A Sárközy Theorem for Finite Fields (Q4655906)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 2142658
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | A Sárközy Theorem for Finite Fields |
scientific article; zbMATH DE number 2142658 |
Statements
A Sárközy Theorem for Finite Fields (English)
0 references
8 March 2005
0 references
differences
0 references
finite fields
0 references
ergodic theory
0 references
0.8496373295783997
0 references
0.8049404621124268
0 references
0.803952157497406
0 references
0.7995583415031433
0 references
0.7980984449386597
0 references
A result of \textit{A. Sárközy} [Acta Math. Acad. Sci. Hung. 31, 125--149 (1978; Zbl 0387.10033)] asserts that a set of integers having positive density contains two integers whose difference is a square. Theorem B of the paper states the following analogue. Given a positive integer \(k\) and an \(\varepsilon >0\), there is a bound \(M\) such that in every finite field with \(| F| >M\), every set \(A\subset F\) with \(| A| >\varepsilon | F| \) contains two elements whose difference is a nonzero \(g\)th power. This is derived from a theorem of \textit{P. Larick} [''Results in polynomial recurrence'', PhD dissertation, Ohio State Univ., Columbus, OH, 1998], of which a new proof is presented. This asserts that for a countable field \(F\), a Følner sequence \(\Phi _n\) in \(F\), polynomials \(r_1, \dots , r_l \in F[x]\) with \(r_i(0)=0\) and a unitary action \((U_g)_{g\in F^l}\) on some Hilbert space \(H\) we have always \[ \lim _{n\to \infty } | \Phi _n| ^{-1} \sum _{g\in \Phi _n} U_{r_1(g), \dots , r_l(g)} f = Pf, \] where \(P\) is the orthogonal projection onto the subspace \( \{h\in H: U_{r_1(g), \dots , r_l(g)}h=h \text{\;for\;all\;} g\in F\}\). The following generalization of Theorem B is announced later. Given polynomials \(r_1, \dots , r_l \) with \(r_i(0)=0\) and an \(\varepsilon >0\), there is a bound \(M\) such that in every finite field with \(| F| >M\), every set \(A\subset F^l\) with \(| A| >\varepsilon | F| ^l\) contains two elements whose difference is of the form \((r_1(g), \dots , r_l(g))\) with \(g\neq 0\). A deduction from Larick's theorem is outlined for the case \(l=1\). The referee remarks that the case of a single polynomial can be handled by the usual method of trigonometric sums, all we need is a decent estimate for the sums \(\sum \gamma (r(g))\), where \(\gamma \) is an additive character of the field, which are available (and we obtain that a bound of the form \(| A| >c\sqrt {| F| }\) suffices). The case when \(r(g)=g^k\) (Theorem B) can even be proved by an elementary counting argument, using only the fact that \(k\) dilatations of the set of \(k\)th powers cover the field; this works for \(| A| >c{| F| }^{1-1/k}\).
0 references