Language families

What is a language family?
Most languages belong to language families. A language family is a group of related languages that developed from a common historic ancestor, referred to as protolanguage (proto- means 'early' in Greek). The ancestral language is usually not known directly, but it is possible to discover many of its features by applying the comparative method that can demonstrate the family status of many languages.

Sometimes a protolanguage can be identified with a historically known language. Thus, provincial dialects of Vulgar Latin gave rise to the modern Romance languages, so the *Proto-Romance language is more or less identical with Latin. Similarly, Old Norse was the ancestral of Norwegian, Swedish, Danish and Icelandic. Sanskrit was the protolanguage of many of the languages of the Indian subcontinent, such as Bengali, Hindi, Marathi, and Urdu.

Further back in time, all these ancestral languages descended, in turn, from one common ancestor. We call this ancestor *Proto-Indo-European (PIE).

Language families can be subdivided into smaller units called branches.

How do linguists establish relationships among languages?
In some cases, it is relatively easy to establish relationships among languages. Let's look at the Romance languages. We know that Italian, for instance, is a descendant of Latin, a language that was spoken in Italy two thousand years ago, and one which left a great number of written documents. The Roman conquest helped spread Latin throughout Europe where it eventually developed into regional dialects. When the Roman Empire broke up, these regional dialects evolved into the modern Romance languages that we know today: French, Italian, Portuguese, Spanish, and others. These languages form the Romance branch of the Indo-European language family. By looking at the word for 'water' in three Romance languages, one can easily see the similarities.

Italian - aqua
Spanish - agua
Portuguese - agua

What if the ancestral language left no records?
The case with Romance languages is unusually easy because their common ancestor — Latin — left many written documents. In most cases, however, the ancestral language was not written. As a result, linguists look at similarities among the modern descendants to establish common origins. Take a look at these examples:

English - water
German - wasser
Danish - vand

Russian - voda
Polish - woda
Czech - voda

It is clear that the word for 'water' looks very similar within each group, but not so similar across groups. Languages in the first group belong to the Germanic branch of the Indo-European language family. Languages in the second group belong to the Slavic branch. Although there are no written records of the ancestral *Proto-Germanic or *Proto-Slavic languages, we have to assume that these two ancestral languages must have existed, just like Latin did.

Where do these mystery languages belong?
Here is the word for 'water' in three more languages. Do you think these languages belong to any of the three branches above?

Latvian - udens
Albanian - uje
Basque - ur

As it turns out, Latvian belongs to the Baltic branch of the Indo-European language family, Albanian has no close relatives and does not belong to any branch within the Indo-European language family, and Basque does not belong to any language family at all. In fact, it is a language isolate, i.e., a language that cannot be reliably assigned to any family.

What if there are no records, and we know little about the languages?
In many parts of the world, there are no written records, and we don't know enough about the languages themselves. Consequently, we have to resort to grouping languages on the basis of geography. This is the case with many of the aboriginal languages of Australia, the native Indian languages of the Americas, the tribal languages of Africa, and countless other languages all over the world.

How many language families are there?
According to Ethnologue, there are 218 language families in the world. This figure is probably an overestimate because of our limited knowledge about the languages of the most linguistically diverse areas of the world. For example, the Austronesian family (Pacific area) has 1,262 languages, many of them little known. So the actual number of families, once these languages are studied and relationships among them are established, will probably be smaller.