One perennial problem in linguistics is how to decide whether a language is a language or dialect. In the fascinating book, Speak: A Short History of Languages, which I read recently, Tore Janson argues that a language can be considered a language when those who speak it decide that it is one, and they give it a name. This often happens when a language acquires a standard written form, and/or becomes the language of a state of other political entity.

He gives the example of Italian in the Chapter Did Dante Write in Italian?: Dante is said to be one of the first authors to write in Italian rather than Latin, however he didn’t see Latin and Italian as separate languages, but just different forms of the same language. Dante refers to Classical Latin as Grammatica (Grammar), the colloquial language of Italy as Latium vulgare (popular/vulgar Latin), and he calls the language he wrote in Latino (Latin). Italian only started to be called italiano or lingua italiana not long after Dante’s death.

Janson gives the another example of the Khoisan languages of South Africa, which have many different names. Speakers of these languages, when asked, might use the name of their area, tribe or some other name for their language – but generally don’t have a particular name for their form of speech. Several hundred names have been collected by linguists, and as a result nobody is quite sure how many Khoisan languages there are and how they are related to one another. None of these languages have a standard written form, and speakers rarely, if ever, write them.