Variability in cognitive strategies in a numerical discrimination task: individual and comparative insights from...

Sensitivity to numbers is a crucial cognitive ability. Numerical discrimination, defined as the ability to distinguish between different numerosities, allows animals to reduce predation risk, increase foraging efficiency, and maximize food intake, and has been documented in primates, amphibians, fish, birds, and insects. Evidence from day-old chicks suggests that basic numerical competencies are part of a precociously available cognitive toolkit that guides animals in their interaction with the environment. To discriminate between two numerosities, animals could employ either an absolute or a relative strategy. Absolute discrimination requires identifying a specific numerosity (e.g., 10) against others. Relative discrimination consists in selecting the higher (or the lower) numerosity in any pair. Which strategy is preferred differs across species: Humans and fish preferentially adopt a relative strategy, while honeybees apply an absolute strategy. This posed the question how strategy preferences evolve phylogenetically and develop ontogenetically. This study explores whether 3-d-old chicks (Gallus gallus) spontaneously adopt an absolute or relative strategy. During rearing, 124 chicks were exposed to a numerical comparison (5 vs. 10; experiment 1) and learned that one numerosity (10) was associated with food. During testing, which consisted of a 6-min unrewarded observation, each chick was presented with the numerosity associated with food (10, the choice of which would indicate an absolute strategy) alongside an unfamiliar numerosity consistent with the relative strategy (20, which is larger than 10). Based on previous evidence, we hypothesized that chicks would spend more time near the numerosity corresponding to their preferred strategy. Chicks did not show a significant preference on a group level, regardless of whether the higher (experiments 1 and 2) or the lower numerosity (experiment 3) had been associated with food during rearing. Interestingly, chicks exhibited significant individual preferences for one or the other strategy that cannot be explained by the novelty of the stimuli (experiment 4) or by the use of physical variables (experiment 2). These findings suggest either that diversity of numerical discrimination strategies was evolutionary favorable in domestic chickens' natural environments, or that, at this early stage of ontogeny, young chicks' strategies are still highly variable and flexible.