Oldest known language(s)

* * * * NOTE FROM THE MODERATOR * * * *

As I noted in a previous post, the author's use of the phrase "harsh revisions of phony history" is evidence not only of bias, but also of unprofessional scholarship, specifically: Failure to submit his work for peer review and proper editing.

I therefore judge String's accusation as reasonable, and your objection is overruled.

I do not find that accusation in String's post. Please be careful to quote people accurately during an argument.

His response appears to indicate that he understands your arguments. If you disagree, please explain your objection in a little more detail, since your own statement is not directed at anything specific in his post.

I see no failure in String's remarks. As for "groundbreaking ideas," so far I see no evidence that Ranganath is an accomplished scholar of linguistic science. His haphazard combination of five languages in two different families into one group in order to bolster his rhetoric is an example of very poor work that borders on intellectual dishonesty.

I would have to read further into this book in order to draw a fair conclusion, but based upon two instances of unscientific and unscholarly rhetoric in such a short exerpt from it, I am obligated to consider the possibility that Ranganath's work is crackpottery, rather than groundbreaking science.

 

I am sure as a Self-Proclaimed Linguist, you know more about the Tamil language than Ranganath. Therefore yours must be a professional opinion. Correct? Have you published any paper yet?

 

I have never claimed to be a professional linguist and I have frequently enumerated my extremely modest credentials for this post. I have two friends who speak Tamil and I know that it is a Dravidian language, not Indo-European.

My criticism of Ranganath on that point was narrowly focused: on his inclusion of Tamil and Brahui, a second Dravidian language, in a list with three Indo-European languages. In a book written for linguists, who would all be expected to know the difference, that would be acceptable.

But this book is clearly not written for linguists. (Again, I'm judging from the excerpts presented and I will be pleased to stand corrected if other excerpts are presented which contradict me.) To casually link two language families, without even mentioning the possibility of a Sprachbund, while pursuing an argument regarding the ancestry of languages and the relationships among them, when writing for people who won't know the difference, is (to put it in the least accusatory terms possible) bad writing.

And regardless of my status as a linguist, I am a professional editor who knows bad writing when he sees it.

The fact that a book which positions itself not just as a work of scholarship, but as the assertion of a hypothesis that is guaranteed to generate controversy (and in fact invites it in a petty criticism of "phony history"), went to press with this quality of writing strongly suggests that it was not properly peer-reviewed or even professionally edited. Hence my accusation of vanity publishing, a practice that has become affordable and commonplace in the internet era.

And one that we all must be on the lookout for.

For example, we recently had to deal with Anita Meyer, a certified crackpot and spammer, promoting a book which insists that the Hebrew abjad was based on numerals. She conveniently ignores (or perhaps is genuinely ignorant of) its well-documented lineage through older Semitic writing systems, going all the way back to the Ancient Egyptians in a time before the Jews as a distinct people existed.

 

You said in another post

"As the only language expert on this website and a professional editor and writer,"

and your BOLD fonts in reply to my post previously, I thought you are a professional linguist and a bully.

That is how it came across...

My bad...never mind...ignore my previous post.

 

I may be a bully, but I'm not a professional linguist.

Please Register or Log in to view the hidden image!

Hey if you misunderstood then maybe someone else did too. On a science forum it never hurts to explain something in too much detail.

The Proto-Indo-European people (probably a collection of related tribes speaking dialects of one language) lived 6-7KYA in the Pontic-Caspian Steppe where Europe meets Asia north of the Black and Caspian Seas.

The Anatolian branch split off first, around 3500BCE, but those languages (e.g., Hittite) are now extinct. Around 3000BCE the Western and Eastern branches split, and by 2500BCE Proto-Greek and Proto-Indo-Iranian were established, representing the two branches.

It's difficult to say "how old" any language is. All languages change over time. The definition of a "dialect" vs. a "language" is that speakers of two dialects can understand each other, perhaps with some effort, and perhaps only when close to each other on a dialect continuum, whereas speakers of two languages cannot.

This rule works only reasonably well in a spatial context, and even then politics frequently overrides it. Flemish is a dialect of Dutch, but don't tell that to the Flemings, who cite their "language" as one of the reasons to demand more independence from Belgium. Dutch and German form a dialect continuum from Berlin to Brussels. Pick any two towns 25 miles apart and the people can talk to each other. Yet few foreigners know this, and everyone calls the German of Vienna and the Dutch of Amsterdam "languages."

If we apply this same rule to a temporal context, it becomes a big headache. If you take 200-year snapshots of Rome, the people in any two adjacent snapshots can talk to each other. But no Ancient Roman could understand Modern Italian, and only the scholars in modern Rome can understand Latin. The same is true of Paris, Lisbon, Madrid and Bucharest, and moreover Latin evolved into disinct languages in each of those places, which are not intercomprehensible with each other or with Latin.

So, would you say those people all speak Latin, so all their languages are as old as Latin?

This is an important question because of Greece. Modern Greek has not diverged as much from Ancient Greek as Romanian has from Latin. According to a friend who spent her adolescence in Greece, Greek universities teach from untranslated ancient texts, and the classes are taught in the Greek of Aristotle, not the modern vernacular. It's assumed that any educated Greek can read Plato in the original, and could speak with the author if he suddenly appeared.

By this measure, Greek is one continuous 2500 year-old language, whereas neither Latin nor Italian is. That judgment will certainly make the Greeks happy, but not the monks in Vatican City who speak Latin among themselves--a dead language that they never spoke with their parents but rather learned in their religious studies.

No language evolves at a steady rate, and we try to find junctures at which major change takes place, to identify one "generation" of a language from the next. In our language this has occurred three times. When the Roman Empire collapsed, the Angles, Saxons and Jutes sailed over to conquer and occupy their abandoned colonies in Britannia. They brought their closely related dialects of Old German with them. But after a couple of centuries of isolation from the homeland, and of the influence of the substratum of British, a now-extinct Celtic language spoken by the natives, they were speaking Anglo-Saxon, a different language than the one spoken by the Germanic tribes on the continent; until recently Anglo-Saxon was usually called "Old English." This is the language in which Beowulf was written; an Anglo-Saxon from 500CE and an Anglo-Saxon from 900CE could probably understand each other.

Then in 1066 the Normans invaded, conquered and occupied Angle-Land. French was the language of government and commerce, and this superstratum had a huge impact on the Anglo-Saxon language: new words, simplified grammar, modified phonetics. Within one or two hundred years the Anglisc people were calling themselves English and speaking Middle English, which would not be understandable to an Anglo-Saxon.

The next generational juncture is not so easy to pinpoint. The French never actually retreated from occupied England. They simply assimilated into the native population and adopted their language. Modern British government and culture can trace an unbroken chain of development back to the Normans. This assimilation had another huge impact on the language, and by the 15th century Modern English was distinctly different from Middle English. The phonetic changes alone would render our language incomprehensible to Chaucer. For example, our long A, E and I are nothing like the cardinal vowels they were in his time; he would hear those sounds as long E, long I and the diphthong AI.

So we can conveniently break English into Modern English, Middle English and Anglo-Saxon (or Old English), each of which existed for a few hundred years before changing into something unrecognizable. Does that make English 500 years old, or 1600 years old? In any case, it's not as old as Greek.

We absolutely do not know this. It's difficult to trace any language back before it was written. All we can do then is correlate it with obviously related languages, and use migration patterns and linguistic analysis to figure out when all the related tribes were one.

Using these methods we can say that people were speaking Proto-I-E around 5000BCE, but there are no convincing relationships to any other language family that indicate whether people were speaking any language at all before that, much less that I-E and any other language family or families are related.

Many hypotheses have been offered, and they are based on good scholarship, using apparent similarities. But none of these is accepted by the community of linguists as a whole. No piece of evidence is sufficent to be conclusively ruled more than coincidence.

This is the question we can't answer. We don't know whether language was invented only once, and, like many technologies which consist primarily of ideas rather than artifacts, it quickly spread. It could just as easily have been invented in multiple places at multiple times.

Language changes slowly but steadily, and in roughly ten thousand years it can turn over completely: vocabulary, grammar, syntax, even the underlying perspective on life and the universe. There's no reason that Japanese, Hopi and English can't have a common ancestor in 25000BCE. But there's also no reason to assume that they do.

It would require a much greater reduction than 75% to reach that point. There are many definitions of language; the one I enforce here is rather generous, but clear. It is a set of symbols (spoken, written, bleeped or tap-danced) with well-defined meanings, accompanied by a set of rules for combining those symbols to express a relationship between their referents. I'll waffle over the number of symbols that can be combined. A set of invariable one-sound commands given to a herding dog is not a language, whereas a parrot that looks at an object and calls it a "red round key" is using language. Combinations of two symbols, that's borderline and I don't know if any animal is at that level anyway so perhaps it doesn't matter.

We keep calling dolphin communication "language," and we know that individuals have names for themselves, but other than that we don't know what any of their sounds mean. We're just assuming from the length of the "sentences" and the variety of the ways they're constructed, that they must be combining more than two "words," but that is a really smart hypothesis, not a fact.

The migration routes we have are pretty thoroughly established and not very subjective, but for the most part they are large migrations. The cost of DNA analysis is dropping steadily, so soon it will be a matter of going around and collecting all the samples, rather than queueing up for your turn in the laboratory.

Many of them were spread by conquest and assimilation. The Romans didn't routinely kill off the native people in their provinces. The people in northern France still have the DNA of the Germanic Franks (as well as their uvular R sound) and the people in southern France still have the DNA of the Celtic Gauls (and their flapped R sound).

But marginalization, exile and extermination have also been powerful forces, as in the United States.

The spread of Aramaic was an interesting phenomenon. The Aramaeans themselves never became a major political force, but for reasons only dimly understood the Syrians and Babylonians adopted their language as a lingua franca for their empires. (Why don't we say lingua aramaica? And in any case why do we say it in Latin instead of French--or Aramaic?) As each empire gave way to the next larger one, Aramaic was so well-established that the new conquerors kept it. Aramaic was the language of the Persian Empire and of the Ottoman Empire, which was still in existence less than 100 years ago. It is still spoken and there are websites in Aramaic.

So to get back to your original question, a strong case could be made for calling Aramaic the oldest continuously-spoken vernacular language.

Please Register or Log in to view the hidden image!

 

To say that Catalan is 2500 years old is to say that Catalan is Latin, or at least a dialect of Latin. This violates our definition of the word "language;" Catalan and Latin are not intercomprehensible, therefore they must be two separate languages, not two dialects of the same language.

We don't use the concept of a dialect continuum in a temporal sense, only spatial. Obviously if you take snapshots of any language two hundred years apart, it will be intercomprehensible across the two hundred-year boundary, so it's not a useful model. Latin and Romanian are as different from each other as Anglo-Saxon and Modern English.

This is an important point because it makes a language like Greek stand out: across a 2500-year temporal displacement, Ancient and Modern Greek are intercomprehensible. That's an important difference between the Greek continuum and the Latin-Italian, Old German-Modern English, Sanskrit-Hindi or Old Slavonic-Polish continuum.

The paradigms that are useful in one science are not always useful to another. In vertebrate biology we have an obvious beginning (and end) to the life of an individual organism, so it's useful to date the age of a human or other mammal to its birth. We have no such beginning for the life of a language, although we sometimes have an end. So we need a better way to date its age.

Every language has different milestones. For Afrikaans it was consensus: the mid-19th century when Afrikaners stoped writing in Dutch. For Middle English it was political: the Norman Invasion in 1066 when it acquired a French superstratum. But the purpose of establishing a milestone is to establish at least a rough date at which an earlier form of a language has clearly changed into a later form, with a loss of intercomprehensibility.

There are several different estimates that represent diligent scholarship. I'll go with the consensus. To express the age in scientific terms we'd calculate the standard deviation among the estimates and say that the date of the split was something lke 5.7±1.2KYA.

Cultural discontinuity is a major factor in change. Language reflects its community's view of the universe. When it is conquered or goes through any other kind of paradigm shift like a qualitative increase in technology, this changes that view of the universe: England was conquered by the Normans. Breaking apart into separate communities also does this: France and Spain broke away from Italy. The Slavic nations did not undergo such wrenching changes as they separated, and today their languages are still much more closely related. I heard a Czech lady work up a pidgin version of Croatian during lunch in Croatia. No German could do that over lunch in Norway.

Greece didn't really undergo any wrenching changes, and its community did not spread out and separate. The Ottoman conquest was surely the greatest trauma in Greek history, but as conquerors go, the Ottomans were fairly tolerant of native cultures. The Greeks were not even converted to Islam.

I’m no expert but I think so. It has more of the grammatical complexity. Although to be fair almost all of the Western Indo-European languages have more grammatical complexity than English. Greek also has not undergone as many phonetic shifts as English. But again, English, like all of the Germanic languages, underwent the shift of Grimm’s Law (D-->T, T-->TH, K-->H, etc.), rendering many Indo-European words nearly unrecognizable, such as “hound” for canis.

I’m sure we all feel the same way, but in science, feelings do not count as evidence. This is a very astute hunch, but a "hunch" is a merely "a promising hypothesis that needs more evidence."

I don’t understand how you define a “modern language.” A language must provide the set of tools its speakers need to communicate effectively. All known languages, extinct and living, do that.

We don’t know how fast languages changed in the Paleolithic Era, when life itself didn’t change very fast. Technological advances did not happen as often as they do today. Archeology and carbon dating inform us that thousands of years passed between one innovation in the way a stone tool was constructed and the next. So it's likely that there was not as much pressure on languages to change as there has been since the Paleolithic Era ended and people began living in larger communities and settling down in permanent villages, and inventing new technologies more often.

The problem remains that we still have no persuasive evidence to date the invention of language, or of any language. Of course it seems unlikely that a language like Proto-Indo-European, with its maze of grammatical rules and its intricate inflections, could have sprung just that way from the collective imagination of a human tribe. But many inflections are just vestiges of separate words that were joined to another word to express relationships. The transition from separate words to inflections might have been a single change, perhaps brought about by teaching the language to another tribe who didn't quite grasp the space between the words. For all we know, that transition in Indo-European might have only taken a few generations to complete.

When Koko saw her first zebra she called it a “white tiger.”

I have still never seen a report indicating that Koko or Washoe or the other apes trained in ASL have been introduced to any deaf people who can only communicate in ASL. I think those people would have a much different perspective on their language ability than hearing people who learned ASL just for the experiment but use spoken language among themselves. I would pay attention to their opinion of the apes’ language skills. For them, it would be the equivalent of an ape using spoken language with one of us. They would be able to communicate with the deaf people better than you or I can.

 

Do you mean the cave paintings? The earliest that have been found are in France and Spain and date back about 32,000 years, although the conditions might introduce errors into the carbon-dating process. There are some in Namibia that are also quite old, about 25KYA (KYA=Thousand Years Ago).

Since we don't know what they mean, and may never know, we can't say whether they are language. More recent cave paintings from the Neolithic Era (10KYA) have been found that are representations of religious rituals, such as people praying to their livestock.

In any case, even if they are language, that still does not establish them as older than spoken language. We don't know when spoken language was invented so we don't know how old it is.

Linguists point out that humans must surely have tried to migrate out of Africa before 60KYA. Why were their migrations suddenly successful after that date? Two in a row occurred that populated the entire planet. There is no evidence of new technology being invented around that time that would have facilitated the project.

Unless that technology was language, one that leaves no archeological evidence. Language makes it easier to plan, coordinate and remember things. Perhaps that's what got us out of Africa. If it's true, it would date language to at least 50KYA, when the second wave of migration occurred that populated Eurasia. (The first wave 60KYA populated only Australia.) That would make language somewhat older than the cave paintings.

Actually we have quite good archeological evidence. The Olmecs founded their civilization around 1500BCE, many thousands of years later than Mesopotamia.

The Maya arose a few centuries later, as did other cities in Mesoamerica. There was considerable cultural exchange so for our purposes they can be regarded as a single civilization. The Olmecs declined in the last centuries BCE and the Maya became the dominant culture in the region. The Maya developed the only written language in the Western Hemisphere.

Maya civilization collapsed around 900CE--some say they clear-cut too many forests to build their temples and changed the region's ecosystem so it couldn't support the population. The Maya people survived and their northern cities remained in use. They still exist today, along with their languages.

The Aztecs migrated into the region (some evidence suggests they came from California) around 1300CE and became the political rulers. Like the Phoenicians, they assimilated the nearby cultures, and they adopted the Mayan written language. The Aztecs were in the Bronze Age but had not yet invented iron technology when the Christian armies came to obliterate their civilization. Like the Maya the Aztec people still exist, although they intermarried enthusiastically with the other peoples. Their language, Nahuatl, is still widely spoken.

Olmec-Maya-Aztec can be considered a single continuum of civilization, which began 3500 years ago. This makes it one of the youngest of earth's six independently created civilizations. Only the Inca was founded later, around 1200CE.

Remember that the New World has only been populated for about 15,000 years. At least three language families have been identified, depending on which linguist you ask. This might suggest that the languages were invented here, except for the fact that there have been multiple (and poorly defined) waves of migration from Siberia, which could mean that different Siberian immigrants brought their own languages.

In particular, evidence has found that may link Na-Dene, the family that includes Navajo, Apache, Hupa, Tlingit and a great many other languages of the western USA and Canada, to the Yeniseian family of Siberia. Many linguists find this evidence persuasive, and AFAIK no one has yet stepped forward to refute the hypothesis.

The Na-Dene people are generally (but by no means unanimously) thought to have arrived in Alaska about 6000BCE. If there is a Dene-Yeniseian language family I think it will be the oldest one we have evidence of.

 

A language is an arbitrary set of symbols used for encoding information. This includes spoken and written language, but also sign language, mathematical notation and C++. The "arbitrary" part is important because a language is invented; the communication of sparrows, using a standard set of sounds every member of their species knows instinctively, does not count under this definition.

Obviously the distinction between a language and a code will be somewhat arbitrary. A phonetic writing system, no matter how perfectly its set of symbols maps one-on-one to the set of phonemes of the corresponding spoken language--for example, Finnish--is still called a "written language," rather than a "code," yet Morse and semaphore, which use the exact same set of letters in a different medium, are called codes. Meanwhile, ASL and other deaf sign languages are called languages; but that may be justified because they do not exactly duplicate the grammar of the spoken language, e.g., ASL has no verb tenses.

One language is distinct from another so long as the speakers cannot understand each other without significant study. If they can with only a little effort, then they are speaking two variants of the same language called "dialects."

A cant is a variant of a language (or combination of two languages) deliberately crafted to thwart understanding by outsiders. The best-known cant is Pig Latin, whereas the most successful is Shelta, the speech of the Irish Travellers.

A "dialect continuum" can exist, in which the people of two neighboring regions can understand each other, but communities farther apart cannot. The two variants at the end are generally counted as distinct languages, and the variants in the middle may be classified as dialects of one or the other for cultural or political more than linguistic reasons. This is true of German and Dutch.

Delimiting languages over time is more difficult than over space, because they change gradually. As has been discussed in this thread, if you arbitrarily pick any 200-year period in England since the Anglo-Saxon invasion, the people at the beginning and end of that period would be able to understand each other. Yet speakers of Modern English and ancient Anglo-Saxon could not possibly converse. We have divided this time continuum into Anglo-Saxon, Middle English and Modern English, since all three variants of English are mutually incomprehensible.

It's all an act, don't be intimidated.

Please Register or Log in to view the hidden image!