Skip to main content

Does Social Learning Need to be Social? Insights from Bumble Bees

This might seem perplexing to some, but Ive just spent two days listening to talks and meeting with people who all work on social insects. And it was great.

This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American


This might seem perplexing to some, but I’ve just spent two days listening to talks and meeting with people who all work on social insects. And it was great. I was at Royal Holloway, University of London, where the IUSSI meeting was taking place. The IUSSI is the ‘International Union for the Study of Social Insects’, although they seem to let people in who work on social spiders too (a nice inclusive attitude if you ask me). This meeting was specifically for researchers who are in the UK and North-West Europe, of which there are a surprisingly large number. The talks were really good, sharing a lot of the recent research that’s happened using social insects, and I thought I’d share my highlight of first day’s events here.

One of my favourite talks from the first day was from Elli Leadbeater who spoke about work carried out primarily by Erika Dawson. I’ve written before about ‘social learning’ in monkeys and whales, where one animal can learn something from observing another animal, normally of the same species.

Dawson and her colleagues were looking specifically at whether there is actually anything ‘social’ about ‘social learning’, or whether it can be explained with the same mechanism as other types of learning.


On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


In the simplest form of learning, associative learning, an animal learns to associate a particular stimulus (for example a particular colour, smell or sound) with a reward (usually food). The classic example of this was Pavlov’s dogs, who learned to associate the sound of a metronome with food. When Pavlov then sounded the metronome, the dogs salivated even when there was no food present.

Pavlov then went on to take this a step further, to ‘second order conditioning’. Here, the animal learns to associate something else, for example a black square in the case of Pavlov’s dogs, with the metronome (already associated with the food). The dogs will then salivate in response to the black square, even though they’ve never had it directly paired with the food.

The recent study by Dawson et al. has found that second order conditioning can actually explain social learning. Bumble bees have been previously shown to be able to watch where other bees get their food (sugar water) from and then copy them. For example, if a bee sees two flowers, red and blue, and there are some bees feeding on the red flower, then when the first bee is given a choice between red and blue, it will go to a red flower, just by copying the bees that it’s seen.

This seems pretty impressive. However, the researchers showed that the bees were solving this through having already learned that other bees were associated with a reward (sugar water). As a bee’s nestmates generally would be going to artificial flowers where there was food, the bee watching them had just learned my nestmates = food, now my nestmates are on the (red) flower, so red flower = food.

The researchers proved this by having bees in different 'treatment' groups that they trained to learn different things. First, some bees were trained to learn that when they saw bees on a feeder, these feeders had yummy sugar water in them. When these ‘observer’ bees then watched other bees in a ‘test’ scenario, they then chose the flower colour that they saw these bees on.

However, in another treatment group, bees were instead trained that when they saw bees on a feeder, this would be full of a distasteful substance, quinine. Instead, the feeders without bees on them were the ones they wanted, the sugar water. Now, when these bees were given the chance to watch bees on the two different flower colours, they avoided the colour of flower they saw the bees on- they had learned to associate bees with the nasty quinine instead of the sugar water.

To really show conclusively what was going on, the researchers had a third group of bees, that were given no training. They were allowed to visit feeders, with quinine and sugar water like the other treatment groups, but with no bees on them, meaning they had no opportunity to learn associations between bees and food. When these bees were tested in the same set-up as the other bees, they did not choose either the colour of flower that they saw bees on, or the other colour.

So, this nice simple experiment has clearly demonstrated that learning from another animal can be as simple as two learned associations (so-called second-order conditioning).

Therefore, any animal that is capable of learning second-order conditioning (an awful lot of animals) should also be capable of learning from observing other individuals, if in the right circumstances. This calls into question whether so-called ‘smarter’ animals (like primates) are using the same mechanism to learn from each other. I’ll bet there will be a study out in the next year that will let us know.

 

Photo credits

Vervet monkeys: Erica van de Waal

Ivan Pavlov: copyright expired

bumble bee: Penny Metal

 

Reference

Dawson, E. H., Avargues-Weber, A., Chittka, L., & Leadbeater, E. (2013). Learning by Observation Emerges from Simple Associations in an Insect Model. Current Biology, 23, 1-4. http://dx.doi.org/10.1016/j.cub.2013.03.035