A familiar question. A search for what stays true.
Late 2025 / My starting question
The spacetime and branching pictures mattered to me because the theories behind them agree so well with experiment. Whatever deeper understanding I reached, it had to preserve what they already got right.
Einstein’s relativity brings space and time into one description, with no single “now” shared by the whole universe. In the block picture, the events of my life belong to that spacetime history.
Quantum mechanics presented a different picture. Schrödinger’s equation contains no collapse mechanism. When that evolution is applied to the system, apparatus and observer together, the quantum state retains the different outcome records. Many Worlds takes those branches as physically real, each containing an observer with a definite experience.
I could not reconcile a single complete history with this multiplicity of realized histories as complete accounts of the same reality. Both grew out of mathematics I had strong experimental reasons to take seriously. Yet I still had to account for the life I was experiencing.
That was what drove me to search for something deeper and more fundamental. What could both descriptions depend on? What would have to remain true through either account of the universe?
I started with the one thing I could examine from the inside: my own experience—“me.”
01 / The question that started it
Where am I in this picture?
You remember arriving here. You are reading these words. Later, you may remember this moment. From the inside, this is your experience.
Start with that familiar experience. Then explore how the spacetime and branching pictures each represent it.
Start from the inside
A moment, with a past.
Reading now includes a memory of arriving here. A later recollection could carry both moments forward.
Opening the pageReading nowRemembering later
A familiar sequence. The later moment is imagined.
Inspect a moment and what it remembers.
Earlier · opening the page
“I have just arrived here.”
This moment comes before reading this passage. Its memory does not yet include these words.
Now · reading this passage
“I remember opening this page.”
The earlier event is part of what this present experience remembers.
Later · an imagined recollection
“I remember reading that question.”
If this memory forms, it will connect a later experience to the reading happening now.
The spacetime / block view
The moments are in the picture together.
The same sequence can be drawn as events along a worldline. Each event has its own local records and memories.
↑ Time along the drawingSpace →
Earlier · lowerReading · middleLater · upper
A schematic history, with one space direction shown. Highlighting an event means inspecting it.
Open the record at one event.
Earlier event · opening the page
“I have just arrived here.”
The later events remain in the drawing. They are not memories held at this earlier event.
Reading event · with an earlier memory
“I remember opening this page.”
The memory belongs to this event. The drawing represents its place in the history.
Later event · an imagined recollection
“I remember reading that question.”
Here the represented memory reaches back to the reading event. The order of the events stays the same.
The Everett / branching view
Both records remember arriving here.
Now imagine this page displays a quantum measurement result. An Everett description contains outcome records A and B, each with the same earlier memory.
A · upper recordRemembers opening the pageB · lower recordRemembers opening the page
Left: shared earlier memory. Centre: quantum measurement. Right: different outcome records.
Read either record. Both stay in the picture.
Record A · result A was observed
“I opened the page, read the question, and saw A.”
This observer record retains the earlier history and a particular result. Record B is also represented.
Record B · result B was observed
“I opened the page, read the question, and saw B.”
This observer record also retains the earlier history, with a different result. Record A is also represented.
What could both descriptions depend on? I began looking for what different descriptions must preserve.
What the two pictures mean
The spacetime drawing represents an ordered history. Relativity supplies no preferred universal present sweeping through it. Our highlight is a reading aid, not a moving cosmic “now.” The block interpretation is a philosophical reading of spacetime; the diagram does not establish that interpretation or refute it. Einstein Online: spacetime and relative simultaneity.
The branching drawing is a schematic measurement interaction in an Everett-style description: the quantum state includes correlated observer and outcome records. It is not branching caused by an ordinary human decision. Each represented observer has a definite record; first-person experience by itself is not a refutation of Everett. A request for one uniquely selected continuation is an additional question whose justification must be made explicit. Hugh Everett III, “Relative State” Formulation of Quantum Mechanics (1957).
The conflict I recall concerns treating one classical history, or multiple realized histories, as the complete account of reality. A block conception of time does not by itself exclude Everettian branching. Wallace and Timpson: quantum mechanics on spacetime.
These pictures introduce the historical question that motivated AASC. They are not a proof of its later results. The interactive highlight changes which record you inspect; it neither changes that record nor identifies a physically preferred observer.
02 / The same events, different coordinates
The numbers change. What stays?
Relativity already offered a concrete example of what different descriptions preserve: observers can disagree about measurements while agreeing on an invariant. That gave me a place to begin.
Imagine holding a watch while moving steadily. You start it, then read it four seconds later. Someone moving past you assigns different time and position coordinates to those readings. They are still the same two events on the same watch.
A
Start the watch
Reading: 0 seconds.
B
Read the watch again
Reading: 4 seconds.
Describe those events from another steadily moving frame.
70% light speed left70% right
With the watch, both readings happen in one place.
In the watch's own frame, the second reading is 4.00 seconds later at the same position. A different frame changes the coordinates, not which readings we are describing.
Frame's time gap
4.00
seconds
Frame's position gap
0.00
light-seconds
Watch's elapsed time
4.00
seconds · unchanged
Spacetime interval
16.00
light-seconds² · unchanged
Keep the events. Change the grid.
The dots are the watch readings A and B. The colored grid belongs to the selected frame. It assigns new coordinates to the same events.
Time axis, ct′
Position axis, x′
The watch between readings
A light-second is a unit of distance: how far light travels in one second. The diagram uses time × the speed of light on its time axis. Changing the selected frame does not accelerate the watch or replay its history.
For a frame moving right at 60% of light speed, the time gap is 5.00 seconds and the position gap is −3.00 light-seconds. The watch still records 4.00 seconds, and the spacetime interval is still 16.00 light-seconds². These are two coordinate descriptions of the same pair of events.
The unchanged interval is a real example of an invariant. It still uses a metric: a rule that combines the time and position coordinates.
What must be settled before we use that description?
The transformation, the units, and what this example shows
This is special relativity in flat spacetime with one spatial coordinate. The watch follows an inertial worldline. In its rest frame, A = (ct, x) = (0, 0) and B = (4, 0), with both coordinates in light-seconds. The chosen frame moves at v = βc relative to the watch; the two frame origins coincide at A.
Using the time-positive sign convention, the squared interval is (Δct′)² − (Δx′)² = 16 light-seconds². Its square root divided by c is the proper elapsed time, 4 seconds, along this inertial watch's path. The drawing holds the rest-frame event positions fixed and maps the moving frame's coordinate grid back onto that drawing with the inverse transformation.
4.00² − 0.00² = 16.00 light-seconds².
This exhibits a Lorentz invariant inside an already specified spacetime description. It does not establish that the Lorentz metric is the deepest possible invariant or derive the foundations from this example.
I had reached Lorentz invariance, but I was still searching for something more fundamental. What did even a metric description depend on?
I deliberately used the hypnopompic state—the transition from sleep to wakefulness—to explore those questions.
Before going to sleep, I would build mental models of the universe and identify the points I could not resolve. On waking, I would return to those questions in the in-between state, when I found my thinking most creative. I made this a working practice.
Over the space of a couple of weeks, I developed the core of AASC through this same process: the AMetric boundary, fixation, bivalence and uniqueness.
We can begin exploring those ideas with the AMetric boundary: when the readings change, what keeps them connected to the same thing?
My recollection of how AASC began in late 2025.
The hypnopompic state: sleep, waking and creative insight
Hypnopompic describes the transition from sleep to wakefulness; hypnagogic describes the transition into sleep. My practice deliberately focused on the waking transition.
Dalí documented his “slumber with a key” in Fifty Secrets of Magic Craftsmanship (1948), an example of deliberately interrupting sleep onset. Poincaré described ideas coming to him in bed in the morning or evening, and emphasized the conscious preparation and verification surrounding an insight.
In a 2021 experiment, participants who briefly entered light sleep were more likely to discover a hidden mathematical rule than those who remained awake. That study concerns sleep onset; it does not directly test the practice after waking described here.
Return to the watch experiment. Different observers assigned different numbers to the same two events. But the events stayed identified: you started your watch, then read it.
The coordinates changed from frame to frame while we kept track of the same events. Put the numbers aside for a moment. That connection is still doing essential work.
Two readings, attached to two events.
In the watch’s rest frame: it starts at 0 seconds, and you read it at 4 seconds. Both events are at the same position in that frame.
A
You start your watch.
The beginning of this interval.
Time 0 sPosition 0
Still the same start event.
B
You read your watch.
The later event on that same watch.
Time 4 sPosition 0
Still the same later event.
The display changes. The events, their order and their connection to the watch stay the same. Position is in light-seconds; the watch faces are schematic.
The step toward the AMetric boundary
A description needs a definite thing to be a description of.
A time, a distance or a label cannot settle its own connection to that thing merely by being written down. We have to establish what the description concerns and why its use is justified.
This is the question the AMetric boundary makes precise. “Before measurement” means a requirement measurement depends on. It does not name an earlier moment or a hidden place in the universe.
What this shows—and what the formal paper establishes
The display keeps the event identities and order supplied by the example. Hiding coordinates does not remove the physical metric or construct an equality-only input. It helps expose the distinction between the target, a representation of it, and the grounds that connect them.
The revised AMetric paper asks which distinctions and answers are supported by the complete authoritative input. A proposed choice must be eligible and remain fixed under every lawful transformation that leaves its supplied input unchanged (Theorem 4.6). Full arbitrary relabeling requires the equality-only premise; absence of a preferred individual alone does not erase other structure (§3.3 and §6).
The fibre test and constructed Axiom 6 concern what a description can establish and the order of qualification (Theorems 7.2 and 7.5). A warranted new record, test or construction can supply additional information. The boundary is not a physical transmission channel, and absence of a metric alone does not establish non-action.
I began by asking how my own life fits into physics. That grew into a question about every definite person, object and event: what must a description preserve to remain an account of that same thing?
The Kernel gives four connected roles precise names. Their work is already present in the questions we have been asking.
Reference
What are we talking about?
This person. This experience. These particular events. A new description must keep its connection to its subject.
Standing
Why does this account count?
Writing a possible description is a beginning. It must meet the requirements for the particular claim it makes.
Irreversibility
Can a later story change what happened?
New evidence can correct our account of an event. A later event does not rewrite the original one.
Admissibility
Does the account hold together?
It must satisfy the relevant conditions together and preserve them through the uses it claims to support.
The person and the events do not acquire these requirements because someone notices them. In the Kernel argument, they are necessary for anything definite, including what goes unobserved.
How the personal starting point becomes a general result
The origin story motivates the investigation. The Kernel paper's derivation proceeds from determinate targets, assessment and use, historical fixation, and answer accountability (Definitions 2.1–2.3, Proposition 2.4 and Theorem 3.1). Consciousness is not introduced as an extra physical force or as a premise needed to govern other objects.
Its conservation results concern independently warranted answers to the same complete determinate question about the same bearer (Theorems 6.2–6.3). Agreement follows under those conditions; agreement alone does not establish that two accounts concern the same thing. Evidence about an unchanged occurrence is distinguished from a new occurrence or a changed construction (Example 3.2 and Lemma 3.3).
Why a complete account needs more than one yes-or-no answer
An account can answer one question correctly while leaving out something needed for another. The revised Structure paper therefore studies the full information required for the intended questions and uses, including whether a requested operation is defined. One qualification answer is binary; the complete account need not reduce to that single bit (Theorems 5.5–5.8).
It then constructs the greatest set that continues to meet the independently specified requirements throughout the permitted uses. The term admissible interior names that complete set. The requirements and permitted uses are specified before the set is found (Theorem 6.3 and §8).
Does a unique complete answer mean only one possibility remains?
Not automatically. A complete answer includes every possibility that meets all the specified requirements. There can be one uniquely complete set with several members—or no members. Omitting a qualifying possibility and including an unqualified one are different ways to give an incomplete or incorrect answer.
Unus Solus Possibilis Est constructs the greatest sound interior from independent qualification, warrants and permitted continuations (§§2–3). A unique individual occupant requires additional conditions that separate identities, together with a positive realization (Theorem 10.2). That further result is not supplied merely by calling the complete set unique.
What does it mean for a theory to describe the world we live in?
The foundation papers develop the requirements precisely. Next, follow how familiar physics takes shape from that common starting point, then explore the tours on time, measurement, matter and experience.
These passages are my recollection of how AASC began in late 2025. The formal notes follow the September 2026 revisions of the four foundation manuscripts. The existing paper records and all-versions DOIs are linked below.
Establishes the necessary roles of admissibility, standing, reference and irreversibility. Independently warranted answers to the same complete question about the same object agree, while later repair leaves the original trial intact.
Kernel governance is necessary for non-degenerate determinate objects and incidences. Conservation compares independently warranted answers about the same bearer and complete determinate question. Later evidence can correct an account; later repair cannot replace the original occurrence or conditions. A complete record also retains the substantive information required by the task.
Shows why labels cannot recover a lost connection to the original target, and how a grounded record can supply the missing information. Distinguishes what a description tells us from which individual it can identify.
The AMetric boundary is an order-of-qualification constraint. Equality-only symmetry permits equality information but supplies no privileged individual. A grounded selection needs the required input association and warrant. Complete interfaces preserve all declared uses and operation domains; these authority conditions do not describe a physical transmission mechanism.
Two configurations can both pass one test and still behave differently on the next. Constructs the complete information needed for the declared uses and the greatest set that remains qualified throughout them.
Qualification is binary for one complete question; the full record can preserve many distinctions. The safe interior is constructed from independently fixed qualification, permitted uses and their warrants. It can contain several configurations or be empty. Safety does not by itself establish eventual completion.
Builds the complete safe set from independently specified requirements and permitted continuations. Every sound candidate set lies inside this greatest interior, and faithful descriptions preserve it even when several configurations qualify.
The greatest sound interior is constructed before its membership supplies robust standing. Exact descriptions preserve local qualification and the required continuation information. One complete interior can contain many occupants; one unique occupant needs additional identity-separating constraints and a positive realization.
Separates the conditions a description depends on, the meaningful content it preserves, and its presentation. These are called Anchor, Tensor and Skin.
Unique fixed-domain role decomposition, importing intrinsic detectability, bivalence, fail-closure and structural exhaustion. Tensor does not necessarily mean a numerical array; anchor is not a physical mechanism.
Asks where any genuinely different contribution could enter the same construction. Reduces relevant variation to its effect on the boundary, classification, identity or continuation.
Theoremic, non-enumerative fixed-domain exhaustion via future-behavior quotient and boundary factorization. Does not exhaust arbitrary richer domains or derive physical dynamics by checking examples.
Even rejecting every candidate requires a stable criterion and a way to identify what was evaluated. Develops the structure forced by meaningful evaluation without inferring that an admitted object exists.
Governance non-vacuity from non-degenerate reuse-relevant evaluability. An all-rejecting finite model explicitly separates the required bundle from nonempty admission or concrete ontology.
When otherwise matching descriptions cannot pick out a unique target, a declared, stable connection can fix the role without selecting the most appealing outcome.
Singleton bivalent outcome-independent trace within the declared domain. Missing, multiple or unstable traces fail closed. Role fixation is not empirical confirmation and is not transmission across the AMetric boundary.
Carries the same-target requirement into differentiation: account for changes in the frame of comparison rather than confusing them with changes in the object. Develops transport, covariant differentiation and metric-compatible closure.
Necessary transport-comparison role; covariant derivative in the smooth finite-linear first-jet branch. Levi-Civita follows in the stated minimal natural metric branch. No numerical connection, gauge group or complete physical model is selected by the Kernel alone.
Pre-admissible classification is not an execution process. Operators and composition require licensed interior standing; no boundary operation or same-lineage repair.
A corollary under the already fixed AMetric architecture. A proposed deeper invariant imports forbidden structure, does no governance work, or renames existing admissibility and standing.
