Human infants are amazing learning machines. Not only are they capable of surprising levels of computation, but, as Celeste Kidd, Steven Piantadosi and Richard Aslin of the University of Rochester have found, they can even direct their attention to improve their learning opportunities.
Because of their relative immobility, infants must do much of their learning by passively watching things. However, they don’t gaze indifferently at whatever’s in front of them. Instead, they direct their attention to things that offer them the best learning opportunities. This means that they don’t choose to pay attention to things that are too complex (out of their mental grasp), or too simple (already mastered).
Just how did the researchers come to this conclusion? Once again, we turn to the ‘ol ‘how long does the baby stare’ test. It’s well documented that babies only a few months old will stare longer at things that surprise them. Kidd and her colleagues were able to fine-tune that test by presenting babies with more or less predictable computer images.
They found that if you show babies the same images for too long, they grow bored and look away. If you then show them something new, you regain their interest. However, you can’t retain a baby’s interest with just anything. If the new imagery is too random and unpredictable, the baby also turns away. In other words, the babies prefer to look at things that are novel, but not completely weird.
You can see an explanation below: