VIGIL CONSILIUM Search
Context. Not headlines.
Contents

Open Professional
Professional methodology
What Vigil Professional is

Chapter 05

Research

Research is a working surface over the corpus, not a search box. A query fills the panels around it, and selecting a result opens what it rests on and what it connects to without leaving the page.

The workspace

The Research workspace with the query “hormuz” returning 45 records grouped by family — watch question, observable records, scenario templates, developments, actors, events, locations — one observable record selected, and three panels beside it: an evidence chain of question, observables, this record and sources; related entities naming six actors and six places; and sector lenses.
One query, and four instruments around it: the results by family, the evidence chain beneath the selected record, the entities it names, and the sector lenses it falls under.

The panels are instruments, like Monitor's. Add or remove them, change the layout, and save the arrangement as a view — the controls are the same row and behave the same way.

Querying

Ask for a thing, not for a document. A place, an actor, an instrument, a system, a phrase from a record.

The query instrument after searching “hormuz”: a count of 45 matching records, a grouping control set to “by family”, and a row of family facets — watch question 1, observable record 5, scenario template 2, development 4, actor 4, formation 4, event 4, logistics system 2, terrain 2, impact chain 1, economic exposure 1, location 10, zone 2, weekly brief 1, source 1, reference 1.
  1. How many records matched, and how many are shown. Nothing is ranked by importance.
  2. Families. Press one to narrow to it — this is the fastest way to get from 45 records to the 5 you want.
  3. Why each result matched: its name, its own text, a source it cites, or a place it names.
Grouping by family is what makes a broad query useful: the shape of what the corpus holds is visible before anything is read.

Filters narrow by module, actor, location, the question a record bears on, confidence, strategic system, sector lens and date. Every filter is in the address, so a query you are pleased with can be sent to a colleague or kept with Save view.

Save the query, not the results. A saved query re-runs against the corpus as it is today. A list of results captured six weeks ago would be a snapshot of a corpus that has since been reviewed — which is the one thing a research tool should never hand you.

Evidence

Every reading in Professional sits on a chain you can walk to a named source.

  1. Watch question

    The standing question the reading bears on.

    Public.

  2. Observable

    What is being checked, where it stands, what would move it, what it cannot establish.

    306 records.

  3. Evidence entry

    A dated line naming the canonical development, event, Note or brief it rests on.

    586 entries.

  4. Source

    The registry entry the basis cites, and every other place it is used.

    149 sources.

Four kinds of canonical record, each with an address of its own. Every reading in Professional can be walked down this chain, and the counts beside each link are counts of records rather than a measure of anything.
One observable record — “War-risk cover written for Hormuz transits at rates charterers accept” — showing its parent question, four counts across evidence entries, countervailing evidence, movement conditions and stated limits, the test being applied, and the current reading marked as a fact.
  1. The register. This reading is a recorded fact rather than Vigil’s assessment.
  2. Three dates: what the reading is as of, when it was last reviewed, and its review band.
  3. Four counts: dated evidence, how much of it bears against, what would move the reading, and how many limits it states.
  4. What is being checked — the test, written before the reading.
One observable record in full. The parent question is public; this layer is what Professional adds beneath it.

Reading an evidence entry

Each entry carries a date, a direction, what it bears, and the canonical record it rests on.

  • Toward — it bears toward the observable being met.
  • Against — it bears against it.
  • Neutral — it bears on the observable without direction.

Directions do not sumFour entries toward and one against is not a score of three. The direction says how one piece of evidence bears; nothing anywhere adds them up, and the reading is the editor's, arrived at against all of them.

The limits are the part to read first

Every observable states what this reading cannot establish. In research that is the most valuable field on the record: it tells you where you would be over-reading if you used this in an argument. Read it before the reading, not after.

The Evidence Explorer

Research answers "what does this rest on". The Evidence Explorer answers the other two directions: where is this source used, and what does this development affect. It is the same chain, walked from a different end, and it is how you check whether a source is carrying more weight across the corpus than you are comfortable with.

A citation count is a count of citations. It is not a measure of a source's quality, not a measure of a claim's strength, and not a ranking of one body of evidence against another.

Sector lenses

A lens reads one theme across every module at once.

Records declare the sectors they belong to. A lens gathers everything that declares one — so Energy shows what all 9 sector vocabularies' worth of records hold on production, processing, export and restoration, wherever in the world it sits.

The Energy sector lens: a definition of what the lens covers — production, processing, export and restoration of oil, gas and power — over the records from every module that declare it, grouped by module and by kind.
A lens holds no records of its own. Every row on it belongs to a module and links back to it.
Security / military
Force posture, defence arrangements, command and the conditions of armed confrontation.
Energy
Production, processing, export and restoration of oil, gas and power.
Maritime
Straits, ports, shipping, cover and the physical, commercial and legal usability of sea routes.
Commercial / supply chain
Trade, freight, insurance, supply-chain continuity and the exposures of firms.
Defence industry
Procurement, production, transfers and industrial capacity.
Humanitarian
Displacement, access, protection and the civilian consequences of the system.
Civil resilience
Electricity, water, communications, transport, health and continuity of government.
Infrastructure
Fixed nodes, corridors and networks, and their disruption and repair.
Political / institutional
Institutions, mandates, agreements and the decisions of governments.

Use a lens when your brief is thematic rather than geographic — an energy analyst who covers four theatres reads the Energy lens first and the modules second.

Timeline

A dated sequence assembled from canonical records.

The Timeline for Ukraine–Russia: date-range and keyword controls, a row of record-kind filters, and a dated sequence of canonical records — events, developments, reviews and briefs — in order, each linking to the record it came from.
  1. Window and keyword. Narrow before you read — a module timeline is long.
  2. Record kinds. Events, developments, reviews and briefs can each be turned off.
  3. The sequence itself, each row linking back to the record it came from.
Nothing on a timeline is authored. It is the module's own dated records, in order.

Three things it is good for:

  • Reconstruction — what actually happened, in order, when an account you have read is vague about sequence.
  • Sequence analysis — whether one thing preceded another, which is usually the question underneath "did X cause Y".
  • Brief preparation — the chronology section of a briefing, drawn from records rather than written from memory.

Hold a chosen set and keep it with Save; the set reopens as a timeline of exactly those records. Scenario Lab and Compare can both send a selection here.

Next: Analyse — two records side by side, and one condition held.