The role of language in forming the concepts same and different

Relational ability – the ability to compare situations or ideas and discover common relations – is key to our capacity for higher-order cognition. The Relational-Match-to-Sample task (RMTS) is the gold standard to evaluate relational ability, yet recent reports conflict on when and how this ability emerges. Here, we focus on a version of the task in which children receive no training or feedback and ask whether possessing the relational concepts same and different is the key to success. We present two experiments that found 4-year-old children succeeded on three RMTS conditions aligned on the vertical axis (N = 79; 39 female and 40 male; 52% White; Wilcoxon's r ≥ 0.48) and one on the horizontal axis (N = 31; 15 female and 16 male; 73% White; Wilcoxon's r = 0.62). Across both experiments, children who justified their choices using relational language tended to perform better on the RMTS task. We conclude that possession of the same/different concept is the key to success on the RMTS and production of relational language to justify performance is a better predictor than age for indicating possession of the concept.