Failure of duality for power weights over finite residue rings

Less than 1 year old · traced to

For an integer ℓ≥2\ell \geq 2, let w(ℓ)w^{(\ell)} denote the corresponding power weight on Z/mZ\mathbb{Z}/m\mathbb{Z}.

Power-weight duality conjecture. The power weights w(ℓ)w^{(\ell)}, ℓ≥2\ell \geq 2, do not respect duality over any Z/mZ\mathbb{Z}/m\mathbb{Z} with m≥4m \geq 4.

The conjecture predicts a uniform failure of the relevant MacWilliams-type duality property for all power weights of exponent at least two and all moduli at least four. The surrounding discussion reports numerical evidence for the hypotheses needed to establish such failures, but does not resolve the conjecture.

References

Primary source

Jay A. Wood, “Weights on finite fields and failures of the MacWilliams identities”, arXiv:2601.02608 (2026).

Progress summary

Refreshed
Claimed solved

A posted, unverified complete-proof claim says the conjecture is true in every stated case, but no independent source has checked it.

Jay A. Wood posed the conjecture in 2026: every power weight of exponent ℓ≥2\ell \geq 2 fails duality over every residue ring with modulus m≥4m \geq 4. His paper presents the statement as open and supplies numerical support for part of a prime-modulus strategy.

Known results

  • Wood (2026): for prime moduli p≥5p \geq 5, some hypotheses of his failure theorem hold, while the remaining coefficient conditions are supported only numerically.
  • Wood (2026): explicit constructions for m=4,6,9m=4,6,9 and a lifting argument to general mm were identified as missing steps.

Posted attempt

A reader-written argument claims a complete proof for all m≥4m \geq 4 and ℓ≥2\ell \geq 2, using Wood’s theorem for primes, explicit pairs for m=4,6,9m=4,6,9, and divisor lifting; it claims differing dual coefficients A1A_1 or A2A_2. The attempt has not been independently verified.

Current status (as of August 2026): A complete proof has been claimed but remains unverified; absent validation, the conjecture is not established.

Sources

Solutions 1

ProofThis solution needs a summarySee full solutionHide full solution

Power weights over integer residue rings

For m≥4m\ge4, write Rm=Z/mZR_m=\mathbb Z/m\mathbb Z and let

wm,ℓ(x)=min⁡{r,m−r}ℓ,x≡r(modm),0≤r<m,w_{m,\ell}(x)=\min\{r,m-r\}^{\ell}, \qquad x\equiv r\pmod m,\quad 0\le r<m,

where ℓ≥2\ell\ge2 is an integer. Extend this weight additively to vectors. For a linear code C≤RmnC\le R_m^n, put

WC(y)=∑c∈Cywm,ℓ(c),Aj(C)=#{c∈C:wm,ℓ(c)=j}.W_C(y)=\sum_{c\in C}y^{w_{m,\ell}(c)}, \qquad A_j(C)=\#\{c\in C:w_{m,\ell}(c)=j\}.

Its dual is C⊥={z∈Rmn:c⋅z=0 for every c∈C}C^\perp=\{z\in R_m^n:c\cdot z=0\text{ for every }c\in C\}. Since all code pairs below have the same length, equality of these one-variable enumerators is equivalent to equality of the homogeneous enumerators in the original definition.

Theorem. For every m≥4m\ge4 and integer ℓ≥2\ell\ge2, there are linear codes C,D≤RmnC,D\le R_m^n, for some nn, such that

WC=WD,WC⊥≠WD⊥.W_C=W_D, \qquad W_{C^\perp}\ne W_{D^\perp}.

Thus the power weight does not respect duality. This proves Conjecture 9.5 of Jay A. Wood, arXiv:2601.02608.

The argument first handles prime moduli using Wood's Theorem 2.8. A singular weight matrix is dealt with directly; in the nonsingular case, a power-sum inequality verifies the remaining hypothesis of that theorem. Explicit constructions cover 4,6,94,6,9, and a divisor-lifting argument then covers every modulus. The divisor-reduction strategy was used by N. Abdelghany and J. A. Wood in their 2020 Lee-weight paper; the low-coefficient lifting argument needed here is proved below.

1. Two coefficients that survive lifting

For every modulus at least four, the only symbols of weight 11 are 11 and −1-1, and every other nonzero symbol has weight at least 2ℓ≥42^\ell\ge4. Consequently a vector of weight 11 is a singleton with entry ±1\pm1, and a vector of weight 22 has exactly two entries, both in {1,−1}\{1,-1\}.

Suppose d∣md\mid m, with d≥4d\ge4, and put q=m/dq=m/d. The map

ι:Rdn⟶Rmn,x⟼qx,\iota:R_d^n\longrightarrow R_m^n, \qquad x\longmapsto qx,

is injective, and the image of an RdR_d-linear code is RmR_m-linear. Moreover,

wm,ℓ(qx)=qℓwd,ℓ(x).w_{m,\ell}(qx)=q^\ell w_{d,\ell}(x).

Thus WC=WDW_C=W_D implies Wι(C)=Wι(D)W_{\iota(C)}=W_{\iota(D)}. If π:Rmn→Rdn\pi:R_m^n\to R_d^n is coordinatewise reduction, then

ι(C)⊥=π−1(C⊥),\iota(C)^\perp=\pi^{-1}(C^\perp),

because qc⋅z=0(modm)qc\cdot z=0\pmod m is equivalent to c⋅π(z)=0(modd)c\cdot\pi(z)=0\pmod d. Reduction gives a bijection between the ambient vectors of weight jj over these two rings for j=1,2j=1,2: it preserves their support and their entries ±1\pm1. Hence

Aj(ι(C)⊥)=Aj(C⊥),j=1,2.A_j\bigl(\iota(C)^\perp\bigr)=A_j(C^\perp), \qquad j=1,2.

It therefore suffices to construct a pair distinguished by A1A_1 or A2A_2 for each prime p≥5p\ge5 and for d=4,6,9d=4,6,9.

2. Every prime modulus

Wood's Example 9.4 already discusses the small primes 55 and 77. The argument here is uniform in the prime and does not assume that every power-weight matrix is nonsingular.

Fix a prime p=2t+1≥5p=2t+1\ge5. The symmetry group of w=wp,ℓw=w_{p,\ell} is H={1,−1}H=\{1,-1\}, and the minimum positive weight is attained on exactly this orbit. Choose representatives r1,…,rtr_1,\ldots,r_t of the nonzero HH-orbits and form the integer matrix

Wij=w(rirj).\mathcal W_{ij}=w(r_i r_j).

Every row and column has sum

Sℓ=∑a=1taℓ.S_\ell=\sum_{a=1}^t a^\ell.

The singular case

If W\mathcal W is singular, take a nonzero integral vector vv with Wv=0\mathcal Wv=0. The constant column sum gives ∑ivi=0\sum_i v_i=0. Choose an integer M>max⁡i∣vi∣M>\max_i|v_i|, and define multiplicities

ηi=M,ηi′=M+vi.\eta_i=M,\qquad \eta'_i=M+v_i.

Let CC and DD be the one-dimensional codes whose generator rows contain respectively ηi\eta_i and ηi′\eta'_i copies of rir_i. All multiplicities are positive, both generator maps are injective, and the codes have the same length. The identity Wη=Wη′\mathcal W\eta=\mathcal W\eta' says that the corresponding scalar multiples of the two generator rows have equal weights, so WC=WDW_C=W_D.

A dual vector of weight 22 has two entries ±1\pm1. For a generator row with no zero entries, these can cancel precisely when their two columns lie in the same HH-orbit. There are two choices of signs for each pair. Therefore

A2(C⊥)=2∑i(ηi2),A2(D⊥)−A2(C⊥)=∑ivi2>0.A_2(C^\perp)=2\sum_i\binom{\eta_i}{2}, \qquad A_2(D^\perp)-A_2(C^\perp) =\sum_i v_i^2>0.

The nonsingular case

Now suppose W\mathcal W is nonsingular. We verify the hypotheses of Wood's Theorem 2.8. The symmetry and minimum-orbit hypotheses have already been checked, and nonsingularity is precisely nondegeneracy in that theorem.

Let gg be a primitive element of Fp\mathbb F_p, write aj=w(gj)a_j=w(g^j) for j∈Z/tZj\in\mathbb Z/t\mathbb Z, and put

ch=∑j=0t−1ajaj+h.c_h=\sum_{j=0}^{t-1}a_j a_{j+h}.

The values aja_j are a permutation of 1ℓ,…,tℓ1^\ell,\ldots,t^\ell. Thus

c0=S2ℓ,ch=ct−h,∑h=0t−1ch=Sℓ2.c_0=S_{2\ell},\qquad c_h=c_{t-h}, \qquad \sum_{h=0}^{t-1}c_h=S_\ell^2.

If all the quantities in the last hypothesis of Wood's theorem were equal, then

pc1=⋯=pc⌊t/2⌋=2Sℓ2.pc_1=\cdots=pc_{\lfloor t/2\rfloor}=2S_\ell^2.

The preceding identities would imply

pS2ℓ=3Sℓ2.pS_{2\ell}=3S_\ell^2.

This is impossible. For real x>0x>0, set Sx=∑a=1taxS_x=\sum_{a=1}^t a^x. The function F(x)=log⁡SxF(x)=\log S_x is convex, since F′′(x)F''(x) is the variance of log⁡a\log a under the probabilities ax/Sxa^x/S_x. Consequently

ddxlog⁡S2xSx2=2F′(2x)−2F′(x)≥0.\frac{d}{dx}\log\frac{S_{2x}}{S_x^2} =2F'(2x)-2F'(x)\ge0.

At x=2x=2, the usual power-sum formulas give

(2t+1)S4−3S22=t(t+1)(2t+1)2(t−1)(t+2)60>0.(2t+1)S_4-3S_2^2 =\frac{t(t+1)(2t+1)^2(t-1)(t+2)}{60}>0.

Since ℓ≥2\ell\ge2, it follows that pS2ℓ>3Sℓ2pS_{2\ell}>3S_\ell^2. Wood's theorem therefore applies. Its proof, using Proposition 4.8, constructs equal-length codes with equal weight enumerators whose dual A2A_2-coefficients differ. This completes the prime case, including the coefficient needed for lifting.

3. Modulus four

Failure for this modulus is already a consequence of Wood's classification; see Corollary 9.6 of arXiv:2404.07154v1, published in Contemporary Mathematics 826 (2025), 361–430. The following explicit pair also supplies the particular low-weight coefficient needed for lifting.

Put a=2ℓa=2^\ell and q=a/2−1q=a/2-1. Thus a≥4a\ge4 is even and q≥1q\ge1. Use length n=a+qn=a+q. Let CC be generated over R4R_4 by the row containing aa copies of 11, followed by qq copies of 22. Let DD be generated by two rows: the first is 22 on the first a/2a/2 positions and zero elsewhere, while the second is 22 on the next a/2a/2 positions and zero elsewhere.

Both codes have four elements, and direct calculation gives

WC(y)=WD(y)=1+2ya2/2+ya2.W_C(y)=W_D(y)=1+2y^{a^2/2}+y^{a^2}.

The generator row of CC has no zero columns. The two-row generator matrix of DD has exactly qq zero columns. Since a vector of weight 11 is a singleton with entry ±1\pm1, we obtain

A1(C⊥)=0,A1(D⊥)=2q=a−2>0.A_1(C^\perp)=0, \qquad A_1(D^\perp)=2q=a-2>0.

4. Modulus six

Put

a=2ℓ,b=3ℓ,h=a+b−1,z=b−a−1.a=2^\ell,\qquad b=3^\ell,\qquad h=a+b-1, \qquad z=b-a-1.

Here z>0z>0, since 3ℓ>2ℓ+13^\ell>2^\ell+1. A generator row over R6R_6 with u,v,wu,v,w copies of 1,2,31,2,3, respectively, has nonzero scalar-orbit weights

(1abaa0b0b)(uvw).\begin{pmatrix} 1&a&b\\ a&a&0\\ b&0&b \end{pmatrix} \begin{pmatrix}u\\v\\w\end{pmatrix}.

The first two orbits, {±1}\{\pm1\} and {±2}\{\pm2\}, have size two; the third is {3}\{3\}.

Take the two multiplicity vectors

η=(2b−a, 0, 3a−2),η′=(2b+a, z, a−2).\eta=(2b-a,\,0,\,3a-2), \qquad \eta'=(2b+a,\,z,\,a-2).

Their entries are nonnegative integers, and each first coordinate is positive. Multiplication by the displayed matrix gives, respectively,

(a(3b−1)a(2b−a)2b(a+b−1)),(a(2b−a)a(3b−1)2b(a+b−1)).\begin{pmatrix} a(3b-1)\\ a(2b-a)\\ 2b(a+b-1) \end{pmatrix}, \qquad \begin{pmatrix} a(2b-a)\\ a(3b-1)\\ 2b(a+b-1) \end{pmatrix}.

Thus the two size-two orbit weights are interchanged and the third is unchanged. The second generator row is zz coordinates longer than the first. Append zz zero coordinates to the first row. The resulting cyclic codes C,DC,D have the same length, each has six elements, and WC=WDW_C=W_D.

A weight-11 vector belongs to the dual precisely when its nonzero coordinate is a zero column of the generator row. Hence

A1(C⊥)=2z,A1(D⊥)=0.A_1(C^\perp)=2z, \qquad A_1(D^\perp)=0.

5. Modulus nine

Put

a=2ℓ,b=3ℓ,c=4ℓ,S=1+a+c,L=S−2b.a=2^\ell,\qquad b=3^\ell,\qquad c=4^\ell, \qquad S=1+a+c,\qquad L=S-2b.

Convexity of x↦xℓx\mapsto x^\ell gives 2ℓ+4ℓ≥2⋅3ℓ2^\ell+4^\ell\ge2\cdot3^\ell, so L>0L>0.

First let C0C_0 be the cyclic code over R9R_9 generated by a row containing bb copies of each of 1,2,41,2,4. It has length 3b3b and size nine. Multiplication by a unit permutes the three unit negation orbits, whose symbol weights are 1,a,c1,a,c. The six unit multiples of the row therefore have weight bSbS. Its two nonzero nonunit multiples, by 33 and 66, have weight 3b23b^2. Thus

WC0(y)=1+6ybS+2y3b2.W_{C_0}(y)=1+6y^{bS}+2y^{3b^2}.

Next construct a two-dimensional ternary code BB. Use the four projective column directions

(1,0),(0,1),(1,1),(1,2)(1,0),\quad (0,1),\quad (1,1),\quad (1,2)

with multiplicities b,b,b,Lb,b,b,L, respectively. Its length is 3b+L=S+b3b+L=S+b. Every nonzero linear functional on F32\mathbb F_3^2 vanishes on exactly one of these four directions, and each direction is the kernel of exactly two nonzero functionals. Consequently its Hamming weight enumerator is

WBH(y)=1+6y(S+b)−b+2y(S+b)−L=1+6yS+2y3b.W_B^{\mathrm H}(y) =1+6y^{(S+b)-b}+2y^{(S+b)-L} =1+6y^S+2y^{3b}.

Embed BB in R9S+bR_9^{S+b} by multiplying all entries by 33, and call the resulting R9R_9-linear code DD. Each nonzero embedded symbol has power weight bb, so

WD(y)=1+6ybS+2y3b2=WC0(y).W_D(y)=1+6y^{bS}+2y^{3b^2}=W_{C_0}(y).

Append LL zero coordinates to C0C_0, obtaining CC of the same length as DD. All projective-column multiplicities are positive, so BB has dimension two and neither its generator matrix nor that of DD has a zero column. Therefore

A1(C⊥)=2L>0,A1(D⊥)=0.A_1(C^\perp)=2L>0, \qquad A_1(D^\perp)=0.

6. All remaining moduli

Every integer m≥4m\ge4 has a divisor that is either a prime p≥5p\ge5 or one of 4,6,94,6,9. Indeed, if its only prime divisors are 22 and 33, then either 4∣m4\mid m, or 9∣m9\mid m, or m=6m=6.

For such a divisor, the preceding sections provide equal-length codes with equal power-weight enumerators and different dual A1A_1 or A2A_2. Section 1 lifts that pair to RmR_m and preserves the distinguishing coefficient. This proves the theorem for every stipulated modulus and exponent. □\square