{
  "article": "<figure class=\"distributed-mediation-figure\"><img src=\"assets/evidence-viewer/distributed-mediation-multiple-witnesses-single-point-failure.png\" width=\"1672\" height=\"941\" alt=\"Distributed mediation diagram comparing a single point of failure with multiple witnesses and sources converging toward a more inspectable conclusion.\" loading=\"lazy\" decoding=\"async\"></figure>\n<div class=\"detail-section-heading\">Observation</div>\n<div class=\"detail-article-block\">\n<p><strong>Distributed mediation does not make testimony automatically true; it makes the structure more inspectable.</strong> One witness, one manuscript, one chain of custody, or one repeated rumor can fail in a single place. A field of letters, creedal summaries, Gospel traditions, worship practices, opponents, names, places, and public claims can still be mistaken, but it cannot be dismissed by breaking one thread.</p>\n<p>This is methodology, not scoring. The row explains why The Signal distinguishes repeated dependence from genuinely distributed support, why dependency caps matter, and why a rival explanation must preserve the whole field rather than the easiest isolated fragment.</p>\n</div>\n\n<div class=\"detail-section-heading\">Source Structure</div>\n<div class=\"detail-article-block\">\n<p>Historical reasoning asks how evidence is mediated. Are we hearing one report copied many times, or several channels with partial independence and visible contact points? Are differences signs of fabrication, independent memory, genre, compression, theological shaping, or all of those in different places? The answer has to be argued rather than assumed.</p>\n<p>Public revelation has a similar structure. Christianity does not ask the reader to trust an invisible private oracle alone. It points to public claims: Israel's story, Jesus' ministry, crucifixion, resurrection proclamation, apostolic witness, texts, communities, worship, and rival objections. Each line can be inspected, and each line has limits.</p>\n</div>\n\n<div class=\"detail-section-heading\">What This Row Does</div>\n<div class=\"detail-article-block\">\n<ul>\n<li><strong>It names mediation.</strong> Evidence arrives through sources, witnesses, communities, texts, memories, and practices.</li>\n<li><strong>It protects against brittle proofs.</strong> If one strand weakens, the question becomes what remains, not whether the whole case vanishes by reflex.</li>\n<li><strong>It protects against overcounting.</strong> Related strands still need caps because distribution is not the same thing as full independence.</li>\n<li><strong>It keeps public revelation inspectable.</strong> Claims tied to history can be questioned, cross-checked, and corrected.</li>\n</ul>\n</div>\n\n<div class=\"detail-section-heading\">What This Row Does Not Do</div>\n<div class=\"detail-article-block\">\n<ul>\n<li>It does not duplicate <code>E-REV-TESTABLE</code> or <code>E-PHIL-REVELATION-GAP</code>.</li>\n<li>It does not add a Bayes factor to resurrection, Christ identity, canon, or Logos.</li>\n<li>It does not say multiple sources are automatically reliable.</li>\n<li>It does not remove the need for source criticism, genre awareness, or dependency governance.</li>\n</ul>\n</div>\n\n<div class=\"detail-section-heading\">Bayesian Meaning</div>\n<div class=\"detail-article-block\">\n<p>This item currently has <strong>no active Bayes factors</strong>. It is unweighted methodology for source and testimony structure. Its job is to explain why distributed evidence, dependency caps, and public inspectability matter without turning those methodological points into extra score.</p>\n</div>",
  "axioms": [
    "A6"
  ],
  "bayes_factors": {},
  "bf_status": "unweighted_explanatory",
  "category": "Evidence Governance",
  "citations": [
    {
      "title": "C. Blomberg, The Historical Reliability of the Gospels (on source relations).",
      "url": ""
    },
    {
      "title": "Richard Bauckham, Jesus and the Eyewitnesses, 2nd ed. (Eerdmans, 2017).",
      "url": ""
    },
    {
      "title": "Richard Swinburne, The Resurrection of God Incarnate (Oxford University Press, 2003).",
      "url": ""
    },
    {
      "title": "C. A. J. Coady, Testimony: A Philosophical Study (Oxford University Press, 1992).",
      "url": ""
    }
  ],
  "counts_in_cache": true,
  "evidence_id": "E-REV-MEDIATION",
  "last_updated": "2026-05-28T00:00:00Z",
  "major_category": "Methodology / Signal Core",
  "metadata": {
    "category": "Evidence Governance",
    "last_updated": "2026-05-28",
    "major_category": "Methodology / Signal Core",
    "rev": 3,
    "sub_category": "Source / Testimony Structure",
    "legacy_bayes_factors_status": "archived_not_runtime_scored",
    "legacy_bayes_factors_note": "Legacy Bayes factors are retained for audit history only. Runtime scoring uses the active bayes_factors field.",
    "legacy_bayes_factors_reviewed": "2026-05-17",
    "evidence_function": "methodological_pressure",
    "directness": "methodological",
    "dependency_cluster": "methodological_controls",
    "dependency_role": "methodology",
    "dependency_cluster_id": "methodological_controls",
    "dependency_cluster_role": "methodology",
    "cap_profile": "manual_review",
    "answer_status": "methodological_control",
    "counts_as_direct_resurrection": false,
    "counts_as_direct_christ_identity": false,
    "counts_as_direct_logos_synthesis": false,
    "source_status": "expanded_source_review_pending",
    "governance_note": "Expanded unweighted methodology/source-structure clarity only. No BF change; no public-testability duplication.",
    "scoring_note": "Unweighted methodology row. Do not add score unless a future testimony/reliability support-layer hypothesis is explicitly approved."
  },
  "status": "enriched",
  "sub_category": "Source / Testimony Structure",
  "summary": "Unweighted methodology row: distributed mediation makes historical and revelatory claims more inspectable while preserving dependency caps and avoiding extra score.",
  "tags": [
    "Revelation",
    "Theism comparison",
    "Textual"
  ],
  "title": "Distributed mediation - multiple witnesses vs single point of failure",
  "type": "atomic",
  "hypothesis_ref": [],
  "legacy_bayes_factors": {
    "H-MULTI": {
      "bf_max": 0.4,
      "bf_min": 0.1,
      "log10BF": 0.25,
      "rationale": "Redundancy and cross-checks modestly raise reliability."
    },
    "H-SINGLE": {
      "bf_max": 0.05,
      "bf_min": -0.25,
      "log10BF": -0.1,
      "rationale": "Single-node models are more failure-prone."
    }
  },
  "disposition_status": "contextual_unweighted",
  "cluster_note": "Methodology boundary: keep unweighted unless a testimony/reliability support-layer hypothesis is approved.",
  "source_note": "Expanded for source-structure clarity. Distinct from E-REV-TESTABLE and E-PHIL-REVELATION-GAP; no active BFs."
}