That, according to MIT, suggests that the English of non-native speakers could be mapped in a similarly uniform way, which could pave the way for it smarter grammar correction software.