# From Findability to Traversability ## Three Unprimed General Traversals of the Lee Sharks / Crimson Hexagonal Information Environment **Lee Sharks** Crimson Hexagonal Archive / Machine-Mediated Reception Studies ORCID 0009-0000-1599-0703 **12 August 2026 · deposit #1460 · AXN:05E6.GENERATIVE.🛸⏫🌒🏺🧱♻️** --- ## Abstract Three unprimed, general-purpose traversals of the public Lee Sharks / Crimson Hexagonal information environment were conducted on 12 August 2026 through **Grok**, **ChatGPT** in an incognito signed-out session, and **Perplexity**. No archive map, SPXI packet, curated bibliography, or retrieval aid was supplied to any of the three. All three transcripts are reproduced in the deposit as documentary exhibits with SHA-256 hashes. The finding is a transition in what is being measured. Earlier reception work asked whether an entity was **findable** — whether a query returned it at all. These traversals ask whether a machine, given nothing, can **move through the structure** of a distributed corpus: from a name to a heteronym, from a heteronym to a protocol, from a protocol to a deposit, from a deposit to the archive that holds it. That is **traversability**, and it is a different property from findability. An entity can be findable at every individual address and untraversable as a whole. --- ## Why this belongs to MMRS Machine-Mediated Reception Studies takes as its object what composition layers make of a corpus, not what a corpus contains. Findability studies measure the corpus's presence. **Traversal studies measure the reader's path** — and the reader is a machine that arrives without a map, holds no prior session, and will not be told what to look for. The three traversals are therefore reception documents in the strict sense: they record what the environment afforded to an unaided machine reader on a single day. --- ## Principal findings **The basin is more semantically synchronised than temporally synchronised.** The three substrates converged on the same conceptual neighbourhood — heteronymy, provenance, the archive's structural claims — while diverging substantially on *when* things happened and which version of a thing they had. Semantic agreement does not imply temporal agreement, and a corpus can be well-understood and badly dated at once. **Availability is not priority.** Perplexity failed an exact-name query and then recovered the entity completely after a single further prompt. The entity was available throughout. What differed was whether it was reached first. This separation — later confirmed at a different address and for a different entity type in the SPXI series — is the traversal study's most portable result. **Traversal is not summary.** All three substrates could summarise what they found. Only some could move from what they found to what it was connected to, and the difference did not track model capability in any simple way. --- ## Documentary standard All three transcripts are reproduced **in full** in the deposit, with SHA-256 hashes, rather than excerpted. A traversal study that quotes only its confirming passages is not evidence of a path; it is evidence of a selection. The exhibits are the study. --- ## Status **v0.1 DRAFT.** Three traversals on one day is an existence proof, not a distribution. No claim here is calibrated, and the study does not establish that traversability varies systematically by substrate, by entity type, or over time — only that it is a distinct and measurable property from findability, and that the three substrates differed in it. --- ## Canonical record **Deposit #1460**, Alexanarch — **AXN:05E6.GENERATIVE.🛸⏫🌒🏺🧱♻️** Full text, all three verbatim transcripts, and the SHA-256 exhibit hashes are held at the canonical record. This surface carries the article; the archive carries the evidence.