The normal speed of language change

A. When linguists studied the northern Australian language Ngan’gityemerri in 1930, they found a language with sentences similar to the following:

1930:

Dudu dam, dam dudu, kinji dinj parl.

Track poke poke track here he-sat camp

“He poked along, tracking it along here to where it made its camp.”

1990:

Damdudu, damdudu, kinyi dinyparl.

Poke-track poke-track here he-sat-camp

Notice that in 1930 the speaker could give the order of dudu and dam (track and poke) in either order; they were separate words. But when linguists returned to the language in 1990, its entire grammar had changed. Now, dudu had grammaticalized into a prefix of dam, such that there was one word dududam, meaning roughly “pokingly tracked.” This had happened with all verbs in the language. Ngan’gityemerri had moved along the path toward becoming a language like Yupik Eskimo, which packs a sentence’s worth of meaning into one word. (Recall the Yupik Eskimo word for “He had not yet said again that he was going to hunt reindeer”: Tuntussuqatarniksaitengqiggtuq.)

B. But English has changed more slowly in the time after the Middle Ages. Shakespeare speaking 500 years ago would have sounded strange to us, but we could converse with him. However, Shakespeare would have found an Old English speaker from 500 years earlier almost as incomprehensible as a German.

C. This phenomenon can be explained by the fact that when a language is written and standardized and literacy becomes widespread, the written form comes to be seen as “The Language,” and it affects people’s speaking habits enough that the language changes more slowly than it would naturally. Standardized languages are “frozen in aspic,” as it were.

D. A contrast: we can easily read presidential addresses from the late 18th century, but a speaker of Saramaccan Creole in Suriname would find the speech of a chief in 1789 extremely peculiar. For example, at that time, the way to say “not” was no, but today, it is just a.