Orientation: The Information Environment You Live In

The information environment stack

The six elements are the periodic table of this course — the complete alphabet from which every defensive read of an operation is built — and the stack is those six elements drawn in the one arrangement that lets a target read them safely. The mechanism it shows is directional. The operator works top-down: some actor with a purpose picks a vector, hires or organises amplifiers, wraps its intent in a narrative, and rides the platform into the reader's mind, while spending real effort to keep the signal — the coordination, the timing, the provenance — hidden, because the moment the reader sees the signal the operation loses its power. The defender reads the same stack bottom-up, in the exact opposite direction, because that is the only direction in which each layer can be checked against something the reader can actually verify. It starts with the effect on you: why did I notice this? Then the platform: where did I see it — a feed, an ad, a forwarded message, a group? Then the amplifiers: who put it in front of me, and what does their public data look like? Then the narrative: what compact story am I being asked to hold, and does it match what independent sources say? Then the vector and timing: when did this appear, on what channel, and does the timing look organic or engineered? And only at the very top, and only if the evidence goes that far, the actor — described at the level a published investigator would sign, and no further. The completeness argument is simple: to defend your judgement in the information environment, you have to know what is being asked of you, what narrative is carrying that ask, on what vector it arrived, who moved it, what signal it leaves and what effect it is having — and every module of the course reads one of these. Module 3 lights the arrow into the reader (the cognitive ground); Module 4 lights the observable signal (what a push looks like from outside); Module 5 lights the narrative layer; Module 6 lights the platform layer; Module 7 lights the amplifier layer with its network structure; Module 8 sets the documented record of who has done this and how far the evidence has been made to go; Module 9 lights the arrows between layers and the confidence tags on them; Modules 10, 11 and 12 draw shields around the reader, the firm and the public square. What falls outside the stack — how to build any of its layers, how to reach an audience, how to disguise a signal — is out of scope in this course, by design and without exception.

Traced instance: the Halden petition. On the day the petition appears, a defensive reader in Halden sees only the bottom of the stack. Their feed shows a neighbour's share of the petition text; a paid post from the Halden Citizens' Voice page; a short piece from the Standard-Bearer two days later. That is the effect and the platform layers, and both can be observed directly (evidence class A1 in the course's grading system, taught in M9). Above that, the amplifier layer takes verification but not much: the Citizens' Voice page's creation date, its posting frequency and its top-100 sharers' public metadata are all readable on the platform — illustratively, 62 of those sharers were created in the last 30 days, 24 use stock or AI-generated profile photos and posting concentrates between 03:00 and 07:00 Meridian time (evidence class B2: mostly reliable source, probably true). Above that, the narrative layer is the compact story — "Halden First: pause the rare-earth refining permits" — which can be read at face value, and its top-shared image (a Halden street with a fabricated air-quality readout) can be reverse-image-searched to a 2023 unrelated original (evidence class A2 — a reliable source, but the caption is manipulated). The vector and timing layer is where the operation begins to be visible as a push: a page created Day −4 posting 41 times in 72 hours, coordinated timing across the top-100 sharers, a fabricated artefact reused across posts. That is the signal, and on Day 9 Meta confirms enough of it to take down a network of 84 accounts and 6 pages for coordinated inauthentic behaviour targeting Meridian civic discourse (illustrative platform statement, evidence class A1). On Day 11 and Day 14 two named research organisations — Halden University's Media Integrity Lab and CivicSignal Europe — publish reports at moderate confidence that a coordinated network amplified the petition, and at low confidence on origin, listing rival hypotheses (hire-for-influence contractor; foreign state cutout; domestic activists using low-cost amplification services; organic community outrage with parasitic bad actors on top). And the actor layer — the very top of the stack — is where the defensive reader stops. Publicly, on the evidence available, the origin is . An anonymous X thread claiming "we've proved it was foreign" is evidence class — unreliable source, truth cannot be judged — and is set aside. What the defensive reader of the course records is exactly this: . And that is the of the course itself — the highest claim any module makes about the Halden case. Every module honours it, and Module 9 makes the discipline of that ceiling explicit.

From target to defensive reader

The gap between the two directions of reading is the whole value of the course. The target — anyone who reads the environment top-down, the way it presents itself — meets the narrative first, then absorbs it before the platform and the amplifier and the vector have registered as anything worth checking. The operation's cost of hiding the signal is spent buying exactly that: a few seconds in which the story feels like the room. In those seconds the target does the work the operation needs — believes a little, shares once, moves on. The defensive reader replaces each of the target's habits with a discipline, and every discipline is a form of reading the stack in the harder direction. The first is pausing before amplifying. A share is an act at the amplifier layer, not the narrative layer: to share is to become a small amplifier for whatever the story turns out to be, and once a share is made it cannot be recalled from the people who saw it. So the defensive reader treats a share as a decision, not a reaction, and defaults to not yet until the triage — source, corroboration, evidence, intent (M4 · O4) — has been run. The second is the four orienting questions. What belief, action or attention is being asked of me? — because until the ask is named, the story is doing its work quietly. What does the evidence actually support? — the observation that a petition exists is not the observation that its origin is known. What would a public investigator require to say more? — the yardstick that stops the reader from confusing a plausible story with a graded finding. What am I going to do? — because reading without acting is a kind of passivity the operation counts on, and not amplifying is itself an action. The third discipline is grading every claim. Every sentence a defensive reader relies on is one of three things: a fact (the petition text, the signature count, the platform's public data); an assessment (that the Citizens' Voice page looks coordinated with the top-100 sharers); or a speculation (that a particular state was behind it). Fact, assessment and speculation each carry their provenance, and speculation is never dressed as fact, even when a story that treats it as fact would be more satisfying. The fourth is attribution restraint — the discipline of Module 9. The words "Russia was behind it", "the Chinese Communist Party did this", "a firm called X ran this" are legal, diplomatic and reputational instruments. They are reserved for qualified investigators who can back them with evidence they can defend in public. A lay defensive reader does not use them; instead, the reader describes what they see and cites the level at which a published investigator has actually spoken. The fifth is the the course teaches — a small vocabulary that stops the two most common failures of the target. , , and , tied to the seven-band probability yardstick published in the US intelligence community's analytic standards (ICD 203; the course adopts it in the form used across Tricellix). Every judgement in the course carries both a probability band and a confidence, because a judgement at confidence and a judgement at confidence are very different objects. The sixth is . A defensive reader does not doxx, does not scrape in breach of platform terms, does not name a private individual as an operator, and — when a judgement they voiced turns out to be wrong — says so, once and clearly, rather than letting it fade.

The method — thirteen rungs

The course is thirteen modules, and the structure follows the four orienting questions and the stack. After this orientation, Module 2, The Discipline & Boundary of a Defensive Reader, installs the six commitments that govern everything else — awareness not tradecraft, lawful and ethical inquiry, fact / assessment / speculation with provenance, attribution restraint, protecting people over networks, and correction in the open. Modules 3 and 4 lay the ground. The Cognitive Ground: Why Minds Are Targetable names the well-documented biases that operations exploit — confirmation, availability, in-group and out-group, motivated reasoning — never to weaponise them but to defend against them; the role of identity and emotion as accelerants; the attention economy as the channel condition on which all of this plays out; System 1 and System 2 as a defence heuristic in the tradition of Kahneman's Thinking, Fast and Slow; rumour and information voids as the terrain silence hands the operator; and the pause-verify-attribute habit that puts your judgement back on solid ground. Recognising an Influence Operation teaches the observable footprint an outsider actually sees, the discipline of base rate — most odd content is not an operation — what a coordinated push looks like in daily use (timing bursts, unusual seeding, mismatched provenance), the four-question triage, what the signal is not, and what to do in the first hour: pause, screenshot for the record, do not amplify. Modules 5, 6 and 7 walk up the stack. Narratives: How They Spread and Take Hold teaches what a narrative is, how narratives travel (framing, repetition, community fit, all documented), the observable life-cycle a reader can see (seeding, amplification, mainstreaming, aftermath), the signals that a story is being pushed rather than emerging, the published research on counter-narratives and pre-bunking, and why "just correcting the facts" often fails — and what evidence-based communication does instead. The Platforms and the Information Environment teaches the recommendation systems and their engagement incentives at the level the platforms themselves publish, the ads and targeted delivery visible in the public ad libraries, search and video and audio as vectors, messaging apps and closed groups as the layer that is harder to see, the wider media ecosystem — legitimate press, partisan press, faux outlets, wire services — and what the platforms actually publish in their transparency reports. Coordinated Inauthentic Behaviour: What It Looks Like teaches CIB as platforms define and act on it (Meta, X, Google), the markers researchers report (creation patterns, timing, network structure, cross-platform reuse), automation versus coordinated humans versus mixed, what CIB is (a viral hashtag, a partisan movement, a large organic community), the published research pipelines the course names — DFRLab (Atlantic Council), Stanford Internet Observatory, Graphika, Bellingcat, EU DisinfoLab — and why calling CIB is a high-bar claim with a real rate of false positives. set the record and the discipline of naming. teaches what "foreign interference" means in public policy, what public investigators and legislatures have documented at their published level (the US SSCI 2019 and 2020 volumes; the UK ISC "Russia" report; EU External Action and EU DisinfoLab published cases; academic case studies) — never over-claim — the documented role of domestic actors and hire-for-influence firms, the interaction of foreign and domestic amplification, the targeting patterns that recur (elections, diasporas, minorities, protests, crises), and the evidence bar you have to clear before you say "foreign". teaches attribution as a professional discipline whose words carry legal, diplomatic and reputational weight; the evidence types (platform data, technical signals, linguistic and thematic patterns, financial trails); the confidence scale in attribution — high, moderate, low, not known; the contrast between what published reports actually claim and what social media claims; the rival hypotheses (state, state-adjacent, hire-for-influence, domestic, organic) and what the analysis-of-competing-hypotheses discipline looks like when it is done honestly; and what a lay reader is to conclude — which is where the course's attribution ceiling lives. are the shields. builds the reader's own posture: the media diet audit, the pause rule, lateral reading in the SIFT tradition of Mike Caulfield, cross-checking with primary sources, emotional signalling as a warning that verification is needed, and the personal correction habit. runs the Kestrel Motion worked case — crisis communications when a brand is targeted, verifying claims about your own firm before responding, a two-page playbook for a communications lead, employees and executives as vectors and targets, the legal and regulatory posture, and the after-action review. reads the platform layer, the state layer, the press layer and the civil-society layer — election-security agencies, regulators, fact-checkers, standards bodies, verified networks, research organisations, media-literacy programmes — what has worked and what has not on the published evaluations, and the honest trade-offs between free expression, over-blocking, transparency and privacy. , the , reads one campaign end to end — the Halden petition — through the whole discipline: noticing, triage, six-element description, verification against public sources, investigator-level attribution and the personal, organisational and institutional response, compressed into a master equation and a mastery self-assessment. The worked universe is fictional so the method can be shown whole without real classified material, without exposing real sources of ours, and without naming any real living private individual as an operator. Real-world anchors are used only at the level well-documented public reports carry them: the platforms' transparency centres and community-standards enforcement reports; the SSCI and ISC volumes; the archived Mueller Report public volumes; the ODNI ICA (2017); the EU DisinfoLab Indian Chronicles case reports; the archived Cambridge Analytica UK ICO findings; NATO StratCom Centre of Excellence public reports; CISA and CISA's Rumor Control; SIFT (Mike Caulfield); the ISD public reports; and the foundational reading — Kahneman, Sherman Kent, Heuer, Tetlock and the Good Judgment Project, Wardle and Derakhshan's (Council of Europe, 2017), Rid's , and Pomerantsev.

How to read this course

This is a dissection course, not a survey. Every technique is broken down to what an outsider actually sees, what the outsider can verify, and what the outsider must not claim — so you finish not with opinions about the information environment but with a repeatable defensive read you can point at any campaign. Three things to hold onto as you go. First, read the stack bottom-up: when you assess anything in this environment, start with the effect on you, your senses and feed, the platform, the amplifier and the narrative — the layers you can actually check — and go up to vector and actor only as far as the evidence takes you; a reading that starts at the top and works down is the operator's direction, not yours. Second, grade what you say, and say only what you can grade: label every claim as fact, assessment or speculation with its provenance; use only the course's confidence words — high, moderate, low and not known — and its probability yardstick; attribute only at the level a qualified investigator would sign; and when you do not know, say not known rather than reaching for a story that fills the gap. Third, protect, and correct: convey a source's weight without exposing its identity or route, keep private individuals out of the read, refuse to doxx or harass under any framing including your own righteousness, and when a judgement of yours turns out to be wrong, say so first and plainly — a defensive reader who corrects in the open is trusted the next time; a reader who goes quiet is not. One note on scope, because a course on influence operations attracts questions it will not answer. The course teaches recognition, verification, attribution at the investigator's confidence, and defence — for you, your organisation and the public square. It does not teach how to construct or spread a narrative, how to build or operate accounts or networks, how to buy or run persuasion ads, how to seed content to look organic, how to game a platform's systems, how to hire or launder work through an influence firm, how to evade detection, how to build audience psychological profiles, or how to attack any named individual. Where a topic touches those, the material stays at the outside-observable, published-advisory level — what a reader sees and what a published investigator reports, and no further. That line is the discipline; Module 2 makes it explicit before the first layer is read.