Many people know the Biblical tale of the Tower of Babel, when God broke apart the world's singular language into 70 different branches.

Most linguists don't give this a second thought, or anything more than folklore.

But I've read some very convincing literature that traces a large portion of many of the world's great languages back to one parent language that is very similar to modern Hebrew. The evidence is pretty convincing.

For example the English word "DiReCtion" is very similar to the Hebrew word "DeReC" meaning "route". Another basic example is the Japanese word "SaMuRai" which is very similar to the Hebrew word "ShoMeR" which means "watchman".

More example are here: http://www.edenics.org/languages-3/

My question is, what is the reason this theory is not more widely recognized? Is it because it tries to legitimize something from the Bible and is therefore seen as unscientific?

  • 15
    Undid the downvote, and humbly request this question not be closed. I may disagree with Edenics and its presuppositions, but then I believe this question was asked in good will. Plus I think it's important questions like this get real, non-patronizing answers as to why Edenics gets rejected - else some people might cry "censorship!" and claim linguists are part of an über big anti-God conspiracy, along with physicists, biologists, etc. – Joe Pineda May 3 '14 at 20:09
  • 2
    Thank you Joe for the sanity and for understanding where I am asking from. – CodyBugstein May 3 '14 at 20:18
  • 4
    You are using the 'romanised' form of Japanese syllables... that's your first mistake. Not taking into account phonetics is your second mistake. – deutschZuid May 16 '14 at 4:06
  • 8
    It would be particularly bizarre for the Ursprache to be more similar to Modern Hebrew that to Biblical Hebrew. – user6726 Oct 24 '16 at 23:13
  • 2
    I don't understand the downvotes. I think people think they are downvoting Edenics, but they are actually downvoting a question about its legitimacy, which is a valid thing to be curious about. – CodyBugstein Oct 27 '16 at 0:23

Modern linguistics does not rule out the possibility that all languages of the world descended from a single language. But the mainstream consensus seems to be an agnostic one: Most think that this hypothesis is not testable, at least not with the current data.

Languages change. More you try to track those changes back in time using scientific methods (comparative and internal reconstruction), less you become certain about the properties of the proto-language.

For instance, now linguists have a pretty clear picture of the properties of the Proto-Indo-European language, the ancestor of English, Spanish, Russian, Hindi an many others, conjectured to be spoken some time between 7th and 4th millennia BC. But there are many unresolved problems and uncertainties.

These uncertainties cause problems in so called long range reconstructions. When you try to apply the comparative method to proto-languages, most of the time you come up with things like "the word for 'foot' must be a consonant followed by a vowel followed by a consonant followed by a vowel (CVCV)". Obviously it doesn't help much.

So, most mainstream linguists seem to be inclined to think that these uncertainties make long range reconstructions unfeasible past a certain point in time. But some linguists do nevertheless work on long range reconstructions using more or less accepted methods. (check here for example)

Besides, we have very good reasons to conclude that Hebrew cannot be the proto-world language. It's a part of the well-studied Semitic language family. It has current (like Arabic) and historical (like Akkadian) relatives. There is ample evidence that together they go back to an older proto-language: Proto-Semitic. And there is also evidence that Proto-Semitic, in turn, goes back to an even older proto-language called Proto-Afroasiatic.

The "convincing" examples you give are not convincing at all. The English word "direction", for instance, has a well established etymology that goes back to Proto-Indo-European, through French and Latin. Those are chance resemblances. If you scan enough words, you're eventually going to come up with some words with somehow similar meanings and that look somehow similar between any two languages. So, this definitely isn't a scientific method. Scientific comparison requires regular correspondences. Many of them. Try to find some material on how Proto-Indo-European was painstakingly reconstructed by linguists with more than a century's work.

In short, no, the Edenics is not rejected because it tries to legitimize something from the Bible. It's rejected simply because there is no scientific evidence that supports Edenics and there is ample scientific evidence that rules it out.

| improve this answer | |
  • "The English word "direction", for instance, has a well established etymology that goes back to Proto-Indo-European, through French and Latin. " But we have manuscripts that date back 3000 years containing the Hebrew word "DeReC" so we know that it preceded Latin by a long shot. – CodyBugstein Apr 29 '14 at 14:53
  • 12
    There are no Hebrew manuscripts that "date back 3000 years". The earliest surviving manuscripts in Hebrew are from about the time of Christ. And Proto-Indo-European is a lot older than that. – fdb Apr 29 '14 at 15:19
  • 19
    But even if there were, this would be irrelevant. Direct, or its Latin ancestor, has a transparent etymology in Latin of di or de + regere. There is nothing in it which could correspond to the Semitic root DRK. – Colin Fine Apr 29 '14 at 17:22
  • @ColinFine Other than possibly some inherent synesthetic connection between the sound derek and direction, which would act to reinforce both de + regere and the Semitic root. – Damian Yerrick Dec 10 '15 at 19:12

As an addendum to cyco130's excellent answer, it's easy to see based on simple math that resemblances like the ones you cite can tell us nothing about language relationship.

From looking at some of the materials in the website you link to, the procedure seems to be: (a) ignore vowels; (b) treat consonants as interchangeable if they are members of the same phonetic class, of which there are six (bilabials, dentals, gutturals, fricatives, liquids, and nasals). (You can also metathesize or add nasals to your heart's content, but I'll ignore that.) This means that there are only six distinctive elements which can be compared. So if we're looking at a word that contains three consonants, there are a total of 6x6x6 = 216 possible types of such words. Within each type, all the words could be related to each other by this method.

Now how many words are there in the world's languages? At a very conservative estimate, let's say there are 4000 languages, each with a lexicon of 10,000 words. That's 40,000,000 words. Each of the 216 classes would then contain an average of 40,000,000/216 = 185,185 words, all of which would be considered "similar to each other". Now all you need is to find some pairs of these words that are also vaguely similar in their semantics, and you have things like shomer - samurai. As many of them as you want, between any two languages you choose.

Even if you make the phonetic rules stricter, so that you can only compare d with d and not with t and so on, the math will still mean that most of these resemblances will be due to chance.

This is why Edenics is not recognized as a serious linguistic theory, and also why scientific historical linguistics does not work with resemblances at all, but with regular correspondences, which are a completely different concept.

(By the way, the older form of the word samurai was saburai, which looks a lot less like shomer. This is usually the case with these accidental resemblances: the more you know about the words' histories, the less similar they look.)

| improve this answer | |
  • 4
    Great answer, TKR. Deriving Proto-World with tools you probably have at home and [How likely are chance resemblances between languages? ](zompist.com/chance.htm) have been my go-to link for cases like this, if somebody should be in need of a smaller tea spoon. – kaleissin Apr 30 '14 at 19:08
  • 5
    @Imray Well, let's keep doing the math. Say there's a 1/1000 chance that two randomly given words will have meanings as close as those of shomer and samurai. (I bet the chances are actually better; those words' meanings aren't really "nearly the same".) That means that by this calculation, for any word you choose in any language, there are likely to be 185 words in other languages that by pure chance happen to be similar (by these criteria) in both sound and sense. (In reality there are probably more, as I'm using low estimates of both total vocabulary and chance of semantic similarity.) – TKR May 4 '14 at 0:37
  • 6
    Regardless of the details of the computation, the real point is that the Birthday Paradox applies. It is the mathematical reason why there is a high probability of weird accidents such as accidentally meeting your neighbours on a different continent so long as you don't try to fix time, space and kind of the accident beforehand. It is also why linguists had to learn to be very careful when trying to prove relations between languages. Without extra care, they can be found between every pair of languages. – user4938 Apr 7 '15 at 8:05
  • 1
    @CodyBugstein - The problem is that meaning veries wildly with time and place, and, on contrary of phonetics, there are no known ways to predict regular changes. Thus, the English word "loot" is a borrowing from Hindi, and does not point to any "proto-Anglo-Hindi". The English cognate for Hindi "lu:t" is "leaf". Now how does one predict that a word meaning violent ramsack is cognate with a word meaning, well, leaf? – Luís Henrique Oct 24 '16 at 12:50
  • 4
    "Even if you make the phonetic rules stricter, so that you can only compare d with d and not with t and so on, the math will still mean that most of these resemblances will be due to chance." - And then you will miss actual relations, such as English "door"/German "Tur", which are effectively cognates. The impressionistic "method" not only gives many false positives, but also false negatives. Who would say that Portuguese "joelho" and English "knee" are related? – Luís Henrique Oct 24 '16 at 16:09

The reason "Edenics" isn't taken seriously is that it isn't, by scientific standards, serious.

To establish a kinship between two languages, linguists need to have whole series of words that can be traced to a common origin.

So, if we did not know Latin, we could reconstruct its lexic from French, Castillian, Catalan, Italian, Portuguese, Rumanian, etc.

For instance:

Here are the words for "horse" in those languages:

Portuguese: cavalo

Spanish: caballo

Italian: cavallo

Rumanian: cal

French: cheval

Catalan: cavall

From here, how can we find the Latin word for horse?

It seems that it should start with either "c" or "ch". But perhaps this is a coincidence? How can we know if it is not?

Because we have lots of words that begin with "c" in Italian, Portuguese, Castillian, Rumanian, etc., and with "ch" in French:

Portuguese - cabra, caminho, canção

French - chèvre, chemin, chanson

Which shows that there are regular transformations between French and Portuguese; some initial consonant in Latin morphed into "c" in Portuguese, Catalan, Italian, Rumanian, but into "ch" in French (and if we look at a map, we see that Portuguese and Rumanian are placed, respectively, at the Western and Eastern extremes of the geographical extension of Romance languages, so it isn't likely that the same innovation happened to both of them; it looks like the Latin consonant was preserved in both - and transformed in French.

The same reasoning helps us establishing that the first vowel of the Latin language should be an "a", not an "e". The second consonant is more difficult, because it went three different directions, morphing into "b" in Castillian, disappearing in Rumanian, and morphing into "v" in the remaining languages. But other considerations have led linguists to believe it was a "b" in Latin. And so on, so that we can reasonably guess that the Latin root for "horse" is something like "caball-".

We do know Latin, however, and we can see that the Latin word was "caballus" (OK; it wasn't; it was "equus" - which gave us the Portuguese, Castillian, Rumanian and Catalan word for "mare". "Caballus" was Vulgar Latin, and comes from the Classic word for "horse pack", which in turn is a borrowing from unknown origin).

Now compare this with the pair "direct"/"d-r-k". Which other English words correspond to a Hebrew (or other Semitic) root filled with "i" and "e"? Or, the other way round, which other English words, stripped from its vowels, correspond to a Semitic three-consonantal root? Do we find that "derelict", or "derailed", "deranged", "correct", "erect", "ferret", correspond to Semitic roots "d-r-k", "d-r-l", "d-r-g", "c-r-c", "r-c-t", "f-r-t", etc.? I do not know Hebrew, but I bet you that none of these happens, with the possible exception of "erect"/"eretz", where anyway the semantic relation is quite loose (upright vs great). And that is not random: it is a consequence of English and Hebrew not being related in any knowable way (they are possibly related, perhaps through a proto-language of both Proto-Indo-European and Proto-Afro-Asiatic - but we do not have the means, as things are, to reconstruct such proto-language). However, we do know the origin of the English words I listed above; they are borrowings from Latin or Old French; and the obvious relation between "direct" and "correct" can only be explained by the fact that "rect-" is a Latin root, and "di-" and "co-" are Latin prefixes. (If "direct" was related to "d-r-k", then "correct" is unexplainable, for in Semitic languages both words would be either unrelated or share the same three-consonant root - and then the English word for "correct" should be something like "daruct" or "deroct", which evidently isn't the case.)

In fewer words, bats are not related to nightingales, albeit the obvious fact that both can fly; on the contrary, nightingales are related to ostrichs, who don't. The flying is coincidental; the egg-laying vs breastfeeding is not.

| improve this answer | |
  • 1
    Just for context, "direct" (adjective) is shr in Hebrew; "direct" (verb) is lehorot. "Route" is maslul, nativ, afik, tur, orach, and, uff, also derech. So Hebrew bought six ticktets, not one, for this particular lottery - and all the others seem completely unrelated to English, except perhaps "tur"/"tower" (or would that be "tur"/"door"?) At this level of imprecision, practically anything goes. And that is why it isn't considered serious, or even convincing. – Luís Henrique Oct 24 '16 at 15:48
  • Also note that this cuts both ways. Consider the word "cabal". It means the actions of of a group of persons secretly united in a conspiracy, or the people involved is such actions. In the latter acception, it is a collective: cabal = group of conjurors. And, oh, Latin "caballus" was also a collective - a group of horses. Could that be a coincidence? According to those who don't believe in coincidences, no: it means that "cabal" and "caballus" are related, so "cabal" must be a borrowing from Latin. Unfortunately, we know that "cabal" is a borrowing... from the Mediaeval Hebrew qabbalah... – Luís Henrique Nov 7 '16 at 19:31
  • ... probably through Castillian cabala. And the original Hebrew word qabbalah meant tradition, received lore, not a collective of anything. So words change their meaning in unpredictable ways, and coincidences between sounds do exist. Only a proper comparative method can elucidate what is coincidence and what is relation. And "Edenics" relies in a completely speculative method - that of cherry-picking similarities between isolated words. But once we attempt to build any system from these coincidences, we fail: coincidences, unlike actual relations, do not happen in a systematic way. – Luís Henrique Nov 7 '16 at 19:38

The idea that all languages derive from Hebrew has been around for a long time. It is a theological construct, not a scientific theory. By the way, the tower of Babel story (Gen 11,1-9) does not say anything about “70 different branches”, nor does it claim that Hebrew was the primeval language.

| improve this answer | |
  • 1
    In my question, I didn't say that all languages came from Hebrew, but perhaps from a language close to Hebrew – CodyBugstein Apr 29 '14 at 14:53
  • 3
    You did say that, but "a language close to Hebrew" would be some known or unknown Semitic language. As cyco has explained the idea of a Semitic origin for all languages of the world is linguistically impossible. – fdb Apr 29 '14 at 15:10
  • 1
    Absolutely right! No mention at all in the Bible of any number of branches! Wish I could upvote this answer more times. – Joe Pineda May 3 '14 at 20:12

Though cyco's answer mostly says it all, wanted to add my 2 cents. How can you distinguish between a scientific approach and one that isn't? If you've read Karl Popper, you know the simplest litmus test is that of "falsifiability": for every theory, there must exist at least one possible experiment that could put the theory in the garbage bin by yielding a result contrary to what was expected.

"Edenics" and its kin (Intelligent Design) are for the most part un-falsifiable. Since they're based off faith, you can get away with any illogicity or irregularity in them, as well as any deviation between their predictions and reality by simply claiming such is God's will. Nice if you want to make a theological point, but useless if you want to do science - we humans do science with the aim of understanding our world and trying to predict events in it.

The reconstruction of proto-Indo-European, as well as proto-Semitic and other last-common-ancestor of various language families has been the result of linguists applying methodic analyses of data, looking further than simple commonalities but to the underlying causes of them and searching for regular processes that could explain them.

Claiming everybody spoke a proto-Semitic language and then, ex-nihilo, full blown and totally different grammars appeared in less than a generation with no solution of continuity is a far-fetched proposition that requires a lot of hard data to be demonstrated - difficult at best. Except if it's said as a statement of faith in which case no proof is required.

| improve this answer | |
  • We know that it's possible for languages to change drastically in a short period of time. There are many examples throughout history – CodyBugstein May 3 '14 at 20:27

The key advance in the study of the history of languages was Rasmus Rask's suggestion in his 1818 Prize Essay that Danish, other Germanic languages, and Latin and Greek, were descended from a language which perhaps no longer existed. This explained why previous attempts to derive the various modern languages from Hebrew or some other known language had been fruitless.

I don't agree with some comments here that the Hebrew-as-progenitor theory is inherently unscientific. In its day, before the 19th century, it was a plausible theory to explain the resemblances among languages that people had begun to notice. It was superceded when scholars realized, under the influence of the developing sciences of geology and Darwinian evolution, that large differences among contemporary languages could result, over long spans of time, from small changes of a sort that can still be observed.

We now tend to trace the historical development of language systems (sets of phonological rules and so on) rather than individual words.

| improve this answer | |
  • 5
    "Edenics" is a modern theory, promulgated by that particular wingnut web site. It is perhaps inspired by the Babel myth, and it didn't exist as a theory in the 19th century.. Before praising the virtues of this blatant rejection of science, I'd suggest determining first what the theory is. – user6726 Oct 24 '16 at 21:00
  • 1
    @user6726, I didn't actually look up Edenics. I've amended my answer to refer to "the Hebrew-as-progenitor theory". I hope to avoid reading about the specific modern version. – Greg Lee Oct 24 '16 at 22:48

Not the answer you're looking for? Browse other questions tagged or ask your own question.