By John McWhorter, Ph.D., Columbia University
Based on the similarities between languages, linguists have now realized that some groups of languages were related and descended from a parent language. For example, we know all Indo-European languages descended from a parent language called the Proto-Indo-European language. There are no written records of it, but we know what it looked like based on other linguistic methods.
How to Reconstruct Proto-Indo-European Language?
There is a whole body of research about the grammar, vocabulary, and even pronunciation of this language. One might think there are still speakers of this language around that have made this whole work possible. But everything we know about this Proto-Indo-European language is based on deduction. By looking at the descendant Indo-European languages, linguists deduce what the parent language would have looked like.
To make such deductions, one has to have a thorough knowledge of how languages change over time, especially in terms of pronunciation. Languages change constantly. Consonants and vowels weaken over time. By looking at these change trends, we can deduce what kind of vocabulary the parent language of Indo-European languages had.
Learn more about Indo-European and the prehistory of English.
The studies conducted on the Proto-Indo-European language over the past 150 years have been so extensive that scholars have been able to build a dictionary of this language. It is also possible to make some sentences in this language. It might seem that getting a clear idea of what the language actually looked like is not possible. There are many issues that make this process difficult. But we have a pretty clear picture of it at the moment.
This is a transcript from the video series The Story of Human Language. Watch it now, on Wondrium.
A Case of Reconstruction in Proto-Indo-European
Let’s take a look at how these deductions are made to reconstruct this long-gone language. As an example, we use an Armenian word to trace it back to the original Proto-Indo-European word. This word is nu, which means ‘sister-in-law’ in Armenian. Over time, and due to semantic drift, it has changed its meaning to ‘bride’.
Interestingly, the cognates of nu in many other Indo-European languages also mean ‘sister-in-law’. In Sanskrit, it was snuşā́; in Russian, it is snokhá; in Old English it was snoru. In a slightly different and justifiable form, other languages have similar words for ‘sister-in-law’. In Latin, it was nurus; in Greek, it is nuós; in Albanian, it is nuse. Using these cognates, we can work out the original word for ‘sister-in-law’ in the parent language. We know that changes are constant in languages, so we are convinced that the original word could not have been one of these words exactly. It has to have changed over time.
But we can reconstruct the word based on these words, one letter at a time. The first letter could be n or sn since some of these words start with n, and some start with sn. The logical choice would be sn because, as a general rule, consonants weaken over time. Consonant erosion has been more prevalent than new ones coming in.
Then we would assume that the word started with sn, and then the s dropped over time. Why is that and not the other way round? Why didn’t the n drop? Because for this to be true, it should have been an n, then came an s, and then the s dropped. This would be unreasonable. Suppose an imaginary word like neeb turned into sneeb. It is counter-intuitive in terms of language change rules. So, the first letters of the original word are assumed to be sn.
Then, we should determine what the first vowel probably was. Based on the sample words that we have in those languages, after the sn comes either u or o. like snuşā́ in Sanskrit, nurus and nuós in Latin, snoru in Old English, and snokhá in Russian. Most of these words have u after sn. We can suppose that based on this majority, the original word’s first vowel was u.
So we have s-n-u. Now we proceed to determine the next letter. Again, the majority of these words have an s after the first vowel. So, the word is s-n-u-s, so far.
This is the root of the word. Now we want an ending. Since ‘sister-in-law’ is a word that has a feminine reference, we assume a feminine ending is needed here. In most Indo-European languages, -a is a feminine ending. But the words for ‘sister-in-law’ in Latin and Greek have a masculine ending. We might be inclined to think that was an accident. There are two explanations for that. One is to assume that these words had feminine endings, and then over time, masculine endings replaced them. The other one is the other way round; it started with a masculine ending, and some languages changed it to feminine. Which one sounds more plausible and reasonable? The second one. So we assume that the original ending of our hypothetical word was –os.
Thus, through the process of comparative reconstruction, we reconstructed a word in the proto-Indo-European language. The original word for sister-in-law in the Proto-Indo-European language was supposedly snusos.
Learn more about how to identify a language family.
Common Questions about Reconstructing the Proto-Indo-European Language
Sanskrit is an Indo-European language that descended from the Proto-Indo-European language. It is related to all other Indo-European languages through this parent language.
Through the process of comparative reconstruction, linguists have reconstructed Proto-Indo-European. It is the most accurate and complete reconstructed version of the language.
All Indo-European languages are related to Proto-Indo-European. They all have features that are shared among all of these languages and originally belonged to their ancestor.
