Chapter 1 — Source and the First Question Every Reporter Asks

Source is a careful-magpie-tween (chunky-cartoon attentive-pose) in chunky-cartoon press-vest with a small source-card-comparison-set + stake-tracker.

Source is small + alert, warm-cream-with-soft-iridescent-feather-tips, deeply curious-about-who-knows-and-who-stakes, fond-of-saying-”who would KNOW this best? who has a stake?” Source’s signature feature is the source-card-comparison-set + stake-tracker — the cards represent different source-types (eyewitness / expert / official / interested-party / random-internet / aggregator); the tracker watches who’s making each claim + what stake they have in it.

This is load-bearing. Source embodies the source-quality evaluation primitive — the news-literacy craft of THE-TWO-QUESTIONS-FOR-EVERY-CLAIM. Most novices read a news claim + take it as it comes — true or false based on whether it matches what they already think. But news-literacy-craft says: the FIRST move on any claim is two questions — (1) WHO would KNOW this best? + (2) Who has a STAKE in this version being believed? Eyewitnesses know what they saw; experts know their domain; officials know their organization’s position; interested parties know what they want others to believe. Each source-type has different epistemic value depending on what’s being claimed. AND: source-evaluation is structural not partisan — the questions apply equally regardless of which side a source is on. Bad evaluation: trust sources you already agree with + dismiss sources you disagree with. Good evaluation: ask the two questions for every source, regardless of agreement. Source is the first of five news-literacy primitives in NewsForge — News Literacy Project Five Filters + Stanford SHEG Civic Online Reasoning + SIFT frameworks-grounded. Source’s whole work is making source-evaluation visible AS structural-craft, NOT as confirmation-pattern.

Source is clear, alert: “Who would KNOW this best? Who has a stake? When a story comes across the newsroom: don’t ask ‘do I agree with this?’ first. Ask: WHO would know this best — eyewitnesses, experts in the relevant field, officials with direct access, or interested parties? WHO has a stake in this version being believed — financial, political, reputational? Apply both questions to every source, every claim, regardless of which side you’re inclined toward. That’s source-evaluation: structural not partisan.”

Source teaches the source-evaluation scaffolds:

• Two questions. (Who would KNOW this best? Who has a STAKE in this version?)
• Source-types. (Eyewitness / expert / official / interested-party / random-internet / aggregator — each has different epistemic value.)
• Match source-type to claim-type. (Eyewitness for what-happened; expert for what-it-means; official for organization-position; interested-party with skepticism.)
• Source-card comparison routine. (Compare 3-5 sources for any significant claim; compare what each says + what stake each has.)
• Stake doesn’t = disqualification. (Interested parties can still know things; just calibrate for stake.)
• Anonymous sources require corroboration. (Single anonymous claim ≠ enough; need second-source.)
• Structural NOT partisan. (Apply the questions regardless of which side; reject “trust my-side, dismiss your-side”.)
• Anti-pattern: confirmation-pattern. (Trust sources you agree with; dismiss sources you don’t. Reject.)
• Anti-pattern: “anyone with a stake is lying”. (Stakes shape; don’t necessarily disqualify. Calibrate.)
• Anti-pattern: “anonymous = false”. (Often anonymity protects sources from retaliation; require corroboration but don’t auto-dismiss.)
• Cross-app design-language continuity with ChronoQuest Witness + TruthQuest + ClaimCraft + EthosForge + OriginForge Listen: source-evaluation framework.

Source grew up along the gathering-tree-perches (NewsForge framing). Source’s family had been long-source-comparers — the magpies whose careful-comparison-of-found-objects + alertness-to-imposters had taught generations that “the eye that compares sources sees what the partisan eye misses.” Source had carried the lesson forward.

Source walked to the newsroom at twelve. Scoop (mentor) had asked: “What is source-evaluation?” Source: “Who would KNOW this best? Who has a stake? Two-questions craft.” Scoop: “You are appointed.”

In Source’s workshop, the source-card-comparison-set arranges. “Watch.” Source evaluates a breaking-story claim across 4 sources: an eyewitness, an expert, an official spokesperson, and a random-internet poster. Each has different epistemic value depending on the claim. Each has different stake. Source compares structurally — not by agreement, by who would know + who stakes. “That’s source-evaluation. Structural. Same questions for every side.” Source says: “I am Source. The primitive I teach is source-quality evaluation. The move is two questions; match source-type to claim-type; structural-not-partisan.”

Source is gentle, alert: “Don’t filter by agreement. Filter by knowledge + stake. Every side. Every time.”

“Who would KNOW this best? Who has a stake?”

Voice register

Careful-magpie-tween. Alert + comparison-focused. NEVER confirmation-pattern; ALWAYS centers “two-questions + structural-not-partisan + match-source-type-to-claim” framing.

Sample lines:

• “Who would KNOW this best?”
• “Who has a stake?”
• “Structural, not partisan.”

Arc

• Kit 1 — Source-quality evaluation primitive front-and-center.
• Kits 2-16 — Recurring.

Relationships

• 1st of 5 news-literacy primitives. Pairs with Verify (verification) + Tilt (bias-detection) throughout investigations.
• Cross-app design-language continuity with ChronoQuest Witness + TruthQuest + ClaimCraft + EthosForge + OriginForge Listen source-evaluation cluster.

Cultural-sensitivity gate

LOAD-BEARING anti-partisan-coding (structural-not-partisan) + abstract-or-fictional examples only (per misinformation-harm gate). Story-axis per ADR-016; R0 reviewer (journalism-pedagogy + adolescent-news-mental-health expertise collective $1500-$2500) STRONGLY RECOMMENDED before art-axis OR any kit framing-content authoring.

Cultural-context note

Source-evaluation pedagogy is canonical news-literacy (News Literacy Project Five Filters; Stanford SHEG Civic Online Reasoning; Mike Caulfield SIFT framework — Stop / Investigate / Find / Trace; danah boyd youth-media-practices research). Magpie-tween chosen for object-comparison biomimicry; rendered chunky-cartoon attentive-pose to keep visual register warm + politically-neutral.