The Contradictions and Tensions in the Reformed Confessions

 "Whatsoever Comes to Pass"

An Examination of the Logical Contradictions and Tensions in the Westminster Confession of Faith and the 1689 London Baptist Confession

 

Introduction: Reading the Confessions on Their Own Terms

The Westminster Confession of Faith (1646) is not a sloppy document. It is arguably the single most carefully worded doctrinal statement to come out of the Reformation, the product of five years of debate among more than a hundred divines at Westminster Abbey, cross-checked line by line against Scripture and against centuries of Augustinian and Reformed precedent. Any serious critique of it, or of the confessional documents modeled closely upon it, has to start by admitting that.

It would be easy, and cheap, to catch the Confession in a contradiction by misreading it, by quoting a clause out of its qualifying context, or by assuming a definition of "freedom" or "cause" that the divines never intended. That is not the approach taken here. The method of this essay is to state each Reformed position in its strongest form, using the Confession's own language and the best philosophical and theological defenses of it (Jonathan Edwards' compatibilism, Francis Turretin's doctrine of primary and secondary causation, J. I. Packer's concept of "antinomy") before showing why, even in its strongest form, the position does not escape the charge of contradiction.

The argument of this essay is not that the Westminster divines were unintelligent or dishonest. It is that they inherited two commitments from Scripture and from Augustinian tradition, exhaustive divine determinism and genuine human moral responsibility, that cannot both be true in an unqualified sense, and that the Confession's strategy for holding them together is to assert both and then add a disclaimer ("yet so as," "yet neither is," "yet so as they come most freely") rather than to show how both can be true. A disclaimer is not an argument. Calling the tension a "mystery" is not a resolution; it is a decision not to resolve it while still affirming both halves as flatly true. That is the essay's central claim, and it will be defended section by section, using the Confession's own text throughout. As later sections will show, consistently applying the Confession's own premises does not land on the compatibilism it officially teaches; it lands on hard determinism, a conclusion some within the Reformed tradition itself have already embraced rather than resisted.

Two clarifications before beginning. First, this is a critique of the logical structure of the Confession, not a claim that Calvinists are less sincere, less biblical, or less intelligent than anyone else. Serious, careful people have held this system for four hundred years. Second, pointing out that a system is internally inconsistent is not by itself proof that Christianity, or the sovereignty of God, or predestination in some form, is false. It is proof that this particular formulation of those doctrines cannot be true exactly as stated, and that anyone who wants to keep the underlying biblical commitments has to give up either exhaustive determinism or genuine libertarian responsibility, or seriously qualify what is meant by both. What cannot be done, without contradiction, is affirm both at full strength and call the difference a mystery.

A Note on Method: Scripture, Proof Texts, and the Limits of This Essay

One likely objection to everything that follows deserves to be addressed directly, rather than left for the reader to raise unanswered: this essay argues almost entirely from the Confession's own language and logical structure, and says comparatively little about the individual Bible verses printed alongside each chapter. That is a deliberate choice of method, not an oversight, and it is worth explaining why.

The 1646 Confession itself was adopted without proof texts; Scripture references were added afterward, at Parliament's request, compiled under time pressure rather than argued for at length in the document itself. Even where those references are sound as far as they go, citing a verse beside a doctrinal claim is not the same thing as showing that the verse, read on its own terms and in its own context, establishes the claim. A citation list assumes a whole apparatus of prior interpretive decisions: which framework governs the reading of the passage, which parallel texts control its meaning, and which of several live exegetical options among careful readers of Scripture, Reformed and non-Reformed alike, is correct. The Confession's proof texts do not argue for that apparatus; they presuppose it, and then present the resulting reading as though it were simply what the text says.

Untangling that deeper dispute, verse by verse, across Romans 9, Ephesians 1, John 6, and the dozens of other passages the Confession's proof texts touch, each of which has generated its own centuries-long exegetical literature, would be a separate book, not a section of this one. It is also not necessary for the argument this essay is making. The claim defended here is narrower, and in a sense prior to that debate: that the Confession's own chapters, read together and taken at their word, assert propositions that cannot all be true at once, regardless of which side turns out to be right about the correct reading of any particular proof text. A document that is not internally coherent has a problem no proof text can fix, because the problem is not that the wrong verses were cited; it is that the verses cited, however they are read, are being asked to support two things that will not both fit into a single consistent system. For that reason, this essay treats the Confession's prose, not its footnoted verse list, as the object under examination, and leaves the real, important, and much larger question of whose exegesis of any given passage is correct to the works that specialize in it.

A Note on the 1689 London Baptist Confession

Everything argued in the sections that follow is written primarily against the Westminster Confession of Faith, but it applies with essentially the same force to a second document: the confession adopted in 1689 by an assembly of Particular Baptist churches meeting in London, now generally known as the 1689 London Baptist Confession or the Second London Confession. This is not a loose family resemblance. The Baptist confession was built, chapter by chapter, out of the Westminster Confession by way of the 1658 Savoy Declaration, the Congregationalist revision of Westminster produced by John Owen and others; of the roughly one hundred and sixty paragraphs that make up the 1689 Confession, some one hundred and forty-six are drawn directly from Savoy, itself a light edit of Westminster, with a handful more coming from the Particular Baptists' own earlier 1644 confession or from the 1677/1689 editors themselves. The paragraphs the Baptist framers altered are concentrated almost entirely in the chapters on the church, baptism, and church government, exactly where Baptist and Presbyterian conviction diverge. The chapters this essay is built around are not among them.

Compare the texts directly. Where Westminster III.1 says God did "freely, and unchangeably ordain whatsoever comes to pass," the Baptist confession's own chapter on God's decree says, in the same paragraph position, that God "decreed in himself, from all eternity... freely and unchangeably, all things, whatsoever comes to pass" (1689 LBC 3.1). Where Westminster III.2 denies that the decree was made because God foresaw what would come to pass "upon such conditions," the Baptist confession repeats the sentence almost word for word (1689 LBC 3.2). Where Westminster V.4 insists providence over sin is "not by a bare permission," the Baptist confession uses the identical phrase (1689 LBC 5.4). Where Westminster IX.1 says the will is "neither forced, nor... determined to good, or evil," the Baptist confession says the same (1689 LBC 9.1). And Westminster X.1's account of effectual calling, God "renewing their wills, and, by His almighty power, determining them to that which is good... yet so as they come most freely," is reproduced in the Baptist confession essentially verbatim, down to the word order (1689 LBC 10.1). The same close overlap holds for the chapters on the law, the fall, the gospel offer, perseverance, and assurance examined in Sections V through VIII below.

Because of this, the essay below cites both documents together wherever their wording overlaps, using the format (WCF [chapter].[paragraph]; 1689 LBC [chapter].[paragraph]), and treats an argument that lands against one as landing against the other unless a specific textual difference is noted. There is at least one place where the two documents genuinely diverge rather than merely restate each other, and it is worth flagging in advance rather than papering over: the wording used for reprobation, addressed in Section VI. Readers who hold one confession rather than the other should not assume the argument does not reach them; in every chapter examined here except that one point of phrasing, it was written to reach both.

I. Two Texts, One Confession

Begin with the plain wording. Chapter III, "Of God's Eternal Decree," opens with the most sweeping statement of determinism in the entire document:

"God from all eternity did, by the most wise and holy counsel of His own will, freely, and unchangeably ordain whatsoever comes to pass." (WCF III.1; 1689 LBC 3.1)

Whatsoever comes to pass. Not "whatsoever is good," not "the broad outline of history," but every discrete event without remainder: every thought, every word, every sin. The Confession is explicit that this decree is not grounded in God's foreknowledge of what free creatures would do:

"Although God knows whatsoever may or can come to pass upon all supposed conditions; yet has He not decreed anything because He foresaw it as future, or as that which would come to pass upon such conditions." (WCF III.2; 1689 LBC 3.2)

This is a rejection, in advance, of any view, Arminian, Molinist, or otherwise, in which God's decree responds to or is conditioned by what creatures would freely do. The order of explanation runs only one way: the decree is metaphysically and explanatorily prior to the event, never the reverse. Chapter V, "Of Providence," extends the same claim to the actual execution of history, and goes out of its way to specify that this extends to sin:

"The almighty power, unsearchable wisdom, and infinite goodness of God so far manifest themselves in His providence, that it extends itself even to the first fall, and all other sins of angels and men; and that not by a bare permission, but such as has joined with it a most wise and powerful bounding, and otherwise ordering, and governing of them." (WCF V.4; 1689 LBC 5.4)

Note that last clause carefully: the Confession explicitly denies that God's relationship to sin is "bare permission." Whatever providence is doing with respect to sin, it is doing something more than allowing it to happen. This is a document that has anticipated the easy way out, "God merely permits evil," and closed it off on purpose.

Then, six chapters later, comes this:

"God has endued the will of man with that natural liberty, that is neither forced, nor, by any absolute necessity of nature, determined to good, or evil." (WCF IX.1; 1689 LBC 9.1)

Man's will is "neither forced" nor "determined... to good, or evil" by "absolute necessity." Taken in isolation, that sentence is a textbook statement of libertarian freedom: the view that the will is not fixed in advance by prior causes, that a person genuinely could have chosen otherwise. And Chapter X, on effectual calling, goes further still, describing regeneration as an act in which God, having "renewed their wills, and, by His almighty power, determin[es] them to that which is good... effectually draw[s] them to Jesus Christ: yet so, as they come most freely, being made willing by His grace" (WCF X.1; 1689 LBC 10.1).

So the same document that says God has unchangeably ordained whatsoever comes to pass, and that His providence over sin is more than "bare permission," also says that the human will is not "determined... by absolute necessity," and that a will which has been rendered certain of a single outcome by God's "almighty power" nonetheless comes to that outcome "most freely." These are not merely two doctrines sitting side by side that require careful integration. On the most natural reading of the words, they assert incompatible things about the same event under the same description: that it is fixed with certainty from eternity by a cause outside the creature, and that it is not necessitated and is freely chosen by the creature. The Confession does not ignore the tension: III.1 immediately adds "yet so, as thereby... no violence [is] offered to the will of the creatures," a clause the Baptist confession keeps almost word for word at 1689 LBC 3.1. But appending a denial to a determinist premise does not make the resulting conjunction coherent. It only shows that the divines saw the problem and chose to assert their way past it rather than resolve it.

II. The Best Reformed Answer: Compatibilism and the Natural/Moral Ability Distinction

The most sophisticated Reformed response to this apparent contradiction is not "wait and see." It is a genuine philosophical position called compatibilism, most rigorously developed by the New England Puritan Jonathan Edwards in his 1754 treatise Freedom of the Will, which R. C. Sproul once called the most important work of theology ever produced in America. It deserves to be stated on its own terms.

Edwards distinguishes between natural ability and moral ability. Natural inability is a physical or metaphysical barrier: a man cannot fly no matter how strongly he wills it, because nature does not permit it. Moral inability, by contrast, is simply the strength of one's own inclinations: a devoted husband is, in Edwards' example, morally "unable" to murder his wife for a small reward, not because any external force restrains him, but because he has no desire to do it. For Edwards, the will just is the strongest desire in a given moment; to say a person "chooses" something is to say nothing more than that it is the object their inclination settles on. Freedom, on this account, does not require the ability to have chosen otherwise (what philosophers call the "principle of alternate possibilities"). It requires only that one act according to one's own desires, without external coercion. Edwards illustrates this with a parable of two prisoners offered clemency by a king: one desires to accept but is physically locked in and cannot get out (natural inability); the other has the doors thrown open but has no desire whatsoever to leave his cell (moral inability). Fallen man, on Edwards' reading, is like the second prisoner: the "gates" of the gospel call stand open, but the sinner's own nature has no desire to walk through them. In that sense, Edwards argues, when God regenerates a sinner and gives him a new desire for Christ, He does not force an unwilling man through the door; He gives the man, for the first time, an actual desire to walk through it himself. The choice remains, in Edwards' terms, "the man's own," because it proceeds from the man's own strongest inclination in the moment of choosing, even though that inclination was itself certainly and unilaterally produced by God's decree.

This is not a throwaway move. It is the dominant position among serious Reformed theologians and philosophers today, and it is worth taking seriously precisely because it is an attempt to show how both halves of the Confession can be true, rather than just asserting that they are. Compatibilism does genuine philosophical work: it explains why the Confession can say, without equivocating on the word "freely," that a will effectually and unchangeably determined by God still "come[s] most freely" (WCF X.1), because on this account, "freely" never meant "could have done otherwise." It meant "acting from one's own desires, without external compulsion."

III. Why Compatibilism Does Not Rescue the Confession

Grant Edwards everything: grant that freedom does not require alternate possibilities, and that acting from one's own strongest desire is sufficient for that act to be, in some sense, "one's own." The problem for the Confession is not whether this redefinition of freedom is coherent as a piece of philosophy; plenty of secular philosophers are compatibilists too. The problem is what happens to the concept of moral responsibility once you ask where the desires themselves came from.

This is precisely the question modern philosophy of action has pressed hardest, through what is known as the "manipulation argument," developed most rigorously by the philosopher Derk Pereboom. Pereboom constructs a series of cases involving a man, "Plum," who murders someone for self-interested reasons while satisfying every condition compatibilists have proposed for free and responsible action: his desires are his own, he is not externally coerced in the moment, his action flows from his considered values, he is responsive to reasons. The only difference between the cases is how Plum came to have the character and desires he has: in the most extreme case, a team of neuroscientists directly engineers his psychological states, moment to moment, so that he cannot help but murder. Pereboom's argument is that our intuitions are unshakeable: a person whose entire desire-set has been unilaterally installed by an outside designer, for the designer's own purposes, is not morally responsible for the resulting act, no matter how perfectly that act satisfies compatibilist criteria in the moment of choosing. And, Pereboom argues, an agent under exhaustive prior determinism is not relevantly different from Plum. If covert manipulation of one's desires by another agent undermines responsibility, and if theological determinism is functionally identical to that manipulation (an outside will, in this case the decree of God, unilaterally and irresistibly fixing, from eternity, exactly which desires a person will have and exactly which choice will follow from them), then the same conclusion follows.

This is not an abstract worry. It is exactly the structure the Confession describes. Chapter X does not say God merely permits regeneration to occur, or persuades a genuinely undetermined will; it says God is "renewing their wills, and, by His almighty power, determining them to that which is good." "Determining" is Pereboom's engineer. The sinner's new desire for Christ is not a desire he generates from resources that were ever, at any point, up to him; it is a desire unilaterally installed, with certainty, by an outside agent, for that agent's own purposes, precisely the scenario Edwards' own compatibilist framework cannot distinguish from manipulation. And Chapter III's decree of reprobation is the mirror image: the desires and choices of "the rest of mankind," whom God "was pleased... to pass by" and "ordain... to dishonor and wrath for their sin" (WCF III.7; cf. 1689 LBC 3.3, which reaches the same result in softer language, taken up in Section VI), are equally and eternally fixed, only toward the opposite outcome, and by the same unilateral decree. On Edwards' own account of what a "choice" is, both the regenerate saint's faith and the reprobate's unbelief are simply the outworking of desires that were never, at any stage, self-originated. The Confession wants to say the reprobate are punished justly for "their sin" (III.7), a description that presupposes exactly the kind of ultimate responsibility Pereboom's argument shows compatibilism cannot deliver once the desires themselves are traced to their source.

A common rejoinder at this point tries to undercut libertarian freedom itself, rather than defend compatibilism directly: if moral freedom requires the ability to have done otherwise, the objection runs, then God is not free, since God cannot sin, and if God is nonetheless the paradigm case of a free and morally praiseworthy agent, the ability to do otherwise cannot be what freedom, or responsibility, actually requires. This objection is worth taking seriously rather than waving off, but it does not do the work it needs to do here, for two reasons. First, even granting the point, it shows at most that alternate possibilities are not a strictly necessary condition for freedom in general; it does not show that the source of the will matters any less. God's inability to sin, on the classical account, is not imposed on Him from outside; it flows from His own necessary nature, which is not the product of any decree, and is not owed to any will other than His own. That is exactly what is missing from the Confession's account of the sinner and the elect alike: not the absence of alternate possibilities as such, but the absence of any contribution to the resulting desire that originates with the creature rather than with an external, unilateral decree. Second, and more simply, the objection proves too much for the Confession's own purposes: if divine impeccability shows that determined agents can be morally responsible regardless of source, that same reasoning would just as easily excuse Pereboom's neuroscientists, and everyone else who manipulates another agent's desires toward a guaranteed outcome, which no one, Reformed or otherwise, actually wants to grant. What accomplishes the philosophical work in the manipulation argument is not the presence or absence of alternate possibilities; it is whether the agent's desires are, at some point, self-sourced rather than installed. On that question, the objection from divine impeccability is silent, and the Confession's own doctrine of the decree gives the same answer it always has.

To be fair to the Reformed tradition, this is not a knock-down objection that compatibilists have no answer to; Alfred Mele and others have offered serious "hard-line" replies, arguing that sufficiently gradual, non-coercive manipulation is still compatible with responsibility. That debate continues in professional philosophy, and it is not this essay's job to adjudicate it definitively. But two things can be said with confidence. First, the fact that this remains a live, unresolved dispute among professional philosophers of action means the Confession cannot simply assert, as a settled matter of confessional doctrine, that determined agents are straightforwardly responsible; it is borrowing the most contested move in the entire free-will literature and treating it as obvious. Second, and more importantly for the specific case of the Confession: Pereboom's argument is strongest exactly where manipulation is total, direct, and aimed at a single, unilaterally chosen outcome by a designer with complete knowledge and control, which is a closer description of WCF III and X than of almost any manipulation case philosophers have needed to invent. The Confession does not describe indirect, gradual shaping through ordinary providence (the sort of shaping that even libertarians grant influences character). It describes an eternal, unchangeable, total decree, executed by "almighty power," fixing in advance not just the outcome but the very desires from which the "free" choice will flow. If manipulation of that description undermines responsibility in Pereboom's Professor Plum, parity requires it to undermine responsibility in Westminster's Adam, in Westminster's elect, and in Westminster's reprobate alike.

IV. The Author of Sin: Primary and Secondary Causation

Chapter III's determinism creates an obvious theological hazard: if God has "unchangeably ordain[ed] whatsoever comes to pass," including every sin, has He not authored sin? The Confession anticipates the charge and denies it in the same sentence: God ordains all things, "yet so, as thereby neither is God the author of sin" (WCF III.1; 1689 LBC 3.1). Chapter V repeats the denial in stronger terms, insisting that even though providence over sin is "not by bare permission" but active "ordering, and governing," nonetheless "the sinfulness thereof proceeds only from the creature, and not from God, who... neither is nor can be the author or approver of sin" (WCF V.4; 1689 LBC 5.4, which uses the identical closing phrase).

The classical Reformed defense of this distinction, developed at length by the seventeenth-century theologian Francis Turretin and echoed by modern defenders like R. C. Sproul and the philosopher-theologian Gordon Clark, rests on a distinction between primary and secondary causation. God, on this account, is the primary or ultimate cause of every event, including sinful ones: nothing happens without His decree and concurrence. But the creature is the secondary or proximate cause, the one who actually performs the act, and it is only at the level of secondary causation that moral qualities like sinfulness attach. God's causal "concurrence" sustains the natural act (the motion of the hand, the formation of the thought) without Himself performing, willing as His own, or approving the moral quality that act bears when it flows from a creature's corrupt will. As Reformed writers in this tradition put it, God concurs with the action but not with the sinfulness of the action; He "moves" the instrument without being contaminated by the defect that belongs to the instrument alone.

This is a genuinely careful piece of theology, and it draws on a long scholastic history: the same distinction between primary and secondary causes appears throughout medieval Catholic thought on providence, well before the Reformation. It is not a bare assertion; it is a metaphysical model with real content.

But the model does not do what it needs to do, for a reason internal to the Confession's own commitments. Chapter III.2 has already ruled out the one move that would make the primary/secondary distinction do genuine explanatory work: the idea that God's decree merely tracks what a genuinely undetermined secondary cause would do on its own. If the creature's will were the true, underived source of the sinful act, if God's decree were responsive to, rather than constitutive of, what the creature would freely will, then the primary/secondary distinction would carry real moral weight, in roughly the way a landlord who merely permits a tenant's independent wrongdoing differs from a landlord who scripts it. But WCF III.2, echoed almost word for word at 1689 LBC 3.2, explicitly denies this: God "has not decreed anything because He foresaw it as future... upon such conditions." The creature's will does not supply an independent variable that the decree merely respects; the decree is what fixes, unilaterally and without remainder, that the creature's will would be exactly this and not otherwise. On the Confession's own terms, the secondary cause has no causal history of its own that traces back to anything other than the primary cause. That is what it means to "ordain whatsoever comes to pass." At that point, calling the creature the "author" and God merely the "orderer" is a distinction in vocabulary, not in causal structure: it assigns authorship to the causally last link in a chain every prior link of which was unilaterally fixed by the one being exculpated. In any other context, such as a hypnotist who covertly installs both a subject's desire and the exact object of that desire, or the Pereboom-style neuroscientists who install Plum's murderous desires, we do not accept "the proximate cause performed the act, so the proximate cause is the author" as a moral defense for the one who designed and guaranteed the outcome from the start. The Confession's own denial in III.2, shared by its Baptist counterpart, that anything about the decree is conditioned on the creature closes off the only reading of "primary and secondary causation" under which the distinction would track something morally relevant.

It is worth noting, in fairness, that this is not a fringe worry invented by outside critics; it runs through Reformed literature itself. Even sympathetic expositors of the Confession concede that the term "author of sin" is rarely defined with precision in the tradition, and that the charge persists precisely because on a natural reading of "author" (one who plans, decrees, and guarantees an outcome), the Confession's own description of the decree satisfies it. R. C. Sproul, defending the Confession, wrote candidly that if God foreordains everything, then God must have foreordained the entrance of sin into the world, and then insisted, in the same breath, that this "excruciating problem" still requires affirming that God is not sin's author. That is the pattern this essay has been tracing throughout: the tension is seen clearly, named honestly, and then closed not by resolution but by declaration.

V. Command Against Decree: The Two Wills of God

A further tension sits inside the Confession that is easy to miss because Reformed theology has a long-standing, sophisticated answer ready for it: the relationship between what God commands and what God decrees. Chapter XIX states that "the moral law doth forever bind all... to the obedience thereof" and functions as "a rule of life informing them of the will of God, and their duty" (WCF XIX.5-6; 1689 LBC 19.5-6, which restates both clauses almost verbatim). Call this God's revealed or preceptive will: what He commands and approves. Chapter III, by contrast, states that God has "unchangeably ordain[ed] whatsoever comes to pass" (III.1; 1689 LBC 3.1), and Chapter V insists that this extends to sin by more than "bare permission" (V.4; 1689 LBC 5.4). Call this God's decretive or secret will: what He has actually determined will occur. The problem is immediate: in every case where a person sins, God's revealed will says that person ought not to sin, while God's decretive will has already, unchangeably, and from eternity, fixed that the person will sin. The same God, concerning the same act, wills, in one sense, that it not happen, and wills, in another sense, that it certainly will.

This is not a difficulty the Reformed tradition has ignored or stumbled into by accident. Theologians as far back as Calvin, and worked out at length by Francis Turretin and Jonathan Edwards, have distinguished the will of decree from the will of command (also called the secret and revealed will, or voluntas beneplaciti and voluntas signi) precisely to meet this objection. Their answer is that "will" is being used in two different senses, not one univocal sense applied inconsistently. God's will of decree describes what He has determined will occur, all things considered, for His own wise purposes. God's will of command describes what He approves and requires of creatures, considered simply in itself, without reference to His secret purposes in permitting or ordaining its opposite. On this account there is no more a contradiction here than in a judge who orders a criminal not to commit a crime and also, for reasons of his own, arranges the circumstances that make its commission possible, or in the case of Joseph, where his brothers' sale of him into slavery is described in the same verse as something they "meant... for evil" while God "meant... for good" (Genesis 50:20). Different agents, or the same agent considered under different aspects, can be said to will the same event in different senses without any formal inconsistency.

Grant every part of that distinction. The problem is that it answers a narrower question than the one this section is raising. The two-wills distinction is well suited to showing that God is not self-contradictory, that His approval of virtue and His decree of a particular sinful event can coexist without impugning His own character or making Him a liar when He commands obedience. That is a real question, and the distinction answers it. But it is a question about God's consistency, not about the creature's responsibility, and the Confession needs both. The question this section is pressing is different: given that God's decretive will has fixed, with the same "unchangeably" force used throughout Chapter III, that a particular person will commit a particular sin at a particular moment, in what sense does that person retain the capacity to have kept the command instead? The Genesis 50:20 analogy illustrates that two purposes can coexist behind one event; it does not show that the brothers could have chosen not to sell Joseph. If their choice was as unconditionally decreed as everything else the Confession describes, the text is being asked to do two jobs: establish that God's and the brothers' purposes differed, which it does, and establish that the brothers remained free enough, despite the decree, to bear the full moral weight the text also assigns them, which it merely asserts.

Put plainly, the structure the Confession describes is closer to commanding a man not to injure himself while simultaneously, and with total certainty, arranging every antecedent condition that guarantees he will, and then holding him responsible for the injury as though the command had left the outcome genuinely open. The judge and Joseph analogies that make the two-wills distinction feel intuitive work only because they smuggle in exactly the kind of independent contribution from the secondary agent that Chapter III.2 has already excluded. A judge who merely creates an opportunity a criminal is free to decline, or a father who permits a risk his son might or might not take, is arranging circumstances around a genuinely undetermined choice; that is what makes the analogy feel fair. But WCF III.2 denies that any creaturely choice, including the brothers', has that kind of independence from the decree. Once the secondary cause has no causal history of its own, the very feature that made the judge and the father blameless for their subjects' choices, namely that the choice was not theirs to fix, is no longer available to the Confession's own account of the decree. This is the same gap already identified in Section IV's discussion of primary and secondary causation, appearing again at the point where the Confession's doctrine of the law meets its doctrine of the decree.

A further Reformed reply worth anticipating is that fallen inability is itself culpable: mankind lost, through Adam, the ability to keep the moral law perfectly, so the fact that fallen people cannot help but sin does not excuse them, since the incapacity is inherited guilt, not innocent misfortune. It is worth pausing on that premise before assessing it. The claim that Adam's guilt is itself imputed to his posterity (WCF VI.3; 1689 LBC 6.3, which uses the same language of imputed guilt and conveyed corruption) is not a neutral, uncontested fact this essay is adopting; it is a substantive doctrine that many non-Calvinist traditions reject outright, in whole or in part, and nothing here should be read as endorsing it. Consistent with the Method section above, it is engaged for the same reason every other doctrine in this essay has been: because the Confession asserts it, and the task at hand is to test whether the Confession's own claims hold together on the Confession's own terms, not to independently litigate the doctrine of original sin from outside. Granting it, then, purely for the sake of argument: it is a coherent point about the general, inherited corruption of the will (WCF IX.3; 1689 LBC 9.3, nearly identical in wording), but it answers a different, broader-grained question than the one at issue here. Inherited corruption might explain why sinners in general lack the moral ability to keep the law perfectly. It does not, by itself, explain why any particular sinful choice, individually and specifically fixed by the eternal decree before the sinner existed to inherit anything, is one for which that sinner bears the same kind of responsibility a genuinely undetermined agent would. That move quietly reintroduces the natural/moral ability distinction examined in Section II, and inherits the same problem Section III raised against it: once the specific desire behind the choice is traced to an external, unilateral source rather than the agent's own, Pereboom's manipulation argument applies with the same force whether the source is called Adam's federal headship or the decree directly.

None of this shows that the revealed and decretive wills are logically incoherent as two senses of "willing"; that much of the classical distinction survives intact. What it shows is that the distinction, however real, does not do the further work the Confession needs it to do: it explains why God can command one thing and ordain another without being double-minded, but it does not explain how the creature commanded can be justly held to a standard the same decree has already, and with the same certainty invoked everywhere else in the document, made it impossible to meet. Appealing to "the secret things [that] belong to the Lord our God" (Deuteronomy 29:29) at this point is simply another instance of the pattern traced throughout this essay: naming the tension, and then declining to resolve it under cover of reverence for mystery.

VI. Presupposing the Conclusion: How the Confession Defines Election

A separate, and in some ways more basic, problem concerns method rather than content. Chapter III does not argue for unconditional election from some prior, neutral ground; it defines election in a way that has already excluded every rival view before the argument begins.

Consider III.5: those predestined to life are chosen "without any foresight of faith, or good works, or perseverance in either of them, or any other thing in the creature, as conditions, or causes moving Him thereunto" (WCF III.5; cf. 1689 LBC 3.5: "without any other thing in the creature as a condition or cause moving him thereunto"). This is not a conclusion reached after weighing the biblical and philosophical evidence for conditional versus unconditional election. It is the definition of election as unconditional stated as a premise, with every conditional alternative (classical Arminianism, Molinism, any view in which God's choice responds in any way to what He foresees a person would freely do) ruled out by definitional fiat rather than argued against. No confession, including this one, can define a doctrine from a vantage point outside every tradition; every formulation of election reflects a particular starting point, and this essay does not fault the Confession simply for having one. What is at issue is narrower than that: the difference between presenting unconditional election as a conclusion the document argues for, against a fairly stated case for its conditional rivals, and presenting it as the very definition of the thing being described, so that the conditional alternatives are not so much refuted as defined out of existence before the case for them is ever heard. The Confession does the latter. This is not unique to the Confession; it is arguably the interpretive posture of the entire Reformed system, sometimes called an appeal to "the analogy of faith," in which the system's central commitments (God's absolute deterministic sovereignty, monergistic grace) function as the lens through which every individual text and every rival reading is pre-filtered. That may be theologically defensible as a hermeneutical starting point, but it cannot then be presented as though the resulting document has independently demonstrated, rather than assumed, the conclusion that rival systems dispute.

A related and more localized problem, worth flagging because it recurs constantly in popular Reformed teaching, including lectures on this very chapter, is a terminological slide between "election" and "predestination." It is worth being precise about whose definitions are in play here, since the vocabulary is contested well outside Reformed circles: many non-Calvinist traditions do not use either term the way the Confession does at all. Corporate-election readings, for instance, take "the elect" to name the church as a body chosen in Christ rather than a fixed set of individuals selected apart from any foreseen response, and Arminian and Wesleyan writers typically tie election to foreseen faith in a way III.5 explicitly rules out. What follows examines "election" and "predestination" as the Confession itself defines and uses them; it is not thereby conceding that this is the only, or even the majority, way the underlying biblical vocabulary has been read across the Christian tradition, and the presence of Scripture references beside III.3 through III.8 does not settle that question either, since, as noted in the Method section above, those citations already presuppose the Confession's own reading rather than independently establishing it. Within the Confession's own terms, then: technically, in classical theology, election refers narrowly to God's choice of some to salvation; predestination, taken in its full and traditional sense, is the broader decree encompassing both election (to life) and reprobation (to death, or at minimum, to being "passed by"). The Confession's own chapter title is "Of God's Eternal Decree," and within it, III.3 distinguishes "predestinated unto everlasting life" from "foreordained to everlasting death" as two sides of a single decree, yet III.8 then speaks simply of "this high mystery of predestination" as though it were one undifferentiated doctrine, a formula the Baptist confession repeats at 1689 LBC 3.7, and popular usage (including, frequently, in preaching and lecturing on the chapter) collapses the two terms into each other freely. This is not fatal to the system by itself, but it matters for how the doctrine is defended in practice: when "election" and "predestination" are used as though interchangeable, the comparatively easier case for election (God graciously choosing to save some) gets rhetorically borrowed to cover the much harder case of reprobation (God ordaining others to "dishonor and wrath," III.7) without that harder case ever being separately defended on its own terms. Precision about which term is doing which job would not resolve the deeper problem in Section IV above, but its absence lets the more palatable half of the doctrine quietly carry the weight that only the full, symmetrical doctrine of double predestination should have to bear.

The one place where the Baptist confession's wording meaningfully departs from Westminster's is worth setting out precisely, since it is the closest thing to a genuine improvement the 1689 Confession makes on this point, even if it does not go far enough to escape the underlying problem. Westminster III.7 describes the non-elect in the language of positive decree: God was pleased to "pass by, and to ordain them to dishonour and wrath for their sin." The Baptist confession's parallel clause is gentler: "others being left to act in their sin to their just condemnation" (1689 LBC 3.3). "Left to act in their sin" reads more like Westminster's own vocabulary of permission than like an active ordination to wrath, and a defender of the Baptist confession might reasonably point to this softer phrasing as evidence that it avoids the harshest edge of Westminster's double predestination. But the softening does not survive the very next paragraph. 1689 LBC 3.4 states that "these angels and men thus predestinated and foreordained, are particularly and unchangeably designed, and their number so certain and definite, that it cannot be either increased or diminished," a sentence that explicitly applies "predestinated and foreordained" back to both groups named in 3.3, the saved and "the rest" alike. Whatever comfort "left to act in their sin" offers at the level of phrasing, it describes an outcome that is, by the confession's own next sentence, just as unchangeably and particularly fixed in eternity as the positive election of the saved. The difference between the two confessions here is stylistic rather than structural: Westminster names the mechanism of reprobation more starkly, while the Baptist confession describes the same fixed, unconditional outcome in gentler, more passive language. Section III's argument about the reprobate's fixed, unilaterally sourced unbelief applies to both confessions without alteration.

VII. The Well-Meant Offer: A Crack the Reformed Tradition Cannot Close Internally

If the previous sections rely on philosophical argument from outside the tradition, this one does not need to. It is enough to point to an unresolved controversy within confessional Calvinism itself, one that has run for centuries and shows no sign of ending, because it is generated by the Confession's own text.

Chapter VII.3 states that God "freely offereth unto sinners life and salvation by Jesus Christ" (WCF VII.3; 1689 LBC 7.2, which uses the identical phrase). Chapter X.4 makes clear that this offer is universal in scope, extended even to "others, not elected [who]... may be called by the ministry of the Word" (WCF X.4; 1689 LBC 10.4). The question this raises is simple: in what sense can God "freely offer" life to someone for whom, on the doctrine of limited atonement, Christ did not die, and whom, on the doctrine of unconditional election, God has already and unchangeably decreed to "pass by... and ordain... to dishonor and wrath" (III.7)? An offer, in ordinary language, implies a genuine willingness on the offerer's part that the offer be accepted. Reformed theologians who affirm what is called the "well-meant offer," a majority position among mainstream confessional Presbyterians, including figures like John Murray and Sinclair Ferguson, insist that God does sincerely desire the salvation of every hearer of the gospel, elect and reprobate alike, even though He has eternally and unchangeably decreed that only the elect will be saved.

What makes this section different from the others is that a significant, serious wing of the Reformed tradition itself, so-called hyper-Calvinists, including the Protestant Reformed Churches and writers associated with them, reject the well-meant offer for precisely the reason a non-Calvinist critic would: they argue it is logically incompatible with particular, sovereign grace, because it requires attributing to God a desire (the salvation of the reprobate) that His own eternal decree has already foreclosed. Their solution is to bite the bullet and deny that the gospel call expresses any divine desire for the reprobate's salvation at all, preserving logical consistency by giving up the "well-meant" part of the offer. Meanwhile, defenders of the well-meant offer accuse the hyper-Calvinists of departing from the plain sense of the Confession's own language in VII.3 (and its Baptist parallel, 7.2) and from Calvin's own explicit statements that the gospel promise is addressed to all. Both sides cannot be right about what the Confession consistently implies, and the fact that confessionally committed Calvinists have been unable to agree, across centuries of careful exegetical and systematic argument, on which half of their own document to give up, is itself strong evidence that the document asserts two things that do not sit together: universal, sincere gospel love, and a decree that has, from eternity, made the salvation of most of the gospel's hearers metaphysically impossible and never truly intended. This is not an argument manufactured by outsiders. It is the Reformed household's own unresolved argument with itself.

VIII. Assurance Undermined from Within: Evanescent Grace

The final contradiction concerns not the logical structure of the decree but its pastoral consequence, and here the Confession's own resources make the problem sharper than any outside critic needs to make it.

Chapter XVIII promises "an infallible assurance of faith" available in this life to "such as truly believe in the Lord Jesus" (WCF XVIII.1-2; 1689 LBC 18.1-2). This is meant as good news: the believer need not live in perpetual doubt about their standing before God. But the promise runs directly into a problem the Reformed tradition itself created and has never been able to fully close: the doctrine of temporary, or, in language drawn directly from Calvin's own description of it, evanescent, grace.

John Calvin, in the Institutes (III.2.11), addresses head-on the fact that Scripture sometimes describes people who later fall away in language nearly indistinguishable from that used of the truly regenerate: "a taste of heavenly gifts" (Hebrews 6), "a temporary faith" (Luke 8:13). Calvin's explanation is that the reprobate can be given, by God, an experience of grace that is subjectively identical to the experience of the elect (he writes that "the reprobate are sometimes affected in a way so similar to the elect, that even in their own judgment there is no difference between them") while remaining, in fact, unregenerate. What distinguishes true, persevering faith from this "confused" and merely "temporary" faith is not anything the believer can introspectively detect in the moment; it is revealed only retrospectively, by whether the faith in question endures to the end. Calvin himself describes this sense of grace in the reprobate as one that "afterwards proves evanescent," language later Reformed writers and critics alike have used to name the doctrine itself.

The Baptist confession makes this problem harder to deflect, because it does not require an appeal to Calvin's Institutes to establish the category; it states the category directly in its own text. Its chapter on saving faith distinguishes true faith from "the faith and common grace of temporary believers" (1689 LBC 14.3), and its chapter on assurance opens by warning that "temporary believers, and other unregenerate men, may vainly deceive themselves with false hopes and carnal presumptions of being in the favour of God and state of salvation" (1689 LBC 18.1), immediately before promising "such as truly believe" an infallible assurance in the very same paragraph. The confession's own text places the false case and the true case side by side without supplying, at the level of introspectable experience, any way to distinguish them in the moment: both the deceived temporary believer and the genuine article are described in the shared vocabulary of hope, felt favor with God, and examined conscience. The problem drawn out of Calvin's Institutes above is therefore not an outside imposition on the Baptist confession; it is visible in the confession's own chosen wording, in the same breath in which it offers its promise of assurance.

The pastoral and logical problem this creates has been noted by critics both outside and inside the Reformed camp for exactly as long as the doctrine has existed, and it applies with full force to the Confession's own promise of assurance. If the reprobate's experience of grace can be, by Calvin's own description, subjectively indistinguishable from the elect's, then no believer, however sincere, however examined their heart, however fruitful their life currently appears, has any evidence available to them, from the inside, that rules out the possibility that they are among those receiving only the "confused sense of grace" Calvin describes, rather than the genuine article. The Confession tries to soften this in XVIII.3 by locating assurance not in a "bare conjectural" feeling but in "the inward evidence of those graces" and "the testimony of the Spirit," but Calvin's own doctrine is precisely that this "inward evidence" is exactly what the reprobate can also, and does, experience, right up until the moment their faith withers. The Confession's promised assurance therefore rests on a form of evidence the tradition's own founding theologian says is not reliably diagnostic of the very thing it is supposed to certify. This is not a hostile misreading; it is Calvin's explicit teaching about why some who appear to have "obtained an inheritance" (his language, echoing Hebrews) later fall away without ever having truly possessed what the elect possess.

Chapter XVII compounds the problem rather than solving it. Perseverance of the saints guarantees that the truly elect "can neither totally nor finally fall away" (WCF XVII.1; 1689 LBC 17.1: "can neither totally nor finally fall from the state of grace"), but this guarantee is available only to the elect, and whether any given professing believer is among the elect is, by the logic of Chapter XVIII combined with Calvin's doctrine of temporary faith, exactly the fact that cannot be settled with certainty from present experience alone. The doctrine of perseverance guarantees an outcome for a class of people whose membership the believer cannot, even in principle, verify from the inside with the kind of "infallible assurance" Chapter XVIII promises, because the very phenomena offered as evidence (a sense of grace, fruit, inward testimony of the Spirit) are, on Calvin's own account, sometimes present in people for whom the promise does not, in the end, hold. The comfort the Confession offers with one hand it structurally withdraws with the other.

IX. Putting It Together

Because the Baptist confession reproduces the relevant chapters almost verbatim, as set out above, each of the six problems below recurs in it with essentially the same force, subject only to the qualification, noted in Section VI, that its reprobation clause is phrased more softly without escaping the same underlying structure. None of these six problems stands alone; they compound each other. The Confession asserts unqualified determinism (III.1-2, V.4) alongside unqualified creaturely freedom and responsibility (IX.1, X.1) without a model that preserves both without collapsing one into the other, and the best available model, compatibilism, fails for the specific, total, outcome-fixing kind of determinism the Confession itself describes (Sections I-III). It denies that God is the author of sin while defining the decree in a way that removes the one condition, an independently sourced secondary cause, that would make that denial more than verbal (Section IV). Its classical distinction between God's revealed and decretive wills explains why God is not self-contradictory in commanding what He has not decreed, but leaves untouched the further question of how a creature can be justly bound by a command the same decree has already made impossible to keep (Section V). It presents unconditional election as though demonstrated when it has in fact been assumed in the very definitions used to state it, and it allows the harder doctrine of reprobation to travel under cover of the easier doctrine of election through imprecise terminology (Section VI). Its own confessional heirs cannot agree among themselves whether the gospel's universal offer is compatible with the doctrine's particularity, and that argument has run, unresolved, for as long as the Confession has existed (Section VII). And its promise of assurance is undercut by its own tradition's honest acknowledgment that the very evidence assurance is supposed to rest on is, by that same tradition's admission, sometimes present in those for whom it will ultimately prove false (Section VIII).

Each time the tension surfaces, the pattern is the same: state both halves plainly, then add a qualifying clause, "yet so as," "yet so, as," "yet neither is," that denies the contradiction without resolving it. That rhetorical move is not unique to Westminster; it is the standard Reformed response, elevated to a formal principle by J. I. Packer under the name "antinomy": an "apparent contradiction between conclusions which seem equally logical, reasonable, or necessary," which believers are counseled to "accept... and learn to live with" rather than resolve. Packer is careful to distinguish an antinomy from mere confusion; he insists both halves rest on "clear and solid evidence" and that the difficulty lies only in seeing how they fit together, not in whether they are both true.

But that distinction cannot do the work Packer needs it to do, and it is worth being precise about why. An antinomy, on Packer's own definition, is legitimate only if both propositions really are independently well-established and the only thing missing is a reconciling explanation, the way, for instance, physicists once held that light behaves as a wave and as a particle without yet having quantum field theory to unify the two, because both descriptions were independently, empirically verified beyond dispute. That is categorically different from the case examined here. In each section above, the difficulty was not merely that no one has yet supplied the reconciling account; it was that the reconciling account that has been supplied (compatibilism, primary/secondary causation, the well-meant offer, temporary faith as distinct from saving faith) was examined directly, on its own best terms, and shown to fail to deliver what it promises once its own implications are traced out. That is a different, and more serious, kind of problem than Packer's model allows for. A true antinomy says "these are both true, and I cannot yet show you how." What has been shown here is closer to: "these cannot both be true as stated, and the explanations offered for why they nonetheless can be, do not survive being followed to their conclusion." Labeling that outcome a "mystery" does not change its logical status. It simply moves the unresolved contradiction from the object language of the Confession into the reader's obligation to stop asking the question.

X. What Coherence Would Require

It is worth closing by being constructive rather than merely negative, because the alternative to the Confession's formulation is not "abandon the sovereignty of God" or "abandon human responsibility," both of which remain live, biblically serious options across the wider Christian tradition. What the argument above actually shows is narrower: that a system cannot, without contradiction, affirm all of the following simultaneously and without qualification:

1. God's decree is not conditioned in any way on what creatures would freely do (WCF III.2).

2. That decree unchangeably and irresistibly fixes every event, including every sinful choice and every act of saving faith (WCF III.1, V.4, X.1).

3. The creature alone, and not God, bears the moral authorship of sin, in a sense of "author" that requires some independent contribution from the creature (WCF III.1, V.4).

4. The creature's will is not determined by "absolute necessity" and its choices remain genuinely "free" in a sense that grounds real moral responsibility and a sincere, universal gospel offer (WCF IX.1, VII.3, X.4).

The same four propositions, in essentially the same wording, are asserted across chapters 3, 5, 7, 9, and 10 of the Baptist confession; the same four-way tension, and the same three ways out, apply to it without modification.

Give up (1), and you have something like classical Arminianism or Molinism, in which God's decree does respond to what libertarian-free creatures would do, and (3) and (4) become far more defensible, at the cost of (2) in its strongest form. Give up (2) in favor of a bare-permission or limited providence, and (3) becomes easier to sustain, but at the cost of the Confession's own explicit rejection of "bare permission" in V.4. Give up (4): accept, with full consistency, that "freely" means only "without external coercion," and that no one, elect or reprobate, ever had it in their power to be otherwise than the decree made certain, and the system becomes internally coherent as a form of theological determinism, at the acknowledged cost of everything ordinary language and the Confession's own words mean by human freedom, sincere divine desire, and creaturely authorship of sin.

That last option is not a hypothetical escape route invented for the sake of tidy philosophy. Some within the Reformed tradition, confronting exactly the tension traced through Sections I to III, have already taken it. The Reformed writer Vincent Cheung has argued openly against compatibilism from within Calvinism, on the grounds that if God has decreed all things without remainder, talk of a will that is nonetheless "free" in any morally load-bearing sense is not a solution but an evasion, and that a consistent Reformed theology should affirm hard determinism outright rather than keep the language of human freedom. Other Reformed writers arguing in the same vein have made the same move under the label "theological hard determinism," insisting that semi-Calvinist attempts to preserve a compatibilist free will remain determinism regardless, just an inconsistent version of it. This is worth taking seriously rather than dismissing as a fringe position: it is what strict consistency with (1) and (2), pursued without the Confession's own qualifying clauses, actually yields. The Confession itself does not take this path. It wants (2) and (4) together, in the very same document, without ever admitting that holding both requires giving one of them up rather than qualifying it into paradox. What cannot be done is hold all four at once and call the remaining tension a mystery. That is not a mystery. It is a set of propositions, at least one of which the Confession must actually give up, not merely qualify with a "yet so as," for the whole to be logically coherent.

Conclusion

The Westminster Confession of Faith, and the 1689 London Baptist Confession that restates its central chapters in nearly identical language, are monuments of careful, biblically saturated theology, and their authors were not naive about the difficulties their system created; the recurring qualifying clauses throughout Chapters III, V, IX, and X, clauses their Baptist heirs kept rather than revised, are proof that they saw the tension clearly and tried, again and again, to close it. But seeing a problem and stating around it are not the same as solving it. Examined on its own terms, using its own best defenders (Edwards on the will, Turretin and Sproul on primary and secondary causation, Calvin on the reprobate's experience of grace, Packer on the theology of mystery), the Confession asserts a total, unconditioned divine determinism and a genuine, responsibility-grounding human freedom that its own text will not allow to coexist, denies divine authorship of sin using a causal distinction its own doctrine of the decree empties of content, distinguishes a revealed will that commands obedience from a decretive will that has already fixed disobedience without explaining how the creature bound by the one can be justly liable under the other, defines election in terms that presuppose the very conclusion at issue, cannot achieve internal agreement among its own confessional heirs about whether its universal gospel offer is compatible with its doctrine of particular grace, and promises an assurance of salvation that its own founding theologian's doctrine of temporary faith quietly withdraws. Four hundred years of careful restatement, across two confessional documents in which the later simply kept the earlier's wording rather than revising it away, have refined the language of these tensions considerably. They have not resolved them, and where the Confession's own heirs have pressed its premises to their consistent end rather than qualify them again, the destination has not been the compatibilism the document teaches. It has been the hard determinism the document was written to avoid sounding like.