Open and Closed Sections in the Hebrew Bible

Anyone who opens a Torah scroll or a carefully printed Hebrew Bible will notice that the text is not continuous. At intervals the writing stops before the end of a line, or a whole line is left empty, or a space of a few letters is left within a line before the writing resumes. These spaces are not ornamental and they are not left to the discretion of the scribe. They mark the divisions known in Hebrew as parashiyyot, or in some sources pisqaʾot, which are of two kinds: the open section, parashah petuḥah, and the closed section, parashah setumah. The division of the text into such sections is among the oldest features of the written tradition, and in the Torah it became a matter of law, so that a scroll in which an open section was written as closed, or a closed one as open, could not be used for public reading.

This essay describes the two kinds of section and the rules for writing them, the evidence for their antiquity, the codification of the Torah sections by Maimonides on the basis of a famous codex associated with Ben Asher, the differences between the Babylonian schools of Sura and Nehardea, the less settled condition of the sections in the Prophets and the Writings, and the recovery in modern times of the section divisions of the Aleppo Codex for those parts of the Bible where the codex itself is lost.

The Two Forms of Section

The clearest classical description of the two kinds of section is given by Maimonides in the eighth chapter of his Laws of the Torah Scroll in the Mishneh Torah. An open section, he writes, has two forms. If the scribe finished the preceding section in the middle of a line, he leaves the rest of the line blank and begins the open section at the start of the next line; this applies when the space left in the line is sufficient for nine letters. If only a little space remained, or if he finished at the very end of a line, he leaves the whole of the following line blank and begins the open section at the start of the line after that. A closed section, by contrast, has three forms. If the scribe finished in the middle of a line, he leaves a space of the proper measure and begins the new section later in the same line, so that the blank space stands in the middle. If there was not enough room to leave the space and still write a word at the end of the line, he leaves the rest of that line blank, leaves a small space at the beginning of the next line, and begins the closed section from the middle of it. If he finished at the end of a line, he leaves the measure of the space at the beginning of the next line and begins after it. Maimonides summarises the principle: an open section always begins at the beginning of a line, and a closed section always begins from the middle of a line.¹

Aron Dotan’s account in the Encyclopaedia Judaica describes the same distinction more briefly. The open section starts at the beginning of a line, the preceding line being left partly or wholly blank; the closed section begins at a point other than the start of a line, whether the preceding section ended on the previous line or on the same one, in which case a space of approximately nine letters is left between the two sections. Dotan notes that in some manuscripts and printed editions the two kinds are indicated by the letters pe and samekh, the initials of petuḥah and setumah.² Christian David Ginsburg, describing the Masoretic rule in his Introduction of 1897, gave the measure of the vacant space at the end of the unfinished line before an open section as that of three words of three letters each, which is the same nine-letter measure expressed in another way. He compared the closed section, which begins with an indented line or after a space in the middle of a line, to what a modern reader would call a new paragraph.³

Ginsburg also insisted that the letters pe and samekh written in the space are a secondary development. In the synagogue scrolls, which in his judgement preserved the most ancient practice, and in the best and oldest manuscripts in book form, the layout of the spaces is the only means by which the two kinds of section are distinguished. The practice of writing a letter in the space, adopted in some manuscripts and editions, he regarded as of later date, and he therefore disregarded it in his own edition.⁴ The distinction matters because the spaces themselves can be ambiguous when a section ends close to the end of a line or at the foot of a column, and the letters were introduced in part to remove that ambiguity; but their presence in a manuscript does not, on Ginsburg’s view, belong to the oldest form of the tradition.

Antiquity of the Divisions

The division of the text into sections is far older than the Masoretic notes or the vowel signs. Dotan observes that it differs from the division of the Torah into weekly readings because it is bound up with the very copying of the text, whether in a scroll or a codex.⁵ The rabbinic sources already treat it as fixed. Dotan cites the Talmud, Shabbat 103b, for the rule that an open section should not be made closed and a closed section should not be made open, and the Sifra on Leviticus 1:1 for a question about their purpose: the sections were given in order to allow Moses an interval for reflection between one section and the next and between one subject and the next.⁶

The physical evidence goes back further still. Geoffrey Khan, in his survey of the Tiberian Masoretic tradition, lists the indication of paragraph divisions, known in Hebrew as pisqaʾot or parashiyyot, as one of the distinct written components of the tradition, alongside the consonantal text, the layout of the manuscripts, the accents, the vocalisation and the marginal notes.⁷ He points out that paragraph divisions of this kind are found in the manuscripts from Qumran, both biblical and non-biblical, and that there is a large degree of agreement between the paragraphing of the Qumran biblical scrolls and that of the medieval manuscripts. That agreement indicates that the tradition can be traced back to the Second Temple period.⁸ The medieval system of open and closed sections is therefore not an invention of the Masoretes but a feature they inherited and protected.

A further sign of the independence of the sections is that they do not always coincide with the division of the text into verses. Khan notes that in a number of places the paragraph divisions in the medieval manuscripts fall within a verse as it is divided by the accents, a phenomenon known as pisqa be-emṣaʿ pasuq, “a paragraph division within a verse”. Among his examples are Genesis 35:22 and 1 Samuel 16:2. His explanation is that the paragraph division of the written text and the division expressed by the cantillation are two different layers of exegetical tradition, which occasionally do not correspond with one another.⁹ The sections, in other words, belong to the written form of the text as it was copied, while the accents record the way in which it was read aloud, and the two traditions were preserved side by side even where they diverged.

Sections and Other Divisions of the Text

The open and closed sections should be distinguished from the other ways in which the biblical text has been divided. Ginsburg listed four divisions of the Pentateuch: the open and closed sections, the triennial pericopes, the annual pericopes, and the verses.¹⁰ Dotan explains the difference between the first and the two kinds of pericope. In the Land of Israel the reading of the Torah in the synagogue was completed once in three years, and the Pentateuch was accordingly divided into one hundred and fifty-four weekly portions, or one hundred and sixty-seven according to another version, called sedarim. In Babylonia the cycle was completed in a single year, so that the Torah was divided into fifty-four weekly portions, the parashiyyot in the sense still used for the annual cycle of Torah reading. The sedarim were later applied to the whole Bible, although outside the Torah they were hardly needed for liturgical purposes, and the manuscripts differ in the location and even the number of these divisions.¹¹

The use of the same word, parashah, for both the weekly portion and the open or closed section can cause confusion, but the two are quite different in kind. A weekly portion is a unit of public reading, marked in many manuscripts by a note in the margin, and its boundaries have no effect on the way in which the text is written. An open or closed section, by contrast, is marked by the arrangement of the writing itself, and as Dotan stresses it is bound up with the very act of copying.¹² The scale also differs: Maimonides counts six hundred and sixty-nine sections in the Torah, against fifty-four annual portions. The chapter and verse numbers familiar from printed Bibles are later still, and the site’s essay on chapter and verse divisions describes how they entered Hebrew Bibles from the Latin tradition. A chapter division may cut across a section, and a section may begin in the middle of a chapter, because the two systems were devised independently and for different purposes.

The verses themselves are an older division than the chapters, and they belong to the Masoretic tradition, since the end of each verse is marked by an accent. Yet even the verse division, as noted above, does not always coincide with the sections. The fact that the scribal tradition of the sections, the oral tradition of the verse division and the later system of chapters could all be preserved together, each without forcing the others into conformity, says something about the character of the Masoretic enterprise. Its aim was not to rationalise the text but to keep each inherited element in the form in which it had been received.

Maimonides and the Torah Sections

Despite their antiquity, the placing and number of the sections were not uniform in the medieval manuscripts. Dotan states that different traditions or customs developed concerning the position and type of each section, and that the great degree of uniformity found in printed editions of the Torah today is mainly due to the halakhic fixing of this matter, and of the layout of the songs, by Maimonides following Ben Asher.¹³ Maimonides himself explains why he undertook the task. He first states the law: a scroll that has errors in full and defective spelling can be corrected, but if the scribe erred in the sections, writing a closed section as open or an open one as closed, or left a space where there is no section, or wrote continuously where a section belongs, or changed the form of the songs, the scroll is invalid and there is no remedy except to remove the entire sheet on which the error occurs.¹⁴

He then describes the state of the evidence. He had seen great errors in all the books he examined in these matters, and the Masoretic scholars who wrote and compiled works to make known the open and closed sections disagreed among themselves according to the books on which they relied. He therefore decided to write out all the sections of the Torah, closed and open, together with the forms of the songs, so that all books could be corrected and checked against his list.¹⁵ The book on which he relied, he continues, was the well-known codex in Egypt containing the twenty-four books, which had been in Jerusalem some years earlier and had served there for the correction of other books. Everyone relied on it, because Ben Asher had corrected it and worked on it with exactitude for many years, checking it many times as it was copied. Maimonides adds that he followed it in the Torah scroll that he himself wrote according to the law.¹⁶ It is generally agreed that this codex is the one later known as the Aleppo Codex, which was kept in Egypt before it was brought to Aleppo.¹⁷

The list that follows names each section of the Torah by its opening words and states whether it is open or closed. For Genesis Maimonides counts forty-three open and forty-eight closed sections, ninety-one in all; for Exodus sixty-nine open and ninety-five closed, one hundred and sixty-four in all; for Leviticus fifty-two open and forty-six closed; for Numbers ninety-two open and sixty-six closed; and for Deuteronomy thirty-four open and one hundred and twenty-four closed. The total for the whole Torah, in the text as transmitted, is two hundred and ninety open sections and three hundred and seventy-nine closed, six hundred and sixty-nine in all.¹⁸ The list preserves the internal structure of the books in a way that the later chapter divisions do not. The Ten Commandments, for instance, are divided in both Exodus and Deuteronomy into a series of short sections, nearly all of them closed, and the casuistic laws of Exodus 21–22 are divided into short closed sections, one for each case, each beginning with its own opening formula.¹⁹

Maimonides also fixed the layout of the two great songs of the Torah. The Song of the Sea in Exodus 15 is to be written in thirty lines: the first line in the ordinary way, and the remaining lines with one or two spaces in the middle, so that the line is divided into three and space stands opposite writing and writing opposite space. The song of Haʾazinu in Deuteronomy 32 is written in sixty-seven lines, each with a single space in the middle in the form of a closed section, so that each line is divided into two. For both songs Maimonides lists the words with which the lines begin.²⁰ These rules complete the legal definition of the written form of the Torah: the letters, the spaces between sections and the layout of the songs together make up what a scribe must reproduce.

Sura and Nehardea

The Tiberian tradition that Maimonides codified was not the only one. The Babylonian Masorah, which survives in manuscripts with Babylonian vocalisation and in lists from the Geniza, attended to the sections more closely than its Tiberian counterpart. Kim Phillips, reporting on the Masoretic notes in a Torah codex in the National Library of Russia, observes that the Babylonian Masorah contains far more references to the sections than the Tiberian, and that the sections play a double role in it: some rules are framed so that they apply to a single section, while elsewhere a section serves as a point of reference for locating a word. Phillips cites Yosef Ofer’s suggestion that the reason is that the Babylonian Masorah specifically attends to the preservation of the sections, noting at each relevant point the presence of a new section and its type, whereas the Tiberian Masorah is seemingly indifferent to them.²¹

Within the Babylonian tradition there were further differences between the scholars associated with the academies of Sura and Nehardea. Israel Penkower, in a study of these differences in the text, the vocalisation and the open and closed sections, collected the Babylonian lists and the marginal notes that record them. One source is a list headed “These are the open, closed and sedurah sections in the Torah”, preserved in a manuscript in the National Library of Russia, Evr. II C 156, and published by Ofer; Penkower notes that he rediscovered the manuscript during a visit to the library in August 1990, following a hint in an earlier article by Hermann Strack.²² Another is a list of about seventy sections disputed by the Nehardeans, preserved in a manuscript of the Jewish Theological Seminary in New York and printed by Ginsburg.²³

The examples Penkower gives show the kinds of difference involved. At Genesis 3:16, “To the woman He said”, where the Aleppo Codex has a closed section, the Nehardean tradition as Penkower reconstructs it also had a closed section, but the Surans did not write a section at all. At Exodus 12:1, where the Aleppo Codex has a closed section, the Nehardeans had an open one. At Exodus 23:26 the Aleppo Codex and the Nehardeans have a closed section, while the Surans had a third type of division, the sedurah.²⁴ This third category is unknown in the Tiberian system, and its presence in the Babylonian lists is a reminder that the binary scheme of open and closed, which the Tiberian codices and Maimonides made standard, was one tradition among several. Penkower also notes, citing Ofer’s work on the Babylonian Masorah, that Exodus 23:26 is among the places where a dot was to be written before the first word to signal a dispute over the section, and that the list of such dots includes the example but omits the dot itself by mistake.²⁵

The Prophets and the Writings

Outside the Torah the sections never received a halakhic codification comparable to that of Maimonides, and the manuscripts show correspondingly greater variation. Ginsburg found that the manuscripts on the whole exhibit uniformity in the Pentateuch and that separate lists giving the catchwords of every open and closed section of the Torah had been preserved. No such care, he wrote, had been taken by the Masoretes over the sections in the Prophets and the Writings, and no separate list of them had been discovered. Although the divisions there are reasonably uniform, it is frequently impossible to say whether a given break indicates an open or a closed section, and some manuscripts often show an open section where others show a closed one.²⁶

Ginsburg made this point in criticism of the edition of Seligman Baer, who had inserted pe and samekh into the text of the Prophets and Writings. In Ginsburg’s judgement these markings had no better warrant than the testimony of one codex from one school of Masoretes, and the greater number of the standard codices told against them. For the Former Prophets he collated six codices in the British Museum, now the British Library, arranging the catchwords of their sections in parallel columns beside Baer’s, and he reported that Baer had omitted twenty-nine sections in Joshua, eighteen in Judges and fifty-one in Samuel that were plainly given in the manuscripts.²⁷ In Isaiah he counted twenty-four omitted sections.²⁸

For the Latter Prophets Ginsburg had the help of the St Petersburg codex of the Prophets dated 916, which had been published in facsimile by Hermann Strack. He described it as strictly observing the rules for open and closed sections, so strictly that when an open section ended with a full line at the foot of a column, which according to the rule required an entire blank line, the scribe placed a pe in the middle of the empty space to show that nothing was missing. From this codex Ginsburg concluded that in early times the open and closed sections were indicated as carefully in the Prophets and Writings as in the Pentateuch, and that the neglect of the rules was due to later scribes.²⁹ He also noted that the editors of the Lisbon edition of 1492, which he collated for Isaiah and Jeremiah, were the first to introduce the letters pe and samekh into the printed text of the Prophets.³⁰

The poetic books present a different case. Ginsburg observed that the Psalms have no sections, each psalm forming a continuous whole, separated from the next only by its heading in the middle of the line, and that the best manuscripts set out the three poetical books in half-verses. He noted that the first complete printed Bible, Soncino 1488, abandoned this arrangement to save space, while the Naples edition of about 1491–93 restored it, and that later editors followed Soncino.³¹ For Proverbs he collated a Paris manuscript dated 1286, which divides the book into thirty-nine sections, thirty-two of them marked as open by a blank line and the letter pe in the margin, and the remainder preceded only by a blank line.³² The variety of practice in these books illustrates the general point: where no binding list existed, each school and each scribe could resolve the layout in its own way.

The Sections of the Aleppo Codex Recovered

The modern effort to restore the Tiberian text has had to confront the same problem. When Mordechai Breuer prepared the text of the Bible for the Daʿat Miqra series of Mossad Harav Kook, completed in 5742 (1981–82), he based it on the Aleppo Codex, which had reached Israel in 5718 (1957–58) and had been entrusted to the Bible Project of the Hebrew University of Jerusalem.³³ Yosef Ofer, Breuer’s student and colleague, explains the difficulty that remained. Maimonides had fixed the list of open and closed sections in the Torah according to the Aleppo Codex, and Breuer considered it right to follow the codex in the sections of the Prophets and Writings as well. But in those parts of the Bible where the codex has been lost, its sections could not be reconstructed with certainty from other manuscripts, because each manuscript has its own tradition in this respect. For want of a better source Breuer relied there on the Leningrad Codex, a decision that, according to Ofer, continued to trouble him.³⁴

The solution came by chance. Ofer relates that in 1987 a house in the Kiryat Moshe neighbourhood of Jerusalem was about to be demolished. Books and manuscript papers had accumulated in its attic over many years, and the owners invited an acquaintance to examine them and take what he wished. Among the papers he found writings on the Masorah connected with Rabbi Shalom Shakhna Yellin, one of the founders of the family that had owned the house, who had settled in the Land of Israel in the 1860s. Ofer knew that Yellin had sent an emissary to Aleppo with a printed Bible, asking him to examine the codex and to record its sections and other details of its text, and that after the mission the printed Bible had been lost. The annotated Bible was not among the papers in the attic, but enquiries revealed that some books of apparently little value had been sent to a bookshop in Mea Shearim, where they had been set aside in a carton for disposal. There, at the last moment, the Bible was found.³⁵

From this copy Ofer compiled a list of the sections of the Aleppo Codex in the places where the codex is now lost, and Breuer used it in the new edition of the Bible that he published with Horev in 1996. Ofer quotes Breuer’s introduction to that edition, which describes the discovery in 1987 of a copy of the Bible in which all the open and closed sections of the codex in the Prophets and Writings had been accurately marked, on the initiative of an expert scribe and corrector and on behalf of the leading rabbis of Jerusalem, and which had lain for many years in an attic until it reached a bookseller who was about to consign it to a genizah.³⁶ Ofer adds that the parts of the codex now missing include portions of Kings, Jeremiah and the Twelve Prophets, most of the Song of Songs, and all of Ecclesiastes and Lamentations, and that many scrolls of the Prophets written on parchment in recent years follow the sections of the codex as recorded in this copy. It also emerged that the rabbis and scribes of Jerusalem about a hundred and fifty years earlier had used the same testimony when writing scrolls of the Prophets.³⁷

Assessment

The open and closed sections are among the oldest features of the written Hebrew Bible. Their agreement with the paragraphing of the Qumran scrolls places their origin in the Second Temple period, and their occasional divergence from the verse division shows that they belong to the written transmission of the text rather than to its oral reading.³⁸ The rabbinic prohibition on turning one kind into the other, and Maimonides’ ruling that such an error invalidates a scroll, gave them a legal status that few other features of layout possess.³⁹ Maimonides’ decision to follow the codex corrected by Ben Asher, and to set out every section of the Torah by its opening words, is the main reason why the Torah sections are now nearly uniform.

Outside the Torah, and outside the Tiberian tradition, the picture is less settled. The Babylonian lists record a third category of division and a number of disagreements between Sura and Nehardea.⁴⁰ In the Prophets and Writings, Ginsburg found no Masoretic list and considerable variation among the manuscripts, though the St Petersburg codex of 916 showed that early scribes had observed the rules there as strictly as in the Torah.⁴¹ The rediscovery of the Yellin Bible in 1987 has allowed the sections of the Aleppo Codex to be restored for its lost portions, and with them the tradition that Maimonides had endorsed for the Torah has been extended, as far as the evidence allows, to the rest of the Bible.⁴²

Notes

1. Moses Maimonides, Mishneh Torah, Laws of Tefillin, Mezuzah and Torah Scroll, Laws of Torah Scroll, ch. 8, English translation at Mechon Mamre, §§1–2. *
2. Aron Dotan, “Masorah”, in Encyclopaedia Judaica, 2nd edn, at Encyclopedia.com, §1.2.1.3. *
3. Christian D. Ginsburg, Introduction to the Massoretico-Critical Edition of the Hebrew Bible (London: Trinitarian Bible Society, 1897), pp. 9–10. *
4. Ginsburg, Introduction, p. 10. *
5. Dotan, “Masorah”, §1.2.1.3. *
6. Dotan, “Masorah”, §1.2.1.3. *
7. Geoffrey Khan, The Tiberian Pronunciation Tradition of Biblical Hebrew, vol. 1 (Cambridge: Open Book Publishers, 2020), pp. 17–18. *
8. Khan, Tiberian Pronunciation Tradition, p. 50. *
9. Khan, Tiberian Pronunciation Tradition, p. 50. *
10. Ginsburg, Introduction, p. 9. *
11. Dotan, “Masorah”, §1.2.1.2. *
12. Dotan, “Masorah”, §1.2.1.3. *
13. Dotan, “Masorah”, §1.2.1.3. *
14. Maimonides, Laws of Torah Scroll 8, §3. *
15. Maimonides, Laws of Torah Scroll 8, §4. *
16. Maimonides, Laws of Torah Scroll 8, §5. *
17. “Nusach Ha-mikra: Accuracy of the Biblical Text (3)”, Torat Har Etzion, Israel Koschitzky Virtual Beit Midrash. *
18. Maimonides, Laws of Torah Scroll 8, §§6–10. *
19. Maimonides, Laws of Torah Scroll 8, §§7, 10. *
20. Maimonides, Laws of Torah Scroll 8, §§11–13. *
21. Kim Phillips, “The Masoretic Notes in RNL EVR II B 80+: An Initial Report”, in Daniel J. Crowther, Aaron D. Hornkohl and Geoffrey Khan (eds), Studies in the Masoretic Tradition of the Hebrew Bible (Cambridge: Open Book Publishers, 2022), pp. 57–58 and n. 44. *
22. Israel S. Penkower, “Ḥilufe Suraʾe ve-Nehardeʿe ba-Tanakh: nusaḥ, niqqud u-farashiyyot petuḥot u-setumot” (in Hebrew), Jewish Studies, an Internet Journal 18 (2020), n. 202. *
23. Penkower, “Ḥilufe Suraʾe ve-Nehardeʿe”, n. 202. *
24. Penkower, “Ḥilufe Suraʾe ve-Nehardeʿe”, nn. 203–205. *
25. Penkower, “Ḥilufe Suraʾe ve-Nehardeʿe”, n. 205. *
26. Ginsburg, Introduction, p. 10. *
27. Ginsburg, Introduction, pp. 10–12. *
28. Ginsburg, Introduction, p. 14. *
29. Ginsburg, Introduction, pp. 13–14. *
30. Ginsburg, Introduction, p. 14. *
31. Ginsburg, Introduction, pp. 17–18. *
32. Ginsburg, Introduction, p. 18. *
33. Yosef Ofer, “Ha-Rav Mordekhai Breuer: makhriʿa nusaḥ ha-miqra” (in Hebrew), Makor Rishon, 26 Adar 5767 (2007). *
34. Ofer, “Makhriʿa nusaḥ ha-miqra”. *
35. Ofer, “Makhriʿa nusaḥ ha-miqra”. *
36. Ofer, “Makhriʿa nusaḥ ha-miqra”. *
37. Ofer, “Makhriʿa nusaḥ ha-miqra”. *
38. Khan, Tiberian Pronunciation Tradition, p. 50. *
39. Dotan, “Masorah”, §1.2.1.3; Maimonides, Laws of Torah Scroll 8, §3. *
40. Penkower, “Ḥilufe Suraʾe ve-Nehardeʿe”, nn. 202–205. *
41. Ginsburg, Introduction, pp. 10–16. *
42. Ofer, “Makhriʿa nusaḥ ha-miqra”. *