Probing the limits of flexible tool use in kea

Dr Megan Lambert & Professor Ludwig Huber Background: Human tool use is indisputably complex. We build skyscrapers, send satellites into space and use technology to communicate instantaneously with people across the globe. These remarkable feats are made possible by several underlying cognitive abilities: for example, we are able to flexibly choose different tools depending on the task at hand, even using tools on other tools (e.g. using percussive stones to shape arrowheads), and actively seek information about how objects might work as tools when we use them. We often take these abilities for granted in our daily life, yet it is unclear to what extent we share them with other animals. Furthermore, we know that our own ability to use tools is impacted by other factors such as memory or attention, but these factors are often overlooked in studies of tool use in other animals. Goals: The purpose of this project is to investigate the different factors that impact problem-solving and tool-using behaviour in a large brained parrot species, the kea. Specifically, it aims to answer three key questions: (i) what are the limits of flexible tool use? (ii) how do these birds learn about and create relationships between objects to solve problems? (iii) do these birds actively seek information about how objects can be used as tools before using them? Methods: Study 1: Kea will be presented with several tool-use tasks that impose different demands on attention, memory and representation to determine the extent to which flexible tool use is impacted by these factors. Study 2: Subjects will be given a new puzzle box where they can learn how different pieces of the box move and interact with one another, and then use this information to create relationships between the objects to solve a problem. Study 3: Kea will be given a choice between two objects that look identical but function differently to solve a task (e.g., one is heavier than the other) to determine whether they explore the objects first to learn about how they function. Contribution: This project uses a new approach of focusing on the basic cognitive elements that may affect tool-using performance in different contexts. Identifying these elements and their impact on problem solving will help us to make more informed comparisons of cognition across species, ultimately generating insight into some of the key similarities and differences between human and nonhuman tool use.

Human tool use is indisputably complex. We build skyscrapers, send satellites into space and use technology to communicate instantaneously with people across the globe. Although we now know that a range of other species use tools, we still know little about the cognitive differences that might explain the vast divide between the tool use of humans and other animals. This project focused on kea, a highly playful New Zealand parrot, to understand how they go about solving new problems, and whether these abilities might be limited depending on the complexity of the task. To do this, the kea parrots were presented with several new puzzles to solve. Through a series of studies, we found that the kea's playful nature allows them to quickly discover solutions to new problems through playful exploration, suggesting that play is an important avenue for innovating tool use. When allowed to choose between different tools to push a food reward out of a tube, the kea were able to choose the right tool for the job (the tool that was long enough or skinny enough to fit in the tube), but showed difficulties when they had to pay attention to more than one feature of the problem at the same time (the tool that was long enough and skinny enough), suggesting that the kea's tool use is limited depending on how many features of the task they have to pay attention to at the same time. The kea also frequently sought information about which tool was appropriate to solve a task, for example by lifting both of the objects to see which was heavier, and also rapidly learned to discriminate between objects based on their weight alone. Together, these results highlight the role of play in tool use innovation and show that kea can flexibly solve problems and seek information about which tools they need, but unlike humans, their tool selection performance may be limited to one relevant feature at a time.