Quantitative evaluations in gray crows: generalization of the relative attribute "larger set"

Area of Science:

• Comparative Cognition
• Animal Behavior
• Avian Intelligence

Background:

• Understanding numerical cognition in non-human animals is crucial for insights into the evolution of mathematical abilities.
• Corvids are known for their intelligence, yet their capacity for abstract numerical concepts requires further investigation.

Purpose of the Study:

• To investigate the ability of gray crows (Corvus corone) to discriminate between sets based on quantity.
• To determine if crows can form a generalized concept of "larger set" independent of other visual cues.
• To assess the transfer of learned numerical discrimination to novel ranges and varying visual parameters.

Main Methods:

• Crows were trained to select the larger of two visual sets within the 1-12 range.
• Stimulus properties (color, form, size, arrangement) were varied, and element area was manipulated to isolate quantity discrimination.
• Performance was tested in the 1-12 range and then transferred to a 10-20 range with similar and varied area-to-number ratios.

Main Results:

• Crows accurately discriminated quantities in the 1-12 range (75.3% accuracy), showing reliance on element number over other cues.
• Successful transfer of the discrimination task to the 10-20 range was observed (71.5% accuracy).
• When element area varied inversely with quantity, only some crows maintained high accuracy (71.9%), suggesting a challenge in abstracting quantity under conflicting visual information.

Conclusions:

• Gray crows demonstrate a robust ability to discriminate quantities and form a generalized concept of "larger set."
• The findings suggest crows possess a higher level of reflective activity and numerical cognition than previously assumed.
• This study contributes to understanding the cognitive flexibility and abstract reasoning capabilities in corvid species.