Was the weekly Sabbath a late innovation?

[Difflugia]

The topic about executing a guy for picking up sticks on the Sabbath led to some further discussion about whether the weekly Sabbath observance was instituted sometime after the mid-sixth century BCE when Judah was defeated by the Babylonians and taken into exile.

Does the Jewish observance of a weekly Sabbath go back to the time of Moses (say, anytime before the tenth century BCE) or was it a late (exilic or post-exilic) innovation that was added to the Bible later?

[Difflugia]

Using Richard Elliott Friedman's The Bible with Sources Revealed as the source delineation, the Sabbath isn't mentioned in J or E at all. In Genesis through Numbers, the Sabbath is mentioned only in P passages or those of a late redactor (Ex. 20:1-11 is "likely to be an independent document, which was inserted here by the Redactor", p. 153). In Deuteronomy, the Sabbath only appears in its version of the Ten Commandments (which could be a late redaction as well, but even if not, it was potentially written just decades from the coming Exile). In the Deuteronomistic History, the Sabbath is only discussed three times (2 Kings 4:23, 11:5-9, and 16:18) and described simply as a ritual requiring one to "go out" and then "come in" (weird in light of Exodus 16:29). The only other pre-exilic mention is in Isaiah 1:13, which mentions it opposite the "new moon" (chodesh) in the same way that 2 Kings 4:23 does (is the pre-exilic Sabbath the full moon?).

Every other mention is exilic or post-exilic. The exilic prophets Jeremiah, Ezekiel, and Third Isaiah (Isa. 56 and following) talk about a weekly Sabbath, as do the Chronicles and Nehemiah. No pre-exilic prophets do.

Hosea 2:11 and Amos 8:5 along with Isaiah make three distinct prophets mentioning a Sabbath in the period even before the fall of the northern kingdom.

I don't know how I missed Hosea and Amos when I was looking before. Both of those actually reinforce the same pattern I mentioned earlier, though. Both times, Sabbath is used opposite the word translated "new moon" as though they're a pair of connected observances, but without further explanation. The context of Amos specifically sounds like chodesh and shabbat may mean the beginning and end of the harvest respectively, each having an associated festival. Note that Exodus 34:21-22 (in the "other" Ten Commandments) only specifically mentions a seventh day of rest at the beginning of the growing season and at the harvest.

The grouping of weekly & monthly holy days doesn't seem unusual - surely less so than having a pair of monthly holy days but having a lunar name for only one of them. But in particular, what about the first chapter of Genesis? I'm not familiar enough with all the ins and outs of how particular passages are believed to come from one source or another to find them particularly persuasive; as an amateur reader the presence of two obviously divergent creation myths would seem to be one of the best reasons for inferring two distinct early sources!

Genesis 1:1-2:3 is attributed by source critics to the P (priestly) source. Most source critics (notably excluding Friedman) date the P source to shortly after the return from the Babylonian exile in the early Persian period. As another datum, note that the creation story in Genesis 1 has a potential relationship to the Babylonian Enuma Elish.

The entire introduction to Friedman's The Bible with Sources Revealed can be read online at Google Books. It offers an overview of the Documentary Hypothesis and its sources that is brief and understandable, but sufficiently broad. I recommend Friedman's book as it reflects much more modern research, but some years ago, I put together a similarly color-coded Pentateuch from early twentieth century books in the public domain. It's online as a PDF. I also added a list of books that I found useful to my member notes.

Similarly with all the references to Sabbath in Exodus chapters 20, 23, 31 and 35; these chapters with such distinctions as makeshift altars of unhewn stone (ch.20) vs. portable altar of wood overlaid with bronze (ch.27) and exhaustive instructions for a portable tabernacle (ch.35ff) to my mind would suggest sources ultimately derived from two distinct and early/semi-nomadic sets of traditions. I'm not sure how they come to be assigned to a P source... and I'm always naturally suspicious of the ever-useful Redactor But my ignorance on those points is more or less tangential to the thread topic in any case.

According to Friedman, Exodus 20:2-10 (the decalogue) belongs to an unknown source that isn't JEDP, with verses 1 and 11 being added by the redactor. The remainder of 20 through 24:15a ("And Moses went up to the mountain.") is from the E source. Chapters 31 and 35 are both from the P source. There are a few differences (notably Carpenter assigning the decalogue itself to E), but Friedman mostly agrees with Carpenter in this section. The PDF above should help visualize what I'm talking about.

Of particular importance to your comment is the apparent reference to a Sabbath in Exodus 23:10-12. I admit that I also missed this one last night (I guess I need to be more attentive), but at the same time, the text as written seems to fit the Canaanite pattern that I mentioned last night. That is, it means that any six-day period of work or six-year period of agriculture must be followed by a day of rest or year of lying fallow, but doesn't necessarily mean that there's an established day or year that everyone observes together. From The Ugaritic Baal Cycle: Volume II by Mark S. Smith and Wayne T. Pitard, p. 615 (I've omitted the many references to specific texts for clarity):

The seven-day unit is, of course, a well-known one in ancient Near Eastern literature. It occurs very commonly in Ugaritic literature itself to convey either a complete or appropriate amount of time. In these instances (where the text is clearly preserved), the seventh day marks a shift in the activity in the scenes. Marching on days one through six gives way to arriving. King Pabil sleeps no more on the seventh day. [O]n the seventh day Baal draws near to Dan’il after six days of lamentation. [T]he Kotharat depart on the seventh day after six days of feasting.

From Stories from Ancient Canaan, by Michael Coogan and Mark S. Smith:

Seven days is the standard length of a journey (Kirta, tablet 1; Gen. 31:23) and of a wedding feast (Judg. 14:12); the firing of Baal’s palace takes a week as well. In these and other examples it is difficult to see more than literary convention in the choice of seven. But the alteration of the traditional mourning period from seven days (as in Job 2:13) to seven years after Aqhat’s murder heightens the connections between Danel’s son and the powers of fertility. Thus, in some cases we can agree with the observation that a period of seven days or years is “a time of great potency and fateful in its meaning” (H.-J. Kraus, Worship in Israel: A Cultic History of the Old Testament [Oxford: Basil Blackwell; Richmond, VA: John Knox Press, 1966], 86).

[Mithrae]

Hosea 2:11 and Amos 8:5 along with Isaiah make three distinct prophets mentioning a Sabbath in the period even before the fall of the northern kingdom. The grouping of weekly & monthly holy days doesn't seem unusual - surely less so than having a pair of monthly holy days but having a lunar name for only one of them.

I don't know how I missed Hosea and Amos when I was looking before. Both of those actually reinforce the same pattern I mentioned earlier, though. Both times, Sabbath is used opposite the word translated "new moon" as though they're a pair of connected observances, but without further explanation. The context of Amos specifically sounds like chodesh and shabbat may mean the beginning and end of the harvest respectively, each having an associated festival.

In 276 occurrences the KJV translates chodesh as month/monthly 255 times and as new moon 20 times. So it seems we've got the possibilities that there was in this period a common habit of comparing/associating the monthly and weekly holy days (consistent with later usage); or a monthly holy day, chodesh, with a holy day of some other indeterminate timeframe, sabbath; or two monthly holy days, one of which had a name apparently synonymous with 'month' and the other not. To my mind the first possibility is the most parsimonious: Why would there be two monthly holy days, or why would a monthly holy day become so regularly associated with an indeterminate one? And why would all of that change so radically further on down the line?

Note that Exodus 34:21-22 (in the "other" Ten Commandments) only specifically mentions a seventh day of rest at the beginning of the growing season and at the harvest.

That's not how I read it:
Exodus 34:21 Six days you shall work, but on the seventh day you shall rest; even in plowing time and in harvest time you shall rest. 22 You shall observe the festival of weeks, the first fruits of wheat harvest, and the festival of ingathering at the turn of the year. 23 Three times in the year all your males shall appear before the Lord God, the God of Israel.
The word "even" is italicized in the NASB and absent from the NKJV. But the NRSV, NIV and NET bibles all include it as part of their presentation of the authors' meaning, which makes sense to me: The subsequent verses speak of three annual holidays, so why would v21 be speaking of only two? It also doesn't specify the beginning of plowing or harvest time as you suggest; so given that they lasted more than a week, what would working for six days and resting on the seventh imply for those periods if not a regular, weekly day of rest?

But in particular, what about the first chapter of Genesis? I'm not familiar enough with all the ins and outs of how particular passages are believed to come from one source or another to find them particularly persuasive; as an amateur reader the presence of two obviously divergent creation myths would seem to be one of the best reasons for inferring two distinct early sources!

Genesis 1:1-2:3 is attributed by source critics to the P (priestly) source. Most source critics (notably excluding Friedman) date the P source to shortly after the return from the Babylonian exile in the early Persian period. As another datum, note that the creation story in Genesis 1 has a potential relationship to the Babylonian Enuma Elish.

The entire introduction to Friedman's The Bible with Sources Revealed can be read online at Google Books. It offers an overview of the Documentary Hypothesis and its sources that is brief and understandable, but sufficiently broad. I recommend Friedman's book as it reflects much more modern research, but some years ago, I put together a similarly color-coded Pentateuch from early twentieth century books in the public domain. It's online as a PDF.

That's very interesting and informative, thankyou. I was particularly interested in Friedman's explanation of the P source's focus on the tabernacle (which I was questioning) in terms of Hezekiah's reforms and centralization of worship in the period of influx of refugees from Israel. If artifacts like the ark of the covenant and tabernacle actually did exist (which seems likely) and still existed towards the end of the first temple period but not into the second (both of which seem plausible), then Friedman's arguments for a Hezekian dating of P seem very persuasive to me.

I'm still not seeing why Genesis 1 should be assigned to P rather than to E however, besides perhaps the belief that it is dependent on Babylonian mythology. Going through Friedman's seven lines of evidence, the only ones which seem potentially relevant to Gen. 1 are possibly linguistic dating; the presence of the phrase "be fruitful and multiply"; and the implied discontinuity of the sources' narratives if chapter 1 and only chapter 1 were instead assigned to E. On the other hand doesn't it seem strange to suppose that the J and E sources, so similar in period and purpose, should begin at such different points in the chronology? From Friedman's description, until the introduction of the priesthood there seems to be relatively little distinguishing P and E in general terms, so does that mean that the designation of the non-J sections in early Genesis to P rather than E depends primarily on specific terminology and language characteristics? If so then the use of particular terminology in particular can obviously be quite questionable on occasion: Friedman rightly notes that in general terms the convergence of many of these lines of evidence makes a powerful case, but I'd be surprised if it's always the case for every particular designation of a passage that it contains multiple clear identifying characteristics.

For example while going down those terminology points he lists to see if/how clearly they weigh against switching the P designation of early Genesis material to E instead, his listing of rekuwsh (property) seems a little questionable. Friedman suggests that five of the thirteen occurrences of the term are anomalous or uncertain to begin with, and glancing at a couple of the other early occurrences (Gen. 12:5 and 13:6) it seems that in your PDF those specific verses alone are assigned to the P source in the midst of J content. Purely because they contain that 'P' word? Surely not; if so it would be on the shaky basis (even assuming that every other instance is clear and indisputable) that six out of thirteen occurrences are definitely P material. Maybe there are some other telltale terms in those verses that I'm missing.

But presumably there've been plenty of scholars considering the same possibility, so there must be good reason for the view that the E source began only with the story of Abraham.

Similarly with all the references to Sabbath in Exodus chapters 20, 23, 31 and 35; these chapters with such distinctions as makeshift altars of unhewn stone (ch.20) vs. portable altar of wood overlaid with bronze (ch.27) and exhaustive instructions for a portable tabernacle (ch.35ff) to my mind would suggest sources ultimately derived from two distinct and early/semi-nomadic sets of traditions. I'm not sure how they come to be assigned to a P source... and I'm always naturally suspicious of the ever-useful Redactor But my ignorance on those points is more or less tangential to the thread topic in any case.

According to Friedman, Exodus 20:2-10 (the decalogue) belongs to an unknown source that isn't JEDP, with verses 1 and 11 being added by the redactor. The remainder of 20 through 24:15a ("And Moses went up to the mountain.") is from the E source. Chapters 31 and 35 are both from the P source. There are a few differences (notably Carpenter assigning the decalogue itself to E), but Friedman mostly agrees with Carpenter in this section. The PDF above should help visualize what I'm talking about.

Of particular importance to your comment is the apparent reference to a Sabbath in Exodus 23:10-12. I admit that I also missed this one last night (I guess I need to be more attentive), but at the same time, the text as written seems to fit the Canaanite pattern that I mentioned last night. That is, it means that any six-day period of work or six-year period of agriculture must be followed by a day of rest or year of lying fallow, but doesn't necessarily mean that there's an established day or year that everyone observes together.

Surely the same could be said of virtually every reference to Sabbath in every text? Maybe the guy in Numbers was caught out gathering wood on his own personal day of rest Exodus 20:8 says that the Sabbath applies to all servants, animals and resident foreigners, so that must be a universal day of rest... but then, 23:12 says the same thing! Another passage which springs to mind as a counter-example is the implication in the story of manna from heaven that everyone was gathering - and refraining from gathering - on the same day: Your PDF has most of Exodus 16 assigned to the P source, but 16:5 assigned to J.

So regardless of what we conclude about the P source and Genesis 1, it seems that both J (Ex. 16:5 and 34:21) and E (23:12 and perhaps 20:8) each suggest the existence of a weekly, universal day of rest prior to the Assyrian conquest of Israel - arguably backed up by references in three of the earliest prophets - which presumably implies a tradition going back before the division of the kingdoms.

[gadfly]

The historical question of when the Sabbath legal restrictions began is impossible to determine. I for one see no reason why it should not be dated to when it first appears. The burden of proof rests on those who insist it is a late invention

[Difflugia]

I'm pretty much out of new arguments. I want to address some of the topics you brought up because I find source redaction and its implications for Israelite history fascinating, but don't want it to come across as trying to stretch more evidence out of the text than what's there.

In 276 occurrences the KJV translates chodesh as month/monthly 255 times and as new moon 20 times. So it seems we've got the possibilities that there was in this period a common habit of comparing/associating the monthly and weekly holy days (consistent with later usage); or a monthly holy day, chodesh, with a holy day of some other indeterminate timeframe, sabbath; or two monthly holy days, one of which had a name apparently synonymous with 'month' and the other not. To my mind the first possibility is the most parsimonious: Why would there be two monthly holy days, or why would a monthly holy day become so regularly associated with an indeterminate one?

After reading through the non-P references some more, my speculation at this point is that the Sabbaths were originally the distinct phases of the moon (new moon, quarter moon, full moon, quarter moon) with the new moon marking the beginning of the cycle. It's a pretty short distance from there to the Sabbaths occurring exactly seven days apart and no longer coinciding with moon phases.

And why would all of that change so radically further on down the line?

As long as we're still speculating, there are a number of upheavals in Israelite culture that are indicated in the Bible. I could very easily see a number of traditions being retained, but original reasons being lost during any of these. First, there was the Canaan-Israel shift. Whether it was a military conquest as the book of Joshua has it or simply a collapse of the Canaanite city states as archaeologists currently think, there would certainly be a mixed bag of continuity or discontinuity between "pre event" and "post event" observance. Then there are the periods of the United Monarchy, its collapse, the reforms of Josiah (likely corresponding to the D sources), the exile, and then the return to Palestine.

If the source critics are even a little bit right, then we have a combination of texts from both sides of each of these periods of change with redactors trying with varying success to assemble the parts into a cohesive whole. Even leaving source criticism aside, there are pretty fundamental differences in reported sacrificial tradition between the Samuel-Kings history, the prophets, and the Chronicles. This isn't exactly added evidence for a shift in Sabbath observance, but it certainly creates a framework in which even relatively large changes are plausible.

Note that Exodus 34:21-22 (in the "other" Ten Commandments) only specifically mentions a seventh day of rest at the beginning of the growing season and at the harvest.

That's not how I read it:
Exodus 34:21 Six days you shall work, but on the seventh day you shall rest; even in plowing time and in harvest time you shall rest. 22 You shall observe the festival of weeks, the first fruits of wheat harvest, and the festival of ingathering at the turn of the year. 23 Three times in the year all your males shall appear before the Lord God, the God of Israel.
The word "even" is italicized in the NASB and absent from the NKJV. But the NRSV, NIV and NET bibles all include it as part of their presentation of the authors' meaning, which makes sense to me:

I think one could read it there or not. I read the two clauses of verse 21 as an example of Hebrew paralellism in which the second clause is a repetition of meaning of the first. So, rather than a sense of "rest every seventh day, including when planting and harvesting," the work is intended to mean specifically agricultural work. I'm not sure that offers enough distinction, though, since the wording may just be a relic from an older time when all of the hard work that one might need to "rest" from was agricultural work and the clauses might be practical synonyms even if not technical ones.

The subsequent verses speak of three annual holidays, so why would v21 be speaking of only two?

Reading it again, I think you're right about this. As an alternate Ten Commandments, I see v. 22 starting a new one. Verse 21 is about work and rest and 22-24 are about agricultural festivals. At the moment, I'm trying to work out if it's meaninful that the "day of rest" here nowhere includes the designation of "Sabbath" as a noun, though one could potentially read "...on the seventh day you shall Sabbath..." with "Sabbath" as a verb.

It also doesn't specify the beginning of plowing or harvest time as you suggest; so given that they lasted more than a week, what would working for six days and resting on the seventh imply for those periods if not a regular, weekly day of rest?

That was me reading something into the terms that I didn't actually express. The verb "chadash" broadly means "to renew," from which sense it's associated with the new moon. The verb "Shabath" broadly means "to stop." I was seeing a pattern that might pair a "renewal" of the growing season by cultivation, which was then paired with the end of the season at harvest. The "seven days" was then (as in the Ugaritic examples) the "correct" amount of time for such an undertaking, after which one rested.

That's very interesting and informative, thankyou. I was particularly interested in Friedman's explanation of the P source's focus on the tabernacle (which I was questioning) in terms of Hezekiah's reforms and centralization of worship in the period of influx of refugees from Israel. If artifacts like the ark of the covenant and tabernacle actually did exist (which seems likely) and still existed towards the end of the first temple period but not into the second (both of which seem plausible), then Friedman's arguments for a Hezekian dating of P seem very persuasive to me.

The important disconnect here is that Friedman thinks that the Tabernacle was real, while most source critics consider the Tabernacle to be a fictional stand-in for the Temple. If the question is something you find interesting, the counterpoint to Friedman's position is still Wellhausen's from the early twentieth century. I suggest Friedman's Who Wrote the Bible first. I found Wellhausen's writing to be both informative and entertaining once I had at least a bit of background in the subject, but Friedman's prose is far more accessible and easier to understand than the nearly literal translation of Wellhausen's dense, jargon-filled German.

Friedman makes a compelling series of arguments, but I think he completely ignores several of Wellhausen's arguments that are just as compelling. Wellhausen notes, for example, a number of textual details surrounding the Tabernacle that make no sense for a temporary structure that is carried and rebuilt, but would make sense as details of the Temple anachronistically projected onto the Tabernacle. Friedman's overall thesis about the Tabernacle seems to be that if he can find enough context in which a real Tabernacle might plausibly fit, then the textual issues become mere annoyances rather than any sort of refutation. In this vein, his relatively recent book The Exodus further explores the role of the Tabernacle in the Exodus.

Joel S. Baden is a contemporary source critic that argues for a late P and he disagrees with Friedman on a number of points. Unfortunately, I find his writing dry and difficult to read.

I'm still not seeing why Genesis 1 should be assigned to P rather than to E however, besides perhaps the belief that it is dependent on Babylonian mythology.

The determination that Genesis 1 is P is based mostly on style. From Brightman (linked in the reading list), p. 204:

The style of P is very marked and is entirely different from that of J or E. It is formal, repetitious, precise, abstract in descriptions of Deity, yet minutely concrete in descriptions of objects, such as the tabernacle; legal, statistical, but usually dignified and elevated, and sometimes sublime, as in Gn. 1.

It is characterized by interest in genealogy. "There is a tendency to describe an object in full each time that it is mentioned; a direction is followed, as a rule, by an accountof its execution, usually in the same words. Sometimes the circumstantiality leads to diffuseness, as in parts of Nu. 1-4 and (an extreme case) Nu. 7" (Driver).

P is a literalist. "Metaphors, similes, etc., are eschewed (Nu. 27:17b is an exception) and there is generally an absence of the poetical or dramatic element’’ (Driver). The factor of "learned" editing, of fitting everything into a theory about the past, is far more highly developed in P than in E (Steuernagel).

The "Driver" reference is to S. R. Driver's The Book of Genesis commentary (Google Books scan; the "P" section begins at the bottom of the linked page).

On the other hand doesn't it seem strange to suppose that the J and E sources, so similar in period and purpose, should begin at such different points in the chronology?

Friedman himself has expressed frustration that we no longer have E's version of the Creation account. He assumes that there must have been one because what is now the beginning of E picks up in the middle of the Abraham story.

From Friedman's description, until the introduction of the priesthood there seems to be relatively little distinguishing P and E in general terms, so does that mean that the designation of the non-J sections in early Genesis to P rather than E depends primarily on specific terminology and language characteristics? If so then the use of particular terminology in particular can obviously be quite questionable on occasion: Friedman rightly notes that in general terms the convergence of many of these lines of evidence makes a powerful case, but I'd be surprised if it's always the case for every particular designation of a passage that it contains multiple clear identifying characteristics.

That your observation is valid can be seen by the number of alternate arrangements offered by different scholars. Each of the sources I offered (Wellhausen, Carpenter, Brightman, Friedman, Baden) has at least minor differences in the divisions. Even if one argues that some of the criteria approach being objective, their application doesn't always provide a clear result, especially when shorter passages are interleaved.

For example while going down those terminology points he lists to see if/how clearly they weigh against switching the P designation of early Genesis material to E instead, his listing of rekuwsh (property) seems a little questionable. Friedman suggests that five of the thirteen occurrences of the term are anomalous or uncertain to begin with, and glancing at a couple of the other early occurrences (Gen. 12:5 and 13:6) it seems that in your PDF those specific verses alone are assigned to the P source in the midst of J content. Purely because they contain that 'P' word? Surely not; if so it would be on the shaky basis (even assuming that every other instance is clear and indisputable) that six out of thirteen occurrences are definitely P material. Maybe there are some other telltale terms in those verses that I'm missing.

My PDF is based on a single source from 1918. I've compared it with Friedman and there are a number of discrepancies. One of the glaring differences is that Friedman is convinced that each source can be read as a continuous narrative, in order (i.e. that the redactors interleaved verses, but didn't rearrange their order from the original), which has become part of his methodology. Whenever there's such a difference, at least some subjective judgement must have been involved.

I confess I haven't spent much time thinking about why specific verses were attributed to which sources in most cases, though. I usually read the other way (i.e. if this chapter is early and that one is late, what does that imply about the evolution of Old Testament theology?).

That is, it means that any six-day period of work or six-year period of agriculture must be followed by a day of rest or year of lying fallow, but doesn't necessarily mean that there's an established day or year that everyone observes together.

Surely the same could be said of virtually every reference to Sabbath in every text? Maybe the guy in Numbers was caught out gathering wood on his own personal day of rest Exodus 20:8 says that the Sabbath applies to all servants, animals and resident foreigners, so that must be a universal day of rest... but then, 23:12 says the same thing!

I concede that particular point, but I will also point out that 23:12 is written with possessive forms (your ox and your donkey and your servan't son, just not your foreigner, if that's meaningful). I take that to mean (whether the Sabbath is the shared by other Israelites) that in addition to you, it also applies to anyone working for you. When you rest, your "employees" (even if they're animals or slaves) rest, also.

Another passage which springs to mind as a counter-example is the implication in the story of manna from heaven that everyone was gathering - and refraining from gathering - on the same day: Your PDF has most of Exodus 16 assigned to the P source, but 16:5 assigned to J.

And Friedman agrees, which is interesting if for no other reason than it is one of the very few verses he attributes to J that even obliquely mentions the Sabbath. I'd say the only verse, but to be fair, I didn't notice that one, either, so I'd better come up with some broader search parameters.

So regardless of what we conclude about the P source and Genesis 1, it seems that both J (Ex. 16:5 and 34:21) and E (23:12 and perhaps 20:8) each suggest the existence of a weekly, universal day of rest prior to the Assyrian conquest of Israel - arguably backed up by references in three of the earliest prophets - which presumably implies a tradition going back before the division of the kingdoms.

I'll have to think about this some more (and I'm a bit muddled after the rest of our discussion), but I think at this point I'm ready to concede that the six-day/one-day, work/Sabbath pattern definitely predates both D and P. I'm still not entirely sold on the continuous, every week thing, mostly because it looks possible to me that prior to the Exile, the Sabbath (as a noun) refers to a lunar event rather than the day of rest (essentially Sabbath as a verb). I'm sure that at some point they were combined, but the question is still whether that was early or late.