FIGURE 2.

Semiotic experiments using conditioned marmoset vocalizations. Does social identity matter for each of the experiments? Learning curves are expected to be steeper and shorter when the interacting animals are from the same family. (A) Mimicking experiment: Marmoset 2 will be rewarded only when able to deliver the same kind of vocalization produced by marmoset 1 (e.g., phee, in blue). (B) Soft generosity experiment: Marmoset 1 will be rewarded when marmoset 2 is rewarded after a correct vocalization. (C) Hard generosity experiment: When marmoset 2 produces the correct vocalization, only marmoset 1 will be rewarded. (D) Envy experiment: Marmoset 1 will get double reward when marmoset 2 is rewarded after a correct vocalization. (E) Collaboration experiment: The two marmosets will be rewarded only when able to deliver the same kind of vocalization in close temporal proximity. Can the task be built with several animals in tandem? (F) Competition experiment: A properly-timed vocalization by marmoset 2 will allow it to steal the reward from marmoset 1, and vice versa. How many turn-taking will be reached as a function of kinship and satiation? (G) Learning a new combination of vocalizations experiment: The marmoset will be rewarded if it manages to reproduce the artificial sequence that will be played back, composed of two different vocalizations (phee + twitter, in red). (H) Simple combinatorial symbolization. The experiment involves learning to produce a different number of phee pulses to obtain rewards of different appetitive value. Shifts in the distribution of phee pulses will be expected depending on the contingency established. (I) Complex combinatorial symbolization. The experiment involves learning to produce different combinations of call types to obtain rewards of different appetitive value (in this example, phees and twitters). Shifts in the distribution of call sequences will be expected depending on the contingency established.