Resurrecting a Dead Language- Part 1
So I’ve written in the past about the evolution of language and how linguistics attempts to make sense of all this. What I’d like to do now is to explain how a dead language- a language no one speaks today (and hasn’t for the past 3,000 years!); a language that was not written down or preserved for us in any form- can be resurrected from across the lost millennia. The process itself is fascinating, like a mystery where every clue contributes to the ultimate answer. It also provides insight into the migration patterns of these ancient peoples and their descendants. But more than that, once those descendants have been identified and the mother language has been reconstructed, its culture, and even some of its mythology, can be retrieved from the mists of time. To me, this is the most fascinating aspect of historical linguistics, the ability peer deep into the past and catch a glimpse of an ancient people whose only legacy is in the daughter languages and cultures they left behind.
Obviously, this is not going to a short post (yeah, surprising), so I’ll break it up into 3 (still not short) parts. This first one will show the process of reconstructing a dead language, using the example of the Romance languages and their parent language of Latin. We’ll also look at how linguists categorize speech sounds and their evolution. Then we’ll see what happens when all of this is applied to Proto-Indo-European (PIE, since I am too lazy to keep typing that out), the mother tongue of many of the world’s languages, including Hindi, Persian, English, Spanish, Russian, and Greek. The second post will then use the results of this PIE reconstruction (and other evidence) to tell us about the culture and history of that ancient people. Obviously, this will be somewhat speculative, but the assumptions are on mostly solid ground and seem to fit the historical record regarding the migration pattern of the original PIE peoples from their homeland in the Pontic Steppe between the Black Sea and the Caspian Sea into most of Europe and parts of Asia. It will also then show how their mythologies (and sayings and expressions) can be reconstructed. The final post will then put all this together to paint a vivid picture of these people, and in particular, what beliefs and stories they found so important and sacred that they survived into most of their daughter cultures and languages. Most interesting of all seems to be a powerful myth they had concerning a hero slaying a serpent or dragon. We’ll look at that myth in more detail and see what it might point to.
A Case Study: Reconstructing a mother tongue
So how do you reconstruct a language when it was never written down and for which there are now no speakers. To see how this can be done, let’s look at the example of Latin and its main daughter languages: Italian, French, Spanish, and Portuguese. Let’s imagine that we didn’t know that the languages were related, or that Latin had been the parent language. So you are an Italian speaker and know how to count to 10.Uno. Due. Tre. Quattro. Cinque. Sei. Sete. Otto. Nove. Dieci.
One day you meet a Spanish speaker who also speaks Italian. As you talk, he tells you about something he noticed. The Spanish numbers, 1-10, sound very similar to the Italian numbers.
Uno, Dos, Tres, Quatro, Cinco, Seis, Siete, Ocho, Nueve, Diez.
There’s no way that can be coincidence. And other words are also similar. Stato, Estado. Scuole, Escuela. Nazione, Nación. He says “cabra” for goat, you say “capra”. He says “gato” for cat, you say “gatto”. Verbs, too, seem to follow the same conjugation rules.
‘To speak’, “parlare”, conjugates into the following forms for the following speakers: "I", "you", "he/she/it", "we", "they/you-plural".
Parlo, parli, parla, parliamo, parlano.
Spanish has similar conjugation rules for verbs. ‘To speak’, in Spanish, is “hablar”. It conjugates this way:
Hablo, hablas, habla, hablamos, hablan.
The similarities concerning verb forms are telling. Both have an “-o” ending for the “I” form of the verb, “-iamo” and “-iamos” for the “we” form of the verb and “-ano” and “-an” for the “they/you-plural” form of the verb. This cannot be chance.
At some point, you meet a Portuguese person who speaks Italian, and after talking with him you notice more similarities.
Um, Dois, Tres, Quatro, Cinco, Seis, Sete, Oito, Nove, Dez.
Estado, Escolar, nação, cabra, gato.
Falar (“to speak”) conjugates to:
Falo, falas, fala, falamo, falam.
You notice the same pattern, though Portuguese is definitely more similar to Spanish than Italian. This can be done with French and Romanian too, as well as with other languages like Neapolitan or Catalan.
So you, like a reasonable person, looking at the location of the countries these languages are spoken and how close they are, finally conclude that these languages must have descended from an original parent tongue.
Now, is that the limit to what you can determine, that they are all descended from a common ancestor? Can we go further and figure out what that language looked like by reverse-engineering the daughter languages? Let’s look at one example, the word for “goat”, and see if that is possible. Here is ‘goat’ in 4 of the proposed child languages (we’ll call them Romance, even though we are supposed to not yet know they came from Roman Latin. We'll call the parent tongue we are reconstructing proto-Romance.) We'll also include the phonetic pronunciation underneath, set off with the “/” symbol to make it clear that this is for pronunciation purposes only and not spelling.
| Italian | Spanish | French | Portuguese |
| Capra | Cabra | chèvre | cabra |
| /kapra/ | /kabra/ | /shevre/ | /kabra/ |
So back to the question, then. Which was it in the original, “b” or “p”? This is where deduction plays a role. Notice the sounds around the plosive consonant (“b” and “p”). There’s the “a” sound and the “r” sound. Make both of those sounds, “ahhhh” and “errr”, with your hand on your throat. Were they voiced? Yes, your vocal cords vibrated for both of them. We see this, then, about both “capra” and “cabra”.
Capra has a voiced-unvoiced-voiced sequence.
Cabra has a voiced- voiced-voiced sequence.
So now we have these two possibilities for the original word, that it either was originally voiced-voiced-voiced (“cabra”) or voiced-unvoiced-voiced (“capra”). Keep in mind that people hear and speak quickly. Sounds run together when made and heard. Because the surrounding sounds (“a” and “r”) were voiced, it makes sense that it wouldn’t take much for a middle unvoiced plosive stop (a “p” sound) to eventually become a voiced plosive stop (a “b” sound). It makes much less sense that all three sounds would originally have been voiced (so that the whole sequence was an uninterrupted voicing of “abr”) and then gradually the voiced plosive sound (the “b”) would turn into an unvoiced plosive stop (“p”) as the speakers stopped vibrating their vocal chords for that one “b”. It’s easier to maintain the vibration, the voicing, through the entire sequence of sounds rather than to interrupt it. Try it. Say “cabra” with your hand over your throat, and notice that the throat vibrates through almost the entire word, in particular the “abr” sound in the middle. Now, do that with “capra”, making sure to emphasize the “p” sound in the middle. Notice the stop? Do both again. Which is smoother and ‘easier’? Which requires no changes in what the throat is doing? Which requires the throat to start and stop and start vibrating? Clearly, “cabra” requires less ‘effort’. So does it make sense that it would start out easy (“cabra”) and then get slightly more complicated with the change to an unvoiced “p” in the middle of the word? Or would it be the other way around- start out a little complicated and then, over time, become more uniform and a bit easier?
Now, none of this doesn’t explain why, exactly, these changes occur. It’s not a conscious change. No one says “Boy, voicing this plosive stop would sure make this word easier to say!” So while a full explanation is not completely known, part of one lies in the way we learn languages from our parents and peers. When we hear the way words are pronounced, they are not always as clear as we think. Nor are they always produced clearly either, with their stops, plosives, nasals, sibilants, fricatives, and glide sounds enunciated clearly. (Don’t worry if you don’t know what those exactly are. You get the point.)
As an example, ask your friends whether the proper word is “supposedly” or “supposably”. The latter is not actually the word, but the “b” and “d” sounds are very similar and people don’t always pronounce it clearly. Thus, others don’t hear it correctly. Our mind hears the word and does its best to link that sound (in this case a "d") to a sound we know and recognize. Anyone who has tried to pronounce a word in another language knows how it feels to keep trying to say the word correctly, only to have the speaker say “no,” and then repeat the word clearly and slowly again. We then repeat what we think we heard, only to fail again. Similarly, on the phone, we may be spelling something and need to say “D, as in David” because they heard a “b”. So, going back to the example, they may hear “supposedly” not enunciated clearly, or there is background noise and their brain pieces together the word “supposably”.
As another example, ask your friends to say “athlete”: Do they say “athlete” or “ath-uh-lete”? In this case, the “thlee” sound is kind of hard to make and has to be said precisely. The tongue begins between the teeth with air blowing over it, making the an unvoiced “th” sound, and then suddenly the tip goes to the roof of the mouth just behind the palate, just as the mouth widens horizontally, to then flick down to make the “lee” sound. (Try, again, paying attention to what your mouth, lips and tongue are doing.) People don’t always make that sound correctly. And even when they do, sometimes they want to make clear the difference between the “th” and the “lee” sound, so they will- subconsciously, usually- insert the “uh” sound between the “ath” and “lete”, by voicing the air flow as they switch mouth positions from the “th” to the “lee” sounds. Those changes then get noticed by the hearer, repeated and thus passed on, and the language changes.
So then, getting back the the “cabra”/”capra” question, the likelihood is that the original word was “capra” and gradually the consonant became voiced by the majority of speakers over time, resulting in the Spanish and Portuguese way of saying the word, while Italian kept the original. Ok. All well and good, but what the heck is going on with French?
To answer that question, let's look at the whole "p" vs "b" situation and our conclusion that the sound was originally a "p" and then became a "b" due to the voiced nature of the surrounding sound elements (or phonemes). In linguistics, this kind of a change is called "directionality", and refers to the direction sound changes go. So, in the above example, the "p" changing to "b" directionality would be written as p>b. And once we have established that "p" is the original sound, it would be written with a * symbol: *p>b. This means that original "p" sounds become "b" sounds.
p>b is not the only directional change there is. There are countless directional changes cataloged. Not all of them necessarily have to make logical sense (as it did when we deduced p>b) but- and this a key point: These changes just have to be consistent in the daughter language. They don’t have to be understood or explainable. They just have to be consistent. In our p>b example, then, if in specific constructions where you have two voiced elements surrounding an unvoiced stop ("a" and "r" surrounding the "p" sound), you must always be able to observe that in that daughter language, the unvoiced stop will always change to a voiced one (p>b).
Well, how can we know this for certain with just one word? We need a larger sampling of words to analyze. Words with the same meaning from languages that are suspected of being related are called cognates. So we need a larger sample of cognates to see what is going on. Let's enlarge our sample. This time we’ll include columns for the English meaning of the word, as well as for the actual Latin word we are trying to reconstruct, so we can check it. In cases were the meaning of the suspected word is slightly different than the others, we'll mention what the word means in that language to show how it still conceptually comes from the original parent word.
| Italian | Spanish | French | Portuguese | Latin | English |
| capra | cabra | chèvre | cabra | capra | 'goat' |
| /kapra/ | /kabra/ | /shevre/ | /kabra/ | ||
| caro | caro | cher | caro | caru | 'dear' |
| /karo/ | /karo/ | /sher/ | /karu/ | ||
| capo | cabo | chef | cabo | caput | 'head, top' |
| /kapo/ | /kabo/ | /shef/ | /kabu/ | ||
| 'main, chief' | ''extremity' | 'main, chief' | 'extremity' | ||
| carne | carne | chair | carne | caro/carn | 'meat, flesh' |
| /karne/ | /karne/ | /sher/ | /karne/ | ||
| cane | can (archaic) | chien | cão | canis | 'dog' |
| /kane/ | /kan/ | /shye/ | /kãw/ |
While that initially may seem like a lot of information, let's notice a few familiar things. First, note that in the word for “head”, we see a repeat of the directional we deduced earlier: p>b. With this larger sample, whenever we have a 'p' sound between two voiced phonemes, that sound is kept in Italian and changes to a “b” in the other examples (with the presumable exception of French.) So we have “capo”, “cabo”, and “cabo” in Italian, Spanish and Portuguese. That’s the consistency I mentioned above. Even when a directional doesn’t make sense (though this one did), as long as it is consistent, we can count on it. But do you notice any other direction changes or consistencies?
Notice that all the daughter languages establish a sound correspondence: Original “c” (written as /*k/ to make clear the sound is a hard "c" as opposed to the soft “c” in “circle”) always maintains its /k/ sound in its daughter languages- with the exception of French. French, instead, seems to have this rule, written phonetically: /k/=/sh/. Whenever a word in the other 3 languages starts with a /k/ sound, the French will start with a /sh/. It's weird, but it is also consistent. That is the important thing. It is consistent so even if it makes no sense we can trust it to be true. This becomes more convincing with larger and larger cognate sets, though we won't enlarge this one any further. This sound correspondence indicates that the original /*k/ somehow became /sh/ in French, another directional we can write as: /*k/>/sh/
Now, to clear things up, this is really what's going on with the French and /k/>/sh/. *k does not immediately become a “sh” sound just like that. Instead, what happens over time is that sometimes the “k” sound, which is made by friction of air flowing between your tongue and the back palate, will gradually become the “ch” sound. “ch” is made when the teeth are put together and the tongue starts at the top of the ridge behind the teeth and then suddenly lets air flow through the teeth, creating a similar fricative sound to the “k” sound. Again, practice making both the “k” and the “ch” sounds and you'll notice this. Both are fricatives made from different positioning of the tongue on the roof of the mouth- either in the back palate or the ridge behind the teeth. Then what happens is that the “ch” sound can gradually change to a “sh” sound- which is made identically to the “ch” sound, with the one exception being the tongue blocking the air flicking down and letting it through. Again, try it, “ch” and “sh”. So this rule can be written as *k>ch>sh.
That’s seems like a pretty long list of changes to infer. Can we really trust this? Does French show this evolution? Yes it does. In fact, we can see it in our very own English! A thousand years ago, in 1066 AD, William the Conqueror came from Normandy, France and invaded England, conquering and assuming the throne. From that time, onward, the English language adopted a number of French words, especially those that related to royal matters. The English word “chief” comes from the French, as does the name “Charles”. Notice that both of those words start with a “ch” sound. But later borrowings from French brought us the words “chef” and “Charlene.” Again, both are from the French, but this time, the “ch” really is a “sh” sound (ignore the spelling). Why the difference? Because the borrowings for “chief” and “Charles” were from the spoken French of a 1000 years ago, whereas the newer ones were only from the last few hundred years. So the directional rule /*k/>/sh/ really did happen! Initially, the language we now call French started “goat” with a /k/ sound but gradually it became a /ch/ sound. Finally, that morphed to /sh/. /*k/>/ch/>/sh/. So now we can say that our initial assumption of a “k” sound for the beginning of the proto-Romance version of “goat” makes sense for all the languages.
That explains the “ch” in “chevre”. But what about the “v”, where all the others have a “p” or a “b”. Well, French was no exception to that p>b directional. It, too, gradually replaced the “p” in that word with a “b”. But at this point, it may not be surprising to find out that “b” will sometimes become a “v”. Play with the two sounds, paying attention to what your lips and teeth are doing, as well as the type of sound it is- a voiced labial sound (a voiced sound made with your lips)- and you can see how that might happen. Thus, the rule extended in French to become: *p>b>v. And there we go. We have (for the most part- we left out the vowels) explained how the proto-Romance reconstruction of “capra” might have changed into the versions of it we see in Italian, French, Portuguese, and Spanish. And now, we look back at Latin, for which we actually do have written records, and we indeed find that our proto-Romance reconstruction fits- the original Latin word for ‘goat’ was ‘capra’.
It seems like a lot of work for just one word, and it is. But as we worked on that one word, we were able to come up with directional rules, such as /*k/>/sh/, that will help in explaining the other words in the cognate set. For example, look at the French “chef”, in contrast to the Italian, Spanish and Portuguese versions (again, ignoring the vowels). Right away, the /*k/>/sh/ explains the “ch” in “chef”. But here is the really cool part: the *p>b>v gives us most of the path to the “f” sound in “chef”. Looking at how “f” and “v” sounds are made (both labial fricatives) allows us to postulate another rule: v>f. So *p>b>v>f. Larger cognate lists show that these directional changes are consistent across the board, which is one of the things that indicate that this is a correct deduction. Thus, though the first word took the longest, once the initial work was done, the explanation for the others went fairly quickly.
With that as background, we can understand that linguists have formalized what we just did into a specific algorithm called the Comparative Method. It is broken into roughly 5 steps, though they are not necessarily done in this order and some steps repeat or influence the others.
1. Assemble a cognate list. The idea is to get a good sample, primarily of words that are the most resistant to borrowing from other languages. These would be terms for body parts, kinship relations (father, mother, etc), low numbers (one, two, three, etc), and basic geographical features.
2. Establish sound correspondences. Remember how we took the beginning “k” sound in Italian, Portuguese, and Spanish and were able to see that it always corresponded with the French “sh”? Thus, we could see that French and the others were related, even though the sound had changed over time in French.
3. Reconstruct the proto sounds. This can be done with the following techniques.
a. Establish directionality: p>b, /k/>/ch/>/sh/, p>b>v were all directionalities that we discovered that helped us to explain why the cognates were different and also gave us clues as to what the original word was. We saw that proto-Romance most likely originally had the word “capra” and the directionality rules showed us how things might have changed in the daughter languages. That conclusion would be consistent with other word changes, which we saw in the larger cognate set chart.
b. Majority wins. This is a tentative step. On the one hand, given that all the words for 'goat'- with the exception of French- started with a “k” sound, we could reasonably conclude that the original language word also must have started with a “k” sound. But majority wins isn't always the case. Remember how (ignoring French), Italian was the only word with a “p” sound in the middle, whereas Spanish and Portuguese had a “b” sound? Majority wins applied there would mean the original proto-Romance word must have been a “b”. But directionality always trumps majority wins. Some changes just don't happen. b>p between voiced phonemes is one of them. Moreover, historical context can help us determine language relationships, which changes the deciding weight we give to each language phoneme. This chart shows the relationship of the Romance languages.
Looking at the chart, and only noting our original 4 languages, we see that Spanish and Portuguese are more closely related than they are to Italian. So their agreement with the “b” sound doesn't give them precedence over that of the Italian. Those possibilities all being equal, the directional rule gives preference to the Italian. Thus, it had to be the “p” from the Italian 'capra' that was in the original proto-language and the immediate ancestor to Spanish and Portuguese (Ibero Romance, from the chart) is likely where the p>b occurred.
c. Factor common features. Phonetic elements from the daughter languages are analyzed and the common features are the ones we try to preserve. Going back to our 'goat' example, the middle consonants were Italian: p, Spanish, Portuguese: b, and French: v. In all four languages, the consonants were 'labial ', that is, made with the lips. In three of the languages, the consonant was made by stopping the air, called a “stop”. Thus, two major phonetic components, the common features, we'd most likely find in the parent tongue would be a labial stop of some sort- which would be either 'b' or 'p'. The directional of p>b and b>v would preserve those common features in the daughter languages, while also explaining the differences.
d. Economy. This simply means that it makes more sense that when you have multiple alternative paths to get from the projected parent word to the daughter word, you chose the shortest path- the one with the fewest independent changes. So if we wanted to postulate that the proto-Romance word for goat started with a “sh” sound, based on the French cognate, than we'd have to then assert three different /*sh/>/k/ changes in the remaining languages. First, that directional is not really common (/sh/ doesn’t go to /k/). Second, you are then requiring that uncommon directional to happen more than once in the remaining languages- three independent changes! The other way around, /k/>/sh/ makes more sense. Only the change in French is necessary, the /k/>/ch/>/sh/. It's simpler that way.
4. See how those reconstructions work on a larger set of words. We did 'goat' and got rules that helped us reconstruct 'capra' in proto-Romance. Afterward, we applied those rules to a larger set. Thus, here we apply discovered rules to a larger set of cognates and see if they explain some of those features. As we do this, we expand our correspondences and directionals to explain more and more words. We repeat for larger and larger sets of words until we’ve done them all
5. See how plausible our reconstructions are from a phonological and linguistic standpoint. Languages are consistent things. Sounds (where they are made and how- see below for an explanation), in particular, are consistent. The changes we find are consistent. The fundamental theory of linguistics is this: Sound change is regular and exceptionless. Therefore, we then must ask the question: Does everything fit together in a logical and consistent manner, accounting for all the changes from parent to daughter language? If so, we can have a high certainty in our reconstruction of the proto-language.
So how do these reconstructive techniques measure up? Actually, quite well. Since we have a well preserved body of Latin literature and documentation, as well as one covering the entire developmental lifespan of Italian, French, Spanish and Portuguese over the past 2000 years, it is easy to check our assumptions and reconstructions of proto-Romance against what actually happened with Latin. In example of the /k/>/ch/>/sh/ (“chief” and “chef”), we can actually see those postulated changes occurring in the written record, so it works.
Moreover, the proto-Romance that linguists have reconstructed as their demonstration of these methods is very close to actual Latin. Not 100% it's true. Certain elements of Latin, such as noun cases, an element of Latin grammar, have been lost in the formation its daughter languages and are not retrievable using these techniques. Still, over all, the reconstruction is quite astounding in its accuracy. Using these reconstructive techniques, we can take just the spoken suspected daughter languages and create a reasonable facsimile of the original mother tongue.
Representing Specific Sounds
You’ll remember that above, when we wanted to represent specific sounds, we generally used the letters which they were spelled with or, we’d use a “/” and then a ‘k’ or “ch” or “sh”.. There is a problem, though, with using simple letters to represent sounds. “c” can actually represent an ‘s’ sound, like the first ‘c’ in “circle”. But it also can mean a hard ‘k’ sound, like the second ‘c’ in “circle”. The ‘r’ in ‘rat’ is distinct from the trilled “r” in “perro” in Spanish and the ‘r’ in the French “cher”, which is an almost silent ‘r’ made in the back of the throat. “l” in “trouble” is different from the “l’ in “land”. Yes, they seem related, but they are made differently. Then, too, you have the sounds we represent with two or more letters. “ch”, “sh”, and “zh” all make use of the ‘h’ to somehow represent the sound of the air passing through the clenched teeth. But ‘h’ when by itself is a fricative in the back of the throat. And what about “th”? Sometimes it is voiceless, as in “thin” and sometimes it is voiced, as in “they”. This is also true of vowels. The ‘e’ in “the” is different from the ‘e’ in “except”. They are made in different places of the mouth. What about the the “i” in “bird”? It’s different than the ‘i’ in “tin”. And what about “g”? Sometimes it is soft, as in "general" and sometimes it is hard, as in "gas”. Now say “painting”. Notice what’s going on with the ‘g’ this time. Are you saying it with the hard ‘g’ in “gas”? No you are not. To confirm this, try to add the hard ‘g’ on to the end of “painting”, emphasizing it. “paintin-guh!” Sounds pretty dumb, huh? Pay attention to your mouth when you make the hard ‘g’. The middle of the tongue is at the palate stopping the air and then drops down while the throat voices the sound and it comes out the mouth. Now notice the "g" in “painting” (or any word ending in “ing”). The tongue is at the palate again, but this time it stays there and all the air is forced through the nose not out the mouth like a hard "g". It is nasal, as is “m” and “n”. (Try all of these and see for yourself.)So what do linguists do with all this? First, they note all the places involved in the creation of sound. A look at a diagram reveals some examples of these.
The alveolar ridge just behind the teeth, the lips, the palate, velum, glottis, and so on, are some of the places sounds are made. Next, they will note what kind of sound is being made. Is it made through stopping the air? That’s a stop. Is it made by friction with the air? We call that a fricative. Is it made by forcing air through the nasal cavity? That’s a nasal. Then they assemble all that information into a chart, using symbols to represent the specific sounds.
What about vowels? There are clearly more than six (even if you include ‘y’). Vowels are not made using friction, in the teeth or on the lips. Instead, vowels are formed without obstruction and use merely the shape of the mouth and where in the tongue is in the mouth. Again, we look at a diagram and chart, showing where in the mouth the sound is made, and how that is represented in IPA.
Though it looks complicated (and sort of is), let’s look at the ‘e’ sound in “bert”. Notice both where your tongue is in your mouth and the shape of your mouth when you make that sound. The tongue is in the middle (central) of the mouth and the mouth is pursed and almost closed. So according to the chart on the left, we are looking at an ɨ symbol for the “e” in “bert”. The “e” in “the” has the tongue central again (though a little more to the back of the mouth), but this time the mouth is slightly more open (close mid). So we have a ɘ for the “e” in “the”. In “bit” the “i” sound puts the tongue near the front (just behind the teeth) and the mouth is pursed (near close), so the symbol is ɪ. To see this in action, alternate between those three vowel sounds- ɨ , ɘ and ɪ in “bert”, “the” and “bit”. Pay attention to the shape and size of your mouth for each. You will notice the “close-mid close-near close” differences the chart shows for those sounds.)
The Reconstruction of Proto-Indo-European
So, finally (finally) we’ve come to the climax. I know it’s taken a while to get to this point, but the background information we have gone through was 300 years in discovering and formalizing. Knowing it will make this part a breeze.Beginning in the late 1700’s, language experts began to notice similarities between Sanskrit and classical Greek and Latin. This was an amazing discovery. Culturally and geographically, Sanskrit (being a language from the Indian subcontinent) seems to bear no relation to the languages spoken in Greece and Italy. While the similarity between Greek and Latin could be that they came from a mother tongue (similar to our noting the similarities between Italian and Spanish), the idea that an ancient language from India was related to them was mind blowing. Over the next 200 years, this relationship was investigated and fleshed out. One outstanding name in linguistics was Jacob Grimm, whom you may recognize from “Grimm’s Fairy Tales”. In investigating the changes that occurred in the Germanic languages 2,500 years ago as it broke away from PIE, he discovered directional changes that are part of what is now called Grimm’s Law. Just two examples of it will suffice: *p>f and *t>Ɵ (Remember Ɵ is the “th” in “thin”). The following table illustrates the changes that went on in the Germanic tongues (of which Gothic and English are part of) that didn’t happen to other PIE language families.
| *p>f | Sanskrit | Greek | Latin | Gothic | English |
| pad- | pod- | ped- | fōtus | foot | |
| pitár | patēr | pater | faðer | father | |
| *t>Ɵ | |||||
| trī-/tráyas | treīs/tría | trēs | Ɵrija (‘j’ is ‘y’ sound) | three | |
| tv-am | tū | tu | Ɵu | thou |
In this chart, you’ll notice the Greek, Latin and Sanskrit languages have a “p” and “t” sound that we recognize even in English words. “Podiatrist” is a doctor who specializes in feet. “Tristate” means three states. But after the proto-Germanic language broke away from PIE, *p>f and *t>Ɵ in the same way that *p>b when Spanish was forming from Latin.
One example is, in my opinion, outstanding and illustrates the power and accuracy of the Comparative Method. As they studied all the daughter languages, it became evident that at one time, PIE had to have consonants called “laryngeals”. Looking at the diagram of the mouth tells you that these are sounds made way back in the throat (near the larynx), similar to sounds in many Semitic languages like Arabic. The thing was, up until 100 years ago, no known daughter language (going back 3,000 years) had any kind of laryngeal in it and no modern Indo-European language today has it. Yet, as the rules and relationships were worked out, it became more and more clear that PIE had to have had them at one time. It was the only way to explain certain changes that went into the formation of its daughter languages. It was the only thing that made sense- but there was no evidence for it! And then, in the early 20th century, Hittite (a language that was spoken in what is now Turkey) was discovered and deciphered and was found to definitely be a part of the Indo-European family, but from the very distant past (~3,500 years ago). In that Hittite language they found them- laryngeal consonants. Because the language was a very early daughter of PIE, it added evidence that originally PIE had them as well but that it was later lost in the daughters. The reconstruction rules had worked!
And so we come to today. PIE, a language spoken more than 3,000 years ago, has been reconstructed. Here is just a partial list of PIE words denoting kinship and their versions in some of its daughter languages. The ‘H1’ and ‘H2’ (or H1 and H2) are a couple of the laryngeal consonants I was talking about. But you can see how it looks, overall. (Here is a link to a bigger list.)
The following is an example of a short story that is written in PIE:
To rḗḱs éh1est. So n̥putlos éh1est. So rēḱs
súhnum éwel(e)t. Só tós(j)o ǵʰeutérm̥ (e)pr̥ḱsḱet: "Súhxnus moi
ǵn̥h1jotām!" So ǵʰeutēr tom rḗǵm̥ éweukʷet: "Ihxgeswo
deiwóm Wérunom". So rḗḱs deiwóm Werunom h4úpo-sesore nu deiwóm
(é)ihxgeto. "ḱludʰí moi, phater Werune!" Deiwós
Wérunos km̥ta diwós égʷehat. "Kʷíd welsi?" "Wélmi súxnum."
"Tód h1éstu", wéukʷet loukós deiwos Werunos. Rēǵós pótniha súhnum gegonh1e.
English translation:
Once there was a king. He was childless. The king wanted a son. He asked his priest: "May a son be born to me!" The priest said to the king: "Pray to the god Werunos". The king approached the god Werunos to pray now to the god. "Hear me, father Werunos!" The god Werunos came down from heaven. "What do you want?" "I want a son." "Let this be so", said the bright god Werunos. The king's lady bore a son.
Once there was a king. He was childless. The king wanted a son. He asked his priest: "May a son be born to me!" The priest said to the king: "Pray to the god Werunos". The king approached the god Werunos to pray now to the god. "Hear me, father Werunos!" The god Werunos came down from heaven. "What do you want?" "I want a son." "Let this be so", said the bright god Werunos. The king's lady bore a son.
And here is my own
translation of a couple phrases into PIE:
*twé (egoh²) leubh-mi
[You (I) love]
*dw-en-elo (tu) h¹és-si
[Beautiful (you) are]
[You (I) love]
*dw-en-elo (tu) h¹és-si
[Beautiful (you) are]
Keep in mind what we are looking at, what has been done. PIE is a completely reconstructed language- there is no written or recorded instance of it. And yet despite that, we can create a reasonable representation it, a resurrection of what those people, 3,500 years ago, were speaking as they lived in the Pontic Steppes of Eurasia. How cool is that!
Additional links:
http://colfa.utsa.edu/drinka/pie/pie.html
http://en.wikipedia.org/wiki/List_of_Indo-European_languages
http://en.wikipedia.org/wiki/Proto-Indo-European_language
http://dnghu.org/Indo-European-Language-Europe/
It's great to see you blogging again. Although I'm bemused as to how you find time to do all this research! 你继续写
ReplyDeleteThanks. This stuff has been gelling in the back of my mind for the last 6 or 7 years as I read different books on the subject. So it was spread out over a long period of time. Sometimes I just like to explain something clearly and completely, as a way of making sure I understand it properly. I appreciate your comments though.
Delete