What the Audit Leaves Open
What the Audit Leaves Open
How much weight the possibilities that survive the audit of classical theism carry, which of them the evidence already reaches, and one attraction declared at zero weight
Companion note · to God Does Not Exist: calibration over the open categories, the survival branch, and the fantastical corner
Version 2 · September 2026
A verdict of nonexistence on one object is not a verdict on everything that object's name has been used to cover. The parent note is explicit about this and holds the open categories open. What it does not say, because its job was the verdict, is how much weight the open categories deserve, whether the evidence already bears on any of them, whether an author who finds some of them attractive has let that attraction into the weights, and how far a given believer stands from the weights once they are set. Those are the four questions here.
1. Open is not weightless, and open is not equal
Two errors sit on either side of the open categories, and they mirror the two the parent note separates in its §2.
The first is assigning zero. A credence of exactly zero cannot be moved by any evidence, because conditioning on anything leaves it at zero. For a proposition that is not logically false, zero is therefore not rigor but a decision never to update, the defect the parent note calls agent-level incorrigibility. The standard name for the rule against it is Cromwell's rule: reserve probabilities of zero and one for logical falsehoods and truths. Every category in the parent note's §15 is logically possible, so each carries a credence above zero.
The second error is treating open as equal. That a possibility cannot be excluded says nothing about its size. The categories differ in how many specific commitments they carry, in whether any evidence already bears on them, and in whether our intuitions about them are trustworthy. Each of those differences moves weight, and the sections below take them in turn.
2. The branch the evidence already reaches: survival of this mind
One family of possibilities is not insulated from evidence, and it is the one most people mean first: that the mind now reading this continues, with its memories, after the brain that runs it stops.
The evidence against it is cumulative and ordinary. The contents of a mind vary with the state of the brain at every scale anyone has tested: anaesthesia switches experience off and on, localized damage removes specific capacities while sparing others, and memory is stored in physical changes that can be disrupted. A thing that can be dimmed by a drug and dismantled capacity by capacity does not behave like a thing that detaches intact when the whole organ stops.
The most direct test of the rival has so far returned nothing that discriminates in its favor. The AWARE II study built audiovisual tests into resuscitation during in-hospital cardiac arrest across 25 sites. Of 567 arrests, 53 patients survived and 28 were interviewed; 11 reported memories or perceptions suggestive of consciousness. Nobody identified the visual image, and one of the 28 identified the auditory stimulus. The study also recorded EEG activity consistent with consciousness emerging as late as 35 to 60 minutes into CPR.
The two results do not weigh the same. The visual image is the test built for the rival: a target placed where it could be seen only from a vantage outside the body, and it returned nothing. Hearing during resuscitation, in a brain intermittently supplied by CPR and showing activity consistent with consciousness, is what the physical account itself expects, so the single auditory identification does not discriminate between the two. The reported experiences are real data about what brains do near death. By the likelihood argument of the parent note's §11, the null on the visual target is evidence against the rival, weak alone and not zero.
The persistence criterion closes the branch from a second side. On What You Actually Are, a person is a particular boundary-maintaining process, and survival is continuity of that work. When the work stops and the configuration disperses, the structure has ended; a later structure carrying its memories, however faithful, is a twin and not a continuation. So even a fantastical rescue that preserved the record would not be this mind surviving unless the labor itself never halted. Evidence against the continuation, plus a criterion on which a preserved record is not the continuation: the branch is closed at the weight the evidence gives it, which is low.
What this does not close is everything else in the class, and the next section is about why.
3. Threshold possibilities: deferred difference, not no difference
The parent note's §5.2 dismisses a hypothesis that makes no possible differential difference: one on which no observation could ever differ depending on whether it is true. That dismissal is correct and is not what is at issue here.
A different shape of possibility makes a difference but locates it behind a threshold. Something beyond current physics exists and matters to an agent, and it becomes observable to that agent only when the agent crosses some threshold. Death is the familiar candidate, the simulation that ends and a player who wakes is one example among many, and the threshold need not be death at all: it could be a capability, a discovery, a stage of development.
Such a possibility is not evidentially idle in the §5.2 sense. It predicts a difference; it predicts it late. Its standing is therefore that of any hypothesis whose test has not been run: current evidence is neutral on it, and credence rests on priors. Two consequences follow.
First, the survival evidence of §2 does not reach the whole threshold class. It counts against branches in which this mind, as currently constituted, continues. It is silent on branches in which the brain is an interface to a process running elsewhere, and on branches in which whatever crosses the threshold is a different system. The criterion sorts these without testing them: a branch in which the constituting labor runs on through the threshold is survival; a branch in which a new structure wakes with the record is twinning, and consoles every system but the one whose loss it was.
Second, not every threshold is sealed. If the threshold is death, no one on this side will see the result. If it is a capability or a discovery, it can be crossed within a lifetime or within the life of a species, and for those members the openness is not merely formal. The right stance toward them is the ordinary one toward any untested claim: no belief now, and a stated observation that would change that.
4. Specific stories cost weight
Every open category can be told as a family of specific stories, and every specific detail lowers the probability of the story that contains it. A simulation, run by players, who inhabit our lives as a game, and who wake at death remembering them: each clause is a further conjunct, and a conjunction is never more probable than its least probable part. This is why any single scenario, however vivid, deserves little weight, while the disjunction of all members of a category can deserve more. Openness is correctly held at the level of the category and not at the level of the story. A reader who asks which story the author believes is asking the wrong question; the answer is none of them, and a small weight on the family they belong to.
5. The fantastical corner is discounted, and why
Within the open space there is a corner with a recognizable shape: possibilities that are not merely more reality but reality that matters to us. A reveal, a continuation, a meaning, an agent that attends. Call it the fantastical corner. The rest of the space, strange structure indifferent to human concerns, is at least as open and far larger.
The corner carries less weight than the rest of the space, and less than it feels as if it carries. There are two reasons, and they do different work.
The first is size. Of all the ways reality could exceed current physics, those that bear specifically on the concerns of one species of primate on one planet are a small subset. Nothing about the open space favors that subset in advance.
The second concerns how large the corner feels, not how large it is. The intuition that whatever lies beyond would be meaningful to us is produced by processes that do not depend on whether it is true. Which processes is contested. One tradition in the cognitive science of religion traces the intuition to perception: minds that bet on the most meaningful reading of ambiguous input and over-detect agents, because for ancestors the opposite error cost more. That account has run into empirical trouble; experiments have given inconclusive support for a dedicated over-detection mechanism, and a leading alternative holds that general beliefs about unseen agents come mainly from surrounding culture, with agency experiences sought afterward to confirm them. The weighting here does not need the dispute settled. On either account the intuition arises whether or not its object exists, and an intuition so generated carries no evidence about that object. So the pull of the corner counts for nothing, and the corner is weighted by its size alone, which is below where it feels.
The inductive record points the same way, weakly. The largest surprises physics has delivered, in quantum mechanics, relativity, and the accelerating expansion of the universe, were stranger than prior imagination and indifferent to human concerns. If there is more, the better bet is that it is strange rather than kind.
6. One attraction, declared at zero weight
The author is drawn to the fantastical corner. The possibility that something exists that fantastical is more interesting to him than its alternatives, and he would find it wonderful to be wrong in that direction.
This is stated as a fact about the author, not as a reason. It carries no evidential weight, for the reason §5 gives: a pull toward a conclusion is generated by the one who feels it, not by the conclusion's object. It is not offered as support for the corner over the indifferent possibilities or over any other part of the open space, and none of the weights in this note were set by it.
It is declared because an undeclared attraction is where motivated weighting hides. Stated, it can be checked: a reader can look at §4 and §5 and ask whether the discounts there bend toward the thing the author wants. They do the opposite. The corner the author finds most attractive is weighted below the indifferent remainder and given nothing for its attraction, and that is the test the declaration exists to make possible.
7. The one gap that is not in the open list
The parent note's §13 records that physicalism has not solved consciousness, and holds that an open problem is no evidence for a supernatural rival. Both points stand. The gap is still the most serious open item in the whole picture, and it is a different kind of item from the categories above: not a further region or agent that might exist, but a missing account of how something known to exist fits the ontology.
If any “something more” carries real weight in this author's picture, it is here, and it is not supernatural. The live options are that experience is physical in a way not yet understood, that the problem as posed rests on a confusion, or that experience is in some sense fundamental. What You Actually Are §8 fences the author's own bet on this question and relies on nothing in it; this note does the same.
8. Grading belief against the weights
Sections 1 to 5 set weights: how much credence each open possibility deserves. A second question is how far a given believer stands from those weights. The parent note's §2 names the measure without giving it a scale: beside the distance of a belief from the truth, there is the distance between the confidence a belief is held with and the confidence the evidence allows. Call the second misalignment. It is a property of a believer's state, not of the object believed in, and it has two components that vary independently: how confidently the belief is held, and how much credence its object warrants.
The second component is where the objects differ. An object that fails on coherence as well as on evidence, as the classical God does on the parent note's audit, warrants credence at the floor. An object that is coherent but unsupported, a finite creator or a god who is not omnipotent, warrants credence that is small but not at the floor: it is one member of the open space, and §4 has already priced its specific versions. An open category held as a category warrants a small credence of its own, larger than any single story inside it.
| State of belief | Confidence held | Credence warranted | Grade |
|---|---|---|---|
| Calibrated openness to the open categories | Small, held as small | Small | Aligned |
| Exactly zero credence in every open category | None | Small | A defect of process, small in distance |
| Confident belief in a specific member of the open space, religious or secular | High | Very low | Severe misalignment |
| Confident belief in a coherent but unsupported god | High | Small, not at the floor | Severe misalignment |
| Confident belief in the classical God | High | At the floor | Near-maximal misalignment |
Aligned. Openness to the open categories, held as a small credence and stated as uncertainty, is where this note's own weights place a careful reasoner. It is not theism: it names no god and asserts none. A person who says there may be something more, attaches no doctrine to it, and holds it with uncertainty is doing what the corpus asks of any claim, and the author regards that as reasoning well, whatever label the person uses for it.
Zero. The opposite error is small in distance but real. A credence of exactly zero in a logically possible category sits close to the correct small value, yet it can never be moved, which is the incorrigibility the parent note's §2 counts as a defect. It is graded here as a failure of process rather than of distance, and it is the error §1 exists to rule out.
Severe. Confident belief in a coherent but unsupported god fails on evidence alone. The confidence is high and the warrant small, so the gap is wide even though the object survives the coherence audit. Where such a god is also said to answer prayer, reveal, or guide events, the nulls of the parent note's §11 count against it directly, and the belief sits at the upper end of the grade. The grade is not reserved for religion: confident belief in any specific member of the open space, a simulation run by particular designers or a particular reveal at death, carries the same conjunction cost from §4 and earns the same grade. The scale grades the shape of a belief, not its subject.
Near-maximal. Confident belief in the classical God keeps the parent note's verdict unchanged. It is the one state in which both components reach their extremes at once, high confidence in an object that fails on coherence and on evidence, and near-maximal stays reserved for that combination. Because the classical package is the mainstream of theism as it is actually believed, this is the grade that applies to most theists as they actually hold their belief; severe applies to the minority whose god is finite or otherwise non-classical.
Two limits keep the scale exactly the size it is. It grades states of belief, not persons: a severely or near-maximally misaligned belief can sit in a careful mind, and general religious belief is acquired mainly from surrounding culture rather than reasoned into, so the grade says how far a belief stands from the evidence, not how well its holder reasons elsewhere. And it grades against the weights of this note, so it moves when they move: evidence that raised a category's warranted credence, of the kinds listed in §11, would lower the grade of confident belief in it by the same amount.
9. Ledger
| Claim | Standing | What carries it |
|---|---|---|
| The classical supernatural God does not exist | Held as published | The parent note, unamended |
| Every category in the parent note's §15 carries a credence above zero | High confidence | Cromwell's rule; each category is logically possible |
| This mind continuing after its brain stops | Low | Brain-dependence of mental contents; the AWARE II null on the visual target; the persistence criterion |
| Threshold possibilities other than naive survival | Small, set by priors | Current evidence neutral; deferred difference, not no difference |
| Any single specific scenario within the open space | Very low | Conjunction cost |
| Calibrated openness to the open categories, held as uncertainty | Aligned | Credence matched to warrant (§8) |
| Confident belief in a coherent but unsupported god, or in any specific member of the open space | Severe misalignment | High confidence against small warrant; the parent note's §11 nulls where the god is interventionist (§8) |
| Confident belief in the classical God | Near-maximal misalignment | High confidence against floor-level warrant; the parent note's verdict, unchanged (§8) |
| The fantastical corner | Below the indifferent remainder, and below its felt weight | Size of the subset; the felt pull generated independently of its object (§5, load-bearing) |
| The author's attraction to the fantastical corner | A fact about the author; zero evidential weight | §6 |
| How experience fits the physical ontology | Open; no basis to choose among the options | §7 |
10. Where this stands to existing positions
Nothing here is new as a position, and the nearest neighbors are worth placing exactly.
Agnosticism in its original sense suspends judgment on questions judged unanswerable. This note does not suspend: it assigns weights, low ones, and it assigns the parent note's verdict on the classical God in full.
Ietsism, from the Dutch for “something-ism”, is the stance that there is, or must be, something more, left unspecified. The stance here shares its openness and denies both the “is” and the “must”.
The closest academic neighbor is J. L. Schellenberg's ultimism: the proposition that there is a metaphysically and axiologically ultimate reality, in relation to which an ultimate good can be attained. Schellenberg argues that in our present state neither belief nor disbelief in ultimism is justified, and recommends a nonbelieving, imaginative faith toward it. The comparison is exact on one point. Ultimism is, by its definition, the fantastical corner: its axiological clause is what makes it matter to us. Schellenberg responds to shared doubt by directing faith at that corner. This note responds to the same doubt by weighting that corner lowest and directing nothing at it. The divergence is not about the evidence, on which the two agree; it is about whether an attraction may be turned into an attitude. Here it may not.
Two further neighbors sit inside the open space rather than beside it. Bostrom's simulation argument is one structured member of the threshold class. Hick's replica theory, on which death is followed by the appearance of an exact replica of the person in another, resurrection universe, is on this corpus's criterion a case of twinning, not survival, which is why §2 treats a preserved record as no rescue.
11. Kill conditions
The weights move on evidence, in both directions. The following would raise them:
- A replicated result, under protocols at least as strict as AWARE II's, in which a patient identifies a visual target placed to be seen only from outside the body, at a time when the brain shows no activity consistent with consciousness. This would reopen the survival branch of §2.
- Information arriving through a channel physics treats as closed, reproducibly and under adversarial checking.
- Evidence that a threshold of the non-death kind has been crossed and something beyond current physics observed on the far side.
- A showing that the felt sense that reality is meaningful to us tracks its object, correlating with independent evidence that it is. This would make the pull of §5 count as evidence and raise the fantastical corner's weight.
What would lower the weights further is more of what has already been found: well-powered tests of specific claims returning nulls, and continued success of physical explanation in domains once assigned to something more.
Standing of this document. A companion note. It amends nothing in God Does Not Exist and assigns weights to the categories that note leaves open. Its parts do not share a status. That open categories carry nonzero credence is near-analytic given Cromwell's rule. The closure of the survival branch is empirical and cumulative, and inherits the persistence criterion of What You Actually Are at the confidence that note gives it; the value-indexing posit marked there is not used here. The treatment of the fantastical corner rests on the independence claim marked in §5, which holds under both live accounts of where the pull comes from and fails only if the pull is shown to track its object. The author's attraction in §6 is a declaration, not a claim, and carries no weight by construction. The consciousness question in §7 is left open and fenced. The misalignment scale of §8 is definitional in its measure, the gap between confidence held and credence warranted, and inherits its placements from the weights above, so it moves when they move. On any conflict with the corpus, the corpus governs. Corrections and counterexamples are welcome and change the note.
Version history. Version 1, 24 September 2026: initial publication. Version 2, September 2026: additive; the original text is preserved. A new §8 grades states of belief against the note's weights, measuring misalignment as the gap between confidence held and credence warranted: calibrated openness is aligned, exact zero is a defect of process, confident belief in a coherent but unsupported god or in any specific member of the open space is severe misalignment, and near-maximal stays reserved for confident belief in the classical God. The ledger adds the matching rows, the introduction and scope note name the fourth question, and the later sections renumber to §9–§11.
References
- God Does Not Exist (September 2026). The parent note: the verdict on the classical God, the supernatural trilemma of §5, the likelihood argument of §11, the physicalism comparison of §13, and the open categories of §15. Internal
- What You Actually Are (June 2026, Version 2 July 2026). The persistence criterion, continuity of the work; copies as twins; death as dissolution; the fenced consciousness bet of §8. Internal
- Epistemic Forge: The Apex Synthesis. The physical floor and evidence discipline, inherited through the parent note. Internal
- Parnia, S., Keshavarz Shirazi, T., Patel, J., et al. (2023). “AWAreness during REsuscitation – II: A multi-center study of consciousness and awareness in cardiac arrest.” Resuscitation 191, 109903. 567 in-hospital arrests, 53 survivors, 28 interviewed, 11 reporting memories or perceptions suggestive of consciousness; nobody identified the visual image, one identified the auditory stimulus; EEG activity consistent with consciousness up to 35–60 minutes into CPR. Checked against the published abstract. Verified · September 2026
- Schellenberg, J. L. (2005). Prolegomena to a Philosophy of Religion. Ithaca: Cornell University Press. Propositional faith without belief. Verified · September 2026
- Schellenberg, J. L. (2009). The Will to Imagine: A Justification of Skeptical Religion. Ithaca: Cornell University Press. Ultimism and skeptical faith toward it; the nearest neighbor of §10. Verified · September 2026
- Guthrie, S. E. (1993). Faces in the Clouds: A New Theory of Religion. New York: Oxford University Press. Religion as systematic anthropomorphism, arising from a perceptual strategy of betting on the most meaningful reading of ambiguous input; the perceptual account named in §5. Verified · September 2026
- Barrett, J. L. (2000). “Exploring the natural foundations of religion.” Trends in Cognitive Sciences 4(1), 29–34. The cognitive approach in which agency detection was theorized to underlie experiences of supernatural presence. Verified · September 2026
- Van Leeuwen, N., & van Elk, M. (2019). “Seeking the supernatural: the Interactive Religious Experience Model.” Religion, Brain & Behavior 9(3), 221–251. Empirical and conceptual difficulties for the over-detection account; general religious beliefs acquired mainly from culture, agency intuitions sought to confirm them. The reason §5 does not rest on any one mechanism, and the cultural-acquisition point in the limits of §8. Verified · September 2026
- Lindley, D. V. (1985). Making Decisions (2nd ed.). London: Wiley. The work commonly credited with naming Cromwell's rule: no prior of 0 or 1 except for logical truths and falsehoods. Edition verified; the passage not checked against the text. Verified · September 2026
- Bostrom, N. (2003). “Are You Living in a Computer Simulation?” Philosophical Quarterly 53(211), 243–255. One structured member of the threshold class. Verified · September 2026
- Hick, J. (1976). Death and Eternal Life. London: Collins. The replica theory, chiefly chapters 15, 20 and 22: at death an exact replica appears in a resurrection universe. Read in §10 as twinning on this corpus's criterion. Verified · September 2026
Comments
Post a Comment