Yesterday, I got a note from lisashafter in which she asked, "How the heck did Earth's languages end up so completely different? I know that languages descend from roots, obviously, but what are the origins of those roots [...] I've got the Biblical story of the Tower of Babel, but I was wondering if modern linguistics had any theories along these lines." This made me very happy because linguistics is pretty much my favorite thing to talk about, especially when the person I'm talking to is at least marginally interested. So while I was writing a reply, I thought, "hey, I haven't blogged in forever (sorry, guys. I'll probably explain at least some of the reasons for that in later posts), and I could say a whole lot more about the subject than what is needed to answer this question." Thus I've decided to at least partially address the topic of one of my two linguistics classes for this semester: Language Change.

There's no agreed-upon theory about how ALL the languages connect. We can get pretty far back, though. We know, for example, that languages including English, German, Swedish, and Icelandic are all descended from one language, so we say these are members of the Germanic sub-family. Germanic is similarly related to sub-families including Italo-Celtic (which includes e.g. Latin and Celtic languages), Balto-Slavic (e.g. Latvian, Czech, Russian), and Indo-Iranian (e.g. Farsi, Romani, most languages from India). These sub-families, along with others, make up the Indo-European family, which is the most well-studied and geographically largest family.



We can even do a pretty good job of figuring out how some of the root languages (linguists call them "mother" languages) actually sounded. Based on its descendants, we can reconstruct large parts of Proto-Indo-European, the mother of the Indo-European family. ("Proto" at the front of a language name means it's been reconstructed, as opposed to us having direct evidence as we do for something like Latin.)

Indo-European is pretty much as far back as anyone can agree on for sure. At this level, there are somewhere between 250 to 300 different families, according to my textbook¹. Some families are diverse enough to give Indo-European a run for its money, others just have one or two languages. People have tried to connect these families in various ways into super-families, but it gets really tricky because a lot of the time they have to start out with reconstructed languages to even begin comparing, and it's hard to tell sometimes which words were borrowed from other families, and many other reasons. Some people even claim that all languages are connected in a "Proto-World" family, but that's even more controversial. Personally, I think the Tower of Babel story is true and thus we'd never be able to reasonably get down to less than several dozen families.

Of course if you're starting from the various root, or "mother," languages, there are all kinds of ways in which they can change and split, and many reasons for them doing so. In class, we've been talking about sound changes. Sounds may change for any number of reasons, including making words easier to pronounce or understand, or because other sound changes have left some sound features unnecessary, or have left gaps that need to be filled or have gotten too close to sounding like other sounds.

Those last two reasons are called "chain shifts." A chain shift that happens when sounds change to fill gaps left by other shifts is called a "pull chain," and a chain shift that happens because one sound gets too close to another is called a "push chain." One famous chain shift happened between Middle English and Modern English. Known as the Great Vowel shift, it happened when the vowels started moving upward (that is, vowels pronounced with the mouth more open, such as "oh" and "ah," began to be pronounced with a more closed mouth), and the highest vowels, not having anywhere else to go became diphthongs. Overall, it went something like this:

This also does a lot to explain why vowels in our writing system are so weird. Many words were originally pronounced how they're still spelled, it's just that being stubborn English speakers we never got around to changing the spellings to reflect the new pronunciations.

1 Campbell, Lyle. Historical Linguistics: An Introduction, 2nd edition. (2004)

Word of the Week:
spirantization (n) this is a name for a common type of sound change in which a stop or affricate becomes a fricative. (Stops are consonants that are made by stopping the airflow entirely, such as p, b, t, d, k, and g in English. Fricatives are consonants that are made by restricting the airflow very tightly without stopping it, such as English f, th, s, z, sh, h. Affricates are like a combination of a stop followed immediately by a fricative, such as the sounds English usually writes as "ch" and "j.")

This is process was involved in another, older chain shift that happened somewhere between Proto-Indo-European and Proto-Germanic, called Grimm's Law. (Yes, as in the Brothers Grimm. Jacob Grimm was an awesome linguist in addition to being an awesome folktale collector!) We're not sure whether the changes Grimm observed resulted from a pull chain or a push chain, but since the pull chain scenario is what we talked about in class, that's the scenario I'll explain here.

In Proto-Indo-European, there were voiceless stops: p, t, and k; voiced stops: b, d, g; and voiced aspirated stops (which means they're pronounced with an extra puff of air at the end): bh, dh, gh. After awhile, spirantization affected the voiceless stops, and they became voiceless fricatives: f, th, and the ch used in German (linguists write voiceless "th" as /θ/ and German "ch" as /x/).
The spirantization meant that with the remaining types of stops (voiced b, d, g and voiced aspirated bh, dh, gh), it didn't matter whether they were voiced or not. So people stopped voicing the unaspirated stops.
But now the aspiration on the voiced aspirated stops wasn't needed anymore, because the voicing was different, so they stopped with the aspiration, and ended up with voiceless fricatives f, θ, x; voiceless stops p, t, g; and unaspirated voiced stops b, d, g.

By this time the Germanic people and the other Indo-Europeans would have found communication pretty much impossible, because they had so many of the same sounds, but they were all in the wrong places! Knowing about this change can lead to new insights when thinking about words that come from other Indo-European languages. For example, take pater, which we're familiar with because of words like paternal, change the "p" to "f" and the "t" to "th," and you get father! Guess which words in Latin and English come from the same Indo-European word for 'father'?