Status: conjectural research synthesis and experiment-routing document; not a theorem, not a theory-promotion record, and not a Public API contract.
Required prior reading: MATHEMATICAL_CORE.md,
which records the present objects, constructions, laws, and boundaries. Read
RESEARCH_STATUS.md next for dated evidence and explicit
nonclaims. This document answers a different question: what unifying
possibility is the project deliberately testing, and why do its apparently
distant research lines belong to one programme?
When this programme, the Mathematical Core, and a detailed proof artifact use different strengths of language, the proof artifact and Core determine what has been earned. This document determines neither maturity nor API status.
Process Geometry is guided by a conjectural synthesis stronger than the claim that many processes admit geometric representations.
Arithmetic generativity. Arithmetic primitives, their free process compositions, and their objectified higher-rank operations may generate not only presentation geometries, but also effective calculi, task-relative information structures, complexity measures, statistical constructions, and new compositional ranks.
Addition and Multiplication provide the first substantial and computable model of this mechanism. They are not assumed to be the final generic ontology. The broader question is whether arithmetic-generated presentations can serve, for a sufficiently wide class of declared process--task systems, as task-sufficient, effectively calculable, and economically responsible process languages.
This is the project's Arithmetic Universality programme. It is an active research direction, not an assumption made by the package.
The programme is intentionally more than a proposal for a common geometric skeleton. Its target is a compatible executable system:
arithmetic primitives
-> free compositions and literal histories
-> presentation freedom, charts, and transported scale
-> geometry and effective change calculus
-> task-relative forgetting, fibres, residuals, and decoders
-> cost, measure, probability, and optimization
-> objectification and higher-rank free composition.
Geometry is one generated layer of this system. A successful theory must also say how quantities are evaluated, how changes propagate, how information is forgotten or retained, how costs are charged, and how a stable process schema becomes a new primitive without losing effective lowering semantics.
The programme keeps four terms separate.
- Arithmetic generativity names the proposed mechanism by which process composition and objectification generate geometry, analysis, information, complexity, and higher rank.
- A/M geometry and calculus name the present Addition/Multiplication base case: the first concrete continuous process geometry and effective function language developed by the repository.
- The arithmetic tower names the candidate vertical pattern in which a stable lower-rank process or action is objectified into a new primitive with free higher-rank composition and compositional lowering.
- Arithmetic Universality names the scope question: how broadly arithmetic-generated presentations satisfy the semantic, analytic, statistical, and economic obligations below.
The phrase arithmetic-generated process geometry is deliberate. The project is not renaming the established number-theoretic field of arithmetic geometry. Nor does arithmetic mean an unrestricted permission to encode a target solver inside a sufficiently large expression. The primitive grammar, admissible charts, adapters, residuals, and costs must be frozen for each claim.
AM is therefore neither demoted to an incidental example nor promoted to an unproved foundation. It is the first hard model organism, the current effective continuous analysis family, and the principal candidate base layer of a broader arithmetic-generative mechanism.
The programme decomposes its strongest conjecture into five obligations and one cross-cutting admissibility condition. These identifiers route evidence; they are not maturity levels, and satisfying one does not imply the others.
| ID | Obligation | What would count as real evidence | Typical trivialization to reject |
|---|---|---|---|
| U1 | generative presentation | an arithmetic process grammar produces a task-sufficient presentation while preserving declared composition and continuation semantics | an arbitrary redundant encoding whose decoder contains the original solver |
| U2 | effective analysis | the presentation supports symbolic and/or numerical evaluation, a native change calculus, closure or controlled extension, certificates, and failure semantics | importing ordinary additive tangent or jet machinery without deriving its transport at the relevant arithmetic rank |
| U3 | information--complexity compatibility | task quotients, residuals, stopping sections, cost cocycles, coding, and optimization coexist on one auditable history presentation | calling distinguishability bits, runtime, entropy, and phase volume the same quantity |
| U4 | statistical--macroscopic compatibility | conditional fibre laws, transverse ensemble weights, correlations, coarse-graining, and macroscopic response are connected by declared adapters and limits | attaching an arbitrary probability law or closure and crediting it to arithmetic structure |
| U5 | objectification and rank closure | a stable process schema becomes a reusable primitive with genuinely new free composition and effective all-composite lowering | treating abbreviation, a quotient point, or a stable fibre as a new rank by name alone |
| E | covariance and economy | the preceding constructions commute under admissible chart or presentation changes up to declared residual/error, with bounded representation, compilation, decoder, and evaluation costs | rescuing every failure by unbounded fibres, residuals, chart data, or hidden oracle work |
The strongest programme would satisfy all six obligations on a broad and independently characterized class of processes. Useful partial theorems may instead establish one obligation locally, prove an obstruction, or identify a maximal domain on which a weaker universality statement survives.
The repository studies complexity, entropy, and statistical mechanics not because all three use logarithms, volumes, or optimization, and not because they have already been identified. They test whether one arithmetic-generated process language can carry increasingly demanding kinds of structure.
arithmetic-generated process history
-> task evaluation and forgetting
-> measured/costed fibres and stopping sections
|-> optimal history and coding laws complexity
|-> counting or measuring forgotten fibres entropy
`-> conditional laws and transverse tilts statistical mechanics
-> coarse-graining and macroscopic response.
- Complexity pressure asks whether process composition supplies a transportable ruler, whether cost is accumulated before or after quotient, whether stopping sections are task-sufficient, and whether Bellman/Huffman optimization remains presentation-covariant after compilation and decoding costs are charged.
- Entropy pressure asks when task-relative multiplicity becomes a counted or measured fibre, which reference measure and units make logarithms legal, and which information remains as a continuation-visible residual.
- Statistical-mechanics pressure asks whether microscopic composition, correlation fibres, conditional laws, ensemble mixing, coarse-graining, irreversibility, and macroscopic closure can remain inside one effective change-and-adaptation language.
The disciplinary nonidentifications in the Mathematical Core are therefore not a retreat from synthesis. They make the endpoints of future bridge theorems well typed. The programme conjectures a common generative origin; it does not infer
complexity = entropy = action volume = thermodynamic state count
from analogy. A bridge must declare its task, measure, units, limiting operation, covariance law, residual, and cost.
The Boltzmann--BBGKY line is especially important in this role. It is not merely an application of a finished geometry to a physics problem. It tests whether exact equivalence can give way to typed semantic adaptation, whether forgotten correlations can be located in measured or filtered fibres, and whether a process-native response calculus survives noninvertible and probabilistic reduction.
This table records why major research lines belong to the programme. It does
not replace the dated achievements and boundaries in
RESEARCH_STATUS.md.
| Pressure route | Primary obligations | Current evidence owners | Programme-level question |
|---|---|---|---|
| AEG translation, multiplication, and mixed lowering | U1, U5 | 50-aeg-translation-objectification-rank-lowering.md, 51-aeg-addition-multiplication-rank-transition.md |
does arithmetic process composition generate a genuine next-rank language rather than an abbreviation? |
| pendulum, moving observers, and AM chart search | U1, U2, E | vignettes/simple-pendulum.md, moving-am-observer, am-conformal-chart-normal-forms |
can a nontrivial physical process acquire a discovered, covariant, and computationally economical arithmetic atlas and calculus? |
| canonical histories, Bellman/Huffman, and downstream pruning | U3, E | 56-am-universal-history-recalibration.md, lonely-runner |
can task semantics and arithmetic rulers change a real computation without hiding quotient, decoder, or implementation cost? |
| local-field/projective fibres, partitions, and change actions | U1, U3, U5 | local-field-projective-process-geometry |
which presentation data descend, which remain fibred, and when does an objectified action support a higher calculus? |
| stochastic first passage and chart transport | U2, U3, E | stochastic-feedback-trap-first-passage |
which change law is forced when nonlinear charts meet stochastic variation? |
| thermodynamic objectification and partition towers | U3, U4, U5 | 63-thermodynamic-objectification-and-partition-towers.md |
do measure, ensembles, partition structure, and objectification share an arithmetic-generative interface? |
| Boltzmann--BBGKY and the H-theorem frontier | U2, U4, E | boltzmann-bbgky-h-theorem |
can correlations, semantic adaptation, response, and macroscopic closure be handled without importing the answer as an oracle? |
The same result may support generic Process Geometry while leaving Arithmetic Universality untouched. Every evidence report must therefore separate:
generic process-geometry result
arithmetic-specific mechanism
universality obligation pressured
claim not earned.
Mere encodability is too weak. For a declared class of process--task pairs ((P,Q)), a substantive theorem would need an arithmetic-generated presentation (A_{P,Q}) and comparison data satisfying, in an appropriately typed form:
- task adequacy: every declared observation and continuation is preserved exactly or within a declared topology, horizon, and error budget;
- compositionality: legal source composition transports coherently to the arithmetic presentation and, across ranks, lowers again;
- native variation: the necessary change law is derived or transported at the relevant arithmetic rank rather than assumed from additive calculus;
- information accounting: forgotten data, non-descent, branch choices, correlations, and reconstruction obligations remain explicit;
- effective evaluation: symbolic or numerical evaluation terminates or fails under a declared certificate and resource contract;
- covariance and economy: competing admissible charts give compatible task values, and total lift/search/decoder/evaluation overhead is bounded.
Global presentation by one AM chart is not the only possible successful form. A more plausible theorem may provide a task-relative arithmetic atlas whose local AM charts are glued by admissible transition data, with global obstructions carried by explicit cocycles, fibres, or residuals. Such a theorem would still have to bound those additions; otherwise the atlas claim would collapse into unrestricted encoding.
No present result proves this package for a generic class of processes.
The programme loses mathematical content if every counterexample can be absorbed by adding a larger fibre, a stronger decoder, another chart, or a new rank. A universality experiment must freeze in advance:
- the process and task class;
- the arithmetic primitive grammar and allowed objectification steps;
- admissible chart and adapter families;
- residual and decoder budgets;
- symbolic and numerical evaluator contracts;
- baseline representations and total cost axes;
- the failure that would reject or narrow the tested claim.
The following outcomes count against a strong universality claim:
- the target dynamics or solution is smuggled into a connection, decoder, closure, measure, or oracle;
- task adequacy requires residual growth comparable to retaining the full source history without compensating structure or economy;
- chart transition or change calculus leaves the declared arithmetic language in an uncontrolled way;
- an allegedly intrinsic ruler, entropy, or ensemble depends on an arbitrary undeclared frame, measure, scalarization, or limiting path;
- higher-rank notation supplies no new free composition or no effective all-composite lowering;
- the arithmetic presentation gives no structural theorem, obstruction, discovery advantage, or net computational economy over honest baselines.
A negative result may reject only one obligation or process class. The
surviving exact laws, algorithms, and obstructions remain assets, as recorded
in RESEARCH_STATUS.md; but they must not be used rhetorically to preserve a
stronger failed conjecture.
Every substantial Sonnet or extraction candidate that bears on this programme should record:
Universality pressure: U1 / U2 / U3 / U4 / U5 / E
Arithmetic-specific mechanism:
Generic Process Geometry result:
Evidence gained:
Universality claim not earned:
Kill condition:
Not every Sonnet must pressure Arithmetic Universality. A problem-local result may be valuable without doing so, and a generic Process Geometry obstruction may show that the arithmetic conjecture was posed at the wrong level. The field must never be chosen merely because arithmetic terminology can be overlaid on it.
The repository keeps three registers distinct:
RESEARCH_PROGRAM.md what unifying possibility is being tested
MATHEMATICAL_CORE.md what mathematical synthesis is currently responsible
RESEARCH_STATUS.md what the dated evidence has actually earned
THEORY_MAP.md locates dependencies and maturity;
ENGINEERING_ARCHITECTURE.md makes the effective-calculation obligations
operational; governance controls promotion. None substitutes for the others.
The repository has earned concrete A/M analysis, exact arithmetic objectification and lowering calibrations, task-relative quotient and fibre laws, several covariance and non-descent obstructions, and one downstream computational transfer. These results justify sustained pressure on the Arithmetic Generativity programme.
They do not yet establish one universal arithmetic carrier, one universal calculus, a canonical complexity ruler, a derivation of entropy from process geometry, a generic statistical closure, or a complete arithmetic rank tower.
The intended stance is therefore:
propose a common arithmetic-generative origin boldly; keep complexity, entropy, action volume, probability, and thermodynamic structure distinct until typed bridge laws connect them; and make every universality claim pay for its semantics, residuals, decoders, and computation.