When two languages share some things in common, they might be related, but they also might
not be. How do we know?
When looking to see if languages are related, we look for words in the two different languages
that are similar in both meaning and sound. But we have to be careful. There are two ways
languages can end up with these pairs without being related at all.
One is borrowing. The word for chocolate is very similar across
a whole range of languages that aren't related to each other. It's similar because all those
languages borrowed the word. It originally comes from Nahuatl, an Aztec language, and
spread to others after explorers brought it back to Europe in the 1500s.
Another way languages end up with words that seem related is by pure dumb chance. Sometimes
languages just happen to hit on the same sequence of sounds for for the same thing. The Spanish
word for "to look" happens to be a lot like the Japanese word for "to look." These coincidences
happen more than you might think. After all, there are only so many sounds the human vocal
tract can make. There's bound to be some accidental overlap.
Language relationship is determined not just by finding similarities, but by finding systematic
similarities. So the English word ship is similar to the German word Schiff. But it's
not just that those two words are similar, it's that you find a p corresponding to an
f through the vocabulary.
And a d corresponding to a t
And th corresponding to d
In this way, English and German show a family resemblance.
So what about all those English words that look like French words? There are tons of
them, but you just don't find the same kind of systematic correspondences for those. Sure,
there are some but not many. Plus our words from French have to do with the domain of
high civilization. They are chocolate words. We borrowed them less than 1000 years ago,
not so long ago in the grand scheme of things.
In contrast, our Germanic words hearken back to the pre-civilized basics. Water, earth,
stone, heart, blood, birth, death.
We did not borrow these essential words from German. Rather, both German and English evolved
from a common ancestor. They are siblings, or cousins. They belong to a whole family
of descendants, including, Dutch, Swedish, Icelandic, and Gothic. By studying the sound
correspondences across these languages and applying consistent sound change rules we
can actually work out what this mystery ancestor, Proto-Germanic, looked like.
In the case of French's language family, the Romance languages, we already know what the
ancestor language looked like: it was Latin.
This method of looking for systematic correspondences allows us to peer even further back in time
and see how whole language families are related to each other. It turns out Latin and Proto-Germanic
themselves have a common ancestor. This was one of the great discoveries of the 18th century
when scholars noticed that the classical Indian language Sanskrit had consistent similarities
to Latin and Greek.
They eventually uncovered systematic correspondences among a wide range of languages that spanned
continents, working out a very ancient common ancestor known as Proto-Indo-European.
It turns out English does have a family relationship to French, but it goes much further back to
a much more distant ancestor. So on the surface French and English words for heart or hound
don't look similar at all. But it is possible to see how the reconstructed Proto-Indo-European
words, kerd and kun, could, through intermediate historical steps, end up as English and French.
How does k become h? K is a breathy sound made toward the back of the mouth. So is H.
They started being pronounced differently in those ancestor words before Proto-Germanic
split from Latin.
These kinds of correspondences are harder to find and confirm, but it helps that the
languages of the Indo-European family have a rich historical record of surviving documents,
going back through centuries of language change. Not every language family is so lucky, but
a lot can still be accomplished using the comparative method in the absence of such
a record. We can show how geographically distant languages are related to each other in families
like Uto-Aztecan in the Americas, Bantu in Africa, or Uralic, which spans from Europe
to Asia and Siberia.
Every language is a history. The comparative method, carefully applied, lets us travel
back in time further than any written record of that history.