SOURCE AND OVERLAP LEDGER -- accessed/reconciled 9 November 2026

B1. Brian Theory, From Observable States to Acquirable Predictive Models,
technical report revised 7 October 2026, site reading copy 1.
https://www.b-theory.com/manuscripts-html/acquirable-predictive-models/
Actual text inspected: Example 8, Proposition 10, Example 11, Sections 4/7/8,
Appendix B. Precedent: acquisition differs from predictive sufficiency and
actions change available information. The finite, RF-constrained alternating
contract is not imported here. The scalar obstruction is proved separately.

B2. Diagnosing While Repairing: Information Loss, Adaptive Order, and Noisy
Readouts, AI-assisted draft prepared for Brian Theory, v0.2.5, 8 November 2026.
Unpublished companion; public persistent location not yet assigned.
Actual scope/adaptive-order text compared against v0.2.4, 7 November 2026.
The adaptive-order chapter is byte-identical; the later scope chapter adds
a warning that observational equivalence need not preserve terminal health.
This does not change the overlap comparison. Its adaptive repair-order
separation is a restricted adaptive-testing realization. This addendum instead
fixes controls in both comparators and changes observation timing, with
continuous disturbances and mandatory next-step forecasts. No unpublished
result from B2 is required to prove any theorem here.

B3. Causal Self-Diagnosis with Noisy Tests: Identifiability, Admissibility, and
Decision Value, archived v0.1 source, 3 November 2026. Later versions exist.
Actual foundations/cumulative sections inspected by the priority reviewer;
not presented as latest research. Transcript indistinguishability and common
random-tape arguments are already antecedents. Fixed physical state plus
support-sensitive test constraints differs from continuously evolving state
with tests always permitted. This source is contextual, not a proof dependency.

L1. Scott, Findeisen, Braatz and Raimondo. Input design for guaranteed fault
diagnosis using zonotopes. Automatica 50 (2014), 1580-1589.
https://doi.org/10.1016/j.automatica.2014.03.016
https://web.mit.edu/braatzgroup/input_design_for_guaranteed_fault_diagnosis_using_zonotopes.pdf
Author PDF inspected: Sections 1-2, Theorem 3/equation (17), and Section 5.
Bounded-error output-set separation is antecedent. It designs inputs across
fault scenarios; it does not establish this fixed-input threshold budget.

L2. Hierons, Jourdan, Ural and Yenigun. Checking Sequence Construction Using
Adaptive and Preset Distinguishing Sequences. SEFM 2009, 157-166.
https://doi.org/10.1109/SEFM.2009.12
https://bura.brunel.ac.uk/bitstream/2438/4120/1/Fulltext.pdf
Institutional PDF inspected: introduction and PDS/ADS definitions, pp.157-158.
State-changing distinguishing experiments are established. Initial-time batch
acquisition here is not the literature's preset input/output experiment word.

L3. Petra van den Bos and Frits Vaandrager. State Identification for Labeled
Transition Systems with Inputs and Outputs. Author preprint, 2019.
https://arxiv.org/abs/1907.11034
https://arxiv.org/pdf/1907.11034
Inspected Definition 20/Figure 3, Section 5 and Theorem 35. Adaptive
distinguishing graphs and state-merging obstructions are antecedents.
Theorem 35's compatibility premise is not imported into this continuous model.

L4. Hu and Tucker. Optimal Computer Search Trees and Variable-Length
Alphabetical Codes. SIAM J. Appl. Math. 21(4) (1971), 514-532.
https://doi.org/10.1137/0121057
Publisher abstract and metadata only; full text not inspected. Historical
context for alphabetic trees, NOT a source proof for the maximum-depth lemma.
That elementary lemma is proved directly in the manuscript.

L5. Pola, Girard and Tabuada. Approximately bisimilar symbolic models for
nonlinear control systems. Automatica 44(10) (2008), 2508-2516.
https://arxiv.org/pdf/0706.0246
Introduction and Theorem 5.1 inspected. Incremental-stability/quantization
conditions give an abstraction framework. No new such framework is claimed.

L6. Reissig, Weber and Rungger. Feedback Refinement Relations for the Synthesis
of Symbolic Controllers. IEEE TAC 62(4) (2017), 1781-1796.
https://doi.org/10.1109/TAC.2016.2593947
https://arxiv.org/pdf/1503.03715
Introduction, Proposition VII.1 and Section VIII growth bounds inspected.
Sound set-valued abstraction does not make abstract infeasibility exact.
The present proofs use exact intervals, not an approximate grid.

L7. Bian and Abate. On the Relationship between Bisimulation and Trace
Equivalence in an Approximate Probabilistic Context. FoSSaCS 2017, 321-337.
https://arxiv.org/pdf/1701.04547
Theorems 4-5 inspected. Finite-horizon 1-(1-epsilon)^H trace bound and general
sharpness are prior results. Stochastic computations are outside this addendum.

SEARCH SCOPE AND LIMITS
Targeted queries: active bounded-disturbance diagnosis; preset/adaptive
distinguishing sequences; ordered threshold decision trees; symbolic control;
probabilistic trace bounds; B-Theory acquisition and self-diagnosis overlap.
No exact duplicate of the signed-binary obstruction was located in the
examined sources. This is NOT a systematic priority result or originality claim.
Lee/Yannakakis survey via CiteSeer and a 2026 outcome-dependent-search publisher
page returned 403 and were not relied on. One Oxford PDF request timed out;
the Bian/Abate arXiv preprint was separately accessible. Full-text Hu/Tucker
coverage remains absent. No access restriction was bypassed.

Attribution: interval reachability, binary leaf bounds, ordered search,
indistinguishable transcripts and support propagation are classical. The
candidate specialized increment is the original-alphabet arithmetic obstruction
and its one-added-input boundary witness. Addendum scope is intentional.
