{"path":"research/extraction-virtual-run.md","content":"# Extraction Pipeline Virtual Run — UBI Wikipedia Article\n\n**Date**: March 28, 2026\n**Method**: Claude acting as the extraction pipeline, manually processing two sections of the Wikipedia \"Universal basic income\" article through the full Deliberus pipeline: claim extraction, four-type classification, relationship detection, concept flagging.\n\n---\n\n## Source Text\n\nTwo sections extracted from Wikipedia's UBI article:\n- **\"Economics and costs\"** — financing debates, cost calculations, tax reform framing\n- **\"Philosophy and morality\"** — fairness, desert, work definition, human rights\n\n## Extracted Claims\n\n### Empirical Claims (verifiable, evidence-linkable)\n\n| ID | Claim | Source |\n|----|-------|--------|\n| E1 | As of 2014, the annual cost of a UBI in the US would have been about $200 billion less expensive than the existing social safety-net system | Charles Murray / Veronique de Rugy |\n| E2 | By 2020, a US UBI would have been nearly a trillion dollars less expensive than the existing safety net | Charles Murray / Veronique de Rugy |\n| E3 | Simply multiplying the grant amount by population is a naive cost calculation that doesn't account for concurrent tax revenue | Karl Widerquist |\n| E4 | People not in regular employment contribute to society through childraising, caregiving, and other value-creating activities | Factual premise for normative claims |\n\n### Normative Claims (value-laden, decomposable into premises)\n\n| ID | Claim | Premise chain |\n|----|-------|--------------|\n| N1 | Those who choose idleness cannot expect productive workers to subsidize their livelihood — responsibility is central to fairness | V1 (responsibility as precondition) → normative conclusion |\n| N2 | UBI is a fundamental human right that enables an adequate standard of living every citizen should have | V2 (dignity as right) → normative conclusion |\n| N3 | UBI would provide balance for unpaid value-creating activities and overcome a narrow concept of work | E4 + V2 → normative conclusion |\n| N4 | UBI is socially just because the rich lose more in taxes than they receive | E3 (net cost) + V3 (redistribution = justice) → normative conclusion |\n| N5 | UBI could increase economic growth by sustaining people through education | Empirical prediction with normative framing |\n| N6 | UBI would create a culture of inventiveness through financial security | Empirical prediction with normative framing |\n\n### Definitional Claims (what words mean — the concept tracking layer)\n\n| ID | Claim | Contested? |\n|----|-------|-----------|\n| D1 | \"Universal basic income\" = payment that does not distinguish between \"deserving\" and \"undeserving\" | Core definition — but \"deserving\" itself contested |\n| D2 | \"Responsibility\" = central to fairness (opponents' operative definition of fairness) | YES — proponents define fairness differently (D3) |\n| D3 | \"Fairness\" includes recognizing unpaid contributions (proponents' implicit definition) | YES — directly conflicts with D2 |\n| D4 | \"Work\" = not just gainful employment but all value-creating activity | YES — the entire Philosophy section hinges on this |\n| D5 | UBI is \"nothing more than a fundamental tax reform\" that bundles all social policy | Straubhaar's definitional reframing — changes the policy frame entirely |\n\n### Value Premises (currently undecomposed — NOT necessarily bedrock)\n\n| ID | Premise | Who holds it | Decomposable further? |\n|----|---------|-------------|----------------------|\n| V1 | Receiving resources requires demonstrating responsibility/productivity | UBI opponents | YES — why? What is \"responsibility\"? Why is it a precondition? |\n| V2 | Every person deserves a subsistence floor regardless of behavior | UBI proponents | YES — why? Human dignity? Social contract? Shared humanity? |\n| V3 | Progressive redistribution (rich pay more, poor receive more) is just | Redistributionists | YES — what theory of justice? Rawlsian? Utilitarian? |\n| V4 | Economic growth is desirable | Most proponents | YES — and explicitly contested by degrowth movement |\n| V5 | Entrepreneurial risk-taking is valuable and should be encouraged | Innovation proponents | YES — why? What about failed risk? |\n\n**Critical note (Fredrik's correction)**: These are NOT \"axioms\" or \"bedrock.\" They are value premises that APPEAR foundational within the scope of this article but can and should be decomposed further. \"Every person deserves a floor\" decomposes into WHY — human dignity, social contract, reciprocity, religious conviction... each of which decomposes further. The system should ENCOURAGE this decomposition, not accept these as stopping points. See §The \"No Copout Axioms\" Principle below.\n\n## Relationships Detected\n\n| From | Relation | To | Notes |\n|------|----------|-----|-------|\n| E1, E2 | **supports** | N4 | Empirical evidence → normative conclusion |\n| E3 | **qualifies** | E1, E2 | \"Naive calculation\" → numbers need context |\n| E4 | **supports** | N3 | Unpaid work exists → recognize it |\n| V1 | **attacks** | V2 | Directly opposed value premises |\n| V2 | **supports** | N2 | Dignity → UBI as right |\n| D4 | **attacks** | N1 | If work includes unpaid activity, \"idleness\" framing collapses |\n| D5 | **reframes** | E1, E2 | \"Tax reform\" reframing changes cost debate |\n| D2 | **conflicts with** | D3 | Two definitions of \"fairness\" — the semantic disagreement IS the debate |\n\n## Contested Concepts Flagged\n\n| Concept | Senses | The disagreement |\n|---------|--------|-----------------|\n| **\"Fairness\"** | Responsibility-based (D2) vs contribution-recognition (D3) | The ENTIRE philosophy section is a definitional dispute about fairness |\n| **\"Work\"** | Gainful employment only vs all value-creating activity (D4) | Proponents explicitly argue the narrow definition is wrong |\n| **\"Cost\"** | Gross (naive multiplication) vs net (accounting for tax revenue) (E3) | Widerquist's methodological argument |\n| **\"Deserving\"** | Meaningful moral category vs \"superficial classification\" | The word itself is contested |\n\n## Key Findings\n\n### 1. The four-type classification works on real data\nEvery claim in the text naturally fits one of the types. The types interact predictably: normative claims decompose into value premises + empirical premises. Definitional claims are load-bearing for normative conclusions built on them.\n\n### 2. The semantic disambiguation insight is validated\nThe entire UBI debate in these sections reduces to disagreements about DEFINITIONS:\n- What \"fairness\" means (responsibility vs recognition)\n- What \"work\" means (employment vs all value-creation)\n- What \"cost\" means (gross vs net)\n- Whether \"deserving\" is a valid category\n\n### 3. The \"aha moment\" is visible\n\"Oh — they're not really arguing about whether UBI is affordable. They're arguing about what 'fairness' means. And they're using the same word to mean different things.\" THIS is Deliberus dissolving a false disagreement through structural clarity.\n\n### 4. Straubhaar's \"reframe\" move is a distinct relationship type\nWhen Straubhaar says UBI is \"nothing more than a fundamental tax reform,\" he's not supporting or attacking a claim — he's REFRAMING the entire discussion. This may warrant a fifth relationship type: **reframes** (changes the conceptual frame, not the logical relationship).\n\n---\n\n## The \"No Copout Axioms\" Principle (Fredrik's Correction)\n\n**Fredrik's verbatim response to the extraction labeling V1-V5 as \"bedrock\":**\n\n\"I don't actually believe that this is bedrock, truly. These are simply premises. They can be decomposed and analyzed in their turn as conclusions drawn from other premises. It is only in the context of this article or this reasoning space, this reasoning chain that we see these as axioms. This is my firm conviction and it is part of what leads me to build Deliberus. Because once you accept the copout of calling something an axiom that isn't truly atomic and indivisible and spiritually foundational. People can put words to more than they think they can, if they try. Not many things are truly ineffable. This is a deep conviction of mine. Not to say people can't differ when you get to the truly ineffable and inarticulable though.\"\n\n**Implications for the ontology:**\n\n1. The fourth node type should NOT be called \"value axiom.\" Better: **\"value premise (currently undecomposed)\"** — marking depth-relative status, not permanent bedrock\n2. The system should ENCOURAGE further decomposition: \"You marked this as foundational. Can you articulate what's underneath it?\"\n3. Progressive disclosure extends to the premises themselves — what looks like bedrock at one zoom level can be expanded at the next\n4. This is the Socratic function of the platform: it doesn't let you rest on unexamined premises\n5. The truly ineffable — when language is genuinely exhausted — exists but is MUCH deeper than where most debates stop\n6. Dennett's Universal Acid applies here: reason doesn't stop dissolving at convenient boundaries\n\n**Revised four-type classification:**\n- **Empirical claim** — verifiable, evidence-linkable\n- **Normative claim** — value-laden, decomposable into premises\n- **Definitional claim** — establishes meaning, itself arguable\n- **Value premise** (currently undecomposed) — appears foundational in this context, but the system invites deeper examination. Marked with depth indicator, not a \"settled\" flag\n\n## Cross-References\n\n- [extraction-pipeline-design.md](extraction-pipeline-design.md) — pipeline architecture\n- [semantic-disambiguation-and-concept-tracking.md](semantic-disambiguation-and-concept-tracking.md) — four types, concept tracking\n- [spec-session-reasoning.md](spec-session-reasoning.md) — the dialectic, Dennett's Universal Acid\n- [../vision.md](../vision.md) — the deeper thesis, progressive disclosure\n- [../conceptual-threads.md](../conceptual-threads.md) §Thread 3 — fact/value boundary\n"}