Takeshi Konno, Junya Morita and Takashi Hashimoto talk about the integrative approach to the emergence of symbolic communication. The talk included details of a hybrid model of cognition for communication that involved a context-free grammar to handle denotation and a neural network to handle connotation. However, the most interesting work was an analysis of the different brain areas used at different stages of the evolution of a communication system. They used an experimental paradigm similar to Galantucci (2005) where two human players played a coordination game using computer terminals. On the screen, players were placed in one of four coloured room, but unable to see their partner in another room. The aim was to move once (or not move) to end up in the same room as your partner. Players were allowed to communicate once before moving using a sequence of abstract shapes. Players could send a sequence of two abstract shapes to their partner. The idea was to set up a communication system whereby, for instance, a square followed by a circle might mean 'move into the green room'.
Konno et al. observe an evolution in the communication system: First, the establishment of common ground (what shapes meant what colour). Next, a symbolic system emerged with a semantics and a syntax. At this stage, players were sending messages simultaneously. Finally, role division (pragmatics) emerged to handle situations where the suggestion of a move by one player was impossible to reach in a single move by the other. Therefore, one player would make a suggestion, and the second player would either modify the suggestion or confirm the suggestion by sending back the same signal. Konno et al. note the emergence of the possibility of the same signal to meaning different things.
Interestingly, a recent experiment used EEG scans of participants' brain activity as they played. Konno et al. observed activity in Wernicke's area at the semantic and syntactic stage, but also increased activation during the pragmatic stage of the evolution of the system in Broca's area, the right frontal cortex and the medial frontal area. Although this finding was not covered in a lot of detail, and the implications were not fully fleshed out, it's an intriguing result, and may usher in a new series of brain-scanning versions of other communication game paradigms. Do participants at a later stage of an iterated learning paradigm used different brain areas to those in the initial stages of the evolution of the language?
Galantucci, B. (2005). An experimental study of the emergence of human communication systems. Cognitive Science, 29(5), 737-767.