We often say that a particular language belongs to a family of languages. French, for instance, is a member of the Romance family of languages, along with Spanish, Portuguese, Italian, and some others. All the languages in the Romance family are derived from Latin. Latin itself belongs to a larger family, the Italic family, which also includes Faliscan, Oscan, and Umbrian—all these, including Latin, are dead languages. English, however, is not derived from Latin and it therefore it doesn’t belong to the Romance family. English belongs to the West Germanic family, along with German and Frisian and Dutch. The West Germanic family belongs to the wider Germanic family, which also includes North Germanic languages, such as Icelandic, Swedish, and Danish.
Both the Romance family and the Germanic family belong to a large group called the Indo-European family; this includes the Balto-Slavic languages, such as Latvian, Lithuanian, Czech, Russian, and others; the Germanic languages; the Celtic languages, such as Gaelic and Welsh; the Romance languages; the Albanian languages, Gheg and Tosk; the Hellenic language, that is, the various dialects of Greek; the Anatolian languages, such as Luvian or Hittite; the Indo-Iranian languages, which are divided into the Iranian languages, such as Persian and Kurdish and Pashto, as well as the Indic languages, such as Hindi, Urdu, Singhalese, and Romany; and the Tocharian languages, with two dialects, Tocharian A and Tocharian B. Some of these are dead languages; Hittite, for instance, has been extinct since sometime in the Bronze Age, some three thousand years ago, but we know about it from written records written on clay tablets.
All these Indo-European languages derive from a single language, which today we call Proto-Indo-European. This language was spoken perhaps six thousand years ago; there are no written records of Proto-Indo-European, but linguists have been able to reconstruct much about it from the evidence of its descendant languages.
All this information can be shown in a kind of tree diagram, in some ways like a family tree, with Proto-Indo-European at the top node; then branches leading out to the major families (Balto-Slavic, Germanic, Celtic, Italic, Albanian, Hellenic; Anatolian, Indo-Iranian, and Tocharian). Then each of these major families divides into smaller sub-families; so Germanic divides into North German, West Germanic, and East Germanic; then West Germanic divides into English, Frisian, Dutch, and German. (The diagram I am consulting for this blog post comes at the very end of The American Heritage Dictionary of Indo-European Roots, edited by Calvert Watkins; similar diagrams can be found in many other sources.)
But languages are not people, and they don’t form family trees the way people do. People have two parents. Your family tree goes back one stage to two people, your mother and father; two stages back you have four grandparents; and so on. A human family tree is organized in generations. You are in one generation, your parents are in the generation before you, and your children will be in the generation after you. (A person can belong to two generations; in ancient Greece, for instance, it was common for an uncle to marry his niece.) Human generations usually overlap; your grandparents and parents may well be alive during your lifetime. Some animal generations, however, don’t overlap; many insects parents die before their offspring hatch. And the generations of a human family tree are made up of a succession of distinct individuals; you are distinct from your parents, and your children are distinct from you.
None of these points apply to languages. Languages change gradually, rather than in distinct generational jumps. Languages don’t have two parents. Italian is not the offspring of Latin and some other language. (Though a language may well be influenced by another language, as English has been influenced by French.) It may be convenient to say that Latin is the mother (or father?) of Italian, but in fact Latin gradually changed over time, a little bit here and a little bit there. There are no distinct generations of a language. At a certain point we might say that the accumulating changes have produced a new language, but we can’t put a finger on the day or even the year or the decade that Latin became French. The speakers of French in the year 2023 would not understand or be understood by the speakers of Latin in the year 100, but there was no specific moment when one generation of speakers suddenly became unable to understand their children.
In some circumstances, however, speakers of a language may become aware that they are no longer speaking the way earlier generations spoke. Often this realization occurs when there are written records of the earlier stages of a language. Shakespeare’s language is recognizably English, but odd to the modern ear. Students first reading Hamlet may have a lot of trouble, unless they have an annotated edition. There are now modernized editions of Shakespeare’s plays “translated” into modern English. Chaucer’s Canterbury Tales is even more different from modern English, and many modern readers are unable to make much sense of it. We say that Chaucer’s language was English, but Middle English. And Beowulf is just written in a different language, sometimes called Old English and sometimes called Anglo-Saxon. Sometimes archaic forms of a language may be preserved in oral literature, such as oral epics, or in religious rituals, and modern speakers may be aware that these special genres preserve older forms.
In situations where archaic forms of language are preserved, we can ask when the speakers of a language become aware that they are speaking a new and different language. When, for example, did people become aware that French isn’t Latin anymore? Again, we can’t pinpoint a moment, and different people may have come to this awareness at different times, but it’s probably reasonable to say that sometime in the 9th century some people began to think of their everyday spoken language as different from Latin—a different dialect or perhaps a different language. The Council of Tours in 813 directed priests to give their sermons in “rusticam romanam linguam” so they would be understood by their congregations.
In the year 842, two of Charlemagne’s grandsons, Charles the Bald and Louis the German, made a military treaty, known as the Oaths of Strasbourg, in opposition to their older brother Lothair. They swore the Oaths in public so that their respective armies could hear and understand. Charles the Bald swore in an early form of Germanic, so Louis’s soldiers would understand, and Louis the German swore in “Romance” so that Charles’s soldiers would understand.
What was “Romance”? Was it a kind of “rustic” Latin? Was it a very early form of French? Or was it something in between? Here is the “Romance” version of the Oaths (as was written down some years later; this record is not likely verbatim), along with an English translation (not by me):
Pro Deo amur et pro christian poblo et nostro commun saluament, d’ist di en auant, in quant Deus sauir et podir me dunat, si saluarai eo cist meon fradre Karlo, et in adiudha et in cadhuna cosa, si cum om per dreit son fradra saluar dift, in o quid il mi altresi fazet. Et ab Ludher nul plaid nunquam prindrai qui meon uol cist meon fradre Karle in damno sit.
For the love of God and Christiandom and our joint salvation, from this day onward, to the best of my knowledge and abilities granted by God, I shall protect my brother Charles by any means possible, as one ought to protect one’s brother, insofar as he does the same for me, and I shall never willingly enter into a pact with Lothair against the interests of my brother Charles.
I must say that this doesn’t look like Latin to me, or French, for that matter; more like a little of each. Because of the educational reforms instituted by Charlemagne, the literate class of 842 could read and write a pretty good approximation to something like classical Latin, but ordinary people often couldn’t understand it. Whoever wrote the “Romance” version of the Oaths almost certainly didn’t say to himself “I have to write this in French”, because the concept of a French language did not yet exist, but he did (perhaps) think “I can’t write this in the official Latin dialect of the schools, I have to write it in a dialect the soldiers will understand”. (The Oaths were a piece of political theatre, and the languages of the Oaths were part of that political theatre, but I will have to leave that topic aside.)
The family tree diagram of language history might suggest that there are no intermediate stages in language change — Latin just becomes French (or Italian, or Spanish, or. . .). But really, every stage of a language is transitional. If we look at the Romance version of the Oaths of Strasbourg from the perspective of classical Latin or Modern French, it looks like a transitional stage between those two languages, and so it is, but classical Latin is a transitional stage between Italic and Romance, and modern French is a transitional stage on the way towards some future version of French, which may not be called French at all.
In my next essay I will continue this discussion of language families. (If you’re interested in the development of the Romance languages, I recommend a wonderful collection of articles, Latin and the Romance Languages in the Early Middle Ages, edited by Roger Wright.)
The Art of Reading Slowly 