# ITHKOR Experiment Atlas v6

This atlas is a navigation layer over the documented ITHKOR experiment families.
It is not a new result and does not override the source documents.

| branch / family | status | tested question | positive signal | boundary / negative result | controls | QPU status | allowed claim | blocked claim | next fair step |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Kernel/C | scoped finite support | Can compression/foliation harnesses distinguish schedule-legal finite dynamics? | schedule-independent selector behavior in tested graph families | physical equivalence not shown | legal schedules, unsafe schedules, no-signalling | some branches QPU-facing later | finite selector diagnostics | Lorentz/spacetime | replication package |
| C3/C3b | operational finite support | Are legal foliations stable under selected compression triggers? | fixed/canonical/probe selectors stabilize active regimes | event-local trigger path-dependence | fixed event/support, matched noise, channel algebra | local/sim mostly | finite foliation sensitivity harness | physical covariance | keep as harness |
| D0-D1 | finite metric/response support | Can pressure define a finite information metric and response? | D0 information metric, D1 response correlation | not physical metric | shuffled/degree/sign controls | local | finite metric seed | spacetime metric | keep bounded |
| D2-D2b | min readiness, strong fail | Does D0/D1 produce transport/front behavior beyond topology? | local front readiness | D0 close to hop/topology, strong transport fails | hop, static pressure, shuffled | no strong QPU | finite front-readiness | physical transport | treat as boundary |
| D5-D7 | finite QPU-facing readouts | Can selected finite witnesses be read on QPU? | multiple frozen readout supports | hardware readout only | layout, shots, backend repeats | supported in scoped cases | QPU-facing finite witness | physical proof | replicate only |
| D8 | source/metric deformation support | Does source pressure deform finite metric/field scores? | source-law finite support | no gravity | shuffled, inverse, no-source | QPU robustness in scoped readout | finite source-law diagnostic | gravity | keep claim-safe |
| D9 | motion bridge fail | Does a probe/geodesic flow follow the metric beyond controls? | designed case only | deconfounded probe-flow fails | hop, degree, shuffled, inverse | blocked | negative result | free fall | closed boundary |
| D10-D11 | field/readout/coupling support | Can source-response field be read and sensed by probe perturbations? | D10/D11 finite/QPU readout support | not motion, not gravity | hop, degree, shuffled, no-field, inverse | supported in frozen readouts | finite field/coupling readout | gravity | package |
| D12 | causal-envelope support | Is there a finite response envelope? | D12/D12b local/QPU smoke support | no physical speed | shuffled layer, global response, no-field | QPU smoke | finite envelope | speed of light/Lorentz | bounded reruns |
| D13 | scoped falloff | Is radial/source falloff visible? | D13/D13c scoped local support | D13b robustness boundaries | hop, degree, shuffled, source-shift | QPU blocked | scoped falloff diagnostic | Newtonian gravity | no escalation |
| D14 | probe-class fail | Do different probe classes agree after normalization? | none broad | D14/D14c fail | class controls | blocked | negative result | equivalence principle | closed/blocked |
| D15-D16 | coarse/scale scoped support | Do fields survive finite coarse-graining and scale stability? | scoped passes and boundaries | not continuum | bad partitions, shuffled, degree | blocked | finite coarse/scale scope | continuum | package |
| D17 | cost/energy fail | Does source budget behave like a specific cost bridge? | amplitude-like hints | no reach/falloff/accounting bridge | degree/hop/random cost | blocked | negative cost boundary | energy/stress-energy | locked unless new cost model |
| D18 | parameter basin | Is support a point or an island? | finite viability islands | radial/falloff boundary | graph/source/falloff controls | no QPU | finite island map input | physical constants | taxonomy only |
| E1 | shared constraint | Can finite readouts show shared correlations without signalling? | local/Aer and QPU robustness support | no physical entanglement proof | product/shared-random/no-signalling | supported | finite shared-information readout | foundational entanglement proof | frozen package |
| E2 | contextuality | Can contextual readout beat hidden-table controls? | local/Aer and QPU support | no foundational proof | hidden-table/product controls | supported | finite contextual readout | proof of QM foundations | frozen package |
| E3 | no-cloning/copy limit | Can finite harness show copy-limit behavior? | local plus frozen QPU repeat support | initial near-miss ledger remains | classical-copy, clone fidelity controls | supported with ledger | finite copy-limit readout | theorem proof beyond QM | ledger |
| E4 | probability calibration | Does readout probability calibrate locally? | local support | QPU not confirmed | angle/shot/linear controls | not confirmed | local calibration boundary | Born-rule derivation | close as boundary |
| E5 | record boundary | Does accessible record destroy/restore conditional patterns? | local/Aer support | no hardware claim | noise, eraser, phase random controls | not run | finite record boundary | measurement proof | package |
| F1-F3 | macro record | Are records stable, irreversible, and observer-consistent? | finite macro-record layer | no physical measurement theory | repeated readout/noise/coarse observers | no QPU | finite macro-record diagnostics | classical world proof | package |
| G1-G1E | record geometry | Can public records define robust geometry? | scoped degree-balanced candidate | held-out generalization mixed; degree/coarse boundaries | same-degree shuffle, coarse partitions | blocked | scoped record-geometry candidate | spacetime | status ledger |
| COND1-COND3 | coherent order | Can finite Bose-Hubbard toy show coherent-order island? | coherent island and phase boundary | high-interaction boundary | disorder/localized/thermal/twist controls | blocked | finite coherent-order toy | real BEC/He II | informatics metric |
| GR1-GR5 | modular response | Does finite modular first-law-like response hold? | GR1-GR3 real-sector support | local-Y phase/quadrature boundary | X/Z/XX/Y, complex references | blocked | scoped modular-response pillar | gravity/Einstein equations | closure |
| COARSE1 | coarse hygiene | Does modular response survive finite coarse maps? | core regions pass | block coarse boundary | block2/block3, region modes | blocked | finite region hygiene | continuum | boundary taxonomy only |
| TIME1-TIME2C | internal clock | Does internal entropy/dark signal order events? | scoped dark/entropy signal | coherence-loss boundary | shuffled entropy, no-exchange | blocked | scoped finite internal-clock signal | physical time/Lorentz | closure |
| MOD1 | modular spectrum | Does spectrum taxonomy separate support/boundary? | none | frozen fail | spectrum/offdiag controls | blocked | negative result | spectrum bridge | closed |
| ISLANDS1 | meta-audit | Do branch results form finite islands and repeated boundaries? | 10 islands, 9 repeated boundary types | no physical claim | source-doc consistency | no QPU | scoped island map | General ITHKOR confirmation | reviewer package |

## How To Read The Atlas

The atlas should be read from left to right:

1. identify the branch;
2. check whether it is support, mixed, fail, or closure;
3. read the boundary column before the positive signal column is used in any
   public statement;
4. check QPU status only as finite readout status;
5. keep the blocked claim attached to every result.

## Atlas-Level Conclusion

The experiment program is not a single linear proof. It is a map of finite
supports and finite stop signals. Its current strength is the combination of
positive islands and explicit boundaries.