Request Analyst Access

Analytical Infrastructure  //  CIE Protocol

The world,
encoded.

Global Narratives encodes geopolitical events in CIE — Compressed Information Expression — a formal specification language that enables syntactically-dense, zero-loss, multilingual political analysis and news coverage at institutional scale. Legacy media covers events. CIE encodes them.

Request Analyst Access Read the methodology →
CIE_ANNOTATOR  //  EVENT_VIEWER 
e.26112025.sam.039

bra.jstc.grtc [rl]
    - $ ap-2668.dkt {sc}
    - {rl} -> bra.1955.061 : [jl] > ∫ dt.*:27y
Plain English: The Brazilian Supreme Court issued a ruling on docket AP-2668, sentencing Jair Bolsonaro to 27 years imprisonment, effective immediately from 26 November 2025.

01 / The Problem

Building the infrastructure for the next phase of global analysis.

Legacy media was fundamentally built for a world of information scarcity. Journalists raced to be first. Editors curated what was knowable. Audiences trusted the hierarchy because there was no alternative.

We now live in a world of information overabundance. The same event is described a hundred different ways across a thousand sources — none of them machine-verifiable, none of them structurally comparable. Analysts spend the majority of their time reconciling conflicting narratives rather than extracting signal.

Global Narratives does not just report the news. CIE encodes it — for transparent, testable, rigorous analytic and explanatory uses, from newsmedia to political risk analysis to predictive planning. The substrate of global power is legible. The interface, until now, has been broken.

02 / How CIE Works

A formal language for political events.

01

Encode

Every political event — a sanction, a treaty, a leadership change — is expressed in CIE: a structured syntax that captures actor, action, institution, and time with no ambiguity and no translation loss.

02

Link

Encoded events are assigned to Scenarios: binary, falsifiable predictions designed for analytic rigor. Each assignment carries a directional weight, building a living probability model updated by real-world developments.

03

Analyze

Scenarios aggregate into Narratives — high-level geopolitical frameworks. Analysts can query across thousands of encoded events to surface patterns invisible to prose-based research. CIE is extensible and transferable across use cases.

03 / Sample Events

Events from the corpus.

e.26112025.sam.039

bra.jstc.grtc [rl]
    - $ ap-2668.dkt {sc}
    - {rl} -> bra.1955.061 : [jl] > ∫ dt.*:27y

Brazilian Supreme Court sentences Jair Bolsonaro to 27 years imprisonment on docket AP-2668.

SAM

e.17112025.weu.102

fra.min.jstc.exc.01, fra.min.gndm.* [s-gp]
    - $ fra.sbs.pac.cpl / *axn.*vrbn, *axn.*psc @ *.loc
        - {a-in} min.jstc, min.gndm > [sr] @ *.loc

French ministers of justice and interior pledge coordinated action on crime in Provence-Alpes.

WEU

e.17012026.weu.086

gbr.cpl.exc.01 [s-op]
    - $ gbr.lh.318-2024-26.lex
    - {s-op} + $ sbs.imp.lex

Mayor of London expresses support for new tax powers under legislation 318-2024-26.

WEU

e.10122025.sea.138

tha.dfns.ary.spx [s-ds] tha.ary {ak} khm
    - 3.{ak} ∫@ ~13.9N,~103.2E # .ary, .arf
    - [c-uv] khm.ary [nu] `+1.tha.ary, [ij] ~8.*

Thai army spokesman confirms three armed actions along the Cambodian border, with approximately eight reported injuries.

SEA

e.27112025.nea.072

chn.sbs.hkg.ntfd.incd.exc.01 [s-ds]
    - $ axn.incd @ *.loc 26112025 UTC 0651
        - `+65.civ {di}, `+270 {ij} ~incd
        - unk.hom, unk.fem, unk.enf

Hong Kong fire chief gives update on incident: 65 civilians displaced, approximately 270 injured at location.

NEA

04 / Alpha Testing

Alpha testing.

We are recruiting professional geopolitical analysts for a closed alpha. Participants will encode real events, test analytical workflows, and help refine CIE syntax through practical use. This is institutional-grade infrastructure in active development — not a finished product.

We are looking for analysts with professional backgrounds in political risk, intelligence, international news coverage, international relations, or adjacent fields. Familiarity with formal methods or structured analytical techniques is valued but not required.

Apply for Analyst Access

01 / The Problem

The analytical tools used to understand geopolitics were built for a world that no longer exists.

"Men are disturbed not by events, but by the view that they take of them." — Epictetus

The End of an Age

The contemporary newsmedia and political risk industries were engineered for an era defined by information scarcity. The premium resided in access — the wire service that moved fastest, the analyst who had the source, the bureau that was present when the event occurred. Information itself was the moat. To get the story, or to get it first, was a defensible positional advantage that justified entire institutional architectures: foreign correspondents, embassy contacts, paywalled intelligence reports, proprietary databases.

That world is gone. We now live under conditions of near-total information overabundance. The same event is described by hundreds of sources in the hour after it occurs — wire services, social feeds, official statements, analysis pieces, counter-narratives, corrections, and retractions arriving in an undifferentiated stream. The scarcity that once gave the established institutional players their edge has dissolved. What has replaced it is not clarity, but noise — and a particular kind of noise: untraceable, opaque narrative distortion that is structurally invisible to existing analytical tools.

The problem facing anyone who needs to understand the world at scale is no longer access to raw information. It is the inability to verify, compare, or aggregate what they already have. Our core institutional infrastructure — media organizations, risk consultancies, intelligence services — was designed to solve the problems of the previous era. In doing so, it has become systematically unable to solve the ones we actually face.

The Prose Problem

Prose-based political analysis is the medium through which almost all geopolitical intelligence is currently transmitted — news articles, risk reports, briefing papers, academic publications. Prose is flexible, expressive, and persuasive. It is also, from an analytical standpoint, largely useless at scale. A prose description of a political event cannot be machine-verified, cannot be systematically compared to another prose description of the same or a related event, cannot be searched across a corpus for structural patterns, and cannot be tested for internal consistency or factual accuracy.

The result is that pattern recognition across even a modest corpus of events — say, fifty judicial decisions in Southeast Asia over three years — is practically impossible. Each event is effectively locked inside its own narrative container, accessible only through the analyst's memory or the imprecision of keyword search. The analytical labor required to extract signal from noise under these conditions is enormous, and the results are not reproducible. Two analysts working from the same sources can arrive at incompatible conclusions with no formal mechanism for adjudication.

Compressed Information Expression addresses this directly. CIE imposes strict, transparent, syntactic guardrails on the encoding of political events — ensuring that actor, action, institution, and temporal reference are recorded in a consistent, machine-readable form across the entire event corpus. The resulting structure enables the kind of cross-corpus analysis, pattern matching, and systematic insight that prose-based methods categorically cannot produce. It is not a taxonomy or an ontology. It is a specification language: one in which meaning is transferred completely and unambiguously, regardless of the analyst encoding it or the language in which the underlying source material appears.

The Prediction Market Failure Mode

Prediction markets have attracted serious attention as a candidate solution to the problem of structured political forecasting. The premise is appealing: aggregate distributed beliefs through a price mechanism, and the resulting probability estimates will outperform expert consensus. And on certain narrow, well-defined questions — election outcomes, for instance — they sometimes do. But prediction markets have a structural misalignment between their incentive architecture and the demands of serious geopolitical analysis. Questions are designed to be resolvable, not important. The liquidity necessary for market function concentrates around high-profile events with unambiguous resolution criteria — which is to say, precisely the events that least require forecasting assistance. The events that matter most to institutional decision-makers are often complex, slow-moving, and impossible to reduce to a binary with a clean resolution date.

More fundamentally, prediction markets offer no transparency into the analytical reasoning underlying any given price. A probability estimate produced by a market conveys a confidence level but provides no insight into the logic, the evidence base, or the structural interpretation generating it. For anyone whose professional task is not to bet on outcomes but to understand the dynamics producing them, this is a critical deficiency. A number without an argument is not intelligence.

What's Needed

What the newsmedia and political risk industries need is the analytic revolution that the financial industry completed over the previous forty years: the shift from proprietary, prose-based analysis to standardized data, structured encoding, interlinked databases, and queryable infrastructure. That revolution — driven by Bloomberg, Reuters Elektron, the emergence of algorithmic trading, and ultimately the full digitization of financial markets — did not make human analysts redundant. It changed what they could do, dramatically expanding the scope and precision of what was analytically possible. The analysts who thrived were those who learned to work with the infrastructure rather than around it.

Global Narratives is building that infrastructure for political intelligence. Not another consulting report. Not another prediction market. Not another journalistic organization doing the same thing differently. A wholly revised informational substrate — one in which geopolitical events are encoded with the precision and interoperability that financial data long ago achieved, and through which the analytical possibilities of that substrate are only beginning to be understood.

How Global Narratives approaches this problem →

02 / Methodology

CIE is a formal specification language for political events.

Design Philosophy

CIE was not designed as a taxonomy or an ontology — two approaches that have been attempted before and found wanting in political intelligence contexts. Taxonomies classify; ontologies define relationships. Both are useful for organizing existing knowledge, but neither provides the precision required for encoding novel events in a consistent, transferable form. What was needed was a specification language: a formal system with defined syntax, a controlled vocabulary, and strict encoding rules, such that any event can be expressed by any trained analyst and yield the same structured output.

The design is informed by standards-based thinking drawn from fields where precision, reproducibility, and interoperability are baseline requirements — formal logic, software specification, legal drafting, and the encoding standards that underpin modern financial data infrastructure. The goal of zero-loss multilingual transfer was foundational from the outset: a CIE-encoded event must carry exactly the same analytical content whether the source material is in English, Arabic, Mandarin, or Portuguese, and whether it is encoded by an analyst in London or Jakarta.

CIE has been in active development since 2017–18 and has evolved through practical application rather than abstract design. Every syntactic decision in the current specification reflects a concrete encoding problem that arose in the field — an event that could not be represented without ambiguity, a category of actor that had no existing code, an action type that fell between established categories. The language is not finished. It is designed to be extended systematically as the corpus grows and new encoding requirements emerge.

The Four-Layer Hierarchy

CIE organizes political knowledge into four nested layers, each with a defined role in the analytical architecture. The structure is designed so that lower layers supply the empirical substrate for higher-layer analysis.

Anatomy of a CIE Expression

Below is a fully annotated CIE expression encoding a real political event. Hover over any underlined component to see its plain-English interpretation. Each element is syntactically required and has an unambiguous definition within the CIE specification.

CIE_ANNOTATOR  //  EVENT_VIEWER 
e.26112025.sam.039

bra.jstc.grtc [rl]
    - $ ap-2668.dkt {sc}
    - {rl} -> bra.1955.061 : [jl] > ∫ dt.*:27y
Plain English: The Brazilian Supreme Court issued a ruling on docket AP-2668, sentencing Jair Bolsonaro to 27 years imprisonment, effective immediately from 26 November 2025.

Action Codes

CIE uses a controlled vocabulary of action codes to classify political behaviors. The vocabulary is hierarchical: broad categories (Legal, Diplomatic, Military, Statement, Claims) subdivide into specific codes, each defined precisely and designed to be mutually exclusive. Every action in CIE has one correct code — the encoding is not interpretive, it is classificatory. The current vocabulary contains over 100 action codes covering the full range of political, institutional, economic, military, and informational behavior observed in the event corpus.

Code Category Name Description
[vu] Legal-Executive Approves The act of an executive or other non-legislative governmental entity approving a formal process
[rc] Political-Diplomatic Recognizes The act of granting formal status under law to a specific person, institution, or act
[nm] Political-Operational Nominates The act of selecting an individual for a designated office with the need for confirmation of this choice by another entity
[pt] Commercial-Operational Capitalizes The act of providing capital to a designated enterprise in exchange for equity
[nu] Military-Kinetic Neutralizes The act of killing or destroying a specified target by the recognized and uniformed armed forces of a state
[s-al] Statement-Inflected Alleges To raise a defined, unverified but falsifiable claim about a specified actor or institution
[s-st] Statement-Neutral States To offer a formal declaration of individual position on a specific entity, individual, or event
[c-or] Claims & Rumors Off-the-Record Claim A specific claim, opinion, or interpretation made without full attribution at the request of the claimant
[c-uv] Claims & Rumors Unverifiable Claim A factual and falsifiable claim or assessment made by an individual or organization whose underlying truth cannot be verified at time of reporting

Scenarios and Probability

Every encoded Event can be assigned to one or more Scenarios — binary, falsifiable predictions about how a specific geopolitical situation will resolve. Each assignment carries a directional weight on a scale from −12 to +12, indicating how strongly the event supports or undermines the Scenario's resolution condition. A judicial decision that directly advances a constitutional crisis scenario might carry a weight of +10; a diplomatic statement that marginally complicates the picture might carry +2. Weights are assigned by the encoding analyst according to a defined categorical framework — the process is structured and transparent, not discretionary.

As Events are encoded and assigned, the system aggregates their weighted inputs to produce a probability estimate for each Scenario. This estimate updates in near-real time as new events arrive. Crucially, the probability is not a black box — every contributing event, its directional weight, and its categorical classification are visible in the audit trail. An analyst who disagrees with a probability estimate can examine the inputs that produced it, identify where their interpretation diverges, and articulate precisely why. This is what distinguishes CIE-based analysis from both prose-based consulting reports and prediction market prices: the reasoning is inspectable and falsifiable at every level.

Apply for analyst access →