Decision-making flexibility in New Caledonian crows, young children and adult humans in a multi-dimensional tool-use task

Abstract

The ability to make profitable decisions in natural foraging contexts may be influenced by an additional requirement of tool-use, due to increased levels of relational complexity and additional work-effort imposed by tool-use, compared with simply choosing between an immediate and delayed food item. We examined the flexibility for making the most profitable decisions in a multi-dimensional tool-use task, involving different apparatuses, tools and rewards of varying quality, in 3-5-year-old children, adult humans and tool-making New Caledonian crows (Corvus moneduloides). We also compared our results to previous studies on habitually tool-making orangutans (Pongo abelii) and non-tool-making Goffin's cockatoos (Cacatua goffiniana). Adult humans, cockatoos and crows, but not children and orangutans, did not select a tool when it was not necessary, which was the more profitable choice in this situation. Adult humans, orangutans and cockatoos, but not crows and children, were able to refrain from selecting non-functional tools. By contrast, the birds, but not the primates tested, struggled to attend to multiple variables-where two apparatuses, two tools and two reward qualities were presented simultaneously-without extended experience. These findings indicate: (1) in a similar manner to humans and orangutans, New Caledonian crows and Goffin's cockatoos can flexibly make profitable decisions in some decision-making tool-use tasks, though the birds may struggle when tasks become more complex; (2) children and orangutans may have a bias to use tools in situations where adults and other tool-making species do not.

Fig 1

Stone-apparatus (left) and stick-apparatus (right), a) shows the stick-apparatus for the crows and b) for the humans, with both tools—the functional tool indicated by a green hook and the non-functional tool with a red cross.

Fig 2. All conditions.

(a) Tool Selection Condition: both tools present, most preferred food (MPF) inside: (b) Motivation Condition: functional tool present, MPF inside and outside; (c) Quality Allocation Condition: functional tool present, either MPF inside (left), and the least preferred food (LPF) outside, or MPF outside and LPF inside (right); (d) Tool Functionality Condition: functional tool present (left) or non-functional tool present (right), MPF inside and LPF outside; (e) Tool Selection Quality Allocation Condition: both apparatuses present with both tools, MPF in stone-apparatus (left), or stick-apparatus (right), LPF in other apparatus; (f) Apparatus Functionality Condition: both apparatuses and both tools present, MPF in stone- (left) or stick- (right) apparatus, other apparatus empty; (g) Apparatus Choice Condition: only one (functional) tool is present, both apparatuses presented and baited with MPF.

Fig 3. Mean percentage of correct trials for each condition for New Caledonian crows, children aged 3–5 years, adult humans (present study), Goffin’s cockatoos [50] and orangutans [5].

* indicate significant selection of correct choice within condition for each age from exact two-tailed binomial tests for the children and adults, and 1-sample Wilcoxon signed rank tests for the crows, cockatoos and orangutans. Dashed horizontal line indicates chance level.