Personal Note [added in 2014]: This is one of the first essays I wrote for this web site, back in 2001. Looking back, I think it is fair to say the anomaly in the population figures I called attention to here was one of the catalysts that prompted the beginnings of my personal journey toward thinking more critically about my religious beliefs. At that time I was committed to the idea that the Bible was the literal word of God and so had to be absolutely free of error in its “original manuscripts,” as the essay makes clear. I no longer hold to this view of the Bible, so the issue I called attention to in this essay is no longer a problem for me—I simply think the Biblical writers had incorrect information at their disposal which they erroneously assumed to be factual and therefore reported as such.
I may eventually find the time to bring the essay up-to-date with my current beliefs, but until then I am keeping it on my web site as originally written. I think the issue it raises is still something that serious thinkers about Judaism or Christianity ought to be aware of, whether or not they believe in Biblical inerrancy as I did at the time.
One of the most problematic issues that critics of the Bible raise is that of the population of Israel during the time of the exodus. The figures given in the Old Testament indicate that there were about 600,000 men, not including women and children. This would make the total population somewhere in the neighborhood of two million or more, which is claimed to be absurd. The entire population of Canaan at that time is not thought to be so great. Furthermore, such a high population is also claimed to be logistically untenable. How could so many people cross the Red Sea in one day, for example? And how could they survive in the desert for so long?
It may be tempting to dismiss such arguments as simple unbelief. After all, God is omnipotent and could easily have worked whatever miracles were necessary to accomplish the logistics. If the Bible says there were 600,000 grown men, should we not simply believe it without asking questions?
As pious as this approach sounds, it is not God-honoring to turn a blind eye to possibly valid arguments. Far better to study them with care and see how they can be reconciled with our understanding of Scripture, even if it turns out to be our understanding of Scripture that has to be modified. We who follow the One who is "the truth" should certainly never be afraid of the truth. In that spirit, the remainder of this essay attempts to investigate the population of Israel during the exodus in an open-minded manner.
The population of Israel is reported in two different censuses during the time of the exodus. One was taken near the beginning of their journey, and the other was taken near the end of their 40 years of wandering. The correct view of the population hinges on a proper understanding of the numbers reported in these censuses. Therefore, it will be helpful to carefully analyze these numbers from a statistical point of view.
In both censuses the men from age twenty and upward were said to be counted "head by head." They were numbered by tribe, but the tribe of Levi was not included. Numbers 1 reports the results of the first census as follows:
Reuben 46,500 Simeon 59,300 Gad 45,650 Judah 74,600 Issachar 54,400 Zebulun 57,400 Ephraim 40,500 Manasseh 32,200 Benjamin 35,400 Dan 62,700 Asher 41,500 Naphtali 53,400 Total: 603,550
Numbers 26 reports the results of the second census as follows:
Reuben 43,730 Simeon 22,200 Gad 40,500 Judah 76,500 Issachar 64,300 Zebulun 60,500 Ephraim 32,500 Manasseh 52,700 Benjamin 45,600 Dan 64,400 Asher 53,400 Naphtali 45,400 Total: 601,730
Since the two censuses are so similar, we can study them together, giving us 24 different population figures to analyze, not including the totals. The following diagram is a histogram of the ten-thousands digit of the 24 different population figures:
This histogram shows that the 24 digits in question are fairly closely grouped around a central average. This is not at all surprising, since we would expect the 24 population figures to be somewhat similar in number. The ten-thousands digit is the most significant digit, and thus will reflect this tendency.
The thousands digit is a different story, however. Since the 24 population figures range over tens of thousands, we would not expect any special grouping of the thousands digit. Rather, we would expect this digit to be uniformly distributed among all the possible values from 0 through 9. This is shown in the following histogram:
A sample size of 24 is not large enough to clearly show that the probability distribution is uniform, but it at least shows that there is no apparent grouping around a central average, as was the case for the ten-thousands digit. Thus this histogram is as unsurprising as the first.
The surprising histogram is the one for the hundreds digit. We would expect this digit to be uniformly distributed, as was the case for the thousands digit. Instead, the histogram is as follows:
Notice the clear grouping of this digit around a central average. This histogram bears far more resemblance to that of the ten-thousands digit than to that of the thousands digit. This is a totally unexpected result for a variable that would be expected to follow a uniform probability distribution.
To get an idea of just how significant this is, consider that all 24 of population figures had their hundreds digit in the range 2 through 7. This represents just six out of the ten possible values. The chances that all 24 hundreds digits would be in such a limited range is extremely low, approximately 1 in 40,000.
What this result shows in all likelihood is that what appears to be the hundreds digit in each of these population figures must not really be the hundreds digit of a five-digit number. Rather, it appears to be the most-significant digit of some other number that stands apart from the thousands and ten-thousands given in each of the population figures.
We cannot get around the fact that the population figures given in the two censuses are not as straightforward as they appear. This does not prove anything about the population totals in and of itself, but it does serve as a warning that the figures as currently stated may very well be erroneous. The Word of God is error-free as it was originally given, but there are indisputable cases of corruption that occurred in the transmission (i.e., copying) of the text. Thus we should be open to the likelihood that the population figures in question are examples of these kinds of errors.
The presence of transmission errors in the current manuscripts should not be viewed as an alarming or hopeless situation. Rather, it is quite reasonable to assume that these errors are innocent mistakes made by copyists who were conscientious and well-meaning, not malicious and intentionally deceptive. Therefore, we can assume that there is a plausible explanation for their mistakes.
For the censuses in question, the most frequently suggested and probably best explanation is that the Hebrew word for "thousand" in the population figures actually refers to an indeterminate-size clan or troop. The word itself is transliterated 'elep (Strong's number 505) and apparently does indeed carry the meaning of a family complex or clan in some cases (probably the most notable being "family" in Judges 6:15). Furthermore, since the original Hebrew manuscripts consisted of consonants only and no vowels, the same spelling could also indicate the word 'allup, meaning "leader" or "chief." Similarly, the word can also be used of a military unit or troop under a leader or captain.
Further information on the technical features of this explanation can be found in Colin Humphreys' article How Many People Were in the Exodus from Egypt? (Science and Christian Belief, Vol. 12, No. 1, pp. 17-34). The view is also discussed by Gleason Archer in his Encyclopedia of Bible Difficulties (Zondervan, 1982, p. 129), although he rejects it in favor of the belief that the population really was as reported in the current manuscripts. Other relevant sources cited by these two writers include G. J. Wenham (Numbers, Inter-Varsity Press, 1981), E. W. Davies (A mathematical conundrum: the problem of the large numbers in Numbers I and XXVI, VT 45, 1996, 449-69), G. Mendenhall (The Census of Numbers 1 and 26, JBL 77, 1958, 52-66), John Bright (History of Israel, Westminster, 1959, 144), R. E. D. Clark (Journal of the Transactions of the Victoria Institute 87, 1955, 82ff.)
If the proposed explanation is valid, then the population figures given for each tribe were actually intended to consist of two parts. For example, where the population of Reuben is given as 46 thousand and 500, the intended meaning was 46 clans or troops and 500 individual men (rounded, of course). This would presumably mean that the 46 clans or troops consisted of 500 men, although it could also mean that there were 46 clans or troops consisting of an uncounted number of men, plus 500 other men who were not part of any clan or troop. Either way, the hundreds digit of the population figures turns out to be the most significant digit of the number of individuals counted. This is totally consistent with the actual grouping of that digit.
It is not known how long the intended meaning of the population figures was correctly understood. At some point prior to any known manuscript, the copyists apparently interpreted the word to literally mean "thousand" instead of "clan," and adjusted their text accordingly. This adjustment evidently included manipulation or addition of the totals given in Num. 1:46, 2:32, and 26:51. The sums would not have worked out as they are currently stated in the original versions. For example, the total for the first census was probably originally stated as 598 clans and 5550 individuals, and a well-meaning copyist took this as 598 thousand and 5550, which he simplified to 603,550. The total is also given as "about 600,000 men" in both Ex. 12:37 and Num. 11:21. These figures may have been introduced by a copyist, or they may have been intended to mean "about 600 clans."
It must be emphasized that there is every reason to believe that these mistakes were made by copyists, not by the original human author (i.e., Moses). Certainly Moses would have known that the numbering was done by clans, and would not have added the results as if they were by thousands. Also note that the numbers show no sign of being intentionally fabricated. There would be no plausible explanation for the unexpected distribution of the hundreds digit if the numbers were randomly fabricated.
If the individual census figure were intended to give the total number of men for each tribe along with the counts of the clans or troops for the tribes, then the first census adds up to 598 clans or troops consisting of a total of 5550 men. The second census would add up to 596 clans or troops consisting of 5730 men. This would probably put the total population including women and children at somewhere around 20,000.
The low population total resulting from this view can be argued against on the basis of the Scriptural reports of the number of people killed in various events during the time of the Israelites' journeys. For example, Num. 16:49 tells of a plague that killed 14,700, and Num. 25:9 tells of another plague killing 24,000. These figures could possibly also be copyist errors, but that is doubtful on the basis of the New Testament. 1 Cor. 10:8 tells of 23,000 being killed in one day, presumably in the incident of the golden calf. (The total number who died at this time is not recorded in the Old Testament.) If we assume this New Testament figure is inspired, then the other two figures of people killed by plagues in Numbers seem to be in a very believable range. Thus these three events account for more than 50,000 deaths, which seems unlikely if the total population was only about 20,000.
This does not mean that this view is impossible, however. In the first place, the 23,000 killed in the incident of the golden calf was before the first census was taken. It is possible that the original population was around 40,000 or more, and only around 20,000 were left for the first census. The population could have then swelled again during the first part of the 40 years of wandering, only to be reduced back to around 20,000 by the aforementioned plagues. Also note that the Israelites were accompanied in the exodus by a mixed multitude of unspecified size (Ex. 12:38). It is possible that the death tolls recorded included a significant portion from among this mixed multitude.
The other possible explanation is that the individuals counted in the censuses were men who were not part of any clan. Thus, the first census would add up to 598 clans of unspecified size and 5550 additional men. The actual number of men would be left unspecified. This may be a legitimate alternative, but it certainly seems like a strange way to conduct a census, especially in light of the Scriptural description of the process as "head by head" (Num. 1:2 and following).
There are a few hints from other Scriptures that may help us decide between these two alternatives. The conquest of Ai in Joshua 8 seems to be particularly helpful in this regard. Ai itself had a total population of 12,000, including women and children (v. 25). Joshua set up an ambush to take over the city. He is said to have put 30,000 soldiers at the rear (v. 3), which apparently meant on the west side, between Bethel and Ai (v. 9). He is also said to have put 5000 men in ambush in what appears to be the same place, namely on the west side, between Bethel and Ai (v. 12). Some commentators take this to mean that there were two different ambushes, but that does not seem to fit the narrative (vv. 14-22). Rather, it is probably better to assume that the reference to 30,000 refers to 30 captains over individual troops, and that the total number of men in ambush was actually 5000.
The remainder of the Israelite army lured the men of Ai out to attack them. Ai assumed they could defeat this group of soldiers as they had during the first attack, which had consisted of about 3000 Israelite soldiers, or perhaps less if this figure actually means 3 troops (Josh. 7:4). At any rate, it seems reasonable to assume there were about the same number of soldiers to lure them out in the second attack. If there were significantly less than 3000, then the men of Ai would probably not all leave the city, and if there were significantly more, then the men of Ai would probably not be so confident of victory. Thus the total size of the Israelite army would seem to have been somewhere around 8000 at this time.
If the Israelite army consisted of only about 8000 men, it is hard to see how the second census could have meant that there were 5730 men in addition to the 596 clans. It is far more likely that the 5730 figure was indeed the total. It may be that the 8000 figure is a bit on the high side, or it may be that the ranks could have swollen to around 8000 in the time between this census and the conquest of Ai. Also, the 8000 or so men who fought against Ai may have included some from the tribe of Levi, who were not included in the results of the census. In any case, it seems that the total population of Israel during the exodus was indeed around 20,000 or so, as first suggested.
The assumption that Israel's population during the exodus was only around 20,000 instead of two million or more greatly simplifies the explanation of a number of difficult Scriptures. Some of the more significant ones are listed below:
In numerous places, Scripture speaks of the whole congregation of Israel listening to Moses or Aaron (e.g., Ex. 16:10, 35:1, Lev. 19:2, Num. 20:8, cf. Josh. 8:35), or gazing at Moses (Ex. 33:8), or assembling at the doorway of the tent of meeting (Num. 16:19). This is hard to fathom for a population of around two million, but probably plausible for a population of around 20,000.
Moses initially acted as judge for the entire congregation of Israel (Ex. 18:13). This was a large task, but evidently not impossible. In v. 25 Moses delegated the job to "leaders of thousands, of hundreds, of fifties, and of tens." This would seem to be a strange hierarchy if the total population of Israel was around two million. There would have to be around 2000 leaders of thousands, directly under Moses. It seems much more plausible if the total population is assumed to be only around 20,000, so that there would only have to be about 20 leaders of thousands under Moses.
In the incident of the golden calf, Moses instructed every man of the Levites to execute God's vengeance with the sword (Ex. 32:27). This they faithfully did (Ex. 32:28-29), and they ended up killing a total of 3000 people. If the tribe of Levi consisted of tens of thousands of men of fighting age, this seems like a rather small total assuming "every man" of them really did as he was told. But if the tribe of Levi consisted of only a few hundred men of fighting age to begin with, then it makes reasonable sense.
The logistics of the offerings in Leviticus seem hard to explain with a population of around two million. Aaron and his sons had to officiate at these offerings. Could so many offerings really be handled by so few people? For example, with a population of around two million, there would probably be several hundred births on any given day. Each of these births required Aaron and his sons to officiate at an offering (Lev. 12:6-7). This seems logistically impossible, unless the total population was actually much less than two million.
In Num. 20:17, Israel requested permission of the Edomites to travel through their land, promising to stay on a certain highway. The same request is made to the Amonites in Num. 21:22. If the population of Israel was around two million, how could they possibly travel through on a single highway, along with all their possessions? Even if they could go 10 abreast with only 10 feet between groups, the line would stretch out for nearly 400 miles! This would be a difficult enough task even for a population of only around 20,000. It seems utterly inconceivable for a population of around two million.
Scripture repeatedly emphasizes that Israel was smaller than the surrounding nations at this time. Deut. 7:1 lists seven nations that were greater than they were, and v. 7 says they were the fewest of all the peoples (see also Deut. 9:1, 11:23, and 20:1). This is consistent with the fearful report of the spies in Num. 13. The surrounding nations were almost certainly not greater in population than two million each, but they could very well have been several tens or hundreds of thousands.
The account of the conquest of Jericho would be hard to explain from a logistical point of view if the army of Israel numbered 600,000. The city would have been totally surrounded by the 600,000 Israelites, unless it were many miles in circumference. But if it was that big, then how could the men march around it 7 times in one day? This problem is even more significant if we accept the archaeological findings that the walls of Jericho encompassed an area of only about 8½ acres, per research cited in The New Unger's Bible Dictionary (Moody, 1966, p. 672). This is a much more reasonable fit for around 6000 soldiers than 600,000.
In spite of all the evidence in favor of this view, there are still admittedly a good number of difficulties with it, in addition to the problem of having to accept that the total population figures are copyist errors, as already discussed. Some of the most significant potential difficulties, along with suggested possible solutions to them, are listed below:
Ex. 38:26 lists the total amount of silver collected for the tabernacle, and correlates it to the 603,550 figure. However, this correlation is suspect to begin with, since the 603,550 figure is obviously approximate, being the sum of the numbers for each tribe, most of which were rounded to the closest hundred. Thus there is reason to suspect that some copyist adjusted this number to make it correlate with what he assumed to be the total population figure.
Num. 3:22, 28, and 34 give counts for the families of the Levites at 7500, 8600, and 6200, and the total is given as 22,000 in v. 39. This figure is compatible with a large total population of Israel, not a small one. However, these numbers are suspect from the start, since the first three figures add up to 22,300, not 22,000. Some Septuagint manuscripts apparently have 8300 instead of 8600 in v. 28. In any case, there seems to be a strong possibility that the numbers here may represent yet another copyist error. The thousands figures could easily mean clans, as in the censuses. This would make the total 21 clans consisting of either 1000 or 1300 men (depending on the figure in v. 28). This is higher than the amount of men numbered in the other tribes, but this is to be expected, since males under the age of twenty were included in the numbering of Levi only.
Num. 3:43 gives the first-born males as 22,273, which does not make sense for a total population of only around 20,000. But note that this figure is hard to reconcile with the usually-assumed population of around two million as well. (One common explanation is that it must have meant only the first-born males born thus far in the journey, but that does not really seem to be implied by the text.) The important detail about this number is given in v. 39, namely that it was exactly 273 more than the population figure of 22,000 for the Levites. But it is highly unlikely that the 22,000 figure is exact, and so the subtraction really does not make sense. Rather, the 22,000 figure is better understood as 21 clans (see above), and the likelihood is that a well-meaning copyist manipulated the 22,273 figure to make the subtraction work out. Colin Humphreys believes that the Levites numbered 1000, which would suggest that this figure was originally 1273. He demonstrates in his article (mentioned above) that this appears to be quite consistent with the other population figures in the census.
Num. 4:34-48 gives the counts for the males ages 30-50 of the Levite families at 2750, 2630, and 3200, totaling 8580. This seems to be compatible with a large population, not a small one. This is perhaps the most difficult problem of all, since there does not seem to be any good reason why a copyist would intentionally come up with these figures here. But the possibility of inadvertent numerical errors in such figures must be considered.
In the second census, Num. 26:62 gives a count of 23,000 Levites. However, as suggested for the count of the Levites in the first census, this could well mean 23 clans instead of 23 thousands.
In the slaughter of Midian in Num. 31, the booty taken included 675,000 sheep and 32,000 virgin women. Although this says nothing about the population of Israel directly, it does seem more compatible with a large population than a small one. But it may be possible that even a small population could have accomplished this; it was a great victory, after all. Alternatively, these figures could also be copyist errors resulting from a misunderstanding of the word for thousand.
In Num. 35:7, the Levites are given 48 cities, which would seem to imply that the tribe of Levi was quite populous. Similarly, Josh. 15:21 and following lists numerous cities for the other tribes of Israel. But there is no reason to assume that the Israelites occupied all of these cities right away. In fact, Deut. 7:22 explicitly indicates that Israel was not able to take over all the territory at once. Furthermore, there is no way to know how large the cities in question were. Perhaps many of them were quite small, with populations of only a few hundred or so.
Josh. 4:13 states that the 2½ tribes beyond the Jordan sent out about 40,000 warriors. But it seems this could well have been intended to mean 40 clans or troops instead of 40 thousands.
The idea that there were were literally 600,000 Israelite men in the exodus as suggested by the current Old Testament manuscripts is fraught with difficulty and probably ought to be abandoned. The alternative explanation explored in this essay and elsewhere is that the thousands digits of the 24 tribal population figures were originally intended to be taken as "clans" or "troops," with the result that the total population of Israel was probably around 20,000 instead of two million. This view is clearly suggested by the distribution of the hundreds digits of the same population figures, and it appears to be the most plausible explanation when all the facts are considered. It does not argue against the inerrancy of Scripture as it was originally written, but it does illustrate the fact that there are copyist errors in our current manuscripts.
This page copyright © 2001 Edward A. Morris. Created July 15, 2001. Last updated July 31, 2001.