Skip to content

Definite atonement examined: Owen's trilemma and its assumptions

The Reformed position

The classic defense is John Owen's trilemma (The Death of Death in the Death of Christ, 1647):

God's will to save is either (1) that all men should be saved by the death of Christ, or (2) that some men only should be saved, or (3) that all men should be saved on a condition which some fulfill and some do not.

Owen argues that (1) is universalism, (2) is definite atonement, and (3) makes the atonement's effect depend on the creature and leaves the unbeliever's sins unatoned. His conclusion: Christ died for all the sins of the elect only.

The argument is powerful because it forces a choice. The non-Calvinist has to say which of the three he holds and why it does not collapse into one of the others. The casebook's answer is (3), rightly understood — and the work is in showing that (3) is coherent and biblical.


The problems

1. The trilemma assumes the atonement is a payment of a debt

Owen's argument works if the cross is a commercial transaction: a fixed number of sins are paid for, and the payment must be applied to be effective. On that model, if Christ paid for a sin, God cannot punish it again; therefore Christ did not pay for the sins of the non-elect.

But Scripture uses a range of metaphors for the atonement: sacrifice, propitiation, redemption, ransom, reconciliation, victory, adoption, and yes, purchase. The commercial metaphor is one among many, and pressing it literalistically produces conclusions the others do not. A sacrifice can be offered for all and applied to those who come; a door can be opened for all and entered by some; a ransom can be paid for all and received by those who leave the prison.

The non-Calvinist does not deny that the cross is a payment; he denies that the payment model is the only model, and that it entails a fixed, secret list.

2. The "double payment" argument proves too much

If Christ's death is a payment for sin, and the non-elect are punished for their sins, then either Christ did not pay for their sins (definite atonement) or God punishes twice (injustice). Owen takes the first horn.

But the same logic would apply to the elect who lived before the cross, to those who never hear, and to the relation between the atonement and faith. If the payment is unconditional, why is faith necessary at all? The commercial model needs a doctrine of application to avoid automatic universalism, and once application is conditional on faith, the model already allows for a universal payment whose benefit is received by some.

The consistent commercial model is actually universalism (if Christ paid for all, all are saved) or hyper-Calvinism (if Christ paid only for the elect, the offer to others is meaningless). Mainstream Calvinism wants a universal offer and a definite payment, which is the unstable middle.

3. Unbelief is not simply "another sin" to be paid for

Owen's argument treats unbelief as one of the sins that must be atoned for. But in the New Testament, unbelief is the rejection of the atonement, not a separate debt alongside it. John 3:18: "whoever does not believe is condemned already, because he has not believed in the name of the only Son of God." The condemnation is for rejecting the provision, not for failing to be included in it.

This is why the free offer can be sincere even on a universal atonement: the provision is made for all, and the unbeliever's sin is that he refuses it. The "double payment" objection assumes that the only reason a person can be lost is an unatoned sin; the New Testament says the reason is unbelief.

4. The definite-atonement texts are about the church as beneficiary, not an exclusion

The classic texts — Matt 1:21; John 10:11, 15; Eph 5:25; Acts 20:28; Titus 2:14 — describe what Christ did for his people. They do not say he did nothing for others. That is the negative inference fallacy. "Christ loved the church and gave himself for her" is an inclusive statement about the church's blessing; it is not a statement that he gave himself for no one else.

The one text that comes closest to an exclusion is John 10:11 ("the good shepherd lays down his life for the sheep"), and the same chapter has "other sheep" (10:16) and John 11:51-52 has "the children of God who are scattered abroad." See 04-sheep-and-other-sheep.md.

5. The universal texts are numerous, varied, and not marginal

  • 1 John 2:2: "not for ours only but also for the sins of the whole world."
  • John 3:16: "God so loved the world that he gave his only Son."
  • 1 Tim 2:4-6: God "desires all people to be saved"; Christ "gave himself as a ransom for all."
  • Heb 2:9: "he tasted death for everyone."
  • 2 Cor 5:14-15: "one died for all, therefore all died."
  • Rom 5:18: "one act of righteousness leads to justification and life for all men."
  • Titus 2:11: "the grace of God has appeared, bringing salvation for all people."
  • 2 Pet 2:1: false teachers "deny the Master who bought them."
  • 2 Pet 3:9: God is "not wishing that any should perish."
  • Mark 16:15: "proclaim the gospel to the whole creation."
  • Rev 22:17: "let the one who is thirsty come."

Each has a Reformed reading; the cumulative weight is what matters. The universal texts are not obscure; they are the church's evangelistic vocabulary.

6. The "sufficient for all, efficient for the elect" formula equivocates

Dort says Christ's death is "of infinite value and worth, more than sufficient to expiate the sins of the whole world" (2.3), but intended to "effectually redeem... those only who were from eternity chosen" (2.8). The formula uses "sufficient" in the sense of value, not availability. A medicine worth enough to cure everyone, but not made available to them, is not "sufficient" in the sense that grounds an offer.

If "sufficient" means "available to all who believe," then the non-Calvinist agrees, and the "efficient for the elect" clause is just a description of who believes — not a limitation of the atonement's intent.

7. The historical evidence is on the non-Calvinist side

The pre-Reformation church, the Eastern church, and the liturgies of the historic churches prayed for "the sins of the whole world." The formula "sufficient for all, efficient for the elect" predates Dort (it is found in Aquinas and the medieval tradition), and it was a way of holding together a universal sufficiency with a particular application. The Westminster Assembly included hypothetical universalists (Davenant, Ussher, Hall). The 1948 OPC free-offer controversy showed the tension is structural. See 08-atonement-in-church-history.md and ../04-history/06-the-free-offer-controversy-in-the-opc.md.


The non-Calvinist position

The atonement is:

  • Universal in provision. Christ's death is sufficient for every person and grounds a genuine offer to every person.
  • Definite in accomplishment. It actually saves those who believe; it is not a mere possibility.
  • Conditional in application. The benefits are received by faith, which is enabled by grace and is not a work.
  • The ground of the offer. Because Christ died for all, the church can say to all: "Christ died for you; believe and live."

This is the classical Arminian and Provisionist position. It is not universalism (not all are saved) and not definite atonement (the provision is not limited). It is the position that makes the free offer coherent.


What this does and does not prove

It proves: that Owen's trilemma rests on a commercial model that Scripture uses only as one metaphor among many; that the double-payment argument begs the question; that the definite texts do not exclude others; and that the universal texts are numerous.

It does not prove: that the atonement accomplishes nothing definite, or that all are saved. The casebook affirms a definite accomplishment for those who believe, and denies that the provision is limited in its intent.


Further reading

  • John Owen, The Death of Death in the Death of Christ (1647) — the classic; read the argument at its best.
  • Canons of Dort, Second Head, Articles 3, 5, 8.
  • David Allen, The Extent of the Atonement (2016) — the most thorough recent critique.
  • Robert Picirilli, Grace, Faith, Free Will (2002), chs. 5-6.
  • I. Howard Marshall, Jesus the Saviour (1990), ch. on the death of Christ.
  • Jonathan Edwards, "The Satisfaction of Christ" — a Reformed account of the commercial model.

See also