Reasoning by exclusion in New Caledonian crows (Corvus moneduloides) cannot be explained by avoidance of empty containers.

Whether animals can reason or merely learn associatively is a long-standing debate. Researchers have approached this question by investigating whether dogs, birds, and primates can reason by exclusion (choosing by logically excluding all other alternatives). However, these studies have not resolved whether animals are capable of inferring which option is rewarded or are merely avoiding options known to be incorrect. Here, we used a forced-choice tubes task, where strategies of "reasoning by exclusion" and "avoidance of empty containers" predicted different responses. Two tubes (1 straight, 1 bent) were presented in 5 types of orientation, varying whether the rewarded location could be inferred. We compared predictions from both strategies with the observed performance of 8 wild-caught New Caledonian crows. Two of the 8 birds' choices were entirely consistent with reasoning by exclusion only. A further 4 birds followed a mixed strategy, where both reasoning and avoidance could have influenced their decisions. Thus, although avoidance plays a role, it cannot fully explain the crows' choices. Confirming how animals naturally solve problems is increasingly important in animal cognition; we demonstrate that NC crows can inferentially reason without explicit training, but, like humans, most do not rely solely on reasoning to make decisions.