Research into animal theory of mind can benefit from the renewed interest in mental state attribution in infants. A series of recent experiments contend that false belief attribution is already present in the second year of life (1-3), much earlier than the 4 years of age previously thought (4). These studies resort to new, spontaneous response, methods such as Violation of Expectation and Anticipatory Looking, which avoid verbal interaction between experimenter and participant, thus making the task simpler. Therefore, they can also be used in non-verbal creatures. However, it is not obvious that what they demonstrate is really false-belief attribution by infants: these methods rely on first creating an expectation through repetition, but the content of this expectation might have to do with keeping track of the agent's views, the places and the object that gets moved (in the change of location task), rather than attributing a propositional content, much less a false one to the agent. If success in the false belief spontaneous response task relies on a tracking mechanism of this kind, it is reasonable to expect that some non-human animals are also able to pass these simpler tasks: those that are likely to understand intentions, and keep track of objects, and gaze direction, such as dophins (5-7). We are currently studying whether bottlenose dolphins (Tursiops truncatus) are able to pass these new false belief tasks using a variation of the anticipatory looking paradigm: an anticipatory pointing task. We will present our results, and if positive, we will discuss as well whether and why this epistemic tracking doesn't amount to false belief attribution, and what else would be required to bridge the gap.