Why does Exodus regulate slavery instead of abolishing it?
Exodus 21 opens the Covenant Code with rules about slaves. Hebrew slaves serve six years and go out free in the seventh. A female Hebrew slave sold by her father has different rules. A master who knocks out a slave's tooth must let the slave go free. A master who beats a slave to death is to be punished, but a master whose slave dies a day or two later 'is not to be punished, for the slave is his property' (Exod 21:21). Read against Leviticus 25:44-46, which authorizes permanent ownership of foreign slaves, and Deuteronomy 15, which restates the seventh-year release, the chapter raises a question readers have asked since at least the patristic period. Why does the same God who freed Israel from Egypt then give Israel rules for keeping slaves?
Exod 21:1-11 sets out the rules for a Hebrew bondservant. Six years of service, release in the seventh, the option of permanent service with the ear-piercing ritual if the slave loves his master. The female Hebrew slave (vv. 7-11) is treated as a concubine with marriage-rights protections. Exod 21:20-21 and 21:26-27 add the slave-injury provisions. Leviticus 25:44-46 then permits permanent ownership of slaves taken from the surrounding nations. Deuteronomy 15:12-18 restates the seventh-year Hebrew-slave release with material additions (provisioning the freed slave; extension to female Hebrew slaves). The chapter does not abolish slavery. It regulates it. Four positions have organized the modern discussion, and each has to handle different texts and different evidence. None of them takes the chapter lightly.
What the texts actually say
Exodus 21:2 opens: 'If thou buy an Hebrew servant, six years he shall serve: and in the seventh he shall go out free for nothing.' The Hebrew word for the slave is eved, which covers everything from indentured servant to chattel slave depending on the context. The six-year limit applies to Hebrew slaves only. If the master gave the slave a wife during his service, the wife and any children stay with the master at the seventh-year release (v. 4). If the slave chooses to remain, the ear-piercing ritual at the doorpost makes the service permanent (vv. 5-6).
Exod 21:7-11 covers the female Hebrew slave. 'And if a man sell his daughter to be a maidservant, she shall not go out as the menservants do.' The daughter is sold by her father, presumably under economic distress. Verses 8-11 protect her: she cannot be sold to a foreign people; if her purchaser-master rejects her, she is to be redeemed; if he marries her to his son, he must treat her as a daughter; if he takes another wife, he must not diminish 'her food, her raiment, and her duty of marriage'; if any of the three is withheld, she goes free without payment.
Exod 21:20-21 covers a slave killed under the master's hand. 'If a man smite his servant, or his maid, with a rod, and he die under his hand; he shall be surely punished.' The next verse contains the most discussed phrase in the chapter: 'Notwithstanding, if he continue a day or two, he shall not be punished: for he is his money' (Exod 21:21, KJV; Hebrew kaspo, 'his silver'). Exod 21:26-27 then balances this with the injury-for-freedom rule: a master who knocks out a slave's eye or tooth must let the slave go free as compensation for the eye or tooth.
Leviticus 25:44-46 extends the picture. 'Both thy bondmen, and thy bondmaids, which thou shalt have, shall be of the heathen that are round about you; of them shall ye buy bondmen and bondmaids. Moreover of the children of the strangers that do sojourn among you, of them shall ye buy... And ye shall take them as an inheritance for your children after you, to inherit them for a possession; they shall be your bondmen for ever.' The seventh-year release applies to Hebrew slaves; permanent ownership applies to slaves from the surrounding nations and resident foreigners.
Deuteronomy 15:12-18 restates the Hebrew-slave release with three additions. The seventh-year release is extended explicitly to the female Hebrew slave ('Hebrew man, or an Hebrew woman'). The freed slave is to be provisioned from the master's flock, threshing floor, and winepress ('thou shalt not let him go away empty'). And the exodus rationale is given: 'thou shalt remember that thou wast a bondman in the land of Egypt.' The ear-piercing ritual is repeated (Deut 15:16-17) with the same option of permanent service.
The four positions
Four families of reading. Each is held by serious readers; none of them denies the difficulty of the texts.
- Christopher J. H. Wright, Old Testament Ethics for the People of God (IVP, 2004); The Mission of God (IVP, 2006)
- Richard B. Hays, The Moral Vision of the New Testament (HarperOne, 1996)
- John Goldingay, Old Testament Theology, vol. 3 (IVP Academic, 2009); Israel's Faith (2003)
- Nicholas Wolterstorff, Justice: Rights and Wrongs (Princeton, 2008)
- Walter Brueggemann, Theology of the Old Testament (Fortress, 1997)
- Daniel I. Block, The Gospel According to Moses (Cascade, 2012)
- Patrick D. Miller, Deuteronomy (Interpretation; John Knox, 1990)
- C. S. Lewis, Reflections on the Psalms (1958) on accommodation as a category
- • Jesus uses the category of accommodation explicitly in Matt 19:8: Moses gave the divorce law 'because of the hardness of your hearts.' The category of accommodated commands is given inside the New Testament canon
- • The seventh-year release (Exod 21:2; Deut 15:12) interrupts ANE practice. Most ANE legal collections do not contain any general manumission provision. The Hebrew-slave law treats slavery as time-limited and ends it for Hebrew slaves periodically
- • The slave-injury provisions (Exod 21:20-21, 26-27) protect the slave's body in a way that has no exact parallel in the ANE legal mainstream. A master who knocks out a slave's tooth loses the slave. A master who beats the slave to death is to be punished
- • The sojourner protections in the same legal corpus (Exod 22:21, 23:9; Deut 10:19) make the exodus the structural memory of how the vulnerable are to be treated. The exodus framing constrains how the slave laws can be applied
- • Goldingay and Wright argue the legal corpus is concerned to keep slavery within bounds (it does not become hereditary for Hebrews, it ends after six years, it cannot be used to permanently dispossess a Hebrew family) while accepting that the institution existed
- • Wolterstorff distinguishes between texts that authorize an institution and texts that regulate it. The Mosaic slave laws fit the regulatory category, not the authorizing category. They do not stand in the canon as commendation; they stand as constraint
- • Leviticus 25:44-46 is the harder text for this reading. It does not just regulate existing foreign slavery; it explicitly authorizes its acquisition and its inheritance. 'Ye shall take them as an inheritance for your children after you' is closer to authorization than to mere regulation
- • Exod 21:21 ('he is his money') treats the slave as property in language that the accommodation reading has to work to handle. The verse is widely read by abolition-era critics and by modern readers as authorizing the master-property relation, not merely tolerating it
- • The accommodation reading depends on a category (regulating vs. authorizing) that is clearer in the modern reader's framework than in the text itself. The text does not flag which of its provisions are accommodated and which are normative
- • If the law is accommodated, the question of what God actually willed becomes harder to answer from the text alone. The accommodation reading often pairs with a progressive-revelation reading (next position), but the two are not the same
The slave provisions across three legal corpora
The Pentateuch contains three major treatments of slave law: Exodus 21 (the Covenant Code), Leviticus 25 (the Holiness Code), and Deuteronomy 15 (the Deuteronomic Code). They overlap, but they are not identical, and the differences are part of what the positions above are arguing about. The Deuteronomic and Holiness restatements expand the Exodus provisions in ways that some readers see as marking a trajectory and others see as harmonizing variation within a unified Mosaic legal corpus.
Reading the rows across the columns shows where the corpora agree, where they diverge, and where each adds material the others do not contain.
The 'he is his money' clause
Exodus 21:20-21 is the most discussed text in the chapter and the one each position has to address most directly. 'And if a man smite his servant, or his maid, with a rod, and he die under his hand; he shall be surely punished. Notwithstanding, if he continue a day or two, he shall not be punished: for he is his money.' The Hebrew kaspo means literally 'his silver,' and the underlying noun is the standard word for money or material wealth.
Three readings of the verse have circulated. The first takes the clause as straightforward property language: the slave is treated as the master's property, the master's loss of the slave's labor is its own consequence, and the law does not impose additional punishment. This is the reading most defenders of chattel slavery in the 19th century used, and the reading most modern critics highlight when they argue the texts authorize property-status for human beings.
The second reading treats the clause as procedural. The master who beats his slave to death immediately is punished because the intent to kill is evident. The master whose slave dies a day or two later is not punished because the master clearly did not intend the slave's death (an intentional killer would have made sure of it). On this reading, the 'his money' clause is the legal grounds for inferring lack of intent: the master would not have wanted to destroy his own economic asset. The reading does not soften the property language, but it locates the legal logic in the question of intent rather than in property as such. This is the reading taken by many rabbinic and patristic commentators and by some modern defenders (Sprinkle 1994, Copan 2011).
The third reading takes the clause in light of the surrounding verses (Exod 21:26-27, the eye-and-tooth manumission rule). The two cluster together as a unit. A master who kills the slave is punished. A master whose slave dies later is not (intent inference). A master who knocks out an eye or tooth loses the slave entirely. On this reading the legal unit is balanced: the slave's body is protected on both ends. The 'his money' clause is part of the procedural logic of the cluster, not a standalone property declaration. Critics respond that the balanced-cluster reading does not change the fact that the underlying property frame is what makes the inference of non-intent legally relevant.
Exod 21:4 and the family that stays with the master
The second provision modern readers most frequently flag is Exod 21:4. 'If his master have given him a wife, and she have born him sons or daughters; the wife and her children shall be her master's, and he shall go out by himself.' A male Hebrew slave whose master gave him a wife during his six years of service is released alone at the seventh year. The wife (presumably a foreign slave, since she is described as the master's property) and any children remain with the master.
This provision is what the next verse (Exod 21:5) gives as the rationale for the ear-piercing ritual that makes the service permanent. The slave who 'loves his master, his wife, and his children' faces a choice: walk away from his family, or accept lifelong slavery. The provision structurally tests the question of what choosing freedom costs. For the slave who has formed a family in his master's house, freedom and family are made into a choice.
Accommodation readers point out that the provision reflects the economic reality that the master had provided the wife and bore the cost of her support and the children's upbringing. The provision constrains the slave's exit but is consistent with a frame in which the master retains his property in the slave-wife. Ethical-critique readers respond that whatever the economic logic, the result is the structural use of family bonds as an instrument of permanent enslavement. The chapter does not soften this, and the New Testament's later language about marriage as an indissoluble bond (Mark 10:8-9) and about no master-slave division in Christ (Gal 3:28) reads on the trajectory framing as setting a different horizon.
The female Hebrew slave
Exod 21:7-11 covers the case of a daughter sold by her father. The provision distinguishes her from the male Hebrew slave: she is not subject to the six-year release ('she shall not go out as the menservants do'). She is, on most modern readings, sold as a concubine or as a wife-candidate for the purchaser or his son. The provisions in vv. 8-11 set protections within this arrangement.
She cannot be sold to a foreign people (v. 8). If her purchaser-master rejects her after taking her, she is to be redeemed by her father or her family; he cannot keep her without taking her or sell her at his discretion. If she is given to his son, she is to be treated as a daughter, not as a slave (v. 9). If he takes another wife, he must continue to provide her 'food, raiment, and the duty of marriage' (her conjugal rights, v. 10). If any of these three is withheld, she goes free 'without money' (v. 11).
Deuteronomy 15:12 then includes the female Hebrew slave in the six-year release ('Hebrew man, or an Hebrew woman'). The relationship between the two passages is contested. Some readers (the trajectory framing) treat Deut 15 as expanding Exod 21 to include the female slave in the general release; this is one of the clearest cases of trajectory inside the Pentateuch. Other readers treat the two passages as covering different cases: Exod 21:7-11 covers the daughter sold by her father into a quasi-marital arrangement; Deut 15:12 covers the woman who has sold herself into service for debt. On this reading, the two provisions are complementary rather than progressive.
The fugitive slave law
Deuteronomy 23:15-16 contains a provision that does not have an Exodus 21 parallel and is striking against ANE norms. 'Thou shalt not deliver unto his master the servant which is escaped from his master unto thee: he shall dwell with thee, even among you, in that place which he shall choose in one of thy gates, where it liketh him best: thou shalt not oppress him.' A fugitive slave is to be granted asylum, not returned.
This provision contrasts directly with the ANE legal norm. Hammurabi §15-§20 covers the harboring of fugitive slaves: death for the harborer, ten shekels for the returner, a brand on the slave to mark recapture. The Hittite Laws and the Middle Assyrian Laws contain similar return-to-owner provisions. The Deuteronomic law inverts this. The ANE-comparative reading treats this as one of the clearest signals that the Mosaic legal corpus is doing something distinctive in its treatment of slavery. The accommodation reading treats it as evidence that the legislation is constraining the institution rather than authorizing it. The ethical-critique reading agrees the provision is striking but argues a fugitive-slave protection sits awkwardly alongside Lev 25:44-46's authorization of permanent foreign-slave ownership.
What each position has to account for
Each of the four positions has to do certain work on the texts. The accommodation reading has to handle Lev 25:44-46 (authorization, not just regulation) and Exod 21:21 (property language). It usually does so by pairing with the progressive-revelation reading and treating Leviticus as marking the limit of the accommodation rather than the trajectory's endpoint.
The progressive-revelation reading has to handle the New Testament's failure to call directly for abolition (the slave-obey-master texts in Eph, Col, 1 Pet, 1 Tim, Titus) and Lev 25:44-46 as a counter-trajectory text. It usually does so by appealing to the social-economic difficulty of direct abolition in the first-century Roman world and by treating Galatians 3:28 as the theological frame within which the obey-your-master texts have to be read.
The ANE-comparative reading has to handle Hammurabi §117 (the three-year debt-slave release is shorter than Exod 21:2's six years) and the property language at Exod 21:21. It also has to handle the underlying ethical question of whether 'better than the surrounding nations' is the right benchmark for divine legislation. It usually does so by pairing with one of the other three positions and limiting its claim to the comparative-historical level.
The ethical-critique reading has to handle the canonical question (why are these texts in the canon if they are morally unresolved?) and the historical-interpretation question (the texts have been read differently in different periods, including being read by ancient and medieval communities as constraints rather than authorizations). It usually does so by distinguishing canonicity from immediate normativity, or by relocating the locus of authority in the New Testament's reframing rather than in the Mosaic texts as freestanding ethics.
Reading the chapter with the question open
Exodus 21 sits in the canon as one of the texts that does not let its readers walk past. The four positions name the families of reading that have emerged across two thousand years of Jewish and Christian engagement with the chapter. They are not airtight cells. Many readers hold a combination: the comparative-historical observations of the ANE-comparative position alongside the canonical trajectory of the progressive-revelation reading, with the accommodation framework holding the parts together, and the ethical-critique reading insisting that the difficulty does not yield to harmonization.
What the texts give the reader is a set of constraints on an institution and a memory of slavery's underlying form. The God who freed Israel from Egypt then gave Israel rules for keeping slaves. Each position handles that pairing differently. The accommodation reading says the constraints are the point, not the institution. The progressive-revelation reading says the constraints are a stage, not an endpoint. The ANE-comparative reading says the constraints are unusually strong against the surrounding norms. The ethical-critique reading says the constraints do not absorb the harm. Each reading has a cost. The texts themselves do not resolve which cost to pay. They sit in the canon as testimony, and each generation of readers takes a position knowing what their position has to handle.
Sources
- Exodus 20:10; 21:1-11, 16, 20-21, 26-27, 32; 22:3, 21; 23:9, 12 (Masoretic Text; Leningrad Codex B19a)
- Leviticus 19:20-22; 25:1-55 (the Jubilee, the Hebrew-slave release, the foreign-slave authorization)
- Deuteronomy 5:14-15; 10:19; 15:12-18; 23:15-16; 24:7 (Mosaic-period legal restatement and expansion)
- Jeremiah 34:8-22 (prophetic critique of slave-release violation)
- Nehemiah 5:1-13 (post-exilic enforcement of debt-relief and slave-release)
- Code of Hammurabi §§14, 15-20, 117, 196-205 (slave provisions and parallel cases; Louvre AO 10237)
- Hittite Laws §§19-24 (fugitive-slave return; Hoffner 1997)
- Middle Assyrian Laws A §§4, 40-41 (slave provisions; Roth 1997)
- 1 Corinthians 7:21-24; Galatians 3:28; Ephesians 6:5-9; Colossians 3:22-4:1; 1 Timothy 6:1-2; Titus 2:9-10; 1 Peter 2:18-25; Philemon 8-21 (New Testament slave-related texts)
- Philo of Alexandria, De Specialibus Legibus II.79-85, 122-123 (c. 30 CE) on Mosaic slave provisions
- Josephus, Antiquities 4.8.28 (on the slave-release law) (c. 93 CE)
- Babylonian Talmud, Kiddushin 14b-22b (rabbinic exposition of Hebrew-slave law)
- Maimonides, Mishneh Torah, Hilkhot Avadim (c. 1170s), codification of slave law
- Walter C. Kaiser, Toward Old Testament Ethics (Zondervan, 1983)
- Phyllis Trible, Texts of Terror (Fortress, 1984)
- Joe M. Sprinkle, 'The Book of the Covenant': A Literary Approach (JSOTSup 174; Sheffield, 1994)
- Patrick D. Miller, Deuteronomy (Interpretation; John Knox, 1990)
- Eugene H. Merrill, Deuteronomy (NAC; B&H, 1994)
- Richard B. Hays, The Moral Vision of the New Testament (HarperOne, 1996)
- Susan Niditch, Ancient Israelite Religion (Oxford, 1997)
- Walter Brueggemann, Theology of the Old Testament (Fortress, 1997); An Unsettling God (Fortress, 2009)
- Christopher Marshall, Beyond Retribution (Eerdmans, 2001)
- William J. Webb, Slaves, Women, and Homosexuals: Exploring the Hermeneutics of Cultural Analysis (IVP Academic, 2001)
- David L. Carr, The Erotic Word (Oxford, 2003); Holy Resilience (Yale, 2014)
- Christopher J. H. Wright, Old Testament Ethics for the People of God (IVP, 2004); The Mission of God (IVP, 2006)
- I. Howard Marshall, Beyond the Bible (Baker, 2004)
- James L. Crenshaw, Defending God: Biblical Responses to the Problem of Evil (Oxford, 2005)
- Mark Noll, The Civil War as a Theological Crisis (UNC Press, 2006)
- Bruce K. Waltke, An Old Testament Theology (Zondervan, 2007)
- Nicholas Wolterstorff, Justice: Rights and Wrongs (Princeton, 2008)
- James K. Hoffmeier, The Archaeology of the Bible (Lion, 2008)
- John Goldingay, Old Testament Theology, vol. 3 (IVP Academic, 2009)
- Eric A. Seibert, Disturbing Divine Behavior (Fortress, 2009)
- Paul Copan, Is God a Moral Monster? (Baker, 2011)
- Hector Avalos, Slavery, Abolitionism, and the Ethics of Biblical Scholarship (Sheffield Phoenix, 2011)
- Daniel I. Block, The Gospel According to Moses (Cascade, 2012)
- John D. Currid, Against the Gods: The Polemical Theology of the Old Testament (Crossway, 2013)
- Paul Copan and Matthew Flannagan, Did God Really Command Genocide? (Baker, 2014)