Versification and Cantillation

This document is a companion to Versification differences. That document deliberately ignores cantillation, defining a verse as simply a span of text between two cv-labels. That definition of a verse is what we will call a numbered verse in this document. In this document, we will not ignore cantillation, and it will be important to make a distinction between a numbered verse and a chanted verse. We define a chanted verse as a span of text between two sof pasuq marks, whose accents are intended to obey the rules of cantillation. (By this definition, a dually-cantillated span between two sof pasuq marks is not a chanted verse, although at least one of its strands is, if both sof pasuq marks belong to that strand.)

In cases that are rare, but central to this document, a numbered verse is not a chanted verse. This happens mostly (and unavoidably) in the Decalogues, where, no matter what versification is used, some numbered verses are not chanted verses in at least one of the two strands of cantillation. Different trade-offs in the face of this "impossible" situation give rise to different versifications of the Decalogues.

Aside: like many numbered verses of the Decalogues, Gen. 35:22 has two strands of cantillation and is only a chanted verse in one of those two strands. (In the other strand, it contains two chanted verses in their entirety.) Unlike in the Decalogues, at Gen. 35:22 all versifications discussed here (MAM, BHS, and Sefaria) make the same trade-off in the face of this "impossible" situation: they all agree to label Gen. 35:22 as a single numbered verse.

Other than the cases in the Decalogues, the one other case in our versifications where a numbered verse is not a chanted verse is the Num. 25/26 break in BHS versification. Beyond that shared trait, this case has little in common with the Decalogue cases: here there is only one strand of cantillation, and, as we will see, no versification is forced to let a numbered verse fail to be a chanted verse — BHS simply does. BHS splits the chanted verse that MAM calls Num. 26:1 into a 25:19 span (a cv-label that does not exist in MAM) and a 26:1 span. BHS 26:1 starts right after an etnaḥta.

The Decalogues: background

Each of the Decalogues has two parallel strands of cantillation. The taḥton (תחתון) strand divides the passage into twelve chanted verses. The elyon (עליון) strand divides the passage into ten chanted verses: the Ten Commandments. Relative to the taḥton, the lengths of the elyon verses are uneven. There are two elyon verses that are quite long, containing, respectively "three and a half" and four taḥton verses. (That "half" will be explained below.) In contrast, there are four consecutive elyon verses that are very short: all four are contained within a single taḥton verse.

Aside: this document is only concerned with the cantillation of the Decalogues in the manuscript tradition; whether the cantillation of the Decalogues in the printed tradition is even grammatical is examined in a separate report, In the printed tradition, are the accents of the Decalogue grammatical?

The Decalogues: two versification policies

In the Decalogues, our versifications differ only in which chanted verses start with a cv-label:

From here on we discuss only MAM and BHS: describing Sefaria in detail adds nothing of interest, since it is merely a mix of the two. The Sefaria details can be found in the companion document.

The early drift

When an elyon verse contains only part of a taḥton verse, the BHS numbering drifts forward relative to MAM. The table below shows the first of two such drifts in the Decalogues. A span of two elyon verses contains four taḥton verses in their entirety, but they are not contained "nicely", for the reasons listed below.

E אנכ֖י…עבדֽים׃ ל֣א יהיה־לך֩…על־פנ֗י ל֣א תעשה־לך֣…לא֗רץ לא־תשתחו֣ה…לשנא֑י וע֤שה…מצותֽי׃
T אנכי֙…עבד֑ים לא־יהי֥ה לך֛…על־פנֽי׃ לא־תעש֨ה לך֥…לאֽרץ׃ לא־תשתחו֥ה…לשנאֽי׃ וע֥שה…מצותֽי׃
M 20:2 20:3 20:4 20:5
B 20:2 20:3 20:4 20:5 20:6
e
t

In the table above, each chanted verse's first and last word is shown in green and red respectively. Also, a schematic version of each chanted verse is shown at the bottom of the table, as a green-to-red color gradient. All examples in the body of this document are from the Exodus Decalogue; the appendix revisits these points in the Deuteronomy Decalogue, whose chanted verse structure is the same, with the numbered verse boundaries falling at the same places in it, despite some differences in the details of its text that are unimportant to our basic points here.

The middle no-drift case

When an elyon verse contains multiple taḥton verses in their entirety, there is no numbering drift between MAM and BHS. This is the case with the four consecutive taḥton verses in the table below, although unfortunately the point is somewhat muddied by a drift that has already happened, which we described above.

E זכור֩…לקדש֗ו ש֣שת…כל־מלאכת֒ך֒ וי֨ום…בשער֔יך כ֣י…ויקדשֽהו׃
T זכ֛ור…לקדשֽו׃ ש֤שת…כל־מלאכתֽך׃ ויום֙…בשערֽיך׃ כ֣י…ויקדשֽהו׃
M 20:7 20:8 20:9 20:10
B 20:8 20:9 20:10 20:11
e
t

The late drift

When an elyon verse contains only part of a taḥton verse, the BHS numbering drifts forward relative to MAM. The table below shows a drift due to four elyon verses each containing only a "quarter" of a taḥton verse.

E ל֖א…תרצֽח׃ ל֖א…תנאֽף׃ ל֖א…תגנֽב׃ לא־תענ֥ה…שֽקר׃
T ל֥א…תרצ֖ח ל֣א…תנא֑ף ל֣א…תגנ֔ב לא־תענ֥ה…שֽקר׃
M 20:12
B 20:13 20:14 20:15 20:16
e
t

Num. 25/26: a break at an etnaḥta, not a sof pasuq

Here there is only one cantillation, and BHS starts a numbered verse in the middle of a chanted verse, right after an etnaḥta.

MAM keeps Num. 26:1 as a single chanted verse with a mid-verse paragraph break inside it — a פסקא באמצע פסוק (in its petuḥah form). The break is shown below as a line break, which is what a petuḥah ("open" section) in fact is (with some qualifications we won't get into). Both lines are part of a single chanted verse.

ויה֖י אחר֣י המגפ֑ה
וי֤אמר … לאמֽר׃

The chanted verse ends, as always, at sof pasuq (לאמֽר׃). The word before the break, המגפ֑ה, has only etnaḥta — the strongest mid-verse disjunctive, but not a verse-ending accent. BHS further emphasizes that break by making its 25:19 end at המגפ֑ה — on an etnaḥta, not a sof pasuq. So BHS's 25:19 is not a chanted verse.

Why did BHS choose to start a numbered verse (and a chapter!) here? The etnaḥta and its accompanying פסקא באמצע פסוק already signal a strong break here; perhaps someone nonetheless felt that a chapter break was needed to emphasize the break at this point. Or perhaps this chapter break has its origin in a non-cantillated (perhaps even non-Hebrew) version of the text, where the cantillation-location of a chapter break is irrelevant (or at least invisible).

Note that the פסקא באמצע פסוק does not explain the chapter break. Even in the petuḥah form it takes here, this Numbers case is the only place in which פסקא באמצע פסוק causes a numbered verse break, much less a chapter break, in any of our versifications. (The פסקא באמצע פסוק phenomenon also has a setumah form, which is considered a weaker break than the petuḥah form.) See, for example, 1 Samuel 16:12 — one of many such verses. (This is just one of nine cases of petuḥah-form פסקא באמצע פסוק in the book of Samuel, out of twenty-four such cases in all of Tanakh. For some reason, this rare type of break is overrepresented in the book of Samuel.)

Why, then, does MAM not follow BHS here? Unlike the question of BHS's motive, this one cantillation can answer. BHS's break creates a numbered verse that is not a chanted verse: its 25:19, which ends at etnaḥta. In the Decalogues no versification can avoid such a verse — that is the "impossible" situation described above — but here avoiding it is entirely possible, and MAM avoids it. So, although ignoring cantillation may be what allowed BHS's break, a refusal to ignore cantillation — in particular, a refusal to have a numbered verse that is not also a chanted verse — is what may have made a versification like MAM's resist that break.

Appendix: the same points in Deuteronomy

Every example above is drawn from the Exodus Decalogue. The Deuteronomy Decalogue (MAM 5:6–5:17) has the same two strands of cantillation with the same chanted verse structure, so the same points hold there. What differs are aspects of the wording (substituted words, added phrases) that leave those cantillation and versification points intact. The numbered verse boundaries still fall at the same places in the chanted verse structure. This appendix shows the two places where that wording difference is visible. In the tables below a word is washed out when it is identical to its Exodus counterpart and shown at full strength when it differs, so the eye lands on the differences; first/last words keep their green/red hue, paled when they match Exodus.

The early split is not shown again here: through MAM's Deut. 5:9 the first and last words of the chanted verses are the same as those in Exodus, so its table would be uniformly washed out. (Words other than those first and last ones do differ — a dropped and an added connective vav, one plene spelling — but those words never reach this boundary-only table.) The differences live almost entirely in the Sabbath commandment, with only a connective vav and one substituted word in the late split.

The Sabbath — a different, longer text

Deuteronomy's Sabbath commandment opens with שמור (not Exodus's זכור) and is considerably longer, so although the taḥton/elyon chanted verse structure is unchanged — one elyon verse spanning four taḥton verses, exactly as in Exodus — most of the boundary words differ:

E שמ֣ור…אלה֗יך ש֣שת…כל־מלאכת֒ך֒ וי֨ום…כמ֑וך וזכרת֞…השבֽת׃
T שמ֛ור…אלהֽיך׃ ש֤שת…כל־מלאכתֽך׃ ויום֙…כמֽוך׃ וזכרת֗…השבֽת׃
M 5:11 5:12 5:13 5:14
B 5:12 5:13 5:14 5:15
e
t

Only the second taḥton verse (MAM 5:12) matches Exodus at both ends. The first (5:11) and fourth (5:14) differ at both ends; the third (5:13) shares its first word (ויום) but not its last word — Deuteronomy runs on past Exodus's בשעריך to כמוך.

The late split — nearly identical

Here the two Decalogues read very nearly alike. Deuteronomy joins the commandments with a connective vav (וְלֹא, "and you shall not") where Exodus sets them without it, and its ninth commandment ends עֵד שָׁוְא where Exodus has עֵד שָׁקֶר. Those wording differences aside, the taḥton/elyon chanted verse structure — four elyon verses nested in one taḥton verse, all ending together — is unchanged, so only וְלֹא and שָׁוְא stand at full strength while everything matching Exodus washes out:

E ל֖א…תרצֽח׃ ול֖א…תנאֽף׃ ול֖א…תגנֽב׃ ולא־תענ֥ה…שֽוא׃
T ל֥א…תרצ֖ח ול֣א…תנא֑ף ול֣א…תגנ֔ב ולא־תענ֥ה…שֽוא׃
M 5:16
B 5:17 5:18 5:19 5:20
e
t