Little is known about how animals represent their own actions in working memory. We investigated whether bottlenosed dolphins could recall actions they had recently performed and reveal those recollections using an abstract rule. Two dolphins were trained to respond to a specific gestural command by repeating the last behavior performed. Both dolphins proved to be able to repeat a wide variety of behaviors on command and were able to generalize the repeating rule to novel behaviors and situations. One dolphin was able to repeat all 36 behaviors she was tested on, including behaviors involving multiple simultaneous actions and self-selected behaviors. These results suggest that dolphins can flexibly access memories of their recent actions and that these memories are of sufficient detail to allow for reenactments. The repeating task can potentially be used to investigate short-term action and event representations in a variety of species.Animals act under the influence of mental representations that result from a variety of internal and external factors. Understanding how animals create, activate, and manipulate such representations is a central concern for studies ofanimal learning and memory. A variety of techniques (e.g., delayed matching-to-sample tasks and maze tasks) have been developed to gain access to "unobservable" neural representational systems (for a review, see Roitblat, 1982Roitblat, , 1987. These techniques have been used extensively to investigate how animals represent external stimuli (e.g., objects, sounds, locations, and events) that they have experienced in the recent past. In contrast, few studies have investigated how animals represent recent internal events, such as the production of behaviors. Consequently, more is currently known about how animals represent environmental conditions in working memory than about how they represent their recent actions. An animal's short-term memory for its own actions can be interpreted as a type ofmetaknowledge; self-reports based on such memories can be used to measure an animal's ability to explicitly recall past behaviors (Shimp, 1982).

Kako (1999)reviews the evidence for syntactic competencies in several animal species exposed to artificial language systems, emphasizing the importance of core syntactic properties such as argument structure and closed-classitems. Wepresent evidence from our dolphin studies for the acquisition ofthe closed-class functionality of demonstratives, prepositions, conjunctions, and locatives. Sensitivityto argument structure is also evidenced by wholly untrained and consistent interpretations of the dolphin to probes of anomalous syntactic structures. These results are generated within our comprehension-based paradigm, which enables us to provide convincingobjective evidence for the development and generalization of concepts by the dolphin subject. Demonstrations of animal language competencies may illuminate certain aspects of human linguisticcompetence by suggestingthat the particular modeled subsets can derive from general cognitive mechanisms, rather than language-specificones.

Over the last quarter century, the dominant tendency in comparative cognitive psychology has been to emphasize the similarities between human and nonhuman minds and to downplay the differences as "one of degree and not of kind" (Darwin 1871). In the present target article, we argue that Darwin was mistaken: the profound biological continuity between human and nonhuman animals masks an equally profound discontinuity between human and nonhuman minds. To wit, there is a significant discontinuity in the degree to which human and nonhuman animals are able to approximate the higher-order, systematic, relational capabilities of a physical symbol system (PSS) (Newell 1980). We show that this symbolic-relational discontinuity pervades nearly every domain of cognition and runs much deeper than even the spectacular scaffolding provided by language or culture alone can explain. We propose a representational-level specification as to where human and nonhuman animals' abilities to approximate a PSS are similar and where they differ. We conclude by suggesting that recent symbolicconnectionist models of cognition shed new light on the mechanisms that underlie the gap between human and nonhuman minds.