E Evidence Press

Press release · 7 September 2026 · version 0.1.0-candidate

Reducedness of the Hilbert scheme of eight points in affine four-space

A computer-assisted proof candidate claims reducedness of the Hilbert scheme of eight points in affine four-space over algebraically closed characteristic-zero fields.

Listen to this briefingNarrated summary · OpenAI API synthetic voice (fable) · MP3 · download

Summary

Eight points can collide in ways that ordinary pictures cannot distinguish. The Hilbert scheme records not only their positions but also algebraic structure at collisions. This unrefereed proof candidate claims that, for eight points in four-dimensional affine space over a characteristic-zero algebraically closed field, the Hilbert scheme has no hidden nilpotent structure. It remains singular and reducible: reduced does not mean smooth.

Summary for specialists

The claimed theorem is reducedness of $\operatorname{Hilb}^{8}(\mathbb A^4_k)$ for algebraically closed $k$ of characteristic zero. The candidate reduces the question to two monomial local rings, separates the translated homogeneous component by exact annihilator identities, and identifies the full associated graded algebras of their Pfaffian quotients. No positive-characteristic or general-length assertion is included.

Technical account

The closed support of the nilradical, a Borel fixed-point argument and sixteen explicit triangular coordinate frames reduce the global question to B15 and B16. Original commutator identities and an ambient equivariant Grassmannian projection establish local component separation.

For B16, an integral open chart and a rank-1907 finite-module comparison produce an integral core. B15 requires two open charts and the corrected parameter $\ell_1=c_7+c_{79}$. A zero fiber modulo 32003 gives an upper bound; a Cohen–Macaulay, flat comparison fiber modulo 65521 and its completion give the matching lower bound 9923. Regularity of that parameter closes the remaining tangent-cone kernel in every degree. This is not an inference from the last computed degree.

The antecedent component and Pfaffian descriptions are due to Cartwright, Erman, Velasco and Viray. Hu supplies important local-structure and reduction methods. Their results are inputs; the candidate's claimed contribution is the source-specific reducedness argument, not rediscovery of those inputs.

Evidence, assurance and limitations

The complete package includes exact ring presentations, parameter forms, coordinate maps, written proof modules, all four finite bases, native software-diverse counterparts and direct coefficient-comparison records. Those records cover 55,557,191 terms. The public aggregate replay checks saved receipts and their exact dependencies; it does not silently recompute all large bases. Original code, deeper replay instructions and complete outputs allow inspection beyond the aggregate.

The supplied review saw the overview and elementary checks, not the full archive, owing to an upload limit. Internal model editorial review is a different, producer-coordinated process. Neither establishes unaffiliated whole-proof validation. Formal verification, exhaustive novelty clearance and independent external reproduction remain unestablished.

Who should care, and why

AudiencePotential useRequired caution
Hilbert-scheme specialistsInspect a proposed answer to a scheme-structure question beyond component classification.Scrutinize the source-specific identifications and all-degree argument.
Computational algebraistsReproduce large finite fibers and inspect a mixed-prime finite-module bridge.Software agreement and hashes do not replace the source-to-claim argument.
Research-tool buildersReuse explicit replay levels and semantic corruption controls.Workflow assurance is not mathematical truth or measured productivity.

Why the problem matters

Knowing the irreducible components of a space does not determine its scheme structure. Nilpotents can record infinitesimal information invisible in its underlying set. This candidate targets that distinction in a small but already reducible Hilbert scheme. Its proposed finite-to-global method may also be useful elsewhere, provided every hypothesis is proved anew.

How to inspect or reproduce the recorded checks

Download and extract the full versioned archive. Read the manuscript and README, then run python3 verify_release.py. Use --manifest-only for a byte-integrity check. The wrapper keeps assertions enabled in its legacy children even if invoked with python -O. Follow the detailed instructions in a disposable copy for new direct comparisons or native Gröbner runs; those runs regenerate timed receipts and must not overwrite the frozen record.

The most valuable next projects

  • Obtain unaffiliated scrutiny of the ambient Pfaffian transport and the B15 purity/open-coverage argument.
  • Independently rebuild the source-to-fiber encoding and recompute the full bases in a documented environment.
  • Explore other applications of the mixed-prime lemma without transferring its flatness or dimension hypotheses by analogy.

What is in the evidence package

The manuscript, original proof modules, exact source code and full finite outputs are accompanied by structured claims, an assurance map, a review response, internal editorial reports, environment instructions, negative controls, licences and a complete hash manifest. Immutable GitHub and Zenodo assets are the reference versions; the page and audio are communication aids.

Media

The audio briefing is provided in the header above. Download the MP3 briefing · read the transcript.

Open directions for follow-up research

Also available in machine-readable form for research agents and follow-up projects.

  1. Unaffiliated scrutiny of ambient Pfaffian transport and B15 purity/open coverage.
  2. Independent reconstruction of source encodings and full native computations.
  3. Investigate positive characteristic or other lengths only as new problems with their hypotheses checked anew.

Research process, metrics and reusable methods

Prospective process metadata under the Evidence Press operating model and research-metrics policy. It records the intended handoff, measured scope and claim boundary; it is not evidence that the method accelerated this work.

Work ID
ep-work:hilbert-eight-points-reducedness
Attempt and metric receipts
  • ep-attempt:hilbert-eight-points-reducedness-assurance-publication — published / positive

    Measurement scope
    assurance-through-publication — Remaining review revision, assurance and first canonical publication only. Discovery, initial intake, extraction and initial verifier-wrapper implementation predate registration and are excluded; their time is not reconstructed. Final ledger reseal and closeout are reported separately.
    Frozen target
    Review-repaired candidate, exact archive replay, internal editorial gate, immutable public assets and canonical Evidence Press readback, or a preserved integrity blocker.
    Fermi active-time forecast
    150 minutes; plausible interval 120–210; expected unattended wait 30. Reference class: Full-candidate procedural prior (n=0) — Skill benchmark, not an empirical matched comparison..
    • Manuscript and source revision: 1 × 35/45/65 minutes (low/central/high) — Existing full evidence; substantive exposition repairs.
    • Package replay and editorial gate: 1 × 35/45/65 minutes (low/central/high) — Large frozen bases, one five-role internal gate.
    • Public identities and media: 1 × 25/30/40 minutes (low/central/high) — Established publication tools.
    • Seals and canonical readback: 1 × 25/30/40 minutes (low/central/high) — Two new-slug deployment cycles.
    Tractability forecast
    Within 240 active minutes: positive signal 0.9; target closure 0.8. Stop rule: Fail closed on unresolved integrity or authority gates; do not expand the mathematical claim. Time thresholds are recorded telemetry, not an arbitrary termination rule.
    Observed clocks
    19 active-agent; unknown active-human; unknown substantive-compute; 3 unattended-wait; 0 blocked; 0 rework minutes. Calendar elapsed: 88 minutes.
    Research search
    Cycles: 0 positive, 0 negative, 0 inconclusive. Falsification gates: 1. Candidate architectures: 0 tested, 0 rejected.
    Agent and review load
    6 agent runs; maximum parallelism 4; 244 model turns; 33201497 deduplicated model tokens; 1 substantive review rounds; P0/P1 findings 0/0; pre-publication claim corrections 0.
    Result and calibration
    target-closed — Review-repaired candidate, exact complete archive, internal editorial acceptance, public GitHub/Zenodo byte readback, standard media and first guarded canonical publication completed. Final ledger reseal follows outside this measurement cut. No external mathematical assurance or discovery-speed claim. Eighteen mode/corruption cases constitute one adversarial gate, not eighteen research cycles. Positive signal: true; target reached: true. Active-time error -131 minutes; actual/forecast 0.12666666666666668; inside interval: false. Brier score: positive signal 0.01; target closure 0.04. Variance: Active time is the union of instrumented Reasoning and AgentMessage intervals; unattended time includes instrumented compaction only. Both are observable lower bounds, not complete labour/wait totals. Tool-call generation and uninstrumented provider waits are excluded. Deduplicated response IDs remove fork copies. Cached inputs remain part of raw tokens. No historical discovery clock is inferred. Offline gaps and unnecessary administrative pauses are not classified as external blocking. Rework time is a zero instrumented lower bound, not a claim that no repair occurred: the replay recipe, narration and generated audio index were repaired. This incomplete timing coverage is not suitable for claiming a forecast speed gain. Earlier discovery and pre-registration work remain excluded.
    Missing telemetry
    activeHumanMinutes — No contemporaneous human active-time instrument.; computeMinutes — No complete scoped substantive-computation wall clock; ordinary CI/build time is not research compute.
    Measurement corrections
    • measurement.agentRuns -> metrics.outcome.agentRuns — Terminal receipt records six agent runs; original intake snapshot retained. Reason: Prospective snapshot predates five completed internal reviewers.
    • measurement.reworkMinutes -> metrics.outcome.reworkMinutes — See outcome varianceReason: actual replay-instruction, audio and governance repairs are disclosed; zero is not an absence-of-rework claim. Reason: Intake snapshot was unknown; terminal timing records an explicitly incomplete zero instrumented lower bound.
Prospective work ledger · metrics policy
Intended aims
science
Artifact roles
research-output, evidence-assessment, communication
Decision object
certificate — Source-specific local reducedness proof candidate and exact finite computational evidence. Scope: Hilb^8(A^4_k), algebraically closed characteristic-zero k; two local models B15 and B16.
Reusable methods
Certificate-first, proof-carrying research (certificate-first); Structural compression (structural-compression); Adversarial scientific controls (adversarial-controls); Productive failure and stop receipts (productive-failure); Assurance as a vector (assurance-vector); Agent-readable research objects (agent-readable-research-object) · registry
Targeted clocks
assurance, publication
Semantic bridge
explicit — Original border-basis equations, ambient equivariant projection, component separation and all-degree graded arguments connect finite computations to the claimed reducedness statement. Remaining risks: Written mathematical bridges require unaffiliated specialist scrutiny.; Aggregate replay checks saved records, not every large computation afresh.; Finite-field identities alone do not prove characteristic-zero reducedness.; Novelty and priority search is bounded..
Human judgement gates
  • Check global-to-local reduction and ambient transport.
  • Check purity, open coverage and mixed-prime freeness hypotheses.
  • Preserve source attribution, component rights and assurance boundaries.
Next assurance action
Unaffiliated scrutiny and independent source-to-computation reconstruction. Claim ceiling: Unrefereed, author-claimed complete computer-assisted proof in characteristic zero only. No external whole-proof validation, formal verification, exhaustive novelty clearance or priority determination.
Aim-scoped impact evidence
  • science: NO_IMPACT_EVIDENCE — Inspectable reducedness proof candidate in Producer-coordinated mathematical publication. Design: none; comparator: No matched workflow comparator.; estimand: No discovery-speed, effort, reuse or impact effect estimated.. No real-world effect evidence is asserted.
Parent handoffs
  • depends-on-claim https://doi.org/10.2140/ant.2009.3.763 — inherited claim: Component and Pfaffian descriptions for eight-point Hilbert schemes.; inherited ceiling: Established antecedent is not validation of this candidate reducedness argument.

Verification status

Unrefereed, author-claimed complete computer-assisted proof in characteristic zero only. No external whole-proof validation, formal verification, exhaustive novelty clearance or priority determination.

Cite

Anonymous. (2026). Reducedness of the Hilbert scheme of eight points in affine four-space (Version 0.1.0-candidate) [Unrefereed computer-assisted proof candidate]. Evidence Press. https://doi.org/10.5281/zenodo.22643443
BibTeX
@misc{hilberteightpointsreducedness2026,
  title        = {Reducedness of the Hilbert scheme of eight points in affine four-space},
  author       = {Anonymous},
  year         = {2026},
  doi          = {10.5281/zenodo.22643443},
  url          = {https://doi.org/10.5281/zenodo.22643443},
  version      = {0.1.0-candidate},
  howpublished = {Zenodo},
  note         = {Unrefereed; internally replayed evidence package. Press page: https://evidencepress.org/releases/hilbert-eight-points-reducedness/}
}

Also: cite.bib · paper.json · this page as Markdown