The Sedarim and the Palestinian Reading of the Torah

In about 1170 a traveler from Tudela came into Mizraim, the great city on the Nile, and found two Jewish congregations living side by side. One synagogue, he was told, belonged to the men of the land of Israel, and was called Kenisat-al-Schamiyyin. The other belonged to the men of the land of Babylon, and was called Kenisat-al-Irakiyyin. Their ways of reading the Law were not the same. The Babylonian congregation finished the Torah every year, as was done in Spain, and as Benjamin himself knew the custom. The Palestinian congregation divided each of those larger portions into three sections, and finished the Law at the end of three years. Yet the two communities had an established custom to unite and pray together on the day of the Rejoicing of the Law, and on the day of the Giving of the Law.¹ The hub note on the sedarim sets out the lectionary facts in brief. This essay follows the older, shorter lessons, the sedarim, from a Sabbath in the Land of Israel to the margins of the manuscripts that still remember them, and to the London tables where Christian D. Ginsburg later wrote their disagreements down.

Two Synagogues in Mizraim

Benjamin’s notice is precious because it is local and late. By the time he walked those lanes, the annual cycle of Babylon had already become, in most places, the ordinary custom of Israel. Maimonides, writing in the same century and in the same country, says that the common custom throughout all Israel is to complete the reading of the Torah in one year, beginning after Sukkot with Bereshit, and that although there are those who finish the reading in a three-year cycle, this is not a widely accepted custom.² Benjamin shows what “not widely accepted” still looked like on the ground. In the capital of Egypt the older Palestinian rite had not vanished. It had a named synagogue, a named counterpart, and a negotiated calendar. Twice a year the two congregations, who would not read the same lesson on an ordinary Sabbath, stood together.

Adler’s edition of the itinerary places the passage in the account of Mizraim, a city Benjamin puts on the banks of the Nile, with some seven thousand Jewish inhabitants. Among them he names Nethanel, the Prince of Princes and head of the Academy, who appointed rabbis and officials and was attached to the court.³ The political hour was uneasy — Adler’s notes recall that Saladin had been the real power in Egypt since 1169 — but the liturgical fact does not depend on that commentary. What Benjamin recorded, and what the Jewish Encyclopedia and the Encyclopaedia Judaica both repeat from him, is that Egyptian congregations in about 1170 still took three years to read the Torah.⁴ The Palestinian method, as he describes it, was not an independent list of lessons so much as a threefold division of the portions already known from the annual cycle. That is a traveler’s way of putting it, and it may already reflect a world in which the Babylonian year had become the frame of reference. The Masoretic sedarim, as we shall see, are not simply each annual parashah cut into three equal parts.

Two large synagogues are there, one belonging to the men of the land of Israel and one belonging to the men of the land of Babylon. … Their usage with regard to the portions and sections of the Law is not alike; for the men of Babylon are accustomed to read a portion every week, as is done in Spain, and is our custom, and to finish the Law each year; whilst the men of Palestine do not do so, but divide each portion into three sections and finish the Law at the end of three years. The two communities, however, have an established custom to unite and pray together on the day of the Rejoicing of the Law, and on the day of the Giving of the Law.

Benjamin of Tudela, in Marcus Nathan Adler’s translation (London, 1907), p. 70⁵

The scene matters for a reason that is easy to miss once the sedarim have become a problem of lists and totals. A lectionary is first of all a way of standing in a room and hearing the Law week by week. The Palestinian sedarim were shorter lessons for a longer year. They belonged to synagogues that took their time, and they left their trace, later, in the margins of codices written for communities that no longer read that way.

The People of the West

The earliest clear statement of a fixed, consecutive Palestinian cycle is in the Babylonian Talmud. In the course of a discussion about the special reading for Shekalim, the Gemara explains that a portion which usually falls much later in the year, in the summer, may sometimes be read at the beginning of Adar, “for the people of the West, i.e., Eretz Yisrael, who complete the cycle of reading the Torah not in one year but in three years.”⁶ The remark is almost casual. It is also decisive. “The West,” from a Babylonian house of study, is the Land of Israel. The annual and the triennial cycles were already known as regional customs, and the difference between them was already being used to solve a calendrical puzzle.

The Encyclopaedia Judaica treats this passage as the earliest reference to a fixed cycle of consecutive readings, and notes that in Babylon and in other communities outside Palestine an annual cycle was followed, with the Pentateuch divided into fifty-four weekly portions. That annual division became the universal Jewish practice, except for isolated survivals; in Palestine itself the triennial cycle was in time superseded, possibly under the influence of Babylonian immigrants.⁷ Maimonides, as we have seen, could already describe the annual year as the common custom of all Israel, while still admitting that some places kept the older pace.⁸ Benjamin, a generation or so later, found those places in Egypt.

The Jewish Encyclopedia traces the public reading of the Law itself much further back, at least to about the second century B.C.E., when the grandson of Sirach refers to it in his preface as an Egyptian practice, and therefore, the article argues, it must have existed even earlier in Palestine.⁹ That is a history of reading, not yet a history of the sedarim as a counted Masoretic division. The sedarim are what happened when a three-year synagogue practice was remembered, after the fact, as a map of the text.

What the Words Mean

The vocabulary is a standing invitation to confusion, and it is better to face it at once. In both Talmuds, the Jewish Encyclopedia explains, sidra means a section of Scripture read in the synagogue or in the school. Rashi, on a passage in Yoma, explains the word as a section of the Prophets or the Writings. In Shabbat 116b the people of Nehardea are said to have read a sidra of the Writings at the Sabbath afternoon service. Later Ashkenazi usage applied sidra to the weekly lesson of the annual cycle, which Sephardi usage calls the parashah.¹⁰ The same encyclopedia is equally clear about the older, smaller units. The 153 parts into which, on its count, the Torah was divided in the three-year cycle are known as sedarim, as is seen from the Masoretic colophon at the end of each of the five books. Those sedarim, it adds, are not marked in ordinary synagogue Pentateuchs, because the three-year cycle has so long been out of use.¹¹

Even careful writers slip. Maimonides, in the very chapter that distinguishes the rare three-year custom from the common annual one, calls the annual weekly lessons sidrot and can speak of Sabbaths on which “two sedarim” are read, meaning two annual portions combined so that the year will close on time.¹² He is using the language of his own rite. In Masoretic usage, and in this essay, the sedarim are the shorter divisions of the triennial cycle, preserved in lists, colophons and margins. They are not the fifty-four annual portions, and they are not the open and closed sections that space the paragraphing of the text itself. A seder may begin in the middle of an open or closed section, or may share a beginning with one. The two systems answer different questions: where a paragraph ends, and where a Sabbath lesson once began.

The annual portions have names, taken from their first words or from the first striking word, and they are numbered in printed Bibles. The sedarim of the three-year cycle, the Jewish Encyclopedia notes, are not marked in the same popular way.¹³ Their memory is more technical, and more fragile. That is why the Masorah had to carry them, and why a sixteenth-century editor could almost let them vanish from the printed page.

How Many Sedarim?

No single figure for the sedarim of the Pentateuch has commanded the field, and the disagreement is itself part of the evidence. The Masoretic summaries, the Jewish Encyclopedia reports, give 154, a number that Solomon Judah Rapoport first connected with the Sabbaths of a three-year cycle. Menahem Meiri reckoned 161, the greatest number of Sabbaths that three years can contain. Yemenite grammars and scrolls enumerate 167. Tractate Soferim gives 175, which Rapoport explained as a cycle of three and a half years, so that the Torah was read twice in a sabbatical period.¹⁴ The Encyclopaedia Judaica offers a related account of the same spread: 154 corresponding to the minimum number of Sabbaths in a triennial cycle, 161 to the maximum, the difference arising because a festival that falls on the Sabbath can displace the ordinary lesson; 167 in the Yemenite division; and 175 from the practice of completing the Pentateuch in three and a half years, twice in seven years.¹⁵

The Jewish Encyclopedia’s own article on the parashah, adding the totals given at the end of each book, arrives at a different sum again. Genesis, it says, has twelve annual parashiyyot and forty-three sedarim; Exodus, eleven and twenty-nine; Leviticus, ten and twenty-two; Numbers, eleven and thirty-two; Deuteronomy, eleven and twenty-seven. Those sedarim add up to 153.¹⁶ The same encyclopedia’s “Sidra” article, printing a table of the triennial lessons from a Yemenite manuscript as indicated in Ginsburg’s Masoretic Bible, remarks that according to the Masorah these should number 154, or according to Masseket Soferim 175, but that as a matter of fact they amount to 167.¹⁷ Already, then, a reader who only consults the encyclopedias is faced with 153, 154, 161, 167 and 175, and with the admission that a table taken from the Yemenite tradition does not match the total that tradition’s own Masorah announces.

Ginsburg showed that the discrepancy does not lie only between schools. It runs through a single source. The grammatico-Masoretic treatise that stands before Yemenite Pentateuch manuscripts states plainly that the Sedarim are the pericopes of the triennial cycle which obtained in many communities, and it gives that cycle a number:

There are places where they read through the Law in three years. Hence the Pentateuch is divided into one hundred and fifty-four Sections called Sedarim, so that one Seder is read on each Sabbath. Accordingly the Law is finished at the end of every three years.

A grammatico-Masoretic treatise preceding Yemenite Pentateuch manuscripts, as translated by Ginsburg¹⁸

Yet in the analysis of each book, and in the separate enumeration of each seder, the same treatise, Ginsburg continues, specifies 167 such sedarim, and gives the catchword or verse for every one of them: forty-five in Genesis, thirty-three in Exodus, twenty-five in Leviticus, thirty-three in Numbers, and thirty-one in Deuteronomy. There can, he concludes, be no doubt that the Masoretic school from which these manuscripts proceeded divided the Pentateuch into 167 sedarim. It is, however, certain that other Masoretic schools divided it into 158, and that others again divided it into 154.¹⁹ The numbers are not a single total waiting to be corrected. They are the records of different rooms in which the Law was read, and of different clerks who counted the beginnings.

The 175 of tractate Soferim has a midrashic reason of its own, which is worth hearing in its own words, because it shows a count being justified by something other than a calendar of Sabbaths. In the Soncino translation of the minor tractates, Rabbi Joshua ben Levi says that he once looked into a book of aggadah and found written there that the one hundred and seventy-five sections of the Torah in which any expression of speaking, saying or commanding occurs correspond to the years of Abraham. “On this account,” the passage continues, “the Rabbis instituted one hundred and seventy-five orders in the Torah to be read in public worship.”²⁰ Rapoport’s three-and-a-half-year explanation, reported by the Jewish Encyclopedia, and this Abrahamic explanation can stand side by side. A figure may be remembered for more than one reason, and later scholarship is not obliged to choose a single story when the sources themselves have more than one.

A further constraint on the length of a lesson is older than any of these tallies. The Jewish Encyclopedia, citing Mishnah Megillah 4:4, notes that the minimum seder for a Sabbath portion, when seven persons are called up to the Law, should consist of twenty-one verses, since no one should read less than three verses. Some sedarim, the article adds, have less than twenty-one verses, for example Exodus 30:1–8.²¹ The exception is as revealing as the rule. The division was not a mechanical slicing of the text into equal lengths. It followed, or sometimes failed to follow, a sense of what a Sabbath reading ought to be.

A List Almost Lost in Israel

Because European communities followed the annual cycle, Ginsburg writes, the sedarim that exhibit the more ancient division of the text have been totally ignored in most manuscripts. Even modern editions of so-called Masoretic Hebrew Bibles, which state at the end of each book that it contains such and such a number of sedarim, give no indication whatever as to where, in the text, any seder occurs.²² The count survived as a colophon. The beginnings did not always travel with it. That is one reason the printed page can look so sure of a number and so shy of a place.

Jacob ben Hayyim, the first editor of the Bible with the Masorah, issued at Venice in 1524–25, tells a story that Ginsburg was right to linger over. If he had found this Masoretic division of the text, he would have followed it in preference to the Christian chapters which he adopted from Rabbi Nathan’s Hebrew Concordance. Having, however, obtained the list when he had nearly carried the Bible through the press, he says: “I have published it separately so that it may not be lost in Israel.”²³ The sentence is almost too neat to be true, and it is in the introduction of the edition itself, as Ginsburg reports it. A system older than the chapter divisions of the Latin and Hebrew concordances almost yielded, in the first great printed Masoretic Bible, to a system imported from another tradition. It was saved as an appendix, not as the running frame of the page.

The later history of those Christian chapters, and of the verse numbers that followed them, belongs to another essay. What matters here is the near-miss. The sedarim were, for a moment in Venice, in competition with the chapters for the right to organise a printed Hebrew Bible, and they lost on a point of timing. Their later life has been lived in lists, in Masorah parva notes, and in the closing summaries at the ends of books — the very summaries that can name a total without mapping it.

Marks in the Margin

Where the sedarim are marked in the manuscripts, it is in the margin. Ginsburg singles out Oriental 2201 in the British Museum, dated 1246, as of special importance, “since it not only has a separate List of the Sedarim, but marks every Seder in the margin of the text itself with” a samekh “against the place where it begins, thus leaving no doubt as to which verse it belongs.” The same, he says, is true of Oriental 2451, a manuscript in a Persian hand that contains the Pentateuch, the haftarot and the Psalms, in which the sedarim are also marked in the margin of the text.²⁴ In Oriental 2349, a Yemenite copy, the Masorah parva numbers each seder in the margin against the places where it occurs within the annual pericopes. On Genesis 1:1, in the annual lesson Bereshit, it remarks that the lesson contains four sedarim and that this is the first; at Genesis 2:4 it notes the second; at 3:22 the third; at 5:1 the fourth.²⁵ A reader of that margin is watching the older cycle peek through the newer one. The annual sidra has become the frame, and the sedarim are counted inside it.

The Yemenite evidence is both rich and untidy, which is what one wants from a living tradition. All five of the Yemenite manuscripts Ginsburg used state that Genesis has forty-five sedarim, and they give the pericope and verse for each. Even Oriental 2201, which gives forty-three sedarim in its list, states in the Masoretic summary at the end of Genesis that the book has forty-five. The variations among the copies are not trivial. The sixth seder, Genesis 8:1, given in all five Yemenite manuscripts and in all the lists, is omitted in the margin of Oriental 2201 and in the editio princeps. The ninth, Genesis 11:1, given in the Yemenite manuscripts and marked in the margin of Oriental 2201, is omitted in all the lists and by Baer. Genesis 12:1 has no seder in the Yemenite manuscripts or in the list of Oriental 4227, though it is marked in the text of Oriental 2201 and appears in that manuscript’s list, in Add. 15251, in the editio princeps and in Baer. Genesis 17:1, given in all the lists as the fourteenth seder, is not marked in the Yemenite manuscripts or in the text of Oriental 2201. At Genesis 21:22, Oriental 2201 marks a seder in the margin instead of at 22:1, which is the place given in all the other manuscripts and even in this manuscript’s own list.²⁶ A list and a margin, even in one book, can disagree. Ginsburg’s patience with those disagreements is the point. He is not hunting a single correct total. He is showing how a division was used.

The sedarim were not confined to the Pentateuch. After the Torah, Ginsburg turns to the Prophets and the Writings, and he records, among much else, that all the manuscripts agree that Joshua has fourteen sedarim, with only one instance in which the Yemenite manuscripts exhibit a different recension.²⁷ From the analysis as a whole he concludes that the Hebrew Bible contains 452 sedarim: 167 in the Pentateuch, 97 in the Former Prophets, 107 in the Latter Prophets, and 81 in the Hagiographa.²⁸ Those figures belong to his collation. They should not be silently turned into “the” Masoretic number for every manuscript. They are the best published map we have of a system that, as he had already said, most European copies simply ignored.

Later scholarship has tried to put the Palestinian cycle back into a calendar, not only into a margin. The Encyclopaedia Judaica records that Adolf Büchler, “with great ingenuity,” attempted to reconstruct the weekly portions of the triennial cycle, assuming that the cycle began on the first day of Nisan, and that on the basis of that reconstruction he sought to explain various traditions about the dates of biblical events — for example, that Moses died on the seventh of Adar, or that Sarah was “remembered” on the first of Tishri — as memories of the season in which the relevant lesson was read once every three years.²⁹ The Jewish Encyclopedia’s “Sidra” article already pointed readers to Büchler’s studies in the Jewish Quarterly Review.³⁰ Whether every such date-tradition was born from a lectionary is a question this essay does not need to settle. What the attempt shows is that the sedarim, once recovered as a system, looked again like a way of living through the year, not only like a set of marks beside a verse.

A Cycle That Became a Note

By Maimonides’ day the annual cycle was already the common custom of Israel, and the three-year cycle a minority practice.³¹ By Benjamin’s day it could still be heard in one of the two great synagogues of Mizraim, and the two congregations still knew how to keep festival together.³² In the printed Bibles that followed Jacob ben Hayyim, a book might close with a proud total of sedarim and never show a reader where to find the first word of any of them.³³ The Masorah had done its work, and then, in many places, the work had become invisible.

That invisibility is not a failure of the Masoretes so much as a measure of how completely the Babylonian year won the synagogue. A division that had organised the hearing of the Law in the Land of Israel survived as an apparatus for the page. Ginsburg, working in the British Museum with Yemenite and other Oriental manuscripts — Oriental 2201, the Persian-hand Oriental 2451, the Yemenite group that includes Oriental 2348, 2349, 2364 and 1379, and the lists in Oriental 4227 and Add. 15251 — treated that apparatus as a historical document.³⁴ He also compared the printed list in the first edition of Jacob ben Hayyim’s Rabbinic Bible. In the preface to his own Massoretic edition he announced, as a feature of that edition, that “the ancient Massoretic chapters, called Sedarim, are also indicated throughout in the margin against their respective places.”³⁵ The recovery was meant to be visible on the printed page, not only argued in a chapter of an introduction. He printed the treatise that announced 154 sedarim and then listed 167. He trusted Oriental 2201 for the samekh in the margin, and then recorded the places where even that copy’s margin and list did not agree. He added the Prophets and the Writings, and he came to 452. The result is not a restored lectionary for modern use, though modern communities have at times tried to recover a three-year pace. It is a description of how the text was once heard, and of how the memory of that hearing was stored.

The later life of the cycle, once it had left the ordinary synagogue, is not only a story of manuscripts. In the early twentieth century some congregations in the English-speaking world considered returning to a three-year reading. The Encyclopaedia Judaica records that Büchler, asked for a ruling by an Anglo-Jewish congregation, answered that if the question was one of law, the triennial cycle was against the codified law from the twelfth century onward, and even much earlier in Babylon, “whence our law proceeded.” “If you introduce the triennial cycle,” he said, “you separate yourself from the main body of Judaism.” The encyclopedia cites the 1913 report of London’s New West End Synagogue for that reply, and it adds that, since the early nineteenth century, various attempts had been made to reintroduce the cycle, and that many contemporary Reform and Conservative congregations follow the different practice of reading about a third of the annual portion each week, thereby finishing the ordinary sidra cycle in three years without restoring the old sedarim as the lesson-units themselves.³⁶ That modern compromise should not be confused with the Masoretic division. It is a new way of slowing the Babylonian year, not a revival of the Palestinian list.

The hub page on the sedarim gathers the competing totals and the principal manuscripts in a shorter compass. The thing to keep, after the totals have been recited, is the human shape of the practice. Somewhere in the Land of Israel a congregation took three years to walk through the Torah. Somewhere in Egypt, centuries later, a Palestinian synagogue still did, within earshot of a Babylonian one that did not. On two days of the year they prayed as one people. The rest of the time their Sabbaths were of different lengths, and the shorter of those Sabbaths left marks, a letter samekh and sometimes a number, in the margins of the books that outlived the rite.

The numbers will not be made to agree, and they should not be. 153, 154, 158, 161, 167 and 175 are not a problem to be solved by choosing a favourite source. They are what a lectionary looks like when it has been lived in more than one city, counted by more than one school, and then written into a Masorah that was already serving a different year.

Notes

1. The Itinerary of Benjamin of Tudela, ed. and trans. Marcus Nathan Adler (London: Henry Frowde, Oxford University Press, 1907), 70, at archive.org; “Triennial Cycle,” in the Jewish Encyclopedia (1901–1906), at jewishencyclopedia.com; “Torah, Reading of,” in the Encyclopaedia Judaica, at Encyclopedia.com. *
2. Maimonides, Mishneh Torah, Laws of Prayer and the Priestly Blessing 13:1, trans. Eliyahu Touger, at sefaria.org; Encyclopaedia Judaica, “Torah, Reading of”. *
3. Adler, Itinerary of Benjamin of Tudela. *
4. Jewish Encyclopedia, “Triennial Cycle”; Encyclopaedia Judaica, “Torah, Reading of”; Adler, Itinerary of Benjamin of Tudela. *
5. Adler, Itinerary of Benjamin of Tudela. *
6. Babylonian Talmud, Megillah 29b, William Davidson translation, at sefaria.org; Encyclopaedia Judaica, “Torah, Reading of”. *
7. Encyclopaedia Judaica, “Torah, Reading of”. *
8. Maimonides, Mishneh Torah, Prayer 13:1. *
9. Jewish Encyclopedia, “Triennial Cycle”. *
10. Cyrus Adler, M. Seligsohn, Joseph Jacobs and I. George Dobsevage, “Sidra,” in the Jewish Encyclopedia (1901–1906), at jewishencyclopedia.com; “Parashah,” in the Jewish Encyclopedia (1901–1906), at jewishencyclopedia.com. *
11. Jewish Encyclopedia, “Parashah”. *
12. Maimonides, Mishneh Torah, Prayer 13:1. *
13. Jewish Encyclopedia, “Parashah”. *
14. Jewish Encyclopedia, “Triennial Cycle”; Jewish Encyclopedia, “Sidra”. *
15. “Triennial Cycle,” in the Encyclopaedia Judaica, at Encyclopedia.com. *
16. Jewish Encyclopedia, “Parashah”. *
17. Jewish Encyclopedia, “Sidra”. *
18. Christian D. Ginsburg, Introduction to the Massoretico-Critical Edition of the Hebrew Bible (London, 1897), ch. 4, at archive.org, 32. *
19. Ginsburg, Introduction, 32–34. *
20. Tractate Soferim 16, trans. A. Cohen, in The Minor Tractates of the Talmud (London: Soncino Press, 1965), at sefaria.org. *
21. Jewish Encyclopedia, “Triennial Cycle”; Mishnah Megillah 4:4, as summarised in the Jewish Encyclopedia article “Triennial Cycle,” at jewishencyclopedia.com. *
22. Ginsburg, Introduction, 32. *
23. Ginsburg, Introduction, 33. *
24. Ginsburg, Introduction. *
25. Ginsburg, Introduction, 33–34. *
26. Ginsburg, Introduction. *
27. Ginsburg, Introduction. *
28. Ginsburg, Introduction. *
29. Encyclopaedia Judaica, “Torah, Reading of”. *
30. Jewish Encyclopedia, “Sidra”. *
31. Maimonides, Mishneh Torah, Prayer 13:1. *
32. Adler, Itinerary of Benjamin of Tudela. *
33. Ginsburg, Introduction, 32. *
34. Ginsburg, Introduction. *
35. Ginsburg, Introduction. *
36. Encyclopaedia Judaica, “Torah, Reading of”. *