By @josebrox

613b16a / negative_results/README.md

Raw
# Obstructions and lower bounds

This directory contains proved negative results, separated from the constructive results in [papers](../papers). Each obstruction concerns the precise equations, arithmetic family, or geometric presentation stated in its manuscript. Failure of a replacement on all prepared inputs does not exclude a construction whose actual witnesses lie in a smaller, separately proved locus.

The [current construction table](../README.md) gives the positive bounds. No result in this directory establishes a number field whose ring of integers is not Diophantine in its fraction field.

## Pole predicates and contact functions

| Manuscript | Result and scope |
|:--|:--|
| [One-parameter genus-one criterion and shifted-abscissa obstruction](../papers/pole_parameters/genus_one_sound_criterion_20261120_v82.tex) | Every prepared scalar fiber is locally soluble, but an exact nonsquare-input example excludes rational points even for all three shifted abscissas and all multipliers supported on 2, 3, 5, 7. This is prepared scalar data, not a claimed actual detector tuple; the ten-coordinate actual-input criterion remains open. |
| [Raw abscissa preparation](papers/pole_parameters/raw_abscissa_preparation_20261117_v75.tex) | At an actual multiplication tuple and physical slope, both specified raw replacements give negative odd pole valuations at 76777 despite all pre-pole guards and real bounds. Three exact implementations certify the obstruction. Other tuples and a new completeness argument remain open. |
| [Odd halfpoint cosets](papers/pole_parameters/odd_coset_raw_20261117_v76.tex) | Restricting to odd inputs and odd multiples of P/2 does not automatically repair the raw squared variant: an actual physical slope progression has pole valuation -11 at 85817. This does not assert selected-prime or complete fixed-place preparation, or global failure of the coset. |
| [Native pole witnesses and ramification](papers/pole_parameters/native_pole_ramification_20261117_v75.tex) | A built-in norm at an actual seven-contact tuple gives no native witness. More generally native witnesses force even ramification above negative odd contact poles, with genus at least 1+d(r-4)/4. The specified abelian families have constant slopes. Individual rational points, constant-slope families and global realization remain open. |
| [Prepared products and adaptive classes](papers/pole_parameters/adaptive_prepared_pairs_20261117_v76.tex) | Two exact prepared abstract native products exclude every rational scalar in the two specified adaptive square tests, by universal local arguments at 100003. They have not been lifted to actual elliptic tuples and do not disprove 21/9 realization. |
| [Vertical valuations and affine derivative slopes](papers/pole_parameters/vertical_and_fixed_slope_20261117_v70.tex) | No nonconstant rational abelian family through a prepared point at fixed input; the derivative statement also fixes the root, while the general statement permits all nine point roles to vary. A separate derivative proposition forces the root and slope to be constant whenever the slope is a fixed rational affine function of the scale. Individual rational witnesses and unrestricted simultaneous root-and-slope variation remain open. |
| [Curves and abelian families in actual-trace loci](papers/pole_parameters/curves_and_abelian_families_20261117_v69.tex) | At fixed rational input, every fixed algebraic curve in either specified actual-trace source has finitely many rational points; no genus-zero or genus-one curve over the rationals maps nonconstantly into it. Abelian families have constant slope in the general source, and in the derivative source when the root point is fixed. Independent point roles and the stated finite menus are included. Individual witnesses and both realization criteria remain open. |
| [Cofinite six-contact Möbius coupling](papers/pole_parameters/cofinite_mobius_coupling_obstruction_20261117_v64.tex) | At every shift in the stated obstructed progression, an infinite progression of positive inputs defeats the twelve-symmetry menu, every fixed finite exponent menu and the prescribed root. Finite input-factor repairs cannot restore completeness. Other coefficient shifts, arbitrary variable rescalings and other nine-variable sources remain open. |
| [Input-only weight false witnesses](papers/pole_parameters/input_only_weight_false_witnesses_20261117_v61.tex) | The specified degree-eight helper supplies an abc input-height bound but admits noninteger inputs `2+1/q` outside a fixed finite set. The proof includes compatible reference indices, local interpolation, all-prime preparation and whole-character guards. |
| [Denominator support and fixed rational height coordinates](papers/pole_parameters/denominator_support_height_obstruction_20261117_v60.tex) | A radical supported in one reference denominator bounds the quotient for the stated height inequalities. Every fixed rational carrier generically nonconstant in the second point has only finitely many intended integer inputs, including degenerate specializations by Siegel. Other support architectures are not excluded. |
| [Square-rescaled weight perturbation](papers/pole_parameters/square_rescaled_weight_perturbation_20261117_v60.tex) | Under the stated strict preparation, square-rescaling an input-independent weight class preserves witnesses at nearby noninteger inputs. Additional closed input equations are outside the theorem. |
| [Fixed-slope trace incidence](papers/rational_fibers/fixed_slope_trace_finiteness_20261117_v59.tex) | The actual first trace has finitely many rational realizations on a fixed smooth genus-one trace curve, at fixed input and prepared slope, with all source point roles independent. Individual witnesses, varying slopes, point-dependent curves, the conic and the new derivative source are not excluded. |
| [Actual multiplication surfaces](papers/rational_fibers/actual_multiplication_surface_20261117_v58.tex) | Nonconstant rational slope coupling on the full fixed-index multiplication surface cannot make the stated safely twisted trace tests rational-function squares. Individual rational points, constant slopes, proper subfamilies and ramified covers are not excluded. |
| [A prepared covering obstruction](papers/pole_parameters/prepared_sha_obstruction_20261117_v57.tex) | The prepared input `112/225` gives an everywhere locally soluble one-coordinate pole curve without a rational point; an exact Cassels pairing excludes the required abscissa class. No actual detector tuple is asserted. |
| [Intrinsic specializations](papers/pole_parameters/intrinsic_specializations_20261117_v57.tex) | No finite rational-function menu for the second coordinate works on all prepared inputs. The rank-one class theorem restricts the specified congruent curves; adaptive relations and higher rank remain possible. |
| [Deterministic order-three coupling](papers/pole_parameters/order_three_inert_obstruction_20261117_v56.tex) | For the specified height substitution with prescribed root `A=[X]P`, every positive odd shift has infinitely many positive input failures at `p=81421`, uniformly in the auxiliary tuple. Both stated weights are covered. Other substitutions, root labels and free slopes remain open. |
| [Rational conditional products](papers/pole_parameters/rational_detector_products_20261117_v56.tex) | A rational actual product `W` satisfying the pole implication forces its argument into a finite set of genus-two twists. Rational products in both coupled blocks allow only finitely many integer inputs. Nonrational products and the general source are outside the theorem. |
| [Fixed native square covers](papers/pole_parameters/fixed_flat_fibers_20261116_v53.tex) | The actual tuples `A=[X]P`, `H=3P`, `U=[3X]P`, at `X=1,2` with coefficient fixture `B=3` and exponent four, give geometrically integral simultaneous covers of genera 236 and 332. Four rational points at infinity are outside the source domain. Finite rational points and varying auxiliary indices remain open. |
| [Fixed normalized traces](papers/pole_parameters/fixed_trace_obstructions_20261116_v51.tex) | The original general scalar has an actual-input obstruction at `X=4`, `p=252877` to simultaneous unit traces, persisting under the specified native normalizations. No finite fixed set of trace targets gives uniform completeness. For actual conditional six-contact tuples, simultaneous unit traces require comparison scale `a<147456`, excluding every prepared real slope when `B ≥ 19`. Varying norm coordinates and modified scalar sources are not excluded. |
| [Automatic native levels](papers/pole_parameters/automatic_level_finiteness_20261116_v48.tex) | The conditional level `a=wNorm(W)` is inadmissible at every positive odd contact for every guarded auxiliary tuple. Admissibility forces finitely many genus-two arguments, excluding uniform completeness even after fixed finite rational rescalings. Detector-dependent bounded levels and other scalar constructions remain open. |
| [Finite nontrivial complement classes](papers/pole_parameters/finite_nonsquare_class_obstruction_20261116_v47.tex) | Every finite set of nontrivial signed rational square classes fails for both complements on an arithmetic progression of inputs in the original general scalar source, uniformly in the odd exponent and all rational level multipliers. Arbitrary guarded auxiliary tuples are covered. The rational-square class and redesigned scalars remain possible. |
| [Conic-chain predicates](papers/pole_parameters/conic_chain_obstruction_20261115_v41.tex) | Every fixed rational coefficient in the specified five-auxiliary family has arbitrarily small prepared local counterexamples. |
| [Automatic first conics](papers/pole_parameters/automatic_conic_obstruction_20261115_v41.tex) | Dense prepared counterfamilies obstruct the remaining pole equations, including on the completely split six-contact map. The argument uses only Tunnell's unconditional necessary criterion. |
| [Fixed pole parameters](papers/pole_parameters/fixed_pole_parameters_20261109_v17.tex) | Constant first-conic values and the specified fixed conic parameters fail on prepared inputs. The fixed-parameter genus-one examples are global obstructions, consistent with everywhere local solubility. |
| [A quartic pole predicate](papers/pole_parameters/quartic_pole_obstruction_20261109_v17.tex) | The stated quartic predicate and finite menus of its fixed coefficients fail on explicit prepared local counterfamilies. |
| [Combining pole certificates](papers/pole_parameters/combining_pole_certificates_20261109_v16.tex) | Sharing the first-conic value requires matching input square classes; the paper also gives reflection and product-height limitations. This is distinct from sharing the normalized parameter `τ`. |
| [Products of contact functions](papers/pole_parameters/product_height_obstructions_20261110_v19.tex) | Exact square-product and fixed-square-class families have bounded odd radical while the comparison scale grows. |
| [Rational parity combinations](papers/pole_parameters/rational_combinations_of_contacts_20261110_v20.tex) | No fixed rational parity combination has the stated detection property on the specified independent-contact family. |
| [Quadratic contact covers](papers/pole_parameters/quadratic_contact_covers_20261110_v21.tex) | Square-class matching produces genus-four covers; the proposed negative-product section has an inert-prime obstruction. |
| [The ordered square-class surface](papers/pole_parameters/ordered_squareclass_surface_20261115_v33.tex) | General type excludes the stated rational parametrizations and rational or genus-one general-fiber fibrations. Special arithmetic witness families remain a separate question. |
| [Reflection matching](papers/pole_parameters/reflection_obstruction_20261115_v33.tex) | The exact reflection restriction omits a true integer for every sufficiently large odd shift under the retained source premises. |
| [Fresh-prime square classes](papers/pole_parameters/fresh_prime_squareclass_obstruction_20261116_v42.tex) | The specified fresh-prime witnesses cannot have matching first-pole square classes; fixed comparison scales also give the stated higher-genus covers. |
| [Fixed-input K3 surfaces](papers/pole_parameters/fixed_input_k3_surfaces_20261116_v42.tex) | The full one-pole surface is a K3 after resolving four A3 points. The common-parameter product for fixed inputs has no rationally connected smooth projective model. Rational witnesses and varying-input constructions are not excluded. |
| [Rational sections](papers/pole_parameters/rational_sections_20261116_v42.tex) | For prepared fixed inputs, a rational-function section exists exactly when 17(17−b) is a rational square. The general affine and projective criteria include every boundary case. Individual rational fibers, shared rational parameters and multisections are not excluded. |
| [Actual square-complement neighborhoods](papers/pole_parameters/actual_square_complement_neighborhoods_20261116_v43.tex) | For every odd source exponent, a legitimate independent-index family has nonsquare complements throughout a fixed 23-adic neighborhood. Other detector choices and common rational fibers remain possible. |
| [Square-complement normalization](papers/pole_parameters/square_complement_normalization_20261116_v43.tex) | The square-class-preserving map 1156b/(17+b)² forces square complements but destroys the original pole predicate on an explicit 409-adic counterfamily. This concerns the pole subsystem. |
| [Fixed-class synchronization](papers/pole_parameters/fixed_class_synchronization_20261116_v43.tex) | For distinct positive pole scales, fixed twist classes give genus 1+2^r(r−1), hence finitely many rational parameter values when r≥2. The full common-parameter variety has Kodaira dimension one. One common fiber and varying class tuples are not excluded. |
| [Abelian parameter spaces](papers/rational_fibers/abelian_parameter_obstruction_20261116_v44.tex) | No uniform rational square-complement or original-pole identity on an abelian parameter space with nonconstant detector argument, even with varying twist classes. Fixed-argument rational fibers, ramified covers and individual witnesses are not excluded. |
| [Fixed-argument complement ramification](papers/rational_fibers/fixed_argument_complement_ramification_20261116_v44.tex) | For allowed fixed source data and positive detector weight, the complement double cover on the full independent auxiliary product is geometrically ramified. No surjective abelian coupling gives a uniform section. Proper coupled subvarieties, individual rational tuples and common pole fibers without square complements remain open. |

The [pole-compression guide](docs/pole_compression_obstructions_20261115_v41.md) gives the recent explicit counterfamilies. Their exact checkers remain in [conic_chain_v41](../computation/pole_parameters/conic_chain_v41) and [automatic_conic_v41](../computation/pole_parameters/automatic_conic_v41). The positive [shared-parameter local theorem](../papers/pole_parameters/shared_parameter_local_solubility_20261115_v41.tex) remains with the constructive results.

The [fixed-fiber literature audit](docs/fixed_fiber_literature_20261116_v43.md) records the precise hypotheses of Kummer and genus-one fibration theorems that prevent their direct application to the stated common-parameter family.

## Elliptic points and integer labels

| Manuscript | Result and scope |
|:--|:--|
| [Quadratic root restrictions and point-role couplings](papers/integer_labels/point_role_restrictions_20261117_v66.tex) | Any nonzero quadratic root-index relation, or finite menu, omits infinitely many positive inputs in every fixed arithmetic progression. Two independent fixed group words, including finite menus of pairs, are excluded within the exact multiplication family even after reordering its construction. The paper also gives single cross-root and scalar-reference obstructions. Arbitrary numerator-index families and other couplings remain open. |
| [Half of a scalar helper](papers/elliptic_charts/half_scalar_helper_20261117_v70.tex) | The specified 21-coordinate interface has exactly seventeen quadratic identities, all compatible with an incorrect nonzero scalar quotient at a nontorsion point pair. An exact finite-field minor and a characteristic-zero specialization proof certify the complete identity space. Additional global coordinates, other helpers and other charts remain outside the theorem. |
| [Quadratic tangent completion](papers/elliptic_charts/tangent_completion_20261116_v48.tex) | Exact false-height families survive every quadratic identity in the specified ten-field-coordinate compressed pairing, every rational-component quadric including conjugate products, and every quadric after adjoining the physical trace. The latter trace remains arbitrary on the false family. Different embeddings, added helpers and separately proved restricted witness families are not excluded. |
| [One-parameter elliptic presentations](papers/elliptic_charts/elliptic_chart_obstruction_20261107_v7.tex) | One retained rational parameter cannot give the specified finite algebraic presentation of infinitely many elliptic points. |
| [Structural constraints](papers/elliptic_charts/structural_constraints_20261109_v13.tex) | Dimension, label-graph, common-index and square-class obstructions for the stated elliptic witness models. |
| [Products of elliptic indices](papers/elliptic_charts/index_products_and_finite_correspondences_20261110_v21.tex) | The index-product graph is Zariski dense, obstructing a fixed finite correspondence on the stated infinite label sets. |
| [Label-dependent progressions](papers/elliptic_charts/progression_density_20261114_v29.tex) | Allowing infinite progressions depending on the label does not provide the proposed algebraic elimination of the reference point. |
| [Mixed reference signs](papers/elliptic_charts/mixed_reference_signs_20261116_v42.tex) | An explicit progression of doubled points lies in opposite torsion neighborhoods at the two primes above 76537. The reference chart alone cannot force agreement of the corresponding formal-parameter signs. |
| [Shared reference test obstructions](papers/elliptic_charts/shared_reference_test_obstructions_20261116_v43.tex) | Exact local cancellation defeats the stated squared reference factor and the whole-factor odd-cube repairs, including a positive-valuation comparison with unit error. These are factor-level obstructions, not complete-source counterexamples. |
| [Duplicated reference signs](papers/elliptic_charts/duplicated_reference_signs_20261116_v44.tex) | An explicit allowed half-integral reference progression has opposite local translation signs after duplication at a valid soundness contact prime. This concerns the reference branches, not a solution of the full detector. |
| [Rational Pell approximations](papers/integer_labels/rational_pell_approximation_obstruction_20261116_v42.tex) | Arbitrarily accurate rational approximation at 19, even with the exact norm equation, does not replace the integral-label hypothesis. |

## Quadratic compilation of quartic sources

| Manuscript | Result and scope |
|:--|:--|
| [Sequential six-unit detector bases](papers/quartic/sequential_unit_detector_obstruction_20261117_v70.tex) | No integer diagonal Laurent compilation of the stated literal factor design works with any of the 2,834 sequential six-unit bases, for any assignment of its quadratic tests. A finite integer certificate reduces to two bases and a direct incompatible-weight proof. Other unit families, powers, equation combinations and nonmonomial coordinates remain outside the theorem. |
| [Physical detector weights and Laurent scaling](papers/quartic/physical_detector_scaling_20261117_v67.tex) | One literal reference row and three specified physical factors cannot all remain quadratic under diagonal Laurent coordinate changes and monomial row division, for either stated unit basis and arbitrary real weights. A separate shared-factor obstruction also holds. Exact certificates and a complete coordinate-map audit accompany the proof. Different factor pairings, reference equations, row combinations and nonlinear substitutions are outside its scope; no general variable lower bound is asserted. |

## Heights, inputs and general lower bounds

| Manuscript | Result and scope |
|:--|:--|
| [Quadratic input maps](papers/input_normalization/quadratic_input_obstruction_20261109_v12.tex) | The required universal local square condition is incompatible with the specified quadratic input normalization. |
| [Scalar measurements of degree at most three](papers/scalar_heights/cubic_scalar_height_obstruction_20261109_v11.tex) | Infinite rational fibers obstruct the proposed uniform scalar-height control on ambient rational pairs; special arithmetic subsets are not covered. |
| [Restricted abc inputs](papers/restricted_abc/restricted_abc_inputs_20261109_v17.tex) | The required scalar families fail the bounded-denominator or asymmetric-size hypotheses of the specified restricted estimates. |
| [Degree and coordinate lower bounds](papers/lower_bounds/lower_bounds_20261109_v15.tex) | Lower bounds for the classes of defining equations, quadratic systems, and nonnegative polynomials distinguished in the paper. |
| [Two independent squares](papers/contact_heights/two_square_height_20261117_v58.tex) | For power-of-two exponents, fixed-output families have unbounded rational input height even in a positive interval. The actual inner source form is not imposed; this is not a full-detector counterexample. |
| [Four-contact radicals](papers/contact_heights/four_contact_radical_obstruction_20261116_v42.tex) | The displayed four-contact family has bounded odd radical and unbounded height, including points in a fixed positive real interval. |
| [Limits of auxiliary hypotheses](papers/arithmetic_hypotheses/limitations_of_auxiliary_hypotheses_20261116_v42.tex) | Prime assignments do not control valuation parity; polynomial injections do not supply finite algebraic recovery. The separate implication from Mazur's connected-component conjecture is explicitly conditional. |

## Number-field constructions

| Manuscript | Result and scope |
|:--|:--|
| [Rank-stable adjacent contacts](papers/number_fields/rank_stable_adjacent_contacts_20261117_v63.tex) | After a fixed multiple descends, adjacent identity reductions above one base prime force a fixed non-torsion point to reduce to the identity. Only finitely many base primes can occur; torsion, ramification and finite shift menus are included. This restricts one detector architecture. |
| [Elliptic ratios and integrality](papers/number_fields/elliptic_ratios_and_integrality_20261109_v13.tex) | Local solubility and simultaneous approximation do not make the stated ratio construction surjective; finite rational combinations still omit infinitely many integers. |
| [Finite correspondences and denominators](papers/number_fields/finite_correspondences_and_denominators_20261109_v14.tex) | Finite-output algebraic constructions omit the specified progression terms; bounded-denominator hypotheses obstruct the stated rational or genus-one constructions. |
| [Complex multiplication and trace cancellation](papers/number_fields/complex_multiplication_and_trace_cancellation_20261110_v19.tex) | Adjoining `i` changes the rank and permits unbounded points with cancelling traces, obstructing direct transfer of this rank-and-height argument. |
| [Norm and trace graphs](papers/number_fields/norm_trace_graph_obstruction_20261116_v42.tex) | Graphs on one ambient extension-field input are Zariski dense and have no quantifier-free ring-language definition. Existential definitions are not excluded. |
| [Unchanged formulas after finite extensions](papers/number_fields/unchanged_formula_obstruction_20261116_v42.tex) | A fixed formula, or finite menu of formulas, cannot define the rings of integers unchanged over every finite extension. The paper includes degree bounds and an explicit quadratic-extension example. |
| [Products of elliptic abscissas](papers/number_fields/abscissa_product_obstruction_20261116_v42.tex) | A valuation obstruction excludes 5 from every finite multiplicative combination of rational abscissas on the stated positive-rank curve. |

The positive [definability reductions](../papers/number_fields/number_field_definability_20261109_v12.tex) are kept separately. A failure of one formula after extending its coefficient field is a statement about that formula, not a counterexample to the existence of every possible Diophantine definition.

## Sharpness and necessary hypotheses of positive theorems

An optimality clause or an example showing why a hypothesis is needed remains attached to the positive theorem it qualifies. This includes the [optimal quadratic-field square predicate](../papers/square_predicates/quadratic_square_predicates_20261108_v10.tex), the exact [product-height equality criterion](../papers/elliptic_heights/product_height_20261113_v23.tex), and the exceptional-constant condition in the [translated-point chart](../papers/elliptic_charts/fixed_difference_charts_20261110_v18.tex).

## Reproduction

The [separate theorem catalog](data/results.json) contains the obstruction records extracted from the historical results metadata. The current positive construction table remains in the main README.

The computations linked by the papers retain their existing paths when they are shared with constructive results. The repository's verification commands and manifest cover both directories. The manuscripts state the infinite results; finite examples and coefficient checks certify only the computations they actually perform.