to human cognition that we're willing to fabricate a good story on demand. And what's worse, we don't always realize when we're doing so. Her examples range from famous experiments involving split-brain patients to classic results in social psychology and even Freud. An interesting read!
* This article and the associated video provide an intriguing glimpse into some of the research being conducted in my own department at UC Berkeley, and some of it in my own lab. The article focuses on recent research that investigates how children solve challenging inductive problems – the sorts of problems on which humans typically outperform computers. For example, even young children are often pretty good at figuring out simple causal relationships from only a few examples. Understanding how children and adults solve these problems so succesfuly can help researchers in Artificial Intelligence to build smarter machines. The article mentions some research I conducted with Elizabeth Bonawitz concerning children's preference for simpler explanations, which I discussed in this post back in January. And the video includes a short clip with Caren Walker (a graduate student working with me and with Alison Gopnik) asking children to explain why a block made a machine light up – part of a cool set of findings that I'll write about in a future post.
If you know of any other relevant reads on the cognitive science of explanation, please share them in a comment. Happy reading!