The article actually originated on Aish HaTorah's website and can be read here. The main points of the article are as follows:
- Moshe wrote the Torah.
- This Torah (or, perhaps later on, others based on this Torah scroll) were kept in the Bais HaMikdash as a model and standard.
- New Torah scrolls that were written would be checked against this Torah.
- Sofrim (scribes) were very careful not to add/delete/change anything since any change makes a Sefer Torah invalid.
- The Torah has a built-in "security system" that prevents invalid Torahs from being used.
- The end result is that, as of today, the only variant that exists of the Torah is the Yemenite Torah, which has nine minor spelling variations from the "standard" version. These variations are all minor spelling differences (as British spelling differs from American spelling) and do not change the meanings of any words. Otherwise, every Torah we use today is the same letter-for-letter that God gave to Moshe.
- This is very impressive because, compared with the Christian Bible (what is commonly called "The New Testament") the Torah is remarkably stable. The Christian Bible has well over 200,000 variant letters. We have nine.
Therefore, you can rely, with a high degree of confidence, that the Torah that we have today is *exactly* the same as the one that Moshe left for us at the end of his life.
The problem with all this is that most of those points are either exaggerations or just plain wrong. Let's go through these points and examine them.
I'm going to grant the author of the article the first two points as given. If we don't agree that Moshe wrote the Torah, then there is really no point in the rest of the article. I am also going to assume that he did, in fact, leave a Torah as a standard.
However, it becomes clear that, at some point, that standard became corrupted. For example, consider the event recounted in Meseches Sofrim. In it, Reish Lakish recounts that three Sifrei Torah were found in the Bais HaMikdash:
One book was called "The Ma'on Book." The reason it was so called was because Devraim 33:27 started out with the word "Ma'on." In the other two, it started out with the word "M'onah."
The second book was called "The Zatutei Book." It was so called because in it, the text of Sh'mos 24:5 says "And he sent to the 'Zatutei' (young men) of the Children of Israel..." In the other two books, the word "Na'arei" replaced "Zatutei."
The third book was called the "Hee Book." It was so called because it had one set of variant spellings of the word Hee in Hebrew, while the other two had a different set of spellings**.
In each case, in establishing the correct reading, the Sages followed the majority. They rejected the reading of "Ma'on" and instituted "M'onah." They rejected the reading of "Zatutei" and instituted "Na'arei." The rejected the spellings in the Hee Book and accepted the spellings in the other two books. Those readings became the standard and, indeed, are in our Sifrei Torah today.
There are several points that need to be made about this story.
- The first point to be made is that there was no single model text that could be used to check against. Indeed, these three texts *were* the model texts that were used. These were the Sifrei Torah that were found in the Temple Courtyard. If there was an alternate authoritative text, the Sages could simply have consulted it to determine the correct text for each of the three cases. The sad fact, however, is that there was no single authoritative text to compare these to -- these *were* the authoritative texts -- and now they were at variance with one another. As a result, the Sages had to establish the correct text and, in each case, went with the majority.
- The second point to be made from this story is that the so-called "built-in security system" failed... and failed miserably. It's one thing if an error creeps into a text in a backwater shul somewhere where perhaps only a few people were even capable of reading the sefer and where, if an error is found in the book, it could be isolated. This, however, was an entirely different matter. Here, textual variants are showing up in the model texts themselves. And, I'd bet dollars to donuts, that these variants didn't just show up in only these three books. I highly doubt the Sages woke up one morning and decided to check the Temple scrolls against each other just for the heck of it. I'd be willing to bet that they were getting numerous reports of variant readings and needed to investigate. And, furthermore, I'd be willing to bet that after an informal survey of the scrolls in their own personal libraries and in the shuls and study halls in Jerusalem (which were probably used on a daily basis), they found variant readings too -- otherwise, why start comparing the Temple scrolls against each other? So, they went to the Temple to get the authoritative reading, and found that even there, there was no single text. Clearly, when the authoritative texts have variants, the "security system" has failed.
- The third, and perhaps most startling point to be made is this -- at the end of the story, we find that *none* of the three authoritative model texts was kosher! Every single one of them was invalid. One had a variant reading in Devarim, one had a variant in Sh'mos and one had variants in the spelling of Hu/Hee. But *none* of the three had the text that we have today! In other words, in the end, the standard text that we have today was based on a combination of these three texts.
You might think that this settled the matter and that, at least, from this point onward, we would have a unified standard text. Alas, such was not the case. There are several places in the Talmud where the Gemara quotes a different text than the one we have. One of the more famous examples is the Gemara in Sanhedrin where one of the three reading of the word "Totafos" has an extra vav -- and that extra vav is used to help determine that there are four compartments in the Tefillin Shel Rosh. However, in the end, even the Gemara attests to the fact that we don't necessarily have accurate spellings for all the words in the Torah. The Gemara states that we are not expert in chasser and malei (i.e. words that have "extra" letters to represent vowels).
The Rambam, in the 12th century, famously went to view the bible today knows as the Aleppo Codex (also known as the Kesser Aram Tzovah), to determine the standard text and spacings in the Torah scrolls. It should be noted that, obviously, the Rambam did not have a scroll at his disposal which he considered authoritative enough.
Likewise, the Rav Mair HeLeivi Abulafia (13th century) writes in the preface to his work Mesores Siyug L'Torah that in his day there were doubts as to the correct reading. He, like the Sages in the Temple, relied on a "majority rules" principle to establish the text which he published in his sefer. Likewise, Yaakov ben Chayim (early 16th century), who published the first edition of the Mikraos Gedolos, noted that there were variant readings in his day. Nonetheless, with the adoption of the Mikraos Gedolos and the invention of the printing press, a standard text was finally adopted.
But even that's not the end of the matter. As noted above, the Yemenites have a slightly different Torah than we do. Although the author claims that the differences are only spelling there is at least one case where the spelling does change the meaning of a word (from a singular to a plural). Furthermore, even setting aside the Yemenite Torahs, there is still at least one textual variant extant today -- the final letter of the word "Dakah" in Devarim 23:2 is spelled in some Sifrei Torah with an aleph and in others with a Heh. But aside from these few cases, the text that we (finally) have today has been standardized.
The article tries to make the case that we can authoritatively state that our Torahs are accurate (vis a vis the Torah gave to Moshe) because of the traditions of the scribes. For example, the article makes the point that there was a Torah in Jerusalem that was used as the model against which others were judged. We've already seen that the model wasn't always accurate either, but let's put that point aside for the moment and assume, for the sake of argument, that the model is 100% accurate. There are still several assumptions that are being made by the author of the article that are not, in fact, in evidence:
The first point to be made is that a model text is only good if it's actually used. There is no indication anywhere that in the centuries after Moshe that scribes and other people *routinely* brought their Torah scrolls to Jerusalem to check them against the model. It's not very difficult to see how an error can creep into a sefer and stay there. Likewise, it's not too difficult to see how an inaccurate version can be copied to other texts. In a place where there aren't very many Sifrei Torah circulating about (as you can imagine would be the situation in Israel between the time of Joshua and the Exile), it's very easy for an inaccurate text to be copied to another one.
Just consider the three variant scrolls that were eventually found in the Temple. Do you think that they were first scrolls to have those variations? Or is it more likely that they were copied from other variant scrolls? I would argue the latter -- especially if you're also going to posit that scribes were generally very careful with their work.
Another assumption that is being made by the author is that the laws regarding the writing of a Sefer Torah (i.e. pronouncing each word out loud before writing, not writing by heart, etc.) were always the same as they are today AND that those laws were universally observed. Neither of those (and certainly not in latter) can be said to be true 100% of the time. It's certainly not inconceivable that there might have been scribes who were less than scrupulous with their work and did, indeed, introduce errors into their work.
Lastly, the author tries to make the comparison between the accuracy of our Torahs and the Christian Bible. He states that there are only nine spelling variants extant today, while showing that there are thousands of variants of the Christian Bible.
I'm not an expert in the Christian Bible, so I can't speak to that point directly. But what I do know is that the author is making a false comparison. The author, in making his point, is outright dismissing any known variant text to our Bible (Yemenite Torahs excepted). He's conveniently forgetting that there are variations of our text that do exist -- and they're still around today. The Samaritan Bible, the Septiguant, the Dead Sea Scrolls, et al are all still extant and can be read to this very day. By forgetting them (or, more likely dismissing them), the author is engaging in a form of the "No True Scotsman" fallacy. In short, he's claiming the to reject any variants he doesn't like as non-authoritative while, at the same time, holding Christians to task for all their variants.
Of course, Nate makes the very same error when he says "UNDISPUTED" regarding the article. What he means is "undisputed by anyone who agrees with it," which, again, is a form of the No True Scotsman fallacy. But it's pretty clear that it can, indeed be disputed whether or not the Torah text we have today is a letter-for-letter copy of Moshe's.
* The author of the article does acknowledge that the Yemenite Torahs are different than ours in nine places.
** In later writings, the word "Hee" is written Hey-Yud-Aleph. However, in the Torah, it is often spelled Hey-Vuv-Aleph, the same as the word "Hu." However, there are a number of places where the Torah uses the first spelling. The differences between the scrolls was in where the exceptional spelling was used.