Methodology: Reconstructing Truth

A method for reconstructing truth where time-segmented institutional processes have weaponised epistemology.

1. Begin from fragmentation, not fabrication

Assume that institutional failure rarely arises from outright falsehood. More often, it arises from fragmentation across time:

  • knowledge distributed across agencies
  • decisions justified locally and retrospectively
  • records created for different purposes at different moments
  • no single actor holding the full picture

If you begin by accusing dishonesty, the method fails. If you begin by assuming fragmentation, the work can proceed.

2. Collate everything before understanding anything

The first practical step is total collation.

This includes, without filtering:

  • FOIA disclosures
  • SAR responses
  • custody records
  • CAD logs
  • court bundles
  • correspondence and internal communications
  • machine extracts and summaries

At this stage, relevance is unknowable. You do not summarise, rank, or discard. If it exists, it enters the corpus.

3. Destroy institutional aggregation

Institutional documents are aggregated for administrative convenience, not epistemic accuracy. That aggregation must be reversed.

Custody records:
  • A single custody PDF often contains multiple arrests and episodes.
  • It must be split into individual custody episodes.
  • Each episode must then be split into procedural moments.
CAD records:
  • CAD logs must be split into atomic communications.
  • They are not incidents, outcomes, or summaries.
  • Gaps, silence, and resumptions are data.
Machine-extracted summaries are never sufficient. They flatten the very detail that carries epistemic weight.

4. Normalise hostile formats without interpreting them

Different records speak different structural languages:

  • custody records are fixed-form and justificatory
  • CAD logs are conversational and contemporaneous
  • disclosures are selective and defensive

They must be normalised conceptually as events in time, without explanation or reconciliation. Meaning is deferred.

5. Build a single, continuous temporal map

Every atomic event is placed onto one timeline governed by a single rule:

When did this exist, and what was known at that moment? (More detail on chronology building here)

No motives are inferred. No contradictions are resolved. Sequence, proximity, overlap, and delay are fixed.

6. Allow contradictions to exist

Contradictions are not errors. They are pressure points.

CAD knowledge may contradict custody narratives. Early records may contradict later justifications. Silence may sit between actions without explanation.

These are preserved, not resolved.

7. Separate record, meaning, and consequence

Only after the temporal map exists is the work separated into domains:

  • Chronology: what existed, when, and how it is anchored
  • Analysis / appendices: what those sequences imply
  • Claims: what consequences follow under recognised frameworks

This separation prevents hindsight, advocacy, and emotion from contaminating record.

8. Apply the system’s own standards back to the whole

The final move is not adversarial.

The system’s own epistemic standards — consistency, sufficiency, procedural fairness — are applied to a dataset that the system itself never held whole.

Time does the accusing. No rhetoric is required.

9. Stop when the map is complete

The endpoint is not narrative satisfaction. You stop when:

  • no document floats unanchored in time
  • no new events can be placed
  • contradictions are visible and stable
  • silence is accounted for as silence
Anything earlier is not progress. It is premature coherence.