Abstract While humans are able to understand much about causality, it is unclear to what extent non-human animals can do the same. The Aesop's Fable paradigm requires an animal to drop stones into a water-filled tube to bring a floating food reward within reach. Rook, Eurasian jay, and New Caledonian crow performances are similar to those of children under seven years of age when solving this task. However, we know very little about the cognition underpinning these birds' performances. Here, we address several limitations of previous Aesop's Fable studies to gain insight into the causal cognition of New Caledonian crows. Our results provide the first evidence that any non-human animal can solve the U-tube task and can discriminate between water-filled tubes of different volumes. However, our results do not provide support for the hypothesis that these crows can infer the presence of a hidden causal mechanism. They also call into question previous object-discrimination performances. The methodologies outlined here should allow for more powerful comparisons between humans and other animal species and thus help us to determine which aspects of causal cognition are distinct to humans.

Citation: Logan CJ, Jelbert SA, Breen AJ, Gray RD, Taylor AH (2014) Modifications to the Aesop's Fable Paradigm Change New Caledonian Crow Performances. PLoS ONE 9(7): e103049. https://doi.org/10.1371/journal.pone.0103049 Editor: Thomas Boraud, Centre national de la recherche scientifique, France Received: January 19, 2014; Accepted: June 25, 2014; Published: July 23, 2014 Copyright: © 2014 Logan et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: CJL was supported by a SAGE Junior Research Fellowship at the SAGE Center for the Study of the Mind at the University of California Santa Barbara, and received a National Geographic Society/Waitt Grant (number 252-12) to conduct field work. RDG and AHT were supported by the Marsden Fund. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Publication costs were provided by the University of California Santa Barbara Open Access Pilot Fund Program. Competing interests: The authors have declared that no competing interests exist.

Introduction Humans have an excellent understanding of the relationships between cause and effect [1]-[3]. Currently, little is known about which aspects of this understanding are unique to our species or how such understanding evolves, despite a number of recent claims for human uniqueness [2], [4]. One reason for the slow progress in this area is that, until recently, the predominant test of causal cognition was the trap-tube task, where food has to be moved out of a horizontal tube while avoiding a hole [5]–[11]. There are a number of issues with this paradigm [12]–[13], including the failure of critical controls by adult humans [14], the large effect that small modifications to the paradigm have on problem-solving capabilities [7], and the inability of large numbers of individuals from each species tested to learn any strategy to solve the task [5]–[6], [11]. The Aesop's Fable paradigm has recently emerged as an alternative test of causal cognition [15]. This paradigm is useful because it requires a novel form of tool use (stone dropping into water-filled tubes), which is not seen in the wild and thus allows cross-species comparisons of causal cognition [16]. In this paradigm, subjects are presented with a water-filled tube that contains an out-of-reach floating food reward. To solve this problem, subjects must drop objects into the tube to raise the water level, thus bringing the food within reach. Rooks (Corvus frugilegus), Eurasian jays (Garrulus glandarius), New Caledonian crows (Corvus moneduloides) and children (Homo sapiens) can discriminate between functional and non-functional substrates (water vs. sand/sawdust/air) and objects (large vs. small, sinking vs. floating, hollow vs. solid) when presented with choices [13], [15]–[18]. However, these species differ in their reactions to the U-tube task, a variant of the Aesop's Fable paradigm. This task examines how subjects respond to an unexpected effect after dropping a stone into a tube, namely that the stone causes water to rise in a seemingly unconnected adjacent tube. Eurasian jays were the first species presented with the U-tube task, in an attempt to determine whether a confusing cause and effect relationship would inhibit their learning of a simple rule [18]. In the task, two large tubes were positioned adjacent to a small middle tube containing food. Since the small tube was too small to drop stones into, subjects had to drop stones into one of the large tubes. Only one of the large tubes was connected to the small tube under the table, thus making this tube the functional option, though because the base was covered, the causal mechanism was hidden. Each large tube was marked with a distinct color cue. To solve the task, subjects had to notice that dropping a stone into a tube marked with one color resulted in the rise of the floating food in the middle tube. None of the jays solved this task, indicating that a confusing causal relationship appears to inhibit the learning of a simple associative rule: ‘drop the stone in the red tube’. All New Caledonian crows [13] and all 5-year-old children also fail this task, though all children 7 years and older solve it [16]. One explanation for the bird's failure is that the distance between the two large tubes was smaller than in the other tasks, which might have interfered with successful discrimination [19]. In other water tube discrimination experiments that involve two tubes, such as the sand vs. water experiment, the tubes are positioned 30 cm apart [13], [15]–[18]. In the U-tube experiment, the two large tubes are only 15 cm apart [13], [16], [18]. It could be that this reduction in distance made it more difficult for the birds to inhibit switching between tubes and so prevented discrimination. A further issue with the current U-tube methodology is that, even if subjects succeed, such as children 7 to 10 years of age, there are two competing hypotheses for their performance. One possibility is that they have inferred the presence of a hidden causal mechanism, namely the connection between the two tubes. Another possibility is that the subjects are highly sensitive to perceptual feedback [16], [18], [20]–[21], and thus notice the effect stone dropping has on the adjacent small tube. An additional limitation of previous studies concerns the ability of crows to make functional object discriminations. While current results suggest the crows make discriminations based on the causal properties of the objects involved, an alternate possibility is that they discriminate simply because they prefer to handle one object type over another. In particular, preferences could be driven by familiarity. For example, crows could prefer to handle solid rather than hollow objects simply because solid objects look more similar to the stones they have previously dropped in the tube and encountered in the wild. While simple preferences for approaching previously rewarded objects were controlled for in one study [17], no study has so far examined if subjects have a preference for handling certain objects. Given the history of the trap-tube task, where small modifications to the procedure such as letting subjects pull rather than push the food leads to large behavioral differences [7], it is critical that the effect of modifications to the Aesop's Fable task be explored. Here, we use the Aesop's Fable paradigm to test causal cognition in New Caledonian crows. This species may have sophisticated causal cognition in the wild since it makes and uses tools to extract hidden food [22] and appears to make inferences about hidden causal agents [23]. We addressed the limitations of the U-tube and object preference tasks outlined above. First, we replicated the object and substrate experiments in Jelbert and colleagues [13] to ensure that our results were consistent with results from previous studies on New Caledonian crows. We also replicated Jelbert and colleagues' [13] volume discrimination experiment. However, instead of providing the crows with 12 objects, which allowed individuals to obtain food from either tube if they persisted long enough, we provided the crows with only enough objects (four) to correctly solve the task by choosing the more efficient tube. We then addressed the methodological limitation of the U-tube experiment by increasing the distance between large tubes, such that it was identical to the distance used in standard tube discrimination experiments. Additionally, we gave each wide tube its own narrow tube to make the distinction between the two sets of tubes clear. To investigate whether object preferences are actually related to the functionality of the task, we presented crows with hollow and solid objects of the same weight when the task required an object with a particular weight, rather than a particular displaceable volume. Finally, to test whether New Caledonian crows make inferences about a hidden causal mechanism, we designed a novel uncovered U-tube task to be given after subjects had attempted to solve the original U-tube task. We removed all arbitrary cues and exposed the hidden mechanism, the connecting pipe between two of the tubes. If subjects successfully solved the original covered U-tube task by inferring that the large and small tube were connected, we expected them to subsequently choose the visibly connected large tube rather than the unconnected one. However, if they had simply learned the association between the arbitrary color cue and the movement of the food, then their learning should be disrupted by the removal of this cue.

Discussion As in past studies, New Caledonian crows preferred to drop stones into water-filled, rather than sand-filled tubes, and they preferred to drop objects that sank or were solid, rather than floating or hollow objects [13], [17]. However, while past work suggested that New Caledonian crows cannot discriminate between water-filled tubes of different volumes [13], here three of the six birds tested not only preferred to drop stones into a narrow tube, but then switched their preference to a wide tube when the water level in this tube was significantly higher. This switch from the narrow to the wide tube shows that the success in the equal water level condition is not simply due to these three crows having a preference for narrow tubes. While previous work has suggested that New Caledonian crows [13], like Eurasian jays [18], and children under 7 years of age [16], cannot solve the U-tube task, here one crow, Kitty, was able to do so by learning the connection between the arbitrary color cue and the movement of the food. However, all crows, including Kitty, subsequently failed our novel U-tube task with the exposed connection, which suggests that Kitty's success on the colored U-tube was due to the color cue association and not to the inference of a hidden causal mechanism or attention to the movement of the food. Crows also showed a preference for solid, rather than hollow, objects of the same weight when presented with a task not involving water, which suggests that they may have a bias for solid objects independent of the task at hand. In first trials, significantly more correct choices were made in the substrate and object discrimination experiments, but not in the narrow vs. wide equal water levels, colored U-tube, and uncovered U-tube experiments. The crows' lack of an initial preference for particular tubes or objects during habituation suggests that first trial performance was not based on a bias to approach specific objects. In the experiments where first choices on first trials were correct, aside from the solid vs hollow condition, crows showed a learning effect in the first 10 trials, with performance improving in trials 11-20. This suggests that, while they had high overall levels of success across 20 trials, the crows needed experience with the objects, substrates, and tubes to settle on the functional option. These results are consistent with previous water tube experiments where rooks, Eurasian jays, and other New Caledonian crows showed some learning effects across the first five trials [13], [15], [17]–[18]. While our results show that the crows did not have a bias to approach objects to gain food, they may have had a bias towards the type of objects they preferred to pick up, given that all three birds without water tube experience chose solid rather than hollow objects of the same weight when presented with a platform apparatus. This finding calls into question the previous object discrimination choices of children [16], jays [18], and New Caledonian crows [13], [21]. The possibility that, irrespective of the task at hand, subjects have preferences for particular objects (i.e., solid and sinking objects), particularly those that look more like objects they have previous experience with (i.e., stones), has not been sufficiently well controlled for in past studies. Interestingly, it was only in the solid vs. hollow condition that no learning effect was shown, which suggests that the crows may have had a specific bias for solid objects, rather than a more general bias for the functional objects in our experiments. However, further work is required to test this hypothesis. It is also important to note that a handling preference has not yet been controlled for in children [16], and it may explain children's performances to date. Our results differ from previous findings where no crows passed the wide vs. narrow equal water level experiment [13]. It appears that providing enough objects (four rather than twelve) for the bird to succeed only in the narrow tube, but not in both of the tubes, provided the motivation to choose the more efficient tube. That three birds then preferred a wide tube when the water level was unequal shows that these crows made discriminations based on water volume because they switched their preference from the narrow to the wide tube, indicating that they did not simply prefer to drop objects into the narrow tube. This finding raises two possibilities. The first is that crows are able to imagine changes in the magnitude of a causal relation. The crows only had prior experience dropping stones into a tube of one volume to raise its water level. They could have used their knowledge of the relationship between this tube size, the object, and the water level to mentally model the effect on the water level of dropping the same object into a bigger or smaller tube. This would have led them to choose the narrow tube. In contrast to this high level explanation, an alternate possibility is that the crows attended to differences in the feedback generated by dropping objects into the narrow tube (the floating reward moved significantly closer to the top of the tube) and the wide tube (minimal movement occurred). Further testing is required, using a tube where visual feedback is not available, to distinguish between these causal relation and feedback explanations The performance of Kitty in solving our modified U-tube apparatus raises the possibility that past failures by corvids [13], [18] may be due to problems with tube discrimination when the stone dropping tubes are close together. Kitty's successful solution of this problem suggests two hypotheses. Like 42% of children between ages 4 and 10 [16], Kitty may have inferred the presence of a hidden causal mechanism linking the two tubes of the U-tube task. Alternatively, consistent with the most likely explanation for children's success on this task, she may have paid more attention to feedback than the other crows: by carefully watching the effect of a stone drop into each tube she may have noticed which tube caused the water level in the small tube to rise. However, these hypotheses both predict that Kitty should have passed both U-tube experiments. Though Kitty chose the correct tube on the first trial of the uncovered U-tube experiment, she did not consistently choose the connected tube. Thus, rather than inferring the presence of a hidden causal mechanism or attending to feedback, Kitty may have instead associated the color cue with the eventual receipt of the food. Further testing is required, given that children have not yet been tested on the novel uncovered U-tube task used here. Past research has shown the pitfalls of assuming how humans will react when faced with causal problems [29]. Despite the last 20 years of work on what animals understand about their physical world, we still have a rudimentary understanding of the cognition behind how animals solve problems. One of the key issues in the past has been the lack of rigorous testing of the experimental methodologies themselves, particularly the relation between small changes in procedure and animals' performance, which has resulted in the formation of strong conclusions about issues, such as human uniqueness, before the hypotheses under examination have been thoroughly tested [2], [6]–[7], [10]. The results here highlight a number of methodological limitations in the Aesop's Fable paradigm. With appropriate modifications to this paradigm, it should be possible to overcome these issues in future studies and so produce useful results for comparative cognition research. Comparisons will be particularly interesting between corvids and apes as they will provide insights into how the very differently structured bird and mammalian brains understand the causality of the world.

Acknowledgments We thank Joe Jablonski for machining the acrylic tubes, Gavin Hunt for field support, and three anonymous reviewers for helpful feedback on the manuscript.

Author Contributions Conceived and designed the experiments: CJL SAJ RDG AHT. Performed the experiments: CJL AJB. Analyzed the data: CJL. Contributed reagents/materials/analysis tools: CJL SAJ RDG AHT. Wrote the paper: CJL AHT. Edited the manuscript: SAJ AJB RDG.