Innovative problem solving in macaws

Laurie O’Neill; Rahman Rasyidi; Ronan Hastings; Auguste M. P. von Bayern

doi:10.3758/s13420-020-00449-y

Innovative problem solving in macaws

O’Neill, Laurie; Rasyidi, Rahman; Hastings, Ronan; von Bayern, Auguste M. P. 2020-12-07 00:00:00 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. . . . . . Keywords Comparative cognition Tool use Parrot cognition Causal understanding Innovation Physical cognition Introduction Visalberghi & Limongelli, 1994; Visalberghi & Trinca, 1989; von Bayern, Danel, Auersperg, Mioduszewska, & In the field of animal cognition, problem-solving tasks of- Kacelnik, 2018; Weir & Kacelnik, 2006). Typically, these ten involve tool use (Auersperg, Huber, & Gajdon, 2011; tool-use tasks present subjects with a novel apparatus that Auersperg, von Bayern, Gajdon, Huber, & Kacelnik, 2011; requires some sort of behavioural innovation to access a Bird & Emery, 2009; Girndt et al., 2008;Köhler, 1917; reward. They are often used as a measure for an individ- Martin-Ordas Call, & Colmenares, 2008, 2012; Mulcahy ual’s behavioural flexibility and examine what a species & Call, 2006; Taylor, Hunt, Holzhaider, & Gray, 2007; understands about the functional properties of those tools, and physical interactions in their environment in general. Of particular interest is comparing the performance of spe- cies that are habitual tool users in the wild and those that * Laurie O’Neill are not, investigating both closely and distantly related [email protected] species (Auersperg et al., 2011; Auersperg, von Bayern, * Auguste M. P. von Bayern Gajdon et al., 2011; Lambert et al., 2017; Teschke et al., [email protected] 2013). In this manner it is possible to examine if the ca- pacity for flexible tool use is a result of a more general Max Planck Institute for Ornithology, 82319 Seewiesen, Germany cognitive capacity evolved in response to selection pres- Max Planck Comparative Cognition Research Station, Loro Parque sures not restricted to the physical domain (e.g., general Fundacion, 38400 Puerto de la Cruz, Tenerife, Spain 3 intelligence; Burkart et al., 2017), or whether more flexible Behavioural Ecology Group, Department of Animal Sciences, tool-using skills have evolved from a more specialised Wageningen University and Research, Wageningen, the Netherlands 4 physical-cognition domain (Kacelnik, 2009; McCormack Department Biology II, Ludwig-Maximilians-University of Munich, et al., 2011; Taylor & Gray, 2014). Martinsried, Germany Learn Behav (2021) 49:106–123 107 Tool use, however, is not the only type of physical envi- bitter taste of certain piths in their outer layers (Villegas- ronmental interaction that might require complex physical Retana & Araya-H., 2017). Palm cockatoos’ (Probosciger cognition. Many species of birds face similar complex object aterrimus) use sticks as part of their mating display, drum- manipulations when they are also required to probe, weave ming them against hollow trees (Heinsohn et al., 2017; and manipulate many different objects, including sticks, in Wood, 1984). Many parrots can face other types extractive order to build their nests (Breen et al., 2016). Furthermore, foraging difficulties such as trying to manipulate and wield there are many types of extractive foraging without tool use very large and tough-shelled nuts, which require complex ma- that can require complex manipulations, for example when nipulations in order to be able to open them (Tella et al., Gorillas prepare thistles in order to eat them they must manip- 2015). ulate them in such a way so as to avoid being hurt by spines Two parrot species in particular have been the focus of (Byrne et al., 2001). In sum, tool use is not the only measure of physical cognition abilities in parrots and have been shown an individual’s physical cognition skill. to be able to use sticks as extractive foraging tools in a number In this experiment we studied two species of parrots, spe- of problem-solving tasks. Kea are able to use stick tools by cifically the macaws Ara ambiguus (great green macaws) and inserting them into holes in puzzle boxes to knock food items Ara glaucogularis (blue-throated macaws). Some species of off a pedestal (Auersperg et al. 2011; Auersperg, von Bayern, parrots have been recorded as having complex and flexible Gajdon et al. 2011), push food items out of a tube (Lambert physical problem-solving skills (Lambert et al., 2018) and et al., 2017), and even to spring traps meant for other animals they share a number of other life-history and social factors in the wild in order to retrieve the bait (Goodman et al., 2018). with other taxonomic groups, such as corvids and primates, Goffin’s cockatoos have been established as probable non- that suggest they might be required to use complex cognition tool users in the wild (O’Hara et al., 2019; but see Osuna- (Emery & Clayton, 2004; Osvath et al., 2014; Van Horik Mascaró & Auersperg, 2018), but remarkably some individ- et al., 2012). These factors include having relatively large, uals began to spontaneously innovate the manufacture and use neuronally dense brains that mean they might have a higher of stick tools in captivity (Auersperg et al., 2012). upper limit in cognitive capacity (Gutiérrez-Ibáñez et al., Furthermore, these cockatoos appeared to show understand- 2018; Herculano-Houzel, 2017; Iwaniuk et al., 2005; ing of some of the complex properties of these tools that made Olkowicz et al., 2016), although this does not determine or them functional, such as making tools of the appropriate imply high cognitive ability. Many parrots are capable of a length (Auersperg et al., 2018) or adding hooks to their tools diverse range of motor actions, which is likely to be a key when necessary (Laumer et al., 2017). Because of these ex- factor for behavioural innovations (Griffin et al., 2014), and amples, it is interesting to expand tool-use tasks to other spe- they have particularly impressive manipulative skills with cies across the parrot order. The Kea and Goffin’scockatoos both their beaks and tongues as well as zygodactyl feet that are members of the more basal parrot superfamilies allow them to hold and manipulate objects in them (Toft & Strigopoidea and Cacatuoidea, respectively, whereas the ma- Wright, 2016). Parrots also live in stochastic environments caws in this study are parrots of the family Psittacoidea, so the that might require behavioural flexibility in order to thrive current experiment provides an opportunity to expand our (Toft & Wright, 2016), and further to this, many different understanding of parrots’ physical cognition capacities across species of parrots have been shown to be successful invaders the parrot order. We can therefore explore what kind of un- to their non-native range (e.g., Menchetti et al., 2016), some- derstanding these various species have for tools and instru- thing that has also been suggested to require elements of be- mental problem solving, and whether there are certain evolu- havioural flexibility (Sol et al., 2002, 2005). In total, all the tionary factors that might drive more complex physical cog- factors of big brains, capable physical manipulation skills and nition abilities. a capacity to respond to changing environments, suggest that The specific individuals we tested have faced tool-use tasks parrots might be candidates for using cognitively complex before. In a single trial-test of a ‘food-out-of-reach’ situation problem-solving skills, such as causal understanding, in order with a stick tool available, they did not interact with a stick to be adapted to their environment (Van Horik et al., 2012). tool (Krasheninnikova et al., 2019). In the stone-dropping A few species of parrots have shown capacities to handle task, a problem-solving task that requires individuals to inno- objects in a variety of interesting ways, some of which are vate the use of a stone tool to get a reward (von Bayern et al., borderline cases of tool-use. Vasa parrots (Coracopsis vasa) 2009), the majority were capable of using stone tools to drop are able to use stones to scrape shells in order to obtain calci- onto a collapsible platform to retrieve a reward, but most in- um powder for consumption (Lambert et al., 2015). Both hy- dividuals were unable to manipulate a stick tool to perform a acinth macaws (Anodorhynchus hyacinthinus) and great green very similar action (O’Neill et al., 2020). As some species of macaws (Ara ambiguus) have been observed wrapping certain macaws have shown borderline tool use abilities in the wild nuts with leaves, which was interpreted as a method to grip the before (Borsari & Ottoni, 2005; Villegas-Retana & Araya-H., nuts better (Borsari & Ottoni, 2005) or perhaps to avoid the 2017), we believed them to be suitable candidates to test the 108 Learn Behav (2021) 49:106–123 functional understanding of instrumental problem solving in were still able to solve the task after making this error, as the the lab. apparatus only needed two stick insertions from one side to The inspiration for the experimental task we used comes reach the reward. This may have meant there was little moti- from Visalberghi and Trinca (1989). In that experiment, a vation for the example capuchin subjects to optimise their group of tufted capuchins (Sapajus apella) were given a trans- multi-tool construction, as there was little to no negative out- parent, long, horizontal tube, with a reward in the middle. To come from the errors they made. To solve this issue in our solve the task they had to insert a stick tool into one side of the replication of the experiment, we required that there were at tube and push the reward out of the other side. After success- least four insertions of a small stone from one side before the fully doing this, they were given the tube apparatus again but reward was available from the other. Additionally, they also only with access to three smaller sticks, none of which could typically required three consecutive actions before they had reach the reward on its own. The subjects had to insert two of visual feedback that the reward was being moved. these smaller sticks, one behind the other, in order to push the The rationale of the experiment was as follows. The first reward far enough to the other end of the tube so that it was stage of the experiment, the pre-test, gave the subjects an accessible. The aim of the task was to see if the subjects opportunity to spontaneously solve the task and see if they recognised the functional property of the stick tool that they were able to insert multiple stones, one behind the other, to initially inserted, or whether they were just ‘performing’ a push a reward out of a horizontal tube (Fig. 1a and b). It was learned action of ‘insert sticks’ as this would lead to rewards. expected that they would not be able to do so, thus after this The modification we made on this task was to simply replace phase the subjects were given an example of the end-state these small sticks with even smaller stones. Thus, more of function of the multi-stone construction they were required these stone tools would have to be inserted into the tube before to make with the stones, ergo a stick was pre-inserted into the reward was available (a minimum of four). This behaviour the tube and the subjects simply had to push it to remove the could possibly be classed as an ‘additive tool’, as the subjects food (Fig. 1c and d). After this, the subjects were once again would need to “join or connect two or more objects to make given the opportunity to solve the task with the stones – crit- one tool that is held or directly manipulated in its entirety ical test 1 (a repeat of Fig. 1a and b). The underlying principle during its eventual use” (Shumaker et al., 2011). However, was that if subjects were able to reason the underlying func- it is not certain whether this multi-component structure would tional mechanism that made the stick effective, then they fit all definitions of tool use (e.g., Fragaszy & Mangalam, would be able to recreate the same effect of this stick out of 2018). Nevertheless, the focus of the experiment was not multiple stones. If the subjects were still unable to solve the ‘whether the subjects could use tools’, but instead it was to task at this stage, then they were given a more direct experi- explore subjects’ causal understanding of the underlying func- ence of the efficacy of the stones, thus cueing them to use tional properties of the initial stick tool they interacted with to stones. They were given a similar apparatus, but the horizontal see if they could recreate the functional properties in another tube was shortened and only a single stone needed to be way. For this reason, we do not refer to the multi-component inserted and pushed to make the reward available on the other construction subjects were manufacturing as an additive tool, side (Fig. 1e and f). If they succeeded in doing that then sub- as we recognize this is not a name that everyone will accept. jects were given another critical test (critical test 2)tosee if Instead, we refer to it as a ‘multi-stone construction’ or just a they could chain this single stone insertion behaviour repeat- ‘stone construction’. edly to make the required multi-stone construction. In essence, One issue with this task is that the capuchin subjects ap- in each of the three test-phases (the pre-test and critical tests 1 peared to be able to find a solution through exploratory, trial- and 2), subjects had to make a smaller innovative ‘leap’,and and-error behaviour (Visalberghi & Trinca, 1989). These use an assumed ‘less complex’ form of cognition, to solve the kinds of persistent behaviours have been noted to predict task. After the subjects were able to consistently solve the many animal subjects’ ability to solve problem-solving tasks task, we also gave them up to two additional follow-up tasks in other experiments (Chow et al., 2016; Thornton & Lukas, in order to triangulate (Heyes, 1993) how much the subjects 2012; Thornton & Samson, 2012; van Horik & Madden, understood about the function of their multi-stone construc- 2016). Specifically, the three capuchin subjects that were able tion (Fig. 1g and h). to solve this additive tool task repeatedly made an error of Although the subjects we tested were naïve to the specific inserting a stick from both sides of the reward before inserting apparatus that we gave them in this experiment, they were not a stick behind one that had been previously inserted (and thus naïve to the experimental protocols in general. For this reason solving the task). This suggested that the subjects did not have we decided to exploit their testing history to also elucidate an end-state mental representation of the multi-tool construc- what kind of heuristics (Hutchinson & Gigerenzer, 2005) tion they were making with the shorter sticks (e.g., Jelbert and behaviours the subjects may extend from successful be- et al., 2018), and instead were just randomly inserting sticks havioural interactions in previous experiments in order to find until they could reach the reward. Nevertheless, the subjects rapid innovations to the novel apparatus in this experiment, a Learn Behav (2021) 49:106–123 109 110 Learn Behav (2021) 49:106–123 Fig. 1 In the pre-test and critical tests a reward (a walnut represented by were housed in groups of two to six individuals, accord- a green circle in the figures) was placed in the middle of a transparent ing to species and age, in seven aviaries. Six of these acrylic tube with ten stones placed around it (a). In the Pre-test and aviaries were 1.8×6×3m(width ×lengthxheight) Critical tests, subjects had to insert five of these into one side of the tube and one was 2 × 6 × 3 m. Windows of 1 × 1 m could to push out rewards on the opposite end (b). In the stick experience phase, a stick was inserted into the side of the tube up until the reward be opened between the aviaries to connect them together. (c). Subjects had to push the stick to push the reward out (d). In the short- One half of each aviary was outside, so the birds followed tube experience a smaller tube was given to the subjects in which only the natural heat and light schedule of Tenerife. The half one stone needed to be inserted into the tube and pushed to make the inside the research station was lit with Arcadia Zoo Bars reward come out (e). In the pre-inserted stone experience, a stone was inserted to be touching the reward in the short tube, so that the stone only (Arcadia 54W freshwater Pro and Arcadia 54W D3 needed to be pushed to make the reward come out (f). In the parallel Reptile Lamp) that followed the natural light schedule. tubes phase, there were two tubes, but only one of them was baited (g), A large variety of fresh fruit and vegetables were pro- subjects had to target the stones only at the baited tube. In the blocked vided first thing in the morning and again in the after- tubes phase, both tubes were baited, but a bung prevented the reward from being removed from one tube (h). For clarity in the schematics, a noon. In the afternoon, each individual received a portion vertical tube in the middle of the apparatuses is not shown. This was of Versele Laga Ara seed-mix that was modified based on featured in all apparatuses except the short tubes (e and f). This vertical their daily weight. The parrots were never starved, and tube can be seen in the supplementary videos showing the subjects portions of nuts and seeds were controlled to prevent interacting with the horizontal tube overeating and obesity. Testing took place at least an hour after the parrots had breakfast to ensure there was suffi- cient motivation to get food rewards. Unless otherwise mechanism thought to be vital in animal innovations (Tebbich et al. 2016). We therefore added a ‘red-herring’ to the apparatus. specified, walnut halves were used as rewards. These were highly prized by all subjects, and they were able to Specifically, this was in the form of a vertical tube, attached to the obtain them on a daily basis through voluntary testing. horizontal tube, that was placed directly over the reward, but was blocked and so had no connection to the reward (see Supplementary Fig. 7 (Online Supplementary Material, OSM) Experimental rooms for a photo of the apparatus). Placing a rock into this vertical tube would therefore achieve nothing. Nevertheless, in the previous experiment they completed, the stone-dropping task (O’Neill Experiments took place in separate testing rooms away from the aviaries. All birds had been previously trained to enter them. et al., 2020), the solution to the experiment was to place a stone tool into a vertical tube. Thus, we could see if they would apply Theserooms were 1.5× 1.5×1.5mandalsolit with Arcadia Zoo Bars to cover the birds’ full visual range. One wall of each this same rule in their exploration of a new apparatus that was drawn from success in a previous experiment. one of the experimental rooms had a 50 × 25 cm window through which an experimenter could place apparatuses into We predicted that at least some of the parrots would be the testing room from a neighbouring chamber. This window likely to solve the task, as we have seen previously that the could be occluded with a white board so that the experimenter subjects were highly explorative and capable in this similar could hide anything they were doing from the subjects’ view, stone-dropping problem-solving task (O’Neill et al., 2020). such as re-setting apparatuses between trials. The experimenter However, we were not certain whether the subjects would always wore mirrored sunglasses during experiments to prevent be able to solve the task consistently nor succeed in the trans- fer tasks as this experiment required more self-control, some- their gaze being a cue for the parrots. A second wall was made of sound-proofed one-way glass so that zoo visitors could thing we knew the subjects might lack (Kabadayi et al., 2017). observe experiments without disturbing the subjects. One individual, Hannibal, was tested in a separate cage con- nected to his home aviary as he was not comfortable going to the Methods experimental rooms. Each aviary had a separate, removable cage attached to it, which was used to transport subjects between their Subjects and housing home aviaries and the experimental rooms. However, the cages were also large enough for testing (1 × 1 × 1 m). They had sliding Nine great green macaws (Ara ambiguus) and eight blue- doors that could separate an individual from the rest of their throated macaws (Ara glaucogularis) were tested. Their aviary, and were also on wheels, which meant they could be age and sex are shown in Table 1. All the birds were moved to an area occluded from the other individuals in their hand-raised and group-reared by the Loro Parque aviaries. Finally, they also had a second door through which Foundation in Tenerife, Spain, and all were housed in testing apparatuses could be placed into the cage with a test the Comparative Cognition Research Station, within the subject. In this way, Hannibal was able to complete the test in Loro Parque zoo in Puerto de la Cruz, Tenerife. The birds a similar manner to the other subjects. Learn Behav (2021) 49:106–123 111 Table 1 The age, sex and success of each subject taking part in the which the subjects first had a successful trial in that particular test horizontal tube task The table also shows which stage of the phase. Dashes indicate that the subject did not take part in that test experiment each individual completed and at which stage they failed or phase. Columns in both shades of grey are the mechanistic experiences completed the task. The numbers in the columns show the session in Apparatus For each test, the subjects were provided with a selection of ten stones. All were natural volcanic rocks taken from the This task required subjects to push a reward out of a horizontal beach in Puerto de la Cruz, Tenerife (they were returned to transparent tube, which was 30 cm long with a diameter of the beach after testing was finished). They all fit into the tube, 6cm (Fig. 1a and b). The tube was fixed in an elevated hor- had a diameter of 4.5–5.5 cm and weighed between 60 and izontal position at a height of 10 cm by a metal stand at either 100 g. Three stones had to be placed into the tube, one behind end (see Supplementary Fig. 7 (OSM) for picture). The stands the other, before the reward moved. At least four stones had to were attached to a solid wooden board (45 × 30 × 2.5 cm) and be inserted to place the reward within reach. This meant that the central tube was wrapped in wire mesh (1 × 2 cm) to show the subject received no perceptual feedback of the reward the solidity of the transparent acrylic. The reward was placed moving until at least three specifically directed actions were in the middle of the horizontal tube, thus the only way to made. obtain the reward was by pushing it to one of the two edges with the aid of a tool. There was an additional 5-cm tube (vertically positioned) of the same diameter attached to the Experimental procedures centre of the top-side of the main tube (see Supplementary Fig. 7 (OSM)). This additional tube was non-functional as The experiment consisted of three test phases (pre-test, criti- no tool could interact with the reward from here. Its purpose cal test 1 and critical test 2), three different experience phases was to provide a place for the subjects to put the stone tools (pre-inserted stick, short tube 1, pre-inserted stone and short onto the top of the apparatus. This was so that the subjects’ tube 2) and two transfer tasks (parallel tubes and blocked tendency and persistence for placing tools in an opening on tubes). An overview of the order in which these phases took top of the apparatus could be assessed, even when it had no place is shown in Fig. 2. Not every subject took part in every effect on the outcome of the task. phase as they either skipped phases due to early success in an 112 Learn Behav (2021) 49:106–123 Fig. 2 The subjects were given progressively more functional experience they followed the green arrow, if they failed then they followed the red if they failed to solve the task. In blue are the ‘test’ phases, in which arrow. Subjects that failed the short-tube experience moved onto the pre- subjects were exposed to the task that required them to use a multi- inserted stone experience, if they succeeded in the preinserted stone ex- stone construction to push the reward out of the tube. In orange are the perience they were given the short-tube experience again. If they failed ‘experience’ phases, which provided the subjects with mechanistic expe- the short-tube experience a second time then testing ended, i.e. they only rience of how the task could be solved. If subjects succeeded in a phase had the pre-inserted stone experience once experimental phase or did not take part due to failure in an stones and the opening of the tube. The critical test phases, experience phase. after each of these mechanistic experience phases, were im- In essence, in the pre-test the subjects were given an op- plemented to test if the subjects could generalise and extrap- portunity to spontaneously solve the task in order to examine olate from these progressively more specific mechanistic ex- if the experiment could be solved without any mechanistic periences to produce the necessary actions to solve the task. experience, just by exploration behaviours characteristic of Finally, the subjects were offered the parallel and blocked the parrot subjects. The experience phases presented the sub- tubes tasks as follow-up tasks to investigate if they could jects with progressively more specific direct mechanistic ex- flexibly solve slightly modified tasks, verifying whether they perience of how the experiment could be solved: the first indeed attended to the functional properties of the task or not. mechanistic experience (pre-inserted stick) provided them Within the protocol descriptions, the term ‘session’ de- with knowledge of the basic mechanism that the food could scribes a single period of testing in the experimental chamber. be pushed from contact with a stick, but without cueing them Typically subjects took part in a single session per day. Within towards stones in particular. The second mechanistic experi- each testing session (pre-test, both critical test phases, and the ence (short tube) allowed the subjects to move the reward by parallel and blocked tubes phases), subjects could be given a just inserting a single stone, thus reinforcing subjects that maximum of three trials (i.e. chances to solve the task and showed exploratory stone-inserting behaviour by letting them obtain a reward). Within each session in the mechanistic ex- experience that their behaviour can move the reward and lead perience phases (pre-inserted stick, short tube and pre-inserted to success. The third mechanistic experience further helped stone), subjects could be given a maximum of six trials. This subjects that would not insert stones (pre-inserted stone)as was because the successful method in the test sessions took it showed that just pushing stones already pre-inserted in the slightly longer and we did not want subjects to lose motiva- tube would move the reward, hence cueing them both towards tion. If subjects failed a trial, the whole session was stopped Learn Behav (2021) 49:106–123 113 and subjects were brought back to their group aviary. A trial board beneath it. The whole apparatus was pushed into the was considered as failed if subjects did not solve the task subject’s chamber, in the centre of the testing area; thus the within 10 min of the apparatus being placed into the experi- subjects were able to access it from all sides. The subjects mental room with them. If subjects succeeded in a trial, then were given a maximum of 10 min to interact with the appara- the apparatus was removed, re-baited and replaced into the tus. If they failed to retrieve the reward in this time, then the subject’s experimental room for another trial (described be- trial and the session ended, the apparatus was removed, and low), unless it was the last trial of a session, in which case the the subject was returned to its social group. If, on the other session ended. All subjects finished testing, from habituation hand, the subject managed to successfully retrieve the reward to final test trial, within 2 months. All testing took place be- by pushing multiple stones into the side of the apparatus, then tween November 2017 and May 2018. the subject was given further trials within that testing session. To do this, the experimenter waited for 30 s after the subject Habituation had retrieved the reward and then removed the apparatus. It was re-baited with another half piece of walnut out of sight of Every subject was first habituated to the apparatus without the the subject, ensuring that there were ten stones available un- presence of the stones. These sessions took place in the exper- derneath the apparatus. The apparatus was then placed back imental rooms, with an experimenter sitting in the into the room with the subject for another trial exactly 1 min neighbouring chamber. This was necessary for re-baiting if after it was removed from the subject’s chamber. Subjects further trials were needed in the session and it also served to received a maximum of three trials in a session. They needed create a more relaxing environment for the subjects to work in. to complete 12 successful trials to be considered consistently Once the subject was in the experimental room, the experi- successful within this experimental phase (although Captain menter introduced the apparatus through the window in be- and Enya, see Table 1, progressed from the second critical test tween them. Subjects had 5 min to approach the apparatus and phase with 11 successful trials due to an experimenter’serror). take a piece of walnut placed at the base of the apparatus. If This repetition was to validate that subjects had not acciden- they took it, the experimenter waited for 30 s then removed the tally succeeded in a task but could replicate the method they apparatus. Next, the experimenter, out of sight of the subject, used to get the reward. Additionally, if subjects had six failed re-baited the apparatus and reintroduced it 1 min after remov- trials in a condition, they moved onto the next mechanistic ing it. The subjects had up to six of these trials per session and experience phase (or stopped testing, depending on which if they took six walnut pieces in a row within a single session, condition they had ‘failed’;see Fig. 2). they moved onto testing. Sessions were only counted as valid if the subjects The subjects had already been habituated with the stone approached and touched the apparatus. Thus unsuccessful tri- als were ones in which subjects had touched the stone tools or tools in a previous experiment and all were eager to interact with them, so they did not need habituating to them again. We the apparatus but had not been able to retrieve the reward. state that the parrots were ‘eager’ to interact with the stones Subjects that were able to succeed in the pre-test or either of because they would frequently pick up the stones and just the critical test phases moved onto the follow-up tasks (par- manipulate them between their feet and beaks with no appar- allel tubes and blocked tube). ent purpose. This was typical for these macaws for any object that was a similar size and shape to these stones. Pre-inserted stick experience Pre-test and critical test procedures In the first mechanistic experience phase, called pre-inserted stick experience, a stick tool was used to show how the reward The pre-test and both critical test procedures were identical. could be pushed out of the tube (Fig. 1c and d). In the first The difference was that subjects had been given different stage of this phase, the stick tool was attached to the tube so mechanistic experience phases after the pre-test and between that it could only be pushed into the tube, i.e. it could not be critical tests 1 and 2, which provided functional information pulled out. We assumed the parrots would be likely to pull the about how the task could be solved. The order of these tests stick tool in their initial exploratory interactions with it, so and mechanistic experience phases can be seen in Fig. 2. attaching it in this way made sure they could only experience Subjects were given a minimum of six 10-min sessions for the ‘rewarding’ outcome of pushing the stick first. The exper- both the pre-test and the two critical tests. Each session had imenter placed a half walnut reward in the centre of the appa- between one to three trials, i.e. they were given more trials in a ratus with the stick tool inserted flush against it. The subjects session if they succeeded until the maximum trial number of were given multiple 10-min sessions to push the stick through three was reached. The apparatus was first prepared out of explorative behaviour alone. If they succeeded within a ses- sight of the subjects. A half walnut reward was placed inside sion, the experimenter gave the subjects more trials. They the middle of the tube and ten of the stones were placed on the removed the apparatus, re-baited it out of sight, ensured the 114 Learn Behav (2021) 49:106–123 stick tool was correctly positioned and gave it back to the removed, and the subjects were taken back to their social subject. If subjects successfully obtained the reward six times group. If subjects failed six trials in the short-tube experience, within a session, they moved onto the next stage of this mech- then they were given another mechanistic experience phase, anistic experience phase. the pre-inserted stone phase. In this following stage, the procedure and setup were If the subjects succeeded in the pre-inserted stone identical, except the stick was no longer attached to the experience, then they still had to succeed in the short-tube apparatus, so it was possible for the birds to pull the stick experience before they could proceed to critical test 2. out of the apparatus. We did this to ensure that the subjects However, the subjects only had one opportunity to succeed recognised that thestickspecificallyneededtobepushedto in the pre-inserted stone experience,i.e. subjects only moved get the reward, and it wasn’t just exploratory interaction between the short-tube experience to the pre-inserted stone with the stick that led them to obtaining the reward. If sub- experience once, not continually until they succeeded. jects pulled and thus removed the stick at this stage, the experimenter would remove the apparatus and the stick, Pre-inserted stone experience reposition the stick inside the apparatus out of sight of the subjects, and then give it back. There was no limit to the The pre-inserted stone mechanistic experience used the same number of times the stick could be re-positioned like this short tube apparatus (Fig. 1f). However, now one of the stone (but there was still a 10-min time limit per trial). To pass tools was already placed inside the tube next to the reward so this final stage of the experience phase, the subjects had to the subjects only needed to push the stone to obtain the re- push out the reward 12 times in two consecutive sessions ward. The purpose of this was to cue the subjects to the correct (six trials per session). All subjects completed the first stage positioning of the stone and that interaction with the stone of the pre-inserted stick experience in a maximum of six when it was in this position would also make the reward sessions and the second stage in a maximum of three move. If subjects succeeded in this stage, then they were given sessions. They then moved onto the first critical test. another chance to succeed in the short-tube experience, which they then had to succeed on in order to proceed to critical test Short tube experience 2. In this experience phase the experimenter placed one of If the subjects failed the first critical test, they were given a the stones in the tube, flush against the reward, before giv- second mechanistic experience phase. In this mechanistic ex- ing it to the subject. This setup was done out of sight of the perience phase, called the short-tube experience, a modified subjects, so they did not receive a cue from the experiment- apparatus was used (Fig. 1e). A 15-cm long tube was mounted er to specifically insert the stones into the tube, they only on a separate wooden board at the same height as the original saw the apparatus with a stone already inserted. For each apparatus. It did not have the additional vertical tube in the trial, the subjects had up to 10 min to obtain the reward. If centre. With this shorter tube, the subjects were still unable to the reward was obtained, the apparatus was removed, re- reach a reward placed in the middle, but now they only needed baited, and given back to the subject. If subjects could to insert one stone to push the reward out of the tube. Some repeat this behaviour six times in one session, they were subjects had inserted one or more stones into the apparatus in given another round in the short-tube experience phase, the initial critical test, but not a sufficient number to move the exactly as described above. If they passed criterion in the reward (see Results section). If they repeated this behaviour latter, thentheyweregiven critical test 2, but if they failed, with a shorter tube it would be enough to move the reward. their participation in the test ended. The short tube was given with the same procedure as the All of the subjects that were given the pre-inserted stone critical test, i.e. the subjects were given the short-tube appa- phase succeeded (nine subjects), but only one of these subjects ratus with ten stones and they then had 10 min to solve the task (Rita) managed to succeed in the following short tube 2 expe- by placing one stone into the tube and pushing the reward out. rience phase (see Table 1). After successful trials, the apparatus was removed, re-baited, and placed back into the subject’s room. However, as this Parallel tube and blocked tube transfer tasks solution to the short-tube experience phase was potentially very rapid, as the subjects only needed to insert a single stone, Two follow-up tasks were devised for subjects that managed subjects were given up to six trials per session to speed up the to reliably succeed in one of the critical test phases. The pur- testing process. To reach the criterion in the short-tube pose of these was to inform on how the subjects had achieved experience, and in turn proceed to critical test 2,they had to their successful methods in the critical test phase. succeed 12 times consecutively over two sessions. If subjects Specifically, the parallel tubes test (Fig. 1g) was used to failed a trial in a session, i.e. they did not obtain the reward see if subjects were inserting stones into the tubes to move the within 10 min, then the trial and session ended, apparatus was reward or whether they had just learnt that inserting stones Learn Behav (2021) 49:106–123 115 into the tubes was rewarding. In this task, two identical ver- camera angle was considered ‘side 1’ and the right side was sions of the horizontal tube were attached to a single board, ‘side 2’. We did not call these ‘left and right’ as occasionally a parallel to each other. The key difference was that only one of second camera angle was consulted to confirm stone inser- the two tubes was baited with a walnut reward, the testing tions and the other cameras were pointed from different an- protocol was otherwise the same as the critical test phases. gles, hence left and right may have made this confusing whilst The subjects were still only provided with ten stones to try and coding. A stone was classified as inserted into the tube if the obtain the rewards and they were still given 10 min per trial to subjects placed it inside the tube, released it from their beak, try and obtain the reward. If they obtained the reward, the and then it stayed inside the tube without falling out. apparatus was removed and re-baited; however, the reward Sometimes, subjects removed stones from the tubes and then was randomly placed in either of the two different tubes in placed them back inside the tube, either on the same or on the different trials in a counterbalanced schedule, i.e. rewards other side. If the stone was picked up from inside the tube, were not just swapped between the tubes on different trials removed from within the tube entirely, then placed back inside but were placed in each of the two tubes an equal number of the tube and released, it counted as another insertion. In this times over all trials. A more stringent success criterion was manner, it was possible for more than ten stone insertions to used in this phase. Subjects had to obtain the reward 12 times be counted per trial even though there were only ten stones in a row in four sessions (three times per session), i.e. without available for each trial. In the parallel tubes and blocked tube unsuccessful trials between the successful trials. This was to transfer tasks, the number of stones inserted on each side of ensure that random placement of stones would not lead to both tubes was counted, i.e. there were four possible places subjects reaching the success criteria in this phase. If they where the stones could be inserted. For both of these tasks, succeeded, they moved onto the blocked tube task,but if they one tube was the ‘correct tube’ and the other was the ‘wrong failed, they stopped testing. tube’, as rewards could not be obtained from the ‘wrong tube’. The blocked tube task used the same apparatus as the In all of the test trials, from the pre-test and critical tests to parallel tubes, but a ‘bung’ attachment was created to block the parallel tubes and blocked tube transfer trials, we also one end of one of the tubes (Fig. 1h). This transfer task was counted the number of times subjects placed stones into the used to see if the subjects could recognise they needed to vertical-tube on top of the horizontal tube. In some trials, push the reward to a location that was then accessible to subjects again removed and re-inserted stones on multiple themselves, not that the rewards just needed to be pushed. occasions into these tubes. The same rule as above was In this task, both tubes were now baited, but the bung would followed: if the stone was picked up, removed entirely from prevent the subjects from pushing the reward out of the tube, then re-inserted and released, it was counted as another blocked tube. Between every trial the bung was randomly stone insertion. Hence, in some trials many more than two repositioned in one of the four available positions, again in ‘stone-in-top’ insertions were counted even though the verti- a counterbalanced fashion. The success criterion would cal tube could only hold two stones at once. In the parallel tube have been the same as the parallel tubes task, but no subject and blocked tube trials, there was no discrimination between succeeded here. This was the last available task in the subjects placing stones in the top of the ‘correct’ or the experiment. ‘wrong’ tube. Ten percent of the videos (28 sessions, 38 trials; these were Behavioural coding randomly selected from all of the pre-test, critical tests and parallel/blocked tubes tests) were then coded by a second All experiments were recorded on four static CCTV cameras. observer to check inter-observer reliability on the number of These covered all angles of the testing room. Two recordings stones inserted into both the top and the sides of the tube. from separate cameras were saved for each experimental ses- There was ‘excellent’ reliability between the two observers sion, but more recordings were saved if it was necessary for (Intraclass correlation = 93.9% consistency, R-package “irr”). specific trials, for example if there was partial occlusion of a Finally, we also coded subjects’ exploration of the stones camera view from the subject standing in the way. Behaviours and the apparatus. This included the latency until subjects were scored from the videos using Solomon coder (András touched the stones and the apparatus in each trial as well as Péter, solomon.andraspeter.com). the amount of time the subjects touched the stones and appa- Firstly, the location of the placement of the stone tools was ratus in each trial. Finally, the amount of time the subjects recorded. For every trial in the ‘pre-test’, ‘critical test’ and spent touching the apparatus with the stones (combining the ‘critical test 2’ the numberofstonesthatsubjectsplacedin objects together) was also recorded. This data was not used in each side of the tube was counted. The apparatus was always the final analysis, but a summary of it is provided in placed into the testing rooms in exactly the same position and Supplementary Table 2 (OSM), and the complete data are orientation so the locations we recorded of the stones was with the raw data to be found on figshare (O’Neill & ‘absolute’, i.e. the left side of the tubes from one specific Bayern, 2020). 116 Learn Behav (2021) 49:106–123 Fig. 3 The individual mean number of stone insertions into each side of phase of testing. There are fewer subjects in critical test 2 as not all the tube per trial. Grey bars indicate that individuals failed in that phase of individuals passed the short-tube experience, which was a requirement testing, black bars indicate that individuals consistently succeeded in that to complete critical test 2 Results glaucogularis), managed this in critical test 2 after the first short-tube experience phase (Fig. 3; Supplementary Figs. 5 In total, three birds (two out of eight Ara glaucogularis and and 6 (OSM)). These two subjects only solved the task 11 one out of nine Ara ambiguus) were able to consistently times (rather than 12) before being moved onto the follow- solve the task in the critical tests (Fig. 3). One Ara up tasks due to experimenter error. Enya solved the task 11 glaucogularis (Lady) managed to do this in critical test 1 timesin12trialsand Captainsolvedthe task 11 times in16 after the stick experience phase (Fig. 3; Supplementary Fig. trials. All three of these subjects had four trials in which 3 (OSM)); she solved the task 12 times in 16 trials. The they obtained the reward without error, i.e. without other two, Enya (Ara ambiguus)and Captain (Ara inserting a counter-productive stone into the opposing end Learn Behav (2021) 49:106–123 117 Fig. 4 Thenumber of stones insertedoneachside ofeach tubeinthe specifically from one direction. Sometimes this meant she inserted parallel tubes task. The top chart of each bird shows the stones inserted in stones into the unbaited tube, sometimes it led to success. Enya (Ara the baited tubes, the bottom shows the unbaited tubes. Lady (Ara ambiguus) appeared to start inserting stones more randomly at this glaucogularis) appeared to insert stones into either of the tubes, but stage, so she had very little success of the tube from their initial stone construction. Of these the ends of the tube in this stage of the experiment three, only Captain could consistently obtain the reward in (Supplementary Figs. 1 and 2 (OSM)). After the stick the follow-up parallel tubes task, but he was unable to do so experience phase, in the first critical test phase, three of the in the blocked tube task (Figs. 4 and 5). Ara glaucogularis started to consistently insert stones into the None of the subjects were able to solve the task in the pre- ends of the tube (Fig. 3;Supplementary Fig. 3 (OSM)), where- test phase (Table 1;Fig. 3; Supplementary Figs. 1 and 2 as none of the Ara ambiguus did so at this stage (Fig. 3; (OSM)). Additionally, very few subjects inserted stones into Supplementary Fig. 4 (OSM)). Fig. 5 The number of stones inserted by Captain into the ‘correct’ and the inaccessible), tube. He did not have an absolute side preference in the ‘wrong’ tube in the parallel tubes and the blocked tube task. The top chart parallel tube task, but had successful trials from inserting stones from of each task shows the stones inserted in the baited tubes, the bottom either side. At this stage he also had many trials in which he didn’t make shows the unbaited tubes. Captain mostly avoided the unbaited tube in the error of inserting rocks from the opposing side which he had started. the parallel tubes task, but began to fail trials in the blocked tube task by He did appear to lose motivation in the blocked tube task inserting tubes into this baited, but blocked (so the reward was 118 Learn Behav (2021) 49:106–123 One more Ara glaucogularis, Mr Huang, had a few suc- also the only individual who consistently, and spontaneously, cessful trials in critical test 2, but was unable to repeat this inserted stones into the horizontal tube in the pre-test (Fig. 3, behaviour consistently (Supplementary Fig. 5 (OSM)). Of the Supplementary Figs 1 and 2 (OSM)), prior to any functional eight Ara ambiguus that failed, six were unable to complete experience examples. Thus it is also possible that Lady did not the short-tube experience, both before and after the pre- need to learn anything functional from the pre-inserted stick inserted stone experience, so they were never able to continue experience in order to be successful. The pre-inserted stick on to critical test 2 (Table 1). The other two Ara ambiguus experience may have just showed her that the walnut reward failed at critical test 2 (Supplementary Fig. 6 (OSM)). Of the was obtainable, and thus motivated her to persist with the six Ara glaucogularis that failed, two were unable to complete same exploratory behaviour of inserting stones that she had the short-tube experience (both before and after the pre- already shown in the pre-test. inserted stone experience; Table 1) and the remaining four Captain (Ara glaucogularis) and Enya (Ara ambiguus) stopped at critical test 2 (Supplementary Fig. 5 (OSM)). were both able to consistently manufacture the stone construc- Thus in total, five Ara glaucogularis and three Ara ambiguus tion in critical test 2, after they had the short tube functional reached critical test 2. All the birds that attempted critical test experience. This experience showed the birds that a single 2 inserted multiple stones into the apparatus, but most failed stone was able to push a reward out of a much shorter hori- because they tried to insert them from both sides of the tube zontal tube. It was a much more direct example of the effect (Fig. 3; Supplementary Figs. 5 and 6 (OSM)). the stones could have on the rewards in the tube. Innovation of There was a decrease in the number of times the subjects the stone construction at this stage of the experiment is much placed stones in the vertical tube on top of the apparatus over less likely to represent an innovation based on functional un- the pre-test, critical test and critical test 2 (Supplementary derstanding. Instead, the birds are likely to have just been Fig. 8 (OSM)). All but two birds inserted stones into this top repeating the learned successful behaviour of inserting a sin- tube in at least one trial in the pre-test, and many inserted gle stone to get the reward, and this in turn led to them making stones in more than one trial. the multi-stone construction, which eventually pushed the Example videos of every phase of the experiment, includ- walnut out of the horizontal tube. The fact that all the subjects ing failed and successful trials of the critical tests, can be that took part in critical test 2 began to insert multiple stones at found on figshare (O’Neill & von Bayern, 2020). this stage (Fig. 3) suggeststhatthiswas the underlyingmoti- vation to all the subjects’ behaviour at this stage. The differ- ence with Captain and Enya is that for some reason they began Discussion to more consistently insert stones from a single side of the tube. This may have started initially as a side preference, but Three subjects (two Ara glaucogularis and one Ara its possible that after multiple successes they began to more ambiguus) were able to consistently manufacture a multi- purposefully manufacture the tool from a single side only. stone construction to obtain the walnut reward. These birds The three successful subjects each manufactured the multi- were able to create a functional stone-construction by combin- stone construction without errors on four occasions during ing multiple stone components one behind the other to push their successful critical test runs. Out of context, the videos the walnut out of the tube. Due to their successes coming at of these trials are compelling and can look fairly convincing different stages of the experiment, and because of some of the that the subjects were building these stone constructions with behaviours associated with their successful manufacture, it is ‘purpose’ to push the walnut rewards (see figshare at likely that these successful subjects had a different under- O’Neill & von Bayern, 2020, for videos). However, this in- standing of their stone constructions and innovated them in terpretation should be considered with extreme caution. different ways. Captain was able to manufacture the stone construction with- Lady (Ara glaucogularis) manufactured her first successful out error on his first trial in critical test 2 (Supplementary Fig. stone construction in critical test 1 after having the functional 5 (OSM); also shown in the supplementary videos) but sub- experience that a stick tool could push the reward out of the sequently produced many error-laden trials. Enya had an ex- tube. At this stage of the experiment, she had not had a specific tremely error laden (but successful) first trial in critical test 2, positive experience that the stones could be used to obtain the but then immediately produced two ‘perfect’ trials in the mi- reward in this specific apparatus (although she did have pre- nutes after this (trials 2 and 3, Supplementary Fig. 6 (OSM); vious experience that putting singular stones into a different also shown in the supplementary videos). She also subse- apparatus could lead to rewards; O’Neill et al., 2020). This quently produced many error-laden trials. Lady also shows kind of innovation suggests its possible she may have had similar behaviour of a mix of ‘perfect’ and error-laden some form of functional understanding of the stick tool exam- trials (Supplementary Fig. 3 (OSM) and supplementary ple, and had positioned the stones to recreate its underlying videos). The fact that these subjects only produced these ‘per- purpose. On the other hand, it should be noted that Lady was fect’ trials erratically makes it very difficult to interpret why Learn Behav (2021) 49:106–123 119 they sometimes manufacture the multi-stone construction standing closer to the reward, and it was likely difficult to without error, and sometimes they do not. inhibit repeating the stone insertion behaviour they had learnt To an extent, all the successful subjects appeared to show led to a reward even though it pushed the reward further away an egocentric understanding (Woodward, 2011) of their be- from them again. Interestingly, this exact same mistake was haviour, i.e. they were able to replicate it consistently and observed in the original experiment that was run with capu- rapidly in trials following their initial success. They may have chins (Visalberghi & Trinca, 1989). It is also an error that has recognised which of their behaviours (repeatedly inserting been noted in other parrots taking part in an experiment with a stones from a single side of the horizontal tube) was producing very similar setup, the trap tube (Liedtke et al., 2011), and is the reward outcome. However, Captain did have one success- something we had to control for when we tested a version of ful trial in critical test 1 but was not able to replicate it at this the trap tube on the same group of parrots tested in the current stage of the experiment and Mr Huang (Ara glaucogularis) experiment (O’Neill et al., 2018). even had four successful trials early in critical test 2, but was This counter-productive stone insertion mistake was also then unable to replicate these successes. Neither of these fail- observed in all of the unsuccessful subjects that were inserting ures appeared to be due to a lack of motivation as both were stones into the tube. All of these unsuccessful subjects went attempting to obtain rewards by inserting multiple stones into through the short-tube experience where it is likely that they the horizontal tube, but they were just inserting them from learned a simple heuristic rule of ‘insert stone, obtain reward’. both sides of the tube at once which led to the reward being Simply expanding this rule in a quantitative manner failed in trapped. It is unclear exactly which cue made the subjects critical test 2 as the subjects had to ensure they inserted suffi- recognise (or not recognise in Mr Huang’s instance) that the cient stones from a single side. It is thus possible that the stones must be more consistently inserted from only a single successful birds did not have a different ‘rule’ to the other side of the horizontal tube. birds; they may have just had a weak side preference whereas Regarding the successful birds, they made a number of the other subjects had no side preference at all. The three mistakes when manufacturing their stone constructions. successful subjects (Enya, Lady and Captain) only collected Each of the successful subjects only had four trials in their the rewards from one side of the apparatus in all of the suc- successful critical test in which they only inserted stones from cessful trials in the critical tests, i.e. they always built their a single side of the tube (Fig. 3; Supplementary Figs. 3, 5 and stone construction from a specific side of the apparatus, they 6 (OSM)). On all their other trials (the majority), they inserted did not flexibly build it from either side. stones in both ends of the tube, thus on both sides of the Considering this side preference, it suggests why Lady and reward. This type of ‘mistake’ is not just inefficient, but is Enya failed in the parallel tubes task, in which only one tube counter-productive. It meant that more than the minimum was baited. Figure 4 shows that Lady was clearly inserting number of stones had to be inserted in the original end of stones into both tubes (both the baited and unbaited tube), just the tube to push out both the erroneous stones and then the from a specific side (also see supplementary videos at O'Neill reward. This suggests that the successful subjects did not have & von Bayern, 2020). Although Enya was also mostly a mental representation of the final multi-stone tool that they inserting stones from one side of the two tubes, her data are were creating. This contrasts with a tool using corvid example, less convincing in supporting this side preference speculation. the New Caledonian Crow (corvus moneduloides), which Interestingly, Captain appears to have stopped having a were able to tear a piece of paper off a larger piece so that it side bias in this first parallel tubes follow-up task, and had resembled a token of the same size to that which they had successful trials in which he made his stone construction from previously learned could be traded for a reward, i.e. they could both directions of the tube (Fig. 5; examplesofsuccess from manufacture something that was similar to a template they had both sides in supplementary videos at O'Neill & von Bayern, mentally stored (Jelbert et al., 2018). Observationally, the 2020). He also vastly improved the efficiency with which he counter-productive stone insertion behaviour often happened made the multi-stone construction, exhibiting many trials in after the subjects had manufactured an almost complete multi- which he did not make the ‘mistake’ of inserting stones from stone construction from one side, then walked to the other side both sides of the tube and he only inserted stones from one to check if they had pushed the reward far enough. When they side of the tube to obtain the reward (11 out of 18 successful could not reach the reward, they often inserted a stone into this trials, although in four of those he also inserted stones into the other side of the tube that they were now standing next to (the un-baited tube as well). Observationally, one way that Captain supplementary videos provide examples of this error; see behaved differently to Enya and Lady is that he showed a figshare at O’Neill & von Bayern, 2020, for videos). This preference to stand on top of the apparatus when he was may have been a self-control issue, as the birds would have interacting with it, as opposed to the surface next to it. From had to walk back to the initial side of the tube in which they personal observations, Captain was a more nervous bird. One had already inserted stones in order to continue their stone behaviour the parrots manifested when they were nervous was construction already on the go. Instead, they were now try to stand on a higher perch and in this case the apparatus 120 Learn Behav (2021) 49:106–123 was the highest point in the testing room, hence he would pre-test, was to insert a stone into this vertical tube stand on top of it. From this standpoint he had to move shorter (Supplementary Fig. 8 (OSM)). Only two subjects did not distances to interact with the two ends of each tube, and was insert stones into the vertical tube in the pre-test and only also able to transfer between the tubes very rapidly. Enya and one further subject did not place stones in this vertical tube Lady, on the other hand, had to walk ‘the long way round’ if in more than one trial. It could be argued that placing a stone in they wanted to access different ends to the tubes. It’s possible the pipe is evidence of a poor means-ends understanding of that standing on top of the apparatus slightly reduced the self- their actions and that their behaviour appeared to be guided control issues, as described above, which is what made more by trial-and-error exploration. However, not dropping Captain more efficient in creating the stone construction in stones into that tube would have required suppressing a pre- the parallel tubes follow-up task. viously reinforced behaviour and, thus, it should not be Lady and Enya failed the parallel tubes transfer task, al- regarded as a clear-cut conclusion. Moreover, any exploration, though it is possible that a lack of attention combined with even if it was misdirected and could not lead to success, lack of self-control (as discussed earlier), rather than a lack of helped the subjects learn more about the apparatus and its means-end understanding of the stone construction, was the functional properties (Reader, 2015). The one subject that reason for this. Lady especially placed many stones in the un- did not place stones in the top was the least explorative bird baited ‘distractor’ tube and Enya especially began to place overall (he was the only bird to not make any combinatory many stones in from both ends of the rewarded tube, although actions between the stones and the horizontal tube both made both errors at this stage. They both regularly ran (Supplementary Table 1)) and had no success in any of the out of stones. This behaviour suggests that Lady and Enya current tasks (Hannibal), nor did he have any success in the were probably not manufacturing their stone construction with previous task where the subjects had learned to drop a stone the target to specifically push the reward, but that instead they into a vertical tube (O’Neill et al., 2020). Further to this, as likely had a more simple understanding of the apparatus, e.g. subjects did not place many stones into the sides of the hori- that inserting multiple stones would – at some point – make a zontal tube in the pre-test but placed many stones in the ver- reward appear. Further, it may be a reflection of the macaw’s tical tube during the pre-test, it suggests that the subjects who relative lack of self-control to suppress a previously reinforced eventually had success in the horizontal tube task did not movement (Kabadayi et al., 2017), since they had never ex- succeed just because of carried-over behaviour from the suc- perienced a cost to inefficient stone inserting and had only cess in the previous experiment. Thus, the most vital aspect to been faced with tubes containing rewards previously. be learned from the addition of this vertical-tube was that Limited use of the stone resources only became an issue in subjects will try a previously used and successful method on the transfer tasks, and therefore subjects may not have a novel task before attempting something new. recognised that they could run out of them. Perhaps if they Such a behavioural strategy, namely to exploit a similar had been given more trials, they would have learnt to be more feeding opportunity to one that is previously known, is efficient and to pay attention where to insert the stones. thought to be a key mechanism characterising good innovators Captain, on the other hand, was able to direct his multi-stone (Tebbich et al., 2016). The innovation structure described in construction making behaviour towards the baited tube, which Tebbich et al. (2016) provides a useful framework for how suggests he recognised that he had to target the stones towards each subject’s behavioural interaction with the apparatus was the reward, not just the tubes in general. However, he was possibly guided. As an example, the most successful subject, unsuccessful on the blocked-tube task, in which both tubes Captain, likely started the experiment with the knowledge were baited but one had a blocked end. This suggests that from previous experimental apparatuses to ‘insert stone in while recognising the need to direct stones towards the re- top’ to access food. When this failed, he explored the appara- ward, he did not grasp that the reward must be pushed to an tus and moved onto ‘insert stones in tube ends’ (see Captain’s accessible area, but only that the reward must be pushed. behaviour in the first critical test, Fig. 3), which he then newly There were clearly limitations to the majority of the macaws discovered as a favourable interaction. He optimised this rule overall understanding of the physical properties of this task. over successive interactions to discover that he had to ‘insert In this test, an additional vertical tube was attached to the stones from a specific side’ (see Captain’s behaviour in his apparatus that had no effect on the reward outcome. This was successful critical test 2), and finally further optimisation led to see if the subjects continued to use a basic heuristic of him to ‘insert stones towards the reward from a single side’ ‘place stone on top of apparatus’ to get the reward, as this (see Captain’s behaviour in the parallel tubes task). All sub- was something that they had learnt led to success in a previous jects appeared to go through this sort of heuristic ‘evolution’, experiment (O’Neill et al., 2020; all except two subjects had but many of them never moved on from the first two stages. learnt this behaviour (see Supplementary Table 2 in this This suggests that almost all of the subjects were capable of paper)). Indeed, one of the first actions that almost all of the recognising the feeding opportunity from the apparatus, and subjects did, in their initial exploration of the apparatus in the even discovering that there was a favourable behavioural Learn Behav (2021) 49:106–123 121 Funding Open Access funding enabled and organized by Projekt DEAL. interaction through exploration. What the majority of subjects struggled with was the repetition and testing required to opti- Compliance with ethical standards mise this favourable interaction to make exploiting this food as a worthwhile opportunity. A critical stage of the innovation Ethics Statement All applicable international, national, and institutional framework (Tebbich et al., 2016) that is missing from our guidelines for the care and use of animals were followed. In accordance experiment is the opportunity to socially learn from other’s with the German Animal Welfare Act of 25 May 1998, Section V, Article behaviour. If the other subjects could have watched Captain, 7 and the Spanish Animal Welfare Act 32/2007 of 7 November 2007, Preliminary Title, Article 3, the study was classified as non-animal ex- Lady and Enya, perhaps they may have improved their inter- periment and did not require any approval from a relevant body. action with the equipment. This is likely what would happen in the wild when conspecifics have discovered a novel feeding Open Access This article is licensed under a Creative Commons opportunity. Nevertheless, it is interesting that none of the Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as subjects ever showed a truly flexible behaviour that would long as you give appropriate credit to the original author(s) and the suggest a functional understanding of the underlying mechan- source, provide a link to the Creative Commons licence, and indicate if ics of the task, preferring always to instead rely on a simple changes were made. The images or other third party material in this article rule or heuristic (Hutchinson & Gigerenzer, 2005). This sug- are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the gests that these parrots are unlikely to use a complex mechan- article's Creative Commons licence and your intended use is not ical causal understanding of these kinds of tasks (Johnson & permitted by statutory regulation or exceeds the permitted use, you will Ahn, 2017;Woodward, 2011). need to obtain permission directly from the copyright holder. To view a Overall, the subjects in this task showed less, or possibly copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. zero, functional understanding of their multi-stone construc- tions when compared to other parrots’ functional understand- ing of tools they can use in problem-solving tasks (Auersperg References et al. 2011; Auersperg, von Bayern, Gajdon et al. 2011, 2018; Auersperg et al., 2012;Laumer etal., 2017). As stated, the Auersperg, A. M. I., Huber, L., & Gajdon, G. K. (2011a). Navigating a only individual that appeared to be using the stones in order to tool end in a specific direction: Stick-tool use in kea (Nestor directly manipulate the position of the walnut was Captain, as notabilis). Biology Letters, 7(6), 825–828. https://doi.org/10.1098/ he was the only one that would place the stones in the direc- rsbl.2011.0388 tion towards the walnut in the parallel tubes task. So although Auersperg, A. M. I., Köck, C., O’Hara, M., & Huber, L. (2018). Tool making cockatoos adjust the lengths but not the widths of their tools both species of macaw in this study appeared to show some to function. PLOS ONE, 13(11), e0205429. https://doi.org/10.1371/ evidence of innovative object manipulation in order to solve a journal.pone.0205429 problem-solving task, we still need to further study the differ- Auersperg, A., von Bayern, A., Gajdon, G., Huber, L., & Kacelnik, A. ences between different parrot species (in this instance: Kea, (2011b). Flexibility in Problem Solving and Tool Use of Kea and New Caledonian Crows in a Multi Access Box Paradigm. PLoS Goffin’s cockatoos, Ara glaucogularis and Ara ambiguus) ONE, 6(6), e20231. https://doi.org/10.1371/journal.pone.0020231 that make only some of them capable understanding the func- Auersperg, Alice M.I., Szabo, B., von Bayern, A. M. P., & Kacelnik, A. tional properties of more complex physical innovations. (2012). Spontaneous innovation in tool manufacture and use in a Goffin’scockatoo. Current Biology, 22(21), R903–R904. https:// Supplementary Information The online version contains supplementary doi.org/10.1016/j.cub.2012.09.002 material available at https://doi.org/10.3758/s13420-020-00449-y. Bird, C. D., & Emery, N. J. (2009). Insightful problem solving and cre- ative tool modification by captive nontool-using rooks. Proceedings of the National Academy of Sciences, 106(25), 10370–10375. Acknowledgements We thank Luisana Carballo and Anastasia https://doi.org/10.1073/pnas.0901008106 Krasheninnikova for their comments on earlier versions of the Borsari, A., & Ottoni, E. B. (2005). Preliminary observations of tool use manuscript. in captive hyacinth macaws (Anodorhynchus hyacinthinus). Animal We thank the Loro Parque and its president, Mr Wolfgang Kiessling, for Cognition, 8(1), 48–52. https://doi.org/10.1007/s10071-004-0221-3 their support, the access to the birds and the research facilities. We thank the Breen, A. J., Guillette, L. M., & Healy, S. D. (2016). What Can Nest- Loro Parque Fundación and its president Mr Christoph Kiessling for their Building Birds Teach Us? Comparative Cognition & Behavior collaboration and the staff of the Loro Parque Fundación, the animal Reviews, 11,83–102. https://doi.org/10.3819/ccbr.2016.110005 caretakers and the veterinary department for their support. Burkart, J. M., Schubiger, M. N., & van Schaik, C. P. (2017). The evo- LO’N is a member of the International Max Planck Research School lution of general intelligence. Behavioral and Brain Sciences, 40, (IMPRS) for Organismal Biology. The study was funded by the Max- e195. https://doi.org/10.1017/S0140525X16000959 Planck Society. Byrne, R., Corp, N., & Byrne, J. (2001). Estimating the complexity of animal behaviour: How mountain gorillas eat thistles. Behaviour, Author contributions The design of the experiment was initially con- 138(4), 525–557. https://doi.org/10.1163/156853901750382142 ceived by LO’N and AvB, with contributions from RR and RH. The Chow, P. K. Y., Lea, S. E. G., & Leaver, L. A. (2016). How practice experiments were carried out by RR and RH. Videos were coded by makes perfect: The role of persistence, flexibility and learning in LO'N and RH. Data were visualised by LO’N. The initial manuscript problem-solving efficiency. Animal Behaviour, 112,273–283. was written by LO’N with feedback from AvB. https://doi.org/10.1016/j.anbehav.2015.11.014 122 Learn Behav (2021) 49:106–123 Emery, N. J., & Clayton, N. S. (2004). The Mentality of Crows: Behaviour, 156(5–8), 391–407. https://doi.org/10.1163/1568539X- Convergent Evolution of Intelligence in Corvids and Apes. 00003527 Science, 306(5703), 1903–1907. https://doi.org/10.1126/science. Lambert, M. L., Schiestl, M., Schwing, R., Taylor, A. H., Gajdon, G. K., Slocombe, K. E., & Seed, A. M. (2017). Function and flexibility of Fragaszy, D. M., & Mangalam, M. (2018). Tooling. Advances in the object exploration in kea and New Caledonian crows. Royal Society Study of Behavior, 50, 177–241. https://doi.org/10.1016/bs.asb. Open Science, 4(9), 170652. https://doi.org/10.1098/rsos.170652 2018.01.001 Lambert, M. L., Seed, A. M., & Slocombe, K. E. (2015). A novel form of Girndt, A., Meier, T., & Call, J. (2008). Task constraints mask great apes’ spontaneous tool use displayed by several captive greater vasa par- ability to solve the trap-table task. Journal of Experimental rots ( Coracopsis vasa ): Table 1. Biology Letters, 11(12), Psychology: Animal Behavior Processes, 34(1), 54–62. https://doi. 20150861. https://doi.org/10.1098/rsbl.2015.0861 org/10.1037/0097-7403.34.1.54 Laumer, I. B., Bugnyar, T., Reber, S. A., & Auersperg, A. M. I. (2017). Goodman, M., Hayward, T., & Hunt, G. R. (2018). Habitual tool use Can hook-bending be let off the hook? Bending/unbending of pliant innovated by free-living New Zealand kea. Scientific Reports, tools by cockatoos. Proceedings of the Royal Society B: Biological 8(1), 13935. https://doi.org/10.1038/s41598-018-32363-9 Sciences, 284(1862), 20171026. https://doi.org/10.1098/rspb.2017. Griffin, A. S., Diquelou, M., & Perea, M. (2014). Innovative problem solving in birds: A key role of motor diversity. Animal Behaviour, Liedtke, J., Werdenich, D., Gajdon, G. K., Huber, L., & Wanker, R. 92,221–227. https://doi.org/10.1016/j.anbehav.2014.04.009 (2011). Big brains are not enough: Performance of three parrot spe- Gutiérrez-Ibáñez, C., Iwaniuk, A. N., & Wylie, D. R. (2018). Parrots have cies in the trap-tube paradigm. Animal Cognition, 14(1), 143–149. evolved a primate-like telencephalic-midbrain-cerebellar circuit. https://doi.org/10.1007/s10071-010-0347-4 Scientific Reports, 8(1), 9960. https://doi.org/10.1038/s41598-018- Martin-Ordas, G., Call, J., & Colmenares, F. (2008). Tubes, tables and 28301-4 traps: Great apes solve two functionally equivalent trap tasks but Heinsohn, R., Zdenek, C. N., Cunningham, R. B., Endler, J. A., & show no evidence of transfer across tasks. Animal Cognition, Langmore, N. E. (2017). Tool-assisted rhythmic drumming in palm 11(3), 423–430. https://doi.org/10.1007/s10071-007-0132-1 cockatoos shares key elements of human instrumental music. Martin-Ordas, G., Jaek, F., & Call, J. (2012). Barriers and traps: Great Science Advances, 3(6), e1602399. https://doi.org/10.1126/sciadv. apes’ performance in two functionally equivalent tasks. Animal Cognition, 15(5), 1007–1013. https://doi.org/10.1007/s10071-012- Herculano-Houzel, S. (2017). Numbers of neurons as biological corre- 0504-z lates of cognitive capability. Current Opinion in Behavioral McCormack, T., Hoerl, C., & Butterfill, S. (2011). Tool use and causal Sciences, 16,1–7. https://doi.org/10.1016/j.cobeha.2017.02.004 cognition. Oxford University Press. https://doi.org/10.1093/acprof: Heyes, C. M. (1993). Anecdotes, training, trapping and triangulating: Do oso/9780199571154.001.0001 animals attribute mental states? Animal Behaviour, 46(1), 177–188. Menchetti, M., Mori, E., & Angelici, F. M. (2016). Effects of the Recent https://doi.org/10.1006/anbe.1993.1173 World Invasion by Ring-Necked Parakeets (Psittacula krameri). In Hutchinson, J. M. C., & Gigerenzer, G. (2005). Simple heuristics and F. M. Angelici (Ed.), Problematic Wildlife (253–266). Springer rules of thumb: Where psychologists and behavioural biologists International Publishing. https://doi.org/10.1007/978-3-319-22246- might meet. Behavioural Processes, 69(2), 97–124. https://doi.org/ 2_12 10.1016/j.beproc.2005.02.019 Mulcahy, N. J., & Call, J. (2006). How great apes perform on a modified Iwaniuk, A. N., Dean, K. M., & Nelson, J. E. (2005). Interspecific trap-tube task. Animal Cognition, 9(3), 193–199. https://doi.org/10. Allometry of the Brain and Brain Regions in Parrots 1007/s10071-006-0019-6 (Psittaciformes): Comparisons with Other Birds and Primates. O’Hara, M., Mioduszewska, B., Haryoko, T., Prawiradilaga, D. M., Brain, Behavior and Evolution, 65(1), 40–59. https://doi.org/10. Huber, L., & Auersperg, A. (2019). Extraction without tooling 1159/000081110 around—The first comprehensive description of the foraging- and Jelbert, S. A., Hosking, R. J., Taylor, A. H., & Gray, R. D. (2018). Mental socio-ecology of wild Goffin’s cockatoos (Cacatua goffiniana). template matching is a potential cultural transmission mechanism Behaviour, 156(5–8), 661–690. https://doi.org/10.1163/1568539X- for New Caledonian crow tool manufacturing traditions. Scientific Reports, 8(1), 8956. https://doi.org/10.1038/s41598-018-27405-1 Olkowicz, S., Kocourek, M., Lučan, R. K., Porteš, M., Fitch, W. T., Johnson, S. G. B., & Ahn, W. (2017). Causal Mechanisms. In M. R. Herculano-Houzel, S., & Němec, P. (2016). Birds have primate- Waldmann (Ed.), The Oxford handbook of Causal like numbers of neurons in the forebrain. Proceedings of the Reasoning. Oxford University Press. https://doi.org/10.1093/ National Academy of Sciences, 113(26), 7255–7260. https://doi. oxfordhb/9780199399550.013.12 org/10.1073/pnas.1517131113 Kabadayi, C., Krasheninnikova, A., O’Neill, L., van de Weijer, J., O'Neill, L., & von Bayern, A. M. P. (2020). Supplementary materials to Osvath, M., & von Bayern, A. M. P. (2017). Are parrots poor at ‘Innovative problem solving in macaws’. figshare. Media. https:// motor self-regulation or is the cylinder task poor at measuring it? doi.org/10.6084/m9.figshare.12349418.v2 Animal Cognition, 20(6), 1137–1146. https://doi.org/10.1007/ O’Neill, L., Picaud, A., Hastings, R., Buffenoir, N., Gahr, M., & von s10071-017-1131-5 Bayern, A. M. P. (2020). Causal Understanding of the Stone Kacelnik, A. (2009). Tools for thought or thoughts for tools? Proceedings Dropping Task in Two Species of Macaw [Preprint]. BiorXiv. of the National Academy of Sciences, 106(25), 10071–10072. https://doi.org/10.1101/2020.08.24.264390 https://doi.org/10.1073/pnas.0904735106 O’Neill, L., Picaud, A., Maehner, J., Gahr, M., & von Bayern, A. M. P. Köhler, W. (1917) 1925. The mentality of apes (translated from the second revised edition by Ella Winter). New York: Harcourt, (2018). Two macaw species can learn to solve an optimised two-trap Brace, & World. problem, but without functional causal understanding. Behaviour, 156(5–8), 691–720. https://doi.org/10.1163/1568539X-00003521 Krasheninnikova, A., Berardi, R., Lind, M.-A., O’Neill, L., & von Bayern, A. M. P. (2019). Primate cognition test battery in parrots. Osuna-Mascaró, A. J., & Auersperg, A. M. I. (2018). On the brink of tool Behaviour, 156(5–8), 721–761. https://doi.org/10.1163/1568539X- use? Could object combinations during foraging in a feral Goffin’s 0003549 cockatoo (Cacatua goffiniana) result in tool innovations? Animal Lambert, M. L., Jacobs, I., Osvath, M., & von Bayern, A. M. P. (2018). Behavior and Cognition, 5(2), 229–234. https://doi.org/10.26451/ Birds of a feather? Parrot and corvid cognition compared. abc.05.02.05.2018 Learn Behav (2021) 49:106–123 123 Osvath, M., Kabadayi, C., & Jacobs, I. (2014). Independent Evolution of Van Horik, J. O., Clayton, N. S., & Emery, N. J. (2012). Convergent Similar Complex Cognitive Skills: The Importance of Embodied Evolution of Cognition in Corvids, Apes and Other Animals. In Degrees of Freedom. Animal Behavior and Cognition, 1(3), 249. Shackelford, T.K. & Vonk, J. (Eds.), The Oxford Handbook of https://doi.org/10.12966/abc.08.03.2014 Comparative Evolutionary Psychology. Oxford University Press. Reader, S. M. (2015). Causes of Individual Differences in Animal https://doi.org/10.1093/oxfordhb/9780199738182.013.0005 Exploration and Search. Topics in Cognitive Science, 7(3), 451– van Horik, J. O., & Madden, J. R. (2016). A problem with problem 468. https://doi.org/10.1111/tops.12148 solving: Motivational traits, but not cognition, predict success on Shumaker, R. W., Walkup, K. R., & Beck, B. B. (2011). Animal tool novel operant foraging tasks. Animal Behaviour, 114, 189–198. behavior: The use and manufacture of tools by animals. Baltimore: https://doi.org/10.1016/j.anbehav.2016.02.006 Johns Hopkins University Press. Villegas-Retana, S. A., & Araya-H D. (2017). Consumo de almendro de Sol, D., Duncan, R. P., Blackburn, T. M., Cassey, P., & Lefebvre, L. playa (Terminalia catappa) y uso de hojas como herramienta por (2005). Big brains, enhanced cognition, and response of birds to parte del ave Ara ambiguus (Psittaciformes: Psittacidae)en Costa novel environments. Proceedings of the National Academy of Rica. UNED Research Journal, 9(2). https://doi.org/10.22458/urj. Sciences, 102(15), 5460–5465. https://doi.org/10.1073/pnas. v9i2.1894 Visalberghi, E., & Limongelli, L. (1994). Lack of comprehension of Sol, Daniel, Timmermans, S., & Lefebvre, L. (2002). Behavioural flexi- cause-effect relations in tool-using capuchin monkeys (Cebus bility and invasion success in birds. Animal Behaviour, 63(3), 495– apella). Journal of Comparative Psychology, 108(1), 15–22. 502. https://doi.org/10.1006/anbe.2001.1953 https://doi.org/10.1037/0735-7036.108.1.15 Taylor, A. H., & Gray, R. D. (2014). Is there a link between the crafting of Visalberghi, E., & Trinca, L. (1989). Tool use in capuchin monkeys: tools and the evolution of cognition?: Is cognitive evolution affected Distinguishing between performing and understanding. Primates, by tool crafting? Wiley Interdisciplinary Reviews: Cognitive 30(4), 511–521. https://doi.org/10.1007/BF02380877 Science, 5(6), 693–703. https://doi.org/10.1002/wcs.1322 von Bayern, A. M. P., Danel, S., Auersperg, A. M. I., Mioduszewska, B., Taylor, A. H., Hunt, G. R., Holzhaider, J. C., & Gray, R. D. (2007). & Kacelnik, A. (2018). Compound tool construction by New Spontaneous Metatool Use by New Caledonian Crows. Current Caledonian crows. Scientific Reports, 8(1), 15676. https://doi.org/ Biology, 17(17), 1504–1507. https://doi.org/10.1016/j.cub.2007. 10.1038/s41598-018-33458-z 07.057 von Bayern, Auguste M.P., Heathcote, R. J. P., Rutz, C., & Kacelnik, A. Tebbich, S., Griffin, A. S., Peschl, M. F., & Sterelny, K. (2016). From (2009). The Role of Experience in Problem Solving and Innovative mechanisms to function: An integrated framework of animal inno- Tool Use in Crows. Current Biology, 19(22), 1965–1968. https:// vation. Philosophical Transactions of the Royal Society B: doi.org/10.1016/j.cub.2009.10.037 Biological Sciences, 371(1690), 20150195. https://doi.org/10. Weir, A. A. S., & Kacelnik, A. (2006). A New Caledonian crow (Corvus 1098/rstb.2015.0195 moneduloides) creatively re-designs tools by bending or unbending Tella, J. L., Baños-Villalba, A., Hernández-Brito, D., Rojas, A., Pacífico, aluminium strips. Animal Cognition, 9(4), 317–334. https://doi.org/ E., Díaz-Luque, J. A., Carrete, M., Blanco, G., & Hiraldo, F. (2015). 10.1007/s10071-006-0052-5 Parrots as overlooked seed dispersers. Frontiers in Ecology and the Wood, G. A. (1984). Tool use by the Palm Cockatoo Probosciger Environment, 13(6), 338–339. https://doi.org/10.1890/1540-9295- aterrimus during display. Corella, 8,94–95. 13.6.338 Woodward, J. (2011). A Philosopher Looks at Tool Use and Causal Teschke, I., Wascher, C. A. F., Scriba, M. F., von Bayern, A. M. P., Understanding. In McCormack, T., Hoerl, C., Butterfill, S. (Eds.), Huml, V., Siemers, B., & Tebbich, S. (2013). Did tool-use evolve Tool Use and Causal Cognition. Oxford University Press. with enhanced physical cognitive abilities? Philosophical Transactions of the Royal Society B: Biological Sciences, 368(1630), 20120418–20120418. https://doi.org/10.1098/rstb. Supplementary data and videos 2012.0418 Data from the video coding and Supplementary videos can be found Thornton, A., & Lukas, D. (2012). Individual variation in cognitive per- on figshare at: https://doi.org/10.6084/m9.figshare.12349418.v1 formance: Developmental and evolutionary perspectives. Philosophical Transactions of the Royal Society B: Biological Sciences, 367(1603), 2773–2783. https://doi.org/10.1098/rstb. 2012.0214 Thornton, Alex, & Samson, J. (2012). Innovative problem solving in wild meerkats. Animal Behaviour, 83(6), 1459–1468. https://doi.org/10. Publisher’snote Springer Nature remains neutral with regard to jurisdic- 1016/j.anbehav.2012.03.018 tional claims in published maps and institutional affiliations. Toft, C. A., & Wright, T. F. (2016). Parrots of the wild: A natural history of the world’s most captivating birds. Berkeley, CA, University of California press. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Learning & Behavior Springer Journals http://www.deepdyve.com/lp/springer-journals/innovative-problem-solving-in-macaws-y1dyAulLs6

Loading next page...

References (66)

(Woodward, J. (2011). A Philosopher Looks at Tool Use and Causal Understanding. In McCormack, T., Hoerl, C., Butterfill, S. (Eds.), Tool Use and Causal Cognition. Oxford University Press.)
Woodward, J. (2011). A Philosopher Looks at Tool Use and Causal Understanding. In McCormack, T., Hoerl, C., Butterfill, S. (Eds.), Tool Use and Causal Cognition. Oxford University Press.
Woodward, J. (2011). A Philosopher Looks at Tool Use and Causal Understanding. In McCormack, T., Hoerl, C., Butterfill, S. (Eds.), Tool Use and Causal Cognition. Oxford University Press., Woodward, J. (2011). A Philosopher Looks at Tool Use and Causal Understanding. In McCormack, T., Hoerl, C., Butterfill, S. (Eds.), Tool Use and Causal Cognition. Oxford University Press.
(O’Neill, L., Picaud, A., Hastings, R., Buffenoir, N., Gahr, M., & von Bayern, A. M. P. (2020). Causal Understanding of the Stone Dropping Task in Two Species of Macaw [Preprint]. BiorXiv. 10.1101/2020.08.24.264390)
O’Neill, L., Picaud, A., Hastings, R., Buffenoir, N., Gahr, M., & von Bayern, A. M. P. (2020). Causal Understanding of the Stone Dropping Task in Two Species of Macaw [Preprint]. BiorXiv. 10.1101/2020.08.24.264390
O’Neill, L., Picaud, A., Hastings, R., Buffenoir, N., Gahr, M., & von Bayern, A. M. P. (2020). Causal Understanding of the Stone Dropping Task in Two Species of Macaw [Preprint]. BiorXiv. 10.1101/2020.08.24.264390, O’Neill, L., Picaud, A., Hastings, R., Buffenoir, N., Gahr, M., & von Bayern, A. M. P. (2020). Causal Understanding of the Stone Dropping Task in Two Species of Macaw [Preprint]. BiorXiv. 10.1101/2020.08.24.264390
(2009)
Tools for thought or thoughts for tools?
Proceedings of the National Academy of Sciences, 106
(2011)
Flexibility in Problem Solving and Tool Use of Kea and New Caledonian Crows in a Multi Access Box Paradigm
PLoS ONE, 6
(2017)
The evolution of general intelligence
Behavioral and Brain Sciences, 40
(2014)
Independent Evolution of Similar Complex Cognitive Skills: The Importance of Embodied Degrees of Freedom
Animal Behavior and Cognition, 1
(2018)
On the brink of tool use? Could object combinations during foraging in a feral Goffin’s cockatoo (Cacatua goffiniana) result in tool innovations?
Animal Behavior and Cognition, 5
(1994)
Lack of comprehension of cause-effect relations in tool-using capuchin monkeys (Cebus apella)
Journal of Comparative Psychology, 108
(2016)
What Can Nest-Building Birds Teach Us?
Comparative Cognition & Behavior Reviews, 11
(Lambert, M. L., Jacobs, I., Osvath, M., & von Bayern, A. M. P. (2018). Birds of a feather? Parrot and corvid cognition compared. Behaviour, 156(5–8), 391–407. 10.1163/1568539X-00003527)
Lambert, M. L., Jacobs, I., Osvath, M., & von Bayern, A. M. P. (2018). Birds of a feather? Parrot and corvid cognition compared. Behaviour, 156(5–8), 391–407. 10.1163/1568539X-00003527
Lambert, M. L., Jacobs, I., Osvath, M., & von Bayern, A. M. P. (2018). Birds of a feather? Parrot and corvid cognition compared. Behaviour, 156(5–8), 391–407. 10.1163/1568539X-00003527, Lambert, M. L., Jacobs, I., Osvath, M., & von Bayern, A. M. P. (2018). Birds of a feather? Parrot and corvid cognition compared. Behaviour, 156(5–8), 391–407. 10.1163/1568539X-00003527
(2002)
Behavioural flexibility and invasion success in birds
Animal Behaviour, 63
(Van Horik, J. O., Clayton, N. S., & Emery, N. J. (2012). Convergent Evolution of Cognition in Corvids, Apes and Other Animals. In Shackelford, T.K. & Vonk, J. (Eds.), The Oxford Handbook of Comparative Evolutionary Psychology. Oxford University Press. 10.1093/oxfordhb/9780199738182.013.0005)
Van Horik, J. O., Clayton, N. S., & Emery, N. J. (2012). Convergent Evolution of Cognition in Corvids, Apes and Other Animals. In Shackelford, T.K. & Vonk, J. (Eds.), The Oxford Handbook of Comparative Evolutionary Psychology. Oxford University Press. 10.1093/oxfordhb/9780199738182.013.0005
Van Horik, J. O., Clayton, N. S., & Emery, N. J. (2012). Convergent Evolution of Cognition in Corvids, Apes and Other Animals. In Shackelford, T.K. & Vonk, J. (Eds.), The Oxford Handbook of Comparative Evolutionary Psychology. Oxford University Press. 10.1093/oxfordhb/9780199738182.013.0005, Van Horik, J. O., Clayton, N. S., & Emery, N. J. (2012). Convergent Evolution of Cognition in Corvids, Apes and Other Animals. In Shackelford, T.K. & Vonk, J. (Eds.), The Oxford Handbook of Comparative Evolutionary Psychology. Oxford University Press. 10.1093/oxfordhb/9780199738182.013.0005
(2001)
Estimating the complexity of animal behaviour: How mountain gorillas eat thistles
Behaviour, 138
(Johnson, S. G. B., & Ahn, W. (2017). Causal Mechanisms. In M. R. Waldmann (Ed.), The Oxford handbook of Causal Reasoning. Oxford University Press. 10.1093/oxfordhb/9780199399550.013.12)
Johnson, S. G. B., & Ahn, W. (2017). Causal Mechanisms. In M. R. Waldmann (Ed.), The Oxford handbook of Causal Reasoning. Oxford University Press. 10.1093/oxfordhb/9780199399550.013.12
Johnson, S. G. B., & Ahn, W. (2017). Causal Mechanisms. In M. R. Waldmann (Ed.), The Oxford handbook of Causal Reasoning. Oxford University Press. 10.1093/oxfordhb/9780199399550.013.12, Johnson, S. G. B., & Ahn, W. (2017). Causal Mechanisms. In M. R. Waldmann (Ed.), The Oxford handbook of Causal Reasoning. Oxford University Press. 10.1093/oxfordhb/9780199399550.013.12
(Toft, C. A., & Wright, T. F. (2016). Parrots of the wild: A natural history of the world’s most captivating birds. Berkeley, CA, University of California press.)
Toft, C. A., & Wright, T. F. (2016). Parrots of the wild: A natural history of the world’s most captivating birds. Berkeley, CA, University of California press.
Toft, C. A., & Wright, T. F. (2016). Parrots of the wild: A natural history of the world’s most captivating birds. Berkeley, CA, University of California press., Toft, C. A., & Wright, T. F. (2016). Parrots of the wild: A natural history of the world’s most captivating birds. Berkeley, CA, University of California press.
(2006)
How great apes perform on a modified trap-tube task
Animal Cognition, 9
(2006)
A New Caledonian crow (Corvus moneduloides) creatively re-designs tools by bending or unbending aluminium strips
Animal Cognition, 9
(1989)
Tool use in capuchin monkeys: Distinguishing between performing and understanding
Primates, 30
(2016)
Birds have primate-like numbers of neurons in the forebrain
Proceedings of the National Academy of Sciences, 113
(2017)
Are parrots poor at motor self-regulation or is the cylinder task poor at measuring it?
Animal Cognition, 20
(2009)
Insightful problem solving and creative tool modification by captive nontool-using rooks
Proceedings of the National Academy of Sciences, 106
(2017)
Can hook-bending be let off the hook? Bending/unbending of pliant tools by cockatoos
Proceedings of the Royal Society B: Biological Sciences, 284
(2005)
Preliminary observations of tool use in captive hyacinth macaws (Anodorhynchus hyacinthinus)
Animal Cognition, 8
(Köhler, W. (1917) 1925. The mentality of apes (translated from the second revised edition by Ella Winter). New York: Harcourt, Brace, & World.)
Köhler, W. (1917) 1925. The mentality of apes (translated from the second revised edition by Ella Winter). New York: Harcourt, Brace, & World.
Köhler, W. (1917) 1925. The mentality of apes (translated from the second revised edition by Ella Winter). New York: Harcourt, Brace, & World., Köhler, W. (1917) 1925. The mentality of apes (translated from the second revised edition by Ella Winter). New York: Harcourt, Brace, & World.
(O'Neill, L., & von Bayern, A. M. P. (2020). Supplementary materials to ‘Innovative problem solving in macaws’. figshare. Media. 10.6084/m9.figshare.12349418.v2 )
O'Neill, L., & von Bayern, A. M. P. (2020). Supplementary materials to ‘Innovative problem solving in macaws’. figshare. Media. 10.6084/m9.figshare.12349418.v2
O'Neill, L., & von Bayern, A. M. P. (2020). Supplementary materials to ‘Innovative problem solving in macaws’. figshare. Media. 10.6084/m9.figshare.12349418.v2 , O'Neill, L., & von Bayern, A. M. P. (2020). Supplementary materials to ‘Innovative problem solving in macaws’. figshare. Media. 10.6084/m9.figshare.12349418.v2
(2008)
Tubes, tables and traps: Great apes solve two functionally equivalent trap tasks but show no evidence of transfer across tasks
Animal Cognition, 11
(2018)
Parrots have evolved a primate-like telencephalic-midbrain-cerebellar circuit
Scientific Reports, 8
(2012)
Individual variation in cognitive performance: Developmental and evolutionary perspectives
Philosophical Transactions of the Royal Society B: Biological Sciences, 367
(Fragaszy, D. M., & Mangalam, M. (2018). Tooling. Advances in the Study of Behavior, 50, 177–241. 10.1016/bs.asb.2018.01.001)
Fragaszy, D. M., & Mangalam, M. (2018). Tooling. Advances in the Study of Behavior, 50, 177–241. 10.1016/bs.asb.2018.01.001
Fragaszy, D. M., & Mangalam, M. (2018). Tooling. Advances in the Study of Behavior, 50, 177–241. 10.1016/bs.asb.2018.01.001, Fragaszy, D. M., & Mangalam, M. (2018). Tooling. Advances in the Study of Behavior, 50, 177–241. 10.1016/bs.asb.2018.01.001
(2016)
From mechanisms to function: An integrated framework of animal innovation
Philosophical Transactions of the Royal Society B: Biological Sciences, 371
(Villegas-Retana, S. A., & Araya-H D. (2017). Consumo de almendro de playa (Terminalia catappa) y uso de hojas como herramienta por parte del ave Ara ambiguus (Psittaciformes: Psittacidae) en Costa Rica. UNED Research Journal, 9(2). 10.22458/urj.v9i2.1894)
Villegas-Retana, S. A., & Araya-H D. (2017). Consumo de almendro de playa (Terminalia catappa) y uso de hojas como herramienta por parte del ave Ara ambiguus (Psittaciformes: Psittacidae) en Costa Rica. UNED Research Journal, 9(2). 10.22458/urj.v9i2.1894
Villegas-Retana, S. A., & Araya-H D. (2017). Consumo de almendro de playa (Terminalia catappa) y uso de hojas como herramienta por parte del ave Ara ambiguus (Psittaciformes: Psittacidae) en Costa Rica. UNED Research Journal, 9(2). 10.22458/urj.v9i2.1894, Villegas-Retana, S. A., & Araya-H D. (2017). Consumo de almendro de playa (Terminalia catappa) y uso de hojas como herramienta por parte del ave Ara ambiguus (Psittaciformes: Psittacidae) en Costa Rica. UNED Research Journal, 9(2). 10.22458/urj.v9i2.1894
(1993)
Anecdotes, training, trapping and triangulating: Do animals attribute mental states?
Animal Behaviour, 46
(2013)
Did tool-use evolve with enhanced physical cognitive abilities?
Philosophical Transactions of the Royal Society B: Biological Sciences, 368
(2017)
Tool-assisted rhythmic drumming in palm cockatoos shares key elements of human instrumental music
Science Advances, 3
(2014)
Is there a link between the crafting of tools and the evolution of cognition?: Is cognitive evolution affected by tool crafting?
Wiley Interdisciplinary Reviews: Cognitive Science, 5
(2018)
Habitual tool use innovated by free-living New Zealand kea
Scientific Reports, 8
(2015)
Parrots as overlooked seed dispersers
Frontiers in Ecology and the Environment, 13
(2012)
Innovative problem solving in wild meerkats
Animal Behaviour, 83
(2017)
Function and flexibility of object exploration in kea and New Caledonian crows
Royal Society Open Science, 4
(2015)
A novel form of spontaneous tool use displayed by several captive greater vasa parrots ( Coracopsis vasa ): Table 1
Biology Letters, 11
(2011)
Navigating a tool end in a specific direction: Stick-tool use in kea (Nestor notabilis)
Biology Letters, 7
(2004)
The Mentality of Crows: Convergent Evolution of Intelligence in Corvids and Apes
Science, 306
(2008)
Task constraints mask great apes’ ability to solve the trap-table task
Journal of Experimental Psychology: Animal Behavior Processes, 34
(2012)
Spontaneous innovation in tool manufacture and use in a Goffin’s cockatoo
Current Biology, 22
(2014)
Innovative problem solving in birds: A key role of motor diversity
Animal Behaviour, 92
(Krasheninnikova, A., Berardi, R., Lind, M.-A., O’Neill, L., & von Bayern, A. M. P. (2019). Primate cognition test battery in parrots. Behaviour, 156(5–8), 721–761. 10.1163/1568539X-0003549)
Krasheninnikova, A., Berardi, R., Lind, M.-A., O’Neill, L., & von Bayern, A. M. P. (2019). Primate cognition test battery in parrots. Behaviour, 156(5–8), 721–761. 10.1163/1568539X-0003549
Krasheninnikova, A., Berardi, R., Lind, M.-A., O’Neill, L., & von Bayern, A. M. P. (2019). Primate cognition test battery in parrots. Behaviour, 156(5–8), 721–761. 10.1163/1568539X-0003549, Krasheninnikova, A., Berardi, R., Lind, M.-A., O’Neill, L., & von Bayern, A. M. P. (2019). Primate cognition test battery in parrots. Behaviour, 156(5–8), 721–761. 10.1163/1568539X-0003549
(1984)
Tool use by the Palm Cockatoo Probosciger aterrimus during display
Corella, 8
(2012)
Barriers and traps: Great apes’ performance in two functionally equivalent tasks
Animal Cognition, 15
(2018)
Tool making cockatoos adjust the lengths but not the widths of their tools to function
PLOS ONE, 13
(Shumaker, R. W., Walkup, K. R., & Beck, B. B. (2011). Animal tool behavior: The use and manufacture of tools by animals. Baltimore: Johns Hopkins University Press.)
Shumaker, R. W., Walkup, K. R., & Beck, B. B. (2011). Animal tool behavior: The use and manufacture of tools by animals. Baltimore: Johns Hopkins University Press.
Shumaker, R. W., Walkup, K. R., & Beck, B. B. (2011). Animal tool behavior: The use and manufacture of tools by animals. Baltimore: Johns Hopkins University Press., Shumaker, R. W., Walkup, K. R., & Beck, B. B. (2011). Animal tool behavior: The use and manufacture of tools by animals. Baltimore: Johns Hopkins University Press.
(2016)
A problem with problem solving: Motivational traits, but not cognition, predict success on novel operant foraging tasks
Animal Behaviour, 114
(McCormack, T., Hoerl, C., & Butterfill, S. (2011). Tool use and causal cognition. Oxford University Press. 10.1093/acprof:oso/9780199571154.001.0001)
McCormack, T., Hoerl, C., & Butterfill, S. (2011). Tool use and causal cognition. Oxford University Press. 10.1093/acprof:oso/9780199571154.001.0001
McCormack, T., Hoerl, C., & Butterfill, S. (2011). Tool use and causal cognition. Oxford University Press. 10.1093/acprof:oso/9780199571154.001.0001, McCormack, T., Hoerl, C., & Butterfill, S. (2011). Tool use and causal cognition. Oxford University Press. 10.1093/acprof:oso/9780199571154.001.0001
(2016)
How practice makes perfect: The role of persistence, flexibility and learning in problem-solving efficiency
Animal Behaviour, 112
(2019)
Extraction without tooling around—The first comprehensive description of the foraging- and socio-ecology of wild Goffin’s cockatoos (Cacatua goffiniana)
Behaviour, 156
(2005)
Simple heuristics and rules of thumb: Where psychologists and behavioural biologists might meet
Behavioural Processes, 69
(2018)
Mental template matching is a potential cultural transmission mechanism for New Caledonian crow tool manufacturing traditions
Scientific Reports, 8
(2017)
Numbers of neurons as biological correlates of cognitive capability
Current Opinion in Behavioral Sciences, 16
(2015)
Causes of Individual Differences in Animal Exploration and Search
Topics in Cognitive Science, 7
(2007)
Spontaneous Metatool Use by New Caledonian Crows
Current Biology, 17
(2005)
Interspecific Allometry of the Brain and Brain Regions in Parrots (Psittaciformes): Comparisons with Other Birds and Primates
Brain, Behavior and Evolution, 65
(Menchetti, M., Mori, E., & Angelici, F. M. (2016). Effects of the Recent World Invasion by Ring-Necked Parakeets (Psittacula krameri). In F. M. Angelici (Ed.), Problematic Wildlife (253–266). Springer International Publishing. 10.1007/978-3-319-22246-2_12)
Menchetti, M., Mori, E., & Angelici, F. M. (2016). Effects of the Recent World Invasion by Ring-Necked Parakeets (Psittacula krameri). In F. M. Angelici (Ed.), Problematic Wildlife (253–266). Springer International Publishing. 10.1007/978-3-319-22246-2_12
Menchetti, M., Mori, E., & Angelici, F. M. (2016). Effects of the Recent World Invasion by Ring-Necked Parakeets (Psittacula krameri). In F. M. Angelici (Ed.), Problematic Wildlife (253–266). Springer International Publishing. 10.1007/978-3-319-22246-2_12, Menchetti, M., Mori, E., & Angelici, F. M. (2016). Effects of the Recent World Invasion by Ring-Necked Parakeets (Psittacula krameri). In F. M. Angelici (Ed.), Problematic Wildlife (253–266). Springer International Publishing. 10.1007/978-3-319-22246-2_12
(2011)
Big brains are not enough: Performance of three parrot species in the trap-tube paradigm
Animal Cognition, 14
(O’Neill, L., Picaud, A., Maehner, J., Gahr, M., & von Bayern, A. M. P. (2018). Two macaw species can learn to solve an optimised two-trap problem, but without functional causal understanding. Behaviour, 156(5–8), 691–720. 10.1163/1568539X-00003521)
O’Neill, L., Picaud, A., Maehner, J., Gahr, M., & von Bayern, A. M. P. (2018). Two macaw species can learn to solve an optimised two-trap problem, but without functional causal understanding. Behaviour, 156(5–8), 691–720. 10.1163/1568539X-00003521
O’Neill, L., Picaud, A., Maehner, J., Gahr, M., & von Bayern, A. M. P. (2018). Two macaw species can learn to solve an optimised two-trap problem, but without functional causal understanding. Behaviour, 156(5–8), 691–720. 10.1163/1568539X-00003521, O’Neill, L., Picaud, A., Maehner, J., Gahr, M., & von Bayern, A. M. P. (2018). Two macaw species can learn to solve an optimised two-trap problem, but without functional causal understanding. Behaviour, 156(5–8), 691–720. 10.1163/1568539X-00003521
(2005)
Big brains, enhanced cognition, and response of birds to novel environments
Proceedings of the National Academy of Sciences, 102
(2009)
The Role of Experience in Problem Solving and Innovative Tool Use in Crows
Current Biology, 19
(2018)
Compound tool construction by New Caledonian crows
Scientific Reports, 8

Publisher: Springer Journals
Copyright: Copyright © The Author(s) 2020
ISSN: 1543-4494
eISSN: 1543-4508
DOI: 10.3758/s13420-020-00449-y
Publisher site: See Article on Publisher Site

Abstract

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. . . . . . Keywords Comparative cognition Tool use Parrot cognition Causal understanding Innovation Physical cognition Introduction Visalberghi & Limongelli, 1994; Visalberghi & Trinca, 1989; von Bayern, Danel, Auersperg, Mioduszewska, & In the field of animal cognition, problem-solving tasks of- Kacelnik, 2018; Weir & Kacelnik, 2006). Typically, these ten involve tool use (Auersperg, Huber, & Gajdon, 2011; tool-use tasks present subjects with a novel apparatus that Auersperg, von Bayern, Gajdon, Huber, & Kacelnik, 2011; requires some sort of behavioural innovation to access a Bird & Emery, 2009; Girndt et al., 2008;Köhler, 1917; reward. They are often used as a measure for an individ- Martin-Ordas Call, & Colmenares, 2008, 2012; Mulcahy ual’s behavioural flexibility and examine what a species & Call, 2006; Taylor, Hunt, Holzhaider, & Gray, 2007; understands about the functional properties of those tools, and physical interactions in their environment in general. Of particular interest is comparing the performance of spe- cies that are habitual tool users in the wild and those that * Laurie O’Neill are not, investigating both closely and distantly related [email protected] species (Auersperg et al., 2011; Auersperg, von Bayern, * Auguste M. P. von Bayern Gajdon et al., 2011; Lambert et al., 2017; Teschke et al., [email protected] 2013). In this manner it is possible to examine if the ca- pacity for flexible tool use is a result of a more general Max Planck Institute for Ornithology, 82319 Seewiesen, Germany cognitive capacity evolved in response to selection pres- Max Planck Comparative Cognition Research Station, Loro Parque sures not restricted to the physical domain (e.g., general Fundacion, 38400 Puerto de la Cruz, Tenerife, Spain 3 intelligence; Burkart et al., 2017), or whether more flexible Behavioural Ecology Group, Department of Animal Sciences, tool-using skills have evolved from a more specialised Wageningen University and Research, Wageningen, the Netherlands 4 physical-cognition domain (Kacelnik, 2009; McCormack Department Biology II, Ludwig-Maximilians-University of Munich, et al., 2011; Taylor & Gray, 2014). Martinsried, Germany Learn Behav (2021) 49:106–123 107 Tool use, however, is not the only type of physical envi- bitter taste of certain piths in their outer layers (Villegas- ronmental interaction that might require complex physical Retana & Araya-H., 2017). Palm cockatoos’ (Probosciger cognition. Many species of birds face similar complex object aterrimus) use sticks as part of their mating display, drum- manipulations when they are also required to probe, weave ming them against hollow trees (Heinsohn et al., 2017; and manipulate many different objects, including sticks, in Wood, 1984). Many parrots can face other types extractive order to build their nests (Breen et al., 2016). Furthermore, foraging difficulties such as trying to manipulate and wield there are many types of extractive foraging without tool use very large and tough-shelled nuts, which require complex ma- that can require complex manipulations, for example when nipulations in order to be able to open them (Tella et al., Gorillas prepare thistles in order to eat them they must manip- 2015). ulate them in such a way so as to avoid being hurt by spines Two parrot species in particular have been the focus of (Byrne et al., 2001). In sum, tool use is not the only measure of physical cognition abilities in parrots and have been shown an individual’s physical cognition skill. to be able to use sticks as extractive foraging tools in a number In this experiment we studied two species of parrots, spe- of problem-solving tasks. Kea are able to use stick tools by cifically the macaws Ara ambiguus (great green macaws) and inserting them into holes in puzzle boxes to knock food items Ara glaucogularis (blue-throated macaws). Some species of off a pedestal (Auersperg et al. 2011; Auersperg, von Bayern, parrots have been recorded as having complex and flexible Gajdon et al. 2011), push food items out of a tube (Lambert physical problem-solving skills (Lambert et al., 2018) and et al., 2017), and even to spring traps meant for other animals they share a number of other life-history and social factors in the wild in order to retrieve the bait (Goodman et al., 2018). with other taxonomic groups, such as corvids and primates, Goffin’s cockatoos have been established as probable non- that suggest they might be required to use complex cognition tool users in the wild (O’Hara et al., 2019; but see Osuna- (Emery & Clayton, 2004; Osvath et al., 2014; Van Horik Mascaró & Auersperg, 2018), but remarkably some individ- et al., 2012). These factors include having relatively large, uals began to spontaneously innovate the manufacture and use neuronally dense brains that mean they might have a higher of stick tools in captivity (Auersperg et al., 2012). upper limit in cognitive capacity (Gutiérrez-Ibáñez et al., Furthermore, these cockatoos appeared to show understand- 2018; Herculano-Houzel, 2017; Iwaniuk et al., 2005; ing of some of the complex properties of these tools that made Olkowicz et al., 2016), although this does not determine or them functional, such as making tools of the appropriate imply high cognitive ability. Many parrots are capable of a length (Auersperg et al., 2018) or adding hooks to their tools diverse range of motor actions, which is likely to be a key when necessary (Laumer et al., 2017). Because of these ex- factor for behavioural innovations (Griffin et al., 2014), and amples, it is interesting to expand tool-use tasks to other spe- they have particularly impressive manipulative skills with cies across the parrot order. The Kea and Goffin’scockatoos both their beaks and tongues as well as zygodactyl feet that are members of the more basal parrot superfamilies allow them to hold and manipulate objects in them (Toft & Strigopoidea and Cacatuoidea, respectively, whereas the ma- Wright, 2016). Parrots also live in stochastic environments caws in this study are parrots of the family Psittacoidea, so the that might require behavioural flexibility in order to thrive current experiment provides an opportunity to expand our (Toft & Wright, 2016), and further to this, many different understanding of parrots’ physical cognition capacities across species of parrots have been shown to be successful invaders the parrot order. We can therefore explore what kind of un- to their non-native range (e.g., Menchetti et al., 2016), some- derstanding these various species have for tools and instru- thing that has also been suggested to require elements of be- mental problem solving, and whether there are certain evolu- havioural flexibility (Sol et al., 2002, 2005). In total, all the tionary factors that might drive more complex physical cog- factors of big brains, capable physical manipulation skills and nition abilities. a capacity to respond to changing environments, suggest that The specific individuals we tested have faced tool-use tasks parrots might be candidates for using cognitively complex before. In a single trial-test of a ‘food-out-of-reach’ situation problem-solving skills, such as causal understanding, in order with a stick tool available, they did not interact with a stick to be adapted to their environment (Van Horik et al., 2012). tool (Krasheninnikova et al., 2019). In the stone-dropping A few species of parrots have shown capacities to handle task, a problem-solving task that requires individuals to inno- objects in a variety of interesting ways, some of which are vate the use of a stone tool to get a reward (von Bayern et al., borderline cases of tool-use. Vasa parrots (Coracopsis vasa) 2009), the majority were capable of using stone tools to drop are able to use stones to scrape shells in order to obtain calci- onto a collapsible platform to retrieve a reward, but most in- um powder for consumption (Lambert et al., 2015). Both hy- dividuals were unable to manipulate a stick tool to perform a acinth macaws (Anodorhynchus hyacinthinus) and great green very similar action (O’Neill et al., 2020). As some species of macaws (Ara ambiguus) have been observed wrapping certain macaws have shown borderline tool use abilities in the wild nuts with leaves, which was interpreted as a method to grip the before (Borsari & Ottoni, 2005; Villegas-Retana & Araya-H., nuts better (Borsari & Ottoni, 2005) or perhaps to avoid the 2017), we believed them to be suitable candidates to test the 108 Learn Behav (2021) 49:106–123 functional understanding of instrumental problem solving in were still able to solve the task after making this error, as the the lab. apparatus only needed two stick insertions from one side to The inspiration for the experimental task we used comes reach the reward. This may have meant there was little moti- from Visalberghi and Trinca (1989). In that experiment, a vation for the example capuchin subjects to optimise their group of tufted capuchins (Sapajus apella) were given a trans- multi-tool construction, as there was little to no negative out- parent, long, horizontal tube, with a reward in the middle. To come from the errors they made. To solve this issue in our solve the task they had to insert a stick tool into one side of the replication of the experiment, we required that there were at tube and push the reward out of the other side. After success- least four insertions of a small stone from one side before the fully doing this, they were given the tube apparatus again but reward was available from the other. Additionally, they also only with access to three smaller sticks, none of which could typically required three consecutive actions before they had reach the reward on its own. The subjects had to insert two of visual feedback that the reward was being moved. these smaller sticks, one behind the other, in order to push the The rationale of the experiment was as follows. The first reward far enough to the other end of the tube so that it was stage of the experiment, the pre-test, gave the subjects an accessible. The aim of the task was to see if the subjects opportunity to spontaneously solve the task and see if they recognised the functional property of the stick tool that they were able to insert multiple stones, one behind the other, to initially inserted, or whether they were just ‘performing’ a push a reward out of a horizontal tube (Fig. 1a and b). It was learned action of ‘insert sticks’ as this would lead to rewards. expected that they would not be able to do so, thus after this The modification we made on this task was to simply replace phase the subjects were given an example of the end-state these small sticks with even smaller stones. Thus, more of function of the multi-stone construction they were required these stone tools would have to be inserted into the tube before to make with the stones, ergo a stick was pre-inserted into the reward was available (a minimum of four). This behaviour the tube and the subjects simply had to push it to remove the could possibly be classed as an ‘additive tool’, as the subjects food (Fig. 1c and d). After this, the subjects were once again would need to “join or connect two or more objects to make given the opportunity to solve the task with the stones – crit- one tool that is held or directly manipulated in its entirety ical test 1 (a repeat of Fig. 1a and b). The underlying principle during its eventual use” (Shumaker et al., 2011). However, was that if subjects were able to reason the underlying func- it is not certain whether this multi-component structure would tional mechanism that made the stick effective, then they fit all definitions of tool use (e.g., Fragaszy & Mangalam, would be able to recreate the same effect of this stick out of 2018). Nevertheless, the focus of the experiment was not multiple stones. If the subjects were still unable to solve the ‘whether the subjects could use tools’, but instead it was to task at this stage, then they were given a more direct experi- explore subjects’ causal understanding of the underlying func- ence of the efficacy of the stones, thus cueing them to use tional properties of the initial stick tool they interacted with to stones. They were given a similar apparatus, but the horizontal see if they could recreate the functional properties in another tube was shortened and only a single stone needed to be way. For this reason, we do not refer to the multi-component inserted and pushed to make the reward available on the other construction subjects were manufacturing as an additive tool, side (Fig. 1e and f). If they succeeded in doing that then sub- as we recognize this is not a name that everyone will accept. jects were given another critical test (critical test 2)tosee if Instead, we refer to it as a ‘multi-stone construction’ or just a they could chain this single stone insertion behaviour repeat- ‘stone construction’. edly to make the required multi-stone construction. In essence, One issue with this task is that the capuchin subjects ap- in each of the three test-phases (the pre-test and critical tests 1 peared to be able to find a solution through exploratory, trial- and 2), subjects had to make a smaller innovative ‘leap’,and and-error behaviour (Visalberghi & Trinca, 1989). These use an assumed ‘less complex’ form of cognition, to solve the kinds of persistent behaviours have been noted to predict task. After the subjects were able to consistently solve the many animal subjects’ ability to solve problem-solving tasks task, we also gave them up to two additional follow-up tasks in other experiments (Chow et al., 2016; Thornton & Lukas, in order to triangulate (Heyes, 1993) how much the subjects 2012; Thornton & Samson, 2012; van Horik & Madden, understood about the function of their multi-stone construc- 2016). Specifically, the three capuchin subjects that were able tion (Fig. 1g and h). to solve this additive tool task repeatedly made an error of Although the subjects we tested were naïve to the specific inserting a stick from both sides of the reward before inserting apparatus that we gave them in this experiment, they were not a stick behind one that had been previously inserted (and thus naïve to the experimental protocols in general. For this reason solving the task). This suggested that the subjects did not have we decided to exploit their testing history to also elucidate an end-state mental representation of the multi-tool construc- what kind of heuristics (Hutchinson & Gigerenzer, 2005) tion they were making with the shorter sticks (e.g., Jelbert and behaviours the subjects may extend from successful be- et al., 2018), and instead were just randomly inserting sticks havioural interactions in previous experiments in order to find until they could reach the reward. Nevertheless, the subjects rapid innovations to the novel apparatus in this experiment, a Learn Behav (2021) 49:106–123 109 110 Learn Behav (2021) 49:106–123 Fig. 1 In the pre-test and critical tests a reward (a walnut represented by were housed in groups of two to six individuals, accord- a green circle in the figures) was placed in the middle of a transparent ing to species and age, in seven aviaries. Six of these acrylic tube with ten stones placed around it (a). In the Pre-test and aviaries were 1.8×6×3m(width ×lengthxheight) Critical tests, subjects had to insert five of these into one side of the tube and one was 2 × 6 × 3 m. Windows of 1 × 1 m could to push out rewards on the opposite end (b). In the stick experience phase, a stick was inserted into the side of the tube up until the reward be opened between the aviaries to connect them together. (c). Subjects had to push the stick to push the reward out (d). In the short- One half of each aviary was outside, so the birds followed tube experience a smaller tube was given to the subjects in which only the natural heat and light schedule of Tenerife. The half one stone needed to be inserted into the tube and pushed to make the inside the research station was lit with Arcadia Zoo Bars reward come out (e). In the pre-inserted stone experience, a stone was inserted to be touching the reward in the short tube, so that the stone only (Arcadia 54W freshwater Pro and Arcadia 54W D3 needed to be pushed to make the reward come out (f). In the parallel Reptile Lamp) that followed the natural light schedule. tubes phase, there were two tubes, but only one of them was baited (g), A large variety of fresh fruit and vegetables were pro- subjects had to target the stones only at the baited tube. In the blocked vided first thing in the morning and again in the after- tubes phase, both tubes were baited, but a bung prevented the reward from being removed from one tube (h). For clarity in the schematics, a noon. In the afternoon, each individual received a portion vertical tube in the middle of the apparatuses is not shown. This was of Versele Laga Ara seed-mix that was modified based on featured in all apparatuses except the short tubes (e and f). This vertical their daily weight. The parrots were never starved, and tube can be seen in the supplementary videos showing the subjects portions of nuts and seeds were controlled to prevent interacting with the horizontal tube overeating and obesity. Testing took place at least an hour after the parrots had breakfast to ensure there was suffi- cient motivation to get food rewards. Unless otherwise mechanism thought to be vital in animal innovations (Tebbich et al. 2016). We therefore added a ‘red-herring’ to the apparatus. specified, walnut halves were used as rewards. These were highly prized by all subjects, and they were able to Specifically, this was in the form of a vertical tube, attached to the obtain them on a daily basis through voluntary testing. horizontal tube, that was placed directly over the reward, but was blocked and so had no connection to the reward (see Supplementary Fig. 7 (Online Supplementary Material, OSM) Experimental rooms for a photo of the apparatus). Placing a rock into this vertical tube would therefore achieve nothing. Nevertheless, in the previous experiment they completed, the stone-dropping task (O’Neill Experiments took place in separate testing rooms away from the aviaries. All birds had been previously trained to enter them. et al., 2020), the solution to the experiment was to place a stone tool into a vertical tube. Thus, we could see if they would apply Theserooms were 1.5× 1.5×1.5mandalsolit with Arcadia Zoo Bars to cover the birds’ full visual range. One wall of each this same rule in their exploration of a new apparatus that was drawn from success in a previous experiment. one of the experimental rooms had a 50 × 25 cm window through which an experimenter could place apparatuses into We predicted that at least some of the parrots would be the testing room from a neighbouring chamber. This window likely to solve the task, as we have seen previously that the could be occluded with a white board so that the experimenter subjects were highly explorative and capable in this similar could hide anything they were doing from the subjects’ view, stone-dropping problem-solving task (O’Neill et al., 2020). such as re-setting apparatuses between trials. The experimenter However, we were not certain whether the subjects would always wore mirrored sunglasses during experiments to prevent be able to solve the task consistently nor succeed in the trans- fer tasks as this experiment required more self-control, some- their gaze being a cue for the parrots. A second wall was made of sound-proofed one-way glass so that zoo visitors could thing we knew the subjects might lack (Kabadayi et al., 2017). observe experiments without disturbing the subjects. One individual, Hannibal, was tested in a separate cage con- nected to his home aviary as he was not comfortable going to the Methods experimental rooms. Each aviary had a separate, removable cage attached to it, which was used to transport subjects between their Subjects and housing home aviaries and the experimental rooms. However, the cages were also large enough for testing (1 × 1 × 1 m). They had sliding Nine great green macaws (Ara ambiguus) and eight blue- doors that could separate an individual from the rest of their throated macaws (Ara glaucogularis) were tested. Their aviary, and were also on wheels, which meant they could be age and sex are shown in Table 1. All the birds were moved to an area occluded from the other individuals in their hand-raised and group-reared by the Loro Parque aviaries. Finally, they also had a second door through which Foundation in Tenerife, Spain, and all were housed in testing apparatuses could be placed into the cage with a test the Comparative Cognition Research Station, within the subject. In this way, Hannibal was able to complete the test in Loro Parque zoo in Puerto de la Cruz, Tenerife. The birds a similar manner to the other subjects. Learn Behav (2021) 49:106–123 111 Table 1 The age, sex and success of each subject taking part in the which the subjects first had a successful trial in that particular test horizontal tube task The table also shows which stage of the phase. Dashes indicate that the subject did not take part in that test experiment each individual completed and at which stage they failed or phase. Columns in both shades of grey are the mechanistic experiences completed the task. The numbers in the columns show the session in Apparatus For each test, the subjects were provided with a selection of ten stones. All were natural volcanic rocks taken from the This task required subjects to push a reward out of a horizontal beach in Puerto de la Cruz, Tenerife (they were returned to transparent tube, which was 30 cm long with a diameter of the beach after testing was finished). They all fit into the tube, 6cm (Fig. 1a and b). The tube was fixed in an elevated hor- had a diameter of 4.5–5.5 cm and weighed between 60 and izontal position at a height of 10 cm by a metal stand at either 100 g. Three stones had to be placed into the tube, one behind end (see Supplementary Fig. 7 (OSM) for picture). The stands the other, before the reward moved. At least four stones had to were attached to a solid wooden board (45 × 30 × 2.5 cm) and be inserted to place the reward within reach. This meant that the central tube was wrapped in wire mesh (1 × 2 cm) to show the subject received no perceptual feedback of the reward the solidity of the transparent acrylic. The reward was placed moving until at least three specifically directed actions were in the middle of the horizontal tube, thus the only way to made. obtain the reward was by pushing it to one of the two edges with the aid of a tool. There was an additional 5-cm tube (vertically positioned) of the same diameter attached to the Experimental procedures centre of the top-side of the main tube (see Supplementary Fig. 7 (OSM)). This additional tube was non-functional as The experiment consisted of three test phases (pre-test, criti- no tool could interact with the reward from here. Its purpose cal test 1 and critical test 2), three different experience phases was to provide a place for the subjects to put the stone tools (pre-inserted stick, short tube 1, pre-inserted stone and short onto the top of the apparatus. This was so that the subjects’ tube 2) and two transfer tasks (parallel tubes and blocked tendency and persistence for placing tools in an opening on tubes). An overview of the order in which these phases took top of the apparatus could be assessed, even when it had no place is shown in Fig. 2. Not every subject took part in every effect on the outcome of the task. phase as they either skipped phases due to early success in an 112 Learn Behav (2021) 49:106–123 Fig. 2 The subjects were given progressively more functional experience they followed the green arrow, if they failed then they followed the red if they failed to solve the task. In blue are the ‘test’ phases, in which arrow. Subjects that failed the short-tube experience moved onto the pre- subjects were exposed to the task that required them to use a multi- inserted stone experience, if they succeeded in the preinserted stone ex- stone construction to push the reward out of the tube. In orange are the perience they were given the short-tube experience again. If they failed ‘experience’ phases, which provided the subjects with mechanistic expe- the short-tube experience a second time then testing ended, i.e. they only rience of how the task could be solved. If subjects succeeded in a phase had the pre-inserted stone experience once experimental phase or did not take part due to failure in an stones and the opening of the tube. The critical test phases, experience phase. after each of these mechanistic experience phases, were im- In essence, in the pre-test the subjects were given an op- plemented to test if the subjects could generalise and extrap- portunity to spontaneously solve the task in order to examine olate from these progressively more specific mechanistic ex- if the experiment could be solved without any mechanistic periences to produce the necessary actions to solve the task. experience, just by exploration behaviours characteristic of Finally, the subjects were offered the parallel and blocked the parrot subjects. The experience phases presented the sub- tubes tasks as follow-up tasks to investigate if they could jects with progressively more specific direct mechanistic ex- flexibly solve slightly modified tasks, verifying whether they perience of how the experiment could be solved: the first indeed attended to the functional properties of the task or not. mechanistic experience (pre-inserted stick) provided them Within the protocol descriptions, the term ‘session’ de- with knowledge of the basic mechanism that the food could scribes a single period of testing in the experimental chamber. be pushed from contact with a stick, but without cueing them Typically subjects took part in a single session per day. Within towards stones in particular. The second mechanistic experi- each testing session (pre-test, both critical test phases, and the ence (short tube) allowed the subjects to move the reward by parallel and blocked tubes phases), subjects could be given a just inserting a single stone, thus reinforcing subjects that maximum of three trials (i.e. chances to solve the task and showed exploratory stone-inserting behaviour by letting them obtain a reward). Within each session in the mechanistic ex- experience that their behaviour can move the reward and lead perience phases (pre-inserted stick, short tube and pre-inserted to success. The third mechanistic experience further helped stone), subjects could be given a maximum of six trials. This subjects that would not insert stones (pre-inserted stone)as was because the successful method in the test sessions took it showed that just pushing stones already pre-inserted in the slightly longer and we did not want subjects to lose motiva- tube would move the reward, hence cueing them both towards tion. If subjects failed a trial, the whole session was stopped Learn Behav (2021) 49:106–123 113 and subjects were brought back to their group aviary. A trial board beneath it. The whole apparatus was pushed into the was considered as failed if subjects did not solve the task subject’s chamber, in the centre of the testing area; thus the within 10 min of the apparatus being placed into the experi- subjects were able to access it from all sides. The subjects mental room with them. If subjects succeeded in a trial, then were given a maximum of 10 min to interact with the appara- the apparatus was removed, re-baited and replaced into the tus. If they failed to retrieve the reward in this time, then the subject’s experimental room for another trial (described be- trial and the session ended, the apparatus was removed, and low), unless it was the last trial of a session, in which case the the subject was returned to its social group. If, on the other session ended. All subjects finished testing, from habituation hand, the subject managed to successfully retrieve the reward to final test trial, within 2 months. All testing took place be- by pushing multiple stones into the side of the apparatus, then tween November 2017 and May 2018. the subject was given further trials within that testing session. To do this, the experimenter waited for 30 s after the subject Habituation had retrieved the reward and then removed the apparatus. It was re-baited with another half piece of walnut out of sight of Every subject was first habituated to the apparatus without the the subject, ensuring that there were ten stones available un- presence of the stones. These sessions took place in the exper- derneath the apparatus. The apparatus was then placed back imental rooms, with an experimenter sitting in the into the room with the subject for another trial exactly 1 min neighbouring chamber. This was necessary for re-baiting if after it was removed from the subject’s chamber. Subjects further trials were needed in the session and it also served to received a maximum of three trials in a session. They needed create a more relaxing environment for the subjects to work in. to complete 12 successful trials to be considered consistently Once the subject was in the experimental room, the experi- successful within this experimental phase (although Captain menter introduced the apparatus through the window in be- and Enya, see Table 1, progressed from the second critical test tween them. Subjects had 5 min to approach the apparatus and phase with 11 successful trials due to an experimenter’serror). take a piece of walnut placed at the base of the apparatus. If This repetition was to validate that subjects had not acciden- they took it, the experimenter waited for 30 s then removed the tally succeeded in a task but could replicate the method they apparatus. Next, the experimenter, out of sight of the subject, used to get the reward. Additionally, if subjects had six failed re-baited the apparatus and reintroduced it 1 min after remov- trials in a condition, they moved onto the next mechanistic ing it. The subjects had up to six of these trials per session and experience phase (or stopped testing, depending on which if they took six walnut pieces in a row within a single session, condition they had ‘failed’;see Fig. 2). they moved onto testing. Sessions were only counted as valid if the subjects The subjects had already been habituated with the stone approached and touched the apparatus. Thus unsuccessful tri- als were ones in which subjects had touched the stone tools or tools in a previous experiment and all were eager to interact with them, so they did not need habituating to them again. We the apparatus but had not been able to retrieve the reward. state that the parrots were ‘eager’ to interact with the stones Subjects that were able to succeed in the pre-test or either of because they would frequently pick up the stones and just the critical test phases moved onto the follow-up tasks (par- manipulate them between their feet and beaks with no appar- allel tubes and blocked tube). ent purpose. This was typical for these macaws for any object that was a similar size and shape to these stones. Pre-inserted stick experience Pre-test and critical test procedures In the first mechanistic experience phase, called pre-inserted stick experience, a stick tool was used to show how the reward The pre-test and both critical test procedures were identical. could be pushed out of the tube (Fig. 1c and d). In the first The difference was that subjects had been given different stage of this phase, the stick tool was attached to the tube so mechanistic experience phases after the pre-test and between that it could only be pushed into the tube, i.e. it could not be critical tests 1 and 2, which provided functional information pulled out. We assumed the parrots would be likely to pull the about how the task could be solved. The order of these tests stick tool in their initial exploratory interactions with it, so and mechanistic experience phases can be seen in Fig. 2. attaching it in this way made sure they could only experience Subjects were given a minimum of six 10-min sessions for the ‘rewarding’ outcome of pushing the stick first. The exper- both the pre-test and the two critical tests. Each session had imenter placed a half walnut reward in the centre of the appa- between one to three trials, i.e. they were given more trials in a ratus with the stick tool inserted flush against it. The subjects session if they succeeded until the maximum trial number of were given multiple 10-min sessions to push the stick through three was reached. The apparatus was first prepared out of explorative behaviour alone. If they succeeded within a ses- sight of the subjects. A half walnut reward was placed inside sion, the experimenter gave the subjects more trials. They the middle of the tube and ten of the stones were placed on the removed the apparatus, re-baited it out of sight, ensured the 114 Learn Behav (2021) 49:106–123 stick tool was correctly positioned and gave it back to the removed, and the subjects were taken back to their social subject. If subjects successfully obtained the reward six times group. If subjects failed six trials in the short-tube experience, within a session, they moved onto the next stage of this mech- then they were given another mechanistic experience phase, anistic experience phase. the pre-inserted stone phase. In this following stage, the procedure and setup were If the subjects succeeded in the pre-inserted stone identical, except the stick was no longer attached to the experience, then they still had to succeed in the short-tube apparatus, so it was possible for the birds to pull the stick experience before they could proceed to critical test 2. out of the apparatus. We did this to ensure that the subjects However, the subjects only had one opportunity to succeed recognised that thestickspecificallyneededtobepushedto in the pre-inserted stone experience,i.e. subjects only moved get the reward, and it wasn’t just exploratory interaction between the short-tube experience to the pre-inserted stone with the stick that led them to obtaining the reward. If sub- experience once, not continually until they succeeded. jects pulled and thus removed the stick at this stage, the experimenter would remove the apparatus and the stick, Pre-inserted stone experience reposition the stick inside the apparatus out of sight of the subjects, and then give it back. There was no limit to the The pre-inserted stone mechanistic experience used the same number of times the stick could be re-positioned like this short tube apparatus (Fig. 1f). However, now one of the stone (but there was still a 10-min time limit per trial). To pass tools was already placed inside the tube next to the reward so this final stage of the experience phase, the subjects had to the subjects only needed to push the stone to obtain the re- push out the reward 12 times in two consecutive sessions ward. The purpose of this was to cue the subjects to the correct (six trials per session). All subjects completed the first stage positioning of the stone and that interaction with the stone of the pre-inserted stick experience in a maximum of six when it was in this position would also make the reward sessions and the second stage in a maximum of three move. If subjects succeeded in this stage, then they were given sessions. They then moved onto the first critical test. another chance to succeed in the short-tube experience, which they then had to succeed on in order to proceed to critical test Short tube experience 2. In this experience phase the experimenter placed one of If the subjects failed the first critical test, they were given a the stones in the tube, flush against the reward, before giv- second mechanistic experience phase. In this mechanistic ex- ing it to the subject. This setup was done out of sight of the perience phase, called the short-tube experience, a modified subjects, so they did not receive a cue from the experiment- apparatus was used (Fig. 1e). A 15-cm long tube was mounted er to specifically insert the stones into the tube, they only on a separate wooden board at the same height as the original saw the apparatus with a stone already inserted. For each apparatus. It did not have the additional vertical tube in the trial, the subjects had up to 10 min to obtain the reward. If centre. With this shorter tube, the subjects were still unable to the reward was obtained, the apparatus was removed, re- reach a reward placed in the middle, but now they only needed baited, and given back to the subject. If subjects could to insert one stone to push the reward out of the tube. Some repeat this behaviour six times in one session, they were subjects had inserted one or more stones into the apparatus in given another round in the short-tube experience phase, the initial critical test, but not a sufficient number to move the exactly as described above. If they passed criterion in the reward (see Results section). If they repeated this behaviour latter, thentheyweregiven critical test 2, but if they failed, with a shorter tube it would be enough to move the reward. their participation in the test ended. The short tube was given with the same procedure as the All of the subjects that were given the pre-inserted stone critical test, i.e. the subjects were given the short-tube appa- phase succeeded (nine subjects), but only one of these subjects ratus with ten stones and they then had 10 min to solve the task (Rita) managed to succeed in the following short tube 2 expe- by placing one stone into the tube and pushing the reward out. rience phase (see Table 1). After successful trials, the apparatus was removed, re-baited, and placed back into the subject’s room. However, as this Parallel tube and blocked tube transfer tasks solution to the short-tube experience phase was potentially very rapid, as the subjects only needed to insert a single stone, Two follow-up tasks were devised for subjects that managed subjects were given up to six trials per session to speed up the to reliably succeed in one of the critical test phases. The pur- testing process. To reach the criterion in the short-tube pose of these was to inform on how the subjects had achieved experience, and in turn proceed to critical test 2,they had to their successful methods in the critical test phase. succeed 12 times consecutively over two sessions. If subjects Specifically, the parallel tubes test (Fig. 1g) was used to failed a trial in a session, i.e. they did not obtain the reward see if subjects were inserting stones into the tubes to move the within 10 min, then the trial and session ended, apparatus was reward or whether they had just learnt that inserting stones Learn Behav (2021) 49:106–123 115 into the tubes was rewarding. In this task, two identical ver- camera angle was considered ‘side 1’ and the right side was sions of the horizontal tube were attached to a single board, ‘side 2’. We did not call these ‘left and right’ as occasionally a parallel to each other. The key difference was that only one of second camera angle was consulted to confirm stone inser- the two tubes was baited with a walnut reward, the testing tions and the other cameras were pointed from different an- protocol was otherwise the same as the critical test phases. gles, hence left and right may have made this confusing whilst The subjects were still only provided with ten stones to try and coding. A stone was classified as inserted into the tube if the obtain the rewards and they were still given 10 min per trial to subjects placed it inside the tube, released it from their beak, try and obtain the reward. If they obtained the reward, the and then it stayed inside the tube without falling out. apparatus was removed and re-baited; however, the reward Sometimes, subjects removed stones from the tubes and then was randomly placed in either of the two different tubes in placed them back inside the tube, either on the same or on the different trials in a counterbalanced schedule, i.e. rewards other side. If the stone was picked up from inside the tube, were not just swapped between the tubes on different trials removed from within the tube entirely, then placed back inside but were placed in each of the two tubes an equal number of the tube and released, it counted as another insertion. In this times over all trials. A more stringent success criterion was manner, it was possible for more than ten stone insertions to used in this phase. Subjects had to obtain the reward 12 times be counted per trial even though there were only ten stones in a row in four sessions (three times per session), i.e. without available for each trial. In the parallel tubes and blocked tube unsuccessful trials between the successful trials. This was to transfer tasks, the number of stones inserted on each side of ensure that random placement of stones would not lead to both tubes was counted, i.e. there were four possible places subjects reaching the success criteria in this phase. If they where the stones could be inserted. For both of these tasks, succeeded, they moved onto the blocked tube task,but if they one tube was the ‘correct tube’ and the other was the ‘wrong failed, they stopped testing. tube’, as rewards could not be obtained from the ‘wrong tube’. The blocked tube task used the same apparatus as the In all of the test trials, from the pre-test and critical tests to parallel tubes, but a ‘bung’ attachment was created to block the parallel tubes and blocked tube transfer trials, we also one end of one of the tubes (Fig. 1h). This transfer task was counted the number of times subjects placed stones into the used to see if the subjects could recognise they needed to vertical-tube on top of the horizontal tube. In some trials, push the reward to a location that was then accessible to subjects again removed and re-inserted stones on multiple themselves, not that the rewards just needed to be pushed. occasions into these tubes. The same rule as above was In this task, both tubes were now baited, but the bung would followed: if the stone was picked up, removed entirely from prevent the subjects from pushing the reward out of the tube, then re-inserted and released, it was counted as another blocked tube. Between every trial the bung was randomly stone insertion. Hence, in some trials many more than two repositioned in one of the four available positions, again in ‘stone-in-top’ insertions were counted even though the verti- a counterbalanced fashion. The success criterion would cal tube could only hold two stones at once. In the parallel tube have been the same as the parallel tubes task, but no subject and blocked tube trials, there was no discrimination between succeeded here. This was the last available task in the subjects placing stones in the top of the ‘correct’ or the experiment. ‘wrong’ tube. Ten percent of the videos (28 sessions, 38 trials; these were Behavioural coding randomly selected from all of the pre-test, critical tests and parallel/blocked tubes tests) were then coded by a second All experiments were recorded on four static CCTV cameras. observer to check inter-observer reliability on the number of These covered all angles of the testing room. Two recordings stones inserted into both the top and the sides of the tube. from separate cameras were saved for each experimental ses- There was ‘excellent’ reliability between the two observers sion, but more recordings were saved if it was necessary for (Intraclass correlation = 93.9% consistency, R-package “irr”). specific trials, for example if there was partial occlusion of a Finally, we also coded subjects’ exploration of the stones camera view from the subject standing in the way. Behaviours and the apparatus. This included the latency until subjects were scored from the videos using Solomon coder (András touched the stones and the apparatus in each trial as well as Péter, solomon.andraspeter.com). the amount of time the subjects touched the stones and appa- Firstly, the location of the placement of the stone tools was ratus in each trial. Finally, the amount of time the subjects recorded. For every trial in the ‘pre-test’, ‘critical test’ and spent touching the apparatus with the stones (combining the ‘critical test 2’ the numberofstonesthatsubjectsplacedin objects together) was also recorded. This data was not used in each side of the tube was counted. The apparatus was always the final analysis, but a summary of it is provided in placed into the testing rooms in exactly the same position and Supplementary Table 2 (OSM), and the complete data are orientation so the locations we recorded of the stones was with the raw data to be found on figshare (O’Neill & ‘absolute’, i.e. the left side of the tubes from one specific Bayern, 2020). 116 Learn Behav (2021) 49:106–123 Fig. 3 The individual mean number of stone insertions into each side of phase of testing. There are fewer subjects in critical test 2 as not all the tube per trial. Grey bars indicate that individuals failed in that phase of individuals passed the short-tube experience, which was a requirement testing, black bars indicate that individuals consistently succeeded in that to complete critical test 2 Results glaucogularis), managed this in critical test 2 after the first short-tube experience phase (Fig. 3; Supplementary Figs. 5 In total, three birds (two out of eight Ara glaucogularis and and 6 (OSM)). These two subjects only solved the task 11 one out of nine Ara ambiguus) were able to consistently times (rather than 12) before being moved onto the follow- solve the task in the critical tests (Fig. 3). One Ara up tasks due to experimenter error. Enya solved the task 11 glaucogularis (Lady) managed to do this in critical test 1 timesin12trialsand Captainsolvedthe task 11 times in16 after the stick experience phase (Fig. 3; Supplementary Fig. trials. All three of these subjects had four trials in which 3 (OSM)); she solved the task 12 times in 16 trials. The they obtained the reward without error, i.e. without other two, Enya (Ara ambiguus)and Captain (Ara inserting a counter-productive stone into the opposing end Learn Behav (2021) 49:106–123 117 Fig. 4 Thenumber of stones insertedoneachside ofeach tubeinthe specifically from one direction. Sometimes this meant she inserted parallel tubes task. The top chart of each bird shows the stones inserted in stones into the unbaited tube, sometimes it led to success. Enya (Ara the baited tubes, the bottom shows the unbaited tubes. Lady (Ara ambiguus) appeared to start inserting stones more randomly at this glaucogularis) appeared to insert stones into either of the tubes, but stage, so she had very little success of the tube from their initial stone construction. Of these the ends of the tube in this stage of the experiment three, only Captain could consistently obtain the reward in (Supplementary Figs. 1 and 2 (OSM)). After the stick the follow-up parallel tubes task, but he was unable to do so experience phase, in the first critical test phase, three of the in the blocked tube task (Figs. 4 and 5). Ara glaucogularis started to consistently insert stones into the None of the subjects were able to solve the task in the pre- ends of the tube (Fig. 3;Supplementary Fig. 3 (OSM)), where- test phase (Table 1;Fig. 3; Supplementary Figs. 1 and 2 as none of the Ara ambiguus did so at this stage (Fig. 3; (OSM)). Additionally, very few subjects inserted stones into Supplementary Fig. 4 (OSM)). Fig. 5 The number of stones inserted by Captain into the ‘correct’ and the inaccessible), tube. He did not have an absolute side preference in the ‘wrong’ tube in the parallel tubes and the blocked tube task. The top chart parallel tube task, but had successful trials from inserting stones from of each task shows the stones inserted in the baited tubes, the bottom either side. At this stage he also had many trials in which he didn’t make shows the unbaited tubes. Captain mostly avoided the unbaited tube in the error of inserting rocks from the opposing side which he had started. the parallel tubes task, but began to fail trials in the blocked tube task by He did appear to lose motivation in the blocked tube task inserting tubes into this baited, but blocked (so the reward was 118 Learn Behav (2021) 49:106–123 One more Ara glaucogularis, Mr Huang, had a few suc- also the only individual who consistently, and spontaneously, cessful trials in critical test 2, but was unable to repeat this inserted stones into the horizontal tube in the pre-test (Fig. 3, behaviour consistently (Supplementary Fig. 5 (OSM)). Of the Supplementary Figs 1 and 2 (OSM)), prior to any functional eight Ara ambiguus that failed, six were unable to complete experience examples. Thus it is also possible that Lady did not the short-tube experience, both before and after the pre- need to learn anything functional from the pre-inserted stick inserted stone experience, so they were never able to continue experience in order to be successful. The pre-inserted stick on to critical test 2 (Table 1). The other two Ara ambiguus experience may have just showed her that the walnut reward failed at critical test 2 (Supplementary Fig. 6 (OSM)). Of the was obtainable, and thus motivated her to persist with the six Ara glaucogularis that failed, two were unable to complete same exploratory behaviour of inserting stones that she had the short-tube experience (both before and after the pre- already shown in the pre-test. inserted stone experience; Table 1) and the remaining four Captain (Ara glaucogularis) and Enya (Ara ambiguus) stopped at critical test 2 (Supplementary Fig. 5 (OSM)). were both able to consistently manufacture the stone construc- Thus in total, five Ara glaucogularis and three Ara ambiguus tion in critical test 2, after they had the short tube functional reached critical test 2. All the birds that attempted critical test experience. This experience showed the birds that a single 2 inserted multiple stones into the apparatus, but most failed stone was able to push a reward out of a much shorter hori- because they tried to insert them from both sides of the tube zontal tube. It was a much more direct example of the effect (Fig. 3; Supplementary Figs. 5 and 6 (OSM)). the stones could have on the rewards in the tube. Innovation of There was a decrease in the number of times the subjects the stone construction at this stage of the experiment is much placed stones in the vertical tube on top of the apparatus over less likely to represent an innovation based on functional un- the pre-test, critical test and critical test 2 (Supplementary derstanding. Instead, the birds are likely to have just been Fig. 8 (OSM)). All but two birds inserted stones into this top repeating the learned successful behaviour of inserting a sin- tube in at least one trial in the pre-test, and many inserted gle stone to get the reward, and this in turn led to them making stones in more than one trial. the multi-stone construction, which eventually pushed the Example videos of every phase of the experiment, includ- walnut out of the horizontal tube. The fact that all the subjects ing failed and successful trials of the critical tests, can be that took part in critical test 2 began to insert multiple stones at found on figshare (O’Neill & von Bayern, 2020). this stage (Fig. 3) suggeststhatthiswas the underlyingmoti- vation to all the subjects’ behaviour at this stage. The differ- ence with Captain and Enya is that for some reason they began Discussion to more consistently insert stones from a single side of the tube. This may have started initially as a side preference, but Three subjects (two Ara glaucogularis and one Ara its possible that after multiple successes they began to more ambiguus) were able to consistently manufacture a multi- purposefully manufacture the tool from a single side only. stone construction to obtain the walnut reward. These birds The three successful subjects each manufactured the multi- were able to create a functional stone-construction by combin- stone construction without errors on four occasions during ing multiple stone components one behind the other to push their successful critical test runs. Out of context, the videos the walnut out of the tube. Due to their successes coming at of these trials are compelling and can look fairly convincing different stages of the experiment, and because of some of the that the subjects were building these stone constructions with behaviours associated with their successful manufacture, it is ‘purpose’ to push the walnut rewards (see figshare at likely that these successful subjects had a different under- O’Neill & von Bayern, 2020, for videos). However, this in- standing of their stone constructions and innovated them in terpretation should be considered with extreme caution. different ways. Captain was able to manufacture the stone construction with- Lady (Ara glaucogularis) manufactured her first successful out error on his first trial in critical test 2 (Supplementary Fig. stone construction in critical test 1 after having the functional 5 (OSM); also shown in the supplementary videos) but sub- experience that a stick tool could push the reward out of the sequently produced many error-laden trials. Enya had an ex- tube. At this stage of the experiment, she had not had a specific tremely error laden (but successful) first trial in critical test 2, positive experience that the stones could be used to obtain the but then immediately produced two ‘perfect’ trials in the mi- reward in this specific apparatus (although she did have pre- nutes after this (trials 2 and 3, Supplementary Fig. 6 (OSM); vious experience that putting singular stones into a different also shown in the supplementary videos). She also subse- apparatus could lead to rewards; O’Neill et al., 2020). This quently produced many error-laden trials. Lady also shows kind of innovation suggests its possible she may have had similar behaviour of a mix of ‘perfect’ and error-laden some form of functional understanding of the stick tool exam- trials (Supplementary Fig. 3 (OSM) and supplementary ple, and had positioned the stones to recreate its underlying videos). The fact that these subjects only produced these ‘per- purpose. On the other hand, it should be noted that Lady was fect’ trials erratically makes it very difficult to interpret why Learn Behav (2021) 49:106–123 119 they sometimes manufacture the multi-stone construction standing closer to the reward, and it was likely difficult to without error, and sometimes they do not. inhibit repeating the stone insertion behaviour they had learnt To an extent, all the successful subjects appeared to show led to a reward even though it pushed the reward further away an egocentric understanding (Woodward, 2011) of their be- from them again. Interestingly, this exact same mistake was haviour, i.e. they were able to replicate it consistently and observed in the original experiment that was run with capu- rapidly in trials following their initial success. They may have chins (Visalberghi & Trinca, 1989). It is also an error that has recognised which of their behaviours (repeatedly inserting been noted in other parrots taking part in an experiment with a stones from a single side of the horizontal tube) was producing very similar setup, the trap tube (Liedtke et al., 2011), and is the reward outcome. However, Captain did have one success- something we had to control for when we tested a version of ful trial in critical test 1 but was not able to replicate it at this the trap tube on the same group of parrots tested in the current stage of the experiment and Mr Huang (Ara glaucogularis) experiment (O’Neill et al., 2018). even had four successful trials early in critical test 2, but was This counter-productive stone insertion mistake was also then unable to replicate these successes. Neither of these fail- observed in all of the unsuccessful subjects that were inserting ures appeared to be due to a lack of motivation as both were stones into the tube. All of these unsuccessful subjects went attempting to obtain rewards by inserting multiple stones into through the short-tube experience where it is likely that they the horizontal tube, but they were just inserting them from learned a simple heuristic rule of ‘insert stone, obtain reward’. both sides of the tube at once which led to the reward being Simply expanding this rule in a quantitative manner failed in trapped. It is unclear exactly which cue made the subjects critical test 2 as the subjects had to ensure they inserted suffi- recognise (or not recognise in Mr Huang’s instance) that the cient stones from a single side. It is thus possible that the stones must be more consistently inserted from only a single successful birds did not have a different ‘rule’ to the other side of the horizontal tube. birds; they may have just had a weak side preference whereas Regarding the successful birds, they made a number of the other subjects had no side preference at all. The three mistakes when manufacturing their stone constructions. successful subjects (Enya, Lady and Captain) only collected Each of the successful subjects only had four trials in their the rewards from one side of the apparatus in all of the suc- successful critical test in which they only inserted stones from cessful trials in the critical tests, i.e. they always built their a single side of the tube (Fig. 3; Supplementary Figs. 3, 5 and stone construction from a specific side of the apparatus, they 6 (OSM)). On all their other trials (the majority), they inserted did not flexibly build it from either side. stones in both ends of the tube, thus on both sides of the Considering this side preference, it suggests why Lady and reward. This type of ‘mistake’ is not just inefficient, but is Enya failed in the parallel tubes task, in which only one tube counter-productive. It meant that more than the minimum was baited. Figure 4 shows that Lady was clearly inserting number of stones had to be inserted in the original end of stones into both tubes (both the baited and unbaited tube), just the tube to push out both the erroneous stones and then the from a specific side (also see supplementary videos at O'Neill reward. This suggests that the successful subjects did not have & von Bayern, 2020). Although Enya was also mostly a mental representation of the final multi-stone tool that they inserting stones from one side of the two tubes, her data are were creating. This contrasts with a tool using corvid example, less convincing in supporting this side preference speculation. the New Caledonian Crow (corvus moneduloides), which Interestingly, Captain appears to have stopped having a were able to tear a piece of paper off a larger piece so that it side bias in this first parallel tubes follow-up task, and had resembled a token of the same size to that which they had successful trials in which he made his stone construction from previously learned could be traded for a reward, i.e. they could both directions of the tube (Fig. 5; examplesofsuccess from manufacture something that was similar to a template they had both sides in supplementary videos at O'Neill & von Bayern, mentally stored (Jelbert et al., 2018). Observationally, the 2020). He also vastly improved the efficiency with which he counter-productive stone insertion behaviour often happened made the multi-stone construction, exhibiting many trials in after the subjects had manufactured an almost complete multi- which he did not make the ‘mistake’ of inserting stones from stone construction from one side, then walked to the other side both sides of the tube and he only inserted stones from one to check if they had pushed the reward far enough. When they side of the tube to obtain the reward (11 out of 18 successful could not reach the reward, they often inserted a stone into this trials, although in four of those he also inserted stones into the other side of the tube that they were now standing next to (the un-baited tube as well). Observationally, one way that Captain supplementary videos provide examples of this error; see behaved differently to Enya and Lady is that he showed a figshare at O’Neill & von Bayern, 2020, for videos). This preference to stand on top of the apparatus when he was may have been a self-control issue, as the birds would have interacting with it, as opposed to the surface next to it. From had to walk back to the initial side of the tube in which they personal observations, Captain was a more nervous bird. One had already inserted stones in order to continue their stone behaviour the parrots manifested when they were nervous was construction already on the go. Instead, they were now try to stand on a higher perch and in this case the apparatus 120 Learn Behav (2021) 49:106–123 was the highest point in the testing room, hence he would pre-test, was to insert a stone into this vertical tube stand on top of it. From this standpoint he had to move shorter (Supplementary Fig. 8 (OSM)). Only two subjects did not distances to interact with the two ends of each tube, and was insert stones into the vertical tube in the pre-test and only also able to transfer between the tubes very rapidly. Enya and one further subject did not place stones in this vertical tube Lady, on the other hand, had to walk ‘the long way round’ if in more than one trial. It could be argued that placing a stone in they wanted to access different ends to the tubes. It’s possible the pipe is evidence of a poor means-ends understanding of that standing on top of the apparatus slightly reduced the self- their actions and that their behaviour appeared to be guided control issues, as described above, which is what made more by trial-and-error exploration. However, not dropping Captain more efficient in creating the stone construction in stones into that tube would have required suppressing a pre- the parallel tubes follow-up task. viously reinforced behaviour and, thus, it should not be Lady and Enya failed the parallel tubes transfer task, al- regarded as a clear-cut conclusion. Moreover, any exploration, though it is possible that a lack of attention combined with even if it was misdirected and could not lead to success, lack of self-control (as discussed earlier), rather than a lack of helped the subjects learn more about the apparatus and its means-end understanding of the stone construction, was the functional properties (Reader, 2015). The one subject that reason for this. Lady especially placed many stones in the un- did not place stones in the top was the least explorative bird baited ‘distractor’ tube and Enya especially began to place overall (he was the only bird to not make any combinatory many stones in from both ends of the rewarded tube, although actions between the stones and the horizontal tube both made both errors at this stage. They both regularly ran (Supplementary Table 1)) and had no success in any of the out of stones. This behaviour suggests that Lady and Enya current tasks (Hannibal), nor did he have any success in the were probably not manufacturing their stone construction with previous task where the subjects had learned to drop a stone the target to specifically push the reward, but that instead they into a vertical tube (O’Neill et al., 2020). Further to this, as likely had a more simple understanding of the apparatus, e.g. subjects did not place many stones into the sides of the hori- that inserting multiple stones would – at some point – make a zontal tube in the pre-test but placed many stones in the ver- reward appear. Further, it may be a reflection of the macaw’s tical tube during the pre-test, it suggests that the subjects who relative lack of self-control to suppress a previously reinforced eventually had success in the horizontal tube task did not movement (Kabadayi et al., 2017), since they had never ex- succeed just because of carried-over behaviour from the suc- perienced a cost to inefficient stone inserting and had only cess in the previous experiment. Thus, the most vital aspect to been faced with tubes containing rewards previously. be learned from the addition of this vertical-tube was that Limited use of the stone resources only became an issue in subjects will try a previously used and successful method on the transfer tasks, and therefore subjects may not have a novel task before attempting something new. recognised that they could run out of them. Perhaps if they Such a behavioural strategy, namely to exploit a similar had been given more trials, they would have learnt to be more feeding opportunity to one that is previously known, is efficient and to pay attention where to insert the stones. thought to be a key mechanism characterising good innovators Captain, on the other hand, was able to direct his multi-stone (Tebbich et al., 2016). The innovation structure described in construction making behaviour towards the baited tube, which Tebbich et al. (2016) provides a useful framework for how suggests he recognised that he had to target the stones towards each subject’s behavioural interaction with the apparatus was the reward, not just the tubes in general. However, he was possibly guided. As an example, the most successful subject, unsuccessful on the blocked-tube task, in which both tubes Captain, likely started the experiment with the knowledge were baited but one had a blocked end. This suggests that from previous experimental apparatuses to ‘insert stone in while recognising the need to direct stones towards the re- top’ to access food. When this failed, he explored the appara- ward, he did not grasp that the reward must be pushed to an tus and moved onto ‘insert stones in tube ends’ (see Captain’s accessible area, but only that the reward must be pushed. behaviour in the first critical test, Fig. 3), which he then newly There were clearly limitations to the majority of the macaws discovered as a favourable interaction. He optimised this rule overall understanding of the physical properties of this task. over successive interactions to discover that he had to ‘insert In this test, an additional vertical tube was attached to the stones from a specific side’ (see Captain’s behaviour in his apparatus that had no effect on the reward outcome. This was successful critical test 2), and finally further optimisation led to see if the subjects continued to use a basic heuristic of him to ‘insert stones towards the reward from a single side’ ‘place stone on top of apparatus’ to get the reward, as this (see Captain’s behaviour in the parallel tubes task). All sub- was something that they had learnt led to success in a previous jects appeared to go through this sort of heuristic ‘evolution’, experiment (O’Neill et al., 2020; all except two subjects had but many of them never moved on from the first two stages. learnt this behaviour (see Supplementary Table 2 in this This suggests that almost all of the subjects were capable of paper)). Indeed, one of the first actions that almost all of the recognising the feeding opportunity from the apparatus, and subjects did, in their initial exploration of the apparatus in the even discovering that there was a favourable behavioural Learn Behav (2021) 49:106–123 121 Funding Open Access funding enabled and organized by Projekt DEAL. interaction through exploration. What the majority of subjects struggled with was the repetition and testing required to opti- Compliance with ethical standards mise this favourable interaction to make exploiting this food as a worthwhile opportunity. A critical stage of the innovation Ethics Statement All applicable international, national, and institutional framework (Tebbich et al., 2016) that is missing from our guidelines for the care and use of animals were followed. In accordance experiment is the opportunity to socially learn from other’s with the German Animal Welfare Act of 25 May 1998, Section V, Article behaviour. If the other subjects could have watched Captain, 7 and the Spanish Animal Welfare Act 32/2007 of 7 November 2007, Preliminary Title, Article 3, the study was classified as non-animal ex- Lady and Enya, perhaps they may have improved their inter- periment and did not require any approval from a relevant body. action with the equipment. This is likely what would happen in the wild when conspecifics have discovered a novel feeding Open Access This article is licensed under a Creative Commons opportunity. Nevertheless, it is interesting that none of the Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as subjects ever showed a truly flexible behaviour that would long as you give appropriate credit to the original author(s) and the suggest a functional understanding of the underlying mechan- source, provide a link to the Creative Commons licence, and indicate if ics of the task, preferring always to instead rely on a simple changes were made. The images or other third party material in this article rule or heuristic (Hutchinson & Gigerenzer, 2005). This sug- are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the gests that these parrots are unlikely to use a complex mechan- article's Creative Commons licence and your intended use is not ical causal understanding of these kinds of tasks (Johnson & permitted by statutory regulation or exceeds the permitted use, you will Ahn, 2017;Woodward, 2011). need to obtain permission directly from the copyright holder. To view a Overall, the subjects in this task showed less, or possibly copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. zero, functional understanding of their multi-stone construc- tions when compared to other parrots’ functional understand- ing of tools they can use in problem-solving tasks (Auersperg References et al. 2011; Auersperg, von Bayern, Gajdon et al. 2011, 2018; Auersperg et al., 2012;Laumer etal., 2017). As stated, the Auersperg, A. M. I., Huber, L., & Gajdon, G. K. (2011a). Navigating a only individual that appeared to be using the stones in order to tool end in a specific direction: Stick-tool use in kea (Nestor directly manipulate the position of the walnut was Captain, as notabilis). Biology Letters, 7(6), 825–828. https://doi.org/10.1098/ he was the only one that would place the stones in the direc- rsbl.2011.0388 tion towards the walnut in the parallel tubes task. So although Auersperg, A. M. I., Köck, C., O’Hara, M., & Huber, L. (2018). Tool making cockatoos adjust the lengths but not the widths of their tools both species of macaw in this study appeared to show some to function. PLOS ONE, 13(11), e0205429. https://doi.org/10.1371/ evidence of innovative object manipulation in order to solve a journal.pone.0205429 problem-solving task, we still need to further study the differ- Auersperg, A., von Bayern, A., Gajdon, G., Huber, L., & Kacelnik, A. ences between different parrot species (in this instance: Kea, (2011b). Flexibility in Problem Solving and Tool Use of Kea and New Caledonian Crows in a Multi Access Box Paradigm. PLoS Goffin’s cockatoos, Ara glaucogularis and Ara ambiguus) ONE, 6(6), e20231. https://doi.org/10.1371/journal.pone.0020231 that make only some of them capable understanding the func- Auersperg, Alice M.I., Szabo, B., von Bayern, A. M. P., & Kacelnik, A. tional properties of more complex physical innovations. (2012). Spontaneous innovation in tool manufacture and use in a Goffin’scockatoo. Current Biology, 22(21), R903–R904. https:// Supplementary Information The online version contains supplementary doi.org/10.1016/j.cub.2012.09.002 material available at https://doi.org/10.3758/s13420-020-00449-y. Bird, C. D., & Emery, N. J. (2009). Insightful problem solving and cre- ative tool modification by captive nontool-using rooks. Proceedings of the National Academy of Sciences, 106(25), 10370–10375. Acknowledgements We thank Luisana Carballo and Anastasia https://doi.org/10.1073/pnas.0901008106 Krasheninnikova for their comments on earlier versions of the Borsari, A., & Ottoni, E. B. (2005). Preliminary observations of tool use manuscript. in captive hyacinth macaws (Anodorhynchus hyacinthinus). Animal We thank the Loro Parque and its president, Mr Wolfgang Kiessling, for Cognition, 8(1), 48–52. https://doi.org/10.1007/s10071-004-0221-3 their support, the access to the birds and the research facilities. We thank the Breen, A. J., Guillette, L. M., & Healy, S. D. (2016). What Can Nest- Loro Parque Fundación and its president Mr Christoph Kiessling for their Building Birds Teach Us? Comparative Cognition & Behavior collaboration and the staff of the Loro Parque Fundación, the animal Reviews, 11,83–102. https://doi.org/10.3819/ccbr.2016.110005 caretakers and the veterinary department for their support. Burkart, J. M., Schubiger, M. N., & van Schaik, C. P. (2017). The evo- LO’N is a member of the International Max Planck Research School lution of general intelligence. Behavioral and Brain Sciences, 40, (IMPRS) for Organismal Biology. The study was funded by the Max- e195. https://doi.org/10.1017/S0140525X16000959 Planck Society. Byrne, R., Corp, N., & Byrne, J. (2001). Estimating the complexity of animal behaviour: How mountain gorillas eat thistles. Behaviour, Author contributions The design of the experiment was initially con- 138(4), 525–557. https://doi.org/10.1163/156853901750382142 ceived by LO’N and AvB, with contributions from RR and RH. The Chow, P. K. Y., Lea, S. E. G., & Leaver, L. A. (2016). How practice experiments were carried out by RR and RH. Videos were coded by makes perfect: The role of persistence, flexibility and learning in LO'N and RH. Data were visualised by LO’N. The initial manuscript problem-solving efficiency. Animal Behaviour, 112,273–283. was written by LO’N with feedback from AvB. https://doi.org/10.1016/j.anbehav.2015.11.014 122 Learn Behav (2021) 49:106–123 Emery, N. J., & Clayton, N. S. (2004). The Mentality of Crows: Behaviour, 156(5–8), 391–407. https://doi.org/10.1163/1568539X- Convergent Evolution of Intelligence in Corvids and Apes. 00003527 Science, 306(5703), 1903–1907. https://doi.org/10.1126/science. Lambert, M. L., Schiestl, M., Schwing, R., Taylor, A. H., Gajdon, G. K., Slocombe, K. E., & Seed, A. M. (2017). Function and flexibility of Fragaszy, D. M., & Mangalam, M. (2018). Tooling. Advances in the object exploration in kea and New Caledonian crows. Royal Society Study of Behavior, 50, 177–241. https://doi.org/10.1016/bs.asb. Open Science, 4(9), 170652. https://doi.org/10.1098/rsos.170652 2018.01.001 Lambert, M. L., Seed, A. M., & Slocombe, K. E. (2015). A novel form of Girndt, A., Meier, T., & Call, J. (2008). Task constraints mask great apes’ spontaneous tool use displayed by several captive greater vasa par- ability to solve the trap-table task. Journal of Experimental rots ( Coracopsis vasa ): Table 1. Biology Letters, 11(12), Psychology: Animal Behavior Processes, 34(1), 54–62. https://doi. 20150861. https://doi.org/10.1098/rsbl.2015.0861 org/10.1037/0097-7403.34.1.54 Laumer, I. B., Bugnyar, T., Reber, S. A., & Auersperg, A. M. I. (2017). Goodman, M., Hayward, T., & Hunt, G. R. (2018). Habitual tool use Can hook-bending be let off the hook? Bending/unbending of pliant innovated by free-living New Zealand kea. Scientific Reports, tools by cockatoos. Proceedings of the Royal Society B: Biological 8(1), 13935. https://doi.org/10.1038/s41598-018-32363-9 Sciences, 284(1862), 20171026. https://doi.org/10.1098/rspb.2017. Griffin, A. S., Diquelou, M., & Perea, M. (2014). Innovative problem solving in birds: A key role of motor diversity. Animal Behaviour, Liedtke, J., Werdenich, D., Gajdon, G. K., Huber, L., & Wanker, R. 92,221–227. https://doi.org/10.1016/j.anbehav.2014.04.009 (2011). Big brains are not enough: Performance of three parrot spe- Gutiérrez-Ibáñez, C., Iwaniuk, A. N., & Wylie, D. R. (2018). Parrots have cies in the trap-tube paradigm. Animal Cognition, 14(1), 143–149. evolved a primate-like telencephalic-midbrain-cerebellar circuit. https://doi.org/10.1007/s10071-010-0347-4 Scientific Reports, 8(1), 9960. https://doi.org/10.1038/s41598-018- Martin-Ordas, G., Call, J., & Colmenares, F. (2008). Tubes, tables and 28301-4 traps: Great apes solve two functionally equivalent trap tasks but Heinsohn, R., Zdenek, C. N., Cunningham, R. B., Endler, J. A., & show no evidence of transfer across tasks. Animal Cognition, Langmore, N. E. (2017). Tool-assisted rhythmic drumming in palm 11(3), 423–430. https://doi.org/10.1007/s10071-007-0132-1 cockatoos shares key elements of human instrumental music. Martin-Ordas, G., Jaek, F., & Call, J. (2012). Barriers and traps: Great Science Advances, 3(6), e1602399. https://doi.org/10.1126/sciadv. apes’ performance in two functionally equivalent tasks. Animal Cognition, 15(5), 1007–1013. https://doi.org/10.1007/s10071-012- Herculano-Houzel, S. (2017). Numbers of neurons as biological corre- 0504-z lates of cognitive capability. Current Opinion in Behavioral McCormack, T., Hoerl, C., & Butterfill, S. (2011). Tool use and causal Sciences, 16,1–7. https://doi.org/10.1016/j.cobeha.2017.02.004 cognition. Oxford University Press. https://doi.org/10.1093/acprof: Heyes, C. M. (1993). Anecdotes, training, trapping and triangulating: Do oso/9780199571154.001.0001 animals attribute mental states? Animal Behaviour, 46(1), 177–188. Menchetti, M., Mori, E., & Angelici, F. M. (2016). Effects of the Recent https://doi.org/10.1006/anbe.1993.1173 World Invasion by Ring-Necked Parakeets (Psittacula krameri). In Hutchinson, J. M. C., & Gigerenzer, G. (2005). Simple heuristics and F. M. Angelici (Ed.), Problematic Wildlife (253–266). Springer rules of thumb: Where psychologists and behavioural biologists International Publishing. https://doi.org/10.1007/978-3-319-22246- might meet. Behavioural Processes, 69(2), 97–124. https://doi.org/ 2_12 10.1016/j.beproc.2005.02.019 Mulcahy, N. J., & Call, J. (2006). How great apes perform on a modified Iwaniuk, A. N., Dean, K. M., & Nelson, J. E. (2005). Interspecific trap-tube task. Animal Cognition, 9(3), 193–199. https://doi.org/10. Allometry of the Brain and Brain Regions in Parrots 1007/s10071-006-0019-6 (Psittaciformes): Comparisons with Other Birds and Primates. O’Hara, M., Mioduszewska, B., Haryoko, T., Prawiradilaga, D. M., Brain, Behavior and Evolution, 65(1), 40–59. https://doi.org/10. Huber, L., & Auersperg, A. (2019). Extraction without tooling 1159/000081110 around—The first comprehensive description of the foraging- and Jelbert, S. A., Hosking, R. J., Taylor, A. H., & Gray, R. D. (2018). Mental socio-ecology of wild Goffin’s cockatoos (Cacatua goffiniana). template matching is a potential cultural transmission mechanism Behaviour, 156(5–8), 661–690. https://doi.org/10.1163/1568539X- for New Caledonian crow tool manufacturing traditions. Scientific Reports, 8(1), 8956. https://doi.org/10.1038/s41598-018-27405-1 Olkowicz, S., Kocourek, M., Lučan, R. K., Porteš, M., Fitch, W. T., Johnson, S. G. B., & Ahn, W. (2017). Causal Mechanisms. In M. R. Herculano-Houzel, S., & Němec, P. (2016). Birds have primate- Waldmann (Ed.), The Oxford handbook of Causal like numbers of neurons in the forebrain. Proceedings of the Reasoning. Oxford University Press. https://doi.org/10.1093/ National Academy of Sciences, 113(26), 7255–7260. https://doi. oxfordhb/9780199399550.013.12 org/10.1073/pnas.1517131113 Kabadayi, C., Krasheninnikova, A., O’Neill, L., van de Weijer, J., O'Neill, L., & von Bayern, A. M. P. (2020). Supplementary materials to Osvath, M., & von Bayern, A. M. P. (2017). Are parrots poor at ‘Innovative problem solving in macaws’. figshare. Media. https:// motor self-regulation or is the cylinder task poor at measuring it? doi.org/10.6084/m9.figshare.12349418.v2 Animal Cognition, 20(6), 1137–1146. https://doi.org/10.1007/ O’Neill, L., Picaud, A., Hastings, R., Buffenoir, N., Gahr, M., & von s10071-017-1131-5 Bayern, A. M. P. (2020). Causal Understanding of the Stone Kacelnik, A. (2009). Tools for thought or thoughts for tools? Proceedings Dropping Task in Two Species of Macaw [Preprint]. BiorXiv. of the National Academy of Sciences, 106(25), 10071–10072. https://doi.org/10.1101/2020.08.24.264390 https://doi.org/10.1073/pnas.0904735106 O’Neill, L., Picaud, A., Maehner, J., Gahr, M., & von Bayern, A. M. P. Köhler, W. (1917) 1925. The mentality of apes (translated from the second revised edition by Ella Winter). New York: Harcourt, (2018). Two macaw species can learn to solve an optimised two-trap Brace, & World. problem, but without functional causal understanding. Behaviour, 156(5–8), 691–720. https://doi.org/10.1163/1568539X-00003521 Krasheninnikova, A., Berardi, R., Lind, M.-A., O’Neill, L., & von Bayern, A. M. P. (2019). Primate cognition test battery in parrots. Osuna-Mascaró, A. J., & Auersperg, A. M. I. (2018). On the brink of tool Behaviour, 156(5–8), 721–761. https://doi.org/10.1163/1568539X- use? Could object combinations during foraging in a feral Goffin’s 0003549 cockatoo (Cacatua goffiniana) result in tool innovations? Animal Lambert, M. L., Jacobs, I., Osvath, M., & von Bayern, A. M. P. (2018). Behavior and Cognition, 5(2), 229–234. https://doi.org/10.26451/ Birds of a feather? Parrot and corvid cognition compared. abc.05.02.05.2018 Learn Behav (2021) 49:106–123 123 Osvath, M., Kabadayi, C., & Jacobs, I. (2014). Independent Evolution of Van Horik, J. O., Clayton, N. S., & Emery, N. J. (2012). Convergent Similar Complex Cognitive Skills: The Importance of Embodied Evolution of Cognition in Corvids, Apes and Other Animals. In Degrees of Freedom. Animal Behavior and Cognition, 1(3), 249. Shackelford, T.K. & Vonk, J. (Eds.), The Oxford Handbook of https://doi.org/10.12966/abc.08.03.2014 Comparative Evolutionary Psychology. Oxford University Press. Reader, S. M. (2015). Causes of Individual Differences in Animal https://doi.org/10.1093/oxfordhb/9780199738182.013.0005 Exploration and Search. Topics in Cognitive Science, 7(3), 451– van Horik, J. O., & Madden, J. R. (2016). A problem with problem 468. https://doi.org/10.1111/tops.12148 solving: Motivational traits, but not cognition, predict success on Shumaker, R. W., Walkup, K. R., & Beck, B. B. (2011). Animal tool novel operant foraging tasks. Animal Behaviour, 114, 189–198. behavior: The use and manufacture of tools by animals. Baltimore: https://doi.org/10.1016/j.anbehav.2016.02.006 Johns Hopkins University Press. Villegas-Retana, S. A., & Araya-H D. (2017). Consumo de almendro de Sol, D., Duncan, R. P., Blackburn, T. M., Cassey, P., & Lefebvre, L. playa (Terminalia catappa) y uso de hojas como herramienta por (2005). Big brains, enhanced cognition, and response of birds to parte del ave Ara ambiguus (Psittaciformes: Psittacidae)en Costa novel environments. Proceedings of the National Academy of Rica. UNED Research Journal, 9(2). https://doi.org/10.22458/urj. Sciences, 102(15), 5460–5465. https://doi.org/10.1073/pnas. v9i2.1894 Visalberghi, E., & Limongelli, L. (1994). Lack of comprehension of Sol, Daniel, Timmermans, S., & Lefebvre, L. (2002). Behavioural flexi- cause-effect relations in tool-using capuchin monkeys (Cebus bility and invasion success in birds. Animal Behaviour, 63(3), 495– apella). Journal of Comparative Psychology, 108(1), 15–22. 502. https://doi.org/10.1006/anbe.2001.1953 https://doi.org/10.1037/0735-7036.108.1.15 Taylor, A. H., & Gray, R. D. (2014). Is there a link between the crafting of Visalberghi, E., & Trinca, L. (1989). Tool use in capuchin monkeys: tools and the evolution of cognition?: Is cognitive evolution affected Distinguishing between performing and understanding. Primates, by tool crafting? Wiley Interdisciplinary Reviews: Cognitive 30(4), 511–521. https://doi.org/10.1007/BF02380877 Science, 5(6), 693–703. https://doi.org/10.1002/wcs.1322 von Bayern, A. M. P., Danel, S., Auersperg, A. M. I., Mioduszewska, B., Taylor, A. H., Hunt, G. R., Holzhaider, J. C., & Gray, R. D. (2007). & Kacelnik, A. (2018). Compound tool construction by New Spontaneous Metatool Use by New Caledonian Crows. Current Caledonian crows. Scientific Reports, 8(1), 15676. https://doi.org/ Biology, 17(17), 1504–1507. https://doi.org/10.1016/j.cub.2007. 10.1038/s41598-018-33458-z 07.057 von Bayern, Auguste M.P., Heathcote, R. J. P., Rutz, C., & Kacelnik, A. Tebbich, S., Griffin, A. S., Peschl, M. F., & Sterelny, K. (2016). From (2009). The Role of Experience in Problem Solving and Innovative mechanisms to function: An integrated framework of animal inno- Tool Use in Crows. Current Biology, 19(22), 1965–1968. https:// vation. Philosophical Transactions of the Royal Society B: doi.org/10.1016/j.cub.2009.10.037 Biological Sciences, 371(1690), 20150195. https://doi.org/10. Weir, A. A. S., & Kacelnik, A. (2006). A New Caledonian crow (Corvus 1098/rstb.2015.0195 moneduloides) creatively re-designs tools by bending or unbending Tella, J. L., Baños-Villalba, A., Hernández-Brito, D., Rojas, A., Pacífico, aluminium strips. Animal Cognition, 9(4), 317–334. https://doi.org/ E., Díaz-Luque, J. A., Carrete, M., Blanco, G., & Hiraldo, F. (2015). 10.1007/s10071-006-0052-5 Parrots as overlooked seed dispersers. Frontiers in Ecology and the Wood, G. A. (1984). Tool use by the Palm Cockatoo Probosciger Environment, 13(6), 338–339. https://doi.org/10.1890/1540-9295- aterrimus during display. Corella, 8,94–95. 13.6.338 Woodward, J. (2011). A Philosopher Looks at Tool Use and Causal Teschke, I., Wascher, C. A. F., Scriba, M. F., von Bayern, A. M. P., Understanding. In McCormack, T., Hoerl, C., Butterfill, S. (Eds.), Huml, V., Siemers, B., & Tebbich, S. (2013). Did tool-use evolve Tool Use and Causal Cognition. Oxford University Press. with enhanced physical cognitive abilities? Philosophical Transactions of the Royal Society B: Biological Sciences, 368(1630), 20120418–20120418. https://doi.org/10.1098/rstb. Supplementary data and videos 2012.0418 Data from the video coding and Supplementary videos can be found Thornton, A., & Lukas, D. (2012). Individual variation in cognitive per- on figshare at: https://doi.org/10.6084/m9.figshare.12349418.v1 formance: Developmental and evolutionary perspectives. Philosophical Transactions of the Royal Society B: Biological Sciences, 367(1603), 2773–2783. https://doi.org/10.1098/rstb. 2012.0214 Thornton, Alex, & Samson, J. (2012). Innovative problem solving in wild meerkats. Animal Behaviour, 83(6), 1459–1468. https://doi.org/10. Publisher’snote Springer Nature remains neutral with regard to jurisdic- 1016/j.anbehav.2012.03.018 tional claims in published maps and institutional affiliations. Toft, C. A., & Wright, T. F. (2016). Parrots of the wild: A natural history of the world’s most captivating birds. Berkeley, CA, University of California press.

Journal

Learning & Behavior – Springer Journals

Published: Dec 7, 2020

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Innovative problem solving in macaws

Innovative problem solving in macaws

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Innovative problem solving in macaws

Innovative problem solving in macaws

References (66)

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies