Portal de Periódicos da CAPES

Animal Tool-Use

2010; Elsevier BV; Volume: 20; Issue: 23 Linguagem: Inglês

10.1016/j.cub.2010.09.042

1879-0445

Amanda M. Seed, Richard W. Byrne,

Human-Animal Interaction Studies

The sight of an animal making and using a tool captivates scientists and laymen alike, perhaps because it forces us to question some of our ideas about human uniqueness. Does the animal know how the tool works? Did it anticipate the need for the tool and make it in advance? To some, this fascination with tools seems arbitrary and anthropocentric; after all, animals engage in many other complex activities, like nest building, and we know that complex behaviour need not be cognitively demanding. But tool-using behaviour can also provide a powerful window into the minds of living animals, and help us to learn what capacities we share with them — and what might have changed to allow for the incontrovertibly unique levels of technology shown by modern humans. The sight of an animal making and using a tool captivates scientists and laymen alike, perhaps because it forces us to question some of our ideas about human uniqueness. Does the animal know how the tool works? Did it anticipate the need for the tool and make it in advance? To some, this fascination with tools seems arbitrary and anthropocentric; after all, animals engage in many other complex activities, like nest building, and we know that complex behaviour need not be cognitively demanding. But tool-using behaviour can also provide a powerful window into the minds of living animals, and help us to learn what capacities we share with them — and what might have changed to allow for the incontrovertibly unique levels of technology shown by modern humans. The notion of 'man the tool maker', that we alone in the animal kingdom manufacture and use tools, has long been known to be false. The first blow was struck by chimpanzees, which Jane Goodall famously described fashioning tools to fish for termites [1Goodall J. The Chimpanzees of Gombe: Patterns of Behavior. Harvard University Press, Cambridge, MA, US1986Google Scholar]. Since then the flood-gates have opened. Chimpanzee technology has received the most attention and is wide-ranging: for example, they use stone tools to crack open hard nuts, strong sticks to dig in the ground for tubers and underground bee hives, and sharpened sticks to spear bush babies sleeping in tree holes (for a recent overview see [2McGrew W.C. Chimpanzee technology.Science. 2010; 328: 579-580Crossref PubMed Scopus (58) Google Scholar]). Perhaps unsurprisingly primates stand out among mammals as the most frequent tool-users. Like chimpanzees, capuchin monkeys use stones both for nut cracking and digging [3Ottoni E.B. Izar P. Capuchin monkey tool use: overview and implications.Evol. Anthropol. 2008; 17: 171-178Crossref Scopus (181) Google Scholar]. Tools can extend the reach of the senses as well as the grasp: gorillas, when wading through water, use stick tools to test its depth. Acts of communication can also be enhanced by tools: orangutans use a handful of leaves to deepen the pitch of one of their calls. Tools can be protective: a chimpanzee may use a stick to investigate a fire (Figure 1), and orangutans use small sticks to rid Neesia fruits of their irritant hairs when extracting the seeds inside. And protection can also be important when catching prey, as when chimpanzees dip with long sticks for ferocious driver ants marching along the forest floor. Dolphins in Shark Bay carry sponges over their rostrums when foraging on the ocean floor, perhaps for a similar reason (Figure 2) [4Mann J. Sargeant B.L. Watson-Capps J.J. Gibson Q.A. Heithaus M.R. Connor R.C. Patterson E. Why do dolphins carry sponges?.PLoS ONE. 2008; 3: e3868Crossref PubMed Scopus (92) Google Scholar].Figure 2A dolphin 'sponging'.Show full captionIt is thought that dolphins at Shark Bay, Australia, have developed a tradition of using sponges to protect their sensitive rostrums when foraging on abrasive parts of the sea floor. (Photo: Michael Kruetzen.)View Large Image Figure ViewerDownload Hi-res image Download (PPT) It is thought that dolphins at Shark Bay, Australia, have developed a tradition of using sponges to protect their sensitive rostrums when foraging on abrasive parts of the sea floor. (Photo: Michael Kruetzen.) Such remarkable behaviour is not the sole preserve of primates or even mammals: some of the most impressive non-human tools are made by New Caledonian crows, a member of the Corvidea, a large-brained family of birds. New Caledonian crows extract grubs from tree holes using two distinct kinds of tool [5Hunt G.R. Manufacture and use of hook-tools by New Caledonian Crows.Nature. 1996; 379: 249-251Crossref Scopus (392) Google Scholar]. One variety is cut from the leaves of Pandanus (Figure 3). The other variety is made from twigs by removing all side branches from the central stem except one at the distal end, which is instead cut near the base and then sharpened to form a hook. A case of tool-use seems intuitively easy to identify, but a suitably precise definition has actually proved hard to pin down because of the problem of borderline cases. For instance, a widely-used definition is the use of an object "to alter … the form, position, or condition of another object, another organism, or the user itself when the user holds or carries the tool during or just prior to use" [6Beck B.B. Animal Tool Behavior: The Use and Manufacture of Tools by Animals. Garland, New York1980Google Scholar]. Several species (chimpanzees, capuchin monkeys and elephants) use branches or leaves to rid themselves of flies or parasites; but some chimpanzees use a vine for this purpose. The vine is not a detached object and would not 'count' as a tool, but is there a meaningful difference? Yet if we admit the vine, then perhaps the use of a scratching post by a cow or horse should count, and tool-use starts to bleed into any behaviour involving the external environment (for example, climbing a tree or building a nest). Water is another slippery case, as revealed by a problem-solving task given to orangutans and later to rooks. In both studies, a food reward was placed at the bottom of a transparent tube, out of reach of the subject, floating in a small volume of water [7Mendes N. Hanus D. Call J. Raising the level: orangutans use water as a tool.Biol. Lett. 2007; 3: 453-455Crossref PubMed Scopus (77) Google Scholar]. Orangutans brought the food within reach by spitting more water into the tube, while rooks added stones, raising the level of the water until they could reach the reward with their beaks [8Bird C.D. Emery N.J. Rooks use stones to raise the water level to reach a floating worm.Curr. Biol. 2009; 19: 1410-1414Abstract Full Text Full Text PDF PubMed Scopus (114) Google Scholar]. Were the rooks displaying tool use but the orangutans not? Archer fish, which spit jets of water to dislodge insects from the vegetation above them, are not usually considered tool-users. Nor are birds, such as seagulls and ravens, that drop encased food (shellfish, nuts, eggs or bones) from a height onto a hard surface to crack them open; but is this really different from capuchin monkeys or chimpanzees throwing sticks and stones, or using stones to smash open nuts [9Lefebvre L. Nicolakakis N. Boire D. Tools and brains in birds.Behaviour. 2002; 139: 939-973Crossref Scopus (227) Google Scholar]? Despite the seemingly arbitrary nature of the distinction, using a tool as an extension of the body may have particular consequences for psychological processes such as perception, attention and cognition, because the periphery of the body is thereby changed in mechanical and sensory capabilities. Recent research suggests that, in humans and monkeys, this extended motor capability is followed by changes in specific neural networks that hold an updated map of body shape and posture [10Maravita A. Iriki A. Tools for the body (schema).Trends Cogn. Sci. 2004; 8: 79-86Abstract Full Text Full Text PDF PubMed Scopus (794) Google Scholar]. As such, the classic definition retains its usefulness. Some borderline cases, such as sea otters smashing shellfish on a hard surface, are sometimes referred to as 'proto-tool-use': this identifies potentially interesting behaviour in which an outcome is achieved via a secondary object or substance, albeit not something defined as a tool. (And the usefulness of this distinction is reinforced by finding that, among birds, 'true' tool-users have larger brains relative to their bodies than proto-tool-users [9Lefebvre L. Nicolakakis N. Boire D. Tools and brains in birds.Behaviour. 2002; 139: 939-973Crossref Scopus (227) Google Scholar].) For the comparative psychologist, the semantics of which behaviour should 'count' seems less important than finding candidate cases of an animal solving a problem for which evolution has not provided a rigid morphological or behavioural adaptation: a context in which general cognitive abilities, like learning and reasoning, may be recruited. Human tool-making is characterized by an understanding of the physics of our bodies and surrounding objects, and an ability to plan a complex sequence of actions to achieve a distal goal. Tool-use can result from far simpler mechanisms, such as inborn dispositions and trial and error learning [11Tebbich S. Sterelny K. Teschke I. The tale of the finch: adaptive radiation and behavioural flexibility.Phil. Trans. R. Soc. B. 2010; 365: 1099-1109Crossref PubMed Scopus (110) Google Scholar]: but can all animal tool-use be explained in that way? Or do we see the evolutionary roots of human cognition in animal tool use? We can recognize an animal's target outcome: when it pursues that outcome directly, and ceases to act when it has been achieved. But does that mean that the animal has an internal representation of the goal 'in mind', and knows that its actions will cause it to follow? The difficulty is captured by two examples. Baboons throw rocks at predators, with the result that the predator is deterred. Do baboons recognize that being pelted with stones would cause the predator to flee, and throw them with that intention? With a primate, the tendency is to assume that the answer is yes, based on their large brains and their close relationship to humans. But what about the ant-lion larva, which flicks grains of sand at prey on the edge of its trap, causing them to fall in? Our intuition is that this behaviour, despite its surface similarities, has a different underlying cognitive mechanism: but does it? An isolated observation of tool-using behaviour cannot tell us much about the animals' appreciation of the underlying causality. A case for goal-directedness has been made for laboratory rats by training a pair of instrumental actions, such as chain-pulling for one food type and lever-pressing for another. When one food was devalued, for example by pre-feeding it to the rats, they homed in on the action that produced the other food: their action was therefore not triggered simply by force of habit, but by the context in which they were previously rewarded. Similarly, if the food linked to one action began to appear just as often when the rat performed the action as when it did not, the rats gave up that action but continued to perform the other; but if the delivery of 'free', non-contingent food was preceded by a signal, both actions continued to be performed. The rats' actions satisfy the criteria for goal-directedness: sensitivity to the incentive value of the rewards on offer and the causal relationship between their actions and their specific outcomes [12Dickinson A. Balleine B. Causal cognition and goal-directed action.in: Heyes C. Huber L. The Evolution of Cognition. M.I.T Press, 2000Google Scholar]. In other words, rats represent what they want and what they have to do to get it. If the behaviour of the ant-lion larva is hard-wired or habitual rather than goal-directed, we would expect the larva to behave differently to the rats: it might continue flicking sand at insects appearing at the edge of its trap even if they had become inedible, or if the contingency between flicking sand and insects falling was disrupted. That would suggest that tool-use in this species is inflexible and non-selective; unfortunately, these sorts of experimental manipulations are difficult to carry out in natural settings and have seldom been done. The tool-use of chimpanzees, capuchin monkeys and New Caledonian crows, at least, displays the hallmarks of goal directedness: selectivity, choosing or tailoring the tool to suit the specific goal at hand, and flexibility, such as using several means to achieve the same end [13Hunt G.R. Introduced Lantana camara used as tools by New Caledonian Crows (Corvus moneduloides) New Zealand.J. Zool. 2008; 35: 115-118Google Scholar, 14Sanz C. Morgan D. Flexible and persistent tool-using strategies in honey-gathering by wild chimpanzees.Int. J. Primatol. 2009; 30: 411-427Crossref Scopus (88) Google Scholar]. But there is an inborn component to the tool-using behaviour of even these accomplished and relatively large-brained tool-users. They begin manipulating tool objects (sticks or stones) from a young age, even in the absence of a goal object [15Kenward B. Weir A.A. Rutz C. Kacelnik A. Tool manufacture by naive juvenile crows.Nature. 2005; 433: 121Crossref PubMed Scopus (157) Google Scholar, 16Visalberghi E. Guidi C. Play behaviour in young tufted capuchin monkeys.Folia Primatol. 1998; 69: 419-422Crossref PubMed Scopus (9) Google Scholar]; and, despite the impressive ability of New Caledonian crows to use up to three tools in sequence in laboratory experiments (using smaller tools to gain tools long enough to reach a reward), they will sometimes use small tools to fish for longer ones when there is no ultimate food reward present [17Wimpenny J.H. Weir A.A.S. Clayton L. Rutz C. Kacelnik A. Cognitive processes associated with sequential tool use in New Caledonian crows.PLoS ONE. 2009; 4: e6471Crossref PubMed Scopus (91) Google Scholar]. Goal-directedness is not a simple all-or-nothing issue. A special case of causal reasoning concerns the physical process by which a tool effects its outcome. Even a goal-directed tool-user might know that her action caused an outcome, without knowing how or why it worked; in fact this is how the less technologically-minded among us use sophisticated tools like laser pointers or microphones. But humans are also capable of physical reasoning, for example unbending a coat-hanger to fish for keys that have fallen out-of-reach: fashioning a tool with the right physical properties to solve a problem. Betty the New Caledonian crow showed the same creative modification in the laboratory, bending (and unbending) wire to make a suitable hook or probing tool [18Weir A. Kacelnik A. A New Caledonian crow (Corvus moneduloides) creatively re-designs tools by bending or unbending aluminium strips.Anim. Cogn. 2006; 9: 317-334Crossref PubMed Scopus (88) Google Scholar]. Selective tool modification is also seen in primates (Figure 4): Gombe chimpanzees, for instance, famously convert a leafy vine into a thin, flexible probe-tool for termite-fishing, but make a stiffer rod for ant-dipping [1Goodall J. The Chimpanzees of Gombe: Patterns of Behavior. Harvard University Press, Cambridge, MA, US1986Google Scholar]. Behaviour like this raises the intriguing possibility that the animals represent the physical properties and forces involved in the tool-using event in an abstract, conceptual way: in terms of properties such as rigidity, continuity, and connectedness. The simpler alternative is that the animals' thinking is grounded in perceptual features of the objects (their shape, feel, or spatial orientation). Psychological experiments have often capitalized on tool-using (or proto-tool-using) behaviour to try to tease these alternative explanations apart. Chimpanzees in the Goualougo Triangle make a range of tools when fishing for termites: they select robust plants for making digging sticks, stripping side branches and sharpening one end before puncturing a fishing hole into subterranean termite nests; they also split stems lengthwise to use as flexible fishing lines for angry termites to bite onto; and form brush-tipped twigs by raking the tip through their teeth, providing a greater surface area for termite attachment. (Photo: Ian Nichols.) For many years, laboratory studies gave results supporting the simpler explanation. Even chimpanzees seemed to be using perceptually-based information rather than an abstract notion of object properties [19Povinelli D.J. Folk Physics for Apes: The Chimpanzee's Theory of How the World Works. Oxford University Press, Oxford2000Google Scholar]; for example, in the 'trap-tube' task, in which the subject needs to push a piece of food out of a horizontal tube away from a trap, the one chimpanzee that learned to do so continued to use this strategy even when the tube was inverted and the trap was non-functional. A capuchin monkey that solved the trap-tube task did likewise [20Visalberghi E. Limongelli L. Lack of comprehension of cause-effect relations in tool-using capuchin monkeys (Cebus apella).J. Comp. Psychol. 1994; 108: 15-22Crossref PubMed Scopus (223) Google Scholar], as did a New Caledonian crow [21Kacelnik A. Chappell J. Weir A.A.S. Kenward B. Cognitive adaptations for tool-related behaviour in New Caledonian crows.in: Wasserman E.A. Zentall T.R. Comparative Cognition: Experimental Explorations of Animal Intelligence. Oxford University Press, Oxford, UK2006: 515-528Google Scholar]. It seemed that even these successful individuals had avoided the trap as a perceptual feature but had not encoded its functional significance. Similarly, although chimpanzees would choose a complete tool over a broken one when the break in the wrong tool was clearly visible, they chose indiscriminately if the ends of the broken tool were aligned in front of them [19Povinelli D.J. Folk Physics for Apes: The Chimpanzee's Theory of How the World Works. Oxford University Press, Oxford2000Google Scholar]. New Caledonian crows, required to make a tool to fish for food in transparent wells, first made a tool of intermediate length and only made a longer one if the first one was too short [22Hunt G.R. Rutledge R.B. Gray R.D. The right tool for the job: what strategies do wild New Caledonian crows use?.Anim. Cogn. 2006; 9: 307-316Crossref PubMed Scopus (50) Google Scholar]. When presented with a tool made from barbed Pandanus leaves, positioned in the baited hole but with barbs pointing in the wrong direction, these crows first attempted to use the tool, and then either abandoned it or only switched it around after several unsuccessful attempts, even over repeated trials. In the wild, it seems that their successful use of the barbed leaves comes from the procedure by which they rip the tool from the leaf and insert it into the hole [23Holzhaider J.C. Hunt G.R. Campbell V.M. Gray R.D. Do wild New Caledonian crows (Corvus moneduloides) attend to the functional properties of their tools?.Anim. Cogn. 2008; 11: 243-254Crossref PubMed Scopus (32) Google Scholar]. Yet when the same experiment was run with hooked stick-tools, most of the same crows immediately repositioned the tools before attempting to use them: the picture is not straightforward. The fact is that the use of simple heuristic procedures does not preclude the capacity for more sophisticated strategies. Quick and simple strategies offer evolutionary advantages for animals in a harsh and competitive natural environment. Humans can also use 'fast and frugal' heuristics to solve problems, rather than slow and effortful reasoning processes, and indeed this approach is often more efficient [24Gigerenzer G. Brighton H. Homo heuristicus: why biased minds make better inferences.Top. Cogn. Sci. 2009; 1: 107-143Crossref PubMed Scopus (780) Google Scholar]: how many of us have tried a key without bothering to check first which way up it needed to be used? Perhaps the most powerful illustration of the danger of reliance on negative results comes from a study in which adult humans were given the trap problem: just like the tool-using animals, most continued to avoid the inverted trap [25Silva F.J. Page D.M. Silva K.M. Methodological-conceptual problems in the study of chimpanzees' folk physics: how studies with adult humans can help.Learn. Behav. 2005; 33: 47-58Crossref PubMed Google Scholar]! And indeed, it does no harm to do so. Although humans would no doubt use more sophisticated strategies in many of the other situations in which animals 'fail', would this reflect a qualitative difference in our cognitive make-up or a difference in our tendency to employ effortful thought? Are there any contexts in which animals go beyond simple strategies? A re-designed trap task, which pitted physical reasoning against perceptually based strategies, was used with rooks [26Seed A.M. Tebbich S. Emery N.J. Clayton N.S. Investigating physical cognition in rooks, Corvus frugilegus.Curr. Biol. 2006; 16: 697-701Abstract Full Text Full Text PDF PubMed Scopus (129) Google Scholar] and chimpanzees [27Seed A.M. Call J. Emery N.J. Clayton N.S. Chimpanzees solve the trap problem when the confound of tool use is removed.J. Exp. Psychol. Anim. Behav. Process. 2009; 35: 23-34Crossref PubMed Scopus (92) Google Scholar]. To avoid a trap, in some conditions subjects had to move the food across a plastic surface to an exit at the side, and in others they had to push it down an uninterrupted channel to an exit at the bottom. In a critical contrast, the same feature (the plastic surface) had to be treated differently depending on whether it played the role of a supporting surface (when the side exits were open) or barrier (when they were closed). Most chimpanzees and rooks did not make this distinction, suggesting reliance on the surface-level appearance of the task; but one animal from each group did, apparently having encoded something about the key physical properties. Chimpanzees (and also New Caledonian crows) were able to transfer their success in one trap problem to another one with very different perceptual features, also implicating abstract representational capacities [27Seed A.M. Call J. Emery N.J. Clayton N.S. Chimpanzees solve the trap problem when the confound of tool use is removed.J. Exp. Psychol. Anim. Behav. Process. 2009; 35: 23-34Crossref PubMed Scopus (92) Google Scholar, 28Taylor A.H. Hunt G.R. Medina F.S. Gray R.D. Do New Caledonian crows solve physical problems through causal reasoning?.Proc. R. Soc. Lond. B. 2009; 276: 247-254Crossref PubMed Scopus (139) Google Scholar]. Nevertheless, in all of these studies there is notable individual variation, the cause of which remains to be investigated. Capuchin monkeys have also been shown to treat the same visual feature differently depending on task-relevant physical properties [29Visalberghi E. Addessi E. Truppa V. Spagnoletti N. Ottoni E. Izar P. Fragaszy D. Selection of effective stone tools by wild bearded capuchin monkeys.Curr. Biol. 2009; 19: 213-217Abstract Full Text Full Text PDF PubMed Scopus (168) Google Scholar]. Wild capuchins were presented with stones at some distance from a nut-cracking anvil: they quickly selected large, heavy, robust stones to transport, seemingly identifying the effective stones through visual inspection. When presented with novel, man-made 'stones', which looked identical but differed in density and weight, the capuchins tapped and gently moved the stones before transporting the heavy, effective tool. They even chose correctly when the heavy tool was smaller than the other alternatives. If animals do know something about the physics of their environment and their actions, can they use this knowledge productively, to reason about new applications for tools or new means to reach their goals? The question of animal insight dates back to Koehler's experiments at the beginning of the 20th century, in which chimpanzees were given the task of obtaining out-of-reach rewards. The rapid emergence of a fully-formed solution, such as stacking boxes to climb on or combining short sticks to produce a long tool, prompted Koehler to apply the term 'insight' as an alternative mechanism to blind trial-and-error. However, the fact that chimpanzees exhibit many of these behaviours when the objects are presented without an out-of-reach reward points towards a simpler explanation: perhaps the 'spontaneity' of the chimpanzees' solution and its suitability to the task at hand was a case of one lucky trial and no error [30Shettleworth S.J. Animal cognition: deconstructing avian insight.Curr. Biol. 2009; 19: R1039-R1040Abstract Full Text Full Text PDF PubMed Scopus (20) Google Scholar]. In a recent revival of insight experiments, rooks, normally a non-tool-using species, were presented with a tube containing a collapsible platform with a food reward resting on it — the platform would collapse under slight pressure so that the food would fall out [31Bird C.D. Emery N.J. Insightful problem solving and creative tool modification by captive nontool-using rooks.Proc. Natl. Acad Sci. USA. 2009; 106: 10370-10375Crossref PubMed Scopus (228) Google Scholar]. At first the rooks did not solve the problem, but after experience of nudging stones into the tube from a ledge next to the opening they began to bring stones to drop into it, and generalised their solution to pushing down on the platform with sticks. In a subsequent experiment they even made a hook to lift a bucket of food from a tube: they put a wire into the tube, bent it over the lip, reversed the wire and used it to pull up the bucket. Impressive as this undoubtedly is, in all cases the solution, as in the rooks' training, involved adding an object to the tube; and the items to be used were placed beside it, increasing the likelihood that the 'correct' action would be the first one attempted in the face of the new problem. Without another group of naïve birds for each stage, it is impossible to know how important that cumulative experience was [30Shettleworth S.J. Animal cognition: deconstructing avian insight.Curr. Biol. 2009; 19: R1039-R1040Abstract Full Text Full Text PDF PubMed Scopus (20) Google Scholar]. Regardless, the rooks displayed sensitivity to the task requirements, quickly developing a preference for heavy stones over light ones and using thin stones in preference to wide ones when the mouth of the tube was narrowed. In the platform-collapsing task, New Caledonian crows generalised from the experience of pecking at the platform through a small-necked tube to dropping objects into it when the neck was longer; though note that these birds, unlike the rooks, already have tool-use in their behavioural repertoire [32von Bayern A.M.P. Heathcote R.J.P. Rutz C. Kacelnik A. The role of experience in problem solving and innovative tool use in crows.Curr. Biol. 2009; 19: 1965-1968Abstract Full Text Full Text PDF PubMed Scopus (93) Google Scholar]. In his later writings, Koehler made a distinction between insight (quickly recognising when the right solution has been stumbled upon in the course of exploration, leading to a sudden disappearance of further trial-and-error), and 'foresight' (sizing up a problem in advance through physical reasoning and bringing a new solution to it fully formed). To date, there seems to be clearer evidence for the former ability than for the latter in animals such as chimpanzees and corvids. The exact cognitive mechanisms underpinning both sorts of accomplishment are yet to be formalised, but examining the sorts of experience needed for the emergence of such seemingly 'insightful' solutions seems to be a fruitful beginning [30Shettleworth S.J. Animal cognition: deconstructing avian insight.Curr. Biol. 2009; 19: R1039-R1040Abstract Full Text Full Text PDF PubMed Scopus (20) Google Scholar]. A powerful illustration of the role of immediate feedback about an action's effect comes from a study investigating one of the most famous cases for insight in animals: birds pulling up a reward on the end of a string tied to a perch in increments, trapping successive lengths under their feet. Strikingly, most naïve New Caledonian crows produced the complete solution in one fluid attempt from the first trial. However, when deprived of visual feedback (through the use of a horizontal occluder with a hole just large enough for the reward to pass through), naïve New Caledonian crows tugged at the string but did not pull up the reward. Even the performance of the experienced subjects was disrupted [33Taylor A.H. Medina F.S. Holzhaider J.C. Hearne L.J. Hunt G.R. Gray R.D. An investigation into the cognition behind spontaneous string pulling in New Caledonian crows.PLoS ONE. 2010; 5: e9345Crossref PubMed Scopus (76) Google Scholar]. This experiment neatly exemplifies how even rapidly emerging solutions need not have been planned out in the mind. What role, if any, did physical knowledge play? Visual feedback was shown to be necessary, but is it sufficient — would any action with a contingent effect on the reward be repeated? One reason that tools have had such a high profile in the study of human evolution is the evidence they give about ancient hominins' ability to plan ahead. Reconstructions of knapped flint from over two million years ago show a detailed sequence of detachments that can only result from mentally 'looking ahead' in the manufacture process [34Roche H. Blumenschine R.J. Shea J.J. Origins and adaptations of early Homo: what archeology tells us.in: Grine F.E. Fleagle J.G. Leakey R.E. The First Humans – Origin and Early Evolution of the Genus Homo. Springer, Netherlands2009: 135-147Crossref Scopus (34) Google Scholar]; later in human evolution, raw material for stone tools has been found