contextual animations available in all talk modes.

[jwgcurrie]

From all talk modes (Text, Voice and LLM) the operator should be able to toggle semantically aware body animations (NAOQI already supports this).

The default behaviour should be defined in a config file.