Recounting the Census: A Military Force of 5,500 (not 603,550) Men
Editor’s note: Over the past two centuries, scholars have questioned much of the Torah’s historicity. Although biblical archaeology began as an attempt to defend the historicity of the biblical narratives, more recently the findings of archaeology have become some of the chief obstacles to believing in the historicity of much of the Torah.
One serious problem with accepting even the broad outlines of the exodus and wilderness-wandering accounts is the great number of Israelites the Torah claims were involved: upwards of 2 million. Even scholars who accept some sort of historical exodus, such as Richard Elliott Friedman, say that the numbers are impossible to accept.
Many Bible scholars dismiss the great numbers as fanciful, or as a late innovation by the Priestly author, not reflecting actual number of early Israelites. Nevertheless, for those scholars who wish to see the Torah as coming from one Author or as internally consistent, the census numbers have remained a serious problem.
We have witnessed a resurgence in attempts to interpret the census data in a way that reduces their numbers significantly. Professor Joshua Berman, for instance, offered a suggestion along these lines in his recent Mosaic essay “Was there an Exodus?” The following piece by Ben Katz is part of this trend.
The Two Censuses
The book of Numbers records two censuses of Israelite males 20 years and older—i.e., military age men. The census taken in the first year of the wilderness period totals 603,550 (Numbers 1:46 [=Exodus 38:26]), while the one in the 40th year of this period totals 601,730 (Num. 26:51). The number of males 20 years and older implies a population of well over 2,000,000, including women and children.
The Problems with Understanding the Term אלף as “Thousand”
Such a large population wandering in the Sinai seems impossible (there isn’t enough water to sustain them), neither does it fit with the archaeological record, since such a massive amount of people living there for forty years would have left a trace. In addition, a number of other biblical passages are internally inconsistent with a population of 600,000-plus adult males and their families.
The First-Born Problem: Too Many Children per Family
The Torah twice states that there were about 20,000 firstborn males over one month of age (22,273 to be exact [Num. 3:39 and 3:46]). Since the total population of adult males over age 20 is a little more than 600,000, there were at least 580,000 non-firstborn males over the age of 20. Since each family could have only one firstborn—and only half of these would be male—this implies families of at least 29 non-firstborn males to first born males (580,000 / 20,000), i.e., families of at least 30 males apiece, unless for some reason mortality of firstborns was very much higher than that of subsequent children, accounting for a lower ration of firstborns to subsequent children. Including females would roughly double these numbers, unless for some reason significantly more than 50% of firstborns were girls.
Other Biblical Passages Assume a Smaller Population
Many passages in the Bible characterize the Israelites as having a small population. Exodus 23:29-30, for example, states that one of the reasons God will not drive out the Canaanites from the Promised land all at once is “lest the land become desolate” because the Israelites will need time to “increase…and possess the land.” Deuteronomy 7:7 states that the Israelites were “the fewest of all people.” Finally, troop numbers mustered in later periods are lower than expected if 600,000 males of military age lived in the period of the wilderness. For example, Deborah is only able to raise 40,000 troops from six tribes (Judges 5:8).
A Numbers Problem for Modern Readers
Pre-modern traditional Bible commentators were not concerned with this issue. In the modern period, however, these demographic arguments could not be ignored.
Birthing Sextuplets and Firstborn Mortality: Aryeh Kaplan
Aryeh Kaplan deals with this issue in his Living Torah. He begins by citing Rashi on Exodus 1:7, who comments on the six verbs used to describe Israelite fecundity in Egypt. Rashi claims that each woman gave birth to sextuplets.This midrash was designed to explain how the Israelites multiplied from a family of seventy to a nation of two million in just a couple of centuries, but it could also be used to explain why families were so large.
Nevertheless, Aryeh Kaplan likely still preferred not to assume the existence of families of 30 male children apiece, so he explained the relative scarcity of firstborns by postulating other assumptions not found in the Torah. For example, he suggested that perhaps many firstborn males did not observe the first Passover in Egypt and consequently died there, thus lowering the ratio of firstborns to subsequent children, or that most of the firstborns were girls.
Many modern Bible commentators—Baruch Levine, for instance—treat the number “600,000” as exaggerated. The number, they claim, is meant to express the great numbers of Israelites in the wilderness. The reason the number is “six hundred thousand” and not, say, a million or five hundred thousand is based on the fact that the Torah works with a sexigesimal (base 60) system (as was common in ancient Mesopotamia).
Another possibility raised almost a century ago by William F. Albright, is that the numbers are anachronistic. Perhaps a later author, used to a larger Israelite population, projected current estimates back into the ancient past. Albright specifically suggests that the figure of 600,000 males derives from the time of David. (Most current archaeologists and historians of the biblical period, however, strongly doubt that the population of ancient Israel ever approached the figure of two million.)
Alternative Approach: אלף means a Contingent
Another modern suggestion—first suggested by Flinders Petrie (1853-1942), is that אלף, usually translated “thousand,” should be translated like אלוף (troop or contingent.) Examples of this use of אלוף include Exodus 15:15, “clans of Edom (אלופי אדום),” and Gen. 36:15-30 (also in regard to Edom). It is this possibility that I plan to explore here.
Aside from the fact that אלוף means contingent, in several instances in the Bible, the word אלף itself may not literally mean “thousand”:
וְאִ֨ישׁ רֹ֧אשׁ בֵּית־ אֲבוֹתָ֛ם הֵ֖מָּה לְאַלְפֵ֥י יִשְׂרָאֵֽל:
They were every one of them heads of ancestral houses of the contingents of Israel. (Josh 22:14)
הִנֵּ֤ה אַלְפִּי֙ הַדַּ֣ל בִּמְנַשֶּׁ֔ה וְאָנֹכִ֥י הַצָּעִ֖יר בְּבֵ֥ית אָבִֽי:
Why, my clan is the humblest in Manasseh, and I am the youngest in my father’s household. (Judg 6:15)
וְאַתָּ֞ה בֵּֽית־לֶ֣חֶם אֶפְרָ֗תָה צָעִיר֙ לִֽהְיוֹת֙ בְּאַלְפֵ֣י יְהוּדָ֔ה…
And you, O Bethlehem of Ephrath, least among the clans of Judah… (Micah 5:1)
Applying אלף as Contingent to the Census
Many scholars believe that this solution is not helpful for the problem at hand. For example, in his TABS essay, “The Torah’s Exodus,” Zev Farber argues that when the term “אלף” is followed by a hundreds number (603 אלף five hundred and fifty in Exod 38:26, for instance), we have no choice but to translate it as “thousand.” But I disagree.
Recognizing the military nature of the census, I believe it is still possible to translate אלף as “contingents” by understanding the two phrases—the one that precedes the word אלף and the one that follows it—as two different kinds of measurement. Thus, to take Numbers 1:21 (the fighting men of Reuben) for example:
פְּקֻדֵיהֶ֖ם לְמַטֵּ֣ה רְאוּבֵ֑ן שִׁשָּׁ֧ה וְאַרְבָּעִ֛ים אֶ֖לֶף וַחֲמֵ֥שׁ מֵאֽוֹת:
Instead of translating 46 אלף and 5 hundred as “46,500”, it could be translated as “46 contingents, or [i.e., equaling] 500 [men].” Standard concordances list “or” as one of the meanings for a connecting vav, as in Exod. 21:17 “He who curses his mother or his father shall surely be put to death.”
Problems with אלף as Contingent
One problem with this suggestion is that the number of people per contingent varies both within each list for each tribe (see Table 1 and note 19, below) and between censuses for the same tribe (as do the populations of many tribes, when Numbers 1 and 26 are compared). Perhaps populations changed over time or each tribe had a different way of organizing its troops.
Another problem is that the word אלף, which I am translating as “contingent”, alternates with the word מאות (‘hundreds’) and other actual numbers in the censuses, so one might argue that contextually the word אלף should also be a number in the census. However, Num. 10:36 provides an example of two words that are often translated as numbers following one after the other, where one is translated as a number and the second is not. The expression there is רִֽבְב֖וֹת אַלְפֵ֥י יִשְׂרָאֵֽל, where the word אלפי means thousands, but the word רבבות, right next to it, usually translated as 10,000, in context means either myriads or hosts.
Support for אלף as Contingent from Statistics
Certain statistical data forcefully argue that in these censuses, אלף may mean “contingent” and not “thousand.”
If the census numbers are assumed to be real numbers (e.g, 46,500 for Reuben), it will be noted that nearly all of the tribes’ populations are rounded to the nearest hundred. Thus, the units and tens digits are zeros (placeholders), while the hundreds digits are “real” numbers (i.e., significant digits).
In the 24 census numbers that appear in the two censuses in Numbers 1 and 26, there is no “0” in any of the hundreds digits. Now, if אלף means “thousand,” and the census numbers are real numbers, the hundreds digit should be random. Statistically, there should be one or two tribes with a zero in the hundreds column, since there is no reason a tribe couldn’t have rounded to exactly, say, 46,000 troops, as opposed to 46,500 . The probability of no tribe out of 24 (because all 12 tribes are counted twice) having a population that rounded to an even thousand is a bit less than 1 in 200 (see Table 1).
However, if אלף means “contingent,” then the hundreds digit is not rounded but a real number, because the hundreds digit is telling you how many hundreds of men are in the contingents of each tribe. Thus, the hundreds digit cannot be a zero.
Another benefit of interpreting אלף as “contingent” is that the total troop strength is 5500, which fits with what we know of at least a few other, ancient Near Eastern armies.
The Total 603,550 Is Based on a Later Scribal Misunderstanding
The above explains the counting for the individual tribes. When it comes to explaining the totals, I believe the best answer is to assume that a later scribe or editor misunderstood the earlier notation. Thus, originally the text at the end of Numbers 1 may have read “598 אל(ו)פים (contingents) totaling 5 אלף (thousand) 550 [men].” Mistakenly assuming the term אלופים (or even אלפים) to have the same meaning as אלף in this context—when in reality the first meant “contingent” and the second “thousand”—a later scribe may have combined the two “numbers” yielding “603 אלפים ,” which was interpreted as 603,550.
The fact that this theory posits a later editor who misunderstood the distinction between אלופים and אלף and combined the terms, while not traditional, is not completely at variance with some of the later literary activity that is postulated to have occurred in the text of the Torah by a minority of traditional Bible commentators.
The Number of Levites and Firstborns
In the case of the Levites, if אלף is interpreted as “contingent,” the number of Levite males per contingent varies in the three Levitical families from 33–71/75 (see Table 2). The number of men per Levitical contingent is higher than the number of men per contingent for the other 12 tribes (8–14; see Table 1), perhaps because these are not military units, and they include all males over one month. (Note how I Chr. 26:8 describes a Levitical unit of 62 men, for example, which is consistent with the number of Levites per contingent listed in Table 2.)
The original total of Levites may similarly have read “21 אלופים of 1 אלף” which later became 22 אלפים, or 22,000. Similarly, the original total of first-borns could have been 21 אלופים of 1 אלף 273, which later became 22,273.
This suggestion of reading אלף as contingent also solves the big problem noted earlier concerning the number of fist-born males relative to the total male adult population. 1273 (male) first-borns is much more consistent with a total population of 5550 males, yielding a ratio of non-firstborns (5550-1273=4277) to firstborns of 4277/1273 = family sizes of roughly 3.5 male non-firstborns to firstborns, or families with about 4-5 males each—a much more reasonably-sized family.
Thus, we have demonstrated that statistical analysis can lend support for an old suggestion of reading אלף as contingent, which in turn brings the Israelite population in the wilderness down to more manageable numbers, and allows for more normal-sized families.
Military Census of Israel, (Numbers 1)
|No. men||Approx. No. of
Levitical and First Born Censuses (Num. 3:21-39,46)
|Approx. No. of
TheTorah.com is a 501(c)(3) nonprofit organization.
We rely on the support of readers like you. Please support us.
May 15, 2015
March 29, 2023
Previous in the Series
Next in the Series
Prof. Ben Zion Katz, M.D. is a Professor of Pediatrics at the Northwestern University Feinberg School of Medicine and an Attending Physician in the Division of Infectious Diseases of the Anne and Robert H Lurie Children’s Hospital of Chicago. He has also written on topics of Jewish interest related to Bible, liturgy and the calendar. He recently published A Journey Through Torah: A Critique of the Documentary Hypothesis.
Essays on Related Topics: