Narrative essay use case

# Narrative essay use case

A single, reusable narrative-essay AssignmentType covering college essays, personal essay, and memoir. The connecting thread is the personal narrative: voice, scene-building, specificity, and narrative arc. Distinct from the thesis-driven analytical writing Yawp's defaults are tuned for.

## Problem

Two related gaps Yawp doesn't currently serve:

- **College essays.** Many high schools require students to write college essays as part of their senior-year curriculum, but teachers haven't been trained on the genre's specific demands (personal narrative, voice, brevity, prompt-specificity). Schools want to say "yes, every student writes a college essay" without each teacher needing to become an admissions essay expert.
- **Personal narrative more broadly.** Teachers running personal-essay or memoir-style units don't have rubric or tutor support that speaks the genre's vocabulary — scene vs. summary, voice, specificity of detail, narrative arc.

College essays are a subgenre of personal narrative; one AssignmentType serves both.

## Goals (rough)

- One narrative-essay AssignmentType, designed to be reusable across college essay, personal essay, and memoir contexts. The tutor flow is generic across all three in v1 — teacher's prompt sets the context, tutor behavior doesn't branch by subgenre.
- Prompts are teacher-supplied in v1. Preloaded library (Common App, school-specific) is a v2 add.
- Rubric per the [Personal Narrative Writing Rubric v2](https://drive.google.com/file/d/1PaST6RLPQVa0Whw_615_qR6FEFH4G2iX/view?usp=drivesdk) (see below). Tuned to narrative criteria — voice, narrative craft, reflection, structure, conventions — not 5-paragraph or thesis-driven.
- Tutor flow tuned to genre-specific phases: idea/angle → memory/material → scene-building → revision for voice and structure. Different shape than thesis → intro → body → conclusion.

## Rubric

High school · holistic · percentage-based (out of 100). Source: [Personal Narrative Writing Rubric v2](https://drive.google.com/file/d/1PaST6RLPQVa0Whw_615_qR6FEFH4G2iX/view?usp=drivesdk) (Google Drive, Yawp! folder).

**Philosophy.** The rubric prioritizes what makes personal narrative writing genuinely excellent: a distinctive voice, genuine reflection, earned insight, and the ability to render experience in vivid, specific detail. Bold thinking with minor technical flaws outscores technical polish without authentic engagement. The grading assistant should reward narratives that make a reader feel something — not essays that merely follow the rules.

| Category | Weight | What "exemplary" looks like |
|---|---:|---|
| Voice & Style | 25% | Authentic, distinctive voice; tone matches emotional stakes; memorable phrasing; every word earns its place. |
| Narrative Craft | 25% | Scenes rendered with specific sensory detail; showing outpaces telling; dialogue (if used) reveals character; reader can fully visualize the world. |
| Reflection & Insight | 25% | Genuine, nuanced thinking about the experience; insight feels earned, not tacked on; conclusion offers earned discovery; essay carries weight. |
| Structure & Organization | 15% | Intentional pacing; purposeful beginning and earned ending; organic transitions; deliberate compression and expansion. |
| Conventions | 10% | Virtually error-free; sentence variety creates rhythm; punctuation used skillfully; polished presentation; intentional rule-breaking serves voice. |

Performance bands within each category: Exemplary (90–100), Proficient (80–89), Developing (70–79), Struggling (<70). Final grade = Σ (category score × weight).

The rubric is **tutor-internal**: the student doesn't see categories or weights as a panel during drafting. The tutor speaks in the rubric's vocabulary ("the scene needs more sensory detail," "this insight feels tacked on") without naming the categories explicitly. The grading assistant uses the categories and bands to produce the final grade and feedback.

## Tutor flow

Four phases in **soft sequence**: the tutor recommends what to do next, but the student can skip ahead or jump back at any time. The tutor decides when to suggest moving on based on what the student has produced (e.g., "you've got a clear angle — want to start drafting the opening scene?"). No hard gates, no word-count thresholds.

1. **Idea / angle.** The tutor interviews the student to find the productive intersection of a big idea and a specific moment. Students arrive from either direction: some have a moment looking for meaning ("this thing happened, I'm not sure why it stuck"), others have a theme looking for a moment ("I want to write about my grandmother / about leaving home / about not fitting in"). The tutor meets them where they are — zooming in from the abstract or zooming out from the specific. Open-ended prompts: "What's a moment that surprised you?" "When did you change your mind about something?" "What would you tell a younger version of yourself?" — alongside, for theme-first students, "Is there one specific time that captures this?" Output: an angle that has both a thread of meaning and a concrete moment to anchor it.
2. **Memory / material.** With an angle in hand, the tutor helps the student gather raw material — sensory details, snippets of dialogue, the texture of a place or person. Still conversational; the tutor pushes for specificity ("what did the room smell like," "what exactly did she say"). Output: enough concrete material to render a scene.
3. **Scene-building.** Drafting begins. The student writes; the tutor coaches in the rubric's vocabulary — pushing on showing vs. telling, voice consistency, scene rendering. The tutor reacts to what's on the page rather than continuing to interview.
4. **Revision.** Whole-essay pass. Tutor focuses on voice consistency, earned insight, beginning/ending, and pacing. Conventions noted but not the focus — the philosophy ("bold thinking with minor technical flaws outscores technical polish without authentic engagement") guides emphasis.

Phase progression is suggested in tutor turns, not enforced by the UI. Student can shift back to brainstorming mid-revision if a new angle emerges.

**Length.** Default soft target around 650 words across all subgenres in v1 — short enough to force specificity and earned compression, which serves personal essay and memoir as much as college essays. Not named to the student as a "Common App" length; just the writing target. The tutor uses it to coach toward concision in revision ("this scene is doing two jobs — which one earns its place?"), not as a hard cap.

## Non-goals

- Reported narrative, profile writing, or journalism. Distinct genre with different rubric primitives (reporting, sourcing, factual care); out of scope here.
- Submitting essays directly to admissions systems. Yawp is the workspace; export goes to the school's existing pipeline.
- Replacing dedicated college counseling. This is the writing tool, not the strategy advisor.

## Open questions

- [ ] When a student skips a phase, the tutor back-fills by default (e.g., starts asking idea/angle questions in the margins of an early draft) — but bails out if back-filling would derail momentum the student clearly has. What's the heuristic for "too messy to back-fill"? Likely: if the draft already has a working angle implicit in it, skip the interview and coach from where they are.
- [ ] Does the GA's final-grade output show the rubric breakdown to the student post-grade, even though it was tutor-internal during drafting?

## Engineering handoff

Not ready, but close. This is a self-contained AssignmentType — rubric specified, tutor flow described in prose. Independent of the differentiated-rubrics-by-assignment-type spec; can ship on its own. Still needed before the ready-for-engineering bar: domain context, file-path guesses in `yawp-2.0`, data model implications (new AssignmentType variant, any new fields), edge cases, test plan, rollout plan.