One of the two successful orangutans, Dahi, was able to spontaneously solve the opaque version of the FPT without any visual feedback. The solution (i.e., collecting water from the faucet and pouring it into the tube) appeared suddenly after a period of active but unsuccessful exploration, and was only used to lift the peanut out of the tube (comparison with control conditions). More importantly, in his first encounter with the opaque task, Dahi spat water into the tube a total of four times even though none of the spits produced visible changes in the peanut’s position until it finally reached the tube’s hole. In other words, his spitting behavior was not maintained by differential reinforcement (i.e., repeating those actions that bring the peanut incrementally closer to the top). This provides evidence for the first time that orangutans can solve the FPT without relying on sensorimotor learning, and it may indicate that orangutans may be capable of mentally anticipating the solution to the problem.
Despite being an intuitive problem, which does not require a complex and artificial sequence of actions, solving the FPT has proved to be highly difficult for both human and nonhuman primates, as well as other species (e.g., Bird and Emery 2009; Ebel et al. 2019a; Hanus et al. 2011). In fact, the proportion of successful human and nonhuman individuals is typically modest. Its difficulty rests, at least, on two main features. First, water is rarely used for instrumental purposes among the tested species, and its nature considerably departs from that of typical tools (liquid vs. solid); consequently, its use implies a different kind of means-end coordination. Second, the water has a strong a priori function (i.e., satiating thirst). This latter aspect has to do with functional fixedness, in the words of Shettleworth (2012), “the enemy of [mental] restructuring” (p. 218). Functional fixedness refers to a persistent reliance on past experience with a particular object in a consistent and specific way that precludes an unusual use of such object. One study suggests that this phenomenon may be responsible for so many great apes failing the FPT (Hanus et al. 2011; see also Ebel et al. 2021 for evidence of functional fixedness in other problem-solving contexts). Considering all this, our orangutan showed a sophisticated representational capacity while using water as a tool.
The only previous study of 24 naïve chimpanzees that has used the same opaque version of the FPT failed to find positive results (Ebel et al. 2019b). Although our findings should be taken very cautiously due to the limited sample size (n = 2), this interspecies difference is surprising as both chimpanzees and orangutans are known to show high innovation rates. Perhaps the orangutans’ unique socioecology, life history, and arboreal lifeway prepares this species for a stronger preference for exploring and using objects with the mouth (e.g., O’Malley and McGrew 2000). The mothers’ arboreal lifestyle forces their infants to use their hands to cling onto them and to use their mouth to explore the environment (Schuppli et al. 2021). These natural predispositions could generate richer and more complex sensorimotor schemes which may have favored the emergence of novel oral actions in the FPT context. Yet, considering the ensemble of FPT studies, the number of successful subjects in both chimpanzees and orangutans may have been too small to detect species differences.
At the very least, these findings indicate that success (or failure) in the FPT cannot be reduced to between-species differences. Indeed, the current study together with the previous FPT studies show substantial individual differences in performance within the species tested. A growing literature on nonhuman (and human) innovation consistently shows high rates of individual variability in innovative problem-solving that results from many sources (see Reader et al. 2016 for a review). One of the key predictors for success is the subject’s motivation to engage in the problem that is generally operationalized as both the latency to approach the task and the number of task-directed actions (Griffin and Suez 2014). Although our study design does not allow us to identify the potential contributors to the inter-individual variation observed, our results are consistent with this finding: the withdrawn subject never approached the experimental setup; and the differences in performance between the two successful orangutans mirrored their differences in active exploration of the task. The potential causes for these inter- and intra-individual differences in motivation are nevertheless unknown. Undoubtedly, the study of individual variability is a promising way to shed further light on the mechanisms underpinning innovative problem-solving across species (Kuczaj 2017).
The other successful individual, Ron, eventually got the peanut in the first trial of the original version of the FPT (wet-experimental condition) and was able to transfer this solution to the more demanding transparent version after a single trial. Also, he discriminated between experimental and control conditions, and only directed his spitting behaviors to lift the peanut out of the tube. Although we cannot conclude, in line with previous findings, that this subject innovated by means of mental combinations, the behavioral pattern displayed by Ron was similar to that of Dahi in the opaque condition, i.e., sudden appearance of the solution and exponential time decrease at getting the reward across trials, once the solution had been discovered. These data lend further support to the hypothesis that even when perceptual-motor feedback is available, great apes’ problem-solving skills seem to go beyond the scope of basic sensorimotor processes, and presumably rely on a basic understanding of the means-end relationship involved.
In conclusion, the present study shows for the first time that orangutans can potentially mentally generate the correct solution for the FPT without receiving visual information about the effect of pouring water into the tube to extract the peanut. This depends on some understanding of the means-end relationship involved. However, it is unclear whether the orangutans’ knowledge about the task is encoded as procedural or practical representations (in terms of “what causes the peanut reach the top of the tube”) or as abstract knowledge (in terms of “how the peanut reaches the top of tube”) (Seed and Mayer 2017). To further shed light on this question, more tests of the opaque version of the FPT with naïve orangutans and other great ape species (especially, chimpanzees) are needed. Those tests should systematically manipulate the causal properties involved (e.g., presenting rewards that do not float) to gain insights on the nature of great apes’ innovative problem-solving.