Individual representation and summation of symbols associated with food quantities

Researchers have disagreed about the willingness and ability of animals to use number as a discriminative cue. Furthermore, few have extended quantity investigations to include mathematical operations such as summation. In each of six experiments, pigeons were presented with a choice between two stimulus cards. Each card contained one or two symbols, and each symbol corresponded to a specific quantity of food reward. If pigeons can associate the individual symbols with each food quantity, then they should choose the symbol that would lead to more food. Pigeons were trained with select pairs from the set 0, 1, 3, 5, 7 and 9, and were subsequently tested with new pairs and summation problems. For the summation problems, each stimulus card contained two symbols, and pigeons received food corresponding to the sum of the symbols. In Experiment 1, when grains were used as reward, pigeons were found to choose the symbol(s) associated with the larger total quantity of food reward. Because pigeons could have been basing their decisions on either number or mass of food reward, further tests were carried out with Noyes pellets of various sizes to try to isolate these effects. When each symbol revealed only one pellet that varied in mass (Experiment 2), results from the new pairs and summation problems closely resembled what was found in Experiment 1, suggesting that variation in mass alone was sufficient to form quantity representations. When mass was constant across symbols and number varied (Experiment 3), pigeons were found to treat these quantities as equivalent. However, Experiment 4 revealed that these representations were fairly specific, because individual pigeons developed distinct preferences for the larger or smaller number when mass was constant. Experiment 6 indicated that when the pellets were dropped individually, pigeons continued to score significantly above chance overall on the novel pairs and summation problems. However, performance was far worse than in the other experiments, suggesting that having a visual image of all the food in the food well at once improved performance. Experiment 5 revealed that the distance between symbols had no effect on summation performance, indicating that grouping was determined more by the fact that the symbols shared a stimulus card. Overall, the most parsimonious explanation for the individual quantity representations is that they were based primarily on mass, and that with food rewards, number is not the preferred discriminative cue.