Practically every account of the relationships between pidgins and creoles has suggested that some such ‘expansion’ must take place, but no account has previously suggested how this might be done. It is obvious that the process must consist of internalizing linguistic rules for which there is no evidence in terms of linguistic outputs. If such rules are not induced from primary data, they must be derived directly from the human 'faculté de language', which must in consequence contain some kind of analog for the instructions.
- Bickerton (1977), a pag.64