Dogs are inclined to follow our lead in many ways, but they don’t go overboard if it does not serve their interests, say the authors of a new study. If people give dogs bad advice, they figure out that it is worth ignoring, according to a new study in the journal Developmental Science. Let’s look into how they arrived at this claim, which I don’t think is supported by the data.
The researchers were investigating whether dogs (and dingoes) would imitate the way people showed them how to get food out of a puzzle box even when there was an easier way to do it. Only one step was required to reach the food, and that was lifting the lid to a box. As part of the experiment, humans added an extra, unnecessary action to the process by pulling a lever that did nothing, and then lifting the lid of the box.
Both the dogs and the dingoes quickly learned to skip the step with the useless lever and just open the box to get to the treat inside. In other words, it looked like they ignored the useless instructions from the humans. This behavior differs from human children, who tend to perform all the steps they have been shown even when some of them are unnecessary. That behavior is called “overimitation” and the uncritical copying of the behavior they observe may allow kids to minimize the amount of trial-and-error learning they must do.
The dogs and the dingoes observed humans opening the box, and were then repeatedly given the opportunity to open the puzzle box. Over time, as they gained experience with it, they were less likely to use the lever. The experimenters consider this evidence that both species learned that pulling the lever was an unnecessary step for opening the box, even though they saw humans doing it. I agree that the data support the idea that they learned that the lever is irrelevant. I just don’t think that observing the humans pull the lever made any difference, and that’s because this study does not find any evidence that dogs imitated the humans at all.
GET THE BARK NEWSLETTER IN YOUR INBOX!
Sign up and get the answers to your questions.
In addition to the experiment in which subjects observed humans pulling the irrelevant lever, there were also a series of trials (with a different set of dogs and dingoes) in which they were presented with the puzzle box without any opportunity to observe a human opening it. In that experiment, the dogs and dingoes were solving the puzzle without having seen anyone else open it, so they were doing it completely on their own. The authors write that, “dogs were equally likely to use the irrelevant lever, regardless of whether they witnessed a demonstration (in Experiment 1) or not (in Experiment 2).
They point out that there was no evidence that dogs were more likely to copy the humans’ actions than the dingoes were, but what’s just as important is that there was no evidence that the dogs were copying humans at all. Therefore, I don’t think that their conclusions about dogs and overimitation hold water. They would first need to show that dogs copy any human behavior, which they do not do, in order to then test whether dogs copy irrelevant human behavior.
There was one interesting conclusion from this study, though it has nothing to do with imitation, social learning, or human influence on dogs’ actions. Evidence from this study, as well as previous research, indicate that dingoes solve problems more quickly and with greater success than dogs. In Experiment 3 in this research paper, a different puzzle box was used. Pulling the lever was an essential step in opening this particular puzzle box. In this experiment, both dogs and dingoes did pull the lever in order to access the treat inside. When compared to the rates of pulling the lever when it was pointless, dingoes showed a greater change in their behavior. That is, they were more likely than dogs to pull the lever only when it was relevant, unlike dogs, who pulled it quite often even when it was not an essential part of the box-opening task.
Questions about the possibility of overimitation in dogs are extremely interesting, and I want very much to know more about this behavior, which I don’t think was adequately addressed by this study.