Scholar Reasoning Distiller

What it solves

When models are asked to reason like a scholar, a common failure mode is surface imitation: the output borrows vocabulary, citations, and tone, but still reads like generic academic language. Scholar Reasoning Distiller targets a more specific problem: reconstructing a scholar’s stable reasoning architecture from existing texts and applying that architecture as a constraint in concrete tasks. It does not focus on sounding like the scholar. It focuses on how the scholar defines the object of study, draws conceptual distinctions, orders analytical moves, evaluates evidence, recruits theories, manages rhetorical rhythm, refuses certain methods, and changes over time.

How it works

The skill works in two phases. In Distillation, it processes supplied materials—papers, monographs, interviews, or lecture notes—using an eight-layer extraction framework:
- Ontology: how the scholar defines the object of study.
- Concepts: recurring distinctions and conceptual tools.
- Analytical operations: the expected sequence of moves.
- Evaluation: what counts as strong or weak evidence.
- Intertextuality: frequent citations and their function.
- Rhetoric: sentence rhythm and argument pacing.
- Boundaries: methods or questions explicitly refused.
- Diachronic change: shifts across periods, if available.

In Deployment, those extracted features become nine hard constraints: scope checks, problem re-framing, procedural order, evaluation thresholds, source discipline, rhetorical nodes, refusal rules, and provenance notes. It supports four task types—peer review, doctoral supervision, lecturing, and panel discussion—and each task output should include quantitative self-assessment plus qualitative commentary. If multiple scholars are distilled, it can also run a differential validation to check whether their fingerprints actually diverge in ontology, concepts, procedure, evaluation, or refusal boundaries.

Boundary conditions

The skill is best for personal study, critical thinking, and drafting support, and its quality depends on the supplied corpus. Low-maturity fingerprints should be treated as provisional, especially when the question lies outside the scholar’s domain or crosses an explicitly refused analytical path. It should not be used to fabricate peer reviews, impersonate real scholarly labor, or conduct academic misconduct.