When Numbers doesn’t add up

The book of Numbers is about … well … lots of numbers. But as I was reading through chapter three this morning, I came across what seemed to be an error. In that chapter, Moses counts the number of service-age men in the tribes of Kohath, Gershon, and Merari. But when the three totals are added together, the text is off by 300 people.





Num. 3:21-26



Num. 3:27-32



Num. 3:33-37



Num. 3:39


At first, I thought it was my own miscalculation, but after using a calculator it became apparent that the total was wrong. As the Bible is inspired by God and thus without error, this was something that could further harden skeptics as to the reliability of the Scriptures. A trip to http://www.blogger.com/www.biblegateway.com revealed that the NKJV, NASB, and ESV kept the number despite the erroneous math. What was the deal? I finally stumbled across a footnote in the NKJV which referred to a textual variant in some copies of the Septuagint, the famous Greek translation of the Old Testament. The variant changes Kohath’s total from 8,600 to 8,300, making the grand total correct as stated in Numbers 3:39.

Why then do major English translations of the Bible not correct the erroneous number? It would seem that they are trying to be consistent with the Hebrew text from which they are translating despite the error. Now that got me wondering. But not having any study materials nearby and not finding anything helpful on the internet, I consulted the next best thing, Pastor Dan Greenfield of Orwell Bible Church. With his permission, I have quoted his reply below:

The Masoretic Hebrew text (the Hebrew text used today) reads שש. The textual apparatus notes that the Septuagint reads triakovsioi (300), reflecting a Hebrew text of שלש. You can see the difference—the Masoretic text does not include a lamed (ל), while the Septuagint reflects an underlying Hebrew text that does include that letter. Ronald Allen in the Expositors Bible Commentary (vol 2, p. 726) notes that “because the totals of all other numbers in these chapters of Numbers are consistent, it is reasonable to assume a textual transmission disfunction in this instance.”

The Septuagint (abbreviated LXX) was initially translated in the third to second centuries B.C. Thus, it represents the earliest of the translations of the OT. When it comes to utilizing the LXX for OT textual criticism, this is a complicated process that I’ll not go into here. Suffice it to say that the Hebrew text we rely on today was produced about A.D. 1000 by Aaron ben Moses ben Asher. The value of the LXX is evident, as it is closer in time to the original manuscripts than ben Asher’s text. That said, the complicated process involves the fact that the LXX is a translation, and at spots a loose one at that. Additionally, Jewish scribes were meticulous in their copying; comparisons between OT manuscripts found among the Dead Sea Scrolls with today’s Hebrew text shows little loss or change.

All this to say, the explanation you came across seems to be the consensus view and an acceptable one. Keep in mind that inspiration (the result of God’s supernatural guidance and protection of the human author) applies to the original manuscript, not resultant copies and translations. The latter can be said to be “inspired” only as they accurately reflect the original manuscripts.

That was a helpful explanation. It satisfies my curiosity regarding the error, but leaves me wondering why the translators didn’t make the change in the English text.


UPDATE: If you would like to see what Numbers 3 actualy looks like in Hebrew, click here. One column is Hebrew (MT) and the other is English.

Print Friendly, PDF & Email

4 thoughts on “When Numbers doesn’t add up

  1. Dan Greenfield

    I guess I forgot to answer that question–sorry! Translators generally are apprehensive to go against the MT solely on the basis of the LXX. In such cases English versions will provide a footnote pointing out the potential difference in translation.

    On another note, your transliteration of the Hebrew letters is a little off–it’s more like s-l-s with an upside down “^” above the “s” to make the “sh” sound.

Comments are closed.