Arutz Sheva and The Great Hebraic Hoax

Now, I love Arutz Sheva Radio as much as the next hot-headed right-wing extremist settler wannabe, but some of the guests they invite on, well, make my orange ribbons pale.

The perpetrator of the hour is one Jim Long, who came on the Tamar Yonah show as a representative of Isaac Mozeson, author of the new book The Origin of Speeches: Intelligent Design in Language, and an older one, The Word: The Dictionary That Reveals The Hebrew Source of English. Listen to the interview (ASX / MP3 / main audio page) [no longer available], if you want to be taken for fool. Otherwise, don't listen, unless you're definitely going to read the rest of this post. If you've already had the misfortune of hearing it, please read on.

Before I comment, I'll briefly disclaim: I have not read the new book, or even glanced at it. But I did hear the interview, and I did once or twice have the displeasure of perusing The Word, and I heard nothing in the interview that was qualitatively different from what I read there. And, for my own credibility, I'll mention here that I have a degree in linguistics.

Here is the essential background information you need to know:

• Languages are constantly changing. Everyone learns a language and uses it in a way that's slightly different from the way the older generation spoke it, and the way his peers speak it. This happens naturally, even in the absence of conscious innovations or outside influences.


• When speakers are split up geographically or even sociologically, their modes of speech will begin to diverge to a greater degree, because there's not enough communication to keep them in sync.


• Once these varieties are different enough, we deem them separate languages and give them distinct names. We refer to the newer varieties as sisters of each other, and descendants or daughters of the earlier form.

Mozeson has set out to prove that all languages are descended from Hebrew, and so he points out one pair after another of Hebrew and English words (and occasionally words from other languages) that are similar in form and in meaning. Like a magician pulling rabbits out of a hat, he produces Hebrew ארץ/eretz and English earth, שקל/shekel and scale, הר/har and hill. He's got tons of these, and we're supposed to gasp at the obvious correspondences, still intact after 6000 years.

If you talk to a linguist though, you'll get a very different picture. Hebrew is a member of the Semitic family, and by extension the larger Afroasiatic family. It's related as a sister or a cousin to languages like Aramaic, Arabic, and Amharic. Hebrew shares with these a common ancestor, and there's no other language that's said to be a descendant of Hebrew in the sense I described above (Yiddish, Ladino, etc. merely have loanwords from Hebrew). Then there are numerous other families of languages, with no known relationship to Semitic. English is part of Germanic, which itself belongs to a larger family called Indo-European.

What's going on here? Is the academic world set to be turned on its head? Has Mozeson discovered in his straightforward pairings something that generations of trained linguists, who spend sleepless nights absorbed in works with titles like Graded Effects of Verb Subcategory Preferences on Parsing: Support for Constraint-satisfaction Models and Excrescent schwa and vowel laxing: Cross-linguistic responses to conflicting articulatory targets, have somehow overlooked?

Why the answer is no is largely explained in the classic article How likely are chance resemblances between languages? by Mark Rosenfelder. Chance resemblances between languages are very likely, and so Mozeson-type similarities don't really tell us anything. Read the article. Even if your eyes glaze over at some of the linguistic terminology as mine do at some of the math, he's a good writer and you'll get the point, and you'll learn a lesson in caution that can apply to many areas in life.

Exactly how likely chance resemblances are depends not only on the languages being compared, but also on how careful or lax the investigator is. And Mozeson takes all the shortcuts he can find:

• Ignore vowels. Of course it's much easier to make a match when you only look at half of each word.


• If a consonant gets in the way, drop that too!


• Be flexible with the meanings of words. If hand doesn't sound right, try arm, finger, grip, give, etc. [NB: there are real cognates whose meanings have diverged much more than this, but they are bonuses to be claimed after the relationship is already established through firmer evidence.]


• Pick English as one of your languages. It's got the largest vocabulary in the world, and so you don't even have to be so flexible with the meaning; you've probably already got 50 very close synonyms for whichever word you need. At least one of them will probably sound right.


• Count loanwords, even though they show nothing about the origin of the language you're looking at.

Anyone one of these methods is liable to lead one's research astray. Employing all of them together, indiscriminately, is a surefire way to get a false positive.

But the carelessness and confusion run much deeper: not only is Mozeson too forgiving with his evidence, he is looking for the wrong kind of evidence, and in all the wrong places.

When sounds in a language change, they change systematically (most of the time). For example, every [h] sound in Latin disappeared along the road to becoming Spanish (that it's still written is irrelevant); today in many varieties of Spanish [s] is turning into [h] whenever it occurs at the end of a syllable, but not at the beginning. If you find that a [g] sound in Language A corresponds randomly to a [k], [h], or [g] in Language B, you haven't proven anything. But if [g] corresponds to [k] at the end of a word, [h] before a vowel, and [g] elsewhere, then it's systematic and you may be on the way to establishing a relationship. From correspondences like that you can posit rules to describe the changes that likely took place between the source language and the daughters, e.g., "[g] becomes [k] at the end of a word."

Of course, the longer the time gap you're looking at, the more opportunity the languages have to change, and change again, until they look extremely different. Even the correspondences are obscured eventually, because the conditions that are necessary to accurately describe them (such as [g] corresponding to [h] only before a vowel) may themselves change or disappear (say, if some vowels are later dropped).

A couple of examples should suffice to illustrate this: Spanish hoja and French feuille ('leaf, sheet') have not one sound in common, but both come to us from the same Latin folium. Similarly, we have Spanish hecho and Romanian fapt ('done'), both from Latin factum. These forms can be derived through regular, language-wide changes, and though they're very different, this is just what we'd expect after two thousand years.

Finally, one mustn't forget that as sound changes are taking place, words are constantly being coined, borrowed, abandoned, and can easily change their meaning beyond recognition; thus the pool of words that we can use to prove relationships grows ever shallower. For all these reasons, historical linguists focus on the oldest varieties of language available, those least altered from the source in question.

Contrast them to Mozeson, who doesn't even deign to use Old English for his comparisons, let alone Greek, Latin, or Hittite, but chooses English as it is currently spoken. By comparing two languages so far apart in time, one of which is known to be singularly infested with loanwords and recent sound changes, Mozeson is setting himself up for failure.

Mozeson never posits any sound laws with which one could derive English from Hebrew, nor do his examples differ to the degree we'd expect after so many millennia. Neither do his linkages imply any sort of directionality; even if they were completely valid, they don't favor English as a descendant of Hebrew, or Hebrew as a descendant of English, or both as descendants of a common source. We can at least rule out the second option because we happen to know that English did not exist in anything like its current form until recent times, but he'd have no such luxury if he'd chosen an older tongue.

What Mozeson has done, then, is no more and no less than to tell us that 1 plus 1 equals 11. To the untrained eye it's compelling, but it's true only in the most cosmetic and uninteresting sense. Using the same methods he could have shown German to be descended from Hawaiian, Russian from Klingon, or Hebrew from !Kung. And it should surprise no one that claims like these are regularly made for many a language, by many a well-intentioned naif.

The reason I felt the need to write this post is that unlike math, science, music, or politics, very few people study even the most basic principles of linguistics. People are often not even aware that linguistics exists as a discipline distinct from language learning, or else fail to grasp the depth of it. Many will spin off pseudo-linguistic theories as a child would paint a scene, unaware of the importance of perspective or subtlety, and many others will swallow whatever linguistic tomfoolery they are told to, with neither question nor dissent.

It both offends and saddens me particularly when I see such erroneous thinking in a Jewish context, because my own and many other people's religiosity is the result of a long and taxing search for the truth, and so much of Judaism is the very process of rational thought and debate. Yet, theories like Mozeson's have few challengers within Judaism, not only because people are ignorant but because everyone wants them to be true, and when I see how easily and willingly people are misled it makes wonder how exceptional this really is.

Before I close, I want to make clear that I have no personal resentment towards Arutz Sheva, Tamar Yonah, Jim Long, or even Isaac Mozeson. I'm sure the latter has no intent to mislead or deceive, and he probably even thinks he's doing a great mitzvah. I believe, though, that he is unwittingly doing Jews a great disservice by enticing them to greater faith through falsehood, and it's that falsehood that I am opposing here, despite the repeated mentions of his name above.

For more on this topic, have a look at Maven Yavin: Edenics--was Hebrew the original language?, and the negative reviews for The Word on Amazon.

---

---

 
Posted at 8:44 AM / Permalink /