MANNAFEST

The Evidence Room · The Reading Room

What about God commanding the killing of the Canaanites?

This is the hardest thing in the Bible, several of the standard responses have real evidence behind them, and none of them makes the texts comfortable — a page that leaves you at ease here has not been honest with you.

The objection goes first, at full strength, in the words its defenders would choose.

The command is to kill non-combatants. Not soldiers — everyone. Women, the elderly, children, infants. It is given as a direct instruction and it is carried out as narrated. If any other ancient document contained this, and it did not belong to our tradition, we would have no difficulty naming what it is. We would call it what we call it when we find it elsewhere.

And the moral discomfort is not a modern squeamishness imported into an old text. The text itself uses the language of total destruction and expects the reader to understand what that means.

A reader who is not troubled here has not read it. This page will offer several responses, each with real evidence behind it, and then it will tell you what remains after all of them have been given — because something does remain, and the response to that is not a further argument.

Walking it through

  1. What the texts actually say

    Put the hardest verses on the table rather than paraphrasing around them.

    Deuteronomy instructs that of the cities of these peoples, nothing that breathes is to be saved alive. Joshua reports the campaigns in the same terms: he left none remaining, but utterly destroyed all that breathed, as the God of Israel commanded.

    There is no softening available at the level of translation. The Hebrew word usually rendered utterly destroy is a specific term — herem — meaning to devote something entirely, putting it beyond ordinary use, and in these contexts it means destruction.

    Notice one more thing before any response is offered. The instruction in Deuteronomy about offering terms of peace to a city applies to distant cities, and the text explicitly exempts the Canaanite cities from that provision. That is a point against the gentler readings, and this page prints it here rather than leaving it out of the summary of the evidence.

  2. The rhetoric reading — and the internal evidence for it

    The first response is that the total-destruction language of the ancient Near East was conventional. Kings said they annihilated their enemies the way a coach says a team was destroyed — as the standard idiom of a victory report, not as a body count.

    This is not a convenient guess. We have the other nations' inscriptions. The Egyptian pharaoh Merneptah, around 1208 BC, recorded that Israel is laid waste, his seed is not — a claim of total extermination made about a people who demonstrably continued to exist for the next thousand years and more. King Mesha of Moab, around 840 BC, carved into a stone monument that Israel has perished forever. The convention is documented in the neighbors' own words, and the neighbors' own words are provably not literal.

    Now the more interesting evidence, which comes from inside the Bible. The same book that reports total destruction goes on to describe the survivors.

    Joshua says that he cut off the Anakim from the hill country and utterly destroyed them with their cities. The very next verse says none of the Anakim were left in the land of Israel — except in Gaza, Gath, and Ashdod, where some remained. That is one sentence later.

    The same book says there remains yet very much land to be possessed, lists the peoples still living in it, and Judges opens with a catalogue of the groups that were not driven out — the Jebusites still living in Jerusalem alongside Benjamin unto this day.

    Consider what that means about the writer. Someone assembled these texts and left both statements standing, adjacent, with no sense of embarrassment. A person trying to conceal an exaggeration does not do that. The most economical explanation is that the writer and his readers understood the total-victory formula as a formula — and that we are the ones importing a literalism they did not have.

  3. The four hundred years before it

    Long before any of this, Abraham is told that his descendants will not receive the land for four generations, and the reason given is startling: because the iniquity of the Amorite is not yet full.

    Read that carefully. It says the judgment is being withheld — deliberately, for centuries — because the case is not yet complete. Whatever else this is, it is not an impulse.

    What was the case? The Bible names it: burning children in fire as an offering, and it is condemned in the same texts as an abomination that Israel is forbidden to imitate.

    That charge has support from outside the Bible, and the support should be stated with its limits. The Phoenicians who descended from the Canaanite world established colonies across the Mediterranean, and at Carthage archaeologists have excavated a burial precinct containing thousands of urns of cremated infant remains, with dedicatory inscriptions to the gods. Greek and Roman writers describe the practice directly. A minority of scholars have argued the precinct is an infant cemetery rather than a sacrificial one; the majority reading is that it is what the ancient writers said it was.

    This is offered as context, not as a settlement. Naming what a society did explains why judgment is described as judgment. It does not by itself answer the question of the children — that question is still coming, and it is the last section of this page.

  4. Driving out, not only destroying — and mercy inside the story

    Alongside the destruction language, the same books use another verb constantly: drive out. Dispossess. Send away before you. Whole passages describe the objective as removal from the land rather than extermination — one text has them driven out little by little, over more than a year, which is not how you describe a massacre.

    Read together, the two vocabularies suggest the goal was the end of a society's presence in a territory, with the fighting aimed at the fortified military and administrative centers. That is also what the archaeology of the relevant sites tends to show: the places named are largely garrison and administrative centers rather than population towns.

    Then there are the exceptions inside the narrative, and they are not marginal. Rahab, a Canaanite woman in Jericho, changes sides and is spared with her whole household — and Matthew's genealogy puts her in the family line of Jesus. The Gibeonites, an entire Canaanite people, sue for terms and get them; Israel keeps the treaty, and centuries later a king who broke faith with them is held accountable for it.

    The narrative shape matters here. In a story about extermination as such, Rahab and Gibeon are unthinkable. In this story they are told approvingly, and one of them ends up in the Messiah's genealogy. Surrender was a live option and the text presents taking it as the right choice.

    The honest counterweight, again: the peace-terms law in Deuteronomy explicitly does not extend to these cities. Rahab and Gibeon are exceptions granted inside the story, not a general policy the law provides for. Both facts are true and both belong on the page.

  5. The same judgment falls on Israel

    This is the point most often missed, and it changes the character of the argument.

    The reason given for the Canaanites' removal is not their ethnicity. It is what was done in the land. And the very same passage that says the land vomited out the previous inhabitants warns Israel, in the same breath, that the land will vomit them out too if they do the same things.

    That warning was not rhetorical. It happened. The northern kingdom was destroyed and deported by Assyria, and the chronicler states the reason in exactly the terms used against the Canaanites — they walked in the statutes of the nations. Then Jerusalem fell to Babylon, the temple burned, and the covenant people went into exile.

    So whatever this is, it is not tribal favoritism. The standard is applied to Israel with more severity, not less, and by the same God, in books written and preserved by Israelites who said so about themselves. A national literature invented to flatter a nation does not read like this.

    That does not make the conquest texts easy. It does defeat one common description of them — that this is a god who protects his own people and destroys everyone else. On the Bible's own account, his own people got it worst.

  6. The residue — and it is real

    Now the part that does not resolve.

    Suppose every response above is granted at full value. The numbers are conventional. The judgment was delayed four centuries and the practices were real. The primary aim was dispossession. Mercy was available and taken. The same standard later fell on Israel.

    The command is still in the text. Nothing that breathes. And there were children in those cities.

    There are answers offered at this point. That death is not the final word for anyone, and a child who dies is not thereby lost. That God, being the author of every life, holds a prerogative over its end that no human being holds. That a society at its endpoint destroys its children either way, and that the alternatives available inside that world were all terrible. Each of these is seriously argued and none of them is silly. None of them is comfortable either, and if one of them settles the matter entirely for you, it is worth asking whether you have felt the question.

    MannaFest is not going to close this with a sentence. Serious Christians who hold Scripture as the Word of God land differently here, and the range is worth describing without ranking it: some read the accounts as historical judgment on a specific society at a specific moment, never to be repeated and never a model; some read the conquest narratives as written in the idiom of their age, carrying a theological account of the end of a civilization rather than a battlefield record; some hold that the texts show God accommodating himself to a violent world he was in the process of moving out of; some conclude that the tension is not resolvable this side of seeing him, and hold it unresolved on purpose. These are not the positions of people who have stopped believing the Bible. They are the positions of people who have read it closely and come out at different places.

    One thing to add, and it must not be used to erase anything. When two of Jesus's disciples asked whether they should call down fire on a village that refused him, he turned and rebuked them. Some manuscripts add the line most people remember — ye know not what manner of spirit ye are of — and the honesty required elsewhere on this site requires the note that the earliest copies do not have it, while the rebuke itself is in all of them.

    That rebuke is not evidence of a different God in the New Testament. The disciples were quoting Elijah, and the rebuke is not for believing the old story; it is for thinking it was theirs to repeat. Whatever the conquest was, it was not a template. It closed. And the same God who is said to have judged a civilization is said, in the same book, to have come and stood in front of the judgment himself.

The objections, at full strength

Each objection below is stated the way the people who hold it state it. If you hold one of these and think it is put unfairly here, that is a failure on our part — tell us.

The hyperbole reading is special pleading invented under moral pressure

The objection

Notice when this reading became popular. It is not a discovery of the last century's archaeology being applied to a neutral question; it arrives precisely as the moral objection becomes unbearable to defend head-on, and it arrives from people who need the text not to say what it says. That is the classic profile of a rationalization. And it is applied selectively: the same reader takes the promises literally, the miracles literally, the resurrection literally, and then discovers that the one category of text he finds embarrassing is idiomatic. A rule invoked only where it rescues you is not a rule of interpretation.

The response

The suspicion is fair and the timing point has some force, so the reply has to be evidence that does not depend on anyone's motives. Three pieces. First, the comparative material was not produced to solve this problem: Merneptah's stele and the Mesha stone were dug up and read for their own sake, and both make total-annihilation claims that are simply false about peoples who kept existing. That is the convention, documented in someone else's writing, about someone else's wars. Second, the internal evidence is not a modern reading at all — the Anakim are destroyed and then found in three cities one verse later, and Judges opens by listing who was not driven out. Those sentences have been sitting next to each other for as long as the books have existed. Third, the selectivity charge can be tested rather than asserted: the standard being applied is that a text's genre governs how it is read, which is why nobody reads the Psalms' mountains skipping like rams as zoology. Whether that standard is being applied honestly here is a fair thing to check, and the check is the internal evidence, which is public. Where the objection lands and stays landed: even if the rhetoric reading is entirely correct about the numbers, it does not touch the command, which is the next objection and the more serious one.

Even if the numbers are conventional, the command is still in the text

The objection

Grant the whole rhetoric case. Grant that far fewer people died than the language implies, that the targets were garrisons, that the aim was dispossession. The instruction still reads save alive nothing that breatheth, and it is presented as coming from God. The moral problem was never arithmetic. It is that a command of this kind is attributed to a being who is supposed to be the ground of goodness. Reducing the casualty figures addresses the historian's question and leaves the theologian's question exactly where it was — and the eagerness with which defenders reach for the numbers suggests they know which question is harder.

The response

This objection is correct that it is the harder question, and the previous section conceded as much before it was raised. Two things can be said and neither is a solution. The first is a genre point that has to be made carefully: if the total-destruction formula is the standard idiom for a decisive campaign, then the command written in that idiom is a command to end a society's hold on a territory decisively — the phrase carries the sense of the convention it is written in, not a separate literal sense the surrounding narrative then ignores. That is a real argument and it is not nothing. It is also not enough, because the text goes further than a formula in places, and a reader who feels that is reading correctly. The second is to say where this actually leaves a believer, which is not at a solution but at a trust decision. Someone who has concluded on other grounds that God is good and that Jesus is who he said he is holds these texts as a difficulty inside a settled relationship rather than as a proof that overturns it. Someone who has not is under no obligation to start here. What would be dishonest is to claim the difficulty has been dissolved. It has been located, and the location is that this is one of the places where trust is doing work that argument is not.

The they-deserved-it argument has been used to justify real atrocities

The objection

This is not a hypothetical cost. The conquest narratives were read directly onto real populations: by Puritan colonists describing the peoples of New England as Canaanites, by settlers in southern Africa who took the Exodus and the conquest as their own charter, and by others besides. Actual people were actually killed by readers of this text who believed they were doing what Joshua did. So when a defender explains that the Canaanites were a uniquely wicked society ripe for judgment, he is reciting the exact reasoning that has always been used — a reasoning that supplies any group in power with a way to describe the people in its way. The template is the danger, and the template is right there in the text.

The response

This is true history and it should be stated by Christians before critics have to. The text has been used that way, the users were often sincere, and people died. No amount of contextual reading undoes that, and a defense of these passages that skips it is not worth reading. Two things follow. The first is a distinction the misuse always has to violate: on the Bible's own account this was a specific judgment on a specific society at a specific moment, announced centuries in advance, executed once, and never authorized again. Nothing in the Old Testament treats it as a standing policy, no prophet calls for its repetition, and the New Testament forecloses it — which is exactly what the rebuke to the disciples who wanted to call down fire is doing. Every historical misuse required treating a closed event as an open permission. The second is that this cuts against a certain style of defending the text. If the practical danger is the deserved-it framing, then defenders should be more careful with it, not less, and should notice that the Bible itself immediately turns the same standard on Israel — which is the one move that makes the framing unusable as a weapon against somebody else. A reading that leaves you feeling secure about your own side has already gone wrong.

An omnipotent God had other options, and the appeal to mystery arrives right on schedule

The objection

Everything in the responses assumes the conquest was the available means. But an omnipotent being is not constrained by available means. He could have changed hearts. He could have made the land uninhabitable for a generation. He could have sent prophets, as he later did to Nineveh — where preaching worked and a city was spared, which the Bible itself reports as something God can do. He could have settled Israel elsewhere. Any of these is open to omnipotence, and the fact that none was chosen is not explained by saying the ancient world was violent. And observe where the appeal to mystery is deployed: never on the promises, never on the miracles, always at the exact point where the moral cost falls due. Mystery invoked selectively is not humility. It is a placeholder for an answer that does not exist.

The response

The last sentence is a genuinely good criticism and it is the reason this page has tried not to reach for mystery early. On the alternatives: two of them are less clean than they look. Changing hearts by overriding them raises the same question that runs through the whole problem of evil — whether a world of real agents can be run by overriding agency — and if the answer is yes, the objection is really an argument against permitting any evil at all, which is a different and larger question with its own page. Nineveh is the more interesting case and the objection is right to raise it, because it shows warning and repentance are on the table and do sometimes work. What Genesis says about the Amorites is that warning-and-waiting is exactly what was happening for four centuries first. That is not a full answer to why it ended as it did, and this page will not dress it up as one. So here is the honest position. There is no argument on this page that establishes the conquest was the only or the best option available to omnipotence. What there is: reasons to think the accounts are not what they are usually taken to be, evidence that the judgment was neither ethnic nor sudden nor repeatable, and a residue that does not resolve. Anyone who tells you the residue has been cleared is selling something. Where a Christian rests is not on having cleared it but on a person — on the conviction that the character of God was made visible in Jesus, and that whatever is unresolved here is unresolved about someone who has already shown what he is like.

Go further

If God is good, why is there so much suffering?→

Doesn't the Bible contradict itself?→

Who decided what goes in the Bible?→

The apologetics room→

Every source we cite→

Sources: The Merneptah Stele (c. 1208 BC) — the Egyptian victory inscription claiming Israel is laid waste, his seed is not; the earliest known extrabiblical mention of Israel, and a documented example of total-annihilation rhetoric. · The Mesha Stele, the Moabite Stone (c. 840 BC) — King Mesha's inscription declaring that Israel has perished forever; the same convention in a neighbor's own words. · K. Lawson Younger Jr., Ancient Conquest Accounts: A Study in Ancient Near Eastern and Biblical History Writing (JSOT Press, 1990) — cite-only; the comparative study of conquest-report conventions. · Richard S. Hess, Joshua: An Introduction and Commentary (IVP, 1996) — cite-only; on the military and administrative character of the sites named. · Paul Copan, Is God a Moral Monster? Making Sense of the Old Testament God (Baker Books, 2011) — cite-only; the fullest popular statement of the responses gathered here. · Christopher J.H. Wright, The God I Don't Understand (Zondervan, 2008) — cite-only; a defender who declines to claim the difficulty is resolved. · Randal Rauser, Let Nothing that Breathes Remain Alive — cite-only; a Christian philosopher arguing that the hyperbole reading does not remove the moral problem. Printed here because objection one deserves its ablest form. · Gregory A. Boyd, The Crucifixion of the Warrior God (Fortress, 2017) — cite-only; a very different Christian landing, included so the range described above is not a range of one. · Richard Dawkins, The God Delusion (Bantam, 2006), ch. 7 — cite-only; the objection in its best-known popular form. · Diodorus Siculus, Library of History XX.14, and Plutarch, On Superstition — public domain; the classical descriptions of Carthaginian child sacrifice. · Lawrence E. Stager and Joseph A. Greene, on the excavation of the Carthage tophet — cite-only; the archaeological evidence and the minority dissent from its interpretation. Every cited work — the sources registry →