Towards the automatic merging of language resources