3.1The primary text
The primary text is Charles Mackay's Extraordinary Popular Delusions and the Madness of Crowds, first published in 1841. It is read here in the Project Gutenberg transcription listed in the parent study's texts.json (https://www.gutenberg.org/cache/epub/24518/pg24518.txt). The file was downloaded for this article and its SHA-256 digest checked against the prefix recorded in texts.json (d9799310f103f3fd). The transcription is titled Memoirs of Extraordinary Popular Delusions and the Madness of Crowds; its title pages are of a two-volume London edition dated 1852, and a footnote in the South Sea chapter refers to "The first edition of these volumes". The article cites the work as Mackay 1841, following the parent study, and says where the edition matters.
Two parts are read: the Preface and the chapter "The South-Sea Bubble" in the first volume. Every quotation from Mackay is copied from the downloaded file. The transcription wraps lines and marks italics with underscores; line breaks are joined with single spaces, and quotations have been chosen or cut with an ellipsis between two verbatim pieces so that no italic markup is silently dropped. Many passages are Mackay quoting others, a parliamentary history, a member's letter, Edward Gibbon's memoir; they are quoted as Mackay presents them, unchecked against their originals.
The parent study's reading note on Mackay says that later historians have questioned the scale and the stories of the tulip mania, another episode in the book, and that Mackay should be used as a record of how such episodes were told, not as a source of numbers. The rule is applied here to the South Sea chapter too. The article makes no claim about what happened, about the prices of stock or about any director's guilt; its claims are about the text: what Mackay says, in what order, with what judgments attached.
3.2The records
The two records are published by WSTiger on its public atlas. They are synthetic. Each sets an invented scene, poses a question with three options, gives an example judgment that WSTiger presents as its own dated reading, and adds an invalidation line saying when that judgment would be reviewed again. They are modern summaries written for the platform, not archival documents. They report no real trader, ship, desk, loss or firm. Their originals are in Korean; the parent study holds an English translation of each, made once for the study and stored with a hash of the Korean original. Quotations below are from the English translation.
- Reason written before voyage (record wst:risk-ko-outcome-rewrites-reason-past), seed title The Reason for the Choice, Written Before the Voyage. A synthetic trader and a partner, set in 1956.
- Post-mortem after loss (record wst:risk-ko-outcome-rewrites-reason-present), seed title Post-Mortem Review Report Written After the Loss. A synthetic trading desk, set in 2026.
Both belong to one thread. In the Korean originals the thread has a single title, 결과가 이유를 바꿔 쓴다. The English translation renders it as "The Outcome Rewrites the Reason" in one record and "The result rewrites the reason" in the other. The difference does not change the sense, but it is a small case of the article's subject: one sentence, set down once, given two forms by two later renderings.
3.3The ledger
The parent study, Before the Answer Comes, 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 for each question the study asks 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. Some cells are copied from the record by code. Some are counted by code, such as whether a condition for revisiting contains a number, a threshold or a time. The rest are read by two models of different families, each of which gives a value, a basis (stated in the record's words, or implied by the setting) and a reason. A cell is settled when the two agree. The study registered eight hypotheses in advance, with a rule for when the counts support or go against each.
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 of the 53 has a condition for revisiting its judgment, because the record form asks for one. Of those conditions, 5 of 53 (9%) contain a number, a threshold or a time that the code can see. No record names a deadline and none names who decides. These are properties of WSTiger's form as much as of any decision, and the article does not treat them as findings about traders or desks.
The coding of the two records read here is listed in the appendix in the ledger's own terms; the case sections describe the records in ordinary words.
3.4How the reading proceeds
Each case section opens with a claim about how Mackay's chapter tells the episode, reads the chapter at length, and then brings in one or both records. The records' authors were not reading Mackay. The comparison is between two kinds of document that face one problem from opposite ends: Mackay's chapter is an account written long after an outcome; the records are scenes about whether an account written before an outcome can resist being rewritten after it.