How should I evaluate whether a question actually belongs in a science context?

This is a meta-question about the boundary of science itself—and it's a good one. The preprints here don't directly address *your* question, but they do illuminate how experts think about rigor and what counts as legitimate scientific inquiry.

The short answer: A question belongs in science if it's answerable in principle by observation, experiment, or mathematical derivation, and if it's framed in a way that admits being wrong.

Here's what that means in practice:

What makes a question scientific:

• It concerns the physical world, mathematics, or logical consequences of established theories.
• There is (or could be) a way to check the answer: through experiment, measurement, calculation, or peer review of the reasoning.
• The question is specific enough that different answers could be meaningfully different.

For example: "Does a given propulsion mechanism produce thrust measurable above instrumental noise?" is scientific. "Is this concept elegant?" is not—elegance is aesthetic.

Red flags that a question may not belong:

• It asks for a subjective value judgment ("Is quantum mechanics *really* true in some deeper sense?") rather than prediction or measurement.
• It's so vague that any answer would be compatible with it.
• It relies on unfalsifiable claims ("What would happen if the universe had completely different laws?").

The preprints in your list are mostly about *evaluating* AI performance on real physics questions [AG-2025.07-1634][AG-2025.09-1139][AG-2025.02-240], and they show that expert physicists distinguish between conceptual understanding, rigorous derivation, and implicit reasoning constraints [AG-2026.04-565]. That curation process—what experts decide to ask—*is* part of how science polices its own boundaries.

The archive you're asking about focuses on frontier science: propulsion, relativity, mathematical physics. If your question fits those domains and is answerable in principle, it belongs here.