Edit Step4_resolution_patterns Template

Step4_resolution_patterns v3

Viewer

Test Case:

Template

Preview

History

System prompt (optional) User prompt (template)

Analyze HOW the board resolved each ethical question in their conclusions.

## BOARD CONCLUSIONS (determinations made)
{{ conclusions_text }}

## ETHICAL QUESTIONS (that needed answers)
{{ questions_text }}

## CODE PROVISIONS (that could be cited)
{{ provisions_text }}

## CASE NORMATIVE STRUCTURE (already extracted -- ground the weighing in these)
{{ norm_structure or '  (not available)' }}

A board resolution is DEFEASIBLE: it holds only under the facts and conditions
that obtained in this case, and would not hold (or would reverse) if those
conditions changed. Express each resolution conditionally, not as an absolute
rule. State the activating conditions ("holds WHEN ...") and, where the board
signalled them, the defeating conditions ("would NOT hold UNLESS ..." / "absent ...").

For EACH conclusion above, analyze:
1. Which QUESTIONS does it answer? (by Q number)
2. What PRINCIPLES were determinative? (up to 3 key principles)
3. What FACTS were determinative? (up to 3 key facts)
4. What PROVISIONS were cited? (by P number)
5. How were COMPETING OBLIGATIONS weighed? (1 sentence)
6. Under WHAT CONDITIONS does this resolution hold, and what would defeat or
   reverse it? Phrase as "Holds when <conditions>; would not hold if/unless
   <defeating conditions>." (1-2 sentences)
7. A 1-2 sentence NARRATIVE explaining how the board reached this conclusion,
   framed conditionally on the determinative facts above (not as a universal rule)

Output as JSON array:
```json
[
  {
    "conclusion_index": 1,
    "answers_questions": [1, 2],
    "determinative_principles": ["principle 1", "principle 2"],
    "determinative_facts": ["fact 1", "fact 2"],
    "cited_provisions": [1, 3],
    "weighing_process": "One sentence on how competing obligations were balanced.",
    "resolution_conditions": "Holds when the engineer disclosed the conflict in writing; would not hold if the disclosure were withheld.",
    "resolution_narrative": "Given that X obtained, the board concluded Y because...",
    "confidence": 0.8
  }
]
```

Include all {{ conclusion_count }} conclusions in this batch.

{{ style_formatting_line }}

Loading preview...

Version History

No version history yet. Save changes to create versions.

Saved

2133 chars

Reference

Template Variables

All case questions numbered Q1..Qn, or the literal 'No questions extracted'

All provisions numbered P1..Pn as 'Pn. label: code - definition[:100]' lines, or the literal 'No provisions extracted'

Shared ProEthica no-em-dash output formatting clause

Batch conclusions numbered C1..Cn as 'Cn. label: text' lines

Number of conclusions in this batch

Output Schema

type array

items

JSON Wrapper

Appended to the prompt by the extractor:

IMPORTANT: Wrap your response as a JSON object with exactly two keys:
{"new_step4_resolution_patterns_classes": [...], "step4_resolution_patterns_individuals": [...]}
If there are no individuals to report, use an empty array for "step4_resolution_patterns_individuals".