Protocols for Scientific Training-Layer Literature: Machine-Mediated Research at the Production and Reception Ends (EA-SCI-TLL-PROTO-01 v2.1)

The paper separates four related objects:

• genre: training-layer literature;
• operational domain: machine-reception literature;
• architecture: dual-layer scientific publication;
• protocol suite: TLL-P, TLL-R, and TLL-G.

Its six reception layers are:

1. training; 2. indexing; 3. embedding; 4. retrieval; 5. composition; 6. agentic use.

A dual-layer publication contains:

• a machine-reception representation;
• a human-interpretation representation.

Neither is merely a supplement to the other.

The protocol families govern:

• TLL-P: production and structural decomposition;
• TLL-R: reception, cross-model review, and confabulation resistance;
• TLL-G: governance, optimization boundaries, provenance, and accountability.

The paper’s “machine hermeneutic profile” refers to observable reception behavior, not machine phenomenology.

The document itself implements only:

• a stable identifier;
• a machine-audience declaration;
• a definitional anchor;
• a minimum claim registry;
• a schema sketch.

Its full machine-reception counterpart and experimental evaluation are deferred.