Tubes, tables and traps: great apes solve two functionally equivalent trap tasks but show no evidence of transfer across tasks

Technical reasoning is important for cumulative technological culture

Human technology has evolved in an unparalleled way, allowing us to expand across the globe. One fascinating question is, how do we understand the cognitive origins of this phenomenon, which is known as cumulative technological culture (CTC)? The dominant view posits that CTC results from our unique ability to learn from each other. The cultural niche hypothesis even minimizes the involvement of non-social cognitive skills in the emergence of CTC, claiming that technologies can be optimized without us understanding how they work, but simply through the retention of small improvements over generations. Here we conduct a partial replication of the experimental study of Derex et al. (Nature Human Behaviour, 2019) and show that the improvement of a physical system over generations is accompanied by an increased understanding of it. These findings indicate that technical-reasoning skills (non-social cognitive skills) are important in the acquisition, understanding and improvement of technical content-that is, specific to the technological form of cumulative culture-thereby making social learning a salient source of technical inspiration.

Learning versus reasoning to use tools in children

Tool behavior might be based on two strategies associated with specific cognitive mechanisms: cued-learning and technical-reasoning strategies. We aimed to explore whether these strategies coexist in young children and whether they are manifest differently through development. We presented 216 3- to 9-year-olds with a vertical maze task consisting in moving a ball from the top to the bottom of a maze. Two tool-use/mechanical actions were possible: rotating action and sliding action. Three conditions were tested, each focused on a different strategy. In the Opaque-Cue condition (cued-learning strategy), children could not see the mechanical action of each tool. Nevertheless, a cue was provided according to the tool needed to solve the problem. In the Transparent-No Cue condition (technical-reasoning strategy), no cue was presented. However, children could see the mechanical actions associated with each tool. In the Transparent-Cue condition (cued-learning and/or technical-reasoning strategies) children saw both the mechanical actions and the cues. Results indicated that the Opaque-Cue and Transparent-Cue conditions were easier than the Transparent-No-Cue condition in all children. These findings stress that children can use either cued learning or technical reasoning to use tools, according to the available information. The behavioral pattern observed in the Transparent-Cue condition suggests that children might be inclined to use technical reasoning even when the task can be solved through cued learning.

Causal Reasoning and Event Cognition as Evolutionary Determinants of Language Structure

The aim of this article is to provide an evolutionarily grounded explanation of central aspects of the structure of language. It begins with an account of the evolution of human causal reasoning. A comparison between humans and non-human primates suggests that human causal cognition is based on reasoning about the underlying forces that are involved in events, while other primates hardly understand external forces. This is illustrated by an analysis of the causal cognition required for early hominin tool use. Second, the thinking concerning forces in causation is used to motivate a model of human event cognition. A mental representation of an event contains two vectors representing a cause as well as a result but also entities such as agents, patients, instruments and locations. The fundamental connection between event representations and language is that declarative sentences express events (or states). The event structure also explains why sentences are constituted of noun phrases and verb phrases. Finally, the components of the event representation show up in language, where causes and effects are expressed by verbs, agents and patients by nouns (modified by adjectives), locations by prepositions, etc. Thus, the evolution of the complexity of mental event representations also provides insight into the evolution of the structure of language.

Do wild raccoons (Procyon lotor) use tools?

Being able to make and use tools was once considered to be an evolutionary hallmark of our species, but has since been documented in other animals. However, for reasons that remain unclear, not all species naturally use tools. Racoons (Procyon lotor) are generalist carnivores that possess many of the physical, cognitive, and behavioural characteristics linked to tool use in other species (e.g. manual dexterity, tactile exploration, relatively large brains, extractive foraging, and sociality). Although raccoons have not been observed using tools outside of experimental captive conditions, wild data involving objective psychometric tests are needed. The current study administered a tool-related task to a wild population of raccoons from 20 locations within the Croatan National Forest, USA. The task required participants to use a stick to extract food from a pipe. To facilitate interpretations of their performances on the task, data were obtained on natural tool availability at the field site and participants' mode of exploring the novel task. None of the participants solved the task despite natural sticks (suitable for solving the task) being widely available across testing locations. Participants were equally likely to smell versus handle novel sticks, which were provided at testing platforms. Limited tactile exploration, but not tool availability, could be at least one factor that reduces these raccoons' opportunities to interact with and learn about novel tools like sticks.

Innovative problem solving in macaws

Behavioural innovations with tool-like objects in non-habitually tool-using species are thought to require complex physical understanding, but the underlying cognitive processes remain poorly understood. A few parrot species are capable of innovating tool-use and borderline tool-use behaviours. We tested this capacity in two species of macaw (Ara ambiguus, n = 9; Ara glaucogularis, n = 8) to investigate if they could solve a problem-solving task through manufacture of a multi-stone construction. Specifically, after having functional experience with a pre-inserted stick tool to push a reward out of a horizontal tube, the subjects were required to insert five stones consecutively from one side to perform the same function as the stick tool with the resulting multi-component construction. One Ara glaucogularis solved the task and innovated the stone construction after the experience with the stick tool. Two more subjects (one of each species) did so after having further functional experience of a single stone pushing a reward out of a shortened tube. These subjects were able to consistently solve the task, but often made errors, for example counter-productive stone insertions from the opposing end, even in some of the successful trials. Conversely, multiple trials without errors also suggested a strong goal direction. Their performance in the follow-up tasks was inconclusive since they sometimes inserted stones into un-baited or blocked ‘dummy tubes’, but this could have been an attention-deficit behaviour as subjects had not encountered these ‘dummy tubes’ before. Overall, the successful subjects’ performance was so erratic that it proved difficult to conclude whether they had functional understanding of their multi-stone constructions.

Neural Processes Underlying Tool Use in Humans, Macaques, and Corvids

It was thought that tool use in animals is an adaptive specialization. Recent studies, however, have shown that some non-tool-users, such as rooks and jays, can use and manufacture tools in laboratory settings. Despite the abundant evidence of tool use in corvids, little is known about the neural mechanisms underlying tool use in this family of birds. This review summarizes the current knowledge on the neural processes underlying tool use in humans, macaques and corvids. We suggest a possible neural network for tool use in macaques and hope this might inspire research to discover a similar brain network in corvids. We hope to establish a framework to elucidate the neural mechanisms that supported the convergent evolution of tool use in birds and mammals.

Trialling Meta-Research in Comparative Cognition: Claims and Statistical Inference in Animal Physical Cognition

Scientific disciplines face concerns about replicability and statistical inference, and these concerns are also relevant in animal cognition research. This paper presents a first attempt to assess how researchers make and publish claims about animal physical cognition, and the statistical inferences they use to support them. We surveyed 116 published experiments from 63 papers on physical cognition, covering 43 different species. The most common tasks in our sample were trap-tube tasks (14 papers), other tool use tasks (13 papers), means-end understanding and string-pulling tasks (11 papers), object choice and object permanence tasks (9 papers) and access tasks (5 papers). This sample is not representative of the full scope of physical cognition research; however, it does provide data on the types of statistical design and publication decisions researchers have adopted. Across the 116 experiments, the median sample size was 7. Depending on the definitions we used, we estimated that between 44% and 59% of our sample of papers made positive claims about animals' physical cognitive abilities, between 24% and 46% made inconclusive claims, and between 10% and 17% made negative claims. Several failures of animals to pass physical cognition tasks were reported. Although our measures had low inter-observer reliability, these findings show that negative results can and have been published in the field. However, publication bias is still present, and consistent with this, we observed a drop in the frequency of p-values above .05. This suggests that some non-significant results have not been published. More promisingly, we found that researchers are likely making many correct statistical inferences at the individual-level. The strength of evidence of statistical effects at the group-level was weaker, and its p-value distribution was consistent with some effect sizes being overestimated. Studies such as ours can form part of a wider investigation into statistical reliability in comparative cognition. However, future work should focus on developing the validity and reliability of the measurements they use, and we offer some starting points.

Great apes selectively retrieve relevant memories to guide action

Memory allows us to draw on past experiences to inform behaviour in the present. However, memories rarely match the situation at hand exactly, and new situations regularly trigger multiple related memories where only some are relevant to act upon. The flexibility of human memory systems is largely attributed to the ability to disregard irrelevant, but salient, memories in favour of relevant ones. This is considered an expression of an executive function responsible for suppressing irrelevant memories, associated with the prefrontal cortex. It is unclear to what extent animals have access to this ability. Here, we demonstrate, in a series of tool-use tasks designed to evoke conflicting memories, that chimpanzees and an orangutan suffer from this conflict but overcome it in favour of a more relevant memory. Such mnemonic flexibility is among the most advanced expressions of executive function shown in animals to date and might explain several behaviours related to tool-use, innovation, planning and more.

Events and Causal Mappings Modeled in Conceptual Spaces

The aim of the article is to present a model of causal relations that is based on what is known about human causal reasoning and that forms guidelines for implementations in robots. I argue for two theses concerning human cognition. The first is that human causal cognition, in contrast to that of other animals, is based on the understanding of the forces that are involved. The second thesis is that humans think about causality in terms of events. I present a two-vector model of events, developed by Gärdenfors and Warglien, which states that an event is represented in terms of two main components - the force of an action that drives the event, and the result of its application. Apart from the causal mapping, the event model contains representations of a patient, an agent, and possibly some other roles. Agents and patients are objects (animate or inanimate) that have different properties. Following my theory of conceptual spaces, they can be described as vectors of property values. At least two spaces are needed to describe an event, an action space and a result space. The result of an event is modeled as a vector representing the change of properties of the patient before and after the event. In robotics the focus has been on describing results. The proposed model also includes the causal part of events, typically described as an action. A central part of an event category is the mapping from actions to results. This mapping contains the central information about causal relations. In applications of the two-vector model, the central problem is how the event mapping can be learned in a way that is amenable to implementations in robots. Three processes are central for event cognition: causal thinking, control of action and learning by generalization. Although it is not yet clear which is the best way to model how the mappings can be learned, they should be constrained by three corresponding mathematical properties: monotonicity (related to qualitative causal thinking); continuity (plays a key role in activities of action control); and convexity (facilitates generalization and the categorization of events). I argue that Bayesian models are not suitable for these purposes, but some more geometrically oriented approach to event mappings should be used.

The elephant in the room: What matters cognitively in cumulative technological culture

Cumulative technological culture (CTC) refers to the increase in the efficiency and complexity of tools and techniques in human populations over generations. A fascinating question is to understand the cognitive origins of this phenomenon. Because CTC is definitely a social phenomenon, most accounts have suggested a series of cognitive mechanisms oriented toward the social dimension (e.g., teaching, imitation, theory of mind, and metacognition), thereby minimizing the technical dimension and the potential influence of non-social, cognitive skills. What if we have failed to see the elephant in the room? What if social cognitive mechanisms were only catalyzing factors and not the sufficient and necessary conditions for the emergence of CTC? In this article, we offer an alternative, unified cognitive approach to this phenomenon by assuming that CTC originates in non-social cognitive skills, namely technical-reasoning skills which enable humans to develop the technical potential necessary to constantly acquire and improve technical information. This leads us to discuss how theory of mind and metacognition, in concert with technical reasoning, can help boost CTC. The cognitive approach developed here opens up promising new avenues for reinterpreting classical issues (e.g., innovation, emulation vs. imitation, social vs. asocial learning, cooperation, teaching, and overimitation) in a field that has so far been largely dominated by other disciplines, such as evolutionary biology, mathematics, anthropology, archeology, economics, and philosophy.

Inhibitory control and memory in the search process for a modified problem in grey squirrels, Sciurus carolinensis

Inhibiting learned behaviours when they become unproductive and searching for an alternative solution to solve a familiar but different problem are two indicators of flexibility in problem solving. A wide range of animals show these tendencies spontaneously, but what kind of search process is at play behind their problem-solving success? Here, we investigated how Eastern grey squirrels, Sciurus carolinensis, solved a modified mechanical problem that required them to abandon their preferred and learned solution and search for alternative solutions to retrieve out-of-reach food rewards. Squirrels could solve the problem by engaging in either an exhaustive search (i.e., using trial-and-error to access the reward) or a 'backup' solution search (i.e., recalling a previously successful but non-preferred solution). We found that all squirrels successfully solved the modified problem on their first trial and showed solving durations comparable to their last experience of using their preferred solution. Their success and high efficiency could be explained by their high level of inhibitory control as the squirrels did not persistently emit the learned and preferred, but now ineffective, pushing behaviour. Although the squirrels had minimal experience in using the alternative (non-preferred) successful solution, they used it directly or after one or two failed attempts to achieve success. Thus, the squirrels were using the 'backup' solution search process. Such a process is likely a form of generalisation which involves retrieving related information of an experienced problem and applying previous successful experience during problem solving. Overall, our results provide information regarding the search process underlying the flexibility observable in problem-solving success.

Orangutans (Pongo abelii) make flexible decisions relative to reward quality and tool functionality in a multi-dimensional tool-use task

Making economic decisions in a natural foraging situation that involves the use of tools may require an animal to consider more levels of relational complexity than merely deciding between an immediate and a delayed food option. We used the same method previously used with Goffin´s cockatoos to investigate the orangutans' flexibility for making the most profitable decisions when confronted with five different settings that included one or two different apparatuses, two different tools and two food items (one more preferred than the other). We found that orangutans made profitable decisions relative to reward quality, when the task required the subjects to select a tool over an immediately accessible food reward. Furthermore, most subjects were sensitive to work-effort when the immediate and the delayed option (directly accessible by using a tool) led to the same outcome. Most subjects continued to make profitable decisions that required taking into account the tool functionality. In a final multidimensional task design in which subjects had to simultaneously focus on two apparatuses, two reward qualities and two different tools, the orangutans chose the functional tool to access the high quality reward.

Two macaw species can learn to solve an optimised two-trap problem, but without functional causal understanding

The trap-tube paradigm is a useful reference for judging whether a species is likely to use advanced physical causal cognition, however it does not have a standardised format. In this study, the design of an optimised two trap-table is described and is then tested on two species of macaw: Ara ambiguus and Ara glaucogularis. Multiple subjects of both species learned a successful method to solve an initial trap-problem and some transferred this success to other apparatus presented. However this transfer was likely achieved without a functional physical understanding of the task. The macaws probably have a preference to use learned rules based on arbitrary properties to solve the trap-problem. We conclude that this setup of the two-trap-problem is a viable benchmark that could be administered to a variety of species with very little modification, thus paving the way for more directly comparative studies.

Cognitive Paleoanthropology and Technology: Toward a Parsimonious Theory (PATH)

Tool use in humans and hominins (i.e., extant relatives to humans) is unique in several respects. To date, no attempt has been made to review the main patterns of tool behavior specific to these species as well as to integrate them into a coherent framework. The aim here is to fill this gap by (a) identifying these behavioral specificities and (b) trying to explain the greatest number of these specificities with the lowest number of cognitive mechanisms. Based on this approach, this article provides a potential solution, namely, the PArsimonious THeory of hominin technology (PATH), aiming to account for the cognitive origins of 4 behavioral characteristics: transfer, complex tool use, secondary tool use, and tool saving. A key hypothesis is that the emergence of 2 breaking mechanisms—technical reasoning and semantic reasoning—could have boosted hominin technology. PATH offers an original framework for understanding the most archaic, human cognitive traits, thereby providing a good starting point for future investigation about the cognitive evolution of technology in the genus Homo.

Technical intelligence and culture: Nut cracking in humans and chimpanzees

OBJECTIVES

According to the technical intelligence hypothesis, humans are superior to all other animal species in understanding and using tools. However, the vast majority of comparative studies between humans and chimpanzees, both proficient tool users, have not controlled for the effects of age, prior knowledge, past experience, rearing conditions, or differences in experimental procedures. We tested whether humans are superior to chimpanzees in selecting better tools, using them more dexteriously, achieving higher performance and gaining access to more resource as predicted under the technical intelligence hypothesis.

MATERIALS AND METHODS

Aka and Mbendjele hunter-gatherers in the rainforest of Central African Republic and the Republic of Congo, respectively, and Taï chimpanzees in the rainforest of Côte d'Ivoire were observed cracking hard Panda oleosa nuts with different tools, as well as the soft Coula edulis and Elaeis guinensis nuts. The nut-cracking techniques, hammer material selection and two efficiency measures were compared.

RESULTS

As predicted, the Aka and the Mbendjele were able to exploit more species of hard nuts in the forest than chimpanzees. However, the chimpanzees were sometimes more efficient than the humans. Social roles differed between the two species, with the Aka and especially the Mbendjele exhibiting cooperation between nut-crackers whereas the chimpanzees were mainly individualistic.

DISCUSSION

Observations of nut-cracking by humans and chimpanzees only partially supported the technical intelligence hypothesis as higher degrees of flexibility in tool selection seen in chimpanzees compensated for use of less efficient tool material than in humans. Nut cracking was a stronger social undertaking in humans than in chimpanzees.

Time Does Not Help Orangutans Pongo abelii Solve Physical Problems

Many questions in animal intelligence and cognition research are challenging. One challenge is to identify mechanisms underlying reasoning in experiments. Here, we provide a way to design such tests in non-human animals. We know from research in skill acquisition in humans that reasoning and thinking can take time because some problems are processed in multiple steps before a solution is reached (e.g., during mental arithmetics). If animals are able to learn through similar processes their decision making can be time consuming, and most importantly improve if more time to process information is allowed. We tested if performance of two Sumatran orangutans (Pongo abelii) increased in a two-choice experiment when they were allowed extra time before making their decisions, compared to when they were forced to decide immediately. We found that the performance of the orangutans did not depend on the time they were allowed to process the information before making their decisions. This methodology provides a potential avenue for empirical tests of mechanisms underlying reasoning in non-human animals.

Transfer of physical understanding in a non-tool-using parrot

Physical cognition has generally been assessed in tool-using species that possess a relatively large brain size, such as corvids and apes. Parrots, like corvids and apes, also have large relative brain sizes, yet although parrots rarely use tools in the wild, growing evidence suggests comparable performances on physical cognition tasks. It is, however, unclear whether success on such tasks is facilitated by previous experience and training procedures. We therefore investigated physical comprehension of object relationships in two non-tool-using species of captive neotropical parrots on a new means-end paradigm, the Trap-Gaps task, using unfamiliar materials and modified training procedures that precluded procedural cues. Red-shouldered macaws (Diopsittaca nobilis) and black-headed caiques (Pionites melanocephala) were presented with an initial task that required them to discriminate between pulling food trays through gaps while attending to the respective width of the gaps and size of the trays. Subjects were then presented with a novel, but functionally equivalent, transfer task. Six of eight birds solved the initial task through trial-and-error learning. Four of these six birds solved the transfer task, with one caique demonstrating spontaneous comprehension. These findings suggest that non-tool-using parrots may possess capacities for sophisticated physical cognition by generalising previously learned rules across novel problems.

Robust retention and transfer of tool construction techniques in chimpanzees (Pan troglodytes)

Long-term memory can be critical to a species' survival in environments with seasonal and even longer-term cycles of resource availability. The present, longitudinal study investigated whether complex tool behaviors used to gain an out-of-reach reward, following a hiatus of about 3 years and 7 months since initial experiences with a tool use task, were retained and subsequently executed more quickly by experienced than by naïve chimpanzees. Ten of the 11 retested chimpanzees displayed impressive long-term procedural memory, creating elongated tools using the same methods employed years previously, either combining 2 tools or extending a single tool. The complex tool behaviors were also transferred to a different task context, showing behavioral flexibility. This represents some of the first evidence for appreciable long-term procedural memory, and improvements in the utility of complex tool manufacture in chimpanzees. Such long-term procedural memory and behavioral flexibility have important implications for the longevity and transmission of behavioral traditions.

New Caledonian crows rapidly solve a collaborative problem without cooperative cognition

There is growing comparative evidence that the cognitive bases of cooperation are not unique to humans. However, the selective pressures that lead to the evolution of these mechanisms remain unclear. Here we show that while tool-making New Caledonian crows can produce collaborative behavior, they do not understand the causality of cooperation nor show sensitivity to inequity. Instead, the collaborative behavior produced appears to have been underpinned by the transfer of prior experience. These results suggest that a number of possible selective pressures, including tool manufacture and mobbing behaviours, have not led to the evolution of cooperative cognition in this species. They show that causal cognition can evolve in a domain specific manner–understanding the properties and flexible uses of physical tools does not necessarily enable animals to grasp that a conspecific can be used as a social tool.

Ways of thinking: from crows to children and back again

This article reviews some of the recent work on the remarkable cognitive capacities of food-caching corvids. The focus will be on their ability to think about other minds and other times, and tool-using tests of physical problem solving. Research on developmental cognition suggests that young children do not pass similar tests until they are at least four years of age in the case of the social cognition experiments, and eight years of age in the case of the tasks that tap into physical cognition. This developmental trajectory seems surprising. Intuitively, one might have thought that the social and planning tasks required more complex forms of cognitive process, namely Mental Time Travel and Theory of Mind. Perhaps the fact that children pass these tasks earlier than the physical problem-solving tasks is a reflection of cultural influences. Future research will hope to identify these cognitive milestones by starting to develop tasks that might go some way towards understanding the mechanisms underlying these abilities in both children and corvids, to explore similarities and differences in their ways of thinking.

Modifications to the Aesop's Fable paradigm change New Caledonian crow performances

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.

Using the Aesop's fable paradigm to investigate causal understanding of water displacement by New Caledonian crows

Understanding causal regularities in the world is a key feature of human cognition. However, the extent to which non-human animals are capable of causal understanding is not well understood. Here, we used the Aesop's fable paradigm – in which subjects drop stones into water to raise the water level and obtain an out of reach reward – to assess New Caledonian crows' causal understanding of water displacement. We found that crows preferentially dropped stones into a water-filled tube instead of a sand-filled tube; they dropped sinking objects rather than floating objects; solid objects rather than hollow objects, and they dropped objects into a tube with a high water level rather than a low one. However, they failed two more challenging tasks which required them to attend to the width of the tube, and to counter-intuitive causal cues in a U-shaped apparatus. Our results indicate that New Caledonian crows possess a sophisticated, but incomplete, understanding of the causal properties of displacement, rivalling that of 5–7 year old children.

Transfer of the nonmatch-to-goal rule in monkeys across cognitive domains

To solve novel problems, it is advantageous to abstract relevant information from past experience to transfer on related problems. To study whether macaque monkeys were able to transfer an abstract rule across cognitive domains, we trained two monkeys on a nonmatch-to-goal (NMTG) task. In the object version of the task (O-NMTG), the monkeys were required to choose between two object-like stimuli, which differed either only in shape or in shape and color. For each choice, they were required to switch from their previously chosen object-goal to a different one. After they reached a performance level of over 90% correct on the O-NMTG task, the monkeys were tested for rule transfer on a spatial version of the task (S-NMTG). To receive a reward, the monkeys had to switch from their previously chosen location to a different one. In both the O-NMTG and S-NMTG tasks, there were four potential choices, presented in pairs from trial-to-trial. We found that both monkeys transferred successfully the NMTG rule within the first testing session, showing effective transfer of the learned rule between two cognitive domains.

Memory for distant past events in chimpanzees and orangutans

Determining the memory systems that support nonhuman animals' capacity to remember distant past events is currently the focus an intense research effort and a lively debate [1-3]. Comparative psychology has largely adopted Tulving's framework by focusing on whether animals remember what-where-when something happened (i.e., episodic-like memory) [4-6]. However, apes have also been reported to recall other episodic components [7] after single-trial exposures [8, 9]. Using a new experimental paradigm we show that chimpanzees and orangutans recalled a tool-finding event that happened four times 3 years earlier (experiment 1) and a tool-finding unique event that happened once 2 weeks earlier (experiment 2). Subjects were able to distinguish these events from other tool-finding events, which indicates binding of relevant temporal-spatial components. Like in human involuntary autobiographical memory, a cued, associative retrieval process triggered apes' memories: when presented with a particular setup, subjects instantaneously remembered not only where to search for the tools (experiment 1), but also the location of the tool seen only once (experiment 2). The complex nature of the events retrieved, the unexpected and fast retrieval, the long retention intervals involved, and the detection of binding strongly suggest that chimpanzees and orangutans' memories for past events mirror some of the features of human autobiographical memory.

Effect of a cognitive challenge device containing food and non-food rewards on chimpanzee well-being

Exploration and problem-solving are highly motivated behaviors in non-human primates, but little research has focused on whether cognitively challenging tasks can enhance primates' psychological well-being, particularly in the absence of food rewards. We evaluated whether a novel cognitive challenge device (CCD) consisting of a maze of opaque tubes enhanced the well-being of a group of six adult chimpanzees housed at ZSL Whipsnade Zoo, UK, over a two-month period. Chimpanzees had the opportunity to interact with two versions of the CCD: the first contained tokens which fell into a transparent chamber when extracted from the CCD and could not be eaten. The second contained unshelled Brazil nuts, which could be extracted and eaten. CCD-use was low over the study, occupying on average 2.5% of observation time. However, compared to baseline levels, chimpanzees exhibited more problem-solving behaviors (directed toward the CCD) and spent significantly more time engaged in social play when the CCD was present. Cage exploration was rare whether the CCD was present or not. Chimpanzees used the CCD (including tool-use) significantly more when it contained tokens. The relationship between the presence of the CCD and self-directed behavior (rough-scratching) was difficult to interpret. Although rough-scratching was significantly higher in the cage when the CCD was present and 18% of these scratching events occurred within one arm's length from the CCD, rough-scratching decreased when device use increased. This study provides a preliminary investigation of the CCD and two reward types, and suggests how the design could be modified to enhance its effects.

Causal reasoning in New Caledonian crows: Ruling out spatial analogies and sampling error

A large number of studies have failed to find conclusive evidence for causal reasoning in nonhuman animals. For example, when animals are required to avoid a trap while extracting a reward from a tube they appear to learn about the surface-level features of the task, rather than about the task's causal regularities. We recently reported that New Caledonian crows solved a two-trap-tube task and then were able to immediately solve a novel, visually distinct problem, the trap-table task. Such transfer suggests these crows were reasoning causally. However, there are two other possible explanations for the successful transfer: sampling bias and the use of a spatial, rather than a causal, analogy. Here we present data that rule out these explanations.

The cognitive bases of human tool use

This article has two goals. The first is to assess, in the face of accruing reports on the ingenuity of great ape tool use, whether and in what sense human tool use still evidences unique, higher cognitive ability. To that effect, I offer a systematic comparison between humans and nonhuman primates with respect to nine cognitive capacities deemed crucial to tool use: enhanced hand-eye coordination, body schema plasticity, causal reasoning, function representation, executive control, social learning, teaching, social intelligence, and language. Since striking differences between humans and great apes stand firm in eight out of nine of these domains, I conclude that human tool use still marks a major cognitive discontinuity between us and our closest relatives. As a second goal of the paper, I address the evolution of human technologies. In particular, I show how the cognitive traits reviewed help to explain why technological accumulation evolved so markedly in humans, and so modestly in apes.

Barriers and traps: great apes' performance in two functionally equivalent tasks

Tool-using tasks that require subjects to overcome the obstacles to get a reward have been a major component of research investigating causal knowledge in primates. Much of the debate in this research has focused on whether subjects simply use certain stimulus features or instead use more functionally relevant information regarding the effect that certain features may have on a moving reward. Here, we presented two obstacle tasks, a trap platform and a barrier platform, to 22 great apes. Although perceptually similar, these two tasks contain two perceptually different but functionally equivalent obstacles: a trap and a barrier. In a pre-exposure phase, subjects either experienced an obstacle task or a task without any obstacle. In the transfer phase, all subjects were presented with an obstacle task, either the trap platform or the barrier platform. Our results show that those subjects who received an obstacle task prior to the second task performed better than those who first received a non-obstacle task. The type of obstacle task that subjects received first did not have any effect on their performance in the transfer phase. We suggest that apes possess some knowledge about the effects that obstacles have on slow-moving unsupported objects.

Understanding the functional properties of tools: chimpanzees (Pan troglodytes) and capuchin monkeys (Cebus apella) attend to tool features differently

We examined whether eight capuchins and eight chimpanzees were able to retrieve a reward placed inside a tube, of varying length, by selecting the correct stick from different sets of three sticks differing in length (functional feature) and handle (non-functional feature). Moreover, to investigate whether seeing the stick inside the tube (visual feedback) improves performance, half of the subjects were tested with a transparent apparatus and the other half with an opaque apparatus. Phase 1 included (a) Training 1 in which each stick had a different handle and (b) Transfer 1 in which the handles were switched among sticks, so that the functional tool had the same length but a different handle than before. The seven chimpanzees and one capuchin that passed Transfer 1 received Transfer 2. The other subjects received (a) Training 2, which used the same sticks from Phase 1 with handles switched in every trial, and (b) Transfer 2 in which the tube was longer, all sticks had the same new handle, and the formerly longest tool became intermediate in length. Eight chimpanzees and three capuchins passed Transfer 2. Results showed that (1) chimpanzees applied relational structures in tool using tasks more quickly than capuchins and (2) capuchins required more varied experience to attend to the functional feature of the tool. Interestingly, visual feedback did not improve performance in either species.

Two strings to choose from: do ravens pull the easier one?

There are simple co-occurrences as well as functional relationships between events. One may assume that animals detect and use causation rather than mere co-variation. However, understanding causation often requires concepts of hidden forces. In string pulling, obstacles may hamper the access to food. Here, I studied whether ravens have an abstract concept of effort. First, in a competitive situation, ravens (Corvus corax) could choose one out of two strings. The strings differed in whether they were baited with meat and in how far away the meat was. Ravens pulled mainly the string containing meat and where the meat was nearer to the perch, respectively. Second, ravens could choose between two strings that had either a functional obstacle or a non-functional obstacle. Optimal performance required the integration of at least two cues: object and height. In 5 ravens, the model that best matched behaviour took into account only that meat was on a string, ignoring the obstacle. However, 2 ravens' performance was best explained by a model that took into account both an object's identity (meat or wood) and its height on the string. Third, one string out of two was loaded with a heavy meat piece. In this overloaded string condition, 5 out of 7 ravens did not try to pull the heavy meat piece but went straight for pulling the smaller piece. The pattern of results indicated that ravens can judge the effort required to pull a string.

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

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.

Comparing the performances of apes (Gorilla gorilla, Pan troglodytes, Pongo pygmaeus) and human children (Homo sapiens) in the floating peanut task

Recently, Mendes et al. [1] described the use of a liquid tool (water) in captive orangutans. Here, we tested chimpanzees and gorillas for the first time with the same "floating peanut task." None of the subjects solved the task. In order to better understand the cognitive demands of the task, we further tested other populations of chimpanzees and orangutans with the variation of the peanut initially floating or not. Twenty percent of the chimpanzees but none of the orangutans were successful. Additional controls revealed that successful subjects added water only if it was necessary to obtain the nut. Another experiment was conducted to investigate the reason for the differences in performance between the unsuccessful (Experiment 1) and the successful (Experiment 2) chimpanzee populations. We found suggestive evidence for the view that functional fixedness might have impaired the chimpanzees' strategies in the first experiment. Finally, we tested how human children of different age classes perform in an analogous experimental setting. Within the oldest group (8 years), 58 percent of the children solved the problem, whereas in the youngest group (4 years), only 8 percent were able to find the solution.

Physical cognition and tool-use: performance of Darwin's finches in the two-trap tube task

The trap tube is a classic test of causal reasoning abilities in animals in the physical domain. Recently, a modified version of this task improved its diagnostic capacity and allowed testing of non-tool-using animals. We used this modified two-trap tube task to compare the cognition of two Darwin's finch species: the woodpecker finch, Cactospiza pallida, a tool-using species, and the small tree finch, Camarhynchus parvulus, a closely related non-tool-using species. Not all woodpecker finches use tools in nature, and we therefore also tested non-tool-using individuals to assess the effect of experience on trap tube performance. No small tree finches and only two non-tool-using woodpecker finches solved the initial task which was operated using a pre-inserted piston. One tool-using woodpecker finch solved the task when allowed to use its own tool instead of the pre-inserted piston. The fact that none of these subjects transferred their knowledge when the features of the task changed, suggests that in this species, neither experience using tools nor the genetic composition of a tool-user are associated with the general physical cognitive skills required to solve the trap tube task.

Spontaneous use of tools as straws in great apes

Great apes can use multiple tools to extract food embedded in substrates and can invent new ways to exploit those resources. We tested five bonobos, five chimpanzees, and six orangutans in a task in which they had to use (and modify) a tool as a straw to drink the juice located inside a container. Experiment 1 showed that four orangutans and one chimpanzee invented the use of a piece of electric cable to get the juice. Experiment 2 investigated whether subjects could transform a non-functional hose into a functional one by removing blockages that impeded the free flow of juice. Orangutans outperformed chimpanzees and bonobos by differentially removing those blockages that prevented the flow of juice, often doing so before attempting to extract the juice. In Experiment 3, we presented chimpanzees and orangutans with four 3-tool sets (each tool set contained a single straw-like tool) and allowed them to select one tool. Unlike chimpanzees, orangutans succeeded in selecting the straw-like tool above chance levels without having to physically manipulate it. We suggest that orangutans' superior performance is related to their greater reliance on mouth actions during foraging. Experiment 4 investigated whether orangutans were also capable of selecting the suitable tool not by its appearance, but by the effects that it produced. After witnessing the experimenter blow bubbles or absorb liquid with a functional tool but fail to accomplish the same thing with the non-functional tool, orangutans failed to select the functional tool above chance levels.

The representation of tool use in humans and monkeys: common and uniquely human features

Though other species of primates also use tools, humans appear unique in their capacity to understand the causal relationship between tools and the result of their use. In a comparative fMRI study, we scanned a large cohort of human volunteers and untrained monkeys, as well as two monkeys trained to use tools, while they observed hand actions and actions performed using simple tools. In both species, the observation of an action, regardless of how performed, activated occipitotemporal, intraparietal, and ventral premotor cortex, bilaterally. In humans, the observation of actions done with simple tools yielded an additional, specific activation of a rostral sector of the left inferior parietal lobule (IPL). This latter site was considered human-specific, as it was not observed in monkey IPL for any of the tool videos presented, even after monkeys had become proficient in using a rake or pliers through extensive training. In conclusion, while the observation of a grasping hand activated similar regions in humans and monkeys, an additional specific sector of IPL devoted to tool use has evolved in Homo sapiens, although tool-specific neurons might reside in the monkey grasping regions. These results shed new light on the changes of the hominid brain during evolution.

Great apes' performance in discriminating weight and achromatic color

Much work has been done on visual discrimination in primates over the past decade. In contrast, very little is known about the relevance of non-visual information in discrimination learning. We investigated weight and achromatic color (color, henceforth) discrimination in bonobos, gorillas and orangutans, using the exchange paradigm in which subjects have to give objects to the experimenter in order to receive a reward. Unlike previous studies, subjects were not trained to lift objects because lifting the objects was an integral part of the exchange procedure. This methodology also allowed us a direct comparison between visual and weight discrimination. We presented 12 subjects (5 bonobos, 2 gorillas and 5 orangutans) with two sets of objects corresponding to two conditions. The objects in the color condition (white/black) differed only in color and those in the weight condition (light/heavy) differed only in weight. Five apes learned to discriminate weight and six to discriminate color. Subjects learned color discrimination faster than weight discrimination. Our results suggest that bonobos and orangutans are sensitive to differences in weight and able to learn discriminating objects that differ in this property.