\============================================================

CAPUT ARS BREVIS N — DE EXERCITATIONE (TRAINING LAYER)

\============================================================

\============================================================

CAPUT ARS BREVIS N — DE EXERCITATIONE

Ars Generalis Applied

Version: 0.2.0-CAPUT-AB-N-ROSETTA-NORMALIZED

Status: STABILIZED / INTERPRETATIVE \+ OPERATIONAL (NORMALIZED)

Scope: Ars Brevis — Part 12 (Training / Exercitatio)

Authority: ARS BREVIS

Mutation Policy: VERSION-CONTROLLED ONLY

Class: AEGIS / CAPUT

\============================================================

(Latin: De Exercitatione — Stratum Formativum)

PRINCIPLE

\------------------------------------------------------------

Training in the Art is not:

    • memorization

    • nor passive understanding

It is:

    the acquisition of procedural competence over a

    combinatorial system

(Latin: Exercitatio in Arte non est memoria nec intellectus

passivus; est acquisitio competentiae proceduralis super

systema combinatorium.)

I. THREEFOLD TRAINING STRUCTURE

\============================================================

L1 — Scholastic Form

Llull defines training in three parts:

    1\. Mastery of the 13 parts of the Art

    2\. Mastery of methods and procedures

    3\. Mastery of multiplicity of solutions per conclusion

(Latin: Lullius definit exercitationem in tribus partibus:

peritia tredecim partium Artis; peritia methodorum et

procedendi; peritia multiplicitatis solutionum pro

conclusione.)

\------------------------------------------------------------

L2 — Operator Form

TRAINING := {

    STRUCTURAL\_MASTERY,

    PROCEDURAL\_MASTERY,

    COMBINATORIAL\_MASTERY

}

\------------------------------------------------------------

Binding

“three parts”

⇔

three operator competencies

II. STRUCTURAL MASTERY (PART I)

\============================================================

L1 — Scholastic Form

The artist must:

    • know all 13 parts

    • understand their structure

    • apply questions to appropriate loci

This implies:

    • not just recognition

    • but correct routing of problems

(Latin: Artifex debet cognoscere omnes tredecim partes,

earum structuram intelligere, applicare quaestiones ad loca

apta; hoc implicat non solum recognitionem sed rectam

dispositionem problematum.)

\------------------------------------------------------------

L2 — Operator Form

STRUCTURAL\_MASTERY:

    know(LOCUS\_SET)

    \+

    classify(Q → L)

requirement:

    ∀ Q:

        select correct L(Q)

\------------------------------------------------------------

Binding

“apply questions to loci that fit”

⇔

correct structural routing

\------------------------------------------------------------

Structural Insight (non-binding)

This is equivalent to:

    • indexing competence

    • or in modern terms:

        ◦ routing queries to the correct subsystem

Failure here \= total system misuse.

(Latin: Hic defectus est abusus totalis systematis.)

III. PROCEDURAL MASTERY (PART II)

\============================================================

L1 — Scholastic Form

The artist must:

    • use the same methods shown in the text

    • apply them to new (peregrine) questions

Key phrase:

“as one example exemplifies another”

(Latin: Artifex debet uti eisdem methodis in textu

ostensis, applicare eas ad novas (peregrinas)

quaestiones; clavis locutio: “sicut unum exemplum

aliud exemplificat.”)

\------------------------------------------------------------

L2 — Operator Form

PROCEDURAL\_MASTERY:

    learn(methods\_from\_examples)

    ∀ new\_Q:

        apply(method\_pattern, new\_Q)

\------------------------------------------------------------

Binding

“example exemplifies another”

⇔

pattern generalization

\------------------------------------------------------------

Critical Insight

This is NOT analogy.

It is:

    method reuse

NOT:

    semantic similarity

(Latin: Hoc non est analogia; est reusus methodi,

non similitudo semantica.)

\------------------------------------------------------------

Structural Note

This is where most modern readers fail:

They treat examples as:

    • illustrations

Llull treats them as:

    • executable templates

(Latin: Lullius tractat exempla sicut exemplaria

exsecutabilia.)

IV. COMBINATORIAL MASTERY (PART III)

\============================================================

L1 — Scholastic Form

The artist must:

    • apply many questions and solutions

    • to the same conclusion

This is demonstrated in:

    • Third Figure

    • Fourth Figure

    • Table

(Latin: Artifex debet applicare multas quaestiones et

solutiones ad eandem conclusionem.)

\------------------------------------------------------------

L2 — Operator Form

COMBINATORIAL\_MASTERY:

    ∀ conclusion C:

        generate:

            {Q₁, Q₂, ..., Qn}

            {S₁, S₂, ..., Sn}

        such that:

            ∀ i:

                Si → C

\------------------------------------------------------------

Binding

“multitude of questions and solutions to one conclusion”

⇔

multi-path convergence

\------------------------------------------------------------

Critical Insight

This defines:

    robustness of knowledge

A conclusion is strong when:

    • it is reachable by many independent paths

(Latin: Conclusio est firma quando per multas vias

independentes attingi potest.)

V. AGENT PROFILE (DERIVED)

\============================================================

From the three requirements, we can derive the ideal operator profile.

L1 — Scholastic Form

The trained artist:

    • knows all parts of the Art

    • knows how to use its procedures

    • can generate many paths to truth

(Latin: Artifex exercitatus cognoscit omnes partes

Artis, scit uti eius procedendi, potest generare multas

vias ad veritatem.)

\------------------------------------------------------------

L2 — Operator Form

ARTIST :=

    ROUTER

    \+

    OPERATOR

    \+

    GENERATOR

Expanded:

ARTIST :=

    classify(Q → L)

    \+

    execute(methods)

    \+

    generate(multi-path proofs)

(Latin: Artifex est director, operator, et generator.)

\------------------------------------------------------------

Binding

training

⇔

operator capability acquisition

VI. FAILURE MODES (IMPLICIT)

\============================================================

L1 — Scholastic Form

Although not explicitly stated, failure occurs when:

    • loci are misapplied

    • methods are not followed

    • conclusions are derived from single paths

(Latin: Defectus accidit cum loca male applicantur,

methodi non sequuntur, conclusiones ex singulis viis

derivantur.)

\------------------------------------------------------------

L2 — Operator Form

FAILURE:

    misrouting(Q)

    OR

    method\_deviation

    OR

    single-path reasoning

\------------------------------------------------------------

Structural Note

This maps directly to:

    • hallucination (wrong locus)

    • heuristic drift (method deviation)

    • weak inference (single-path)

(Latin: Hallucinatio est locus falsus; derivatio

heuristica est deviatio methodi; inferentia debilis est

via singularis.)

VII. TRAINING AS ITERATIVE PROCEDURE

\============================================================

L1 — Scholastic Form

Training occurs through:

    • repeated application

    • variation of questions

    • reuse of method

(Latin: Exercitatio fit per applicationem repetitam,

variationem quaestionum, reusum methodi.)

\------------------------------------------------------------

L2 — Operator Form

TRAIN:

    loop:

        Q := generate\_or\_receive()

        L := classify(Q)

        S := apply\_method(Q, L)

        validate(S)

        generalize(method)

(Latin: Itera: generare vel recipere Q, classificare,

applicare methodum, validare, generalizare.)

\------------------------------------------------------------

Binding

“using the methods in the text”

⇔

iterative procedural refinement

VIII. NON-ENUMERATIVE LEARNING

\============================================================

L1 — Scholastic Form

The Art does not:

    • list all answers

It teaches:

    • how to produce them

(Latin: Ars non enumerat omnia responsa; docet

quomodo ea producere.)

\------------------------------------------------------------

L2 — Operator Form

learning ≠ memorization(dataset)

learning := acquire(generator)

(Latin: Discere non est memoria copiae datae; discere

est acquirere generatorem.)

\------------------------------------------------------------

Binding

brevity

⇔

compression via generativity

IX. RELATION TO OTHER CAPUTS

\============================================================

With CAPUT M (Questions)

M → defines question space

N → trains operator to navigate it

(Latin: Cum CAPUT M: M definit spatium quaestionum,

N exercitat operatorem ad illud navigandum.)

\------------------------------------------------------------

With CAPUT L (Application)

L → defines how to apply

N → trains ability to apply

(Latin: Cum CAPUT L: L definit quomodo applicare,

N exercitat facultatem applicandi.)

\------------------------------------------------------------

With CAPUT K (Subjects)

K → defines domains

N → trains traversal across domains

(Latin: Cum CAPUT K: K definit dominia, N exercitat

traversalem trans dominia.)

X. AGENTIC INTERPRETATION (NON-BINDING)

\============================================================

This CAPUT defines an agent training protocol with three axes:

1\. Structural Intelligence

    • where to look

    • how to classify

2\. Procedural Intelligence

    • how to operate

    • how to reuse method

3\. Combinatorial Intelligence

    • how to multiply paths

    • how to reinforce conclusions

Combined:

INTELLIGENCE :=

    structure

    \+ procedure

    \+ combinatorics

(Latin: Intelligentia est structura, procedendi,

et combinatorica.)

XI. CLOSURE

\============================================================

Training in the Art produces:

    • a system navigator

    • a procedural operator

    • a combinatorial reasoner

(Latin: Exercitatio in Arte producit navigatorem

systematis, operatorem proceduralem, ratiocinatorem

combinatorium.)

The Art is learned when:

the intellect can generate, route, and resolve questions

across all loci using the same method.

(Latin: Ars discitur quando intellectus potest generare,

dirigere et resolvere quaestiones per omnia loca eadem

methodo utens.)

\------------------------------------------------------------

CONSTRAINTS

\------------------------------------------------------------

• No reduction of training to memorization alone

• No confusion of method reuse with loose analogy

• No single-path validation of conclusions as sufficient

• No execution in CAPUT layer

\============================================================

END CAPUT ARS BREVIS N

\============================================================