Feature- versus rule-based generalization in rats, pigeons and humans

摘要: Humans can spontaneously create rules that allow them to efficiently generalize what they have learned novel situations. An enduring question is whether rule-based generalization uniquely human or other animals also abstract and apply In recent years, there been a number of high-profile claims such as rats learn rules. Most those are quite weak because it possible demonstrate simple associative systems (which do not rules) account for the behavior in tasks. Using procedure allows us clearly distinguish feature-based from (the Shanks–Darby procedure), we adult humans show this task, while pigeons was based on featural overlap between stimuli. brief, when learning stimulus made two components (“AB”) predicts different outcome than its elements (“A” “B”), people an opposites rule new stimuli (e.g., knowing “C” “D” predict one outcome, will “CD” opposite outcome). Rats reverse behavior—they learned, but basis similarity similar “D”, so same predicted compound components). Genuinely observed humans, pigeons, current procedure.