For most of my adult life, I believed the Reformed doctrines of grace, as they’re often called, were the most God-glorifying theological system a person could possibly hold. Total depravity, unconditional election, limited atonement, irresistible grace, the perseverance of the saints, five points built to protect one central claim. God’s purposes cannot fail. I heard Isaiah 53:11 preached often during those years; Christ seeing the travail of his soul and being satisfied, nothing wasted, nothing falling short of what the cross was meant to accomplish. I believed that account, and I believed it was the most God-honoring one on offer.
Turns out, I was wrong about that and this essay is intended to show why.
I am not proposing a new doctrine. (That sort of thing used to get you killed, now it just adds to the noise.) The concept of Christian universalism has real ancestry, older and more diverse than people raised in American evangelicalism often realize. Gregory of Nyssa, one of the architects of Nicene trinitarianism and a canonized saint in both East and West traditions, held it.10 Isaac of Nineveh held it, and the Eastern churches still read him extensively.
George MacDonald preached it and lost his pulpit for it (not his head, thankfully). Karl Barth built a doctrine of election that leans hard in this direction and then deliberately, almost inexplicably, refuses to state it as a conclusion. Hans Urs von Balthasar argued we’re permitted to hope for it, which might get him the award for hedging.4
In the last few decades, however, three men with very different approaches have made the strongest modern cases I’ve found. David Bentley Hart from the standpoint of metaphysics and moral force, Thomas Talbott from pure logic, and Robin Parry from inside evangelical exegesis itself. What follows takes the best, in my opinion, of what those three built, adjusts where I think each falls short of his own premises, and adds what I don’t think anyone has quite said in the way I’ll say it before. That last proposal is always a dangerous one, of course, but I’m in exceptionally good company, and I want to be forthright about exactly which parts of this are borrowed and which parts are my own, or at least I believe them to be.
The foundational claim
God redeems and restores the whole of his creation. Every part of it. Nothing left over; nothing left out.
I want to argue this premise on three separate grounds: logic, metaphysics, and the biblical text itself, with a single account of human agency running underneath all three, because without such a mechanism the other two arguments conveniently float free of anything that actually explains how it all works. That account of agency comes from the same man, incidentally, who spent his career defending and preaching the idea that some people are in fact damned for eternity. Jonathan Edwards is about to end up arguing for something he explicitly rejected, so we might as well start with him. (I can feel him rolling over in his grave already, spider in hand, ready to dangle me over the pit for what I’m about to do to his argument.)
The strongest inclination
Edwards wrote Freedom of the Will in 1754 to defend the accusation of Calvinist determinism against Arminian objections, and his central claim fits in a single line, his own: “I suppose the will is always determined by the strongest motive.”1 Nothing coerces a person from the outside. A person still chooses, still wants what they want, still acts on their own desires. But the desire itself is never uncaused, never random or arbitrary. Temperament, circumstance, and perception combine to make one option appear more compelling than the rest, and the will always goes where the strongest inclination points. Edwards used this to explain how the reprobate remain fully responsible for their own damnation even though God, on his account, never intended to save them in the first place. Real choice, real culpability, and an outcome that is fixed from before the world began by God’s decree.
I intend to keep every piece of that argument and direct it somewhere Edwards did not. If the will always follows whatever it is most inclined toward, and if God knows every person completely, then God already knows, completely and from the beginning, exactly what would move any given soul toward himself. He simply remains that reality for as long as it takes each soul to finally turn toward it, no override required, no coercion anywhere in the process. The same account of the will Edwards spent an entire book defending, only followed further than he was willing to follow it himself.
I have searched for a reference to someone who has already made this argument, in this way, and I’ve come up empty so far. That isn’t proof no one has made it, of course, and a two-hundred-and-seventy-year-old book about the human will has had a great many readers over the centuries, and I’d be genuinely surprised if the tension has escaped every one of them. What I can say with some level of confidence is that it isn’t sitting in the visible literature as a known argument I’ve discovered. Maybe it’s a bad argument; maybe I’m onto something. Either way, I suppose it deserves a name, so it seems to me that the strongest inclination, Edwards’s own phrase, is largely doing the argument’s work in only three words.
Talbott, Hart, and a third place to stand
That frame gives the logical case for universalism something it often lacks, a specific answer to the how question. Thomas Talbott’s version plays out as a trilemma. God is omnipotent, certainly capable of guaranteeing everyone’s salvation. God is perfectly loving, genuinely willing everyone’s salvation. And some people are lost forever anyway, the traditional view so very many of my good friends and loved ones affirm. Three premises, and Talbott’s claim is that they cannot all be true together, and traditional theology usually keeps the third by quietly weakening the first or second, redefining omnipotence as somehow bound by human choice or God’s own self-restriction, or redefining love into something compatible with overlooking someone permanently. Talbott’s move is to keep all three premises intact and let the logical conclusion follow on its own.
Talbott puts the stakes bluntly. Reject universalism, and you’re left denying either that God wills “the redemption of all sinners,” or that this will ever actually gets what it wants.2 Reformed systematics has a ready-made answer here, of course. God’s permissive will can allow what his effectual will doesn’t directly cause, which is typically meant to let God stand behind evil’s existence without being the author of it. The distinction doesn’t rescue the trilemma, though, once the strongest inclination is already on the table. In a fully determined universe, permission and causation are effectively the same thing at the level that really matters. Nothing exists, resists, or persists outside what God, at minimum, sustains and could theoretically reshape. Calling reprobation “merely permitted” in some passive fashion doesn’t change what a being with unlimited patience, power, and complete knowledge of every soul still chose not to do.
David Bentley Hart pushes even harder and argues eternal conscious torment is metaphysically impossible, not merely wrong, because a rational nature is ordered toward the Good by what it is, and a fixed, everlasting rejection of the Good isn’t a viable condition for a rational creature to occupy.3 Balthasar goes the other direction and claims almost nothing, except that Scripture holds warning and promise in genuine tension, which it does, and thus we’re permitted to hope that tension resolves toward mercy.
The strongest inclination argument gives me a third place to stand, likely a bit closer to Talbott’s logic than to either of the other two. I don’t require Hart’s claim that eternal resistance is logically impossible, though I don’t rule it out either. What I need is more modest: on my account resistance remains entirely possible, moment to moment, for as long as anyone is willing and able to hold it (pun intended), it simply never gets held forever. I also don’t need Balthasar’s hedging, because I’m not merely hoping God’s patience outlasts human resistance. I’m thoroughly confident it can and does, on the same grounds I’d be confident that water, given enough time, reaches every low point in a landscape, not because water is searching for the way down, but because reaching every low point is simply what water inevitably does. Call this functional certainty, I guess, positioned between necessity and hope. A confidence earned by patience (or, perhaps, perseverance) and by God’s infinite knowledge of every soul he created, certain for creation as a whole, even though the shape of any given person’s path stays hidden from us until it unfolds.
Relation, not isolation
Logic is useful to explain why the outcome holds. Metaphysics explains what kind of universe makes that outcome possible in the first place, and this is where I want to build something different from Hart, Talbott, and Parry alike.
Hart’s metaphysics leans on classical theism, God as the simple, transcendent Good, evil as privation with no independent reality of its own. That’s a serious and coherent picture by itself. It’s also a somewhat heavier metaphysical commitment than I currently hold, so I’m doing some of my own hedging here on purpose. In my own thinking and theological evolution, I’ve moved toward relational trinitarianism and panentheism, and I want a robust account of why resistance can’t be the final word that arises out of those commitments rather than borrowed wholesale from metaphysics alone.
John Zizioulas helps give me that, and he builds it from the same Cappadocian soil Gregory of Nyssa’s own universalism grows out of. Boethius defined a person as an individual rational substance, complete in itself, standing alone. Zizioulas argues the Cappadocians never actually worked with that precise definition; that for them personhood exists only in relation, constituted by communion rather than achieved in isolation from it. “To be is to be in communion,” as he puts it.5 If that’s correct, a soul sealed off permanently from every relation has stopped being the kind of thing a person is, whatever else remains. Permanent self-isolation describes a state that has run out of anything left to describe, the same conclusion Hart essentially reaches through the notion of privation, arrived at through a doctrine of personhood I firmly hold.
Panentheism significantly raises what’s at stake in getting this idea right. If creation exists within God’s own life rather than standing outside at arm’s length from an unaffected sovereign, any permanently unredeemed corner of it isn’t simply an acceptable loss God absorbs, overlooks, and moves past. It’s something closer to an unhealed wound, an imperfection inside God’s own life. Jürgen Moltmann has argued this shape eschatologically at length in The Coming of God, articulating a relational trinitarianism and panentheistic bent paired with hopeful universalism.6 Robin Parry’s own historical anthology on this subject is subtitled from Origen to Moltmann.7 Parry treats him as the modern endpoint of exactly this lineage of thought I’m describing, not a leap or connection I’m merely concocting on my own.
The fountain
Here is the heart of the argument I actually want this essay to be memorable for, if it is remembered, and it comes from Edwards yet again, in a book he wrote to answer a completely different question.
A Dissertation Concerning the End for Which God Created the World asks why God made anything at all, given that God lacks absolutely nothing and needs nothing whatsoever external to himself. Edwards’s answer is that God’s own nature includes a disposition to communicate his fullness outward, not because he is incomplete without doing so, but because a full fountain overflows as a natural expression of its very fullness, rather than evidence of some deficiency in it. He returns to this image more than once, in fact, writing of “an infinite fountain of light” whose very fullness moves it to pour itself out.8 Creation exists to receive the overflow. That, for Edwards, is the whole point of there being a world.
I want to run that argument forward in the same way I ran the will argument forward. If God’s self-diffusive fullness is a product of the perfection of his nature rather than a response to a lack in it, then the same nature that explains why creation exists at all is the nature that would have to explain why any part of creation stays permanently outside the overflow. A God whose very reason for creating is the outward pouring-forth of his own goodness, and who then allows some portion of what he made to remain forever unreached by that goodness, is a God whose own purpose for creation went unfulfilled in exactly the place it mattered most. Again, Edwards built this argument to explain why the world exists and he never took it forward in this way. I don’t see how it can be kept from also explaining where the world ends up, however.
It’s reassuring to know I’m not alone in reading Edwards’s metaphysics this way, because it means this isn’t simply a private reinterpretation with nothing more behind it. Walter Schultz, writing in the Journal of the Evangelical Theological Society in 2016, works through the Dissertation in close technical detail and concludes that Edwards’s own argument logically entails panentheism, alongside several other commitments Edwards never used that vocabulary for but structurally built anyway.9 Schultz is a conservative evangelical philosopher, publishing in a conservative evangelical journal, finding a concept of panentheism already sitting inside Edwards’s own metaphysics before I ever got around to it. Hart’s account of evil as parasitic on the Good remains useful here too, as a second voice reinforcing the same point from a different tradition, without needing to carry the whole argument in isolation.
How long
None of this tells you how long the process takes for any particular soul, and I don’t really think it’s supposed to. This is where the notion of theosis is exceedingly helpful. If salvation is participatory transformation rather than a single legal verdict delivered once, purification continuing past death isn’t a particularly strange addition that needs much defending on its own. It’s simply theosis, that slow union with God’s own life and nature Eastern Christianity has always centered, continuing past the point the vast majority of Western theology arbitrarily stopped the clock. Gregory’s own doctrine of epektasis assumes perpetual growth into an infinite God, an ascent with no ceiling, which means nobody was ever going to arrive all at once in the first place.10
Hell, on this account and by this definition, is real, and it lasts exactly as long as it takes. Purification aimed at bringing a person back, nothing more theatrical than that. It isn’t the pit or the lake of fire stocked with pitchforks and brimstone most evangelicals grew up envisioning, as that version owes more to fanciful medieval poetry and revivalist preaching habits than it does to anything found in the text, which is a topic for a separate essay.
All means all
Everything I’ve said up to this point explains why the exegesis holds up rather than merely positing that it does, which is where Robin Parry’s contribution finally comes in. It’s difficult to argue against the claim that the Pauline texts read most naturally as universal in scope. As one man’s trespass led to condemnation for everyone, one man’s righteous act leads to justification and life for everyone (Romans 5:18); as in Adam all die, so in Christ all will be made alive (1 Corinthians 15:22); God reconciling the world to himself in Christ (2 Corinthians 5:19); all things reconciled through him, things on earth and in heaven (Colossians 1:20); God desiring everyone to be saved and to come to the knowledge of the truth (1 Timothy 2:4).
Reformed exegesis has to import a qualifier into every one of these texts; all the elect, all without distinction rather than all without exception, although the qualifier never arises from the passage itself. Drink Wisely has examined elsewhere how thin that same move really looks when applied to the Cappadocians’ own reading of restoration.
The Greek behind these texts is telling, since the qualifier has to come from outside the words themselves. Romans 5:18 uses the same phrase, pantas anthrōpous, in both halves of the sentence, all people to condemnation, all people to justification. First Corinthians 15:22 does the same with pantes, in Adam all die, in Christ all are made alive, the identical word carrying both clauses. Nobody reads “all die” as anything short of everyone, and the grammar gives no reason to read the matching “all” differently three words later. Not every instance here is that clear, however. First Timothy 2:4 sits two verses after Paul asks for prayer “for kings and all who are in authority,” and Greek scholars have debated for years whether that context narrows “all people” toward “people of every kind.” Romans 5:18 and 1 Corinthians 15:22 carry the weight the argument actually needs.
The logic leg of the argument is what finally gives the exegesis teeth against the standard Reformed reading. If God’s love genuinely cannot be defeated by our sinful resistance, forever, there’s no longer a theological reason to read “all” in these passages as anything less than what it says.
The devil
I feel compelled to follow this line of thinking all the way to its least comfortable place. If the fallen powers are also God’s own creation rather than co-eternal rivals standing outside his authority, and I think most would agree they are, I don’t see a principled line where the argument is permitted to stop short of their inclusion. A universalism that redeems every human soul while leaving Satan and the fallen angels outside it hasn’t actually solved the problem it claims to have solved. It has just shuffled aside the unredeemed remainder somewhere less visible. These are the ones we don’t speak of. Hart himself doesn’t stop at the human race either. He frames his whole argument around “the ultimate salvation of all rational beings,”11 a phrase that was never going to stay confined to one particular species once he’d committed to it.
Trust me, I know exactly how radioactive this idea can be. The restoration of the devil was specifically among the propositions attributed to Origen and condemned in the sixth century, which is better admitted here than discovered later in a comment section.12 Trust me, I don’t like it either.
The standard remedy for keeping fallen angels out of this argument comes from Aquinas, who holds that angels choose through a single, complete, non-discursive act of intellect rather than the developmental, time-bound reasoning humans rely upon, which is exactly why their choice is treated as instantly and permanently final in a way human choice isn’t.13 I don’t find that argument compelling, because it needs angelic cognition to work nothing like the compatibilist account of the will this whole essay is built on. If even a non-discursive intellect settles on whatever appears most compelling to it, and if God can, given unlimited time, always eventually become what’s most compelling, the Thomistic firewall never actually engages. Put much more simply, this remedy still falls short.
My own answer to the human question turns out to answer the angelic one too, I think, without needing a separate argument built from scratch and more than a little Aristotelian influence. Uncomfortable, yes, but I don’t see how we can stop short of it and remain consistent.
What’s mine here
The claim that hell functions as correction belongs to Gregory of Nyssa and Isaac of Nineveh long before it belongs to anyone else I’ve mentioned in this essay. The logical trilemma belongs to Talbott. The exegetical case belongs to Parry. The account of evil as parasitic on the Good belongs to Hart, in a lineage running back through Augustine’s own account of evil as privation, oddly enough, to the same tradition that produced the rather fanciful and verifiably destructive doctrine of eternal conscious torment I’m arguing against. What I think is actually new here is the specific architecture. Edwards’s own compatibilism, the idea that real choice and fully determined causes were never actually in conflict, run forward instead of backward, doing the infrastructure work neither Talbott’s logic nor Hart’s metaphysics quite supplies on its own; Zizioulas’s relational personhood and Moltmann’s panentheism carrying the metaphysical weight instead of classical theism; and Edwards’s fountain, read by his own careful modern interpreters as already quite panentheistic (not suggesting he intended it to be), turned into an argument about where creation ends up rather than only about why it began. Two different books by the same eighteenth-century Puritan defender of eternal damnation, each pressed past the conclusion their author intended, and met again in the same place.
Satisfied
Which brings me all the way back to where I started this essay. The doctrines of grace never lacked for confidence. Every rhetorical resource in that tradition points toward triumph, a God whose purposes cannot fail, Christ’s soul-travail producing exactly what it was sent to produce, nothing whatsoever wasted, nothing coming up short. Postmillennial and optimistic amillennial eschatology extend that same instinct outward, with history itself bending ultimately toward an undisputed victory in Christ. I spent decades learning and believing this was the highest view available, because on its own terms the argument is genuinely compelling. (Who can’t sign off on “soli deo gloria” after all?)
And then, I find that it quietly bows out from its own promise. Some portion of what God made inexplicably remains lost forever in this otherwise triumphant systematization, by design, and the system still asks you to call that complete and total victory. For me, at least, it’s the account failing the exact test it built for itself, conceding that an eternal remainder didn’t get in for reasons we do not and cannot know, except to know it must be justified.
The standard line is that this isn’t a failure at all. Romans 9 calls some people vessels of wrath, prepared for destruction, so that God might make known the riches of his glory on vessels of mercy, and on this reading, permanent ruin isn’t a loss God absorbs, it’s a display that God intends, his wrath glorifying him exactly as much as his mercy does. I don’t think that answer survives its own calculus, however.
If permanent ruin of created beings glorifies God, more of it glorifies him more, and I’ve never met a Reformed theologian willing to actually say that, which suggests even they don’t fully believe the argument they must assert. Note that Paul doesn’t leave the vessels of wrath sitting where Romans 9 puts them, either. Just two chapters later, describing the same hardening he’d just defended, Paul writes that God consigned everyone to disobedience so that he might have mercy on everyone, the identical Greek word, pantas, doing the condemning and the saving in both instances. Whatever Romans 9 is doing, Romans 11 doesn’t let it stand as the last word.
Paul also writes that God will be all in all (1 Corinthians 15:28), and I want to be cautious when citing that verse rather than basing my whole conclusion on it. Some serious readers, including non-universalist ones, take this reference cosmically rather than personally; God’s final supremacy over every rival power rather than a claim about every individual soul’s reconciliation. I think the personal reading is the stronger one still, for all the reasons already laid out above, but the verse alone doesn’t clearly settle it one way or the other.
What I can say without any additional hedging is this. The theology I spent decades inside promised a God whose purposes land exactly where he sets them and then built in an eternal exception to its own promise. The account I am postulating herein keeps the full promise instead. God gets all of his creation back.
All in all, with nothing left outside it, because a fountain that has been overflowing since before the foundation of the world was never going to leave part of what it made permanently dry.
Endnotes
1. Jonathan Edwards, Freedom of the Will (1754), in The Works of Jonathan Edwards, vol. 1, ed. Paul Ramsey (New Haven: Yale University Press, 1957).
2. Thomas Talbott, The Inescapable Love of God, 2nd ed. (Eugene, OR: Cascade Books, 2014).
3. David Bentley Hart, That All Shall Be Saved: Heaven, Hell, and Universal Salvation (New Haven: Yale University Press, 2019).
4. Hans Urs von Balthasar, Dare We Hope “That All Men Be Saved”? With a Short Discourse on Hell, trans. David Kipp and Lothar Krauth (San Francisco: Ignatius Press, 1988).
5. John D. Zizioulas, Being as Communion: Studies in Personhood and the Church (Crestwood, NY: St. Vladimir’s Seminary Press, 1985).
6. Jürgen Moltmann, The Coming of God: Christian Eschatology, trans. Margaret Kohl (Minneapolis: Fortress Press, 1996).
7. Gregory MacDonald [Robin A. Parry], ed., All Shall Be Well: Explorations in Universalism and Christian Theology, from Origen to Moltmann (Eugene, OR: Cascade Books, 2011). See also Parry’s The Evangelical Universalist, 2nd ed. (Eugene, OR: Cascade Books, 2012).
8. Jonathan Edwards, A Dissertation Concerning the End for Which God Created the World, in Ethical Writings, The Works of Jonathan Edwards, vol. 8, ed. Paul Ramsey (New Haven: Yale University Press, 1989), 433–435.
9. Walter J. Schultz, “The Metaphysics of Jonathan Edwards’s End of Creation,” Journal of the Evangelical Theological Society 59, no. 2 (June 2016): 321–339.
10. Gregory of Nyssa’s doctrine of epektasis, perpetual growth into an infinite God, is developed across his Life of Moses and Homilies on the Song of Songs; see also his On the Soul and the Resurrection for the broader argument for universal restoration. [Direct passage citation still to be pinned down for final copy.]
11. Hart, That All Shall Be Saved, 66.
12. Fifth Ecumenical Council (Constantinople II, 553), The Fifteen Anathemas Against Origen, anathema 14, condemning universal apokatastasis; anathemas 4–6 describe demons as fallen rational creatures within the class the anathema addresses. Historians debate whether these anathemas were formally ratified by the ecumenical council itself or originate from the local synod that met under Justinian shortly before it; see Richard Price, trans., The Acts of the Council of Constantinople of 553, 2 vols. (Liverpool: Liverpool University Press, 2009), for the fullest treatment of the conciliar-status question.
13. Thomas Aquinas, Summa Theologiae I, q. 64, a. 2, on the fixity of the angelic will after its first choice.




This is…challenging. My formal background is Lutheran, and I've always chafed against the Calvinist Elect. So I've generally gone with a type of nihilism, e.g. "Hell is eternal separation from God," (my version of Sheol) and happily left the rest to mystery. But this series has forced me to think again. (How dare you! 😉) Your conclusion resonates with my understanding of the grace and love of God. Thank you for putting in the work and sharing it here.
A comedian asks his Christian audience, "What will you do when Satan gets saved?"