The claim
C1 Appendix D: “A directed information-retention profile can record I(G_j(U_0);U_n) for pragmatic features G_j, while a distributional drift profile uses D[P(G_j(U_n)),P(G_j(U_0))]. The quantities answer different questions: information about a particular ancestor versus evolution of the population’s feature distribution. Separating them prevents equating transmission with degradation.”
A59, from CASE-017: the chains “distinguish two phenomena that are easy to conflate: forgetting the original and forgetting which initial state the chain started in. In one contractive model, the separation between chains initialized in different pragmatic states approached zero, while the resulting distribution remained substantially different from the original … transmission converged toward a common convention rather than faithfully preserving the starting message.”
Manuscript §11: “An iterated chain may converge in surface form while diverging in ancestral information”; §8: the data-processing inequality “is not a theorem that messages get less comprehensible”.
A86 §3 (U31) gave it in information terms: the telephone game is progressive garbling, and “a message can become more legible while becoming less informative about what originally happened”. “Increasing conventionality alongside decreasing information about the origin” would be “a natural information-theoretic explanation of cultural attractors”, and the same reply added the caveat: “The Markov assumption matters here.” The manuscript keeps the caveat (§8) and drops the attractor explanation.
What it does not say
Which profile matters for fidelity: that depends on whether the question is about the original or about the convention (CLAIM-068).
The crystallized argument’s form (A110 §20): “The emergence of stable meaning within a community need not imply faithful retention of ancestral meaning.” The manuscript keeps the point and adds a caveat the argument lacked: “adding independent historical evidence to later interpreters changes the Markov graph and invalidates a naive application of the inequality” (§8).
Note of 2026-10-13: the general content is filed separately
Under ADR-040, the distinction between retention and drift is filed on its own as CLAIM-270, in the formal-apparatus line, resting on no claim. This claim now rests on it for the general content, and keeps CLAIM-056, for reading the chain as a reconstruction chain. Claims in other lines that need only the general content cite CLAIM-270, so setting pragmatic-transport aside does not take it from them.