Languages are territorial. They tend to occupy homogeneous, well bounded areas. When they do not, they lessen their chances of survival, especially if they are languages of minority groups. Reaching beyond the usual sociological causes of this phenomenon, this article searches the neurophysiological and the psychological literature for explanation of the tendency of closed, equalitarian systems with a high density of communication to move toward unilingualism. The search is guided by the questions: are bilinguals less brain-lateralized than unilinguals? Are different languages stored in different "containers" in the bilingual memory? Are the reaction times for coding and encoding slower in a second language than in a dominant language, slower in multilingual compared to unilingual settings? What are the psychological costs and benefits of bilingualism? The hypothesis that the bilingual brain is different from the unilingual brain is not supported by the literature, but some fascinating studies keep the question open. Only two sets of findings emerge to offer likely explanations of language territoriality: the findings that measure the declining level of performance in a second as compared to a first language when the complexity of the task is increased, and the findings that show multilingual communication to be less efficient, due to interferences and delayed reaction times, than the same communication in a single language.