A. Alternatives considered

What wouldn't be built, grouped by what each rejection protects. Strategy-level rejections are in Section 4.11.

Where the feature lives

• A section on the Quest page. The featured slot is already contested by Weekend Quest and Friends Quest, any placement is an implicit ranking, and a section appearing and disappearing with relationship status fails clean removal.
• An item in the Daily Quests list. Draws on the quest-point economy for actions that already earn XP. The daily prompt lives on the Path tile instead, with the same pull and no touch on the quest economy.
• Re-skinning existing Daily Quests. Modifies what already exists, and a re-skin reverting when a pair ends is a visible loss where a dedicated surface disappearing is not.
• A persistent floating quiz entry on the Path. Collides with the scroll-to-bottom control and would be a third entry point competing with the tile and marker.
• A per-row add-a-buddy icon on course selection. No extra clicks for a solo learner, but an unfamiliar affordance repeated down a scrolling list at the moment someone is picking a language.
• A solo-or-together choice after picking a course, or a "how do you want to learn?" step. Both tax every solo learner with a decision on the most common path in the app.
• A private buddy feed. The feed is a notification surface, not a history surface. The record lives on the buddy's profile, and a private feed would duplicate it and be one more thing to dismantle.
• Pending invites in the profile section. A proposal is not a pair, and a pending indicator on a slot already carrying an avatar and a flag would overload it.
• A shortcut from the review range to the solo position. Competes with the existing scroll-return. "Progress is never blocked" concerns permission, not convenience, and scrolling isn't prevention.

Starting and structure

• A required fresh start at Unit 1. Superseded by the shared start point. Requiring a reset would force engaged learners to discard real progress and misrepresent what they've learned. Fresh start was one way to guarantee both people occupy the same territory; the guarantee is what matters, not the freshness.
• Deferring pairs at different levels. No longer applicable. The start point doesn't match content across a gap; it closes the gap, so no matching system is needed.
• Eligibility gates on progress or recency. Willingness to rewind is the learner's judgment about their own progress and their own friendship. Any threshold is arbitrary, self-selection already sorts it, and protection comes from the invite being a proposal rather than an obligation.

Social pressure and consent

The largest group, deliberately.

• Periodic re-commitment prompts. The app inserting itself into a relationship the learners own. A dormant pair costs nothing, ending is always available in two places, and a prompt to think about the relationship is the wrong reactivation mechanism; the quiz notification prompts the activity itself, which is what brings people back. The accepted consequence is that reactivation rests entirely on the quiz loop.
• Mutual-consent exit. Traps someone whose buddy won't respond. Free exit is what makes the pair safe to enter.