Dogs' insensitivity to scaffolding behaviour in an A-not-B task provides support for the theory of natural pedagogy

Abstract

Executive function plays a critical role in regulating behaviour. Behaviour which directs attention towards the correct solution leads to increased executive function performance in children, but it is unknown how other animals respond to such scaffolding behaviour. Dogs were presented with an A-not-B detour task. After learning to go through gap A to obtain the reward, the barrier was reversed, and the dogs had to inhibit their learned response and enter through gap B on the opposite side. Failure to do so is known as the perseveration error. In test trials, dogs taking part in one of two scaffolding conditions, a pointing condition, where the experimenter pointed to the new gap, and a demonstration condition, where the experimenter demonstrated the new route, were no less likely to commit the perseveration error than dogs in a control condition with no scaffolding behaviour. Dogs' lack of responsiveness to scaffolding behaviour provides little support for suggestions that simple social learning mechanisms explains scaffolding behaviour in humans. Instead, our results suggest that the theory of natural pedagogy extends to the development of executive function in humans. This suggests that human children's predisposition to interpret ostensive-communicative cues as informative may be an innate, species-specific adaptation.

Figure 1

Dogs in all conditions were equally likely to commit the perseveration error on the first trial and as likely to learn to avoid committing it over all four trials. Each separate block represents a condition (with 10 dogs in each condition.) Each row represents an individual dog and each column represents an individual trial. 9/10 dogs in each condition showed the perseveration error on the first trial.

Figure 2

Set up of experimental room. The subject was placed on the starting position opposite the experimenter. In the learning trials, the experimenter would get the dog’s attention, show it the treat, place the treat in a bowl, and place the bowl on the target. After the experimenter gave the release command, the owner would release the dog and the dog would have to enter through the gap in the barrier to reach the treat. The test trials were the same as the learning trials, but the gap was moved to the other side. The experimental procedure was the same except that in the scaffolding conditions, the experimenter would either point to the gap or demonstrate the route before giving the release command.