A study published in Cerebral Cortex may be of interest to those who followed the recent harmonic theory bruha. The junior research group at the Max Planck Institute for Human Cognitive and Brain Sciences measured the electric response potentials of listeners' brains when harmonic expectation was violated, combined with another factor of syntactic or semantic language violation. They found that both types of language violation affected different locations of brain waves, a strong argument that harmonic tension-resolution is a grammar, but also has associated meanings. The study does not reveal whether these meanings are universal, merely that individuals do interpret harmonies beyond a syntactic level. The authors presented preliminary research of this at the Neurosciences and Music conference I attended in 2005. And Dave Munger has written about this line of research.
Steinbeis, N. and S. Koelsch. "Shared neural resources between music and language indicate semantic processing of musical tension-resolution patterns." Cerebral Cortex 2007 (eprint).