[Numerousness judgement in hooded crows: generalization by the "more than" concept]

Area of Science:

• Animal cognition
• Comparative psychology
• Avian behavior

Background:

• Numerical cognition is a key area in understanding animal intelligence.
• Previous studies have explored numerical abilities in various species, but research on corvids is ongoing.
• Understanding the limits and mechanisms of numerical discrimination in birds provides insights into cognitive evolution.

Purpose of the Study:

• To investigate the numerical discrimination abilities of carrion crows.
• To determine if crows can discern quantity independently of other visual attributes.
• To assess the crows' capacity for abstract numerical concepts like 'more than'.

Main Methods:

• Four carrion crows were trained to select the larger quantity from pairs of arrays (1-12 items).
• Irrelevant visual cues (shape, size, color, distribution) were systematically varied during training.
• Control procedures, including varying element size relative to quantity, were used to isolate numerical processing.
• Performance was tested with novel quantities (10-20 items) and under conditions manipulating cumulative area cues.

Main Results:

• Crows achieved high accuracy (75.3%) in discriminating quantities from 1-12, indicating reliance on numerosity.
• Successful transfer of the learned task was observed for quantities up to 20 (71.5%) without further training.
• When non-numerical cues (cumulative area) were manipulated, only two crows maintained high accuracy (71.9%), suggesting reliance on numerosity for robust discrimination.

Conclusions:

• Carrion crows can accurately discriminate quantities based on numerosity alone.
• Crows demonstrate the ability to generalize numerical discrimination to larger quantities (up to 20).
• The findings suggest crows can form an abstract concept of 'more than', supporting advanced numerical cognition in birds.