Way back when, when the world was young, and I was first studying psychology, the hot new thing was “social learning”—people learning by watching other people. Today, this seems pretty obvious, and even back then it wasn’t exactly big news to anyone seriously studying human language skills or growing children, for instance. But we were coming out of the behaviorist period, which was interested in reinforcement learning, so experimental demonstrations of imitation learning were galvanizing.
Of course, as so often is the case with “advanced” cognitive skills, social learning was viewed as a uniquely human ability. I mean, this is pretty complicated cognitive processing, and obviously foundational to the creation and promulgation of culture. And, by the way, in my anthropology classes, we were confidently taught that “culture” is the thing that makes us human, that separates us from all the other animals. So, from these perspectives, we wouldn’t a priori expect “lower” animals to be able to do social learning [2].
Consequently, it was a thunderbolt when other primates were observed doing it. And, in the decades since, many mammals and birds have been shown to learn by imitation. (And, by the way, animals may exhibit group specific, socially promulgated behaviors that look pretty darn like “culture”.)
We still feel that social learning is a really clever skill that needs serious brain power. Mammals and birds have complicated brains and behaviors that are overall pretty similar to humans. This makes it easy to imagine them doing the cognitive processes involved in imitation learning. But other species—probably not.
So—thunderbolt time, again.
This winter, British researchers report that bumblebees can learn by watching other bees [1]. Whoa!
The research tested the bees with an experimental puzzle. Bumblebees learned to solve a puzzle by choosing one of two options. Other bumblebees observed them, and when faced with the puzzle chose the option they saw succeeding. Furthermore, the ability to solve the problem spread through a group of bees, i.e., the knowledge diffused through the social group.
The researchers note that the bees showed a distinct preference for socially acquired knowledge. Even when the bee learned more than one correct answer, they preferred the one they learned by observing the other bees in their group. This is similar to behavior seen in monkeys and birds, which choose the behavior modeled by others, especially higher status individuals. Dare we note that people tend to imitate high status or socially salient role models, too.
This is a fascinating result for many reasons. For one thing, it shows that the behavior of these bees is not exclusively driven by genes, or at least may not always be purely genetic. This opens questions about the emergence of complex behavior such as nest building and foraging strategies. And it opens chicken-and-egg questions about the possible mutual influence of social learning on genetic selection. E.g., can socially learned behaviors create advantages for certain queen’s genes, which are passed on to descendants? And vice versa, can certain genes promote social diffusion of learned knowledge, which is advantageous to the hive?
As the researchers note, this artificial experiment only shows that this kind of “cultural” evolution is possible. We still don’t know that it happens in natural settings. Bees have short lives, several generations die out in a single year. At that pace, cultural diffusion seems unlikely. But we now know it is possible, so we wonder if it happens. If so, can we observe it?
I think it is also interesting to think about the cognitive abilities that social learning implies.
Human social learning seems to involve implicit and unconscious recognition of other humans, including an implicit “model” of human behavior. I.e., imitation implies recognition that other people are models, and also an ability to correlate observed behavior of another person with out own behavior, i.e., the ability to mimic what you see and hear. Computer science has discovered that these abilities are pretty hard to program, so whatever is going on in natural brains is non-trivial.
Other mammals seem to have similar brains to us, and also similar cognitive abilities. Bird brains are large and complex, but quite different from mammals. Still, it is plausible that birds have similar cognitive abilities. In short, any animal with a big, complex brain might well have the underlying capability to learn through imitation.
However, bees have totally different and much, much smaller brains than any bird or mammal. In fact, the most complex behaviors of bees are generally group behaviors. Bees seem to be a hive brain, with each individual having a fairly simple brain.
This social learning implies that individual bees can recognize other bees—visually—and can correlate the behavior of other bees with their own behavior. Wow! If this is a correct interpretation, then there are some serious questions to explore about how the bees perceive other bees, how they perceive puzzles, and how they make decisions.
And, while we’re at it, how does this kind of social learning mix with swam behaviors? Bees cooperate in complex ways. Perhaps some of the cooperation is facilitated by social learning. learn to work together partly by following genetic templates and part by following the leader. How does all that work?
Some of the classic swarm behaviors involve foraging, which often includes an important diffusion of knowledge about the environment. I have to wonder if the same brain structures are involved in the rapid diffusion of foraging information and this puzzle solving behavior? And if so, are these brain structures analogous to how birds and mammals learn, or is this a kind of convergent development of a similar skill from different cognitive machinery?
Interesting stuff.
- Alice D. Bridges, HaDi MaBouDi, Olga Procenko, Charlotte Lockwood, Yaseen Mohammed, Amelia Kowalewska, José Eric Romero González, Joseph L. Woodgate, and Lars Chittka, Bumblebees acquire alternative puzzle-box solutions via social learning. PLOS Biology, 21 (3):e3002019, 2023. https://doi.org/10.1371/journal.pbio.3002019
- Emily McGarvey, Bumblebees learn to solve puzzles by watching peers, study finds, in BBC News, March 8, 2023. https://www.bbc.com/news/uk-64884107
Robert McGrath's Blog 