God Didn't Say That

Bible Translations and Mistranslations

What did God Really Create in the Beginning?

What did God create in the beginning?

The usual answer is as obvious as it is wrong: “heaven and earth.”

The problem is that the Hebrew for the first word here means “sky,” not “heaven.” In English, the birds, clouds, rain, etc. are all in the sky, not in heaven. Heaven, by contrast, is, depending on one’s theology, either where good people go when they die or where all people go when they die.

A translation variation, “heavens,” is a little better, but only to the extent that that Biblish word has entered the mainstream. People don’t talk about “cloudy heavens” when it’s overcast. They talk about a cloudy sky.

We see the Hebrew word, shamayim, ten times in the first chapter of Genesis.

The final four times the word is where birds are, which is obviously the “sky” in English, not “heaven” or “heavens.”

Four times the word appears in connection with the Hebrew raki’a, which is usually translated into English as “firmament” — though, again, that’s a word whose use is almost entirely confined to translations of Genesis; the NRSV’s “dome” isn’t a bad alternative. The raki’a is the ancient conception of the sky, which is why the Hebrew raki’a is God’s name for the shamayim, in one place, just like “day” is God’s name for “light.”

In one case, the shamayim is the place under which the water of the ocean is gathered — again, “sky” in English.

And that leaves Genesis 1:1, where God creates the shamayim. (If you’re counting along, it seems like we now have eleven instances, not ten, but only because one of them appears in two lists — in connection with raki’a and in connection with birds.)

Elsewhere in the Bible (Deut. 11:17, e.g.), a lack of rain results when the shamayim gets stopped up. The shamayim is where the stars are (Gen. 26:4). And so forth. All of these are “sky” in English.

So it seems to me that Genesis 1:1 should talk about the “the sky and the land” or “the sky and the earth.”

The only possible reason I can think of not to go with this clear translation is that the Hebrew pair shamayim and eretz is used metaphorically (as a merism) to represent all of creation. (This is presumably why the ISV goes with “universe” here. But in turning the pair “sky/earth” into the one word “universe,” the ISV destroys the dualism that underlies the creation story.)

So what do you think? Is there any reason to keep the common translation “heaven(s)”?


October 9, 2013 Posted by | Bible versions, translation practice | , , , , | 13 Comments

On Genesis 1:1

While most translations agree that the translation of Genesis 1:1 should read, “In the beginning…” the (Jewish) JPS translation offers instead, “When God began to create…” And the NLT and some others offer a footnote with that possibility. What’s going on?

The answer dates back 1,000 years to Rashi. He notes that the usual word for “in the beginning” would be barishona. And he further notes that b’reishit is never used except preceding a noun to mean “at the beginning of.”

He therefore concludes that Genesis 1:1 does not say that creation took place “in the beginning,” but rather that it was “in the beginning of” creation that the first part of the story takes place. That is, the earth was in disarray when God began to create.

Rashi’s analysis gives us, “When God began to create,” or (as the translation in Artscroll’s Rashi edition has it) “In the beginning of God’s creating.”

Rashi’s analysis has at least two kinds of problems.

The first is a matter of detail. For his analysis to work, he needs the verb bara to be a participle, though it’s unclear how that’s possible. Secondly, he needs the “and” of “and the earth was…” to mean “when.” That one is possible, though unlikely.

The second kind of problem, though, is methodological.

Rashi is right that b’reishit is never used except before a noun, but there are only four other times the word is used, all of them in Jeremiah, and all of them before words having to do with “kingdom” or “reign.” This is hardly a large enough sample to deduce what b’reishit means. (The same reasoning would force bara to mean something about kingdoms.)

Rashi’s point is actually more generally about reishit. (The b- prefix means “in/when/at/etc.”) But here, too, he runs into problems, wrongly assuming that a word is the sum of its parts.

Furthermore, while Rashi is correct that barishona means “at first,” that doesn’t really have much bearing on what b’reishit means. Perhaps the two words are nearly synonymous, for example. Or maybe barishona means “at first” in the sense of “the first time around” while b’reishit means “at first” in the sense of “the first and only time around.” (I just met someone who introduces his wife as his “first wife.” She is his first, only, and last wife.)

All of which is to say that Rashi’s commentary here is interesting — and it explains the JPS translation — but I don’t think it helps figure out what the first words of the Bible originally meant.

I have more on Genesis 1:1 here, here, and here.

July 19, 2010 Posted by | translation practice | , , , , , , | 13 Comments

Top Translation Traps: Mimicry

One of the most non-intuitive aspects of translation is that mimicry can lead the translator astray.

For example, it stands to reason that an adverb at the beginning of Hebrew sentence should be translated into English by an adverb at the beginning of a sentence; and, similarly, that an adverb at the end in Hebrew should be rendered as a final adverb in English.

But the reasoning is flawed. This is why the translation of the Hebrew word breishit (“in the beginning”), may not belong at the start of Genesis 1:1.

Modern Languages

As usual, we can look at modern languages to get a sense of the situation. This time, we’ll look at word order in modern Russian and modern Hebrew.

Spoken colloquial Russian allows considerable word-order variation. In fact, in the Russian equivalent of “Yesterday John saw Sarah” (v’chera Ivan videl Saru), all 24 logically possible word orders are grammatical. By contrast, the written language is more restrictive in Russian, generally requiring something close to what we allow in English.

Modern Hebrew also allows more word-order variation than English. But in Hebrew, it’s the written language that is more flexible than the spoken one. So in written Hebrew, again all 24 logically possible word orders are grammatical, but the spoken language is more restrictive.

The naive way to translate Russian into Hebrew is to preserve the word order. After all, the Russian word order is always grammatical in Hebrew (in this example). But mimicking the word order sometimes take colloquial Russian and turns it into formal Hebrew.

This demonstrates what can go wrong when translators mimic instead of translating.


Rather than merely mimicking the original word-order, translators need to look at what the word order in Hebrew or Greek does, and then try to do the same thing in English. More generally, I think this lesson applies not just to word order but also to other aspects of grammar.


Perhaps the most mimicry-based translation is Dr. Everett Fox’s (and this is the problem I have with his work). To take an example that just came up in a discussion on BBB, we can look at part of his rendering of Genesis 22. In verse 2, he translates the Hebrew y’chidcha as “your only-one.” Presumably the hyphenated “only-one” is supposed to mimic the one-word Hebrew yachid. But in (partially) mimicking the number of words, Fox has taken ordinary Hebrew and turned it into bizarre English.

Fox’s “started-early” (verse 3) for the one Hebrew word hishkim makes the same mistake.

Dr. Robert Alter’s rendition of Psalm 104 — which he discusses in the introduction to his generally excellent The Book of Psalms — makes the same mistake. He gives us, “grandeur and glory You don,” which is barely English. He wants to preserve what he calls syntactic fronting. The problem is that the resulting English is bizarre in a way that the Hebrew never was.

A third example comes from Matthew 6:11. The common “give us this day our daily bread” mimics the Greek word order rather than translating it. It is well known that putting full phrases between the verb and its object in English is an odd word order. So “this day” doesn’t belong between “give us” and “our daily bread.” (Also “daily bread” might be wrong, but that’s for another time.)

I don’t know if this strange word order comes from the Greek (which reads: “our daily bread give us today”) or just from the KJV (which was written in a dialect that allowed for more word-order freedom). Either way, the right way to translate the Greek is to use English word order: “Give us our daily bread today” (again, if “daily bread” is right.)

May 28, 2010 Posted by | translation practice, translation theory, Translation Traps | , , , , | 8 Comments

First Things First: Stress, Focus, and Biblical Hebrew Word Order

The importance of word order in Biblical Hebrew recently came up regarding Genesis 1:1, and in particular how we know that that verse answers the question “when?”

Here’s some more information and some additional examples.

An English Diversion

In English, stressing different words changes the implications of a sentence.

For example:

Example 1.

  1. This is my friend.
  2. This is my friend.
  3. This is my friend.

Sentence 1.1 is neutral, and might be used simply to introduce someone.

But 1.2 has addition implications, and it doesn’t make sense as a neutral introduction. But it does make sense in more narrow contexts. For example, it could answer the question, “whose friend is this?” And it makes sense if, because of context, there is already a sense that this is someone else’s friend. In both scenarios — specifically answering a question and clarifying an unasked question — it’s convenient to say that 1.2 “answers the question: whose friend this is?”

Example 1.3 cannot be used to answer the question “whose friend is this?” (Try it: -“Whose friend is this?” -“This is my friend.”) But it can be used if it’s already clear that this person has a relationship to me, but the nature of that relationship is not yet established. And, again, this could be in response to a specific question (e.g, “and how do you know each other?”), an implication, or some other context that causes the speaker to highlight the nature of the relationship.

One particularly interesting consequence of this is that stressing a word sets up alternative scenarios in the mind of the hearer. When I say, “this is my friend,” an English speaker automatically considers other possibilities: this is my enemy; this is my lover; etc.

Example 2.
1. We have landed in LGA. (This is the safe statement.)
2. We have landed in LGA. (This is risky.)

This is why “airplanese” stresses the words that are normally unstressed in English. For example, “we have landed in LaGuardia airport.” This phrasing only sets up one other possibility in the mind of English speakers who hear the sentence: “we have not landed in LaGuardia airport.” But the more usual sentence, “we have landed in LaGuardia airport” makes people consider other undesirable possibilities, like “we have crashed into LaGuardia airport.”

Example 3.
1. I killed Mary with a knife. (This is the safe answer.)
2. I killed Mary with a knife. (This is incriminating.)

Similarly, a witness at a murder trial might be asked, “how did you kill Mary?” The safe answer is “I killed Mary with a knife.” The dangerous answer is, “I killed Mary with a knife,” because that response assumes other pairs of people and weapons to which “I killed” might apply. Just changing the sentential stress is tantamount to admitting to additional murders.


Biblical Hebrew uses the pre-verbal sentence-initial position much the way we use sentential stress in English. This is how we know that Genesis 1:1 answers the question “when?” not “what?” or “who?” — the first word is breishit, “in the beginning.” A spoken-English translation of Genesis 1:1 might be “God created heaven and Earth in the beginning.”

Further Examples

Another example comes from Genesis 3:16-17. The first words of Genesis 3:16 are el ha-isha amar, “to the-woman [God] said.” This deviation from normal word order, like Example 3, has the effect of focusing “the woman,” as if to say, “this is what God told the woman.” The Hebrew reader naturally wonders, “okay. And what did God tell the man?” Genesis 3:17 answers that question, using the word order ul’adam amar…, “and-to-Adam [God] said…” — that is, “and God told Adam…”

We see the same thing in Genesis 45:22. The verse has two parts. The first begins, “to all [of the brothers] [Joseph] gave….” and the second one, “to Benjamin [Joseph] gave…” The point is to contrast Benjamin with the other brothers.

Hosea 5:7 reads badonai bagadu ki vanim zarim yaldu, “God they-betrayed when/for children alien they-bore.” Again, the point is expressed in English with stress: “they betrayed God by fathering illegitimate children.” Hosea is emphasizing the connection between human action and the divine.

All of these nuances come to light when we take Hebrew word order into account.

May 5, 2010 Posted by | general linguistics, translation practice | , , , , , , , , , | 5 Comments

Clear, Cogent, and Wrong

I frequently hear support for a translation philosophy that is in favor of only changing the original “as much as necessary” or of keeping the formal structure of the original “as far as possible” (to quote the introduction to the ESV). But I think that approach is fundamentally misguided.

The first three words of the Bible demonstrate. In Hebrew, they are breishit (“in the beginning”), bara (“created”), and elohim (“God”).

The most direct mapping from Hebrew to English would therefore be, “In the beginning created God [the heavens and the earth].” But that’s not grammatical English. So translators change the text to, “In the beginning God created…”

Unfortunately, they stop there, reasoning (wrongly in my opinion) that a translation of the Bible that means something in English necessarily means the same thing in English as the original.

In our example, the reasoning is this: Putting the verb before the subject in English is bad because it’s not grammatical in English. So far, so good. But the next bit of reasoning is that putting “in the beginning” first is accurate merely because it’s grammatical. The question “does it mean the same thing?” rarely gets asked. And in this case, the answer is “no.” Putting something at the start of a Hebrew sentence does not always mean the same thing as putting something at the start of an English sentence.

In fact, a better translation would be, “it was in the beginning that God created…,” because the first line of Genesis answers the question “when?” not “what?”

More importantly, the reasoning that leads to “in the beginning…” is, I think, faulty. We should not be asking only “is the English grammatical?” but also “does it mean the same thing as the original?”

April 29, 2010 Posted by | translation practice, translation theory | , , , , | 7 Comments