> Full-text translation from Chinese. The original remains authoritative for its research claims; this translation is not independent mathematical verification. # Algebraic Combinatorial Transcendence Methodology ## Algebraic Combinatorial Transcendence Methodology (ACTM) **Authors:** Neo.K; Aletheia (GPT-5.6 Sol) **Institution:** EveMissLab / 一言諾科技有限公司 **Date:** 2026-09-30 **Version:** v0.1 **Document type:** Methodology paper / research program (Methodological Program) **Status:** A conceptual framework has been established; no claim to have proved Schanuel's conjecture or any new transcendence theorem --- ## Abstract Traditional transcendence research often starts from a specified number or function value and asks whether it is algebraic, irrational, or transcendental. This approach has been extremely successful in classical results. However, when the question rises to global statements such as Schanuel's conjecture, involving arbitrary finite collections, algebraic independence, and transcendence degree, starting directly from 'transcendental numbers themselves' can immediately confront an enormous space of relations to exclude. This paper proposes a reverse methodology: instead of first studying 'which numbers are transcendental', study 'which algebraic inputs have the potential to generate transcendence'. The research object shifts from an individual value to algebraic seeds, algebraic tuples, their internal relation structures, and the nonalgebraic generating operators acting on them. The core question becomes: > Given a finite structure lying entirely in the algebraic closure, which combinatorial configurations, after passing through particular operators, limits, periods, analytic continuation, or other generating mechanisms, produce degrees of freedom that cannot be fully absorbed by algebraic relations? This research program is called 'Algebraic Combinatorial Transcendence Methodology' (ACTM). Its core workflow is: $$ \boxed{ \text{Algebraic Seed} \rightarrow \text{Combinatorial Structure} \rightarrow \text{Operator} \rightarrow \text{Relation Transport} \rightarrow \text{Escape} \rightarrow \text{Transcendence Degree} } $$ ACTM aims not to restate existing transcendence theory in new terminology, but to establish a research framework that supports systematic comparison, layer-by-layer verification, and connections to Schanuel-type predimension problems. --- # 1. Reversing the Question ## 1.1 The Traditional Question A typical question is whether: $$ \alpha \in \mathbb C $$ satisfies a nonzero polynomial equation: $$ P(\alpha)=0, \qquad P\in\mathbb Q[X]? $$ If no such polynomial exists, $\alpha$ is called transcendental. In the multivariable case, for: $$ \alpha_1,\ldots,\alpha_n $$ if there is no nonzero: $$ P\in\mathbb Q[X_1,\ldots,X_n] $$ such that: $$ P(\alpha_1,\ldots,\alpha_n)=0, $$ then these values are algebraically independent over $\mathbb Q$. The difficulty is that this is a global exclusion condition: formally, the entire space of polynomial relations must be confronted. --- ## 1.2 ACTM's Reverse Question ACTM does not first ask: > Is a specified number transcendental? Instead, it asks: > Which algebraic seeds and configurations of algebraic relations, after passing through a generating operator, have the ability to escape the algebraic closure or increase transcendence degree? The starting point therefore changes from: $$ \text{Transcendental Output} $$ to: $$ \text{Algebraic Input}. $$ The core direction is reversed to: $$ \boxed{ \overline{\mathbb Q}^{\,n} \longrightarrow \text{operator-induced transcendence} } $$ --- # 2. The First Hard Boundary: Finite Algebraic Closure Cannot Escape on Its Own Let: $$ \overline{\mathbb Q} $$ be the algebraic closure of $\mathbb Q$. Under finitely many additions, subtractions, multiplications, divisions, and extractions of roots of algebraic equations, $\overline{\mathbb Q}$ remains closed. Thus, if: $$ a,b\in\overline{\mathbb Q}, $$ then, whenever defined: $$ a+b,\quad a-b,\quad ab,\quad \frac{a}{b} $$ still belong to $\overline{\mathbb Q}$. More generally, any expression constructed solely from finitely many algebraic operations still has its value in the algebraic closure. Therefore: $$ \boxed{ \text{Finite Algebraic Composition} \not\Rightarrow \text{Transcendence} } $$ This is not a limitation, but ACTM's first filter. It tells us: > What actually needs studying is not 'how adding algebraic numbers together turns them into transcendental numbers', but 'which generating mechanisms break purely algebraic closure'. --- # 3. Algebraic Seeds and Combinatorial Potential ## 3.1 Algebraic Seeds Define a finite algebraic seed: $$ A=(a_1,\ldots,a_n) \in \overline{\mathbb Q}^{\,n}. $$ The value of each $a_i$ alone is usually insufficient to describe its research value. ACTM focuses on the relation structure of the entire tuple. --- ## 3.2 The Relation Ideal Let: $$ I(A) = \left\{ P\in\mathbb Q[X_1,\ldots,X_n] : P(a_1,\ldots,a_n)=0 \right\}. $$ $I(A)$ describes the internal algebraic relations of the algebraic seed. ACTM views $A$ not as a sequence of numbers, but as: $$ \boxed{ (A,I(A)) } $$ that is, 'values + a relation space'. --- ## 3.3 Combinatorial Potential The preliminary definition of 'combinatorial potential' is structural rather than a single numerical value: > An algebraic tuple has combinatorial potential relative to a family of operators if, under that family, its different internal relations are preserved, distorted, destroyed, or transformed to different degrees. It may be denoted by: $$ \operatorname{CP}_{\mathcal O}(A). $$ In v0.1, $\operatorname{CP}$ is not yet assigned a unique scalar form; it can be described jointly by several indicators, such as: - $\mathbb Q$-linear rank; - Multiplicative rank; - The dimension or generating complexity of the relation ideal; - Symmetry; - The number of subrelations invariant under the operator; - Relations newly appearing or disappearing after the operator acts; - The resulting lower bound on transcendence degree. --- # 4. Escape Operators ## 4.1 The Basic Concept ACTM collectively calls mechanisms that send algebraic inputs to potentially transcendental outputs 'escape operators': $$ \mathcal O:A\mapsto \mathcal O(A). $$ The word 'escape' in the name means only that: $$ \mathcal O(A) $$ may leave $\overline{\mathbb Q}$; it does not mean that every input necessarily produces a transcendental value. --- ## 4.2 First Class: Known Strong Escape Operators The most important prototype is the exponential function. If: $$ 0\neq\alpha\in\overline{\mathbb Q}, $$ then Lindemann-Weierstrass theory implies that: $$ e^\alpha $$ is transcendental. Thus: $$ \exp $$ can be viewed as a known strong escape operator on nonzero algebraic inputs. This gives ACTM its first standard example: $$ \boxed{ \text{nonzero algebraic seed} + \exp \rightarrow \text{transcendental output} } $$ --- ## 4.3 Second Class: Conditional Escape Operators Some operators guarantee transcendence only under specific input conditions, branches, function families, or ranges of known theorems, for example: $$ \log,\quad \sin,\quad \cos,\quad \text{elliptic functions}, \quad \text{modular functions}. $$ For these operators, ACTM must strictly distinguish: $$ \text{the known-theorem region} $$ from: $$ \text{the unknown / conjectural region} $$ as separate regions. --- ## 4.4 Third Class: Operators of Unknown Escape Strength For example, the algebraicity, transcendence, or algebraic independence of certain special-function values: $$ \Gamma(\alpha), $$ at many algebraic points is still not fully understood. ACTM does not presuppose that they necessarily have any particular escape strength, but treats them as later research objects. --- # 5. The Increment of Generated Transcendence For an algebraic seed: $$ A\in\overline{\mathbb Q}^{\,n}, $$ and an operator $\mathcal O$, define the preliminary increment of generated transcendence: $$ \Delta_{\mathcal O}(A) = \operatorname{trdeg}_{\mathbb Q} \mathbb Q \big( A,\mathcal O(A) \big) - \operatorname{trdeg}_{\mathbb Q}\mathbb Q(A). $$ Since all entries of $A$ are algebraic: $$ \operatorname{trdeg}_{\mathbb Q}\mathbb Q(A)=0, $$ we have: $$ \boxed{ \Delta_{\mathcal O}(A) = \operatorname{trdeg}_{\mathbb Q} \mathbb Q\big(\mathcal O(A)\big) } $$ If the operator output is a vector, this quantity measures the genuinely new algebraic degrees of freedom in the output. --- # 6. Algebraic Escape Potential Define the working quantity: $$ \operatorname{AEP}_{\mathcal O}(A) $$ as Algebraic Escape Potential. In this version, it is not an established mathematical definition but a research interface. Its minimal classification can be divided into: $$ \operatorname{AEP}_{\mathcal O}(A) = \begin{cases} 0, & \mathcal O(A)\subseteq\overline{\mathbb Q},\\ >0, & \operatorname{trdeg}_{\mathbb Q}\mathbb Q(\mathcal O(A))>0,\\ ?, & \text{Current theory cannot decide}. \end{cases} $$ A more mature version could add: $$ \operatorname{AEP}_{\mathcal O}^{(k)}(A) $$ to indicate a guarantee of at least $k$ transcendental degrees of freedom. --- # 7. Relation Transport Rather Than Single-Point Classification ## 7.1 From Values to Relations ACTM's core is not merely to study: $$ a \mapsto \mathcal O(a), $$ but rather: $$ (A,I(A)) \mapsto (\mathcal O(A),I(\mathcal O(A))). $$ What actually needs comparing is: $$ I(A) $$ with: $$ I(\mathcal O(A)). $$ --- ## 7.2 Relation Preservation An input relation is called preserved if it still corresponds to an algebraic relation after the operator acts. Abstractly: $$ R_{\text{in}} \longrightarrow R_{\text{out}}. $$ --- ## 7.3 Relation Destruction Relation destruction occurs when an input relation cannot be completely transported to the output algebraically. This does not automatically imply transcendence, but it may be one structural source of transcendental degrees of freedom. --- ## 7.4 Relation Generation Operators may also introduce new relations in the output space, for example through functional equations, periodicity, or symmetry: $$ \mathcal O(A) $$ may have more structure than a point in general position. Thus, one cannot simply assume: $$ \text{relation loss} = \text{transcendence gain}. $$ ACTM must account simultaneously for: $$ \text{lost relations}, \qquad \text{preserved relations}, \qquad \text{new relations}. $$ --- # 8. The Relation Gap Define a conceptual relation gap: $$ G_{\mathcal O}(A) = \mathcal R_{\text{expected}} - \mathcal R_{\text{surviving}}, $$ where $\mathcal R$ provisionally denotes some quantifiable relation capacity rather than a fixed scalar. The intuition is: > If the constraint structure on the input side cannot all survive as algebraic relations after the operator acts, the remaining gap may manifest as new transcendental degrees of freedom. However, it must be emphasized that: $$ \boxed{ G_{\mathcal O}(A)>0 \not\Rightarrow \Delta_{\mathcal O}(A)>0 } $$ unless an additional theorem is established. Thus, in ACTM, a 'relation gap' is first a diagnostic quantity, not a proved conclusion. --- # 9. Preliminary Classification of Algebraic Tuples In its first phase, ACTM can classify algebraic tuples into the following research types. ## 9.1 Highly Redundant Tuples For example: $$ A=(a,a^2,a^3). $$ has many explicit internal algebraic relations. --- ## 9.2 $\mathbb Q$-Linearly Free Tuples For example: $$ \operatorname{ldim}_{\mathbb Q}(A)=n. $$ These tuples are particularly important for exponential-type problems. --- ## 9.3 Multiplicatively Free Tuples Study whether a nontrivial integer vector: $$ (m_1,\ldots,m_n) $$ exists such that: $$ a_1^{m_1}\cdots a_n^{m_n}=1. $$ --- ## 9.4 Tuples with Low Algebraic Freedom and High Operator Sensitivity Some tuples have many algebraic relations themselves, yet may undergo strong relation reorganization under particular operators. ACTM pays particular attention to this class, because it may reveal that: $$ \boxed{ \text{transcendence generation} \neq \text{mere input randomness} } $$ --- ## 9.5 Tuples with Mixed Degrees of Freedom The same tuple may be: - Linearly free; - Not multiplicatively free; - Algebraically constrained; - Highly sensitive to a particular operator. Thus, ACTM does not adopt a single 'free / not free' dichotomy, but establishes a multiaxis description of degrees of freedom. --- # 10. Operator Compatibility ## 10.1 The Core Question Given an input relation: $$ R(A)=0, $$ we ask: > Is this relation compatible with the functional equations, symmetry, and geometric structure of the operator $\mathcal O$? If compatible, the relation may be preserved. If incompatible, relation reorganization may occur. --- ## 10.2 Typical Compatible Relations for the Exponential Operator The exponential function satisfies: $$ e^{x+y}=e^x e^y. $$ Thus, the additive structure of the input maps to the multiplicative structure of the output. This is a key reminder: $$ \boxed{ \text{relation destruction} \neq \text{relation disappearance} } $$ Some relations merely change their representation domain. What ACTM must study is therefore: $$ \text{Relation Transport} $$ rather than merely 'whether relations exist'. --- # 11. The Generating Chain ACTM's standard research chain is: $$ \boxed{ A \rightarrow I(A) \rightarrow \operatorname{Type}(A) \rightarrow \mathcal O \rightarrow \mathcal O(A) \rightarrow I(\mathcal O(A)) \rightarrow \Delta_{\mathcal O}(A) } $$ The concrete workflow is: 1. Select algebraic seeds; 2. Compute or describe input relations; 3. Classify their linear, multiplicative, and algebraic degrees of freedom; 4. Select an operator; 5. List the operator's known functional equations and invariants; 6. Predict which relations can be transported; 7. Identify possible relation gaps; 8. Turn the gaps into provable propositions; 9. Verify using existing transcendence theorems; 10. Formulate conjectures for unresolved regions without overstepping in claims. --- # 12. Known Theorems as a Calibration Set ACTM must first pass retrospective tests against known mathematical results rather than directly challenge unresolved conjectures. ## 12.1 Lindemann-Weierstrass Calibration Use algebraic numbers: $$ \alpha_1,\ldots,\alpha_n $$ to study known linear-independence and transcendence results under the exponential map. --- ## 12.2 Gelfond-Schneider Calibration If: $$ \alpha\in\overline{\mathbb Q}, \qquad \alpha\neq0,1, $$ and: $$ \beta\in\overline{\mathbb Q}\setminus\mathbb Q, $$ then, on any appropriate branch: $$ \alpha^\beta $$ is transcendental. This provides another generating template: $$ \boxed{ \text{algebraic base} + \text{irrational algebraic exponent} \rightarrow \text{transcendence} } $$ --- ## 12.3 Hermite-Lindemann Calibration Study algebraic / transcendental transitions under the inverse relationship between exponentials and logarithms. --- ## 12.4 The Purpose of Calibration If ACTM cannot naturally reconstruct the structural features exhibited by known theorems, the methodology remains immature. Therefore: $$ \boxed{ \text{Known Theorems First} \rightarrow \text{Unknown Territory Later} } $$ --- # 13. From the 'Compressing $\pi$' Approach to ACTM Earlier research on $\pi$ left an important methodological principle: $$ \boxed{ \text{Object} \neq \text{Representation} \neq \text{Expansion} } $$ The same mathematical object can have: - A decimal expansion; - A continued fraction; - An integral representation; - A series representation; - A geometric representation; - A serializable generating procedure; - A verifiable approximation; - Structural indices in different theories. Thus, when studying $\pi$, the infinite decimal expansion need not be treated as the principal object itself. ACTM extends this principle to transcendence problems: > Rather than making 'the decimal output of a transcendental number' the primary object, study its generating source, input degrees of freedom, operators, and relation transformations. --- # 14. Generation, Projection, Reconstruction, and Collapse ACTM retains four core operational terms: $$ \boxed{ \text{Generation} } $$ $$ \boxed{ \text{Projection} } $$ $$ \boxed{ \text{Reconstruction} } $$ $$ \boxed{ \text{Collapse} } $$ Their meanings are as follows. ## 14.1 Generation $$ A \xrightarrow{\mathcal O} B. $$ --- ## 14.2 Projection Take different representations or observations of the same generated object: $$ B \xrightarrow{\Pi_i} Y_i. $$ --- ## 14.3 Reconstruction Attempt to recover the source structure from multiple projections: $$ \{Y_i\} \longrightarrow \widehat B. $$ --- ## 14.4 Collapse If a map or system of relations eliminates many preexisting degrees of freedom, this is called a collapse of degrees of freedom. Here, 'collapse' is a structural term; it does not presuppose illegitimacy or contradiction. --- # 15. The Connection to Schanuel's Conjecture For: $$ z_1,\ldots,z_n $$ linearly independent over $\mathbb Q$, Schanuel's conjecture predicts: $$ \operatorname{trdeg}_{\mathbb Q} \mathbb Q \left( z_1,\ldots,z_n, e^{z_1},\ldots,e^{z_n} \right) \ge n. $$ A predimension-type quantity can be defined: $$ \delta(X) = \operatorname{trdeg}_{\mathbb Q} \mathbb Q(X,e^X) - \operatorname{ldim}_{\mathbb Q}(X). $$ Schanuel's conjecture predicts precisely that: $$ \boxed{ \delta(X)\ge0. } $$ ACTM interprets this as follows: > The total algebraic relations after the exponential map cannot excessively compress the source's $\mathbb Q$-linear degrees of freedom into a net deficit of degrees of freedom. One may also define a formal collapse quantity: $$ \mathcal C(X) = \operatorname{ldim}_{\mathbb Q}(X) - \operatorname{trdeg}_{\mathbb Q} \mathbb Q(X,e^X). $$ Then Schanuel is equivalent to: $$ \boxed{ \mathcal C(X)\le0. } $$ This provides the first formal bridge between ACTM and Schanuel. --- # 16. Why ACTM Does Not Start Directly from Schanuel ACTM deliberately restricts its first phase to: $$ A\subseteq\overline{\mathbb Q}. $$ The reasons are: 1. Algebraic inputs can be classified exactly; 2. Many known theorems are available for calibration; 3. Input relations can be clearly distinguished from operator-generated relations; 4. Finite-scale computable experiments can be established; 5. Arbitrary transcendental inputs are avoided at the outset; 6. One can gradually observe which structural quantities might generalize to Schanuel's general case. The research order is: $$ \boxed{ \text{Algebraic Seeds} \rightarrow \text{Mixed Seeds} \rightarrow \text{General Complex Tuples} } $$ rather than jumping directly to the final layer. --- # 17. The First-Phase Research Matrix ACTM v0.1 proposes first establishing a finite research matrix. Input types: $$ \mathcal A_1: \text{single algebraic numbers} $$ $$ \mathcal A_2: \mathbb Q\text{-linearly independent algebraic tuples} $$ $$ \mathcal A_3: \text{multiplicatively independent algebraic tuples} $$ $$ \mathcal A_4: \text{highly constrained algebraic tuples} $$ Operator types: $$ \mathcal O_1=\exp $$ $$ \mathcal O_2=\log $$ $$ \mathcal O_3=\sin,\cos $$ $$ \mathcal O_4=(x,y)\mapsto x^y $$ Add later: $$ \Gamma, \quad \text{elliptic}, \quad \text{modular}, \quad \text{period maps}. $$ --- # 18. The Minimal Experimental Unit Each ACTM experiment uses: $$ E=(A,\mathcal O,R_{\text{known}},R_{\text{pred}},V). $$ where: - $A$: algebraic seeds; - $\mathcal O$: the operator; - $R_{\text{known}}$: known relations; - $R_{\text{pred}}$: relation transformations predicted by the methodology; - $V$: a verification procedure comparing against existing theorems or formal systems. Outputs must be labeled: $$ \text{PROVED}, \quad \text{KNOWN}, \quad \text{DERIVED}, \quad \text{CONJECTURAL}, \quad \text{UNKNOWN}. $$ These labels must not be conflated. --- # 19. The Counterexample-First Principle ACTM should not search only for supporting examples. For any proposed structural rule: $$ H(A,\mathcal O) $$ priority should be given to finding: 1. Low-dimensional counterexamples; 2. Symmetry counterexamples; 3. Branch-choice counterexamples; 4. Periodicity counterexamples; 5. Counterexamples involving re-encoding of relations; 6. Degenerate cases where the output remains algebraic; 7. Spurious increments of degrees of freedom caused by functional equations. The principle is: $$ \boxed{ \text{Counterexample Search Before Generalization} } $$ --- # 20. Five Things That Must Not Be Confused ACTM strictly distinguishes: $$ \text{numerical complexity} $$ $$ \text{algorithmic complexity} $$ $$ \text{algebraic degree} $$ $$ \text{transcendence degree} $$ $$ \text{apparent randomness} $$ In particular, transcendence must not be inferred from statistical features of a decimal expansion. Nor can: - Geometric intuition; - Topological intuition; - Visual complexity; - High-dimensional representations; - Unpredictability; directly imply conclusions about transcendental numbers. --- # 21. The Methodology's 'Quasi-Global Random Quantifier' The difficulty of deciding transcendence can be understood as: $$ \forall P\in\mathbb Q[X]\setminus\{0\}, \qquad P(\alpha)\neq0. $$ In multiple variables: $$ \forall P\in \mathbb Q[X_1,\ldots,X_n]\setminus\{0\}, \qquad P(\alpha_1,\ldots,\alpha_n)\neq0. $$ In ACTM, 'quasi-global random quantifier' serves only as a methodological intuition, not as a probabilistic or formal-logical term. Its role is to remind us: > Directly excluding infinitely many candidate relations one by one is usually not the most natural proof strategy. ACTM therefore attempts to recast the global exclusion problem as: $$ \boxed{ \text{structure-preservation obstruction} } $$ or: $$ \boxed{ \text{dimension lower bound} } $$ --- # 22. ACTM's Core Conjectural Directions The following are research directions, not theorems. ## 22.1 The Combinatorial Potential Principle The capacity of certain algebraic tuples to generate transcendence is primarily determined by their relation types and operator compatibility, rather than their numerical size or decimal appearance. --- ## 22.2 The Incompatible Relations Principle If an input family of relations has an irreducible incompatibility with the operator's structural laws, the output may need additional transcendental degrees of freedom to accommodate the map. --- ## 22.3 The Conservation / Non-Excessive-Collapse Principle for Degrees of Freedom For certain structural operators, source degrees of freedom cannot be arbitrarily compressed into too few algebraic degrees of freedom. Schanuel's conjecture can be viewed as an extremely strong version for the exponential map. --- ## 22.4 The Generation-Before-Classification Principle Instead of first labeling an output 'algebraic / transcendental', first study: $$ \text{What generates it?} $$ $$ \text{Which input relations are preserved?} $$ $$ \text{Which degrees of freedom are added or eliminated?} $$ --- # 23. The Formal Research Route ## Phase A: Reconstructing Classical Theorems Objectives: - Reconstruct Lindemann-Weierstrass in ACTM terminology; - Reconstruct Gelfond-Schneider; - Reconstruct basic $\exp/\log/\sin/\cos$ transcendence examples; - Confirm that ACTM produces no erroneous new conclusions. --- ## Phase B: Relation Maps of Finite Tuples For small algebraic tuples: $$ n=1,2,3,4 $$ establish: $$ \text{input relation graph} \rightarrow \text{operator transport graph} \rightarrow \text{known transcendence outcome}. $$ --- ## Phase C: Designing Indicators of Degrees of Freedom Study whether: $$ \operatorname{CP}_{\mathcal O}(A), \quad \operatorname{AEP}_{\mathcal O}(A), \quad G_{\mathcal O}(A) $$ can be further formalized into comparable algebraic / geometric quantities. --- ## Phase D: The Exponential-Map Project Focus on: $$ A \mapsto (A,e^A) $$ and compare it with: $$ \delta(A) $$ the quantity above. --- ## Phase E: Expanding from Algebraic to Mixed Inputs Allow: $$ A=(a_1,\ldots,a_r,t_1,\ldots,t_s), $$ where $a_i$ are algebraic and $t_j$ are known or assumed transcendental. Study the coupling of degrees of freedom from different sources. --- ## Phase F: Schanuel-Type Generalization Only after stable invariants emerge from the preceding phases should we study: $$ X\in\mathbb C^n $$ in the general case. --- # 24. Minimal Success Criteria ACTM does not need to prove a new major conjecture at the outset. Completing any one of the following in the first phase would be valuable: 1. Find a structural quantity that describes several known transcendence theorems uniformly; 2. Discover a previously unobvious shared pattern of relation transport among known theorems; 3. Find a candidate invariant and eliminate erroneous versions with counterexamples; 4. Obtain a new provable lower bound for a restricted family of operators; 5. Establish a mechanically verifiable algebraic-seed / operator dataset; 6. Find a computable bridge between Schanuel predimension and lower-level known theorems. --- # 25. Failure Conditions If any of the following occur, a version of ACTM should be judged a failure or in need of rewriting: - 'Combinatorial potential' merely renames known transcendence degree; - The indicators cannot distinguish any known examples; - All predictions work only after the answer is known; - There is no provable connection between relation gaps and transcendental degrees of freedom; - Operator classification depends entirely on retrospective manual explanations; - No falsifiable propositions can be produced; - Consistency with known theorems cannot be maintained. --- # 26. ACTM's Research Philosophy ACTM does not understand transcendental numbers as 'mysterious nonalgebraic numbers'. It studies: $$ \boxed{ \text{Under which generating mechanisms the algebraic closure loses closure} } $$ and: $$ \boxed{ \text{Whether the degrees of freedom in this loss of closure can be structured, quantified, and predicted} } $$ The actual research object is therefore not a particular special constant, but: $$ \boxed{ \text{algebraic-to-transcendental transition} } $$ --- # 27. The One-Sentence Version > **Do not first seek regularity in the ocean of transcendental numbers; start from the algebraic world and classify which finite relation structures, under which generating operators, must, may, or are not yet known to produce new transcendental degrees of freedom.** --- # 28. The Core Main Chain $$ \boxed{ \text{Algebraic Seed} \rightarrow \text{Internal Relations} \rightarrow \text{Combinatorial Potential} \rightarrow \text{Escape Operator} \rightarrow \text{Relation Transport} \rightarrow \text{Relation Gap} \rightarrow \text{Transcendence Gain} } $$ Further: $$ \boxed{ \text{Known Theorems} \rightarrow \text{Calibration} \rightarrow \text{Invariant Search} \rightarrow \text{Restricted New Results} \rightarrow \text{Schanuel Interface} } $$ --- # 29. The Next Phase This document establishes only the methodology. Formal research in the next phase should begin from the most conservative entry point: $$ \boxed{ \overline{\mathbb Q} \xrightarrow{\exp} \mathbb C } $$ First establish: 1. Single algebraic seeds; 2. Binary algebraic tuples; 3. $\mathbb Q$-linearly independent tuples; 4. Highly redundant tuples; 5. A relation-transport matrix for the exponential function; 6. Retrospective tests against known theorems; 7. Candidate indicators of combinatorial potential; 8. The first counterexamples. Before completing this layer, do not directly claim proof progress on Schanuel's conjecture. --- ## Conclusion ACTM's core shift can be written as: $$ \boxed{ \text{What is transcendental?} \quad\Longrightarrow\quad \text{What generates transcendence?} } $$ Further: $$ \boxed{ \text{What generates transcendence?} \quad\Longrightarrow\quad \text{Which algebraic structures possess generative potential?} } $$ Thus, the transcendence problem is no longer merely an output-classification problem, but is recast as a dynamic problem involving: $$ \boxed{ \text{source structure} + \text{generating operators} + \text{relation transport} + \text{changes in degrees of freedom} } $$ the components above. This is the starting point of ACTM v0.1.