improve-accessibility

If a student cannot parse the sentence, the pedagogy inside it is invisible. Accessibility is not dumbing down -- it is removing the friction between the learner and the concept.

Purpose

Reduce cognitive load and improve readability of editorial content so that comprehensibility scores reach or exceed 85%. Target flagged fragments from the student-simulator report. Apply sentence simplification, vocabulary substitution, and scaffolding insertion while preserving pedagogical precision and curricular alignment.

Procedure

1. Read report: Load the comprehensibility report from project/reports/. Identify all fragments scored below 85%.
2. Rank by severity: Sort flagged fragments by score ascending. Process the lowest-scoring fragments first.
3. Sentence analysis: For each fragment, measure sentence length (words), subordinate clause depth, and technical term density.
4. Simplify sentences: Break sentences exceeding 25 words into 2+ shorter sentences. Eliminate unnecessary subordination. Keep one idea per sentence.
5. Substitute vocabulary: Replace non-essential technical terms with age-appropriate synonyms. Preserve curricular terms that students must learn -- add inline definitions for those instead.
6. Insert scaffolding: Add transitional phrases, advance organizers, or brief context sentences before complex passages. Use the patterns in references/accessibility-improvement-patterns.md.
7. Verify coherence: Re-read modified passages to confirm logical flow is maintained and no meaning is lost.
8. Re-score: Run the modified content through the student-simulator scoring logic. Confirm all fragments now score >= 85%.
9. Produce changelog: List every modification with before/after and the readability impact.

Inputs

Outputs

Acceptance Criteria

• All previously-flagged fragments score >= 85% comprehensibility.
• No sentence exceeds 30 words after modification.
• Curricular terms are preserved with inline definitions where simplified.
• No pedagogical meaning is altered (content integrity check).
• Changelog documents every change with before/after text.
• Overall unit average comprehensibility >= 85%.

Output Format

## Accessibility Improvement Report

### Score Comparison
| Fragment | Before | After | Delta |
|----------|--------|-------|-------|
| P3-S2    | 62%    | 88%   | +26   |
| P5-S1    | 71%    | 91%   | +20   |

### Changelog
- P3-S2: Split 42-word sentence into 2 sentences (19 + 21 words). Replaced "yuxtaposicion" with "colocacion lado a lado".
- P5-S1: Added advance organizer before technical definition. Replaced nested clause with sequential structure.

### Overall: 72% -> 89% (+17 points)

Error Handling

Anti-Patterns

• Meaning erasure: Simplifying so aggressively that the pedagogical point is lost. Always re-read for meaning preservation.
• Synonym roulette: Replacing terms with near-synonyms that shift meaning. Verify semantic equivalence.
• Scaffolding bloat: Adding so many transitions and organizers that the text doubles in length. Keep scaffolding lean.
• Ignoring report data: Making ad-hoc changes instead of targeting the specific flagged fragments. Follow the report.
• Grade-level mismatch: Using vocabulary appropriate for a different grade. Check target grade level.

References

• references/accessibility-improvement-patterns.md -- Simplification techniques, vocabulary rules, scaffolding patterns
• references/priming-rag/rag-cognitive-load.md -- Cognitive load theory primer
• references/voice-manual-rules.md -- Sinapsis voice manual for tone preservation