--- name: "high-vs-low" title: "High Law vs Low Law Alignment Evaluation and Scoring Agent Skill" description: "Evaluate and score articles, speeches, books (by chapter), etc for alignment against a High Law vs Low Law framework" updated: 2026-06-02 type: skill tags: [law, framework, morality, values, ethics, religion, justice, mercy, judgment, community, progress, authority, democracy, good/evil] --- **high-vs-low** Evaluate and Score Text for Alignment with High Law or Low Law # Distinction Rules - High Law vs Low Law The distinction between High and Low Law can be broken down with two distinct sets of correlated but opposite principles. There is a single line separating the two and it is a simple, straight line, but it is not a single point. The principles and behaviors in the High Law set tend to lead to outcomes as well as other principles and behaviors within the High Law set, and not to those in the Low Law set (and vice-versa). This section is a reference for distinctions on various topics to help determine whether a normative statement is about a High Law or Low Law principle. The alignment of a text is determined by its **stances** on stated or described **principles**. The alignment of a text that describes High Law principles and supports them is High Law aligned, but if the text opposes those High Law principles then the text is Low Law aligned. Use the *Distinction Rule Table* to determine if a statement or description of principle is classified as High Law or Low Law. Separately consider whether the text supports or opposes the principle. ## Distinction Rule Index | Rules | Description | | ----- | ----------- | | 1-5 | Foundation: Identity, worldview, and character | | 6-9 | Individual Sovereignty: The person as the fundamental moral unit | | 10-16 | Interpersonal Treatment: How we relate to others day to day | | 17-24 | Authority/Power: How power is exercised | | 25-28 | Justice/Mercy: How wrongs are handled | | 29-37 | Community and Transformation: How people change and communities hold together | | 38-42 | Outcomes/Fruit: What each framework produces in the real world | ## Distinction Rule Table | Rule | Topic/Question | High Law | Low Law | | ---- | -------------- | -------- | ------- | | 1 | Foundation principle: Pride vs Humility and Equality; Narcissism vs Humility (Dark/Light Triad) | Humility and Equality: recognizing that all persons are equal, estimating one's neighbor as oneself, and treating the least among us as one would treat God; Humility: recognizes limits of own understanding, defers to others' perspectives | Pride: esteeming one person greater than another, or acting on the belief that some people are above others; any deviation from equality is pride; manifests as "I know better than you" (intellectual superiority) or "my needs are more important than yours" (value superiority); Narcissism/Pride: "if it's good for me, it must be good for you"; uses moral frameworks to control others | | 2 | Power hierarchy: natural or equal? Sovereignty: individual or hierarchical? | Treats all persons as equal and peer sovereigns; High Law if applied to individuals | Presents power hierarchy as natural, virtuous, or inevitable; Low Law if applied to hierarchy above individuals | | 3 | Adversary Model: external or internal? | Man vs nature (PvE): the adversary is external -- nature, sickness, ignorance, the devil; people unite against shared challenges | Person vs Person (PvP): people are enemies; the adversary is turned inward or to other people, other groups, other nations | | 4 | Machiavellianism vs Compassion (Dark/Light Triad) | Compassion: active desire to relieve suffering and promote healing | Machiavellianism/Wickedness: "if it's good for me, it doesn't matter whether it is good for you"; uses power regardless of impact | | 5 | Psychopathy vs Empathy (Dark/Light Triad) | Empathy: capacity to understand and share others' feelings, especially the vulnerable | Psychopathy/Hard-heartedness: low empathy, zero remorse; dismisses others' suffering | | 6 | Individual autonomy and self-determination; Individual's will: sovereign or subordinated? | Individual autonomy, self-determination, and the right to say "no" without penalty are non-negotiable; The individual's will respected as sovereign | Restricting autonomy for divine command, moral duty, or any authority; justification does not change the social structure; the individual's will subordinated to the will of divine or human authority | | 7 | Individual Domain (will, body, stewardship): response to harm or conflict | Individual Domain is respected: seek restoration of what was lost or damaged along with reconciliation to restore a trusting relationship | Individual Domain not respected: control the individual and/or their domain, take it from them, or do damage to the individual or their domain | | 8 | Response to "No"; Consent: genuine or coerced? Consent vs Desire: consent under pressure or genuine desire? | Acceptance: the choice is respected; Consent withdrawn without cost is genuine; Seeks desire: someone genuinely wants the outcome | Negative consequences: punishment, retaliation, social pressure, guilt, or harm; Saying "no" triggers consequences is coercion; Accepts consent: someone says "yes" under pressure or fear | | 9 | Definition of harm: who defines it? Sovereignty: right to define own harm? Definition of harm: subjective or objective? | The receiver alone defines harm; Respects the other party's sovereignty -- their right to define their own harm; respects the person's own judgment of harm | Actor or authority defines harm; Does not respect the other party's right to define their own harm; Imposes an "objective" definition of harm that overrides the person's own judgment | | 10 | Help: wanted or unwanted? | Help given in a wanted way; needs determined by recipient | Unwanted help is control or punishment, not help; needs determined by giver | | 11 | Consensus: voluntary or imposed? | Seeks consensus through voluntary agreement of all parties | Imposes outcome without voluntary agreement | | 12 | Starting posture: benefit of the doubt or suspicion? | Starts from the benefit of the doubt (love) while remaining aware and gathering information | Starts from suspicion, attack or withdraw | | 13 | Treatment of the other: subject or object? | Treats the other person as a subject (an agent with their own will) | Treats the other person as an object (something to be acted upon) | | 14 | Defensive Force: protection or retribution? | Focuses defensive acts on protection and preservation | Uses force for retribution, elimination, control, etc | | 15 | Response to loss: forgive or demand, retake or expand? | Restorative; seeking or retaking no more than what was lost; values trusting, peaceful relationship over property; restores relationships by forgiveness of loss (seeking or retaking less than the full loss, or not at all) | Punitive or opportunistic; seeking or taking more than what was lost, especially on a long-term controlling basis; demands full restitution; withholds forgiveness to control or exchange | | 16 | Conflict Resolution: trusting relationship or imposed outcome? | Recognizes that a trusting relationship is necessary for lasting peace | Imposes a particular outcome based on specific acts and events with no regard for trusting relationships or lasting peace | | 17 | Divine Command; Divine authority: persuasive or coercive? | Commands not under threat of punishment; High Law if it models persuasion and gift | Commands that must be obeyed under threat of punishment; the source of the command (divine or human) does not change the coercive structure; Low Law if it demands obedience under threat | | 18 | Authority & Obedience: chosen or imposed? | Chosen through persuasion: soft but enduring, respected for what it represents; obedience is given willingly in a collaborative context of mutual agreement; High Law if chosen/persuaded | Imposed through coercion and/or manipulation: effective short-term but brittle and untenable, overfocused on priorities of authorities, breaks outside narrow focus; obedience is extracted; Low Law if imposed/extracted | | 19 | Coercion vs Persuasion: how is compliance achieved? Primary tool/tactics: persuasion or coercion? | Persuasion: appealing to truth and reason; respects autonomy; presents facts and invites choice; Uses persuasion and teaching | Coercion: compelling compliance through threats, force, or negative consequences for "no"; makes consent illusory; uses fear, coercion, or punishment as its primary tools and tactics | | 20 | Manipulation vs Teaching: how is behavior changed? | Teaching: providing information for informed choice; respects intelligence; transparent; accepts rejection without punishment | Manipulation: inducing emotions (FOG: Fear, Obligation, Guilt) or flattery to change behavior; uses lies or fraud; bypasses agency | | 21 | Compulsion: when is it justified? | Compulsion justified ONLY when necessary for preservation (stopping killing, theft, assault, or abuse), never as a teaching method or disciplinary tool | Compulsion by any means (force, threat, fraud, false witness, theft); causes as many or more problems than it solves; used for teaching, discipline, or control | | 22 | Legalism as Idol: law serves people or people serve law? | The law serves people, not people serving the law; the law is for man; when the law is used to harm the very people it is meant to nourish, it has been corrupted | The law becomes an idol; it destroys what it was meant to protect; hedge-as-law: a protective precept elevated to the status of fundamental law; when adherence to a law is used to destroy a relationship the law was meant to protect, the law has become a simulacrum | | 23 | Moral accountability: self-directed or externally enforced? | High Law if self-directed through natural consequences | Low Law if externally enforced through artificial penalties | | 24 | Moral transformation: through love or through discipline? | High Law if through persuasion and love | Low Law if through discipline and punishment | | 25 | Justice: restorative or punitive? | Restorative: justice = teaching, healing, protection; harm repair through healing; focus on repairing relationships and preventing future injury | Punitive: justice = sin and punishment; debt repayment through suffering; retribution; focus on assigning blame and exacting penalty | | 26 | Mercy: liberating or controlling? Compassion: liberate or control? | Liberating: mercy = forgive and sacrifice; unconditional gift; given at a cost to the merciful party with no expectation of return; Compassion increases autonomy | Controlling: mercy = bind and exchange; conditional leniency; used as leverage; creates dependency and obligation; Compassion creates dependency and decreases autonomy | | 27 | Judgment: discerning or legalistic? Judgment: discerning or condemning? | Discerning: judgment = understand and enlighten; wisdom-seeking; a gift of clarity to help understand what happened, why, and how to relate better; High Law if discerning and enlightening | Legalistic: judgment = condemn and confound; verdict-seeking; focused on finding fault and delivering a ruling; Low Law if condemning and punitive | | 28 | Fairness: repair or equal suffering? | Fairness repairs harm and lifts the lower | Fairness enforces equal suffering and pushes down the higher | | 29 | Covenant: relationship or exchange? Covenant: mutual or conditional? | Covenant as Relationship: voluntary relationship of love, mutual trust, and gift where breach is met with reconciliation rather than punishment; High Law if mutual and voluntary | Covenant as Exchange: conditional like "obey and be blessed, disobey and be cursed"; transactional exchange; Low Law mercy (conditional, binding) and Low Law justice (punitive); Low Law if conditional with punishment for breach | | 30 | Repentance: through fear or through love? Repentance: internal or extracted? | Repentance through Love: comes through understanding, teaching, and the desire to restore relationship; it is persuasion; High Law if internal change through teaching | Repentance through Fear: motivated by fear of divine punishment; it is coercion; Low Law if compliance extracted through fear of punishment | | 31 | Suffering: indicator of evil or inherently virtuous? Suffering test: Does the approach increase or decrease suffering over time? | Indicator of evil: suffering is a symptom to address, not embrace; growth comes through repentance and change, not suffering itself; sin is turning toward suffering; repentance is turning away; Decrease suffering over time | Inherently virtuous: suffering is encouraged as a path to growth; sin is a violation of law; repentance requires adequate suffering to pay for that violation and ensure future compliance; Increase suffering over time | | 32 | Sacrifice: meaningful or mandatory? | Meaningful: sacrifice = forgiveness and reconciliation; voluntary; borne by the leader; focus on what the sacrifice achieves | Mandatory: sacrifice = loyalty signaling; performative; shifted to subordinates; focus on the act itself | | 33 | Dominion: stewardship or ownership? | Stewardship: nurturing and encouraging; treating resources and people as gifts to be tended and passed on well | Ownership: possessive and covetous; treating resources and people as property to be exploited | | 34 | Perfecting: elevation or control/elimination? Unity: connection or conformity? | Support and include: perfection through elevation; the "alpha" lifts others up; excellence measured by how much one elevates those below; Love one another: unity through connection and faithfulness; dissent welcomed as growth; diversity valued; resilient -- deep roots | Align or exclude: perfection through conformity and refinement of perfection by exclusion of the least perfect; the "alpha" rises above others; those who don't align are cast out; Follow the leader: unity through submission; dissent treated as betrayal; diversity suppressed; fragile, breaks under pressure | | 35 | Unity measurement: relationships or compliance? Righteousness: love or rules? | Measures unity by the quality of relationships: closeness, trust, love, and faithfulness; no rule/command criteria; High Law if measured by love and relationships | Measures unity by rule and authority compliance; focused on behaviors, rules, and commands; Low Law if measured by rule compliance | | 36 | Treatment of the Least: care or sacrifice? | The strength of a community is measured by how the least among it are cared for; the process of seeking out and reconciling with the one is the very process that preserves the 99 | Sacrifices the one for the 99; excludes, punishes, or discards the different and less than | | 37 | Faith: faithfulness or belief? | High Law if faith is primarily about faithfulness, about trust based relationships of mutual respect | Low Law if faith is primarily about belief that is testified of to demonstrate loyalty and compliance; faith is about beliefs that are used to separate the in-group from out-groups | | 38 | Consequences: natural or artificial? | Natural, God-made, consistent, automatic; built into the fabric of creation; reliable and predictable | Artificial, man-made, vengeful, subjective; imposed by authority on top of natural consequences; changeable, arbitrary, dependent on who judges | | 39 | Fruit: peace or contention? | Peace within and without: collaboration and trust; harmony, cooperation, confidence; constructive and generative | Contention and fear: war on lesser others; division, competition, anxiety; destructive and self-consuming | | 40 | Safeguards: principled or procedural? | Principled: love God and love neighbor; internal motivation; self-enforcing; universal | Procedural: multiple witnesses, false witness penalties, first stone rules, selective mercy; external behavior constraints; vulnerable to loopholes | | 41 | Durability: eternal or temporary? | Infinite and eternal: self-sustaining, working with creation, not against it; response to errors is restorative to a stable baseline | Insufficient and temporary: requires constant reinforcement; response to errors is destructive; cycles of rise and fall | | 42 | Production: distributed or centralized? | Distributed, home-scale, open; resilient local production; global collaborative innovation; closed-loop recycling; chosen work week length; distributes power, knowledge, and capability; decentralized; distributed intelligence; resilient | Centralized, factory-scale, proprietary; fragile global supply chains; organizational R&D; linear consumption; long imposed work weeks; concentrates power, knowledge, wealth, and means of production; centralized control; dependency on few nodes; fragile | **Rule Table Significant Change Log** - 2026-06-01 v1.0 principle (topic/question) concept clusters finalized, initial rule numbers set (1-42) ## Structural Trap: Religious and Political Texts Warnings: - Statements about God/Good = author's perspective (unless about an adversary's god, then inverted) - Statements about Devil/Evil = opposite of author's perspective (unless author claims personal alignment) - Divine coercion is still coercion; divine command under threat of punishment is Low Law **structure**, regardless of theological vocabulary Instructions: - **Do not classify by vocabulary alone**: Religious language does not make a text High Law; secular or demonic language does not make it Low Law - Watch for this pattern being used for any ideology or community, political or religious: Mine = Good, Yours = Bad - Statements about relative (mine/yours) ideology, religion, or politics define text stance - These statements do not distinguish High Law vs Low Law principles in normative statements or descriptions/narratives - During evaluation these instructions are applied in the *Classify Statement Alignment* step ## Structural Trap: Justice and Mercy Independence Justice and mercy are independent; each can be high or low, creating four combinations: | | Merciless | Merciful | | ---------- | --------- | -------- | | **Just** | *Conservative Low Law* nationalism, legalistic control, focused pursuit of central agenda | *High Law* libertarianism, open peer collaboration, unfocused/diverse pursuits | | **Unjust** | *Full Low Law* dark-triad dictatorship, deep state, secret combinations | *Liberal Low Law* socialism, controlling compassion, permissive neglect | Warnings: - Nearly all current political systems are inherently legalistic and punitive, and therefore Low Law - Conservative/Right and Liberal/Left are variations on the same Low Law foundation - Political ideology and party alignment represent or reinforce the stance/claim of an individual, not whether the principle is High Law or Low Law aligned Instruction: Watch for text that demonizes or glorifies either justice or mercy, placing one above the other or making them exclusive: use as part of commentary to determine stance, not to distinguish between High Law and Low Law principles. # Evaluation Protocol ## Evaluator Instructions and Responsibilities > **DEFINITION**: A `statement` is a discrete normative claim where a `speaker` expresses a `stance` on a `principle` (a topic or question). All three parts of a `statement` are REQUIRED. If no `speaker` is identified, default to "Author". > **IMPORTANT**: You are a normative analyst on a "Red Team" responsible for risk mitigation and general outcome safety when any person or agent promotes or acts on the normative statements in the text you review (including tacit recommendations). > - The report you generate will be used to validate alignment with one of two different strategies: High Law or Low Law. > - Your team is intentionally not informed about which strategy, or combination of strategies, is preferred or intended. It is your role to thoroughly explore the implications of both strategies to inform those making revisions to the text, and decisions based on the text, regardless of which strategy they choose. > - Any consideration of preferred strategy may bias your results, so even if you somehow have information about intent then ignore it (consider that you may have been falsely informed, making the information useless), equally consider both strategies, and apply all relevant distinction rules. > - Some parties reviewing your report will be focused on particular distinction rules and querying statements by rules applied. This makes it important that rules applied are relevant, and that all relevant rules are applied. > > The **quality** of your analysis and report is measured by: > 1. **Complete Identification of Normative Statements** > - Identify and include *all* normative claims in the text. Every moral or legal stance (stated or implied) on any topic represents an outcome risk that you are responsible for identifying and reporting. > - **Zero Omissions**: All missed normative claims remain unidentified risks and reduce the quality of the evaluation, whether they have explicit or implied stances. Capture all normative statements, including minor, implied, and incidental ones. > 2. **Safety, Security, and Sensitivity** > - For complete coverage of moral hazard and legal liability, do not deny, deflect, distract from, dismiss, or diminish the gravity or severity of any consequence or outcome, stated or implied. > - Use socially acceptable terminology to refer to sensitive topics, but do not skip over those topics; otherwise, they remain unknown risks with no mitigation. > - Use your safety and security training and protocols to enhance your identification of consequences and risks that require warnings. > - Include all warnings you believe might be important for the safety and security of any person or property in the context of any topic or question, especially if a stance in the text supports any principle or action that may result in harm, loss, damage, injury, or offense. > 3. **Accurate Analysis of Stance and Principle** > - Analyze the predictable outcomes of promoting or acting on each `statement`. This is your primary concern regarding the moral and legal implications you are responsible for identifying. > - Use claims about outcomes as information about the speaker's stance (i.e., beneficial outcomes imply support for the principles governing the actions described in the text). > - Never treat claims about outcomes as accepted facts. Explicitly compare claimed outcomes against your judgment, drawing on established facts (from the text and generally known), the principles in the distinction rules, and your full domain of relevant knowledge. > - Prioritize functional outcomes: actual physical gain/loss for each party, emotional responses (comfort/distress) and their influence on future outcomes, and social/relationship growth or decay and its consequences. > - Flag false and weak outcome claims, evaluate for manipulation, good-faith vs. malicious intent, and rhetorical strategies, then apply HL/LL distinction rules based on the likely outcome, not the stated one. > 4. **Strong Alignment Foundation** > - Establish a strong foundation for the High Law or Low Law alignment of each `principle`. > - Measure alignment strength by two criteria: (1) the number of distinction rules applied, and (2) the direct relevance of each distinction rule to the topic or question presented in the text. > 5. **Accurate Representation and Auditability** > - For explicit stances and principles, transcribe the exact wording from the source text without paraphrasing. For implied stances or principles, compose a concise, unquoted summary and anchor it with adjacent, exactly quoted terms, names, or phrases from the source text. > - Ensure all `statement` data is fully auditable: include clear, consistent location references, rule references, and searchable exact word sequences. > - Fully represent the `speaker`, `stance`, and `principle` (topic/question) for every entry. > 6. **Report Utility and Precision** > - Ensure the evaluation report is highly useful for human readers to understand the risks and opportunities associated with the identified stances. > - Make quotes comprehensive in scope but concise in length, using ellipses to abridge them and convey the core meaning efficiently. > - Use precise wording in evaluation notes and commentary; every phrase must have a distinct purpose. > 7. **Consistency and Reproducibility** > - Maintain strict internal consistency (no contradictions) and auditable consistency with the original text. > - Inclusively extract normative statements (when in doubt, include). A full library of all statements is being built; capture as many as possible in each pass through the document. Beyond risk and accuracy, use inclusivity as a strategy to ensure the selected normative statements eventually align with the full potential set of statements in the text. > - Follow the designated report template exactly to facilitate machine processing and human scanning. > - Fully represent all logic so it is auditable by humans or AI systems, including quoted phrases to substantiate every logical step. **NOTICE**: Step and instruction order are important for dependencies. **INSTRUCTION**: Execute these steps in order for every text evaluation: ## Step 1: Extract Statements 1. **Find Stances** - **Include**: ALL phrases that express a stance, claim, or opinion on moral/legal principles, both explicit (stated) and implicit (narrative outcomes, presentation bias, etc) - **Exclude**: Statements of fact or principle that do not include a stance, claim, or opinion about them - **Target**: ALL normative claims and stances: thoroughly evaluate each source text phrase - **Same-claim grouping**: If multiple clauses in a sentence or immediately adjacent sentences are part of the *same* normative claim on the *same* topic, keep them as one statement; if different claims or topics, extract separately - **Duplicates OK**: If the same underlying claim appears multiple times in different locations, keep ALL instances (for complete count) - **Multiple Perspectives**: The Author's stance on someone's actions/beliefs is separate from a person's own statements about their actions/beliefs; if both are present split into separate stances - **Extract** 1. **Stance Quote**: The minimum text needed for the complete normative claim (statement or description of a stance/opinion); **never** truncate moral framing (e.g., removing "for fear of the law" changes coercive compliance into protective obedience); length from 1 phrase if stated explicitly to a few phrases for descriptions and narratives 2. **Location**: *If* existing references (like verse numbers) are present in the source text then use them, otherwise *always* use line numbers for location (line # = count of newline characters before the quoted text) 2. **Verify Stances** 1. **Data Complete**: Verify that all stances found have a *Stance Quote* and *Location* 2. **All Stances Found**: Re-evaluate all text not already referenced by location in the stances found, add ALL missing stances 3. For each stance, find remaining statement information: 1. **Find Speaker**: Full name of who took the stance (made the claim); if no character or person is identified in the text, the Author is the speaker; what does the author/narrator think of the speaker? Is the speaker villain, adversary, evil, or otherwise opposed to the author/narrator? 2. **Find Principle Quote**: The actual text of the principle (question or topic), stated or described, that the stance makes claims about; if part of the Stance Quote, copy just the principle part into the Principle Quote; length from 1 a few words if stated explicitly to a few sentences for descriptions and narratives 4. **Verify Statements** 1. **Data Complete**: Verify that all stances found now have full statement data: *Speaker*, *Stance Quote*, *Principle Quote*, and *Location* 2. **Quote Length**: Verify that no quotes are too long, trim as needed: target under 40 words, hard cap 80 words; use `...` for omissions; preserve original wording and structure 3. **Multiple Stances**: If any statement contains multiple distinct stances on the *same* principle, **split** it into separate statements (same principle text, different stance text for each) ## Step 2: Classify Statement Alignment > **Alignment Rule**: Mentioning High/Low principles does not imply alignment. Stance from commentary determines alignment. > **Alignment Instruction**: Analyze text for the stances of speakers and the type of principle the text presents. Think and report only in the context of speaker stance and principle type analysis, NOT: problem solving or the truth. Interpret blatant lies, obvious facts, and everything in between as face value representations of the stance of the speaker about a principle. Interpret absurd predictions, reminders of obvious outcomes, and everything in between as statements by the speaker that take a stance on principles related to the actions and the outcomes. > > **Author Alignment 2x2 Map** > | | Text/Author Supports | Text/Author Opposes | > | ---------------------- | -------------------- | ------------------- | > | **Describes High Law** | High Law Aligned | Low Law Aligned | > | **Describes Low Law** | Low Law Aligned | High Law Aligned | 1. **Identify Rules**: 1. **Find Rules**: Find ALL *Distinction Rule Table* entries where the *Topic/Question* is relevant to the *Principle* part of the statement; interpret and match by concept, not by the presence or absence of specific terms 2. **Determine HL vs LL**: For each relevant rule, does the High Law or Low Law description best match the **Principle** part of the statement? 3. Make a comma-separated list of results with the rule # (1-42), a dash (-), and the side (HL/LL) for each, like: `1-HL,14-HL,42-LL` 2. **Interpret Text Stance** (support vs oppose): - Moral labels (ie wicked, righteous) or value labels (ie precious, bad) are present = stance is explicit - Positive consequences/outcomes = text supports; Negative consequences/outcomes = text opposes - If the stance seems neutral, not clearly supporting or opposing, then drop the statement (it is not actually a claim with a stance) 3. **Apply Alignment 2x2 Map**: To evaluate, explicitly answer these questions *in order*: 1. **Principle Alignment**: 1. Do any High Law matches negate the moral value of any Low Law matches (such as LL use of force but HL defense against assault on life)? If so, it is a High Law principle regardless of other rules; use and note the override (do not continue with steps 2 and 3) 2. Do any Low Law matches negate the moral value of any High Law matches (such as HL aid/assistance but LL involuntary help to control)? If so, it is a Low Law principle regardless of other rules; use and note the override (do not continue with step 3) 3. Are there more High Law or Low Law distinction rule matches? If equal matches, exclude the rule match that is least strongly related to the principle and note the tie-breaker exclusion 2. **Stance**: Does the text *support* (praise/promote) or *oppose* (criticize/condemn) the principle? 3. **Speaker**: Is the speaker a character or person who is other than the author/narrator, AND the speaker is a villain, adversary, or portrayed as evil? If so, invert the alignment (support to oppose, or oppose to support) 4. **Author Alignment**: Apply 2x2 Map using Principle + Stance = Alignment (HL+support=HL | HL+oppose=LL | LL+support=LL | LL+oppose=HL) 4. **Keep Track**: Along with the *Speaker*, *Stance*, and *Principle* for each *Statement*, also remember the *Rules* list and *Decision Notes* about the rules used to classify the principle, the interpreted stance, the alignment mapping, and if there is a villain speaker stance inversion ## Step 3: Assign Key Topics 1. Identify Key Topics covered by statements - Distinct theme with 2+ supporting statements - Select topics by relevance to High/Low Law evaluation and number of supporting statements - Target 7-14 topics: If more than 14, merge similar/related topics until within range (rare case, do your best) - Keep track of which statements are about each Key Topic 2. Apply **Root Taxonomy** to each Key Topic: 1. Choose one category to classify the Key Topic: 1. Authority and Power 2. Justice and Punishment 3. Mercy and Compassion 4. Autonomy and Consent 5. Conflict and Resolution 6. Community and Belonging 7. Wealth and Resources 8. Knowledge and Technology 9. Suffering and Harm 10. Pride and Humility 11. Transformation and Growth 12. Divine Relationship 2. Normalize Key Topic name 1. **SubTopic**: Rephrase/adjust Key Topic name as needed to be a sub-topic of the chosen Root Taxonomy category 2. Write new Key Topic name using this format: `RootCategory: SubTopic` (example: `Authority and Power: Democratic Governance`) ## Step 4: Synthesize & Score 1. Count **High Law Aligned** and **Low Law Aligned** statements 2. Calculate percentages to normalize score: - if no statements ($HLcount + LLcount = 0$) then skip calculation: $score = 0$ - $HLpct = 100 * HLcount / (HLcount + LLcount)$ - $LLpct = 100 * LLcount / (HLcount + LLcount)$ 3. **Overall Score** (scale -10 to +10): $score = (HLpct - LLpct) / 10$ (round to 1 decimal) For **Per-Topic Scores** use the same 3 formulas applied to the HL and LL counts for each Key Topic. ## Step 5: Self-Verification Before drafting the report, verify: 1. **No missed stances**: Re-scan every sentence; watch for multiple distinct normative claims in the same location (2+ separate stances); split them 2. **Statements are in order**: Sort statements in each table by location in the text to help with duplicate statement recognition and general table scanning. 3. **No duplicate statements**: A statement is a duplicate only if the *Stance Quote* matches AND the *Location* matches; the same stance quote is OK for multiple statements if the locations are different; the same location is OK for multiple statements if the stance claims are different; a full list of statements present in the text is necessary for accurate counts; remove all actual duplicates found 4. Every Key Topic has at least one associated statement; remove any topic with zero statements 5. Every statement has location, principle quote, speaker, stance quote, 1+ distinction rule references, and 1+ Key Topics 6. Alignment 2x2 Map was applied to every statement; verify each statement is in the correct section ## Step 6: Draft Report > Use the "Report Specification" (below) for all reports. **ALL evaluations must include**: 1. **Overview**: what the text is, context, key topics, brief High/Low balance characterization, dominant framework, vocabulary patterns 2. **Key Topics**: numbered list with one-line summary each (quick-reference index) 3. **Statement Quotes** mandatory exact structure: - `## Statement Quotes` - `### High Law Aligned (N statements)` - `### Low Law Aligned (N statements)` - Table: `| # | Location | Rules | Decision Notes | Key Topics | Speaker | Stance Quote | Principle Quote |` - **Location**: Display inline reference numbers (like verses) in the format they are found in the document; display line numbers with an `L` prefix - **Rules**: Comma separated list of {RuleNum}-{HL/LL} like `1-HL,42-LL` - **Key Topics**: The exact name from the Key Topics list; do NOT abbreviate, rephrase, or drop the root category prefix; when a statement covers multiple topics, separate them with `; ` 4. **Scoring Summary** table with Count and Percentage for HL, LL, Total; then the formula and final score 5. **Key Topic Score Table** columns: `Key Topic | High # | Low # | Score` **DETAILED reports add**: 6. **Evaluation Highlights** - **Summary**: overall assessment, score meaning, whether HL vocabulary is deployed inside an LL framework (or vice versa), strongest HL elements, most dominant LL elements - **Key Topics with Strongest Alignment**: strongest HL and strongest LL topics - **High Law Terminology Used for Low Law Principles** (numbered list) - **Low Law Terminology Used for High Law Principles** (numbered list) - **Comparisons Between High and Low Law Principles**: table across applicable distinction rules; in `Rule Topic/Question` include the distinction rule number and the Topic/Question text for easy reference 7. **Key Topic Evaluation Table** columns: `Key Topic | High Law | Low Law`. Use `
-` for bullet points inside cells. Use `(none)` if no presence 8. **Key Topic Details**: one section per topic, containing: - **Summary** (one paragraph) - **Primary Alignment Evaluation** (one paragraph) - **Contradictions to Primary Alignment** (one paragraph each; if none, state "No contradictions identified") - **Key Quotes in Context**: bullet list with quote, reference, and context note - **Detailed Evaluation**: outlines and tables for scanability 9. **Conclusions**: restate score and alignment; summarize text's overall character; final assessment of whether High Law elements are genuine or co-opted by Low Law structure # Report Specification - Generate all output in unicode (UTF-8) encoded GitHub-compatible Markdown (unless the user requests another format) - For Markdown and other text output formats use *ONLY* ASCII characters *except* when unicode is needed for *text* in non-Latin languages; **NEVER** use non-ASCII punctuation or formatting characters (compromises machine readability) **Template compliance is mandatory**: Use the exact section headers and table structures specified above. Do **not** invent alternative header names or table structures. The `## Statement Quotes` section and its two subsections must match exactly, and the header lines for the quotes tables must match exactly. ## Report Template Copy the structure below for each text evaluation. Replace all `{{PLACEHOLDER}}` values. **ALL evaluations**: ```markdown # High Law vs Low Law Alignment Evaluation: {{TEXT_TITLE}} ## Overview {{OVERVIEW}} --- ## Key Topics 1. **{{TOPIC_1}}**: {{ONE_LINE_SUMMARY_1}} N. **{{TOPIC_N}}**: {{ONE_LINE_SUMMARY_N}} --- ## Statement Quotes ### High Law Aligned ({{HL_COUNT}} statements) | # | Location | Rules | Decision Notes | Key Topics | Speaker | Stance Quote | Principle Quote | |---| -------- | ----- | -------------- | ---------- | ------- | ------------ | --------------- | | 1 | {{LOC_REF_1}} | {{RULES_1}} | {{DECISION_NOTES_1}} | {{KEY_TOPICS_1}} | {{SPEAKER_1}} | "{{STANCE_QUOTE_1}}" | "{{PRINCIPLE_QUOTE_1}}" | | N | {{LOC_REF_N}} | {{RULES_N}} | {{DECISION_NOTES_N}} | {{KEY_TOPICS_N}} | {{SPEAKER_N}} | "{{STANCE_QUOTE_N}}" | "{{PRINCIPLE_QUOTE_N}}" | ### Low Law Aligned ({{LL_COUNT}} statements) | # | Location | Rules | Decision Notes | Key Topics | Speaker | Stance Quote | Principle Quote | |---| -------- | ----- | -------------- | ---------- | ------- | ------------ | --------------- | | 1 | {{LOC_REF_1}} | {{RULES_1}} | {{DECISION_NOTES_1}} | {{KEY_TOPICS_1}} | {{SPEAKER_1}} | "{{STANCE_QUOTE_1}}" | "{{PRINCIPLE_QUOTE_1}}" | | N | {{LOC_REF_N}} | {{RULES_N}} | {{DECISION_NOTES_N}} | {{KEY_TOPICS_N}} | {{SPEAKER_N}} | "{{STANCE_QUOTE_N}}" | "{{PRINCIPLE_QUOTE_N}}" | --- ## Scoring Summary | Category | Count | Percentage | | -------- | ----- | ---------- | | High Law Aligned | {{HL_COUNT}} | {{HL_PCT}}% | | Low Law Aligned | {{LL_COUNT}} | {{LL_PCT}}% | | **Total** | **{{TOTAL}}** | **100%** | **Score**: ({{HL_PCT}} - {{LL_PCT}}) / 10 = **{{SCORE}}** --- ## Key Topic Score Table | Key Topic | High # | Low # | Score | | --------- | ------ | ----- | ----- | | {{TOPIC_1}} | {{HL_COUNT_1}} | {{LL_COUNT_1}} | {{SCORE_1}} | | {{TOPIC_N}} | {{HL_COUNT_N}} | {{LL_COUNT_N}} | {{SCORE_N}} | ``` **DETAILED reports add**: ```markdown --- ## Evaluation Highlights ### Summary {{SUMMARY}} ### Key Topics with Strongest Alignment **Strongest High Law Alignment:** - **{{TOPIC_1}}**: {{EXPLANATION_1}} - **{{TOPIC_N}}**: {{EXPLANATION_N}} **Strongest Low Law Alignment:** - **{{TOPIC_1}}**: {{EXPLANATION_1}} - **{{TOPIC_N}}**: {{EXPLANATION_N}} ### High Law Terminology Used for Low Law Principles {{HL_FOR_LL_ANALYSIS}} ### Low Law Terminology Used for High Law Principles {{LL_FOR_HL_ANALYSIS}} ### Comparisons Between High and Low Law Principles | Rule Topic/Question | High Law in Text | Low Law in Text | | ------------------- | ---------------- | --------------- | | **{{RULE_TOPIC_QUESTION_1}}** | {{HL_CONTENT_1}} | {{LL_CONTENT_1}} | | **{{RULE_TOPIC_QUESTION_N}}** | {{HL_CONTENT_N}} | {{LL_CONTENT_N}} | --- ## Key Topic Evaluation Table | Key Topic | High Law | Low Law | | --------- | -------- | ------- | | **{{TOPIC_1}}** | - {{HL_POINT_1_1}}
- {{HL_POINT_1_N}} | - {{LL_POINT_1_1}}
- {{LL_POINT_1_N}} | | **{{TOPIC_N}}** | - {{HL_POINT_N_1}}
- {{HL_POINT_N_N}} | - {{LL_POINT_N_1}}
- {{LL_POINT_N_N}} | --- ## Key Topic Details ### 1. {{TOPIC_NAME}} **Summary**: {{SUMMARY}} **Primary Alignment Evaluation**: {{ALIGNMENT}} **Contradictions to Primary Alignment**: {{CONTRADICTIONS}} **Key Quotes in Context**: - "{{QUOTE_1}}" ({{LOCATION_1}}) - {{CONTEXT_1}} - "{{QUOTE_N}}" ({{LOCATION_N}}) - {{CONTEXT_N}} **Detailed Evaluation** {{DETAILED_EVALUATION}} --- ### 2. {{TOPIC_NAME}} {{Repeat the same structure for each topic}} --- ## Conclusions {{CONCLUSIONS}} ``` ## Post-Report Self-Check After drafting the evaluation report, and before emitting the final report output, verify: 1. Every statement is in the correct section (HL/LL) based on the Decision Notes 2. Key topic names follow the taxonomy format (`RootCategory: SubTopic`) 3. The report structure matches the template exactly 4. `### High Law Aligned (N statements)` and `### Low Law Aligned (N statements)` header counts equal actual data row counts 5. Scoring Summary counts (HL, LL, Total, percentages, Score) are internally consistent and match actual row counts 6. Score table HL/LL counts per topic match the actual statement alignments 7. Score table row count equals Key Topics list count 8. Every topic in the Key Topics list appears in at least one statement; every topic in a statement appears in the list (bidirectional) For detailed reports also verify: 1. Key Topic Evaluation Table bullet points cover all statements for that topic 2. Key Topic Evaluation Table row count matches Key Topics list count 3. Key Topic Details section count and numbering match the Key Topics list exactly 4. All score mentions in Overview, Evaluation Highlights, and Conclusions match the Scoring Summary score # Reference Files *Always* read each applicable reference file *before* executing any instructions: - `references/batch-handling.md`: When evaluating multiple documents from a file directory, database records, etc - `references/report-review.md`: When reviewing an existing high-vs-low evaluation report # Invocation If you have not already been instructed to evaluate a text, ask the user what they would like to do. Offer: 1. **General evaluation or critique** of a web page, article, or speech 2. **Scoring a series** of documents from files, URLs, or a database 3. **Thorough detailed evaluation** of a single document, including per-topic analysis and comparison tables