Use the convention that level counts the nonzero shifts, hence an arithmetic progression of length . The SZ property at level means that every measurable set of positive measure satisfies
This is the averaged multiple-recurrence property, stronger than merely finding one positive intersection. Shifting the indexing convention from terms to terms does not change the all-levels assertion. We work with probability measure-preserving systems.
Here is the Furstenberg correspondence principle construction proving the Szemerédi theorem. Let have positive upper density . Put equal to the indicator of on the positive coordinates and zero elsewhere, and let be the left shift. The binary full shift is a compact metric space in the product topology, and is a homeomorphism. Choose attaining the upper density and define the empirical measures
We use the standard weak compactness fact that every sequence of Borel probability measures on a compact metric space has a weakly convergent subsequence, where weak convergence means convergence of integrals of continuous functions. Pass to a subsequence with limit . For every continuous , telescoping gives
Passing to the limit proves . The cylinder set is clopen, so its indicator is continuous and
If every measure-preserving system has the SZ property at level , there is some for which , where . This is also a clopen cylinder set, so . For some sufficiently large , at least one term of the empirical average is positive: for some . Unwinding the coordinates gives
Therefore contains an arithmetic progression of length with positive common difference. Having the SZ property at every level gives the infinite-density form of the Szemerédi theorem.
The finite uniform form follows from the same argument. If, for fixed and , arbitrarily large admitted sets with and no such arithmetic progression, replace above by the indicator of , extended by zero outside that interval. Their empirical measures still have an invariant weak limit with . For every fixed positive , however, , since a point contributing to that cylinder would be an actual progression in . Thus for all positive , contradicting the SZ property. Consequently a sufficiently large depending only on forces the required progression.
A compact measure-preserving system is one for which every observable has relatively compact forward orbit under the Koopman operator . Equivalently, each of these orbits is totally bounded in the global norm; compactness refers to observables, not just to the underlying topological space. For an invertible system it is usual to use all integer times. One may also define compactness by density of almost periodic observables: the two definitions coincide because , so a finite net for an approximating observable is a finite net, with slightly larger radius, for .
To prove the SZ property, fix with , put , and choose
Compactness gives a finite -net for its orbit with centers , where . Let . For every integer , some center satisfies . Since is an isometry, cancellation gives
Thus the return set meets every interval for . In particular, is a syndetic set with lower density at least . This proves the needed syndetic near returns of an almost periodic observable without any assumption of ergodicity.
For , telescoping and the isometry property give
As is an indicator, . The union bound therefore gives
All other multiple-intersection measures are nonnegative. Averaging over the bounded-gap set yields
Thus every compact measure-preserving system has the SZ property at every level, including nonergodic systems and, under the forward-orbit definition, noninvertible transformations.
For an invertible probability measure-preserving system, an almost periodic observable is an whose Koopman operator orbit is totally bounded:
Equivalently, for every it has a finite -net in . For a noninvertible system use the forward orbit .
A factor map satisfies and almost everywhere. Disintegrate over this map and write
for the conditional L2 norm. A relatively almost periodic observable is an such that, for every , there are finitely many with
The center may depend on both and ; the finite list itself does not. Since is countable, the exceptional null sets can be combined. A compact extension of a measure-preserving system is a factor map for which these relatively almost periodic observables are dense in . This density definition should not be replaced by a claim that every observable already satisfies the uniform fiberwise finite-net condition.
For a proper factor map example, let carry independent fair coordinates and the two-sided Bernoulli shift , with . Set
using the product of the Bernoulli measure and the equal two-point measure. These are invertible measure-preserving systems, and the projection preserves measure and intertwines the transformations, so it is a two-to-one factor map.
Take . Then . Distinct coordinates are independent and have mean zero and variance one, hence
Its orbit has no finite sufficiently small net. Thus the Bernoulli coordinate observable is not almost periodic in global . In contrast, the conditional measure over is the equal measure on , and is constant on that fiber with value . The two centers give zero error on every fiber for every . Therefore is a relatively almost periodic observable for this map.
The example is also a finite-fiber compact extension. For a bounded observable on , each orbit value on a fiber is a pair in a fixed bounded subset of , which has a finite net. Its centers can be chosen as functions constant in , with prescribed values on the two labels. Bounded observables are dense in , so the extension is compact. The Bernoulli shift is mixing: cylinder events involving disjoint coordinate sets are independent for all sufficiently large shifts, and approximation by cylinder events extends this to arbitrary events. Thus and are ergodic. If an invariant function on is written , invariance gives and . The first function is constant; the second is -invariant and hence constant, and then its sign equation forces zero. This proves that the example's source is also ergodic.
For the positive-subset assertion, interpret the printed as , or as an invariant full-measure domain of the factor map; the prime is otherwise undefined and does not change the measure-theoretic argument.
First establish conditional norm covariance under a factor map. Invariance of the measures and uniqueness in the disintegration theorem for a probability measure give
on common full-measure sets. For the first equality, is another disintegration over the same fibers: it is supported on and integrates to . Uniqueness identifies it with , and iteration gives the norm formula.
Two elementary facts explain localization of relative almost periodicity to base sets. If is a relatively almost periodic observable and is measurable, then is also relatively almost periodic. Indeed, on a fiber,
When that indicator is one, use a center for ; when it is zero, use the additional center . Secondly, relative almost periodicity is closed under convergence in the uniform conditional L2 norm
If this norm of tends to zero, covariance bounds the corresponding error between every pair of orbit values by the same number. A finite -net for one sufficiently close is consequently an -net for .
Now take . The compact extension supplies relatively almost periodic observables with
Put . The Tonelli theorem gives
Hence for -almost every . Since , choose for which
has positive measure. The Egorov theorem gives a measurable of positive measure on which uniformly. Define
Then
For , the localization fact makes every relatively almost periodic, while
The conditional-norm closure fact therefore proves is a relatively almost periodic observable, with and positive measure. This is the positive-measure almost periodic indicator in a compact extension. No ergodic component or uncountable intersection of exceptional sets is needed in this construction; it in fact works under the same disintegration and compactness hypotheses without using the assumed ergodicity.
For every positive , require finitely many global centers such that the displayed inequality holds for every integer on almost every fiber of the factor map. A center can depend on the fiber and time; its finite list is fixed. This is weaker than almost periodic observable in global norm. For noninvertible systems use nonnegative times.