Smirnova Aa

Crows spontaneously exhibit analogical reasoning

Analogical reasoning is vital to advanced cognition and behavioral adaptation. Many theorists deem analogical thinking to be uniquely human and to be foundational to categorization, creative problem solving, and scientific discovery. Comparative psychologists have long been interested in the species generality of analogical reasoning, but they initially found it difficult to obtain empirical support for such thinking in nonhuman animals (for pioneering efforts, see [2, 3]). Researchers have since mustered considerable evidence and argument that relational matching-to-sample (RMTS) effectively captures the essence of analogy, in which the relevant logical arguments are presented visually. In RMTS, choice of test pair BB would be correct if the sample pair were AA, whereas choice of test pair EF would be correct if the sample pair were CD. Critically, no items in the correct test pair physically match items in the sample pair, thus demanding that only relational sameness or differentness is available to support accurate choice responding. Initial evidence suggested that only humans and apes can successfully learn RMTS with pairs of sample and test items; however, monkeys have subsequently done so. Here, we report that crows too exhibit relational matching behavior. Even more importantly, crows spontaneously display relational responding without ever having been trained on RMTS; they had only been trained on identity matching-to-sample (IMTS). Such robust and uninstructed relational matching behavior represents the most convincing evidence yet of analogical reasoning in a nonprimate species, as apes alone have spontaneously exhibited RMTS behavior after only IMTS training.

A large outdoor radial maze for comparative studies in birds and mammals

For a comparative neurobiological analysis of spatial learning and memory, a large outdoor eight-arm radial maze was constructed which permits behavioral assessment of many avian and mammalian species both from the laboratory or the wild, using the same metric space and session schedules. It consists of a central part of 250cm diameter, and has arms of 650cm length, 170cm height and 80cm width. In order to determine appropriate training schedules for comparison of different species, we tested four mammalian and two avian species during 9-15 sessions: 18 albino rats (Rattus norvegicus), nine outdoors and nine in a conventional small indoor maze; six guinea pigs (Cavia porcellus); six rabbits (Oryctolagus cuniculus); five hedgehogs (Erinaceus europaeus); seven hooded crows (Corvus corone cornix) and six chickens (Gallus domesticus). Rats learned fast in both mazes yet significantly better in the large one. Good-to-excellent learning was also observed in juvenile rabbits and wild-caught crows, although the latter tended to avoid arms in the vicinity of the observer. Hedgehogs and chickens did not show significant learning as a group, but some individuals appeared to learn the task. Guinea pigs remained continuously passive and could not be trained. Thus, in spite of species-specific demands for reward, adaptation and pre-training, this type of radial maze permits to directly compare a wide variety of species. Such comparability is essential for an analysis of underlying neurobiological mechanisms.

Analogical reasoning in amazons

Two juvenile orange-winged amazons (Amazona amazonica) were initially trained to match visual stimuli by color, shape, and number of items, but not by size. After learning these three identity matching-to-sample tasks, the parrots transferred discriminative responding to new stimuli from the same categories that had been used in training (other colors, shapes, and numbers of items) as well as to stimuli from a different category (stimuli varying in size). In the critical testing phase, both parrots exhibited reliable relational matching-to-sample (RMTS) behavior, suggesting that they perceived and compared the relationship between objects in the sample stimulus pair to the relationship between objects in the comparison stimulus pairs, even though no physical matches were possible between items in the sample and comparison pairs. The parrots spontaneously exhibited this higher-order relational responding without having ever before been trained on RMTS tasks, therefore joining apes and crows in displaying this abstract cognitive behavior.

Corvidae Can Understand Logical Structure in Baited String-Pulling Tasks

The ability of the Corvidae to understand the logical structure of a task consisting of pulling a string attached to a bait was studied in a series of experiments with different relative positions of multiple strings. Some hooded crows (Corvus cornix L.) and common raven (Corvus corax L.) could successfully solve tasks in which the strings did not cross each other but were arranged in such a way that the bait was opposite the end of an “empty” string. Hooded crows also solved a task in which the bait was attached to each of two strings, but one string was broken, preventing it from pulling the bait. A task in which two crossing strings, in which the bait was again opposite the end of the empty string was not solved by hooded crows. These data lead to the conclusion that some members of both species are able to pick out the logical structure of tasks of this type.

Prototype symbolization in hooded crows

Studies were undertaken to determine whether four crows previously trained to an image-based abstract selection rule could establish a relationship between the number of elements in a group and initially indifferent symbols (arabic numerals) and operate with these symbols, i.e., studies addressed the ability of these birds to symbolize. Unlike other similar studies, there was no use of special development of associative connections between the symbols and the corresponding arrays of elements, but conditions were created in which birds could observe these relationships on the basis of comparison with previously obtained information. Demonstration series were performed for this purpose, in which correct solutions by the crow resulted in receipt of a number of larvae corresponding to the number of elements in the array pictured or numeral imaged on the stimulus selected by the crow. Images belonged to the same category as the corresponding stimulus for selection (the second stimulus was another category): if the image was an array, then the corresponding picture for selection also bore an array, and vice versa. Crows could solve the task successfully by using an image-based selection rule. In test series, the image and both selection stimuli were initially from different categories: if the image was a numeral, then both selection stimuli were arrays, and vice versa. All four crows successfully coped with this task. Despite the absence of any similarity between the image and “correct” stimulus, they selected the array corresponding to the numeral and vice versa. It is suggested that the birds could achieve this result by comparing the information obtained during the presentation series – about the number of reinforcement units corresponding to each stimulus. Similar experiments showed that crows could use numbers to perform operations analogous to arithmetic addition.

A comparative assessment of birds’ ability to solve string-pulling tasks

A set of string-pulling tasks was used to compare the cognitive abilities of birds with different levels of brain complexity, which was judged using Portmann’s index for the hemispheres. Varying the number and relative position of strings and bait, we investigated whether birds of different species (hooded crows (Corvuscornix), red crossbills (Loxia curvirostra), Eurasian blue tits (Parus caeruleus), and great gray owls (Strix nebulosa)) are capable of comprehending the logical structure of such tasks based on cause-effect relationships. Among the observed representatives of Passeriformes, only hooded crows, which possess the most complex and highly differentiated brain, were shown to be able to solve correctly the most difficult versions of the string-pulling tasks and, therefore, to understand their logical structure. Small passerine birds and owls, which are characterized by a similarly high value of Portmann’s index, proved to be incapable of comprehending the cause-effect relationship between the components of the tasks.

Studying the ability of hooded crows (Corvus cornix L.) to solve trap tube test

The ability of hooded crows (Corvus cornix L.) to use tools has been studied. We determined whether these birds can extract bait from a transparent tube using a piston while avoiding a trap. Six out of eight crows learned to use the piston to extract a food reward from a transparent nontrap tube. One out of these six birds successfully performed the task in which it had to avoid the trap to retrieve the reward, showing spontaneous comprehending of the task structure in the first trial. Another four crows learned to perform this task using the trial-and-error method. Then, these four crows were faced with two similar tasks (tests), in which we changed the relative positions of components in the experimental installation. Thus, we were figuring out what exactly the birds have learned, i.e., whether they learned a set of particular rules of choice, or caught the general structure of these problems. The test results showed that the birds were trying to solve these problems by applying previously learned rules of choice.

Crossbills (Loxia curvirostra) Are Able to Form the "Larger Than" Concept

An original method for studying cognitive abilities in largely wild passerine birds was developed. Studies of five crossbills (Loxia curvirostra) showed that this approach could be used to assess their ability to form concepts. All five crossbills learned selection rules for the “larger number” feature over the range “1–10.” The birds successfully transferred this to stimuli significantly different from those used in training, which were not comparable in terms of all quantitative properties, but only some (only area or number). Only one bird was able to transfer the learned choice rule to multiples in a new range (“10–20”). Thus, the ability of representative small forest passerine birds to form the “larger than” concept was demonstrated.