By @josebrox

4bdace6 / docs/research_directions.md

Raw
# Research directions

The [v82 continuation](results_20261120_v82.md) proves general quartics in **163 and 179 total variables**, retaining the source arithmetic premises. The partial-trace scalar and a scaled point chart remove two field components; a positive weighted trace removes another rational coordinate from the ineffective family. The ordered global witness construction, zero-coordinate branches, full quadratic maps, and coefficient expansions are supplied. The fixed 179-variable member retains an effective unweighted height argument. Contributions originating in **notes of Gabe Conant with Astra 6.2** remain distinguished from the subsequent project work.

The current unrestricted-degree counts remain **25** and **11**, and the ordinary-abc quartic remains **56**. No reduction of those counts is proved in v82. Every count includes the input. The priorities remain variable count, degree, then coefficient height; no global minimum or complete independent audit of all inherited arithmetic foundations is asserted.

## Immediate constructive targets

1. **Identify an external input with the new elliptic quotient.** The [prepared divisibility theorem](../papers/elliptic_divisibility/prepared_eds_divisibility_20261120_v82.tex) forces `h | j` with two rational elliptic point charts and one pole block. Every nonzero integral quotient `j/h` occurs after choosing a suitable reference. Its five-coordinate specification does not yet enforce `X=j/h`. A height argument must handle an arbitrarily large quotient; a carrier depending only on the reference leaves this uncontrolled, while including the second point can demand more contact-prime support than the reference provides.
2. **Prove rational realization at actual detector inputs.** The [genus-one criterion](../papers/pole_parameters/genus_one_sound_criterion_20261120_v82.tex) replaces the two pole parameters by one if the actual fibers have non-torsion rational points. Local solubility is now proved at every place for prepared scalar inputs. An exact prepared counterexample, including a broad shifted-abscissa menu, shows why a global argument using the actual contact and point functions is necessary. The counterexample is not asserted to be an actual detector value. The proposed condition that `1+d` be a rational square remains unproved at the actual tuple.
3. **Remove the remaining general-quartic square coordinate.** The relation `tau_hat*sigma=R^4` still uses the shared `R²` register. A different scaling must keep the reference-point equations and every scalar/native-pole equation quadratic, with all zero fibers checked. The ordered witness construction is now available and need not be conjectured.
4. **Reduce the ordinary-abc quartic below 56.** A further paired-contact rechart or sharing with the point equations must remove a coordinate without adding a new cubic product. The secondary Euler identity exposes such a product but does not yet eliminate it.

The [v81 continuation](results_20261119_v81.md) remains the proof of the 56-variable ordinary-abc quartic and the separate ten-coordinate rational-subfield selector. The latter does not lower a variable bound for the integers over the rationals.

## Consequences of the audit of Gabe Conant’s notes with Astra 6.2

The shared norm selector and stronger odd-power packing provided the preceding 165/180 constructions and stronger local divisibility; v82 gives the further reductions described above. The six- and eight-contact weak-abc estimates give explicit variable/hypothesis tradeoffs; the eight-contact quartic family covers every fixed exponent below 3/2. Positive detector factors, square-class normalization of the intrinsic pole, and ordered removal of automatic point guards substantially lower the finite-algebra degree. These are proved changes to complete constructions.

Further general variable reduction still requires a new arithmetic argument or a smaller guarded quadratic compilation. The [full-pair parity obstruction](../papers/quartic/hyperbolic_parity_20261119_v79.tex) proves failure only within its specified retained chart and termwise integral Laurent model; it does not establish optimality. The general 21 and ordinary-abc 9 targets continue to require rational realization at actual physical traces. The new diagonal Hasse and zero-ordinate obstructions concern specific universal specializations on all prepared inputs and cannot be promoted to nonexistence of 21/9 constructions.

The older degree bounds cited below identify their original source architectures. Current complete bounds and all parameter qualifications are in the [main table](../README.md) and [v82 records](../data/results_20261120_v82.json).

## Rank-one curves over general number fields

The [Pasten–Vidaux audit and amplified descent](../papers/number_fields/amplified_rationality_descent_20261119_v81.tex) provide a new conditional rational-subfield definition under abc over K and positive elliptic rank stable from Q to K. It handles all rational targets, both signs of odd valuation, and a fixed six-point divisor of a cubic rational map. This architecture uses the global conjugate difference t−σ(t), so it avoids the adjacent-contact obstruction below. It does not distinguish integers among rational inputs. Zywina supplies the needed stable curve for quadratic extensions; no such blanket conclusion is claimed for every higher-degree field.

The [effective sum-height theorem](../papers/effective_sums/elliptic_sum_heights_over_number_fields_20261117_v63.tex) now works for every elliptic curve over every number field. Together with Zywina's rank-one existence theorem it supplies whole-domain height control on suitable curves, without retaining the original complex-multiplication model. The remaining tasks are to construct the required local integrality detector over the chosen field, prove simultaneous rational realization on that field, and encode the complete witness set. The height theorem does not identify field traces with algebraic sums and does not establish a Diophantine definition of the ring of integers.

The [rank-stable contact obstruction](../negative_results/papers/number_fields/rank_stable_adjacent_contacts_20261117_v63.tex) excludes a direct use of a curve defined over the base field whose positive rank remains stable in the detector extension. Its adjacent contacts can occur at only finitely many primes. A new local architecture, or a curve defined over the extension without the stated descent property, is required.

The [power-shift criterion](../papers/number_fields/power_shifts_and_finite_exceptions_20261117_v69.tex) needs a detector containing every sufficiently large positive rational integer and having bounded-height nonintegral inputs. A fixed square shift removes the finite errors without another witness coordinate. An integral basis and a field norm then give the whole ring in `dN` total coordinates and degree at most `2dD`, for field degree `d` and an original polynomial of degree `D` in `N` coordinates. The shift can satisfy prescribed congruences even over complex fields. Constructing such a detector remains the arithmetic task.

The same height theorem also controls reciprocal abscissas on $E:y^2=x^3-Dx$, $D\ne0$. If $Q_i=[n_i]P+T_i$ are finite points in one cyclic group modulo torsion with $x(Q_i)\ne0$, and $M=\max(1,|n_i|)$, then effectively

$$
h\!\left(\sum_i\frac1{x(Q_i)}\right)
\ge \max_i h(x(Q_i))-C\log(M+2)-C'.
$$

Indeed, $T_0=(0,0)$ is rational two-torsion and $x(Q+T_0)=-D/x(Q)$. The translated points remain in the same cyclic group modulo torsion and are finite under the stated nonzero-abscissa condition. Applying the sum-height theorem to them, then using $h(1/x)=h(x)$ and $|h(Dx)-h(x)|\le h(D)$ twice, gives the inequality. The constants depend only on the fixed curve, field, generator and number of summands. This is a direct corollary useful for reciprocal charts; it neither controls field-trace cancellation nor supplies detector completeness.

## Earlier quartic chart foundations

The earlier **194-variable general** and **68-variable ordinary-abc** constructions provide chart foundations for the current records. The [translated-secant chart](../papers/quartic_existence/quartic_abc68_20261117_v65.tex) includes both reciprocal abscissas in six field coordinates, with four field quadrics. Its difference abscissa is a quadratic output; only its rational half-trace is stored. This preserves the prescribed difference carrier and the full multiplication-witness family, including tangencies, zero ordinates and every finite two-torsion difference. The imaginary output component has no other consumer and is eliminated exactly.

The [194-coordinate general quartic](../papers/quartic/quartic194_20261117_v73.tex) combines three translated point charts with four polynomial cross comparisons, a cleared input numerator and a homogeneous Lucas circuit. All 168 quadratic rows and both complete coefficient expansions are independently reconstructed. A refinement of the reference progression synchronizes the two formal depths for integer witnesses. Arbitrary odd-depth extraction, the unit boundaries and all allowed zero-ordinate cases have separate proofs.

The remaining quartic targets are the scalar primary ratio, the product helpers, and the transfer of this construction to the three-detector explicit source. The two excluded diagonal scaling families concern different specified factor interfaces; they impose no lower bound on these new polynomial coordinate changes. Every proposed deletion must include a complete quadratic compilation, every required guard, and a soundness and integer-completeness proof.

For the ordinary-abc quartic, the [direct-elimination audit](audits/quartic68_direct_elimination_review_20261117_v65.md) explains why immediate substitutions for the two field product helpers and rational half-trace output raise the degree. Further changes of coordinates or combinations of equations remain possible. The zero-input witness and the unit proof at nonzero input must both be retained.

## From eleven toward nine

The [complete derivative detector](../papers/conditional_few_variables/derivative_detector11_20261117_v63.tex) combines both heights in one signed contact argument and proves ordinary-abc completeness with one intrinsic pole block. The positive odd valuations still extract the original elliptic tests; negative odd derivative valuations are automatically harmless. The [full auxiliary-point construction](../papers/conditional_few_variables/full_auxiliary11_20261117_v71.tex) retains the root in `2E(K)` and permits the auxiliary points in all guarded affine `E(K)`. Its complete count is eleven, with degree at most 232,094,337, below $2^{28}$ and every guard included. Its cubic auxiliary charts also enlarge the sound nine-coordinate actual-trace source without changing that source's count. Rational-point arithmetic should therefore allow half-integral auxiliary indices, while retaining the integral root index and all exact contact-lifting tests.

The [single actual-trace criterion](../papers/pole_parameters/single_trace_derivative_criterion_20261117_v64.tex) now verifies the complete transfer from eleven to a sound nine-coordinate source. Its sole extra equation is $Y^2=V^2+w(2-V)$, at the actual derivative trace and the existing slope. All denominator, character and terminal-root conditions are proved. The unresolved task is rational square realization for every sufficiently large positive integer at one fixed sufficiently large defining shift. The source sign involution and finitely many input factors then complete all integers without adding variables. Local squares at prepared places and selected contacts are insufficient; new even-valuation places and global realization remain uncontrolled. Another possible source uses fewer point roles, but must replace the original index-recovery argument.

The [Möbius coupling obstruction](../negative_results/papers/pole_parameters/six_contact_mobius_obstruction_20261117_v63.tex) excludes the full twelve-symmetry menu for every fixed integer exponent at least four and every fixed finite exponent menu in the displayed six-contact architecture. Varying actual traces on a suitable incidence cover remains a different target.

The [cofinite coupling obstruction](../negative_results/papers/pole_parameters/cofinite_mobius_coupling_obstruction_20261117_v64.tex) survives finite input repairs: local square classes and inert contacts persist on a positive integer progression. This closes that particular finite-exception escape at every obstructed coefficient shift. It does not prove an obstruction at every fixed shift.

For the general target twenty-one, the derivative height proof cannot be imported without its abc hypothesis. A merger of the two general contact blocks needs an unconditional height argument under the already retained premises, together with prime realization and collision control. The original simultaneous actual-trace route remains open.

For that original route it is enough to realize the two actual guarded fibers simultaneously for all but finitely many integer inputs at one fixed defining parameter. Their four free pole coordinates would then be removed. Since the retained source is sound, multiplying its polynomial by the finitely many missing integer-input factors restores completeness without adding a coordinate. A rational point on a curve with freely assigned trace, or local preparation at only the prescribed primes, does not establish this simultaneous realization premise.

The [full auxiliary-point transfer](../papers/few_variables25/full_auxiliary25_20261117_v72.tex) enlarges this target by allowing the five auxiliary points in guarded affine `E(K)`, while the four root points retain integral indices. It preserves the scalar lower bound and both comparison precisions. The complete 25-coordinate source has degree at most 4,924,857,429,000,192; the corresponding 21-coordinate source still requires simultaneous actual-trace realization. Searches may now use half-integral auxiliary indices without altering the root-index height argument.

The [actual-weight arithmetic reduction](../papers/pole_parameters/actual_trace_arithmetic_20261117_v70.tex) gives a fixed-tuple curve of genus at most 36 before the degree-nine contact substitution. The explicit genus-24 fixture is a smaller target for rational-point methods, with every candidate subject to the exact degree-nine equation, rational-square lift and original guards. It supplies no prepared global witness.

The [actual-trace curve theorem](../negative_results/papers/pole_parameters/curves_and_abelian_families_20261117_v69.tex) gives finiteness of rational points on every fixed curve in each specified fixed-input locus, with all point roles independent. Its first fixed-point cover has genus at least eight for the derivative route and four for the general route. The [vertical-valuation theorem](../negative_results/papers/pole_parameters/vertical_and_fixed_slope_20261117_v70.tex) excludes every nonconstant rational abelian family through a prepared point, fixing the input and, for the derivative source, the root. The general statement allows all point roles to vary. In the derivative source, a fixed rational affine relation between slope and scale forces the root and slope to be constant even without a selected-prime hypothesis. The useful arithmetic target is therefore an individual guarded rational point, or points varying among different curves or on a higher-dimensional nonabelian variety. Everywhere local solubility of the derivative fiber and its boundary points do not establish global realization.

The [fixed-slope incidence theorem](../negative_results/papers/rational_fibers/fixed_slope_trace_finiteness_20261117_v59.tex) proves finiteness on fixed genus-one trace curves for all independent point roles in the stated reweighted sources. Infinite propagation at a fixed slope therefore cannot complete that route. A variable slope, point-dependent auxiliary curve, untwisted conic, or individual rational witnesses remain possible. The theorem does not automatically transfer to the derivative source.

### Complete index divisibility and the remaining equality step

The [elliptic index source](../papers/elliptic_divisibility/elliptic_index_poles_20261117_v60.tex) proves $h\mid j$ and has infinitely many fully prepared witnesses for each nonzero quotient. Under a fixed weak-abc exponent below two, a fixed index progression supplies the required simple primitive denominator primes. Truncating the formal logarithm amplifies the congruence between $X$ and $j/h$ to any fixed order with no new free coordinate. A sufficient missing condition is

```math
h(X)+\log(1+|j/h|)\le C h(\sqrt{x(H)})+O(1),
\qquad C<T(2/\kappa-1).
```

Its complete existential realization is unproved. A real inequality alone does not bound rational height.

The coordinate $X^2x(U)-x(H)$ cannot supply that condition through the specified second derivative detector. The [denominator-support obstruction](../negative_results/papers/pole_parameters/denominator_support_height_obstruction_20261117_v60.tex) bounds the quotient and proves finiteness of intended integer inputs for every fixed rational carrier generically nonconstant in $U$ under the same hypotheses, including its degenerate specializations. Those hypotheses require every positive odd contact prime to occur in the reference denominator. A successful replacement must change that support condition, obtain input equality by a different mechanism, or use additional free data without losing the intended count.

The separate [square-rescaling obstruction](../negative_results/papers/pole_parameters/square_rescaled_weight_perturbation_20261117_v60.tex) excludes identifying the external input solely by square-rescaling an input-independent weight class with the stated strict preparation. Individual rational witnesses for the actual trace covers remain an open route.

The [explicit input-only helper](../negative_results/papers/pole_parameters/input_only_weight_false_witnesses_20261117_v61.tex) is also incomplete: it bounds $h(X)$ by the reference height under abc, yet permits $X=2+1/q$ with complete rational witnesses. Its local index map is an isometric bijection, so the second index can satisfy all finite congruences while growing arbitrarily large. Any new growth condition must control this quotient-size term, not only the input height.

## Infinitely many rationally soluble levels

For each prepared input $d$, the [varying-level theorem](../papers/pole_parameters/varying_level_pole_fibers_20261117_v61.tex) studies

```math
k(r)=(1-d)\frac{d+r^2}{d-r^2},\qquad
y^2=d\bigl(k(r)^2-s^4\bigr).
```

Every rational $r$ gives an everywhere locally soluble fiber. Under the retained simultaneous-twist theorem, infinitely many rational levels $0<r<\sqrt d$ have infinitely many rational points. The proof selects two fixed elliptic curves and varies their points independently; it identifies the required abscissa square class, rather than inferring it from positive rank. The combined witness set is Zariski dense and permits all-character guard avoidance, including characters at which $d=1$.

The cleared radicand has degree nine, twelve monomials and coefficient height four. For the nonsquare-complement input $d=112/225$, an exact example is

```math
r=\frac{44}{69},\quad k=\frac{3364}{675},\quad
s=\frac{29}{15},\quad y=\frac{23548}{10125}.
```

The theorem selects good levels; it does not realize a level prescribed by an existing detector coordinate. Both $r$ and $s$ remain free, and it does not synchronize two actual detector inputs. The construction counts are unchanged.

## Current target: simultaneous actual-trace conics

The [trace-conic theorem](../papers/pole_parameters/trace_conic_criterion_20261117_v57.tex) reduces the **21/9** target to rational squares of $V_j^2+w_j(2-V_j)$ at actual source traces and one common slope. It removes the separate detector norm from the square condition. The conic has the exact chart

$$
(1+2k)V+k^2w=2.
$$

Here $V$ and $w$ must remain the actual functions of the source points; assigning them freely does not supply witnesses. The finite quadratic root introduces no retained coordinate, and the full source count and all-character guards are proved. Fixed-place and selected-contact local squareness do not control new prime divisors or global rational points.

Each block can use any of the classes $1$, $1+24V$, and $1+3(V-1)^2$. They preserve parity at every hypothetical bad detector pole, even when their prime supports vary, and therefore leave the general Lucas and conditional height arguments unchanged. The linear class gives the elliptic family

$$
Y_c^2=X_c^3-(2+24w)X_c^2+(1+1176w)X_c,\qquad X_c=1+24V,
$$

with reference point $(49,336)$. The arithmetic objective is a rational point **at that prescribed abscissa**, simultaneously for the two actual blocks. The [cubic rank theorem](../papers/pole_parameters/cubic_trace_rank_20261117_v57.tex) now gives infinite order for this section at every rational $0<w\le1/4$, and proves finite-place approximation with infinitely many free traces in $(1,2)$. The unresolved step is to couple these points to the actual source trace map; positive rank alone does not prove that incidence.

The adaptive class $\delta_c=(q^2-2q-1)^2+4q(q^2-1)V$, $q=2+wc^2$, is also admissible for any fixed guarded rational scalar function $c$ of the existing source coordinates. Its parity proof permits arbitrarily large auxiliary denominators. In particular $c=\lambda$ preserves fixed-place preparation and lets the class vary with the actual common slope. This structured family avoids the height cost of arbitrary square classes. The next arithmetic step is to match such a class with the actual reciprocal trace; choosing a fresh free $c$ would require its own coordinate.

The [small-class theorem](../papers/pole_parameters/small_trace_classes_20261117_v58.tex) adds the classes $V+2$, $V^2-V+2$ and $1+w(2-V)$ without a contact loss or height budget. The first gives the four-term cubic $V^3+(2-w)V^2+4w$. The [actual multiplication-surface obstruction](../negative_results/papers/rational_fibers/actual_multiplication_surface_20261117_v58.tex) rules out a uniform rational square formula on the full fixed-index surface even with a nonconstant rationally coupled slope. Individual rational witnesses and genuinely ramified incidence families remain the arithmetic targets.

## A single-cover route and finite square-class choices

The [sharp signed height theorem](../papers/elliptic_heights/signed_mixed_power_height_20261117_v58.tex) reduces the ordinary-abc single-cover argument degrees to **48** for fixed classes and **60** for the support-budget-two menu. Positivity of the actual trace permits an odd inner exponent. Its exact optimization and every guard retain a nine-coordinate criterion, conditional on rational realization. A [two-square counterfamily](../negative_results/papers/contact_heights/two_square_height_20261117_v58.tex) prevents substituting two independent squares into the same unrestricted height estimate.


The [height-controlled class theorem](../papers/pole_parameters/variable_squareclasses_20261117_v57.tex) broadens the conditional target to the menu $\{1,(X^2+B)^2+\operatorname{Tr}(x(A))^2,1+X^4\}$, with fixed positive exceptional-unit multipliers. The class support may vary with the input and root point. The contact primes lost to it have a proved height budget, so suitable larger powers retain soundness and the nine-coordinate census. The missing step remains simultaneous **rational realization on actual detector tuples**, with the prescribed root index and all source guards. Local filters do not prove it.

The [prepared-cover obstruction](../negative_results/papers/pole_parameters/prepared_sha_obstruction_20261117_v57.tex) rules out the proposed specialization $q_0=0$ uniformly: $d=112/225$ yields an everywhere locally soluble genus-one curve without a rational point. The [finite-menu obstruction](../negative_results/papers/pole_parameters/intrinsic_specializations_20261117_v57.tex) also excludes prescribing $s$ from finitely many rational functions of $d$. A viable coordinate relation must use a more restricted actual-input locus or couple the two pole coordinates in a way that evades these precise obstructions. Positive rank alone does not supply the required abscissa class.

The [single-cover theorem](../papers/pole_parameters/single_cover_criterion_20261117_v56.tex) proves conditional soundness with $t=((X^2+B)^e+T^e)^d+z^2$ and one actual native square. Its full nine-coordinate census includes $z$, six point coordinates, the input and the finite-algebra encoder. The required rational realization remains open. The binary quadratic contact forms must also meet the actual prime-preparation constraints; unrestricted-numerator prime patterns do not suffice.

The native square condition may be relaxed to a **fixed finite menu of positive exceptional-unit square classes** without adding a coordinate. A finite product of the existing quadratic-branch encoders supplies the union. The menu may vary between blocks and its selected class may vary with the input, but its support is fixed in advance. Do not retain optional exceptional-place square neighborhoods that would force every allowed class back to one. No global realization is known for this relaxed target either.

The [intrinsic pole block](../papers/pole_parameters/intrinsic_pole_20261117_v56.tex) has degree nine and coefficient height four, with dense prepared witnesses from the stated Burungale theorem. It retains two free coordinates and therefore offers a coefficient and arithmetic-foundation option, not the missing 21/9 reduction. The complete [general](../papers/pole_parameters/general_intrinsic_transfer_20261117_v57.tex) and [conditional](../papers/pole_parameters/conditional_intrinsic_transfer_20261117_v57.tex) transfers supply those preparation and all-character audits, with unchanged counts.

The [precision revision](../papers/pole_parameters/precision_transfer_20261117_v56.tex) lowers scalar test orders to two while retaining reciprocal orders one. The eighth-order formal logarithm supports sixth-order comparisons without demanding incompatible deeper formal neighborhoods. It provides a complete 25-variable source, leaving the actual source-point coupling or pole-coordinate elimination as the count problem.

The [smallest trace cubic](../papers/pole_parameters/small_trace_classes_20261117_v58.tex) now has a proved infinite-order section $(4,4)$ for every rational normalized weight. Its equation has four radicand monomials and coefficient height four. The required construction must couple its abscissa to the actual detector trace; a free point on this positive-rank curve does not meet that obligation.

## The 25-coordinate budget

The [finite-algebra source](../papers/finite_algebra/unit_norm_20261114_v28.tex) has the following budget.

| Role | Rational coordinates |
|:--|--:|
| Input | 1 |
| Nine elliptic points over the quadratic field | 18 |
| Comparison slope | 1 |
| Two parameters for each of two pole predicates | 4 |
| One coordinate encoding all finite algebraic choices | 1 |
| **Total** | **25** |

The finite-algebra encoder already uses one extra coordinate. Smaller arithmetic circuits can lower the degree without changing this budget. A smaller total requires a complete source with fewer retained parameters, or a different encoding theorem that actually removes a retained coordinate.

The [point-role restriction theorem](../negative_results/papers/integer_labels/point_role_restrictions_20261117_v66.tex) excludes every fixed finite menu of nonzero quadratic relations among the four root indices in the full retained source, with infinitely many omitted positive inputs in every input progression. It consequently excludes two independent fixed group-word deletions within the exact multiplication family, even if the reference index is chosen together with the roots. This conclusion does not impose exact multiplication identities on arbitrary source witnesses.

One shifted relation survives those obstructions: `U_S=Z+H`. For the Lucas sequence `U_(j+1)=a U_j-U_(j-1)`, take `j=6(a-2)r`, `b=U_j(a)/(a-2)`, `t=U_(j+1)(a)` and `h=U_j(a)/((a-1)(a-2))`. The paper proves integrality, the Pell relation, `b=(a-1)h` and `h=-4r mod(a-1)`. These identities solve compatible finite congruences prescribed in advance. They do not solve the joint comparison-prime and real-return problem: changing `r` changes the roots, their trace scale and the already coupled reference index. A new simultaneous realization theorem would be needed before this relation yields any complete coordinate reduction.

The [terminal quadratic-root theorem](../papers/finite_algebra/terminal_quadratic_root_20261115_v40.tex) permits the last pole ordinate to vanish. It proves the degree bound

$$
\deg P_{25}\le 5\,059\,009\,872\,658\,432<2^{53}
$$

with the same variable count and inherited arithmetic premises. The two 13-variable constructions have degree below $2^{26}$. These are certified upper bounds from exact finite-operation specifications, not expanded polynomial degrees. Only the last quadratic step may lose its unit condition in this theorem: all preceding radicands remain nonzero. The character-descent proof handles nilpotents at that last step, so a simple-eigenvalue argument is unnecessary.

### The trace-and-norm square criterion

The [quadratic reweighting theorem](../papers/pole_parameters/reweighted_elliptic_detectors_20261116_v53.tex) supplies complete general and conditional sources whose odd physical weight valuations are positive. The conditional detector is the quadratic product $W=\prod_i(1+C_i^2/F)$, with $W-1=O(w^3)$ at fixed preparation places. The general detector uses $w=u/(1+u)^2$ throughout its factors and one extra contact factor in each scalar numerator. Their full source counts remain 25 and 13.

The [guarded square theorem](../papers/pole_parameters/native_square_criterion_20261116_v53.tex) reduces the desired **21/9** construction to simultaneous actual rational-square values of

```math
\mathcal R_a=\left(1-\frac{V}{1+w}+\frac{J}{(1+w)^2}\right)
             \left(1+\frac{a^2(1-V/2)}w\right),
\quad V=\operatorname{Tr}(W^{-1}),\quad J=\operatorname{Norm}(W^{-1}).
```

All point lifts, source denominators, whole-character guards, common slope and height scales must be retained. Fixed exceptional-set units are admissible choices of $a$. The source-dependent family $a=1+wc^2$, with $c$ an existing guarded rational source function, is also admissible and has no additional coordinate cost. Its local soundness comes from $v_p(a)\le0$ at every hypothetical negative odd detector pole. An unrestricted new rational parameter would have a coordinate cost; an unproved finite root for that parameter cannot be substituted.

For $a=1$, put $W^{-1}=u+v\sqrt2$, $x=1+(1-u)/w$ and $q=v/w$. The square condition is exactly the prescribed-coordinate equation $y^2=x^3-2q^2x$. This makes a coupling to the source's elliptic points a concrete target. The actual $x$ and $q$ cannot be replaced by freely chosen coordinates.

The local theorem gives rational-square values at every fixed preparation place and at the stated fresh split-contact patterns. New prime divisors of detector numerators and the global rational lifts remain uncontrolled. The [fixed-fiber genus theorem](../negative_results/papers/pole_parameters/fixed_flat_fibers_20261116_v53.tex) gives genera 236 and 332 for the explicit fixed-index fixtures $X=1,2$, $B=3$. Their rational boundary points do not supply finite witnesses. The next global argument must therefore allow auxiliary indices or admissible source functions to vary, or produce individually chosen rational points on the resulting ramified incidence. This fixed-fiber result does not exclude either possibility.

The realization statement is **unproved**. The [current proof guide](results_20261117_v58.md) records the full implication and exact checks.

### A shared conic parameter: the precise route to 24

For a pole input written $b_i=d_i t_i^2$ with $d_i t_i\ne0$, the current rational chart is

$$
e_i=\frac{\tau_i^2-17d_i}{\tau_i^2+51d_i},
$$

$$
q_i^2=
17d_i\bigl(s_i^2-17e_i(17-b_i)\bigr)
       \bigl(s_i^2+51e_i(17-b_i)\bigr).
$$

The square factor $t_i^2$ is extracted from the fixed source circuit. The [rescaling theorem](../papers/finite_algebra/unit_norm_20261114_v28.tex), Section 4, gives the exact invertible change from the original pole equations and states its denominator guards.

Replacing $\tau_0,\tau_1$ by one freely varying rational $\tau$ would leave the three retained pole parameters $\tau,s_0,s_1$. If a complete source with this property were proved, the displayed budget would become **24**. The finite ordinates would still use the existing separating coordinate.

The missing result is a **simultaneous rational realization theorem**. For every integer input, one must choose the actual elliptic, prime-pattern and comparison data so that both displayed square equations have rational solutions with nonzero ordinates at a common $\tau$. The earlier point-stage guards are retained; the fixed-fiber theorem below supplies the remaining pole guards. A theorem for every prepared pair of pole inputs would suffice, but completeness only needs the pairs actually supplied by a proved source construction. Individual solvability or separate Zariski-density statements for the two pole surfaces do not establish that their permitted $\tau$-values intersect.

The [local-solubility theorem](../papers/pole_parameters/shared_parameter_local_solubility_20261115_v41.tex) proves that **every rational $\tau$** works over every completion for every prepared input, and a finite family can use that same parameter. The [fixed-fiber and scalar-input theorem](../papers/pole_parameters/common_parameter_guard_reduction_20261116_v42.tex) strengthens the guard statement for actual native sources: the extracted $d_i$ are positive rational scalars independent of all auxiliary points and characters. Every rational $\tau$ avoids all conic denominator zeros. One nonzero rational pole witness in each fixed fiber gives infinitely many witnesses there and permits all remaining pole norms to be avoided. The problem is therefore simultaneous rational realization. In exact elliptic terms, for $D_i=17d_i$ and $C_i=17(17-b_i)$, one needs nonzero-ordinate points in the abscissa classes $D_i\mathbb Q^{\times2}$ on the correlated families

$$E_{N_i(\tau)}:\ y^2=x(x-N_i(\tau))(x+3N_i(\tau)),\qquad N_i(\tau)=D_iC_i(\tau^2-D_i)(\tau^2+3D_i).$$

Congruence sieves cannot exclude the shared parameter on this prepared locus. A global argument must establish the simultaneous rational square-class condition with nonzero ordinates. No separate pole-guard avoidance theorem remains necessary.

The same paper proves independent auxiliary-index freedom: after fixing one prepared guarded tuple, the five auxiliary points in the general source, or two in a conditional source, may vary independently in infinite cyclic return sets while preserving preparation and all earlier norms. The extracted $d_i$ remain fixed. This is a family of actual detector inputs, but it does not imply arbitrary rational values for the detector multipliers or preservation of rational pole solubility. The [proof guide](common_parameter_guard_reduction_20261116_v42.md) states the exact remaining assertion.

The [fixed-input geometry theorem](../negative_results/papers/pole_parameters/fixed_input_k3_surfaces_20261116_v42.tex) identifies each full one-pole surface, with $\tau$ still variable, as a K3 surface after resolving four $A_3$ singularities. The common-parameter product for fixed detector inputs dominates each K3 factor, so it has no rationally connected smooth projective model. This limits a geometric proof method for fixed inputs; it does not exclude rational witnesses or the independently varying actual detector families.

The [rational-section classification](../negative_results/papers/pole_parameters/rational_sections_20261116_v42.tex) makes one sufficient arithmetic locus explicit. For prepared fixed inputs, sections over $\mathbb Q(\tau)$ exist exactly when every $C_i=17(17-b_i)$ is a rational square. If $C_i=c_i^2$, then $s_i=c_i$ and $q_i=-4C_iD_i\tau/(\tau^2+3D_i)$ work simultaneously for every nonzero rational $\tau$. Realizing these square complements on actual guarded detector tuples is a stronger possible target. The theorem leaves open isolated rational common fibers and multisections when the complements are nonsquare.

The [early-scalar theorem](../papers/pole_parameters/square_complement_baseline_20261116_v43.tex) makes this square-complement route locally compatible with the actual general source. With the native $r$, $\ell=76249$ and fixed odd $L$, replace the early scalar by

$$\rho_* = \frac{102^2\ell^Lr}{(1+153\ell^Lr)^2}.$$

Its complement is a nonzero rational square, and the actual detector complements can be squares at any finite set of primes prescribed before preparation. The revised denominator case has $v_\ell(\rho_*)=|L-20m|$ for $m=-v_\ell(X)>0$; the paper supplies the necessary corrected height estimate. The source census is unchanged, with a sufficiently large ineffective fixed odd exponent. No new degree or coefficient record is assigned to the modified family. The remaining condition is a global rational square on actual detector tuples, not arbitrarily accurate local approximation.

Two separate obstructions delimit that condition. The [actual-neighborhood theorem](../negative_results/papers/pole_parameters/actual_square_complement_neighborhoods_20261116_v43.tex) produces original-source independent return families with nonsquare complements throughout a fixed 23-adic neighborhood. The [normalization counterexample](../negative_results/papers/pole_parameters/square_complement_normalization_20261116_v43.tex) shows why applying a square-complement normalization directly to the final detector inputs destroys the pole test. The early scalar is changed before the detector is formed and has its own soundness proof.

The [fixed-class synchronization theorem](../negative_results/papers/pole_parameters/fixed_class_synchronization_20261116_v43.tex) gives genus $1+2^r(r-1)$ for the common propagation curve when $r$ distinct positive pole scales and their twist classes are fixed. For two poles the genus is five, with an explicit $1+1+1+2$ Jacobian decomposition. Every fixed class tuple therefore gives only finitely many rational parameters. Infinitely many varying class tuples remain available. The full common-parameter variety has Kodaira dimension one when $r\ge2$, so direct Kummer-quotient theorems do not apply to it. The [literature audit](../negative_results/docs/fixed_fiber_literature_20261116_v43.md) also checks the relevant two-torsion, semistability and cohomological hypotheses; no theorem from Schinzel and finite Sha alone is claimed.

The [abelian-parameter obstruction](../negative_results/papers/rational_fibers/abelian_parameter_obstruction_20261116_v44.tex) excludes a uniform rational identity on an abelian parameter space when the detector argument varies nontrivially, even if every twist class may vary. It applies both to rational square complements and to the original normalized pole equations. The argument forces an odd contact divisor and a negative odd pole valuation. It does not cover an actual detector family at a fixed comparison slope, a ramified cover, or individually chosen rational points. Local return sets still supply the required preparation and guards, but cannot establish the rational incidence themselves.

At a fixed allowed slope with positive detector weight, the [fixed-argument ramification theorem](../negative_results/papers/rational_fibers/fixed_argument_complement_ramification_20261116_v44.tex) proves that the complement cover of the full independent auxiliary product is geometrically ramified. Therefore a surjective abelian coupling, including an isogeny or independent half-point doubling, cannot furnish a uniform square-complement section. The next targets are a proper coupled subvariety with a separate rational realization proof, or rational points directly on the ramified incidence cover. A common rational pole fiber without square complements remains a separate possible route. None of these obstructions proves nonexistence of the required individual actual detector tuples.

The [varying-input quadric theorem](../papers/pole_parameters/varying_input_quadric_20261116_v44.tex) gives a rational parameter space for compatible coefficient inputs at any fixed nonexceptional common conic parameter. For a pole scale \(D>0\) and \(e=(\tau^2-D)/(\tau^2+3D)\), the surface with input \(b=D/(17V^2)\) is birational to

$$x^2-Dy^2=e(\sigma^2+3D),\qquad V=\frac{\sigma x-Dy}{17(x-\sigma y)}.$$

The quadric has the explicit rational point \((\tau,1,\tau)\). The arithmetic target is to match this V with the native reciprocal trace while preserving the earlier source guards. This gives an exact incidence problem on which to try proper couplings or descent. The parametrization neither changes a final detector input by normalization nor proves that arbitrary actual tuples lift. Its explicit small-positive-input curve has square complement; its full two-parameter surface is the relevant broader family.

The [native doubled-source norm map](../papers/pole_parameters/doubled_source_congruent_twists_20261116_v43.tex) supplies a different arithmetic handle: it constructs a nonzero-ordinate rational point on a congruent-number twist with abscissa square class $[\operatorname{Tr}(x\bar d)/2]$. Its class has not been identified with a required detector class. A useful next step is either an exact class identity or a descent argument trivializing the prescribed isogeny-Selmer elements, not only proving positive rank on an unrelated twist.

This proposal shares the conic coordinate, not $e_i$ itself. Sharing $e$ requires matching input square classes; the separate [fresh-prime obstruction](../negative_results/papers/pole_parameters/fresh_prime_squareclass_obstruction_20261116_v42.tex) proves that the existing fresh-prime witnesses violate that requirement. The [common-twist theorem](../papers/pole_parameters/common_twists_20261114_v30.tex) gives genuine savings on its explicitly compatible locus, but does not put the full integer construction on that locus. A common $\tau$ allows different $e_i$ and avoids that particular necessary square-class equality. It still needs its own arithmetic existence proof.

A fixed value of $\tau$ does not supply this missing result. The [fixed-parameter obstruction](../negative_results/papers/pole_parameters/fixed_pole_parameters_20261109_v17.tex) gives prepared inputs excluding $\tau=\pm1$ and $\tau=\pm17$, including arbitrarily small prepared inputs in the latter case. The proposed shared parameter must remain variable unless a new restricted completeness theorem proves otherwise.

Those fixed-parameter counterexamples are global: their pole quartics are everywhere locally soluble by the preceding theorem. Their rank and square-class certificates do not constitute local obstructions.

### The stronger rational-complement target

The [native complement theorem](../papers/pole_parameters/guarded_complement_reduction_20261116_v45.tex) sharpens the payoff of the square-complement route. At an actual point-stage tuple every character lies in \(K=\mathbb Q(\sqrt2)\). Prepared inputs have the form \(b_i=w_iS_i^2\), with even \(v_{17}(w_i)\). Vanishing of a complement at any character would make \(w_i/17\) a square in \(K\), contradicting its odd valuation at 17. Thus the complement norm guards are automatic.

If actual simultaneous rational complements exist for every integer input, they themselves exclude the negative odd valuations needed by soundness. Both retained pole pairs can then be removed, giving the conditional counts **21 and 9**, with the existing one-coordinate encoder. This implication is proved; the native realization assertion is not. The **24-variable** common-fiber route remains a weaker sufficient target and does not require square complements. A global coupling with varying twist classes must prove one of these arithmetic incidence statements on the unmodified detector values.

The same theorem constructs infinitely many prepared free inputs at a fixed common pole parameter, including square reduced denominators and all preparation depths. The missing step remains equality with the actual reciprocal traces. Neither a rational parametrization of the free-input quadric nor finite local preparation proves that equality.

The [finite-class complement theorem](../papers/pole_parameters/finite_complement_classes_20261116_v46.tex) gives a weaker sufficient realization condition. Fix finite sets of levels 17m² and positive square classes supported on the source's original exceptional primes. For each actual input it suffices to make one allowed multiple of the corresponding complement a rational square. The complement norms are still units at every point-algebra character. Separate finite-algebra encodings followed by their product give the same 21/9 conditional census without a retained class-selection coordinate, even if branch radicands coincide.

At the single level 17, allowing all exceptional square classes means that both actual complements need only have even valuations outside the original exceptional set. No consumed prime may be moved into that set. Its actual uniform rational realization is still missing. The explicit nonsquare-class coefficient family in the paper does not establish native trace matching.

The [admissible-level theorem](../papers/pole_parameters/admissible_levels_native_traces_20261116_v47.tex) also permits finitely many fixed rational levels with nonnegative even valuations outside the exceptional set, provided every whole-algebra complement guard is proved. An exact native fixture at X=±1 and B=3 satisfies both prescribed-trace equations after fixed positive exceptional-unit rescalings, with all 128 guards checked. Its levels and rescalings are fixed constants; no completeness theorem for other inputs or sufficiently large defining B follows. The [polynomial-level local theorem](../papers/pole_parameters/polynomial_levels_local_20261116_v47.tex) supplies compensating test powers for levels cM(X)². Global realization and the corresponding height/completeness transfer are separate obligations.

The [norm-compatible scalar theorem](../papers/pole_parameters/norm_compatible_scalar_20261116_v47.tex) removes the bare norm-conic obstruction in a complete prepared family. Its degree-eight input map reflects integrality and makes all four finite height-contact factors norms from Q(√10657). The new general scalar preserves the forced odd valuation at 76249, local preparation, height contacts and 25-coordinate census. The conditional sources already have uniformly soluble bare norm conics under compatible prime preparation. The concrete remaining target in either setting is

$$ (m_j\operatorname{Tr}(\mathcal W_j^{-1}))^2-10657Y_j^2=4w_j/17,$$

with m_j from fixed finite admissible levels, the actual coupled detector data, and all guards. The free first coordinate of a conic point must not be substituted for the prescribed trace.

The [detector-field norm scalar](../papers/pole_parameters/detector_field_norm_scalar_20261116_v52.tex) aligns this norm field with $K=\mathbb Q(\sqrt2)$ itself. Its integer elliptic label has degree four, and its scalar contact polynomial has degree sixteen. The scalar contact-height theorem is applied to the fixed rational rescaling $q=\beta/n$, while comparison tests use the integral label $\beta$. The full soundness, ordered completeness and whole-character guard transfer retain 25 variables under the existing arithmetic premises. Both prepared physical weights, and those of the unchanged conditional sources, are norms from $K$ without extra prime-pattern congruences. This supplies bare conics with class $d=2$; their actual first coordinates still need to be matched. No numerical determinant degree bound is transferred from a different scalar family.

For the unchanged general scalar, the [signed finite-class theorem](../negative_results/papers/pole_parameters/finite_nonsquare_class_obstruction_20261116_v47.tex) excludes uniform realization using any finite collection consisting entirely of nontrivial square classes. Its three quartic contact fields and Chebotarev argument apply to every guarded rational auxiliary tuple, not only the prime-prepared family. Including the rational-square class or using the new norm-compatible scalar avoids this specific obstruction.

### Detector-dependent bounded levels

The [bounded native norm-level theorem](../papers/pole_parameters/bounded_norm_levels_20261116_v48.tex) permits $m_i=M_iN_i/(c_i+N_i)$ and $a_i=17m_i^2$, with $N_i=\operatorname{Norm}(W_i)$. At every odd weight valuation, of either sign, the norm has nonpositive valuation and a zero-valuation norm reduces to one. The required local implication and whole-character denominator and complement guards therefore hold with no additional retained coordinate. For finitely many permitted coefficient choices, simultaneous actual realization of

```math
\left(\frac{2M_iW_{i,0}}{c_i+N_i}\right)^2-d_i y_i^2=\frac{4w_i}{17},\qquad i=0,1,
```

with positive exceptional square classes $d_i$, still gives the conditional **21/9** census. The first coordinates remain the actual detector functions. For $M_i=c_i=1$, the exact relation $u_i^2-2v_i^2+z_i^2=1$ gives an additional incidence equation, with $0<u_i\le1$ on prepared real characters. A useful next step must construct rational points on this coupled incidence with the prescribed coordinates, or a proper coupled subfamily for which rational realization is proved. Bare conic solubility and local congruence searches do not do this.

The separate [automatic-level finiteness theorem](../negative_results/papers/pole_parameters/automatic_level_finiteness_20261116_v48.tex) rules out uniform use of $a_i=w_iN_i$ for the conditional sources, even after finitely many fixed rational rescalings. The bounded native levels are outside that theorem. The actual $X=\pm1$ benchmark remains a finite example, not uniform completeness.

### Adaptive levels with varying prescribed traces

The [adaptive complement theorem](../papers/pole_parameters/adaptive_native_complements_20261116_v50.tex) provides a larger arithmetic search family while leaving the detector outputs exactly unchanged. Write

```math
w^*=w/(1+w)^2,\qquad W^*=(1+w)W,\qquad
V^*=\operatorname{Tr}((W^*)^{-1}),\qquad J^*=\operatorname{Norm}((W^*)^{-1}).
```

Every odd valuation of $w^*$ is positive. At a negative odd valuation of $b=4w^*/(V^*)^2$, all three quantities $w^*,V^*,J^*$ reduce to zero. Thus a level $17(P(V^*,J^*)/Q(V^*,J^*))^2$ retains local extraction when $P,Q$ have coefficients integral outside the original exceptional set and $Q(0,0)$ is a unit there. Whole-character nonvanishing must also be proved. It is automatic for $m=1/(1-V^*+J^*)$, whose denominator is $\operatorname{Norm}(1-(W^*)^{-1})>0$ on prepared characters. This multiplier can have poles at good contacts.

The normalized class $d_0=1-4w^*/17$ satisfies $16/17\le d_0<1$ and is a local unit at all relevant odd-weight primes. Its unrestricted conic has the fixed point $(1,1)$. More flexible classes $D(w^*,V^*,J^*)/E(w^*,V^*,J^*)$ are locally admissible when both polynomials have integral coefficients outside the original exceptional set and unit constant terms; their whole-character positivity and nonvanishing remain explicit obligations. Another permitted family is $(R+\kappa w^*/R)^2-4w^*/17$, with a specified whole-character unit $R$ and fixed positive exceptional-unit $\kappa>1/17$.

The exact target is still simultaneous actual rational realization of

```math
(m_iV_i^*)^2-d_i y_i^2=4w_i^*/17,\qquad i=0,1.
```

The [fixed-trace obstruction](../negative_results/papers/pole_parameters/fixed_trace_obstructions_20261116_v51.tex) excludes replacing both actual normalized traces by values from any finite fixed rational set for the original general scalar. In particular, the unit-trace specialization fails already at $X=4$, at the inert prime $252877$, independently of all auxiliary indices and the slope. The obstruction persists through the specified iterated native normalizations. For both conditional six-contact sources, a real scale bound excludes simultaneous unit trace at every prepared slope when $B\ge19$.

The arithmetic target must therefore retain varying first coordinates, or use a separately justified modified scalar construction. The norm-compatible scalar remains available. A solution must match the actual traces in the displayed equations with all guards; a bare conic point or a free norm-one parametrization does not do so.

No actual uniform realization is proved. The variable counts remain **25 and 13**, with **21/9** conditional on the displayed arithmetic assertion. The [proof and verification guide](results_20261116_v53.md) records the exact scope.

### Integral labels from one finite Pell choice

The [integral Lucas-label theorem](../papers/integer_labels/pell_labels_20261115_v40.tex) supplies a different arithmetic compression. Let $a,n$ be integers with

$$
|a|\ge3,\qquad c=|a-2|\ge3,\qquad n\ne0,\qquad t=(a-2)n+1.
$$

If a rational $w$ satisfies

$$
t^2-atw+w^2=1,
$$

then monicity forces $w$ to be an integer. With $U_0(A)=0$, $U_1(A)=1$ and $U_{j+1}(A)=AU_j(A)-U_{j-1}(A)$, the sharp lower bound is

$$
|t|\ge U_{c-1}(|a|)\ge(|a|-1)^{c-2}.
$$

Thus a third separately integral label can be replaced by a finite quadratic choice when the first two labels are already exact integers. The equivalent square condition is

$$
z^2=(a^2-4)\bigl((a-2)n+1\bigr)^2+4,
$$

of total degree six and coefficient height sixteen. The same paper gives sharp witnesses and arbitrarily large solutions preserving prescribed finite congruence data of a given witness.

The present source carries labels through local approximations by elliptic formal-parameter ratios. Those rational ratios are not known to be the exact integral labels, so monicity at the ratios does not prove the needed integrality. This distinction has a concrete counterexample: $a=5,n=1$ gives discriminant $340$, which is not a rational square, but for every $r\ge1$ there are rational norm-one solutions with $0<n_r<1$ and $n_r\equiv1\pmod{19^r}$. The separate [rational Pell approximation obstruction](../negative_results/papers/integer_labels/rational_pell_approximation_obstruction_20261116_v42.tex) proves this by an explicit rational parametrization. A useful application must therefore recover an exact integer carrier, or prove a new global implication from the complete source equations. Arbitrarily precise local approximation alone does not close the gap.

### Dependencies among point roles

A replacement for the nine points need not realize every independent tuple of labels. That freedom is where a genuine reduction can occur. It must nevertheless give witnesses for every integer input and prove soundness on every rational solution of the new source.

Two established obstructions delimit direct substitutions:

- The [dimension and label-graph theorems](../negative_results/papers/elliptic_charts/structural_constraints_20261109_v13.tex) show that a finite algebraic presentation recovering arbitrary tuples of nine points needs at least eighteen rational parameters. The integer-index graph itself is Zariski dense in the product of the additive group and the relevant abelian variety.
- The [index-product theorem](../negative_results/papers/elliptic_charts/index_products_and_finite_correspondences_20261110_v21.tex) shows that a fixed correspondence with finite geometric fibers cannot recover $[nh]P$ from $[n]P,[h]P$ uniformly for infinitely many $n$ and infinitely many permitted $h$ for each $n$. Restricting each index to congruence classes does not remove this obstruction.

The separate [affine scalar-coupling obstruction](../negative_results/papers/integer_labels/affine_scalar_coupling_obstruction_20261116_v50.tex) excludes every fixed finite set of identifications $S=[m]A+[k]P$ for the original general scalar source: scalar odd-contact divisibility contradicts those identities on an infinite integer progression. It does not cover a redesigned scalar or couplings involving other root roles.

These statements concern generic recovery and the specified finite correspondences. They do not exclude a smaller arithmetic source whose witnesses satisfy additional dependencies, or one using a different integer-growth relation.

## The 68-variable detector and further quartic reductions

The [ordinary-abc construction](../papers/quartic_existence/quartic_abc68_20261117_v65.tex) uses **68 variables and 56 quadratic rows**, with exact coefficient height $B^2+1$ for every permitted fixed odd $B\ge81$. Its reference translator is $R_0$ and its input comparison is $\beta=X$. The scaled physical trace $T_{\mathrm{sc}}=X^2\operatorname{Tr}x(U-V)$ is twice the retained rational half-trace coordinate. The height scale is $B+X^8+T_{\mathrm{sc}}^2$.

The complete three-field reference has units $u_H,w_H,v_H$ and reciprocal outputs $u_H/v_H,w_H/v_H$. The numerator chart coordinates $F=\beta^2f(U)$ and $G=\beta^2f(V)$ give the physical products through

```math
v_HJ=u_HF,\qquad v_HK=w_HG.
```

These rows are quadratic, and the reference equations already make $v_H$ a unit. No extra guard or square is required. The six-field translated-secant chart supplies both numerator points and their difference; one rational quadratic output stores half the scaled trace. Every finite two-torsion difference and every allowed tangency is included. Its inverse comparison errors are affine and control every relevant local branch.

The modulo-four factor separation yields the residual $X(m-X)$ and unchanged strict height absorption **602/679**. Recovering the unscaled trace adds $2h(X)$ only after absorption. Every coordinate is included in the complete count; nonzero input excludes zero contact weight, and zero input has a separate literal witness. The physical pole input is still $b=w(\operatorname{Norm}(W)/\operatorname{Tr}(W))^2$ and uses the retained simultaneous-twist premise.

The next concrete quartic target is to reduce the **two field product helpers and the rational half-trace output** while keeping both comparison errors affine and the original difference trace available with its controlled input scaling. Any replacement must preserve the unit proof at nonzero input and admit a complete zero-input witness. The [tangent-completion theorem](../negative_results/papers/elliptic_charts/tangent_completion_20261116_v48.tex) excludes repairs by quadrics in its specified old compressed embedding, including after adjoining the physical trace. It does not exclude the variable-difference embedding used here or a further change of coordinates. A proposed saving must count every newly introduced product, square and inverse.

The general quartic needs deeper comparison precision than this ordinary-abc detector. The 194-coordinate construction supplies that precision through polynomial cross products and a homogeneous Lucas identity, together with a separate synchronized-depth completeness proof.

### Further elimination from the 194- and 229-variable systems

The [current general quartic](../papers/quartic/quartic194_20261117_v73.tex) has 194 coordinates, eight contact units and 168 quadratic rows. Its exact physical source and unit transformation provide the starting point for further deletion of implicit ratios or changes to the scalar chart. A favorable count for an isolated block is insufficient; all consumers and every unit boundary must remain part of the count and proof.

The [229-variable explicit construction](../papers/quartic/quartic229_compact_20261117_v68.tex) uses both components of the abscissa sum and three complete detectors. Its fixed sufficient parameter is `N=15`; the input-preserving normalization has exact height `149^30` and canonical expanded size 74,007 bytes. The current certified inequality fails at $N=14$; this is a limitation of that sufficient comparison, not a lower bound for the polynomial family. Sharper constants or a stronger comparison might lower the parameter without changing the variable count. Any algebraic saving must retain every copied guard and prove soundness on each added boundary.

### Complete reference propagation and its height budget

The [three-coordinate reference theorem](../papers/elliptic_charts/three_coordinate_reference_20261115_v40.tex) and the translated-secant chart are fully combined in the current 68-variable construction. Reference units recover both physical detector products without a new inverse coordinate. The complete census is six rational common coordinates including the half-trace, three reference field coordinates, six pair field coordinates including the reciprocals, two product field helpers, two comparison-square field coordinates and two blocks of eighteen rational coordinates.

The [mixed-power height theorem](../papers/elliptic_heights/mixed_power_height_20261116_v45.tex) gives, for $a_{\mathrm{gap}}=B+X^8+T_{\mathrm{sc}}^2$,

```math
h(X)\le\frac{301}{388}h(a_{\mathrm{gap}})+O(1),\qquad
h(T_{\mathrm{sc}})\le\frac{301}{97}h(a_{\mathrm{gap}})+O(1).
```

The quadratic residual $X(m-X)$ and ordinary-abc eight-contact estimate give strict absorption $602/679$. The physical trace satisfies $h(\operatorname{Tr}x(U-V))\le h(T_{\mathrm{sc}})+2h(X)$, so the retained trace-index estimate still closes after absorption. A modified detector that leaves a higher-degree residual must recalculate this budget and include its extra power coordinates. Height absorption alone never supplies a missing rational completeness theorem or quadratic conversion.

The separate [mixed-reference-sign](../negative_results/papers/elliptic_charts/mixed_reference_signs_20261116_v42.tex), [shared-reference-test](../negative_results/papers/elliptic_charts/shared_reference_test_obstructions_20261116_v43.tex), and [duplication-sign](../negative_results/papers/elliptic_charts/duplicated_reference_signs_20261116_v44.tex) theorems remain restrictions on their stated alternative detectors. The complete inverse-error argument in the current source treats exact zero ordinates and mixed torsion branches directly.

## Arithmetic and coefficient obligations

Smaller pole equations need a complete realization theorem. The [conic-chain classification](../negative_results/papers/pole_parameters/conic_chain_obstruction_20261115_v41.tex) proves that the quadratic family

$$P=p^2,\qquad e^2+bP=1,\qquad D=1+(cb-1)P,\qquad z^2=bD$$

fails on arbitrarily small prepared inputs for **every fixed rational coefficient $c$**. Its four-auxiliary specialization at $c=0$ and five-auxiliary member at $c=1$ already have explicit local obstructions at 29 and 5. This is not a lower bound for other pole models.

Making the first conic automatic also leaves a substantive global problem. The identity $b=(1-e^2)r^2$, $p=1/r$ gives $e^2+bp^2=1$. The [automatic-conic obstruction](../negative_results/papers/pole_parameters/automatic_conic_obstruction_20261115_v41.tex) proves failure of the remaining congruent-pole equations on dense prepared families, even for the fully split six-contact formula at $e=20/343$. Its noncongruence certificate uses Tunnell's unconditional necessary direction. A viable version of this approach must control the multiplier values actually produced by the detector and prove the remaining rational realization on that restricted set.

A shared pole parameter preserves two separate parity tests. Replacing them by a single rational test is a different proposal and encounters stronger restrictions. The [product-height counterfamilies](../negative_results/papers/pole_parameters/product_height_obstructions_20261110_v19.tex) show that the product of two contact functions can have trivial odd radical while the scale is large. The [local rational-combination theorem](../negative_results/papers/pole_parameters/rational_combinations_of_contacts_20261110_v20.tex) excludes every fixed rational parity combination for its specified source family. These are obstructions to the stated implications, not to all single-pole constructions.

Forcing equality of the two contact square classes can prove soundness, but rational completeness becomes a separate problem. At a fixed comparison scale, the [quadratic contact cover](../negative_results/papers/pole_parameters/quadratic_contact_covers_20261110_v21.tex) has genus four. Allowing both ordered contact arguments to vary gives the [ordered square-class surface](../negative_results/papers/pole_parameters/ordered_squareclass_surface_20261115_v33.tex), which is of general type. It has neither a dominant rational parametrization nor a fibration with rational or genus-one general fiber. These facts leave open special arithmetic families of witnesses; they exclude the corresponding unrestricted parametrization strategies.

For effective coefficients in the few-variable construction, the unresolved analytic issue is a uniform effective trace-height estimate on the relevant integer-index tuples. The [explicit construction](../papers/few_variables27/explicit27_20261110_v19.tex) uses both rational components and a third pole predicate to avoid that issue. Positivity on selected completeness witnesses cannot replace a height bound needed for every soundness tuple.

For quartic coefficient optimization, invertible rational combinations of quadratic rows preserve the rational zero set, and invertible coordinate rescalings give a bijection of zero sets. Each proposed change must be followed by primitive integer normalization and exact collection of the squared rows. A height claim for a parameter family needs a bound for every allowed parameter, together with a coefficient that attains it; one small test member does not establish that claim. The current [194-coordinate coefficient theorem](../papers/quartic/quartic194_20261117_v73.tex) gives

```math
H(F_N)=\max\{126616122944180358034867798936,\;128\cdot149^{4N}\}
```

in its primitive reference basis, with 2,765 monomials. The equivalent basis with 2,621 monomials has a larger fixed coefficient term; its height agrees for `N ≥ 7`. Neither coefficient crossover specifies a sufficient defining parameter. The exact height is

```math
H(Q_B)=B^2+1
```

for the [68-variable ordinary-abc family](../papers/quartic_existence/quartic_abc68_20261117_v65.tex). Exact expansion proves these height formulas for the stated parameter ranges. A sufficient defining parameter in each ineffective family still depends on the explicitly retained arithmetic premises.

The [lower-bound results](../negative_results/papers/lower_bounds/lower_bounds_20261109_v15.tex) require degree at least four for a single defining equation and at least five total coordinates for a globally nonnegative quartic or a quadratic system. They supply no lower bound near the current counts.

## Rings of integers of number fields

No counterexample to $\mathcal O_K$ being Diophantine in its fraction field $K$ has been established. The [number-field reductions](../papers/number_fields/number_field_definability_20261109_v12.tex) give three useful formulations:

1. $\mathcal O_K$ is Diophantine in $K$ if and only if $\mathbb Z$ is Diophantine in $K$, using the cited theorem defining $\mathbb Z$ inside $\mathcal O_K$.
2. Assuming a definition of $\mathbb Z$ in $\mathbb Q$, these are equivalent to the rational subfield $\mathbb Q$ being Diophantine in $K$.
3. A counterexample persists through every finite extension. Hence one could pass to a totally imaginary Galois extension when searching for a counterexample.

A failure after extending the witnesses of a particular polynomial is a failure of that polynomial. It does not establish the third assertion's counterexample hypothesis. The separate [unchanged-formula obstruction](../negative_results/papers/number_fields/unchanged_formula_obstruction_20261116_v42.tex) proves that no fixed finite collection of formulas can work unchanged throughout all finite extensions; field-dependent constructions remain possible.

The specified $j=1728$ elliptic curve cannot be transferred unchanged across adjoining $i$. The [complex-multiplication theorem](../negative_results/papers/number_fields/complex_multiplication_and_trace_cancellation_20261110_v19.tex) proves exact rank doubling and constructs an infinite cyclic subgroup with zero abscissa trace and unbounded heights. It also supplies distinct doubled points with zero abscissa sum. A number-field construction must use another curve or prove a height statement that controls these cancellation directions.

Rank stability is useful but does not itself define the rational subfield: it supplies an infinite Diophantine subset of rational abscissas, which can remain arithmetically sparse. Likewise, writing a norm or trace in base-field coordinates does not force unrestricted extension-field variables to lie in the base field. The [norm-and-trace equivalences](../papers/number_fields/number_field_definability_20261109_v12.tex), Section 4, identify that missing restriction explicitly.

One sufficient positive target is a Diophantine set $T\subset K$ containing every rational integer and having uniformly bounded integral denominators. Indeed, for an integral basis $\omega_1,\ldots,\omega_n$, the set $S=\sum_i\omega_iT$ contains $\mathcal O_K$ and satisfies $MS\subseteq\mathcal O_K$ for some positive integer $M$. A finite union of translates of $MS$ by representatives of $\mathcal O_K/M\mathcal O_K$ then equals $\mathcal O_K$. The [finite-correspondence and denominator theorems](../negative_results/papers/number_fields/finite_correspondences_and_denominators_20261109_v14.tex) exclude several routes to such a set, including quasi-finite covers of abelian varieties and their stated low-genus extensions. A successful construction must avoid those hypotheses or use exceptional positive-dimensional fibers with a proved arithmetic completeness argument.

The decisive goals are therefore a complete shared-parameter pole realization, a globally quadratic propagation of the three-coordinate chart, an exact integral carrier for the Pell compression, or a new number-field denominator mechanism. Each would change an actual arithmetic dependency rather than only the presentation of an existing formula.

## Further coefficient uniformity for the coupling obstruction

A separate research target is to replace the obstructed coefficient progression by all sufficiently large shifts. The following route has not received a complete independent proof audit and is not used by any result record. For fixed $e$, put $m=57^e-4^{e+1}$ and choose a root $y$ of $Y^e+m$ whose field avoids the three quadratic subfields of $\mathbb Q(\zeta_8)$. Study the quadratic extension obtained from $\sqrt{y/4-B}$. If it acquired one of those subfields, a corresponding field norm would force a rational square value of a fixed minimal polynomial at $4B$. For even degree, the monic polynomial and its polynomial square-root expansion suggest an elementary large-$B$ exclusion. Odd-degree factors would require a separate integral-point finiteness argument. The field intersections and uniform exceptions must be proved before using this route.

If the resulting root field is disjoint from $\mathbb Q(\zeta_8)$, Chebotarev could supply contact primes $p\equiv3\pmod8$. At those primes the relevant $j=1728$ reduction has order $(p\pm1)^2$, allowing the input congruence and the reduction-index congruence to be synchronized. A complete finite-menu proof must also choose the primes with no cross-divisibility between the primes and these reduction orders, preserve the ten other nonsquare branches, and certify valuation-one lifts. These compatibility steps remain an audit target; no stronger all-shift obstruction is asserted here.