The Setting
A real function on a product of prime fields can be tested by placing it at every vertex of a random configuration and taking the resulting product average. The quantity of interest is the root of this diagonal average, together with its behavior when some field coordinates are averaged out.
Let be a finite nonempty vertex set of cardinality , let be a state space, and let be a real linear functional on real functions on . The multilinear integral of a family of real functions on , its diagonal integral , and its diagonal root are defined by (1). (1) The positivity and multilinear Hölder hypotheses on are the conditions in (2), for every real and every family . (2)
The Tuple Laws
A permutation word on a block set is a function assigning a permutation of the positions to each block of .✦ The tuple vertices are the permutation words on the block set with the parameters in (3); is half the vertex count and is the list base. (3) For , an inactive-block address is a permutation word on .✦ An addressed list entry , for , is a pair of an inactive-block address and an index in , as in (4).✦ For , the inactive-block projection is the map that restricts a tuple vertex to . (4) The cyclic shift rotates the positions, returning the last position to zero.✦ For a finite block set , a group , an element , and a threshold , two addresses are threshold shift adjacent when at least blocks satisfy .✦ Write for the list constraint edges specifying the square-difference constraints.
An ordered position pair is a pair with .✦ For a tuple vertex , a block , and an ordered position pair , the position transposition is the tuple vertex obtained by replacing by and leaving the other block permutations fixed.✦ For , the transposition interaction is the real matrix on given by the counting sum in (5). (5)✦ For a real matrix on a finite set , a label set , and labelings , the label coincidence pairing , index energy , and index Boltzmann weight are defined by (6); both sums are counting sums. (6)
For a field , the nonzero square indicator is one on the nonzero squares of and zero elsewhere; write when is prime.✦ Given list values and a labeling , the square-constraint indicator and sampled tuple configuration are defined in (7), together with the tuple Boltzmann weight . (7) For a real observable on , the tuple component averages over uniform list values and uniform labelings, as in (8). The unnormalized tuple mixture weights each component by ; the full tuple law divides this mixture by its total mass, as in (8). (8) Here the outer sum is a counting sum. The reciprocal is taken to be zero when its argument is zero.
For a real linear functional on a finite configuration space, its point mass at is .✦ The tensor product of functionals on is the counting sum in (9), where the and the coordinate set are finite. For a family of primes , the product residue space is ✦, and the product tuple law is .✦ (9)✦
Let be the leading graph edge count.✦ The Boltzmann weight constant and interaction lower bound are given in (10). The same display defines the marked and unmarked reflection constants and , and the mass and reflection thresholds and . (10)
Coordinate Averaging and Residue Lifts
For a finite product , a subset , and points , the coordinate replacement is the point with coordinate for and otherwise.✦ The coordinate averaging projection on a finite product , for , is the operator on real functions on given in (11), with uniform probability average over . (11)✦
Keep the prime family fixed. For a prime , the centered residue delta is the real function on given in (12), with argument ; for , the centered residue product is the real function on given there, with argument . The residue lift component associated with a real sequence , a nonnegative integer length , and a support is the real function on given by the normalized counting sum in (12). (12) The interval used here is ; a sequence on the public interval enters through . At the reciprocal in (12) is zero.
The conductor of a support is . For a nonnegative integer cutoff and , the conductor-bounded support family and remaining family are families of subsets of given in (13); for a real sequence and a nonnegative integer length , the conductor-truncated residue lift is the real function on given there. Each term keeps the complete exact support . (13)
The Claim
Theorem 1. Let be finite and nonempty, and let satisfy the positivity and multilinear Hölder hypotheses (2). If is even, the map in (14) is a seminorm on the real functions on . (14)✦
Let be finite and prime for every . If for every , mean domination holds as in (15), for every real . (15)✦
For any finite prime family, choose with for every , and put . The product tuple law has the square-affine invariance (16) for every real observable on . (16)✦
Suppose with finite additive groups , and suppose is invariant under simultaneous translations for every . Under (2), every real and every satisfy the projection contraction (17). (17)✦
For the residue lifts, let be nonnegative integers, , and . The surviving-support identity is (18). (18)✦ If for every , these lifts also satisfy the diagonal-integral inequality (19). (19)✦
In the paper, the result reads: for the finite prime sets and real functions on of this section, is a seminorm with , square-affine maps with and nonzero square satisfy , and averaging the coordinates in any gives .
What Is Checked
The abstract statements concern real configuration functionals on finite nonempty vertex sets under the positivity and Hölder hypotheses, with even vertex count for the seminorm and simultaneous translation invariance for projection contraction. The product statements use finite prime fields with the stated thresholds, and the residue lifts keep the fixed ambient prime family.
The tuple mixture functional and the tensor product functional are real linear. For a finite nonempty vertex set , a state space , and a real linear functional on real functions on satisfying the positivity and Hölder hypotheses, the diagonal-root triangle inequality (20) holds for real functions on . (20)✦