How Languages Evolve, Part I: The Road to Idiocracy?

In Monty Python’s Life of Brian, a pedantic centurion forces a graffitist to stay up all night correcting his Latin. A few centuries later, the vandal’s mistakes were standard grammar in Romance languages.

Anyone who speaks any language knows someone who is assiduously, perhaps annoyingly, insistent on observing the proper rules of grammar, pronunciation, et cetera. Language snobbery is nothing new: thousands of years ago, pedantic Romans wrote entire books indexing common errores in how the hoi polloi spoke Latin.

Much to the chagrin of their punctilious peers, the uneducated rabble won: today, linguists look to these lists of Roman slang terms, lazy grammar, and mispronunciations as the germ of modern Romance languages. These languages are now often considered among the most beautiful and sophisticated in the world, but they still bear traces of their humble origins. Italians, for example, might be surprised to learn that their words for “head” (testa) and “house” (casa) come from Vulgar Latin slang for jars and peasant shacks, respectively.

After centuries as a seldom-recorded peasant language, English made its comeback in Geoffrey Chaucer’s Canterbury Tales. But between 1066 and the 1390s, the language had changed completely

English underwent a similar transformation. Like Latin, Old English had an intricate system of declensions, with nouns, adjectives, and articles taking on different suffixes depending on their roles in a sentence. However, by the 14th Century, when Geoffrey Chaucer published the most prominent work in Middle English, the Canterbury Tales, virtually all of this had been dropped. Today, grammatical case survives only in pronouns, e.g., he, him, and his.

With each passing year, English speakers jettison strong (i.e., irregular) verb conjugations in favour of regular ones. When our great-grandparents were kids, they clomb trees, and if they fell, their injuries swoll up. Someone might even have holpen them back up. These strong conjugations are deeply rooted in Germanic languages, and they were traditionally the correct ones. But if you used them today, people would think you had hit your head.

Given the common trend toward simplification across most of the major languages of Europe, as well as others like Sanskrit and Persian, many early linguists suspected it was something of a universal law. On the other hand, simple logic poked a hole in that theory: language has been around at least for tens of thousands of years, so if humans were on a perpetual slippery slope of grammatical degradation, by now we would be reduced to a lexicon of a few dozen grunts. So, what was happening?

From Antiquity until 1799, the Ancient Egyptian language, and with it 3,000 years of history, were lost. That changed when a scholar travelling with Napoleon’s invasion of Egypt discovered the Rosetta Stone, a bilingual stele in Ancient Egyptian and Greek.

In 1970, Professor Carleton T. Hodge published a study of the Ancient Egyptian language in which he argued that, over millennia of history, its grammar started complex, then simplified, then became complex again. The scope of Hodge’s study was limited, but the idea of oscillation between simplicity and complexity was catchy.

At the core of Hodge’s schema was the classifying of languages as synthetic versus analytic. Synthetic languages are ones like Old English or Latin, with a complex system of grammatical cases, using suffixes or other morphemes to pack a lot of information into a word about its grammatical role. By contrast, in analytic languages like Spanish or English, words do not change that much and you interpret the action based on syntax.

To illustrate with a popular example, in Latin, canis hominem mordet and hominem canis mordet both mean “the dog bites the man,” because canis, dog, is in the nominative, while hominem is in the accusative (you can also move the verb around). In English, however, to say “the dog bites the man” or “the man bites the dog” are not quite the same.

Some well-to-do residents of Pompeii commissioned a mosaic to warn their neighbours to “cave canem,” beware of the dog. Perhaps if it read “cave canis,” the dog would have known to beware of Vesuvius.

In Ancient Egypt as in Western Europe, the shift from their synthetic to analytic is a key feature of languages “simplifying,” eroding tricky grammar to make them easier to learn. Other features that can disappear are grammatical gender, or the use of different vocabulary to address people according to a speaker’s relationship to them. Phonology can also change. Most of the errors recorded by the Roman grammarian Probus were things like mensa non mesa for a table, oculus non oclus for an eye, or aurum non orum for gold.

From a Western European perspective, it is easy to look at a lot of the languages with which we are most familiar and come to the same conclusion as the early linguists. But venture farther from home, and we find that a lot of languages are still really eccentric. Djirbal, an Aboriginal Australian language, has four genders (including one for women, fire, and dangerous things); Tsez, a Caucasian language of Georgia, has 42 grammatical cases; Javanese has three sets of vocabulary to use for people of different social standing. And it is not just minor languages: Russian still has a complex system of declensions, and Tamil, spoken by 75 million Indians, is pretty similar now to how we find it in the earliest records from 2,000 years ago.

A temple in Tamil Nadu, in South India. If the ancient Tamils who built it came back today, they would have little trouble conversing with their modern descendants.

So what makes a language smooth out its rough edges? The key is simple: people learning it badly. A language of a conquering empire or successful merchants is very useful, so a lot of people will want to learn it. But adults who are set in their linguistic ways will inevitably butcher it and leave out the confusing grammatical niceties.

In this way, Latin became a victim of its own success: so many people learned it that they eventually garbled it so badly that it was no longer Latin. English was on the other end of conquest, with the same result. After generations of crude Vikings and contemptuous Norman conquerors ignoring the three genders and four cases of Old English, they eventually disappeared.

Outside of Europe, the transformation of Old Persian to modern Farsi was the same as that from Latin to Italian. Chinese is a very different language, but the trend was the same. Old Chinese words could, for example, end in all kinds of consonant sounds, whereas Mandarin ones must end in a vowel or an N. Many modern dialects have up to nine tones, whereas Mandarin, the official register everyone is forced to learn, has just four.

I believe it

If you showed a Roman grammarian a copy of Dante, the experience for him would be a bit like us watching the movie Idiocracy, a 2006 satire about a future in which IQs have plummeted and all the stupid trends we lament today have plunged to ridiculous extremes. How and why many major languages become radically simpler over time is well established. For those of us who care about proper language, it is easy to bemoan this as the dumbing down of the core feature of human society (and even our cognition).

But as we observed above, after a few thousand generations of grammatical carelessness, we have not been reduced to grunting. Evidently, it is not just a perpetual downward slope. There must be something more to how languages evolve.