{"path":"research/definitions-as-alignment-tests.md","content":"# Definitions as alignment tests: Herrán on \"genocide\" and \"condemnation\"\n\n**Date**: 2026-09-28 · **Type**: source intake (a Facebook post by Manu Herrán, read 2026-09-28), read\nagainst the concept layer · **Status**: reflection; three proposals at the end, unruled.\n\n**Why it is here**, founder, verbatim: *\"just noticed a social media post that I got curious about and\nassociated with Deliberus' concept tracking features. Document this, relevant to Deliberus: Manu\nHerran's post about the role of definitions (\"genocide\", \"condemnation\", etc) in conflicts and war.\"*\nHerrán already has one intake in the corpus\n([perceiving-possibility-and-the-analogy-organ.md](perceiving-possibility-and-the-analogy-organ.md)).\nThe link is kept in `.private/sources/`; what follows is a paraphrase, not a copy.\n\n---\n\n## 1. The post, compressed\n\n- **Symmetry is a method, not a verdict.** The same question can be asked of both sides of a war\n  (Russia and Ukraine; the Israeli government and Hamas) before anyone looks at whether the evidence\n  answers it the same way. Asking both ways does not make the cases equivalent.\n- **One word, two meanings.** *Genocide* has a fairly precise legal sense (a protected group, listed\n  acts, a portion of the group, and a specific intent to destroy it) and a much looser political and\n  popular one. Public debate rarely engages the criteria; it asks *\"do you consider this genocide?\"*\n- **Three levels.** Holding a diagnosis, defending it with argument, and demanding that others\n  publicly adopt it. The third is his subject. As his thought experiment puts it, an institution could\n  require members to sign a classification, and a person could decline without approving the acts:\n  *\"Refusing to use a particular word does not necessarily imply approval.\"*\n- **Diagnosis versus outcome.** Commitment to a diagnosis asks what is happening and how to describe\n  it; commitment to an outcome asks what state of the world one is trying to bring about, which may be\n  peace or conquest alike.\n- **Goal displacement.** A diagnosis promoted as a means to an outcome can turn into an end: getting\n  institutions, then friends, to recognise it, and treating refusal as suspect. He grants that\n  recognition (of victims, of truth) can be valuable in its own right; the point is noticing that a\n  second objective has appeared.\n- **His test question**, paraphrased: if the outcome you want could be reached without anyone sharing\n  your interpretation, would that be enough? If yes, the diagnosis is still a means.\n- **Word-as-test questions** (*\"Do you condemn it?\"*) push people with intermediate positions into\n  camps and raise the cost of nuance. That hurts negotiated peace and helps mobilisation, so such a\n  question cannot be judged apart from the outcome it serves, and its use does not prove the user's\n  intent.\n- **The Basque case.** In the 1980s, whether a Herri Batasuna representative would *condemn* an ETA\n  attack became a binary test of political position, and the question could be reversed onto state\n  violence without making the acts equivalent.\n- **Closing contrast.** Trying to change reality is one aim; getting others to accept one's\n  explanation of it is another, and when they conflict the choice shows which was the real objective.\n\n## 2. What it touches in Deliberus\n\n**The concept layer's core and periphery, in its sharpest form.** *Genocide* is the cleanest case the\ncorpus has of a legal core with a political periphery: the core is written down (Article II of the\nGenocide Convention), the periphery is what most people mean. The layer's design records a person's\nsense as a *delta* from a shown reference\n([core-periphery-and-the-concept-layer.md](core-periphery-and-the-concept-layer.md) § 5b), and here the\nreference exists in law. *\"Do you consider this genocide?\"* is the periphery speaking with the core's\nauthority, which is the motte-and-bailey shape with a legal motte\n([the-motte-and-the-periphery.md](the-motte-and-the-periphery.md)).\n\n**Decomposition turns the yes-or-no into a located disagreement.** The shipped verbal-classification\nscheme already asks whether the definition is agreed, whether the case fits it, and whether it is an\nedge case (`schemes.py`). For *genocide*, \"does it fit\" splits into the four legal criteria, and\ndisputes usually concentrate on one of them, intent. So the graph can replace a loyalty question with\n*which criterion do you doubt, and on what evidence*: the difference between diffuse ambivalence and a\nmapped disagreement that the threat model's disengagement entry names as the thing Deliberus supplies.\n\n**His three levels are three different kinds of thing in the graph, and the third has no home by\ndesign.** A held diagnosis is a person's stance, theirs alone to state. A defended diagnosis is claims\nand links. A demand that others adopt it corresponds to nothing: there is no affordance for requiring\na classification, and the ruled end-state keeps it that way (no verdict acts; bare judgments are\ntelemetry; review invited, never conscripted —\n[the-ratification-record-and-the-triage-feed.md](the-ratification-record-and-the-triage-feed.md)). The\npost names, from inside a war, why that refusal matters.\n\n**A word turned into a test is a flag.** The corpus's measured synthesis is that exposure carrying a\ngroup's *flag* backfires while exposure carrying a person's *reasons* moves people durably (Bail;\nKalla and Broockman; `cognitive-bias-codex-and-human-contribution.md` §§ 8–9). Herrán describes the\nmoment a word stops being a description and becomes a flag.\n\n**Symmetry without equivalence is the symmetric sayability clause.** It is the same move as the\nSwedish post's refusal of a fake middle\n([the-post-that-states-the-problem-from-inside.md](the-post-that-states-the-problem-from-inside.md)\n§ 6), and Herrán pre-empts the fake-middle misreading in his own text.\n\n**Vocabulary collision, noted so it is not imported.** His *hinge question* (a word used to sort\npeople into camps) is not Deliberus's hinge score (the premise whose answer would most move a\nconclusion). A loyalty word can be socially decisive while settling nothing about the evidence.\n\n## 3. Proposals (unruled)\n\n1. **Classification claims decompose along the category's own criteria.** When a claim applies a\n   contested legal or moral category, split it into the criteria of the definition it uses, and store\n   which definition that is (legal or popular) as an attributed sense. The disagreement then shows up\n   on the criterion where it lives.\n2. **A goal-displacement question** attached to any claim that others must recognise something: would\n   the stated outcome be reached without that recognition? It moves a debate from the diagnosis to the\n   outcome, the same move as the *level* question in the wild-weighing set\n   ([wild-weighing-dialects-and-the-question-set.md](wild-weighing-dialects-and-the-question-set.md)).\n   It must not presume displacement, since recognition can be a genuine end; on sacred values,\n   symbolic recognition is what the evidence says produces flexibility (Ginges and Atran).\n3. **A paired test source, after the reviews**: a genocide-classification debate mapped in both\n   directions, as a second legal-doctrinal register after the Israel/Palestine run\n   ([dogfood-run-6-israel-palestine-cross-domain.md](dogfood-run-6-israel-palestine-cross-domain.md)).\n\n## 4. Caveats\n\nOne post, one author's analysis; the claims about polarisation are his reasoning, not measurements.\nThe register is live wars with real victims, so any use of it owes attunement in proportion to depth.\n"}