ORIGINAL PAPER Object-specific and relational learning in pigeons Leyre Castro Edward A. Wasserman Joe ¨l Fagot Anaı ¨s Maugard Received: 4 April 2014 / Revised: 16 June 2014 / Accepted: 22 July 2014 Ó Springer-Verlag Berlin Heidelberg 2014 Abstract Abstract or relational stimulus processing requires an organism to appreciate the interrelations between or among two or more stimuli (e.g., same or dif- ferent, less than or greater than). In the current study, we explored the role of concrete and abstract information processing in pigeons performing a visual categorization task which could be solved by attending to either the specific objects presented or the relation among the objects. In Experiment 1, we gave pigeons three training phases in which we gradually increased the variability (that is, the number of object arrays) in the training set. In Experiment 2, we trained a second group of pigeons with an even larger number of object arrays from the outset. We found that, the larger the variability in the training exemplars, the lesser the pigeons’ attention to object-specific information and the greater their attention to relational information; nev- ertheless, the contribution of object-specific information to categorization performance was never completely elimi- nated. This pervasive influence of object-specific infor- mation is not peculiar to animals, but has been observed in young children and human adults as well. Keywords Relational learning Á Abstract concepts Á Animal cognition Á Pigeons Introduction Learning about the abstract properties of our environment is foundational for higher-order cognition. According to Morgan (1896), abstract information processing requires that we ‘‘neglect all that is variable and focus the attention on the uniform relation. We have [then] reached a con- ception, and this conception is not concrete, particular, and individual, but abstract, general, and of universal applica- tion (p. 263).’’ So, a concept is a general idea that virtually ignores the specific properties of the individual stimuli and focuses instead on the general aspects of the concept itself. It is on this interaction between learning individual items and mastering a general concept that our research project will focus. In relational learning, for example, the organism must learn about the relations between or among items (e.g., sameness, differentness, less than, greater than). Relational learning in animals has been studied, primarily, with tasks that require the discrimination of items that are the same as one another and items that are different from one another (Blaisdell and Cook 2005; Katz et al. 2002; Wasserman and Young 2010; Katz and Wright 2006). In order to speak of a true same-different relational concept, animals must not only acquire the discrimination, but show robust transfer performance to new collections of same and dif- ferent items. And, they do. A variety of species, including monkeys (e.g., Katz et al. 2002; Wasserman et al. 2001a, b), pigeons (e.g., Cook and Wasserman 2012; Katz and Wright 2006), rats (e.g., Wasserman et al. 2012), and harbor seals (e.g., Scholtyssek et al. 2013) can form a relational concept, suggesting that this ability is not uniquely human, but widespread among animals. Concepts have traditionally been specified as abstract rules that are separate from perceptual systems (Anderson L. Castro (&) Á E. A. Wasserman (&) Department of Psychology, The University of Iowa, E11 Seashore Hall, Iowa City, IA 52242, USA e-mail: leyre-castroruiz@uiowa.edu E. A. Wasserman e-mail: ed-wasserman@uiowa.edu J. Fagot Á A. Maugard Laboratory of Cognitive Psychology, CNRS, Aix-Marseille University, Fe ´de ´ration de Recherche 3C, 3 Place Victor Hugo, Ba ˆt B, Case D, 13331 Marseille Cedex, France 123 Anim Cogn DOI 10.1007/s10071-014-0790-8