I chanced upon an interesting paper the other day about how surprising events enhance learning. What struck me at first was the imaginative study design. Memory studies tend to be much of the same, having changed little since Ebbinghaus - you can only examine someone’s ability to recall digits, letters, words and word pairs for so long and not expect to get the same result. This obsession with so-called verbal learning experiments can only yield so much. George Miller is purported to have considered the field as ‘perhaps the dreariest area of the whole of psychology,’ at least according to Baddeley. I tend to agree.
The study in question looks at whether or not very young children learn better after witnessing impossible events compared to possible, yet improbable ones. I suspect that at this point, some people might scoff at the very idea that we can learn better if surprised. I mean, whatever happened to the tried and tested rote memorisation of nonsense words? (See my previous comments). Anyway, if you find yourself scoffing, just wait for the study’s design. You’ll love that!
Aimee Stahl and Lisa Feigenson recruited 335 children with an average age of 36 months. They divided the children into two groups and showed them a gumball-type machine, only it was filled with toys rather than gumballs. In the first condition, the researchers had filled the machine with a mix of pink and purple toys, but in the second condition, filled it with only purple toys. They gave each child a coin to put in the slot and retrieve a toy. They then told them the toy was called ‘Blick’ (a made-up word so as to avoid any pre-existing knowledge that might help them recall the name).
Now, because psychologists are just plain devious, they’d rigged the second condition so that the child might receive a pink toy, even though the machine only contained purple toys. In condition 1, the chance of receiving a pink toy was probable; in condition 2 it was impossible.
Later, the researchers asked the children the name of the toy. Those in the second condition (the impossible one) were more likely to remember the name than those children in condition 1.
While the design might be unique, the assumptions aren’t. In a 2022 paper in the Journal of Neuroscience, the authors used fMRI to test if expectation violation (or surprise to you and me) before or during learning triggers adaptive encoding mechanisms that result in better memory for unexpected events. They concluded that it does this by shifting the brain towards an encoding state, enhancing perceptual processing, and engaging the hippocampus and midbrain. This mechanism not only prioritises the encoding of unexpected events but also has lasting effects on the retrieval of related information, particularly when those events share perceptual similarities.
Surprise, then, might well enhance learning, but how that could impact real classrooms is yet to be determined. Stahl and Feigenson are pretty clear about the need for more research before teachers start getting too imaginative in lessons.
Like any good research, of course, the study throws up several interesting questions (while answering very few). The obvious is how might the impossible condition have resulted in better retention? And how does this relate to what we currently understand about how memory and learning work?
We know from previous research and cognitive models that enhanced learning, including greater elaboration, indicates deeper processing. For example, Craik and Lockhart’s levels of processing theory posits that different types of information are encoded in different ways (or at different levels): structural, phonemic, and semantic. Deeper levels require higher cognitive engagement. In the study described above, the impossible condition most likely triggered a more profound cognitive engagement, facilitating memory formation and word association (a kind of WTF! moment).
Our natural human tendency to find explanations leads to active processing, where we attempt to reconcile what we’ve witnessed with what we already know. We know that if the machine only contains purple toys, we can’t receive a pink one. Similarly, if we know the machine contains both pink and purple toys, there is a greater possibility we will receive a pink toy.
Surprise may also help increase attention. The impossible condition captured the children’s attention more than the mere probable, leading to more focussed observation and analysis. Furthermore, faced with an impossible outcome, we might explore various hypotheses to explain the discrepancies between our observation and our currently held understanding about what is possible and impossible. We do this when we see a magician perform a trick - we think about a hidden compartment, mirrors, trapdoors, and so on.
However, as the authors are quick to point out, simply being surprised isn’t enough to guarantee enhanced learning. Improbable events are also surprising, but don’t always trigger the same level of cognitive elaboration. Improbable events already have a readily available explanation - chance. If we draw a rare object from a pool of mostly frequent objects, we can attribute this to luck, with no need to revise our understanding of the objects or the machine. Similarly, winning the lottery might be surprising, but can be explained by probability and luck. Impossible events demand a more significant cognitive shift to accommodate any new information. The result of this is deeper learning.
In other news: I’m now exclusively on Bluesky