The ability to identify and retain logical relations between stimuli and apply them to novel stimuli is known as relational concept learning. This has been demonstrated in a few animal species after extensive reinforcement training, and it reveals the brain's ability to deal with abstract properties. Here we describe relational concept learning in newborn ducklings without reinforced training. Newly hatched domesticated mallards that were briefly exposed to a pair of objects that were either the same or different in shape or color later preferred to follow pairs of new objects exhibiting the imprinted relation. Thus, even in a seemingly rigid and very rapid form of learning such as filial imprinting, the brain operates with abstract conceptual reasoning, a faculty often assumed to be reserved to highly intelligent organisms.

A major goal for the next generation of artificial intelligence (AI) is to build machines that are able to reason and cope with novel tasks, environments, and situations in a manner that approaches the abilities of animals. Evidence from precocial species suggests that driving learning through suitable priors can help to successfully face this challenge.

Tool use, though rare, is taxonomically widespread, but morphological adaptations for tool use are virtually unknown. We focus on the New Caledonian crow (NCC, Corvus moneduloides), which displays some of the most innovative tool-related behavior among nonhumans. One of their major food sources is larvae extracted from burrows with sticks held diagonally in the bill, oriented with individual, but not species-wide, laterality. Among possible behavioral and anatomical adaptations for tool use, NCCs possess unusually wide binocular visual fields (up to 60°), suggesting that extreme binocular vision may facilitate tool use. Here, we establish that during natural extractions, tool tips can only be viewed by the contralateral eye. Thus, maintaining binocular view of tool tips is unlikely to have selected for wide binocular fields; the selective factor is more likely to have been to allow each eye to see far enough across the midsagittal line to view the tool's tip monocularly. Consequently, we tested the hypothesis that tool side preference follows eye preference and found that eye dominance does predict tool laterality across individuals. This contrasts with humans' species-wide motor laterality and uncorrelated motor-visual laterality, possibly because bill-held tools are viewed monocularly and move in concert with eyes, whereas hand-held tools are visible to both eyes and allow independent combinations of eye preference and handedness. This difference may affect other models of coordination between vision and mechanical control, not necessarily involving tools.

Pigeons (Columba livia) display reliable homing behaviour, but their homing routes from familiar release points are individually idiosyncratic and tightly recapitulated, suggesting that learning plays a role in route establishment. In light of the fact that routes are learned, and that both ascending and descending visual pathways share visual inputs from each eye asymmetrically to the brain hemispheres, we investigated how information from each eye contributes to route establishment, and how information input is shared between left and right neural systems. Using on-board global positioning system loggers, we tested 12 pigeons' route fidelity when switching from learning a route with one eye to homing with the other, and back, in an A-B-A design. Two groups of birds, trained first with the left or first with the right eye, formed new idiosyncratic routes after switching eyes, but those that flew first with the left eye formed these routes nearer to their original routes. This confirms that vision plays a major role in homing from familiar sites and exposes a behavioural consequence of neuroanatomical asymmetry whose ontogeny is better understood than its functional significance.

No abstract