On rhetorical typology: Evidence from the translation of 'path'