What cognitive strategies do orangutans (Pongo pygmaeus) use to solve a trial-unique puzzle-tube task incorporating multiple obstacles?

Abstract

Apparently sophisticated behaviour during problem-solving is often the product of simple underlying mechanisms, such as associative learning or the use of procedural rules. These and other more parsimonious explanations need to be eliminated before higher-level cognitive processes such as causal reasoning or planning can be inferred. We presented three Bornean orangutans with 64 trial-unique configurations of a puzzle-tube to investigate whether they were able to consider multiple obstacles in two alternative paths, and subsequently choose the correct direction in which to move a reward in order to retrieve it. We were particularly interested in how subjects attempted to solve the task, namely which behavioural strategies they could have been using, as this is how we may begin to elucidate the cognitive mechanisms underpinning their choices. To explore this, we simulated performance outcomes across the 64 trials for various procedural rules and rule combinations that subjects may have been using based on the configuration of different obstacles. Two of the three subjects solved the task, suggesting that they were able to consider at least some of the obstacles in the puzzle-tube before executing action to retrieve the reward. This is impressive compared with the past performances of great apes on similar, arguably less complex tasks. Successful subjects may have been using a heuristic rule combination based on what they deemed to be the most relevant cue (the configuration of the puzzle-tube ends), which may be a cognitively economical strategy.

Biological and artificial cognition: what can we learn about mechanisms by modelling physical cognition problems using artificial intelligence planning techniques?

“…It also broadly agrees with the findings in Tecwyn et al [10] that the simulated models involving the configuration of the open ends (OEALG and OELG in figure 3) fitted the behaviour of Amos and Sandy better quantitatively (percentage agreement ranges from 64.06 per cent to 73.44 per cent for Amos and Sandy compared with OEALG and OELG) than the alternative simulated models.…”

“…10], but the general concepts and workflow (figure 1) could be implemented in a number of different ways, using different modelling techniques. In §4, we use this approach to model a task previously presented to orangutans (Pongo pygmaeus), in which an orangutan must push a nut through a horizontal tube to an opening, while avoiding allowing the nut to fall into inaccessible traps [10]. AI planning is a technique that searches for a series of actions that an agent can execute to achieve a goal (see also Shanahan [15] for a detailed example of how searches of action-trees can be implemented).…”

“…We have used AI planning to explore a range of models for a puzzle tube task previously used with orangutans [10], in a way that allows us to compare our simulated results with those obtained from real animals. The goal for the orangutan was to roll a nut through a horizontal tube to a point at which it could access the nut.…”

“…This tends to happen when researchers are focussed on testing a hypothesis about a particular functional label (and therefore neglect the possibility that other processes may be at work), or because the problem posed to the animal has not been thoroughly analysed in advance to discover the possible ways in which it might be solved. (v) Once designed, experiments can also be difficult to analyse because they involve limited numbers of trials or first-trial performance only, because there is often substantial inter-individual variation (see recent studies [4,6,10,11] among many others), or because animals show unanticipated behaviours (see point iv). Quantitative analysis is desirable, but it would also be useful to compare patterns of behaviour qualitatively against a set of models describing the possible ways in which the task might be solved.…”

Task-specific modulation of adult humans’ tool preferences: number of choices and size of the problem

“…These constructs-cognitive effort, usability, and margin of safety-and related variables, processes, and rules may be why adult humans' tool selections differ from those of crows on similar tasks. In short, people's and other animals' behavior on tool-use tasks may be due to things (e.g., features of the task, associative learning, or procedural rules) other than simply the causal structure of a task and a subject's understanding of physical causality (Silva, Page, & Silva, 2005;Silva & Silva, 2006;Tecwyn, Thorpe, & Chappell, 2012).…”