Shirat Haazinu: 70 or 67 lines?

One of the few columns that stand out in a Sefer Torah is Shirat Hayam, with its “brick and mortar” shape, and Haazinu, with its “two towers” shape. As a general rule, the Torah Scroll has small blank spaces scattered around every column, and they serve to delineate paragraphs and to provide to the reader a moment of reflection.

I have written about the importance of the correct placement of these blank spaces, called in Rabbinic parlance Parshiot Petuchot and Setumot, in an older post and I encourage you to look there too.

But as a whole, the Torah layout is a continuous prose in all its columns, save the two instances mentioned above. Both are songs, and it seems that the unusual layout is intended to highlight their poetic structure. Commentators offer more esoteric explanations, that the two towers Haazinu layout allude to the downfall of the wicked, which are mentioned in one of the stanzas (this explanation is also applied to the two tower layout of Haman's wicked sons in Esther's Scroll, discussed here).

The Rambam dedicates many pages to the correct layout of all Parshiot in the Torah, and he writes that Shirat Haazinu should be divided in 70 lines. Look at the text in Sefaria:

צוּרַת שִׁירַת הַאֲזִינוּ - כָּל שִׁיטָה וְשִׁיטָה יֵשׁ בָּאֶמְצַע רֶוַח אֶחָד כְּצוּרַת הַפָּרָשָׁה הַסְּתוּמָה. וְנִמְצָא כָּל שִׁיטָה חֲלוּקָה לִשְׁתַּיִם. וְכוֹתְבִין אוֹתָהּ בְּשִׁבְעִים שִׁיטוֹת. וְאֵלּוּ הֵן.

That’s indeed how our Torahs (see example pic at the top of this post) are structured - both Ashkenazi and Sephardi scrolls - in accordance to the Rambam’s account and we would expect that to be the case, as the Rambam had in his possession the prized Aleppo Codex - the most authoritative codex according to our tradition.

The Yemenite Jews have a handful differences in their Mesora of the Torah text, minor differences in the spellings but one very visible variance stands out. Their parshat Haazinu is written in 67 lines, unlike Ashkenazi and Sephardic scrolls.

When looking closely, they have a different arrangement in three stanzas, which are merged together forming a longer, more squeezed, line. Because of that, the layout of their Haazinu column is much less homogenic and the “two towers” are not perfectly aligned. See a picture of the Yemenite tikkun:

We all know the Teimanim follow the rulings of the Rambam closely, which in turn begs the question - how do they reconcile their Mesora with the Rambam?

Let’s turn to the Aleppo Codex again. As I discussed elsewhere, this codex is attributed to the Masorete Ben Asher, and was salvaged from the Aleppo synagogue pillaging in the 1947 Arab protests against the establishment of the State of Israel..

The local Sephardi community guarded the Codex closely, and very few outsiders managed to find a way to look at it. One of the few was Humberto Cassuto, a famous scholar who wanted to investigate if this Codex was indeed the one attributed to the Ben Asher lineage. Professor Cassuto was granted limited access and couldn’t study it throughly, but he cast doubt at the provenance of the Codex because he saw that the Haazinu of the Codex had 67 lines, and not 70 lines as discussed in the Rambam’s Mishne Torah.

Many scholars started to investigate this finding. It turned out that the Yemenites have a different reading of the Rambam and in their manuscripts it states that Haazinu has 67 lines - just like Professor Cassuto observed in the Codex, except he wasn’t aware that his own Rambam’s edition was corrupted. The very feature Prof Cassuto found to be suspicious turned out to be the best proof of the authenticity of the Codex. An early manuscript of the Rambam from Oxford's collection has the same version as the Yemenites, and that's how Mechon Mamre has it in their online Rambam:

יא  צוּרַת שִׁירַת הַאֲזִינוּ (דברים לב,א-מג)--כָּל שִׁטָּה וְשִׁטָּה, יֵשׁ בְּאֶמְצָעָהּ רֵוַח אֶחָד כְּצוּרַת הַפָּרָשָׁה הַסְּתוּמָה, וְנִמְצֵאת כָּל שִׁטָּה חֲלוּקָה לִשְׁתַּיִם; וְכוֹתְבִין אוֹתָהּ בְּשֶׁבַע וְשִׁשִּׁים שִׁטּוֹת.  וְאֵלּוּ הֶן

Although almost all the Chumash part of the Codex was destroyed (or hid away, as claimed by Matti Friedman’s great book discussed here), the Haazinu pages observed by Prof Cassuto have survived and can be seen in the Israel Museum and online. See it here:

The Yemenites kept the Rambam’s proposed Mesora (save one puzzling, small variance towards the end of Haazinu in the stanza starting with "Gam Betula" which the Yemenites start with the preceding "Gam Bachur" - the similar words seemed to have caused this confusion but perhaps there's a better explanation I'm not aware of). 

The Ashkenazi and Sephardi did not, and there was an obvious attempt to cover up the discrepancy between their tradition (70 lines) and the Rambam’s (67), and while a few expert scholars (like 16th century Menachem di Lonzanu, in his popular work Or Torah - see here at the bottom) eventually noted conflicting versions of the Mishne Torah, this caused much confusion and eventually most scholars became convinced that the versions of the Mishne Torah with 67 lines were simply wrong because they didn't comply with the vast majority of the existing scrolls.

The Ashkenazi and Sephardi structure of 70 lines has its source in the Masechet Sofrim 12 (exact link here, where it states the first word of every line totaling 70), which is one of the handful small tractates found in the Babylonian Talmud and is generally attributed to the Gaonic period. Even though the Codex was housed in Aleppo - a major Sephardic center - for a very long time, the Sephardic world adopted the 70 line tradition which was the most prevalent and based their text in Rabbi Meir Aboulafia’s (who was an opponent of the Rambam) authoritative compendium Masoret Seyag Latorah - not the Aleppo Codex. Ironically, the Aleppo community guarded the Codex as its prized relic while following another Mesora for the Haazinu parsha (credit for the great Prof Marc Shapiro for this insight).

A few scholars have attempted to conduct studies of Torah Scrolls from different pre-war communities in regards to their Haazinu structure, in order to discover how prevalent was the 70 line structure. Scholars have found that there were more than two options - some scrolls had a little more than 70 lines while others fewer than 67, some had no unique structure at all, while others had Haazinu in the brick and mortar layout of Shirat Hayam! It seems like the scribes generally knew that Haazinu had a special layout but had limited knowledge of how to write the special structure.

The difficulty in regards to Haazinu stems from this Talmudic passage in Megillah:

אָמַר רַבִּי חֲנִינָא בַּר פָּפָּא, דָּרֵשׁ רַבִּי שֵׁילָא אִישׁ כְּפַר תְּמַרְתָּא: כּל הַשִּׁירוֹת כּוּלָּן נִכְתָּבוֹת אָרִיחַ עַל גַּבֵּי לְבֵינָה  חוּץ מִשִּׁירָה זוֹ וּמַלְכֵי כְנַעַן, שֶׁאָרִיחַ עַל גַּבֵּי אָרִיחַ וּלְבֵינָה עַל גַּבֵּי לְבֵינָה. מַאי טַעְמָא — שֶׁלֹּא תְּהֵא תְּקוּמָה לְמַפַּלְתָּן — 

Said Rabbi Hanina bar Pappa, Rabbi Shila, a man of the village of Temarta, expounded: all songs -- all of them -- are written a small brick (writing) above a brick (blank space), and a brick above a small brick, except this song (Sons of Haman) and [the song of] the kings of Canaan, which are a small brick above a small brick and a brick above a brick.

Note that Haazinu is not mentioned as one of the exceptions, and you could infer from this passage that Haazinu should be written like all other songs - in a brick and mortar fashion! That is the likely explanation of why some older scrolls have this feature - perhaps some scribes based themselves in the simple understanding of this Gemara. The Noda Biyuda discusses the Halachic status of this layout and based on this understanding he tries to find a way to not invalidate these scrolls. See below how a Haazinu in brick-and-mortar shape would look like:

However, the Masechet Sofrim is categorical, and clearly states that Haazinu is not to be written like Shirat Hayam, and that's the normative Halacha - even though the Masechet Sofrim is from the Gaonic period and hence, theoretically less authoritative than the Talmud which seems to imply that Haazinu should be written like all songs - in brick and mortar layout.

Professor Mordechai Breuer, one of the leading experts of the Aleppo Codex, attempted to harmonize the Talmudic passage above with the ruling of the Masechet Sofrim by developing the idea that Haazinu is not a real Song/Shira, and therefore not the subject of the Talmud's discussion above. In other words, Haazinu is in a category of its own and it's unlike Shirat Hayam, Bnei Haman and Shirat Devorah (see here page 23 for further discussion and a great resource in this topic).

The complexity of this Talmudic passage is the best explanation of why there's not one single option when it comes to writing Haazinu - the Talmud is ambigious and the scribes had a tough time getting it right.

However, scholarly research has shown that in both Ashkenazi (Prof Goshen-Gottstein) and Italy (see Prof. Orlit Kolodny here with more details), the most common layout was undoubtedly the 70 line structure, as per the Masechet Sofrim. Less than 10% of the 250+ scrolls surveyed have the 67 layout, which means that the Rambam/Yemenite Mesora was actually not very popular. While the Rambam tried to push for the 67 Mesora in his very detailed account of how Haazinu should be written, it seems clear that already in his time this Mesora was not dominant and his initiative did not gain much traction in the wider Jewish world. The fact that the Rambam's manuscript was censored to conform with the 70 line Mesora is an indication that there was a push back to the Rambam's directive, and the censorship (see more about this in Prof Marc Shapiro's "Changing the Immutable") was a very efficient way to safeguard the prevalent Mesora of the Masechet Sofrim - it even fooled an expert scholar like Prof Cassuto.

Rabbi Meir's Torah

"בתורתו של רבי מאיר מצאו כתוב, "והנה טוב מאד" (בראשית א, לא) - "והנה טוב מות" (בראשית רבה וילנא, ט, ה)

"ויעש ה' אלקים לאדם ולאשתו כותנות עור וילבישם" (בראשית ג, כא). בתורתו של ר' מאיר מצאו כתוב: "כותנות אור" (בראשית רבה תיאודור-אלבק, כ, כא)

This Medrash says that Rabbi Meir’s Torah had some variant readings distinct from our mainstream Mesora. Instead of טוב מאד, his text was טוב מות; instead of כותנות עור, he had כותנות אור. 

This is a puzzling and difficult concept to understand. The Talmud (Eiruvim 13a) says that Rabbi Meir was an expert Sofer, who learned by the foremost leaders of his generation - Rabbi Yishmael and Rabbi Akiva. 

While a small variant reading of עור and אור is a relatively minor issue, Rabbi Meir’s other variant - טוב מות - seems completely different and unrelated to the mainstream text. What can be the connection between טוב מאד and טוב מות, which was highlighted by Chazal as a point of variance between two traditions ?

I usually like to follow a somewhat scholarly approach in my posts, but to answer this question I will turn to Derash. 

There are many comments about  the connection between עור and אור, mostly based one the famous Zohar that originally the skin of Adam was translucent, full of light, but after his sin it turned like our skin, hence the connection between the words. 

Exploring this concept further, I've heard in the name of Reb Tzadok Hacohen |(please comment if you have the written source)  that specifically Rabbi Meir had the unique ability to understand the ultimate purpose of everything in this world and how all connects in a meta-physical reality. In his perception, כותנות עור was very clearly not just leather clothes but clothes hiding a spiritual light and Rabbi Meir could perceive that in all creation at any given time - not only before Adam’s sin. For Rabbi Meir, all creation was connected and he saw how that worked.

What about the connection of מאד and מות? 

If we take Reb Tzakok’s insight a step further, that our traditions and Rabbi Meir's reflect two different worldviews, let’s analyze why this specific variant has been highlighted. Both מאד and מות start with the Mem, the middle letter of the Hebrew alphabet and the letter representing the present time. We can see that Rabbi Meir could start from the Mem and perceive the very end-objective of everything, and this is codified in the word מות, going from the Mem directly to the Tav - the final letter of the alphabet and the ultimate goal. 

However, our perception is not like Rabbi Meir’s, and we cannot connect all the dots of the world around us. The best we can do is try to go back to how things started and from there try to find meaning. That’s the מאד - starting from the Mem, going to back to the Aleph which is the symbol of Hashem’s unity and then to the Daled, which is the letter highlighting how Hashem interacts with our world. That’s our approach to dealing with this world (see more about this concept in Ari Bergmann's podcast here).

Hence we find a possible connection between the two readings and how they represent differing worldview approaches, as explored by Reb Tzadok. It turned out to be that Rabbi Meir’s approach was not tenable, and the mainstream text is indeed טוב מאד. 

Of course, this discussion leads to the question of how the Torah text can have variant readings, which in turn challenges the Rambam’s view that our Masoretic text is the “immaculate text”, without any changes through time. To read a great piece on this, which requires a more scholarly approach, see this great post at the Kotzk Blog, discussing what would happen if we would find an authoritative old scroll that differs from our accepted Masoretic text. 

One possible conventional answer is brought by the Torah Temima (source), who writes that some understand the Medrash to be referring not to Rabbi Meir’s actual Torah Scroll but his written novelea, where he expounded the meaning of the Torah text. Or perhaps his marginal glosses written around his personal Torah Scroll. In other words, he had no actual variant Mesora. 

Be it as it may, as for the connection between מאד and מות, we have found that these variant readings can be understood not as a mere curiosity; it’s a hidden message highlighted by the Medrash, and up to us to understand its message.

The Mitzva of Writing Your Own Sefer Torah

The very last commandment of the Torah, number 613, is the obligation of every Jew to write a Sefer Torah to himself. That's right, every Jew. You probably know very few individuals who actually did this and there are many reasons why this Miztva is not so widespread.

First, Maimonides says, quoting the Talmud: "Every Jew must write a Sefer Torah (...). If he writes it himself, it's like he has received the Torah from Mount Sinai; if he however doesn't knows how to write it, others may write on his behalf" (source). In fact the Rambam wrote a Torah for himself, as he states in Hilchot Sefer Torah: " (...) I have relied on it in the Torah Scroll that I wrote according to the Halakha".

The Rambam is not the only well-known personality who fulfilled this Miztva with his own hands. The first person to do it was Moshe Rabbeinu, who according to our tradition wrote thirteen (one for each tribe plus one extra for verification purposes) Sifrei Torah. Other famous sofrim include Ezra, (400 b.c.e.), Rabbi Meir, (2nd century), the famous masorete Aharon Ben Asher (9th century), the Meiri, Avraham Sofer (1800's) and many others.

Over the centuries, many Sifrei Torah were written but, expectably, very few survived the test of time. Countless were burnt in the various pogroms and persecutions Jews suffered throughout the ages, but we do have a few scrolls that are almost one thousand years old.

A notable example is the ספר תורה that the Ran, or Rabbeinu Nissim of Gironda (1320-1380), wrote for himself. One of the most revered Rishonim and the Gadol of his time, the Ran wrote a beautiful scroll in Gvil that is still intact, housed in the National Library of the Hebrew University. This scroll is not displayed in public, although I have a friend who saw it a few years ago in a special exhibition in the National Library. In my visit to Hebrew University just last month, I saw a "duplicate" of this Torah which looks exactly like the original and you can see my snapshots below.

Another famous Rabbi who wrote his own Torah is Rabbi Yitzchak Abuhav of Toledo (1300's) and this scroll is still in use today in Tzfat, in the city's famous Kabbalistic Abuhav Synagogue (read the legendary story here).

Today, some Jews write their own Sefer Torah, either via a hired scribe or rarely, by themselves. Most of the Sifrei Torah sold today go to communities who need it for the communal services, but such purchases are "public" and the nature of this Mitzva is that every individual should have a Sefer Torah for himself. The community's scroll for the weekly Torah reading is not part of this Mitzva, to the extent that an individual will lose this Miztva if he writes a Torah for himself but later relinquishes it to the Shul without any explicit condition of retrieving it back.

So why so few Jews fulfill this Mitzva?

Firstly, the reason why 99.9% of the Jews don't write a Sefer Torah by their own hand is simply because is really hard to do it. This requires a great deal of time, will, skill and patience. But why most Jews don't hire scribes to do the work for them?

The most obvious reason a financial one. It's really expensive to hire a good סופר and a Torah could cost up to U$50,000, a figure that is too high for most of us. But there's also an Halachic loophole.

The Rosh (Rabbeinu Asher, 1259 – 1328) writes that a Jew can fulfill this Mitzva by buying a Chumash, Talmud or any other Torah book. According to this, it would seem that you can fulfill this mitzva by buying any Feldheim or Artscroll chumash in a Judaica bookshop. But there is a discussion about how to interpret this Rosh. Some say go so far to say that there is no Miztva to write a Sefer Torah today (an opinion that is very hard to understand, as there is a principle that no one can annul any Mitzva of the Torah), while some say he was merely saying that in addition to writing your Sefer Torah you can also fulfill this Mitzva in another way - by purchasing Torah books.

So if you take the financial issue added to this Halachic loophole, it understandable why most Jews don't write Sefer Torahs - they feel like they do fulfill the Miztva in some way by buying books and learning them.

But it is also important to note another consideration, that in the Rosh's time, there was no printing press and a Chumash was in fact very similar to a Torah Scroll - a person who wanted it had to buy it or commission it from a scribe, who would have written it by hand. So I can understand that in the Rosh's mind there was a lot in common between buying a Torah Scroll or buying a Chumash at that time. The only difference was the Halachot of Stam, which a Sofer must observe when writing a scroll. But I'm not sure the Rosh was comparing writing a Sefer Torah to swiping your credit card in the local Judaica shop and buying the latest Artscroll Chumash in print, so perhaps the Halachic loophole is not as big as people assume it to be.

Other commentators bring the argument that we are not proficient in the exact spelling of the Torah's text (Chaseiros and Yeteirot) and therefore we can't fulfill the Mitzva properly in our days.

However, most authorities are not willing to subscribe to the notion that the Mitzva of writing a Sefer Torah is not applicable anymore, and they assert that the best way to fulfill this Miztva is indeed by writing a Sefer Torah yourself or via a hired scribe. I'm planning to start writing my own Sefer Torah soon, although this project will likely take a few years. But if the Ran, Rambam and so many other busy Gedolim had the time to do it, I should be able to find some time to do it as well!

The Ran's Sefer Torah

Left: Detail of Shirat Hayam

Right: Steel cover of the scroll, saying "Ani Nissim MiGerundi Katavti Sefer Ze lekehilat (...)"

Left: Account of the whereabout of this scroll, written by the Ran's son in the outer side of the scroll. See transcript of this revealing short story here.

Right: Explanation note from Hebrew University