I. The Instrument Named
There is a use of the counterfactual that is a game and a use that is an instrument, and the two are easily confused, to the discredit of the second. The game asks what would have happened if the arrow had missed, if the messenger had ridden faster, if the frost had held one more week — and it asks for the pleasure of the asking, spinning alternative histories the way a man idly rearranges the furniture of a room he does not intend to change. The game has its charms and its uses, but it is not what this paper means. This paper means the counterfactual as an instrument: a device for furnishing memory honestly, whose function is not to entertain alternatives but to correct a specific distortion in how the actual past is stored.
The distortion is the illusion of fatedness. A past that has happened presents itself to memory as the only thing that could have happened. What occurred acquires, by the mere fact of occurring, an air of necessity — as though the road taken were the only road, the outcome reached the only possible outcome, the whole sequence locked from the start into the shape it finally assumed. This air of necessity is a property of the storage, not of the events. The events, while they were occurring, were open; men did not know how they would turn out; the outcome hung, at more than one point, on a margin that could have fallen the other way. But once the outcome is known, the openness drains out of the memory of the sequence, and the sequence is re-stored as a thing that marched to its conclusion. The counterfactual, disciplined, is the instrument that restores the drained-out openness. It re-furnishes the past with the contingency that the knowledge of the outcome removed.
So the counterfactual belongs in a cluster on memory because it is a memory instrument in the strict sense this project has given the word. It does not tell the researcher what happened; the record does that. It tells him how to store what happened — whether as a fated march or as a contingent sequence that could have gone otherwise — and the difference between these two storings is the difference between a laundered past and a textured one. This paper’s argument is that the fated storing is a laundering, that both of the great verdicts on the past secretly require it, and that the counterfactual, by breaking it, disciplines both verdicts with a single stroke.
A boundary must be set at once, because within the framework of this project the counterfactual raises a question the merely secular treatment never has to face, and to leave it unaddressed would be to launder the paper’s own difficulty. The biblicist holds that God works all things after the counsel of his own will, that not a sparrow falls to the ground without the Father, that the lot cast into the lap is disposed wholly of the LORD. If God has ordained what comes to pass, in what sense could the past have gone otherwise? Does not the counterfactual, which trades on the past’s openness, deny the very providence the framework confesses? The question is real and is answered in Section II, not deferred; the short form of the answer is that the certainty of God’s decree and the contingency of second causes are not rivals but a distinction the Scripture itself maintains, and that the counterfactual operates at the level of second causes, where the openness is genuine, rather than at the level of the decree, where the certainty is absolute.[^1] The instrument does not deny that God ordained the outcome. It denies that men, reasoning after the fact, may read the outcome’s certainty in God’s counsel off the outcome’s appearance in their filtered memory — which is a different thing, and a laundering.
II. The Illusion It Breaks
The illusion the counterfactual breaks is the sense that the actual past was fated. This sense is not a considered belief; almost no one, asked directly, will assert that history was locked from the beginning into exactly the course it ran. It is a property of storage that operates beneath assertion, coloring the memory of a sequence with a necessity the sequence did not have while it was live. And it operates by a mechanism worth naming precisely, because naming it is the beginning of correcting it.
The mechanism is the knowledge of the outcome. A sequence of events, stored after its outcome is known, is stored toward that outcome. The knowledge of where the road led reaches back and reorganizes the memory of the road, so that the turns that led toward the known destination are remembered as significant and the turns that led away are remembered as detours or forgotten, and the whole acquires the shape of a path to a destination rather than a walk through a country with many forks. This is not deliberate. It is what memory does with a sequence whose ending it knows: it stores the sequence as an explanation of the ending, and an explanation of an ending makes the ending look explained, which is to say necessary, which is to say fated. The counterfactual interrupts this by forcing the researcher back to a fork and requiring him to stand at it as the men who stood at it stood — not knowing which way the road would go — and to reason honestly about how nearly it might have gone the other way. Where the reasoning shows the other way to have been close, the fatedness is exposed as an artifact of hindsight, and the openness is restored.
This is the point at which the theological question must be met rather than evaded, for the biblicist reader will rightly ask whether “the openness was restored” is not simply false — whether, God having decreed the outcome, there was ever any openness to restore. The answer requires a distinction that Scripture itself supplies and that the Reformed tradition, at its most careful, has articulated: the distinction between the certainty of the divine decree and the contingency of the second causes through which the decree is accomplished. That God ordained Joseph’s exaltation does not mean the brothers did not freely and wickedly sell him; the text holds both, in one verse — ye thought evil against me; but God meant it unto good. The selling was a free act of second causes, contingent in the sense that the brothers were under no compulsion and might, without any contradiction in the nature of things, have done otherwise; and it was certain, in the sense that God had ordained the end and the means. The contingency and the certainty occupy different levels and do not compete. The decree is certain. The second causes are contingent. Both are true at once, and the biblical narrative depends on their both being true, for it holds men genuinely responsible for acts God genuinely ordained.[^2]
The counterfactual operates entirely at the level of second causes, and this is what licenses it within the framework. When the researcher stands at the fork and asks how nearly the messenger might have arrived too late, he is reasoning about the contingency of second causes — about the openness that was really there in the created order of horses and roads and weather and human decision — and not about the counsel of God, which is hidden from him and into which he is not inquiring. He is not asking whether God could have decreed otherwise, a question that is idle and beyond him. He is asking whether the created sequence, considered at the level on which men act and observe, contained a margin at which it might have run differently — and the Scripture’s own maintenance of contingent second causes under a certain decree is what tells him that it did. The illusion the counterfactual breaks is therefore not the truth of providence but a counterfeit of it: the hindsight-manufactured sense that the appearance of necessity in one’s filtered memory of a sequence is a reading of the divine decree. It is no such thing. It is a property of the storage, and the biblicist has particular reason to distrust it, since it claims to read off the surface of events a certainty that belongs to the hidden counsel of God and is not given to men to read that way at all. To store the past as fated because it happened is not piety. It is the presumption of reading providence backward from outcomes, and it is a laundering that flatters whatever the outcome happened to be.
So the counterfactual, far from being in tension with the framework, is licensed by it and even required by it. The framework forbids the researcher to read the necessity of what happened off the fact that it happened, because that necessity belongs to a decree he cannot see; and the counterfactual is the instrument that enforces the prohibition, restoring to his memory the contingency of the second causes that the decree worked through — the only contingency he has access to, and the real one.
III. The Hidden Assumption in Both Verdicts
The illusion of fatedness would be a minor matter — a private tic of memory, correctable at leisure — were it not that the two great verdicts on the past both secretly depend on it. This is the paper’s central claim, and it is the claim that makes the counterfactual not merely a useful instrument but the specific corrective this cluster needs. The progress verdict and the decline verdict, which appear to be opposites and to exhaust the field between them, share a buried premise, and the premise is fatedness. Neither can do its work without quietly assuming that the past could only have gone as it went.
Take the progress verdict first. The progressive stores the past as an ascent — a sequence of stages, each an advance on the last, arriving at the present as the ripe fruit of a maturing process. The verdict on any earlier stage is fixed by its position on the ascent: it was worse than what followed because it was earlier, and its defects were the defects appropriate to its place on the climb, destined to be outgrown. Now observe what this storing requires. It requires that the ascent be the only path the past could have taken — that the earlier stage led necessarily to the later, that the maturing was a real maturing and not a contingent sequence of turns that happened to arrive where it arrived. For if the past could have gone otherwise — if at some fork it might have turned toward a worse present, or toward a present incommensurable with this one — then the present is not the ripe fruit of a maturing process but the residue of choices that might have been made differently, and the credit the progressive claims for the present, as the earned outcome of an ascent, dissolves. He needs the fatedness. Without it, his ascent is merely one path among the paths the past might have taken, and the present at its summit is merely where this path happened to end, carrying no verdict on the stages behind it.[^3]
Now take the decline verdict, which appears to be the progressive’s opposite and is in fact his twin. The declinist stores the past as a descent from a golden age — a sequence of losses, each a falling-away from a former excellence, arriving at the present as the impoverished residue of a squandered inheritance. The verdict on the golden age is fixed by its position at the top of the descent: it was better than what followed because it was the height from which the falling began, and its goods were stable possessions, lost by the fault of those who let them go. And observe what this storing requires. It requires that the golden age’s goods were a secure possession rather than a contingent arrangement — that they were the kind of thing one has and can lose, rather than the kind of thing that held for a while under conditions that might have failed sooner or differently. For if the golden age could have collapsed earlier, or need never have cohered at all, if its excellences were a fortunate configuration that the margin might easily have denied — then its goods were not a possession squandered but a contingency that ran its course, and the declinist’s grief, which mourns a stable good betrayed, has nothing stable to mourn. He too needs the fatedness, from the other end: he needs the golden age to have been a fixed height rather than a lucky and precarious arrangement, so that the fall from it can be a fall rather than the ordinary dissolution of an arrangement that was never guaranteed to hold.
The two verdicts thus share the fatedness premise and use it from opposite ends. The progressive needs the ascent to have been necessary, so that the present is earned. The declinist needs the golden age to have been secure, so that the loss is culpable. Strip the fatedness from either, and its verdict loses its ground. The progressive’s earned present becomes a contingent endpoint; the declinist’s squandered inheritance becomes a lapsed arrangement. And here is the thing to see: the fatedness they share is the very thing the counterfactual removes. This is why one instrument disciplines both. The counterfactual does not argue against progress and then, separately, against decline. It removes the single premise on which both verdicts silently stand, and both fall for the same reason, at the same moment, by the same stroke.
IV. The Counterfactual as Corrective to Both
The economy of this is the heart of the matter and deserves to be stated as an economy. A cluster on the honest furnishing of memory must correct two opposite distortions: the gilding that stores the past as a golden age fallen from, and the flattening that stores the past as a crude stage climbed out of. These are the two directions in which memory is laundered, and Paper 1 named them as the two halves of the double corrective — Ecclesiastes 7:10 against the gilding, Ecclesiastes 1:9 against the flattening. One might expect that two opposite distortions require two distinct correctives, held in tension, each guarding against the excess of the other. The counterfactual shows that they do not. Because the two distortions rest on a common premise, a single instrument that removes the premise corrects both, and the researcher is not obliged to balance two correctives against each other but has one, which he applies in a single direction against a single shared error.[^4]
The correction runs as follows. Against the progressive, the counterfactual stands at the forks of the supposed ascent and asks how nearly each turn might have gone otherwise — and where it shows the turns to have been contingent, it converts the ascent from a maturing into a walk, and the present at its summit from an earned outcome into a place this walk happened to reach. The progressive verdict on the earlier stages, which depended on their being necessary steps toward a destined present, loses its warrant: they were not steps toward anything destined; they were forks, and the present is where the forks happened to lead. Against the declinist, the same instrument stands at the conditions of the golden age and asks how nearly they might have failed sooner, or failed to cohere at all — and where it shows the golden age to have been a contingent configuration, it converts the descent from a fall into a dissolution, and the golden age from a secure possession into an arrangement that held for a while. The declinist verdict, which depended on the golden age being a stable good culpably lost, loses its warrant: there was no stable good to lose; there was a fortunate arrangement that ran its course, as arrangements do.
Note that in neither case does the counterfactual deliver the opposite verdict. It does not prove that the past was worse than the progressive thinks, nor that the golden age was less good than the declinist thinks. It is not a verdict-machine that outputs the reverse of whatever it is fed. It removes the fatedness that let the verdict be read off the sequence’s shape, and it leaves the researcher facing the evidence about the period’s actual goods and defects — which is to say, it hands him back to Paper 2’s problem of texture, now unobstructed by the fatedness that had been doing the verdict’s work for it. The earlier stage may still turn out, on the evidence, to have been worse than what followed; the golden age may still turn out to have been genuinely better. But the verdict, if reached, will now be reached from the weighing of goods and defects rather than from the stored shape of an ascent or a descent, and this is the whole of the gain. The counterfactual clears away the shape that was pre-empting the weighing. It does not do the weighing. It makes the weighing possible by removing the fatedness that had made it seem unnecessary.
This is why the counterfactual is a corrective and not a verdict. A verdict furnishes a conclusion; a corrective removes an obstruction to enquiry. The fatedness premise is an obstruction: it lets a verdict be installed in the stored shape of the past, exactly as the forbidden question of Paper 1 installs its verdict in the interrogative, so that no enquiry can dislodge it because the enquiry has nowhere to stand that is not already inside the verdict. The counterfactual gives the enquiry somewhere to stand. It re-opens the fork, restores the state in which the outcome is not yet known, and thereby restores the state in which a verdict can be an output of enquiry rather than a property of storage. This is the third demand of Paper 1 — that the researcher specify what would move him and go looking for it — performed against the fatedness that would otherwise ensure nothing could move him, because a fated past admits no evidence that it might have been otherwise.
V. The Tie to the Suite
This project has a counterfactual suite already — the sequence of studies collected under the title Hinges on the Margin — and the present paper is, in one aspect, the statement of what that suite was doing all along. The suite worked case by case, reconstructing the margins at which particular historical sequences might have run otherwise: the hinges on which large outcomes turned, and the narrowness of the turning. It was easy to read the suite as a series of exercises in the counterfactual game — clever reconstructions of roads not taken, spun for the interest of the spinning. This paper’s argument is that the suite was never a game but an instrument, and that what it was instrumentally doing was restoring, case by case, the openness of pasts that both laundered verdicts require to be denied.
Read in the light of Sections III and IV, each case in the suite performs the same office. It takes a sequence that memory has stored as fated — as a necessary ascent or a fall from a secure height — and it stands at the hinge and shows the margin, and in showing the margin it strips the fatedness, and in stripping the fatedness it disables whichever verdict had been living on the sequence’s stored shape. A case that shows a supposed triumph to have hung on a narrow margin disables the progressive reading of that triumph as the destined fruit of an ascent, for a destined fruit does not hang on a margin. A case that shows a supposed golden age to have cohered on conditions that might easily have failed disables the declinist reading of that age as a secure possession lost, for a secure possession is not a lucky configuration of conditions that might have failed. The suite is thus not a collection of alternative histories but a collection of de-launderings, each restoring to a particular past the contingency that its fated storage had drained out, and thereby freeing the verdict on that past to be an output of the weighing of its real goods and defects rather than a property of its stored shape.
And this restoration is exactly the openness that both verdicts deny, which returns the paper to the charge that governs the cluster. The progressive denies the past’s openness from the front, storing it as a necessary ascent to an earned present. The declinist denies it from the back, storing it as a fall from a secure height. Both deny that the past, at the level of the second causes through which it ran, might have gone otherwise — and both must deny it, because the openness, once admitted, dissolves the fatedness their verdicts stand on. The counterfactual, disciplined by evidence about the margins and confined to the level of second causes where the contingency is real, restores the openness, and in restoring it enforces the charge to enquire wisely against the one form of laziness that neither Paper 2’s texture nor Paper 3’s steelmanning reaches: the laziness of reading a verdict off the shape a past acquired merely by having happened. Texture corrects the weighting of goods and defects; steelmanning corrects the furnishing of an opponent; the counterfactual corrects the storage of contingency itself, restoring to the past the openness it had while it was live and losing only the false necessity that hindsight manufactured after the outcome was known.
For the biblicist this last correction has a particular fitness, and the paper ends by naming it. The framework holds that the outcome was certain in the counsel of God and contingent in the acts of men, and it forbids the collapse of the first into the second — forbids, that is, the reading of a fated necessity off the surface of events, since the necessity belongs to a decree that is hidden and the surface belongs to second causes that were genuinely open. The progressive and the declinist both perform exactly the collapse the framework forbids: they read a necessity off the surface, the one calling it an ascent and the other a fall, and both thereby claim to know from the outcome what only the hidden counsel could contain. The counterfactual refuses the collapse. It holds the second causes open, as the Scripture holds them open, and leaves the certainty where the Scripture leaves it, in the counsel of God into which the researcher is not inquiring. It is, in this precise sense, the instrument that keeps the researcher’s memory of the past honest to the doctrine of providence rather than a counterfeit of it — declining to store as fated what was, at the level given him to observe, a walk through a country with many forks, ordained in its end and contingent in its every step, and stored honestly only when both are held at once.
Endnotes
[^1]: The distinction invoked here — between the certainty of the divine decree and the contingency of second causes — is not introduced to this paper as a novelty but drawn from the settled expression of the biblicist tradition on the point, which holds that God from all eternity ordained whatsoever comes to pass, yet so as thereby neither is God the author of sin, nor is violence offered to the will of the creatures, nor is the liberty or contingency of second causes taken away, but rather established. The formula is confessional in its origin and is used here for its analytic clarity, not to import a confessional authority alongside the Scriptural one; the Scriptural ground is given in Endnote 2. The point for the present argument is only that the framework already maintains the two-level structure the counterfactual requires — a certain decree accomplished through genuinely contingent second causes — so that the counterfactual, operating at the level of second causes, neither denies the decree nor competes with it.
[^2]: Genesis 50:20; with Genesis 45:5–8, where Joseph holds together, without any sense of contradiction, that his brothers sold him and that God sent him. Compare Acts 2:23, where the crucifixion is delivered by the determinate counsel and foreknowledge of God and, in the same clause, is the act of wicked hands that are held guilty for it; and Acts 4:27–28, where those gathered against the Lord’s anointed do whatsoever God’s hand and counsel determined before to be done, and are nonetheless the doers of it. These texts are the ground of the distinction stated confessionally in Endnote 1: they maintain, in single sentences, both the certainty of what God ordained and the free and culpable agency of the second causes through which it came about. The proverbial texts on providence over apparent chance — Proverbs 16:33 on the lot, Proverbs 16:9 on a man’s device and the LORD’s direction of his steps — establish the same structure at the level of ordinary events, where the outcome is disposed of the LORD and the deliberation is genuinely the man’s. The counterfactual reasons about the second causes these texts hold contingent, and does not reach toward the decree these texts hold certain.
[^3]: The argument that the progress verdict requires the fatedness premise is developed against the strongest form of progressivism, in keeping with Paper 3’s strictures: not the crude form that merely assumes improvement, but the considered form that grounds the present’s superiority in its being the matured outcome of a developmental sequence. It is this considered form that most needs the fatedness, because it is this form that claims credit for the present as an earned arrival rather than merely asserting the present is better. A progressivism that claimed only that the present happens to be better, without grounding the betterment in a necessary development, would not need the fatedness premise — but it would also forfeit the developmental warrant that distinguishes it from a bare assertion of present superiority, and would be answerable directly to Paper 2’s demand that the comparison be earned from the weighing of goods and defects. Either way the counterfactual does its work: against the considered progressivism it removes the fatedness the developmental warrant requires, and the bare assertion it hands to the texture problem, where an unwarranted comparison cannot survive.
[^4]: That Ecclesiastes 7:10 and 1:9 form a single corrective rather than a balanced pair of opposed cautions is argued in the prolegomenon and assumed in Paper 1; the present paper supplies its methodological vindication. If the gilding verdict and the flattening verdict rested on different premises, they would require different correctives held in tension, and the double corrective would be a balance. Because they rest on the same premise — the fatedness that lets a verdict be read off a stored shape — the double corrective is not a balance but a single instrument applied against a single shared error, and the two verses are revealed as two applications of one charge rather than two charges pulling against each other. This is the sense in which the counterfactual completes the argument of Paper 1: it shows why the charge against gilding and the guard against the conceit of the new are one charge, by exhibiting the one premise they jointly attack.
References
The Holy Bible: King James Version. (1769/2017). Cambridge University Press.
Koselleck, R. (2004). Futures past: On the semantics of historical time (K. Tribe, Trans.). Columbia University Press. (Original work published 1979)
