Deut 32:17 (BHS | ESV):

יִזְבְּחוּ לַשֵּׁדִים לֹא אֱלֹהַ
yizbĕḥû laššēdı̂m lōʾ ʾĕlōah
They sacrificed to demons that were no gods

אֱלֹהִים לֹא יְדָעוּם
ʾĕlōhı̂m lōʾ yĕdāʿûm
to gods they had never known,

חֲדָשִׁים מִקָּרֹב בָּ֔אוּ
ḥădāšı̂m miqqārōb bāʾû
to new gods that had come recently,

לֹא שְׂעָרוּם אֲבֹתֵיכֶם
lōʾ śĕʿārûm ʾăbōtêkem
whom your fathers had never dreaded.

This seems a little bit odd because, according to the ESV, the šēdı̂m ("demons") are called "no gods" (line 1), and then their character is elaborated (lines 2-4) using the label "gods".* However, the NET does something different with the first line:

They sacrificed to demons, not God...

This reflects the singular ʾĕlōah in the fist line. It also removes the unmarked relative clause, and with it the suggestion that the šēdı̂m are anything other than fully ʾĕlōhı̂m.

Are the šēdı̂m called "no gods" here?

* At least I think these are the same beings, though one would prefer to see the preposition lĕ- repeated for clarity.

† For those who care, Deut 32:17 is indeed among the unmarked RC's compiled by Robert Holmstedt. There are three of them in that verse, but 17a is explicitly included on p. 110 (though he translates "demons [who] are not divine").

  • 3
    Check out this paper by Michael Heiser also. I think he nails it
    – Dan
    Commented Oct 14, 2016 at 15:59
  • He does indeed, thanks Dan! (Also helpful for 32:21, the syntax of which has bothered me before.)
    – Susan
    Commented Oct 14, 2016 at 16:09
  • Daniel Strange also address this passage in his book 'For their Rock is Not as Our Rock' (IVP, 2014).
    – LiamM
    Commented Oct 18, 2016 at 22:21
  • To give some idea how this entire translation is awful, the word for "Demons | לַשֵּׁדִים" is actually one of the names for God, (El Shaddai | singular). It might seem the LXX affirms this interpretation, (δαιμονίοις), but it is more complicated, (Perseus Definition, δαιμονίοις). It is hard to reasonably frame a response given the mistranslations in this passage. Another question might be helpful, for what "לַשֵּׁדִים" means - in other contexts. Commented Jul 2, 2017 at 22:35
  • @elikakohen There is no connection between שדים and אל שדי. And certainly שדים is not a name of God anywhere in the OT and certainly the pointalization לַשֵּׁדִים, "to the demons", is not a name of God.
    – user17080
    Commented Jul 3, 2017 at 9:16

2 Answers 2


The NET is apparently translating with the LXX and the Samaritan Torah, against the Masoretic tradition and Onkelos and the Jerusalem Targum. This is a difficult translation in the Masoretic view. The NET choice might be based on the opinion that the LXX and the Samaritan versions reflect an earlier, more authentic tradition rather than on an actual translation of the Masoretic text that the OP quotes.

The Masoretic interpretation

The Masoretic reading is that "lo eloha" is an adjectival phrase, parallel to "lo y'daum" in the second line and "mikarov bau" in the third line, and similar in construction and meaning to "b'lo el" ("with a no-god") in Deut 32:21. So "eloha" is not really singular nor plural in this context.

To be read as "not to God" we would expect "eloha" to have a "lamed" prefix meaning "to" in parallel with the "lamed" prefixed to "shedim" that gives us "to demons". English supports an assumed "to", there being little difference in meaning between "not God" and "not to God", but Hebrew has no such implied "to". You really need the "lamed".

If "eloha" were a corrupted form of "Elohim" in the sense of God then then there would be a problem with the following word, also "elohim", but in the sense of "gods". Juxtaposing these two senses of the same spelling in the same verse would be jarring and would break the thematic unity of the verse which is "useless foreign gods that your ancestors were never beholden to".

Holmstedt's "demons [who] are not divine" is a spot on Masoretic translation, similar to Onkelos's impotent or useless gods,

שֵׁידִין דְּלֵית בְּהוֹן צְרוֹךְ

Gods that are no gods is a common theme, as in the text following the OP quote, Deuteronomy 32:21

הֵם קִנְאוּנִי בְלֹא אֵל כִּעֲסוּנִי בְּהַבְלֵיהֶם

and as in Jeremiah 16:20,

הֲיַעֲשֶׂה-לּוֹ אָדָם, אֱלֹהִים; וְהֵמָּה, לֹא אֱלֹהִים

and as in Isaiah 37:19,

נָתֹן אֶת אֱלֹהֵיהֶם בָּאֵשׁ כִּי לֹא אֱלֹהִים הֵמָּה

and as in the imagery of Deuteronomy 4:28,

וַעֲבַדְתֶּם שָׁם אֱלֹהִים מַעֲשֵׂה יְדֵי אָדָם
עֵץ וָאֶבֶן
אֲשֶׁר לֹא יִרְאוּן וְלֹא יִשְׁמְעוּן וְלֹא יֹאכְלוּן וְלֹא יְרִיחֻן

and Psalm 115:5

פֶּה לָהֶם וְלֹא יְדַבֵּרוּ עֵינַיִם לָהֶם וְלֹא יִרְאוּ

so there is no need to see a possible doctrinal issue here.

Thematic considerations

When reading Deut 32:15-17 together as a single paragraph broken into two parts, we see a list of actions in consecutive clauses:

  1. rejected the "Eloha" who made him (15)
  2. spurned the Rock of his salvation (15)
  3. made [Him] jealous with foreign gods (16)
  4. angered [Him] with abominations (16)
  5. sacrificed to demons, not "eloha"/"Eloha" (17)

followed by a denigration of the gods and demons,

  1. gods with whom they were not familiar (17)
  2. new ones, recent arrivals (17)
  3. that their fathers had never feared (17)

In each of the clauses of the first group there is either an explicit or third person reference to God. The group begins and ends with "eloha", apparently an intentional symmetry. In the second group there is no reference to God, only to the gods and demons. This supports the NET reading.

Spelling considerations

The spelling of "eloha" in Deut 32:17 is without a "vav" following the "lamed", like the "l'elohim" in Deut 32:3. This form of abbreviated spelling, where a letter representing a vowel sound is dropped, is very common, although it is unique for the word "eloha" (not counting Kings II 17:31, which is probably a copyist mistake). In the case of Deut 32:17 as in Deut 32:3 it is not necessarily a mistake. It might simple preserve an older scribal tradition for this particular verse. There is no cognate basis for taking this spelling as a hint for reading "eloha" rather than "Eloha".

The scribal tradition

The scribal tradition is that the "eloha" of Deut 32:17 is a doubtful instance of a divine name, which means that it must be written with the intent of writing a divine name to be on the safe side. This gives some credence to the NET translation in that it preserves a masoretic understanding that this "eloha" might mean "God".

In the scribal tradition, verses 1-43 of Deut 32 are written in two columns of equal width separated by an empty column of the same width. The total width of the three columns is the same as the width of a regular column of the scroll. The text in each column is fully justified by lengthening specific letters. The reading order is right-to-left, one line from each column (rather than reading all of the first column then the second). Verses can start at the right margin of either column and end at the left margin of either column. No verse fits into one column only. The longest verses require two-and-a-half rows. The longest row of either column is 22 letters, including the letter width spaces required between words. The first clause of Deut 32:17 usually starts in the second (left) column and has 18 characters and spaces. It is one of the more cramped clauses in terms of the lettering because many of the letters are wide letters. However, it is not the most cramped clause and the addition of a "vav", which is a narrow character would still not make it the most cramped. So a typographical explanation for the the missing "vav" is probably ruled out.

Also, the first and third clauses of Deut 32:17 and the first clause of Deut 32:18 all end in consecutive rows the left column, each with a tri-gram, "eloha" (אלה), "ba'u" (באו), and "teshi" (תשי). However, the letters do not form any acrostic nor do the words have any meaning when combined, so this is likely a typographical coincidence rather than an explanation for the missing "vav" in "eloha". There is one run of four left-column ending tri-grams in Deut 32:27-30, that also appears to be coincidental.

The reading tradition

In the reading tradition (Yemenite) the "eloha" of Deut 32:17 without the "vav" is pronounced in exactly the same way as all other instances of "eloha" with a "vav", with the accent on the second syllable, "lo". That is, the dropped "vav" does not affect the pronunciation or move the accent to the last syllable, "ha".

Further considerations

Daniel 11:37-38 contains a parallel construction to Deut 32:17 (according to the Daat Mikra) and contains three instances of "eloha", all of which are references to the gods. The intent is apparently to use well known pejorative language of Deut 31:17 to denigrate the Northern King.

A fragment from Qumran that contains Deut 32:17 could shed light on the question either by showing us an alternative spelling, or by the use of the ancient Hebrew script for "eloha". I know of no such fragment.

  • 1
    Thank you for pointing that out about the LXX + SP, although I think we can count on the NET to have told us if they were intentionally translating from something other than MT. I think you're probably right that the parallelism with the following two lines is the best argument for maintaining lōʾ ʾĕlōah as a modifier of what precedes. +1 from me. (If you could elaborate on "a rhyme which would be inappropriate" -- I'm very curious what's behind that.)
    – Susan
    Commented Oct 13, 2016 at 15:14
  • @Susan On second look, the problem is more thematic than too many "eem"s in the same verse. It wouldn't break the meter or sound that strange. See my edit.
    – user17080
    Commented Oct 15, 2016 at 17:50
  • @Susan Please reset the edit count on this post if possible, I might have a few more things to add. Thanks.
    – user17080
    Commented Oct 22, 2016 at 23:05
  • Thanks for your ongoing attention to this Q. I'm not able to reset the edit count, but there's no need; you can edit as many times as you'd like. (But please note that every time you do so, it bumps this Q&A to the top of the "active" page which can be annoying for others who use the page to view new content. For that reason, we recommend that you batch your edits. You can see a full preview of what the post will look like just below the editing window, so there's no reason to repeatedly save and the re-edit.)
    – Susan
    Commented Oct 23, 2016 at 0:11
  • Correction: "Gods that are no gods is a common expression, as in Deuteronomy 4:28". This really, really, needs support. Especially since Deuteronomy 4:28 doesn't actually have that wording, as claimed here. So, the subsequent conclusions get increasingly invalid - if that statement is not actually true. Commented Jul 2, 2017 at 22:54

I would render the verse like this:

enter image description here

A couple of things to note:

  1. The שֵּׁדִים֙ is a subset of the אֱלֹהִ֖ים. The need for a different word springs from the author's desire to communicate that they are a different class of Elohim (spirit beings) who are malevolent ones, in contrast to their אֱלֹ֔הַ who is a benevolent one.

  2. חֲדָשִׁים֙ is an adjective being deployed as a noun, and should be translated as "new ones", not "new gods". New gods or new demons? It doesn't matter, since the author has told us they're from the same stable.


The second לֹ֣א is an adverb modifying the verb יְדָע֑וּם, which becomes "Elohim not known to them". Just like the first לֹ֣א modifies the verb יִזְבְּח֗וּ, i.e. "they did sacrifice to shedim" and "they did not sacrifice to Eloha".

So, no. The shedim are not being called "no gods" here.

* For those who noticed my answer had disappeared for a bit: the app I've developed for analyzing the text, omitted the first word of the verse.

  • On #3, see the first word in the verse: יזבחו. Also, what is meant by "pseudonym" here? Maybe you mean "synonym"?
    – Susan
    Commented Oct 22, 2016 at 2:42
  • I had edited the answer significantly before your comment. Please refresh to get the edited version.
    – enegue
    Commented Oct 22, 2016 at 2:46
  • @enegue - Correction: - There is no "Malevolent" connotation in "שֵּׁדִים֙" - in Hebrew texts. After all, the singular version is used as one of the names of God. That "Malevolent" connotation is purely "pragmatic", a "guess" as to the author's intent. It is certainly not a linguistic / semantic inference that "שֵּׁדִים֙" means "demons/malevolence" at all. That inference can only be made IF AND ONLY IF the context supports it. Modern "Christian" cosmology does not reflect Biblical cosmology, at all. There is a lot more than "God, Satan, Angels, and Demons" in Scripture. Commented Jul 2, 2017 at 22:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.