My question is closely tied to a previous thread, but I would like to stick to the linguistic field. Let's suppose we are searching for classifying an unknown or poorly known language that shares many features, but not all with the language A which has been cleared up. Can we decide whether this is an A dialect or a B language ? In many cases studying phonological, grammatical, lexical similarities and discrepancies are more than enough to make up his mind. Undoubtedly Catalan is not a dialect of Spanish (or vice-versa ) and Occitan is'nt either a dialect of Catalan. But in case data is poor, native speakers are few, is it possible to settle the question ? Where does the linguistic boundary goes through ( I don't mean a geographical one ) ? We can perhaps hang on to the mutual intelligibility ; people who understand each other , whatever the dialectal forms, use the same language. If we are in agreement with this definition, we can say, for example, that Moroccan, Egyptian and so on are really different languages, by the fact that an Egyptian speaker cannot make out the Moroccans' language ( Common, written Arabic is rightly called M.S.A, Modern Standard Arabic ).