3.1The primary text
This method lets the study set Bagehot's account of the one reserve beside two synthetic records, and ask what a shared measure really removes.
The primary text is Chapter VII of Walter Bagehot's Lombard Street: A Description of the Money Market (1873), with two sentences of Chapter II. It is read in the Project Gutenberg transcription listed in the parent study's texts.json (https://www.gutenberg.org/cache/epub/4359/pg4359.txt, eBook 4359). The file was downloaded for this study on 2026-10-07, and its SHA-256 digest was checked against the prefix recorded in texts.json (fb0f41f842b1c868). It is not marked as uncorrected optical character recognition. The passages this study reads are saved with it.
This study paraphrases Bagehot rather than quoting him, and says so each time it reports what the text says. Block quotes are the records' alone. The only words of Bagehot set between quotation marks are two sentences of Chapter II, each inside a sentence of this study's own and each word for word as the transcription has it. Dates and figures that appear in the paraphrases are Bagehot's, and this study computes nothing from them.
Chapter VII is chosen because it is Bagehot's account of the one reserve: who keeps it, how its adequacy was measured, and how it was used in the panics he reports. It contains an Economist article that Bagehot reprints and the reply of one director, Mr. Hankey; this study names whose view it reports each time.
3.2The records
The two records are published by WSTiger on its public atlas. They are synthetic. Each sets a scene, poses a question with three options, gives an example judgment that WSTiger marks as its own dated reading, and adds an Invalidation line saying when that judgment would be revisited. They are modern summaries written for the platform, not archival documents, and report no real bank, supervisor, firm, account or event. The originals are Korean; the parent study holds an English translation of each, made once and stored with a hash of the original. This study quotes the English.
- The same risk measure adopted by all banks (record wst:risk-ko-model-monoculture-past), dated "1996 · synthetic setting". It belongs to a thread WSTiger titles "When everyone uses the same model" and carries that thread's Insight.
- Screen Balance and Funds Available for Immediate Use (record wst:wt-atlas-expansion-b02-p24-q), dated "2026 · Synthetic Scenario". Its Scenario is set out as an assumption, a list of required evidence, and tags. Its Invalidation has a change condition and a "Next verification" line.
The first record's date places it in the past. It reports nothing that happened in 1996, and this study connects it to no real rule or banks. The first record's revisit condition contains the Korean word 초과 ("exceeding"), which is part of the term 한도 초과 ("limit breach"); the record's English renders the term as "limit breaches". It returns in 3.3.
3.3The ledger
The parent study reads 53 of WSTiger's records through a decision ledger. In plain words, the ledger is a table with one row per record and one column per question asked of every record: who decides; on what grounds, and with what missing; by when, and who waits meanwhile; whether the decision can be undone; what would make it change; and when it was judged. Each box is a cell. Some cells are copied or counted by code. Others are read by two models of different families, called coders, each giving a value, its basis (stated in the record, or implied) and a reason. A cell is settled when the two coders agree. The parent study registered its hypotheses in advance.
Of the 53 records, 35 are set in the present, 9 in the past and 9 in the future; all 53 are synthetic. Every one has a condition for revisiting its judgment (53 of 53). Only 5 of 53 (9%) of those conditions contain a number, a threshold or a time that code can see. None names a deadline, none names who decides, and none lists what is missing.
One caution comes from this study's own records, and it is stated here plainly. The code counts the first record's revisit condition among the five only because its Korean text contains 초과 ("exceeding", part of 한도 초과, "limit breach"), a word on the code's list. The condition holds no number, no threshold and no time. The count of 5 of 53 is a count of words matched by code, and this study does not treat it, for this record, as a numeric threshold.
This study uses the ledger sparingly: here, in one section of the discussion (7.4), and in the appendix.
3.4How the reading proceeds
Each case chapter opens with a claim about what Bagehot's text shows and reports that text at length. Only then does a record come in, to see whether the record's own words take a position on the question Bagehot's pages raise. This study reads; it does not judge which option in either record is right. Read in this order, the records can be shown to name a shared reaction and a gap in usable funds, while naming no holder who stands apart.