The reconstruction of Proto-Indo-European

This discussion is a spinoff from posts by rscwc in the Hindi-Devanagari thread.

The relationships among the languages of the Indo-European family were first recognized in the 16th century. Filippo Moretti, a Florentine merchant (Italy was not yet a nation so there were no "Italians"), was the first European known to write about Sanskrit, and he noticed many words that are similar to Italian words. (Florentine was the dialect that provided most of the basis for the modern Italian language). Devah/dio = "god," sarpah/serpe = "snake," sapta/sette = "seven," nava/nove = "nine." These four words are excerpts from a list so long and varied that the similarities cannot possibly be attributed to coincidence or even borrowing.

A few decades later Thomas Stephens, an English Jesuit priest, noticed striking commonalities among Konkani (the Indic language of Goa), Latin and Greek. Unfortunately his letter to his brother, in which he pointed this out, was not published for several hundred years, but the point is that the existence of the Indo-European language family has been recognized since early in the Renaissance/Reformation/Enlightenment era, when European scholarship became less constrained by religion and science as we know it began to be developed.

The organization of the multitude of Indo-European languages into a taxonomy has been underway for more than a century. Today we recognize at least four branches of the family, many of which are broken into subgroups.

• Armenian and Albanian each comprise a single-member branch, although they may have had relatives in the past which are now extinct.
• The Western Branch includes the Celtic, Germanic, Italic and Hellenic subgroups.
• The Eastern Branch includes the Baltic, Slavic (some linguists combine those two into a single subgroup) and Indo-Iranian subgroups.

The observation that at each taxonomic level the member languages clearly descended from dialects of an older "proto" language, inevitably led to the conclusion that all of the subgroups, groups and branches ultimately descended from a single Proto-Indo-European language that was spoken in the homeland on the Pontic Steppe before 2500BCE. This was the time at which all linguistic, anthropological and archeological evidence tells us that the original Indo-European tribes began to migrate in disparate directions. Without the reinforcement of communication with each other, their dialects slowly diverged into mutually incomprehensible languages.

The existence of a Proto-Indo-European language is accepted by the community of professional linguists in Europe, India, Persia, America, and wherever Indo-European languages are spoken. This is not a hypothesis waiting to be proven. The evidence is overwhelming and incontrovertible.

There are several areas in which similarity is used to establish a linguistic relationship. The most obvious is vocabulary. There are entire websites where you can see the huge glossary of words common to many or even all Indo-European languages. I'll just present a couple of the most convincing.

• Numbers. Latin unus, duo, tres, quattuor, quinque, sex, septem, octo, novem, decem. Sanskrit eka, dvi, tri, catur, pañca, shash, sapta, ashta, nava, dasa. Russian odin, dva, tri, chetyre, pyat, shest, syem, vosyem, dyevyat, dyesyat. Swedish en, två, tre, fyra, fem, sex, sju, åtta, nio, tiu.
• Pronouns. The majority of Indo-European languages retain the T-V paradigm of singular and plural "you," such as French tu/vous and Russian ty/vy. Almost all have some form of the first-person singular pronoun starting with M, such as English "me," "my" and Russian moi, mnye.

These are good examples because numbers and pronouns are such fundamental components of language that they are rarely borrowed from other languages. Sure, you might find one or two or even a handful, but never all of the various sets of fundamental words. Japanese uses Chinese numbers in certain contexts, the way we use Latin and Greek numbers in scientific terminology, but just as our children count their toes using Germanic numbers, Japanese children use Japanese numbers.

RW offers dissimilarities as a reason to doubt this relationship. But this challenge does not hold up under the scrutiny that we can apply to the Indo-European languages just since written evidence has been left for us. The differences between French and Spanish have only had a little more than one millennium to accrue, yet they are astounding. It's easy to see why Latin and Greek, which had two millennia to diverge, can be so different, and it becomes easier to realize that the plethora of obvious similarities which they nonetheless do have is irrefutable evidence of their relationship. The same holds for Latin and Sanskrit, which had considerably more time to diverge.

RW specifically brings up the words for "sun" and "moon." I don't know what they are in Sanskrit, but the Slavic languages are closely related to the Indo-Iranian languages and their words for "sun" are various versions of solntzye/slunche, not much different from Latin solis. As for "moon," well gee the Germanic languages lost the common Indo-European root lun- and managed to come up with a new one. It's easy for people to throw out an old word and replace it with a new one. The Germans, Dutch and Scandinavians all call a certain familiar animal hund/hond, but somewhere along the way we decided to start calling them "dogs," even though we still have the ancient word "hound" to describe certain breeds.

Words also shift meanings over the centuries. In German a Tier is any animal, but in English a "deer" is just one genus of ruminants. A German Knecht is a farm hand, but an English "knight" is a nobleman.

RW asks about apples. Apples originated in Turkey and it took a long time for cultivated varieties to be developed that would grow in other regions, so it's no surprise that the people of the Ancient Indus Valley were not familiar with them. There is no question that there were no "apples" in Mesopotamia at the time when the famous legend about the Garden of Eden originated. That fable has been translated many times and the original is lost in prehistory before the invention of writing, but archeologists and biologists tell us that the fruit was surely a pomegranate, a staple crop in that region in the Neolithic Era, because there was simply nothing else there that was remotely similar to an apple.

Finally RW mentions phonetics. I hope by now everyone who reads the Linguistics board understands that sounds are the most ephemeral components of any language. They change rapidly, capriciously, illogically, and even since the invention of writing the standardizing effect of written records on phonetics has been very weak. Americans have only existed as a people for a few hundred years, but look at all the different accents we have in our country. This difference was much greater in the past, before the pervasive influence of radio and TV. The difference between the Standard American accent and the standard Received Pronunciation of the UK is enormous; sometimes it’s very hard for us to understand each other.

So the notion that two Indo-European languages might have two entirely different sets of sounds is hardly remarkable. Using comparative methods we can trace phonemes back to the Proto-Indo-European language and say, “this phoneme was originally a B and this one was a T, this was an A and this was a U,” but we don’t really know exactly how they pronounced those sounds. If someone went back in a time machine and stepped out speaking Proto-Indo-European the way we’ve reconstructed it, he would have many of the words and most of the grammar right, but they might not understand his pronunciation at all.

 

I'm not sure. Let's wait and see what he says.

 

The word and the symbol for zero cannot possibly go back to Indo-European because the concept did not exist. It was invented by the Arabs at the end of the first millennium CE; that's why we call our number system Hindu-Arabic instead of just Hindu: nine of the symbols are Hindu, but zero is Arabic. Of course mathematicians struggled with the concept of zero in earlier eras, but they never invented a satisfactory way of writing it.

However, in response to your specific question, the word "zero" is in fact derived from Arabic sifr, through the word zephirum that was added to the Latin vocabulary by European scholars.

Oddly enough, when Fibonnaci brought the Hindu-Arabic numerals to Europe, only scientists and mathematicians adopted them. The rest of the population ignored them for about 200 years.

One of the things you have to be careful about when tracing linguistic relationships is to understand the level of civilization or pre-civilization in ancient eras. The Indo-Europeans were a Neolithic people (agriculture, permanent settlements) or possibly even Mesolithic (farming but no animal husbandry). They had no words for numbers larger than one hundred because they did not need them.

The Indo-European word for "hundred" was kmtom. In fact we use this word as a shorthand naming convention for the two branches of the Indo-European family. We call the Western Branch the Kentum Branch (usually spelled "centum") because the K in kmtom was preserved. Greek (he)katon, Irish céad, Latin centum. We call the Eastern Branch the Satem Branch because the K was palatalized into S. Russian sto, Sanskrit satem, Lithuanian šimtas, Farsi sad.

K is a very fragile phoneme in Indo-European. Only in Irish and Greek is the K of kmtom still pronounced K. In the other "Kentum Languages" it has changed wildly. It's become H in English "hundred," TH in Spanish ciento, CH in Italian cento and S in French cien. (I'm not sure about Albanian, I can't find it anywhere, but in Armenian it starts with H. Albanian and Armenian are isolates, neither Kentum nor Satem languages.)

Most modern languages have hundreds of thousands of words, and ancient languages probably did too. Even the languages of pre-civilized people surely had tens of thousands. But they were not all retained in their descendants. As I noted in my last post, we can observe the changes in vocabulary in languages over the past couple of thousand years, because the people who spoke them helpfully left written records for us. If you extrapolate this rate of change back into the Stone Age, and then remember that without the stabilizing influence of writing languages change much faster, you'll realize it's a miracle that we have any remnants of the old languages at all.

As times change, people make up new words to talk about new ideas, relationships, objects and activities, and they discard old words that they don't use any more--or change their meaning. They also sometimes simply make up or borrow new words just because they're new and exciting--why do you suppose the Anglo-Saxons started calling their companion animals "dogs" instead of "hounds," like they did before they left Germany? Why do we Americans call a portable battery-operated lamp a "flashlight," when the British retained the venerable old word "torch"?

This is why we can't trace the history of any language or language family back much more than six or eight thousand years. In a longer timespan than that, literally everything about a language can change completely: vocabulary, phonetics, grammar, syntax, even the way of looking at the universe. Just recently it was discovered that the Na-Dene people of North America (e.g., Navajo, Tlingit, Apache) are genetically related to the Yenisei people of Siberia; those North Americans are the descendants of the people who walked across Alaska into the New World, and the Yenisei are the descendants of the ones who decided to stay home. (There are many other groups of North, Central and South Americans and we don't know their genealogy yet.) Motivated by this discovery, linguists began comparing the Yenisei language (which had not been exhaustively studied, since there aren't millions of linguists around) to the Na-Dene languages. They found a relationship! These languages are separated by 15,000 years of migration, so this is a milestone in the history of linguistics. But don't hold your breath waiting for a similar revelation about the Mongolic and Ural-Altaic languages, or Austronesian and Dravidian.

Please Register or Log in to view the hidden image!

I don't really know how many Proto-Indo-European words have been catalogued. I'm not a member of any scholarly organization so I don't have access to their files. Wikipedia lists several hundred of the ones most commonly encountered in etymologies of multiple languages. In general, each one has multiple descendants in the modern and medieval Indo-European languages.

The Proto-Indo-Europeans didn't have pomegranates, so they didn't give us a word for it.

The Proto-Indo-Europeans didn't have coconuts or bananas either. Bananas are from New Guinea. Coconuts are probably from the Ganges (not everyone agrees on that) so only the Indo-European tribes who migrated south encountered them.

Neither do I, at least not a professional. Most of my knowledge was not gathered formally.

So am I. I have been a software engineer since 1967, and I studied at the California Institute of Technology (although that is not where I graduated from).

 

Please read posts more carefully before you respond to them. I did not say that the Arabs invented numerals. Each of the world's civilizations invented numerals. They needed them to record business transactions between strangers in cities who did not instinctively trust each other, the way they would have trusted a fellow villager in the Neolithic Era or a member of the same extended-family unit in the Paleolithic Era.

And I also did not say that the Arabs invented the concept of zero, just the symbol. Here is a website that provides the best explanation of the history of zero that I have ever seen. I'm sorry I had not found it earlier; it gives more accurate information than my own post. It tells us that some of the ancient civilizations had also struggled with ways to record zero and had devised their own symbols. It mentions the early Indian symbol and describes it as a smaller version of the modern zero. But other sources I have read say that it was really just a large dot.

The O symbol that we use today was used by the ancient Greeks, something I had certainly never heard before. The article seems to imply that the Indians might have been inspired by it when they invented their own symbol, which occurred at a later date. If not then perhaps the Arabs uncovered it, since they were meticulous about rescuing all of the science, philosophy and literature from Ancient Greece as Christian Europe was descending into the Dark Ages (approximately 500-1500CE).

So let's stop arguing. Both explanations are plausible. Sometimes the information we need just isn't available.

Actually most Westerners simply call them Arabic numerals. They were brought to Europe by scholars who had traveled to the Arab lands during the Dark Ages. No one in Europe knew anything about India in that era.

It's linguists who are attempting to popularize the term "Hindu-Arabic numerals." But no one listens to us. They still insist that Dutch and Flemish are two different languages.

Please provide a link to their website. Another Indian came to SciForums two or three years ago and presented "scholarship" by an Indian who called himself a linguist, but it was preposterous. It was self-published with no peer review, and it had names, dates, and other easily-checked details quite wrong.

He was trying to prove his own theory. That is not what scientists do. They test their theory to see if it's correct. This fellow was no scientist, he was a crackpot.

There is a mountain of peer-reviewed, scholarly evidence supporting the theory of the Indo-European proto-language. It has never been falsified by proper use of the scientific method. It is almost as well-established as the theory of evolution or relativity.

I have to admit that I'm utterly astounded that even without reviewing any of this scholarship, a person could look carefully at Bengali, Dari, Polish, Latvian, Armenian, Greek, Portuguese, Norwegian, Albanian and Welsh, and insist that they are not all descendants of one earlier language.

 

I never claimed to have vast knowledge of Sanskrit. I merely cite the works of professional scholars who do have that knowledge. Their works have been peer-reviewed and their hypotheses have been judged true beyond a reasonable doubt, which elevates them to the status of theories.

I am not the scientist here, and have never claimed to be one. My knowledge of languages and linguistics qualifies me as a high-class amateur among Americans, but I have never presented myself as a professional.

It is not I whom you are challenging, but the canonical, peer-reviewed theories of linguistic science, which are accepted universally by the overwhelming majority of respected linguistic scientists. You are going to have to offer some very convincing evidence if you expect to challenge them successfully, or to even have your arguments regarded with honor. The Rule of Laplace teaches us that extraordinary assertions must be supported by extraordinary evidence before anyone is obligated to treat them with respect, and so far all the "evidence" you and your buddy have offered are unfocused ramblings that are difficult to even follow, much less understand. You often base your arguments on entirely unscientific premises, such as the superiority of a mother's instincts over an accumulated body of scholarship.

In this particular post, you have asserted that a very long post full of examples, correlations and citations is incorrect, yet you offer no evidence to the contrary. This is not how science is performed.

I have not made a big fuss over this since it is apparent that the other members are intelligent enough and well enough educated to realize that your objections are lacking in merit and that your understanding of science is not well advanced.

But in the future, if you wish to continue posting in this subforum, you will need to try to behave like a scientist, or at least like a student of science.

You seem to feel that linguistics is not a science at all, but some silly thing like Feng Shui or Tarot cards. Well, it is a recognized science, albeit a "soft science," and if you intend to remain here you will have to accept that fact and deal with it.

 

There is no Indo-European word for gold because the Indo-Europeans were a Neolithic people who did not practice mining or metallurgy. Gold was a substance they learned about (and perhaps acquired) from the Bronze Age civilizations south of their homeland in the Pontic Steppe.

So every Indo-European tribe developed its own word for gold as it entered the Bronze Age and began mining and working gold. Russian zolot, Greek khrusos, Latin aurum, Germanic geld.

Apples are native to the region we now call Turkey. It took a long time for the fruit to spread to Mesopotamia and other regions. The "apple" of the Garden of Eden mentioned in the English translation of the Bible was almost certainly a pomegranate.

"Sun" is a common word throughout the Indo-European family, e.g. Latin sol, Greek helios (Latin S is often H in Greek), Russian solntzye. But the other heavenly bodies were named and renamed by the individual tribes, often after their own gods.

Poppycock. Just look at the different pronunciations of the first consonant of the Latin word centum (pronounced kentum) in the modern Romance languages: French sen (cent), Spanish thyen (cien), Italian chento (cento). The K in Indo-European kmtom has undergone dramatic changes in all of its daughter languages, including the H in our Germanic form "hundred." This is in fact why the two main branches of the family are colloquially called the Kentum Branch and the Satem Branch, from the Latin and Avestan pronunciations of the same word.

You are really embarrassing yourself with these ridiculous assertions. Every respectable linguist on Earth recognizes the immense volume of evidence attesting to the incontrovertible history of the Indo-European languages. There is no controversy at all. It is the most thoroughly studied language family on earth! We can trace the phonetic, grammatical and lexical evolution of every Indo-European language back through the millennia in surprisingly precise detail. There are no major gaps. The individual subgroups--Celtic, Italic, Germanic, Hellenic, Indic, Iranic, Baltic, Slavic, Armenian, Albanian, etc.--are as familiar as our own brothers and sisters, and the relationships among them are quite well defined.

Contrary to your bizarre and unscientific assertion, many linguists postulate a superfamily that links Indo-European to other families such as Sino-Tibetan, Dravidian, Finno-Ugric, Mongolic, Uto-Aztecan, etc. The usual name given to this hypothetical superfamily is Nostratic, from the Latin word for "ours." There is no persuasive evidence for this superfamily, so it remains an intriguing hypothesis rather than a theory.

We would all like to know whether language is a technology that was invented once in one group of humans and then learned by their neighbors (like the domestication of dogs, all of whom are descended from one small pack in Mesopotamia), or one that was reinvented by many different groups in many different times and places (like the cultivation of plants, pottery, bronze, the bow and arrow and the domestication of cattle).

For decades it was an accepted principle that language changes so much over the centuries that literally everything that comprises a language can change completely in about six or seven thousand years--vocabulary, phonetics, grammar, syntax, even its people's entire perspective on the universe as they migrate to new regions or invent new ways of living. (Compare our familiar subject-verb-object syntax to the topic-description syntax of Japanese and ponder what that says about how differently we understand the world and our place in it.)

However, in this century very strong evidence was discovered that links the Na-Dene language family of North America (including Navajo, Tlingit and several other Native American languages) to the Yeniseian family of Siberia (of which only one member survives but there were a few others in historical times). If this relationship turns out to be real, it links languages that have been separated for about fifteen thousand years! It gives new hope to our efforts to identify larger language families than the ones we have so far pulled together.