How statement and proof provenance work
The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.
- Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
- AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
- AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.
These labels describe origin, not correctness: citations and verification chips remain separate evidence.
A stationary chain need not be ergodic
Statement refuted
A strictly stationary Markov chain need not be ergodic. On with the identity transition matrix and , the chain started from is strictly stationary, but the strictly shift-invariant path event "zero occurs infinitely often" has probability ; the canonical shift therefore fails to be ergodic (Ergodicity relative to an invariant measure). The example is also reducible, so it does not contradict the ergodicity theorem for irreducible positive-recurrent chains.
Facts & Assumptions
Given: The two-point state space , the identity transition matrix , the probability , and the -chain started from on the canonical path space .
A transition matrix has nonnegative entries with every row summing to one. (Transition matrices and n-step probabilities)
A probability vector is invariant for a countable transition matrix exactly when for every . (Invariant and stationary distribution for a Markov kernel)
A process is strictly stationary when its finite-dimensional laws are unchanged by nonnegative time shifts; its canonical path law is the pushforward under the coordinate map, the left shift is , and a strictly stationary process is ergodic when is ergodic for that path law. (Stationary process and canonical path shift)
A measure-preserving system is ergodic for exactly when every strictly invariant event (that is, ) has or . (Ergodicity relative to an invariant measure)
Counterexample
Given: , the identity matrix , the law , and the chain started from .
Proof technique: identify the canonical path law explicitly, exhibit a strictly shift-invariant event of intermediate probability, and conclude non-ergodicity.
The identity matrix is a transition matrix, and is invariant: and likewise at , which is exactly the identity of [F2].
The event is a countable Boolean combination of coordinate events and is therefore measurable.
Both states are absorbing, so for every ; hence every finite-dimensional law of is the law of the constant tuple , which is unchanged by any nonnegative time shift, and the chain is strictly stationary in the sense of [F3]. Its canonical path law is , where and .
The event is strictly shift-invariant: for infinitely many , because deleting the first coordinate of a sequence does not change whether infinitely many of its entries vanish.
The left shift preserves : by step 2.1 the path law is supported on the two fixed paths , and , , so for every measurable .
Evaluating at : and , so , and by step 2.2 this is the measure of a strictly invariant event; since , [F4] shows that the canonical shift is not ergodic, even though is shift-invariant by step 3.1.
Boundary and axiom cases: the event is strictly invariant, not merely invariant modulo null sets, and step 2.2 verifies the identity on the whole path space; the value is neither nor , so the criterion of [F4] genuinely fails; and the events "infinitely many ones" behave the same way; if were concentrated on or on the chain would be ergodic, so the mixture is essential; the chain is reducible with two communicating classes and , which is exactly why the ergodicity result for irreducible chains does not apply; the explicit description of in step 2.1 makes no selection and no choice principle is used; and no convergence claim is made, the example refuting only the implication "stationary ergodic".
Depends on
Used by
Nothing in the library uses this result yet.
Dependency tree · two levels
13 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.