Hypothesis

Statement

For bounded explanatory texts, communicative purpose, audience prerequisites, and domain proof obligation predict the order and explicitness of progression moves better than a writer's surface style label.

Mechanism

Proof obligations determine what counts as a warrant and which dependencies must be exposed; audience prerequisites and use mode determine orientation, examples, navigation, and closure.

Predictions

• Different writers in the same purpose/proof regime will share more progression operations than one writer across sharply different genres.
• Surface-style imitation without warrant architecture will not preserve task performance.

Evidence that would support it

Larger preregistered corpus coding, within-writer genre comparisons, and reader studies in which architecture predicts outcomes after controlling for lexical and sentence-level variables.

Evidence that would contradict it

Stable author-specific progression across genres, or surface transformations that preserve or improve outcomes independent of proof architecture.

Tests performed

Five-work comparative desk analysis using one frozen rubric and a 15-unit repeated-coding stability sample.

Results

All works manage orientation, prerequisites, warrants, and closure, but warrant type and sequence align with task and proof regime. Support is provisional because author and domain are confounded.

Falsification attempts

Looked for author-like prose features that could explain the matrix. The RFC's collective authorship and Brown's institutional voice weaken author essence as a sufficient explanation; corpus size prevents decisive rejection.

Current assessment

Provisionally supported at Medium (0.63 judgment).

Next experiment

Code two genres by each of five authors and two authors per genre; fit a multilevel model separating author, genre, proof obligation, and audience.