The Layered Composition of Dream of the Red Chamber: Evidence and Theories
《红楼梦》的成书层累：证据与诸说
This is the second topical paper in a series on an interpretive reconstruction of the last thirty chapters of Dream of the Red Chamber. It settles an account left open in §3.4 of the overview: a full register of the evidence for the novel's layered composition, together with a survey of the theories about it. The paper accepts the layered-composition thesis: the surviving first eighty chapters are not a single-layer text written in one pass. It depends on one layer only: the main body of the text of the surviving first eighty chapters went through layering, and the one who integrated it into a book was the Cao Xueqin named in the chain of titles in chapter 1 and in the Zhiyan Zhai commentary. "Cao Xueqin" is taken here as a role in the text; the paper takes no part in the debate over who he was historically. Other layers are recorded separately: the commentators' change of title and their order to delete a passage; the unfinished, damaged and patched end of chapter 22; the supplying of missing chapters; and the changes made in the Cheng–Gao edition. The paper offers no new research on the history of the novel's composition. It does three things. First, it registers by provenance the material in the project's base text that bears on composition: the chain of titles and the general principles (fanli); the commentary's own accounts of deleted drafts, lost drafts and unfinished passages; inconsistencies between earlier and later parts of the text; the form of the manuscripts and the dates recorded in them; and records outside the novel. Second, it reports the three scholarly lines on how the book was composed, lists separately the dispute over whether the Zhiyan Zhai manuscripts or the Cheng editions came first, and then reports popular theories. For each it gives only the claim, the reasons offered and what remains unproven, without judging which is right or whether the materials are authentic. Third, it discloses the layering conjectures the project used in its backstage (its internal working notes): a six-stage working chart, whose chain of hand-offs turned out on checking to have been assembled by the project itself; a "composition signature" inferred from two solar-term dates, withdrawn after recalculation failed to match; and the few points that entered the reconstruction, each with its source, use and present status.
Keywords Dream of the Red Chamber; composition history; textual layering; Zhiyan Zhai commentary; chain of titles; survey of theories; interpretive reconstruction
Notes on the text
- This is the English version of a paper first published in Chinese (《红楼梦》的成书层累：证据与诸说, same concept DOI); it corresponds to the Chinese version 0.3. It is a rewriting for English-speaking readers, not a line-by-line translation: background has been added where an English reader needs it (item 8 below, and short glosses in the text), while the argument, the tables and the item-by-item entries are unchanged. Where the two differ in substance, the Chinese is authoritative.
- This is a pre-release. Materials and theories will be added, and items awaiting verification will be revised as they are checked (Appendix 5). The paper says only as much as has been done: what has been checked against the base text is marked as checked, what has been seen only at second hand is marked as such, and what has not been found is marked as not found.
- This paper is one of the supporting papers of the overview, People Have Qing (concept DOI 10.5281/zenodo.22985375; cited here as "the overview"), and a companion to An Audit of Forward References in the Zhiyan Zhai Commentary on the First Eighty Chapters of Dream of the Red Chamber (concept DOI 10.5281/zenodo.23007235; cited as "paper 1"). Where this paper differs from the overview, this paper is authoritative, and the next version of the overview will be revised accordingly (Appendix 6).
- Notes are numbered as in paper 1, "chapter.sequence": the n-th bracketed passage 【】 within a chapter of the project base text, so that 13.21 is the twenty-first bracketed passage in chapter 13. A note number is to be read as text, not as a decimal. Where one number holds notes from several manuscripts recorded one after another, each is quoted separately by manuscript. Manuscript attributions and quotations are from the project base text unless otherwise stated.
- Sources are graded by what was actually seen. Where a publication or the person's own writing has been seen, it is quoted directly; where it has been seen only in an online reprint, it is marked "reprint"; where it is known only from someone else's account, "second-hand"; where only a catalogue record has been seen, "catalogue"; where it could not be found, "not found" or "to be checked."
- "The reconstruction" means chapters 81–110 as written by the project, cited as "Rec. ch. n." The reconstruction is revised round by round; where this paper mentions its plot, it states only the choices made, and is not tied to the wording of any one round of drafts.
- All material and claims concerning the Guiyou manuscript (the Wu shi Shitou ji 吴氏石头记) are left to paper 3; this paper mentions it in a sentence at most.
- A word of background for readers new to the field. Dream of the Red Chamber (Honglou meng, also known as The Story of the Stone, Shitou ji) is conventionally held to have circulated in manuscript from the 1750s, in about eighty chapters, together with a running commentary whose writers present themselves as close to the author and sign with names such as Zhiyan Zhai ("Red Inkstone Studio") and Jihu Sou ("the Old Man of the Odd Tablet"; the notes also sign Hu Sou, "Old Tablet," and Jihu Laoren); how many hands these names stand for is itself debated. The commentary manuscripts are named by a cyclical year recorded in them, by a former owner or by a preface: Jiaxu, Jimao, Gengchen, Mengfu and Qixu; the original of the Jing manuscript is reported lost, and its notes are known only from a transcription (its authenticity is disputed; see §3.2). In 1791 and 1792 Cheng Weiyuan and Gao E printed a 120-chapter version (the Chengjia and Chengyi editions), and its last forty chapters have travelled with the book ever since. The notes date themselves by the sixty-year cycle; the conventional equivalents in the Qianlong reign are jiaxu 1754, jimao 1759, renwu 1762, yiyou 1765, dinghai 1767 and jiawu 1774, but this paper does not rest anything on them, and one theory reported in §5.2 places renwu sixty years earlier. All translations are the author's own: of the text, of the commentary, and of the scholarly and popular writings quoted. As in paper 1, Hawkes's published translation is not used, because the argument often turns on the wording of the project base text; Chinese is given in parentheses where it does. Personal names are given in pinyin, except for scholars known in English by another form (Eileen Chang, Wu Shih-ch'ang); some place names and epithets are translated, a few following Hawkes (for example Greensickness Peak). Works in Chinese are referred to in the text by an English rendering or a pinyin title; the original titles are in Appendix 1 and the References.
1 Stance
First, where this paper stands. Each point below follows rulings the author has made over time.
The layered-composition thesis is accepted. Dream of the Red Chamber is not a single-layer text written in one pass. Chapter 1 itself says that the book changed its title several times and passed through the hands of several title-givers; the commentary records drafts that were deleted, lost or left unfinished. Section 3 registers these materials one by one.
The paper depends on one layer only. The main thesis of this paper, and of the whole series, depends on one thing only: the main body of the text of the surviving first eighty chapters went through layering, and the one who integrated it into a book and fixed its centre was the Cao Xueqin named in the chain of titles and in the commentary.
This sentence needs three qualifications.
- "Cao Xueqin" is taken as a role in the text. It means the one named in the chain of titles and the commentary, who "read it over for ten years and revised it five times," and whom the notes also call Xueqin, Qinxi or Qin. Who this person was historically is a question on which each view has its reasons; the project has no new evidence, and so takes no part in that dispute.
- "Main body," because not every character of the surviving first eighty chapters is from this one hand. Commentators changed the title and ordered a passage deleted; the end of chapter 22 is unfinished, damaged and patched, and this paper does not rule on who wrote the supplied text or on what it was based; chapters 64 and 67 were missing and were supplied; and the Cheng–Gao edition altered the first eighty chapters. Each of these is recorded as another layer and is not included here (§7).
- "Fixed its centre" is the project's working judgement, which this paper does not argue. What the centre is belongs to the overview's interpretation; it is for the overview to argue and to test, and it is not brought back here as evidence for this layer.
No new research on the history of composition; the paper reports and does not judge. How many layers there were, and whose hand produced each: scholarly and popular theories alike are reported as they stand, each with its claims, its reasons and what remains unproven. The paper does not rule on who is right and takes no part in the dispute. Quotations, sources and chains of citation are nonetheless checked item by item; reporting is not copying.
The popular theories each have their reasons, and they have helped the project think. Section 6 sets out, item by item, what they helped with, divided into method and material. Wherever the project has used one, the paper says where, and whether it has since been withdrawn.
The commentary is testimony whose provenance must be kept. This is as in paper 1. When two notes disagree, it may be because two drafts differed; the paper records this only as a possibility, and does not use it to judge which note is wrong.
No part in disputes over authenticity. The project chooses to believe that the book is layered and existed in several drafts; a binary of genuine and forged does not help the reconstruction. Where the provenance of a material is disputed, as with the notes of the Jing manuscript, certain manuscripts or certain outside records, this paper registers it by its provenance as a material, together with the layer it claims, notes that it is disputed, and does not rule on its authenticity. To register a layer is not to confirm that the layer historically existed; where an inference about the history of composition is involved, the paper states which source it is conditional on. The paper can stay out of the dispute because it has no need to enter it: the one layer it depends on stands even without the commentary being genuine (§7).
The Guiyou manuscript is left to paper 3.
2 What This Paper Means by "Layer"
Different theories use the word "layer" in different ways. This section first sorts out where layers come from, so that the theories are not run together.
| Where the layer comes from | Meaning | Examples |
|---|---|---|
| The author's own revision | The same author rewrote the book several times, and traces of old drafts lie under the new | "read it over for ten years and revised it five times"; Baoyu's age lowered from revision to revision (Eileen Chang) |
| Merging of old drafts | Two or more drafts combined into one book | A Mirror for the Romantic and The Story of the Stone combined (Zhou Shaoliang, Du Chungeng) |
| Another's old draft | Someone else's draft, reworked by Cao Xueqin | Yurui's record; Dai Bufan's theory of "Brother Stone's old draft" |
| Commentators and copyists | Notes, recorded dates, records of collation, and the splitting and joining of manuscripts form a layer of their own; commentators also altered the main text | Change of title: "when Zhiyan Zhai copied it and commented on it again in the jiaxu year, he went back to the title The Story of the Stone"; ordered deletion: 13.89, "and so ordered Qinxi to delete it"; collation: 75.2, "collated on the seventh day of the fifth month of the twenty-first year of Qianlong" |
| Loss and patching | Drafts lost, left unfinished or damaged, and what was supplied later | 20.20, "lost by a borrower"; 26.24, "lost, no manuscript"; 22.104, "from here on damaged and lost; what follows is a later supplement"; 22.106, "this chapter was not finished when Qin [Xueqin] died"; the missing and supplied chapters 64 and 67 |
| Later hands | Continuations, edited texts, and alterations to the first eighty chapters | The last forty chapters of the Cheng–Gao edition, and its alterations to the first eighty; the Guiyou manuscript (paper 3); the edited base text the project uses |
The table should not be read too strictly.
- It classifies where materials come from. It is not a sequence of six stages already proven, and its rows are not mutually exclusive.
- The row on merging says how a layer was formed; the other rows say whose hand it came from and what it went through. They are not on one axis. What was merged might be the author's own draft, as when 1.61 is read as an old work of Xueqin's; or someone else's, as Yurui records.
- Stages of composition, layers of transmission and later re-creation must be marked separately. The existence of the last kind does not directly prove the particulars of the first. That today's edited texts carry the hands of modern editors proves that the material before us has an editorial layer; it cannot show how many authors the book had in the eighteenth century or before.
- Paper 1's ground for non-adoption N6 (original lost or deleted by the author) and its entries "filed as a record" (kept without a ruling) mostly record layers of the loss-and-patching kind.
The project's reconstruction is also a layer, and a self-aware one. The overview calls it interpretive reconstruction, with every stroke marked for the layer it stands on. What it proves is the project's own addition, not the prehistory of the original.
3 Materials on the Making of the Book: The Text's Own Account, the Commentary, and Outside Records
This section is the core of the paper. It gathers the materials on the making of the book that fall within the paper's scope: the chain of titles and the general principles in the project base text, the commentary's own statements, inconsistencies within the text, the form of the manuscripts and the dates recorded in them, and records outside the book. The works used are listed in the References; searches closed in September 2026, and what is still missing is listed in Appendix 5. The section does not exhaust the textual variants, the arguments of each scholar, or the popular theories.
Each item gives the original, its location and how scholars read it, and then the one layer this paper takes from it. As many readings are given as can be found. Where no second scholarly reading can be found, the paper says "no specific alternative seen," and does not invent one on scholarship's behalf.
3.1 The chain of titles and the general principles
Chapter 1 has a chain of titles:
From then on the Daoist Kongkong, from emptiness seeing form, from form giving rise to feeling, passing feeling into form, and from form awakening to emptiness, changed his name to the Passionate Monk, and changed the title from The Story of the Stone to The Record of the Passionate Monk. Then Wu Yufeng titled it Dream of the Red Chamber. Kong Meixi of Eastern Lu titled it A Mirror for the Romantic. Later, because Cao Xueqin, in the Mourning-the-Red Studio, read it over for ten years and revised it five times, compiled a table of contents and divided it into chapters, it was titled The Twelve Beauties of Jinling.
After an intervening title poem comes one more sentence:
When Zhiyan Zhai copied it and commented on it again in the jiaxu year, he went back to the title The Story of the Stone.
The general principles (fanli) of the Jiaxu manuscript (1.2) say: "This book has very many titles. Dream of the Red Chamber is the name that sums up the whole. It is also called A Mirror for the Romantic, as a warning against giving way rashly to romantic feeling. It is also called The Story of the Stone, being a self-likening: the things the stone recorded (自譬石头所记之事)."
This is the book's most direct account of how it was made. Every school has to explain "revised it five times," and there are broadly three readings.
- The titles correspond to drafts. Chen Qinghao holds that of the five titles only three, A Mirror for the Romantic, The Story of the Stone and Dream of the Red Chamber, name drafts of different stages (second-hand). Shen Zhijun matches each layer to a title: the first draft was called A Mirror for the Romantic; once the story of the stone was strengthened it became The Story of the Stone; after further revision it may have been renamed The Record of the Passionate Monk; after the Prospect Garden was enlarged and the characters' ages lowered it became The Twelve Beauties of Jinling; and finally it was named Dream of the Red Chamber (second-hand). Li Qingyu reads the five titles as four stages in the development of the author's conception of fiction: a vehicle for admonition; a true record of events; a tale of wonders offered to the world; play and dream as one (reprint).
- The titles correspond to two books. Du Chungeng, relying on the phrase "the name that sums up the whole" in the general principles, sets out a theory of two books combined (§4.2). Zhou Shaoliang, answering Lu Xun's view that the different titles were a flourish of shifting appearances (摇曳见态), holds that they have a factual basis and reflect a real history of compilation (reprint).
- The chain of titles is the author's arrangement. The Jiaxu upper-margin note 1.62 says that the prologue (the "wedge," 楔子) was written by the author himself, and that "Xueqin read it over and revised it" is the author's "cunning" (式猾) (§3.2). On this reading, the narrative posture of the chain of titles cannot be taken, without interpretation, as a file recording the book person by person and stage by stage.
The chain of titles presents the book as a text that passed through several titles, through reading over and revising, and through a second commentary. The project takes this as a working understanding: the last title was given by the one who "read it over for ten years and revised it five times," and a commentator then changed the title back. That is the chain's own account, and the reading that it is the author's arrangement still stands alongside it. The sentence in which a commentator changes the title is recorded in this paper as a record of the commentators' layer.
There is one more layer to record. It has been said that the Wu Yufeng sentence and the sentence on the second commentary of the jiaxu year are both peculiar to the Jiaxu manuscript and absent from most others. This paper has not yet seen the collation notes themselves, nor checked manuscript by manuscript; this is to be checked. If it is so, the chain of titles itself has a layer between manuscripts, which can be read in two ways: added in the Jiaxu, or deleted in the others. The various attempts in §5 to match Wu Yufeng to a historical person rest on exactly this sentence.
3.2 What the commentary says about the making and alteration of the book
The commentary contains a set of remarks on how the book was written, how it was altered, and where parts of it were lost. The table below registers them under the note numbers of paper 1. "Tier 3" is the tier that paper 1 does not rule on note by note; N6, N8 and N9 are paper 1's numbered grounds for not adopting a note (original lost or deleted by the author; conflict with the text of the first eighty chapters, the text being followed; two notes at odds, one being taken); "filed as a record" (报存) means reported and kept without a ruling.
| Note | Manuscript and type | Gist of the text | Entry in paper 1 | Remarks |
|---|---|---|---|---|
| 1.61 | Jiaxu upper-margin | "Xueqin formerly had a book called A Mirror for the Romantic, with a preface by his younger brother Tangcun. Now Tangcun is gone; seeing the new, I remember the old, and so I keep it." | Tier 3 | The watershed between the second and third lines |
| 1.62 | Jiaxu upper-margin | "If it is said that Xueqin read it over and revised it, then who wrote this prologue, from the opening of the book to this point? This shows how very cunning the author is. There is much of this kind later. This is just where the author paints clouds and mist and blurs the outlines; readers must by no means let themselves be taken in by the author: only then do they have a great eye." | not a forward reference | The base text reads 式猾; the common reading is 狡猾, "cunning" |
| 1.63 | Jiaxu upper-margin | "On the last night of the renwu year, the book not yet finished, Qin [Xueqin] wept his tears dry and died … written in tears, eighth month of the jiawu year." | Tier 3 | — |
| 1.140, 2.5 | Jiaxu double-line interlinear | "I say that in writing this book Xueqin also meant to hand down his poems"; "talent and feeling of this kind were naturally what Xueqin excelled at all his life" | 1.140 fulfilled (referent inferred); 2.5 Tier 3 | To be read with 1.62 |
| 13.3 | Jiaxu | "The Tianxiang Tower episode was suppressed because one could not bear to set brush to it" | fulfilled; suppressed-episode layer N6 | This page of the Jiaxu is damaged; the missing characters are supplied from the Gengchen and the Jing (note in the base text) |
| 13.8 | Jing | "I have pardoned her for now, and so ordered Qinxi to delete the passages on 'the hairpin left behind' and 'changing clothes'; that is why this chapter has only ten pages: with the Tianxiang Tower episode deleted, four or five pages are gone" | fulfilled; layer of the "hairpin" and "changing clothes" passages N6 | Jing only |
| 13.21 | Jiaxu upper-margin | "Nine characters tell the whole Tianxiang Tower affair: this is writing by not writing." | Tier 3 | The base text notes "the Jing manuscript adds the signature 'Tangcun'," referring to this note |
| 13.21 | Gengchen upper-margin | "This note may be followed." | Tier 3 | — |
| 13.21 | Jing upper-margin | "This note may be followed. That the whole chapter deliberately hides how Keqing died is my great act of compassion. Alas, alas! Late spring, renwu. Hu Sou." | Tier 3 | The long form with "my" (余), "late spring, renwu" and "Hu Sou" (笏叟) is found only in the Jing; see also 13.90 |
| 13.51 | Jiaxu side note | "This supplies undeleted text of the Tianxiang Tower." | Tier 3 | Under the sentence in which Ruizhu dashes herself against a pillar |
| 13.52 | Jing side note | "This too is a stroke left undeleted." | Tier 3 | Jing only |
| 13.87 | Jiaxu upper-margin | "This chapter has only ten pages, because with the Tianxiang Tower episode deleted, four or five pages are gone." | Tier 3 | — |
| 13.89 | Jiaxu | "I have pardoned her for now, and so ordered Qinxi to delete it." | fulfilled; deleted-Tower layer N6 | — |
| 13.90 | Gengchen | "That the whole chapter deliberately hides how Keqing died is a great act of the compassionate heart. Alas, alas! Spring, renwu." | Tier 3 | The same purport as the Jing upper-margin note on 13.21, in different words |
| 17.1 | Gengchen | "This chapter would properly be divided into two." | Tier 3 | In the Jimao and Gengchen, chapters 17 and 18 are not divided |
| 18.17 | Gengchen double-line interlinear | The three verdicts of the Tertiary Register (又副册) are given as Qingwen, Xiren and Xiangling; "the rest are not much touched on; presumably they are … Ping'er and others, without doubt" | N8; Ping'er layer N9 | At odds with 6.48; the Secondary Register layer of 6.48 is taken |
| 18.19 | Gengchen upper-margin | "The twelve beauties cited here are none of them certain; all are guesses. Only at the end (回末), with Jinghuan's Roster of Feeling, will one know the names in the principal (正), secondary (副), further secondary (再副), and third and fourth secondary (三四副) registers." | transposed fulfilment | Points to the Roster at the end (回末; see paper 1, note 1); does not say explicitly that the commentator saw it himself |
| 20.20 | Gengchen upper-margin | "Xiren's proper text bears the heading 'Hua Xiren, constant from beginning to end.' I saw it only once, at the time of a fair copy; it and five or six drafts, among them 'Consoling Baoyu at the Prison Temple,' were lost by a borrower" | fulfilled; lost-drafts layer N6 | The commentator's own statement that he saw the draft ("I only saw," 余只见) |
| 22.22 | Gengchen upper-margin | "Xifeng choosing the plays and Zhiyan holding the brush: few now know of these things. Is that not cause for regret?" | Tier 3 | — |
| 22.23 | Gengchen upper-margin | "The earlier note said 'few know of it'; now, in the summer of dinghai, only one decrepit thing is left. How can I not grieve?" | Tier 3 | Follows 22.22 |
| 22.24 | Jing upper-margin | "The earlier note said 'few know of it'; Qinxi, Zhiyan, Xingzhai: these gentlemen have all departed one after another; now, in the summer of dinghai, only one decrepit thing is left. Is it not grief enough to kill?" | Tier 3 | Jing only |
| 22.104 | Gengchen upper-margin | "From here on damaged and lost; what follows is a later supplement (系再补)." | filed as a record | The common reading is 俟再补, "awaiting supplement" |
| 22.106 | Gengchen | "This chapter was not finished when Qin [Xueqin] died. Alas, alas! Summer, dinghai. Hu Sou." | filed as a record | At the chapter's end |
| 26.24 | Jiaxu upper-margin | "The whole long chapter on Hongyu and Qianxue at the 'Prison Temple': alas, lost, no manuscript." | fulfilled; original-form layer N6 | Unsigned |
| 26.24 | Gengchen upper-margin | As above, with eight more characters: "Alas, alas! Summer, dinghai. Jihu Sou." | as above | According to a note in the base text |
| 26.153 | Jiaxu | "Alas, the writing on 'Wei Ruolan at the archery ground' is lost, no manuscript. Alas, alas!" | fulfilled; original-form layer N6 | At the chapter's end; unsigned |
| 26.120 | Gengchen upper-margin | "Alas, there is no manuscript of the writing on 'Wei Ruolan at the archery ground.' Alas, alas! Summer, dinghai. Hu Sou." | fulfilled; original-form layer N6 | Reads "no manuscript" (无稿), without "lost" (迷失) |
| 28.82, 28.83 | Gengchen upper-margin; Jiaxu side note | "Drinking from a great bowl: the day the Western Hall brought forth nine-stalk lingzhi"; "Who has lived through this? Alas, alas! The old affairs of the Western Hall." | Tier 3 | "Western Hall" is usually read as Cao Yin's study; the first proponent and the source are to be checked |
| 75.2, 75.3 | Gengchen | "Collated on the seventh day of the fifth month of the twenty-first year of Qianlong"; "Mid-Autumn poem missing; awaiting Xueqin" | 75.2 Tier 3; 75.3 filed as a record | 75.2 is recorded as the year of collation |
What follows sets out, group by group, how scholars read these notes.
Old draft and title (1.61, 1.62).
1.61 is the watershed between the second and the third lines. Read as an old work of Xueqin's own, with a preface by his younger brother Tangcun, it is the basis of the line of two books (or two sources) combined. Wu Shih-ch'ang's theory of "Tangcun's short prefaces" comes from it: he held that some of the chapter-head general comments in the surviving Zhiyan Zhai manuscripts were originally short prefaces written by Tangcun for A Mirror for the Romantic, preserved by Zhiyan, in the words "and so I keep it" (according to Liang Yusheng's account of a poem Wu wrote on his own book: "Tangcun's short prefaces are plainly there; the Redology experts, alas, do not know"; chapter and page in Wu's book to be checked). Read as an old book by someone else that Xueqin obtained, it connects with Yurui's "Cao Xueqin obtained it" (§3.5), and belongs to the line of another's old draft. There are also two readings of what was "kept" (故仍因之): the title, or the preface.
1.62 is evidence pointing the other way, and what it points against must be stated precisely. The commentator asks: if Xueqin only read the book over and revised it, who wrote this prologue at the opening? So "Xueqin read it over and revised it" is a device the author played on purpose, and "readers must by no means be taken in by the author." The line that holds Cao Xueqin to be the original author cites this note. What it tells against is "Xueqin was only a reviser," and the reading that takes each link in the chain of titles literally as a different author. It does not tell against "the same author revised the book again and again": the commentator says that the prologue is from the author's hand, not that the book was never revised. Elsewhere in the same manuscript the commentary has 1.140, "I say that in writing this book Xueqin …," and 2.5, "naturally what Xueqin excelled at all his life," which can be read alongside it. As for how the third line answers this note, this paper has seen no specific alternative reading, and says so.
Deleted draft: the Tianxiang Tower (13.3, 13.8, 13.21, 13.51, 13.52, 13.87, 13.89, 13.90).
This is the alteration the commentary records in most detail. The Jiaxu chapter-end general comment 13.89 says that the two matters of the family's future which Qin Keqing entrusts to Xifeng command admiration, and so "I have pardoned her for now, and so ordered Qinxi to delete it"; the Jing 13.8 has the same text, with an added passage on "'the hairpin left behind' and 'changing clothes'" and "that is why this chapter has only ten pages." The Jiaxu upper-margin note 13.87 records the page count separately ("four or five pages are gone"), so that the page count of the deleted draft does not rest on the Jing manuscript alone. The Jiaxu 13.3 says that the Tianxiang Tower episode was suppressed because one "could not bear to set brush to it."
By the commentator's own account, the one who deleted it with the brush was Qinxi, and the one who ordered the deletion was the commentator.
The Jiaxu side note 13.51 stands under Ruizhu's "she too dashed herself against a pillar and died," and says that this "supplies undeleted text of the Tianxiang Tower"; the Jing side note 13.52 adds, "this too is a stroke left undeleted." In the commentators' own words, the deletion was incomplete, and traces of the layer remain in the text. Of what the commentators record about chapter 13, this is the most direct.
13.21 must be read manuscript by manuscript. Within one bracketed passage the base text records three manuscripts. There is a Jiaxu upper-margin note, "nine characters tell the whole Tianxiang Tower affair: this is writing by not writing," with the base-text note "the Jing manuscript adds the signature 'Tangcun'," which refers to this note. There is a Gengchen upper-margin note of only four characters, "this note may be followed." And there is a Jing upper-margin note, which after "this note may be followed" goes on: "That the whole chapter deliberately hides how Keqing died is my great act of compassion. Alas, alas! Late spring, renwu. Hu Sou." In this place, this long form, with "my," "late spring, renwu" and "Hu Sou," is marked as the Jing manuscript only. There is also a Gengchen note, 13.90, which records a shorter text to the same effect: "That the whole chapter deliberately hides how Keqing died is a great act of the compassionate heart. Alas, alas! Spring, renwu." It has no "my," adds the word "heart" (心), is dated "spring, renwu," and is not signed "Hu Sou." The two notes are set side by side as separate entries, and neither the signature nor the date of one is transferred to the other.
The commentator gives the reason for the deletion as "pardon." The surviving text indeed lacks this episode, so Cao Xueqin did not disobey the order. Why Cao Xueqin himself went along with it is another question; the project has a conjecture, disclosed as it stands in §6.3.
Lost and unfinished (20.20, 26.24, 26.120, 26.153, 22.104, 22.106, 75.3).
Several notes record that drafts of the later part existed and were lost: 20.20, "lost by a borrower"; 26.24, a Jiaxu upper-margin note, on the long chapter at the Prison Temple, "lost, no manuscript"; 26.153, at the end of a Jiaxu chapter, on the writing about the archery ground, "lost, no manuscript"; while the Gengchen upper-margin note 26.120 says only "no manuscript." The notes signed with a dinghai date and a Jihu name (Jihu Sou or Hu Sou) are 20.20, 26.120, and the Gengchen version of 26.24; the Jiaxu 26.24 and 26.153 are unsigned. Cai Yijiang uses these notes to explain why what came after chapter 80 did not survive; Mei Jie says that Cao Xueqin had already "delivered the goods" to the small circle headed by Zhiyan, so that in this sense the book was finished, and its later part did not survive because it was lost in lending (both: reprint).
22.104 has two literal readings. The base text reads "from here on damaged and lost; what follows is a later supplement" (此后破失，系再补), meaning that what follows is supplied text; the common reading is "awaiting supplement" (俟再补), meaning that what follows was originally missing and awaits supplementing. Paper 1 filed the note as a record without a ruling; that is a way of handling it, not a third reading. In the base text the chapter continues after this note to its end, and at the end stands 22.106, "this chapter was not finished when Qin [Xueqin] died." These materials concern the unfinished, damaged and patched end of the chapter; this paper does not rule on who wrote the supplied text or on what it was based. This is one of the places outside the "main body" of §1. 75.3, "Mid-Autumn poem missing; awaiting Xueqin," is likewise the manuscript's own record of something unfinished.
Commentators and the brush (22.22 to 22.24, 1.63).
The Gengchen upper-margin note 22.22 speaks of "Xifeng choosing the plays and Zhiyan holding the brush"; 22.23 adds later, "now, in the summer of dinghai, only one decrepit thing is left"; the Jing upper-margin note 22.24 has the same text with one more sentence, "Qinxi, Zhiyan, Xingzhai: these gentlemen have all departed one after another."
Three points about this group.
- "Zhiyan holding the brush" is not to be translated as co-authorship. Whether holding the brush means writing, copying, or noting down one occasion on which plays were chosen: each interpretation has its proponents, and what this paper has not checked is marked as not checked.
- 22.23 and 22.24 are later notes following an earlier one. They can be used to discuss the order of the notes, but they are not two independent witnesses.
- The group provides material on who held the brush and on later notes added to earlier ones. "The earlier note" does not exclude the same commentator adding to his own old note; this group alone does not settle how many commentators there were.
The Jiaxu upper-margin note 1.63, "On the last night of the renwu year, the book not yet finished, Qin [Xueqin] wept his tears dry and died," signed at the end "written in tears, eighth month of the jiawu year," is again a record of date and identity left by the commentators' layer.
Drafts of the later part, and two notes at odds (18.17, 18.19, 20.20).
20.20 contains the statement "I saw it only once, at the time of a fair copy" (余只见有一次誊清时): the commentator says plainly that he saw a draft of the later part. 18.19 says that only "at the end, with Jinghuan's Roster of Feeling, will one know" the names of the registers; it points to the design of the later part, without saying explicitly that the commentator had seen it himself. The material on the number of chapters has been tabulated in the overview, §3.5, and is not repeated here.
18.17 and 6.48 are an instance of what §1 calls two notes at odds, registered only as a possibility. The Gengchen double-line interlinear note 18.17 conjecturally assigns Ping'er to the Tertiary Register ("the rest are not much touched on; presumably they are … Ping'er and others, without doubt"), while the Jiaxu double-line interlinear note 6.48 says of Ping'er, "presumably she too is in the Secondary Register." Paper 1 takes the Secondary Register layer of 6.48: Rec. ch. 110 has been supplemented according to Addendum 003, so that Ping'er's page falls within the Secondary Register. In paper 1's words, it "records only which side was taken, and does not judge the other side wrong."
Division into chapters and collation (17.1, 75.2).
17.1, "this chapter would properly be divided into two," corresponds to the form of the Jimao and Gengchen manuscripts, in which chapters 17 and 18 are not divided, and is a record that the division into chapters was still being worked out. 75.2, "collated on the seventh day of the fifth month of the twenty-first year of Qianlong," is a record of the collation of a manuscript.
The Western Hall (28.82, 28.83).
The Gengchen upper-margin note 28.82: "Drinking from a great bowl: the day the Western Hall brought forth nine-stalk lingzhi. Commenting on the book up to here, how can I not grieve? The Double Ninth of renwu." The Jiaxu side note 28.83: "Who has lived through this? Alas, alas! The old affairs of the Western Hall." The "Western Hall" (西堂) is usually read as Cao Yin's study; this paper has not yet found who first proposed this, or where. Among the popular theories, some take the "Western Hall" to be You Tong, and some, on the strength of a note in a disputed manuscript, propose a "late elder brother of the Western Hall" (西堂故兄); the latter involves the Guiyou manuscript and is left to paper 3.
Two general rules.
- The entries of paper 1 given above are the project's records of how it handles each note, not a historical authentication of the materials. They are set side by side here so that the reader can see how the same note is used in each of the two papers; the audit's labels are not to be taken as supporting evidence for the history of composition.
- Several notes in the table are found only in the Jing manuscript (13.8; the "Tangcun" signature of 13.21 and the Jing upper-margin note of 13.21; 13.52; 22.24; for a shorter text with the same purport as the Jing note on 13.21, see the Gengchen 13.90), and in 13.3 missing characters are supplied from the Jing. The authenticity of the Jing manuscript is disputed: Gao Shuwei's A Refutation of the Jing Manuscript of Honglou meng (Zhonghua shuju, 2024) holds that it is a forgery; Mei Jie's "Another Look at the Jing Manuscript" (Honglou meng xuekan 2002, no. 1) holds that it is genuine. Following §1, this paper takes no part in the dispute and registers these notes only by provenance. Paper 1 gives the Jing manuscript reference weight (参级), which is "the weight the project gives to each manuscript's notes; it is not a judgement of the truth of any note or of the worth of any manuscript."
3.3 Inconsistencies within the text
Within the first eighty chapters there is a set of places where earlier and later passages do not agree, which scholars use to argue for traces of revision or of merged drafts. In 1979 Dai Bufan listed four kinds of internal evidence: Beijing speech mixed with the Wu dialect; the Jia mansion moved from south to north, and its garden from west to east; disorder in the story's chronology; and two Baoyus of different ages (from a summary in a history of scholarship; second-hand). Later discussions of layering cite these kinds again and again. For each kind this paper gives representative examples, first checked against the base text, then with their sources and the counter-arguments.
| Kind | Representative example | Who uses it | Counter-argument or other reading |
|---|---|---|---|
| Age | The two Baoyus; in ch. 25, "since we parted at Greensickness Peak, in the blink of an eye thirteen years have passed," which does not agree with Baoyu's age elsewhere | Dai Bufan (another's old draft); Shen Zhijun, "'The New Baoyu' and 'the Old Baoyu'"; Eileen Chang on Baoyu's age being lowered from revision to revision (the author's revision) | Zhang Jinchi, Cai Yijiang and others refuting Dai Bufan's reading in the authorship debate (each article to be checked) |
| Chronology | "Time running backwards" | Dai Bufan | Liu Mengxi: the festivals and the weather run consistently through the chapters |
| Duplicated persons | Caiyun and Caixia; Jia Cong and Jia Lan in the clan list of ch. 13 | Liu Shide on "two components, the first draft and the revised draft" | Copyists' errors, incomplete renaming, two persons existing side by side |
| Genealogy | The generation of Grandmother Jia relative to the Duke of Rongguo | Zhu Danwen (cited by Hu Lianhao) | No specific refutation found |
| Place names and titles | Jinling and Chang'an; the Prince of Dong'an and the Prince of Dongping | Dai Bufan (south moved to north); Hu Lianhao (2024; Jinling changed to Chang'an) | Readings in which the geography is blurred on purpose; the configuration of princely titles awaits comparison |
| Dialect | Beijing speech mixed with the Wu dialect | Dai Bufan | Chen Xizhong, Hou Zhongyi; Wu Shih-ch'ang, Zhang Jinchi |
A few of these examples need more comment.
Caiyun and Caixia. The maid who is on good terms with Jia Huan is Caixia in chapter 25 of the base text ("only Caixia still got on with him"), Caiyun in chapters 30, 60 to 62 and 70 (ch. 70: "because Caiyun had lately fallen out with Jia Huan"), and Caixia again in chapter 72 ("though in her heart she had an old attachment to Jia Huan"). These names have been checked against the base text. Why Caiyun and Caixia should be taken as earlier and later names of one character, rather than as two persons existing side by side, depends on each scholar's particular argument; this paper has not obtained that argument, and lists the case provisionally as a textual phenomenon. Whether variant readings in the manuscripts have regularized some of the instances remains to be checked manuscript by manuscript.
The clan list in chapter 13. After Qin Keqing's death, "Jia Dairu led Jia Chi, Jia Xiao, Jia Dun, Jia She, Jia Zheng, Jia Cong … Jia Lan, Jia Jun, Jia Zhi and others, who all came." The list includes even Jia She and Jia Zheng, and Jia Cong and Jia Lan are in it; the plainest reading is that these are the same characters appearing. Liu Shide holds that in the first draft Jia Cong was just an ordinary clansman, and became Jia Lian's younger brother only in the revised draft (second-hand). This paper has not yet read, in the original studies, why proponents take these to be two different persons; the case is recorded provisionally as "a doubt raised by a particular theory," and the list itself is not taken as established evidence of layering.
Dong'an and Dongping. In chapter 3, the inscription under the ebony couplet in the Hall of Glorious Felicity reads "respectfully written by your fellow townsman and junior in family friendship, Mu Shi, hereditary Prince of Dong'an" (同乡世教弟勋袭东安郡王穆莳拜手书). "Dong'an" appears only here in the whole book; among the four princes of chapters 11 and 14 it is the Prince of Dongping, and Mu Shi's name is not given. That the two titles coexist is registered here only as a configuration of princely titles awaiting comparison; whether the manuscripts agree remains to be collated manuscript by manuscript, and the paper does not assume in advance that one man's title was changed. The project's backstage (its internal working notes) once read this as a "trace of renaming" left by the Cao Xueqin layer; this paper has not found who in scholarship first raised the point (§6, Appendix 3).
Thirteen years. In chapter 25 the monk says, "since we parted at Greensickness Peak, in the blink of an eye thirteen years have passed," which scholars take to disagree with Baoyu's age elsewhere. Eileen Chang's account of Baoyu's age being lowered from revision to revision is the most often cited explanation of such inconsistencies (reprint). This paper has checked only the chapter 25 sentence; why it counts as a disagreement depends on the chronological links each scholar adopts, and it is listed provisionally as a textual phenomenon.
Of the above, those checked against the base text are Caiyun and Caixia, the chapter 13 list, the "thirteen years" of chapter 25, and the examples from chapter 3 and chapters 11 and 14. The others will be checked against the base text case by case before the next version. Where the earliest proponent cannot be found, the paper says "not found"; finding the earliest proponent is not a condition for an inconsistency to hold.
3.4 The form of the manuscripts, and the dates recorded in them
On the form of the manuscripts, this paper goes only as far as kinds, and does not examine each case:
- Division into chapters: in the Jimao and Gengchen manuscripts chapters 17 and 18 are not divided; note 17.1 says, "this chapter would properly be divided into two."
- Missing and supplied chapters: the Gengchen lacks chapters 64 and 67, which later manuscripts supply (as commonly recorded). Scholars differ on the presence and source of these two chapters in each manuscript; the paper does not judge.
- Unfinished: the end of chapter 22 (22.104, 22.106).
- Variant chapter titles: the chapter titles differ among the manuscripts; to be listed manuscript by manuscript.
- Dates in the notes: the winter of jimao, renwu, yiyou, dinghai, jiawu. Yiyou appears only once, in the Gengchen upper-margin note 25.61: "Winter of yiyou, by the snowy window. Jihu Laoren."
- Record of collation: 75.2, "collated on the seventh day of the fifth month of the twenty-first year of Qianlong," a bingzi year (1756).
Dates must be distinguished by what they date. "Renwu," "dinghai," "copied and commented on again in the jiaxu year," and the collation of the twenty-first year of Qianlong are first of all dates and acts recorded in the materials. To read them as the years in which the surviving manuscripts were copied, or in which some original draft was finalized, needs a further layer of evidence. This paper does not date the manuscripts here.
3.5 Records outside the book
Several records outside the book bear on its making. For each, the paper states from whose writing it actually quotes.
Mingyi, "Twenty Poems on Dream of the Red Chamber" (题红楼梦), with a short preface. The preface begins: "Master Cao Xueqin brought out a copy of the Dream of the Red Chamber he had written, recording in full the splendour of romance and luxury …" (as quoted in Eileen Chang, Nightmare in the Red Chamber, in the essay "Er xiang Honglou meng"; online text at Nunu Shufang, checked; to be checked again against Honglou meng juan, compiled by Yi Su). The copy Mingyi saw does not fully agree with the present text. On this basis Wu Shih-ch'ang proposed his "first draft" theory: the first draft of Cao Xueqin's Dream of the Red Chamber that Mingyi saw had "the story entirely within the Prospect Garden," and Wu concludes that "Xueqin was the author of Dream of the Red Chamber, and also the one who worked it over after the materials recorded by the stone had been combined with Dream of the Red Chamber" ("On the First Draft of Honglou meng Seen by Mingyi," Honglou meng xuekan 1980, no. 1; reprint).
Yurui, in Zaochuang xianbi (枣窗闲笔), records: "There was an old book, A Mirror for the Romantic, also called The Story of the Stone, by whose hand no one knows. Cao Xueqin obtained it, and because what it recounted was broadly the same as the history of his own family, he took it as a vehicle for his own purposes, and deleted and altered it as many as five times." He also records that the manuscripts he saw "every one had the notes of his uncle Zhiyan." This paper follows the quotation by Zhou Jianyu from the 1957 facsimile published by Wenxue guji kanxingshe (pp. 9 and 21); a second-hand quotation elsewhere reads "his uncle Zhiyan Zhai"; the two versions stand side by side, to be checked against the facsimile. This is the earliest documentary basis for the theory that the book was another's old draft, reworked by Cao. Liu Mengxi holds that Yurui's records rest mostly on hearsay and are not necessarily reliable (second-hand). The authenticity of Zaochuang xianbi itself is the subject of separate discussion: Gao Shuwei's book of 2024, for example, has a chapter on the authenticity of Yurui's Zaochuang xianbi; this paper has seen only the book's table of contents, and its conclusion is to be checked against the book.
The Chengjia edition (1791), preface by Cheng Weiyuan: "Accounts of the author differ, and in the end no one knows from whose hand it came; only, the book itself records that Mr. Cao Xueqin deleted and altered it several times." (Common transcription; this paper has not seen a transcription it can name, and the page of the Chengjia edition itself is to be checked.) The preface also describes where the last forty chapters came from: "Over several years I had gathered only some twenty-odd juan. One day, by chance, I obtained ten-odd juan from a street pedlar's load (鼓担)." (As quoted in Hu Shi, "A Textual Study of Honglou meng," reprint at Aisixiang; checked.) The foreword to the Chengyi edition (1792): this paper has not seen the page; to be checked. These sentences of Cheng's preface are the Cheng lineage's record of the "reading over and revising," corresponding to the chain of titles.
Duncheng, in "Sent to Cao Xueqin, Thinking of Him" (寄怀曹雪芹), writes "better to write books in the Village of Yellow Leaves" (不如著书黄叶村), a line often cited as evidence that Cao Xueqin wrote a book, though it names none; in "Lament for Cao Xueqin" (挽曹雪芹), the phrase "the 'ox-demon' writings left behind" (牛鬼遗文) has been read by some as evidence of drafts left behind (source to be checked; not yet seen by this paper). In the poems of Dunmin and Zhang Yiquan, this paper has found no line that directly mentions the writing of a book. The pages in the original collections are all to be checked.
The above reports only the records and how scholars have read them. Where the provenance of a historical source is disputed, the dispute is noted, and no ruling is made on authenticity.
4 Scholarly Theories of Composition
This paper presents the scholarly theories of composition in three lines, and lists separately the dispute over whether the Zhiyan Zhai manuscripts or the Cheng editions came first. Scholars are grouped below by the main claims of the works cited. The whole of a scholar's work is not assigned to a single camp, and different stages or different questions in one person's work may be discussed in different places. Each scholar's claims, evidence, representative works, opponents and reliability are given in Appendix 1; the main text gives only the gist. The anchor for the history of scholarship is Du Zhijun's "Thirty Years of Redology in the New Period, 1978–2008" (Honglou meng xuekan 2009, no. 2).
4.1 One book, revised many times
This line holds that the book was produced by revising one and the same draft many times, and that traces of the old drafts are layered into the present text.
Hu Shi is the baseline. The first version of his "A Textual Study of Honglou meng" was published in May 1921 as the preface to the Yadong Library's punctuated edition of the novel, and was revised in November of the same year. He wrote that "the author of Honglou meng is Cao Xueqin" and that "the forty chapters after chapter 80 were supplied by Gao E" (reprint). He posits no layer of old drafts: the whole book has only two layers, Cao's eighty chapters and Gao's supplement of forty.
Zhou Ruchang, Feng Qiyong and Cai Yijiang, building on this, accept that "revised it five times" refers to Cao Xueqin's own repeated revisions, and that A Mirror for the Romantic was an earlier work of his. Feng Qiyong, comparing the format of the copying, the chapter titles and the commentary of the Zhiyan Zhai manuscripts, argues that the Gengchen is "the most complete of the early manuscripts of The Story of the Stone made in Cao Xueqin's lifetime" (reprint). Cai Yijiang says that "the possibility is very great" that A Mirror for the Romantic "was the basis on which The Story of the Stone was later worked over and revised"; he opposes the theory of two books combined, on the ground that "a novel is not a theoretical essay; with their different characters, [two novels] are very hard, or one should say impossible, to combine" (reprint).
Eileen Chang reconstructed the layers of rewriting through close reading of the variants. She says that Cao Xueqin "was all along the one in charge of rewriting this book," that the first eighty chapters were rewritten over more than twenty years, that early versions from different periods had already circulated, and that the variants are the strata left by the rewriting; Baoyu's age was lowered from one revision to the next (reprint). She also held that The Story of the Stone and A Mirror for the Romantic were originally two books, and that the Land of Illusion entered the book only after A Mirror for the Romantic was taken in. She thus straddles the first two lines; she is placed here, under one book revised many times, according to where her main effort lay.
Those who divide stages by the titles. Chen Qinghao, Shen Zhijun and Li Qingyu read the chain of titles as stages of drafting or of the conception of fiction (§3.1). Liu Shide, from collating the manuscripts, discerns "two components, the first draft and the revised draft," and takes inconsistencies such as the two Jia Lans, the two Jia Congs, and Caiyun and Caixia as traces of two drafts lying one on the other (§3.3; second-hand).
Xia Wei takes A Mirror for the Romantic to be Cao Xueqin's first draft, which the author "sublimated and renewed on the basis of the first draft, A Mirror for the Romantic," and from the opposite journeys of the two Baoyus, Zhen and Jia, reads a dualist political outlook of taking office and retiring from the world ("From A Mirror for the Romantic to Dream of the Red Chamber: The Evolution of the Book's Composition and Creative Thought," Wenxue pinglun 2022, no. 6; reprint).
Zhu Danwen is placed in this line provisionally. Xia Wei's article opens by quoting p. 197 of Zhu Danwen's Honglou meng lunyuan (Jiangsu guji chubanshe, 1992): the various titles "name drafts of one and the same novel at different stages of composition"; there were not five drafts before jiaxu, and "the first two revisions were made on the first draft, because the drafts we can at present be sure really existed are only two" (Xia Wei's article read in a reprint; her quotation checked). This is a statement of the one-book view. Both sentences are thus known only second-hand, through Xia Wei; pp. 197 and 200 of the book itself have not been checked, and whether the book argues elsewhere for two books combined is to be checked. She also argues that the genealogy in the old draft differed from the present text, the original draft having made Grandmother Jia the wife of the Duke of Rongguo (cited by Hu Lianhao).
Counter-arguments. The line of two books combined holds that the one-book view cannot explain the difference in the writing of the Ning and Rong mansions (§4.2). Du Zhijun judges both the one-draft-many-revisions view and the two-books-merged view to be "still immature in some respects, while objective, systematic, detailed and in-depth micro-level demonstration is still lacking" (reprint).
4.2 Two books (two sources) combined
This line holds that two drafts, such as A Mirror for the Romantic and The Story of the Stone, were combined into one book by Cao Xueqin.
Zhou Shaoliang proposed early on that two drafts were joined. He holds that A Mirror for the Romantic was an old draft of Cao Xueqin's that really existed, and that the present text joins two first drafts, the original Story of the Stone and A Mirror for the Romantic. From the latter come a series of "romantic" stories, such as Xifeng and Jia Rui, Qin Keqing and the Tianxiang Tower, Jia Lian with You Erjie and with the Duo girl, Qin Zhong, and Xue Pan, whose theme turns on "a warning against giving way rashly to romantic feeling" ("'Xueqin formerly had a book called A Mirror for the Romantic'," Honglou meng xuekan 1979, no. 1; reprint).
Wu Shih-ch'ang has two theories. His English monograph of 1961 proposed the theory of "Tangcun's short prefaces" (§3.2). An article of 1980 proposed two sources combined. One source is the first draft of Cao Xueqin's Dream of the Red Chamber that Mingyi saw; the other is "the story of the Consort's visit home, what the stone recorded: The Story of the Stone, which Zhiyan commented on. The commentator of The Story of the Stone was also the author of The Story of the Stone. That author was not Cao Xueqin." Cao Xueqin was the one who combined the two and worked them over (reprint). In Honglou meng tanyuan waibian, p. 483, he has the sentence "it is not by Xueqin" (cited by Zhou Jianyu); what "it" refers to must be checked in the book.
Du Chungeng is the clearest proponent of two books combined. He holds that the present text was combined from two original drafts: one centred on the Rongguo mansion, which was The Story of the Stone, and one centred on the Ningguo mansion and Jia She's branch, which was A Mirror for the Romantic. Cao Xueqin made a "second creation," merging the two books and altering the relations between characters, "making the Dukes of Rong and Ning into full brothers." The evidence adduced includes the different standards for choosing wives in the two mansions; the difference in scale between Jia Jing's and Jia Zheng's birthdays; the scale of Qin Keqing's funeral; the wording "the Rongguo branch" and "the two mansions of Ning and Rong" in Leng Zixing's account of the family; and the phrase "the name that sums up the whole" in the general principles, together with 1.61 ("The Rong and Ning Mansions, Two Books," Honglou meng xuekan 1998, no. 3, reprint; also "On Cao Xueqin's 'Read It Over for Ten Years and Revised It Five Times'," same journal, 2007, no. 3, catalogue, second-hand).
Tan Dejing sorts earlier research into three kinds: many revisions, combination, and cut-and-paste. He holds that Dream of the Red Chamber is a selective integration of A Mirror for the Romantic, its theme shifting from a warning against lust to the expression of women's beauty and of a sense of sorrow (reprint).
Counter-arguments. Cai Yijiang holds that two novels with different characters "cannot be combined." He also criticized the theory of two books combined as "suspecting the neighbour of stealing the axe" (a proverb, from the story of the man who, having lost his axe, saw a thief in everything his neighbour's son did); the phrase is known from Du Chungeng's account, and Du replied that the same logic applies equally to the view of one draft revised many times (the original source of Cai's phrase has not been found).
4.3 Another's old draft, reworked by Cao
This line holds that the original author was someone else, and that Cao Xueqin was the one who revised and reworked the book.
Yurui's record is the earliest documentary basis for this line (§3.5).
Dai Bufan published "Unveiling the Mystery of the Author of Honglou meng" in Beifang luncong 1979, no. 1, subtitled "On how Cao Xueqin skilfully recast Brother Stone's old draft, A Mirror for the Romantic, and reworked it into the book" (second-hand); it was followed by "Brother Stone and Cao Xueqin" (same journal, no. 3) and "On Qin Keqing's Late Death" (first publication to be checked). He holds that the "Brother Stone" of the book was another person, whose old draft was A Mirror for the Romantic, and that Cao Xueqin was not the author in the sense of originator, only a reviser. His internal evidence is given in §3.3. On dialect he says: "its old draft was written by someone who could hardly shake off a Wu accent … while the revised (new) draft is the work of someone thoroughly versed in the Beijing dialect" (reprint).
Dai's theory drew nearly thirty rebuttals, and the articles of the debate were later collected as Collected Papers of the Debate on the Authorship of Honglou meng (compiled by the editorial office of Beifang luncong; Taiyuan: Shanxi renmin chubanshe, 1985). The rebuttals had three main points.
- Chen Xizhong and Hou Zhongyi pointed out that Dai "confused dialect with dialect vocabulary" (as reported by Zhan Dan); Wu Shih-ch'ang and Zhang Jinchi also held that the method of using dialect to prove more than one author was unscientific.
- Liu Mengxi wrote "Qin Keqing's Death and Cao Xueqin's Authorship: After Reading Comrade Dai Bufan's 'On Qin Keqing's Late Death'" (Wenyi yanjiu 1979, no. 4; catalogue, second-hand). He also argued that the festivals and the weather are "on the whole consistent throughout, following on from chapter to chapter," so that no conclusion of time running backwards can be drawn (reprint).
- Zhang Bibo and Zou Jinxian rebutted Dai's inference that Brother Stone was Cao Zhucun.
The titles and issue numbers of the individual articles follow the table of contents of the Collected Papers, and remain to be checked one by one.
Pan Chonggui held that Dream of the Red Chamber "had another, anonymous original author; Cao Xueqin was only the one who held the brush in revising, supplementing and correcting it" (cited by Zhou Jianyu). His view comes from the nationalist strand of allegorical decoding (suoyin); this paper takes only his claim about composition.
Pi Shumin notes of his own work: "on the inference that The Story of the Stone had two authors, for details see my Honglou meng kaolun ji, pp. 37–39, 111–115" (reprint; Honglou meng kaolun ji, Taipei: Linking, 1984). His view is commonly summarized as: Zhiyan Zhai was Li Ding, and Li Ding took part in the early drafts of The Story of the Stone (second-hand). Whether he claims that Li Ding had a complete old draft and that Cao Xueqin was only a reviser, this paper cannot say, not having seen the book, and does not say so.
4.4 The dispute over whether the Zhiyan Zhai manuscripts or the Cheng editions came first
Ouyang Jian argues that the Cheng editions came first and the Zhiyan Zhai manuscripts later: the Zhiyan Zhai manuscripts are only later copies, "Zhiyan Zhai" is "the collective name of a succession of forgers," and the Chengjia edition is the true text (second-hand). His series of articles began with "Doubts about the Theory of 'Two Great Textual Systems' of Honglou meng, with a Discussion of Zhiyan Zhai as a Forgery by Liu Quanfu" (Fudan xuebao 1991, no. 5); his monographs include Honglou xinbian (Huacheng chubanshe, 1994), Hongxue bianwei lun (Guizhou renmin chubanshe, 1996) and Huanyuan Zhiyan Zhai (Heilongjiang jiaoyu chubanshe, 2003).
Rebuttals came from Ying Bicheng, "On the Textual Systems of Honglou meng: A Discussion with Comrade Ouyang Jian" (Fudan xuebao 1992, no. 1), and from Chen Zhao, Cai Yijiang, Song Mouyang, Yang Guanghan, Feng Qiyong, Xu Naiwei and others (according to the survey by Lin Yi and Cai Zhengxing; second-hand). Du Zhijun's assessment is that this "exposure of forgery" "appears at present … hard to sustain," but that "the very rise of the doubts shows, from the other side, … the complexity of the relation between the Zhiyan Zhai and Cheng texts" (reprint).
This dispute does not lie on the same axis as the three lines above. The three lines dispute how the book was written; this dispute concerns the order of the materials. What it challenges is that part of the argument on composition which presupposes that the Zhiyan Zhai manuscripts and their commentary are earlier than the Cheng editions. The three lines also cite other materials, such as the text of the chain of titles in the Cheng lineage, Cheng Weiyuan's preface, and the inconsistencies in the text, and so are not wholly affected by it. This paper records the theory and its rebuttals as they stand, and takes no part in the dispute over authenticity.
4.5 Scholarship in English
On studies of composition in English this paper adds only a sentence. Wu Shih-ch'ang's monograph of 1961, On The Red Chamber Dream: A Critical Study of Two Annotated Manuscripts of the XVIIIth Century (Oxford: Clarendon Press), is an early monograph dealing directly with composition; it was reviewed in the same year by P. D. Hanan in BSOAS and by C. T. Hsia in JAS. David Hawkes's introduction to the first volume of the Penguin The Story of the Stone (1973) discusses A Mirror for the Romantic and the rewriting; page numbers to be checked. Andrew H. Plaks's work is mainly on structure and allegory, and is not assigned to any line.
4.6 Where the three lines agree
The three lines dispute whether there was one book or two, and whether it was one and the same author throughout or an old draft by someone else. This paper does not decide between them.
Among the three kinds of claim set out in this paper that explicitly discuss the part played by Cao Xueqin's rewriting, the common ground is that they acknowledge his role in writing or reworking the first eighty chapters of the present text: the one-book view says he revised his own draft; the two-book view says he combined two drafts; the view of another's old draft says he reworked someone else's. Hu Shi is the baseline and posits no layer of old drafts, so he is not counted here. This summarizes only the intersection of the claims described, and does not use it to prove any one particular process of composition.
The one layer this paper depends on has three segments. The first two are that the main body of the text of the first eighty chapters went through layering, and that the one who integrated it into a book was the Cao Xueqin named in the text. They correspond to this common ground, and can also point to the text's own account in the chain of titles ("compiled a table of contents and divided it into chapters"); the project takes them as a working understanding, and only in the sense of a role in the text. The third segment, "fixed its centre," is the project's working judgement, which this paper does not argue, and it is not what the three lines agree on (§1, §7). The dispute over whether the Zhiyan Zhai manuscripts or the Cheng editions came first challenges the order of the materials; the one layer this paper depends on also has something to go on in the Cheng lineage (§7).
5 Popular Theories
5.1 Coverage
| Theory | Proponent and source | Relation to the layer this paper depends on | Use in this project (see §6) |
|---|---|---|---|
| Xu Xinggong | Liao Yihu: the column "Xu Xinggong and Honglou meng" on Douban Read; a Bilibili column (2022); "Hu shuo Honglou meng" on Zhihu, and others | Takes "Cao Xueqin" to be a pen name of Xu Xinggong himself | The Xiangyun item (the Dongxiyang kao in Rec. ch. 110); the other prototypes had no influence on the writing of the current draft (author's statement, §6.1); see also Appendix 3 |
| Fang Yizhi | Chen Xianfu, as reported by China News (chinanews.com), 2011-10-31; Chen's book not seen | Takes Cao Xueqin to be Fang Yizhi's invention | One stage of the six-stage chart in the backstage; withdrawn |
| Hong Sheng | Tumore, Tumore hongxue (Jilin renmin chubanshe, 2005), and others | Takes Cao Xueqin as the reviser; compatible | To be reconciled |
| Wu Meicun | Fu Bo and Zhong Changshan; Wu Chengmou; and others | Takes Cao Xueqin (or Yan Shengsun under that name) as the reviser | To be reconciled |
| Mao Pijiang | Mao Lianquan, Xuan'an bu xuan (Huaxia chubanshe, 2016); Gu Hao; Huang Weimin | Takes "Cao Xueqin" to be a pen name of Mao Pijiang | To be reconciled |
| Cao Fu | Chen Lin, Poyi Honglou shijian mima (Jiangsu meishu chubanshe, 2006) | Takes Cao Xueqin to be Cao Fu's invention; also holds the Zhiyan Zhai manuscripts to be forgeries | To be reconciled |
| Zhu Xiangyu | Huo Guoling and others, Honglou jiemeng | Takes Cao Xueqin as the author, and posits a hidden history written beneath the story | To be reconciled |
| The Ming royal house and other figures | The Zhu Cizhao theory (Sina blog "Honglou yinshi die lian hua"), and others; other proponents to be checked | Takes someone else as the author, the dispute being over who; argued mostly from homophones | To be reconciled |
| Qin studies | Liu Xinwu, Hualiang chun jin luo xiangchen (2003), and others | Takes Cao Xueqin as the author, with deletions and alterations made to avoid disaster; compatible | Qin Keqing's origins agree with the reconstruction; the project did not rely on Liu's work, and Liu's view came first (2003); similarities and differences in paper 8 |
| Identifications of the title-givers and commentators | See Appendix 2 | The table of identifications only, not discussed item by item | Kong Meixi = Kong Shangren: The Peach Blossom Fan in Rec. ch. 109 |
| General assessments | Yan Zhong; Peng Guangqian; Ouyang Jian | — | — |
Sources are graded as found. Publications that can be traced are cited; readable writings by the proponent are cited; where only a report or a summary is available, the entry says "as reported" or "title or summary only"; where no source was found, that is stated. The column "use in this project" is to be filled in after reconciliation with the table of twenty-four influences in Appendix 4; for now only the known items are entered, and the rest are marked "to be reconciled."
5.2 The theories in brief
Each theory is given with its claim, its reasons and what remains unproven. The tone is neutral, and proponents are referred to only by the names under which they publish.
Xu Xinggong (Liao Yihu). The claim is that the author of The Story of the Stone was the late-Ming Fuzhou book collector Xu Bo (courtesy name Xinggong, 1570–1642). "Cao Xueqin" is a pen name whose three characters stand for Confucian, Buddhist and Daoist: Cao for Cao Xuequan, Xue for Xuefeng Temple in Fuzhou, Qin for Mount Qin in Fuzhou. The prototypes of the characters are mostly late-Ming literati of Fujian, such as Xie Zhaozhe for Baochai, Ye Xianggao for Yuanchun, Shao Jiechun for Tanchun, Zhang Xie for Xiangyun, and Lin Chong for Daiyu. There are three reasons: Xu Xinggong "died before the last night of the renwu year," which fits the "last night of renwu" of 1.63; the author must meet several "hard criteria," namely book collecting, fame as a poet, social circle, standing as a man of letters, and experience with fiction; and item-by-item identifications of names, sobriquets and lines of verse. Unproven: this paper has found no outside document before Liao that attaches The Story of the Stone to Xu Xinggong; Liao also attributes Jin Ping Mei, Qingshi and other works to Xu Xinggong; and on the last forty chapters he has two views, supplementation by Xu Ling and composition "advancing side by side." "Kong Meixi, Wu Yufeng" are in his account a split of the characters 东吴 ("Eastern Wu") indicating that the author was from Fujian; they are not real title-givers. (From the sample chapters on Douban Read and the Bilibili column; of the Zhihu pieces only the titles have been seen.)
Fang Yizhi (Chen Xianfu). According to a China News report of 31 October 2011, Chen Xianfu claims that Fang Yizhi was the author, "and even Cao Xueqin was invented by him," the purpose of the book being "to mourn the Ming and oppose the Qing." His reasons: the three author-clues of chapter 1, the stone, the Passionate Monk and Cao Xueqin, relate to Fang Yizhi's alias "Wu Shigong" (吴石公) and to his taking the tonsure in Wuzhou; "Xueqin" is a homophone of xueqin 学勤, "diligent in study"; "red" hints at the Zhu (朱, "vermilion") house of the Ming, and "moon" (月) hints at "Ming" (明). The report says that his research had not been formally published and that he planned a book; this paper has so far found no such book. It has been pointed out that Fang Yizhi died in 1671, which is rather early (as reported by Ouyang Jian). A theory of collaboration, with Fang Yizhi writing and Mao Pijiang revising, also circulates; its original proponent is to be checked.
Hong Sheng (Tumore). The claim is that Hong Sheng wrote the first version, first called A Mirror for the Romantic among other titles and later changed to The Story of the Stone, and that Cao Xueqin was the one who read it over and revised it. The prototypes of the "four great families" are the four Hangzhou families of Hong, Gu, Huang and Qian; the women of the Banana Garden Poetry Society are the prototypes of the Twelve Beauties; Xixi is the prototype of the Prospect Garden; and Zhiyan Zhai is a pseudonym of Hong Sheng's wife Huang Hui (as summarized in Pang Guanqing's account of Tumore's "ten theses"). Unproven: this paper has found no direct record of Hong Sheng writing a novel.
Wu Meicun. Fu Bo and Zhong Changshan claim that Wu Meicun was the original author, and that Yan Shengsun revised the book under the name Cao Xueqin. Their reasons are that "Wu Yufeng," "Kong Meixi" and "Tangcun" together conceal "Wu Mei(cun)," and that the renwu of 1.63 should be the renwu of the Kangxi reign (1702), the very year of Yan Shengsun's death (from a reprint; original publication to be checked). Wu Chengmou claims that Wu Meicun wrote the original, with Cao Xueqin, Cheng Weiyuan and Gao E as its editors, and calls this "one scholar's view." Unproven: the splitting and joining of characters and the rereading of dates are both interpretations; this paper has found no outside evidence. A note in a manuscript on which this theory relies involves the Guiyou manuscript and is left to paper 3.
Mao Pijiang. Mao Lianquan wrote An Unsolved Case Solved: Seventy-Three Proofs that Mao Pijiang Wrote Dream of the Red Chamber; Gu Hao holds that the book was completed in the jiaxu year of Kangxi (1694), for reasons including the correspondence of the Shuihui Garden with the Prospect Garden, the Rugao dialect, and the theme of mourning the Ming; Huang Weimin takes Dong Xiaowan as the prototype of Lin Daiyu. Against this, Yan Zhong points out that chapter 54 mentions Sequel to the Lute (续琵琶), which Cao Yin wrote in 1703, whereas Mao Pijiang died in 1693.
Cao Fu (Chen Lin). The claim is that the author of the 120-chapter Story of the Stone was Cao Fu, that Cao Xueqin is a character invented by Cao Fu, and that the Jimao, Gengchen and other Zhiyan Zhai manuscripts are later forgeries (as reported). The dispute involves the authenticity of the Zhiyan Zhai manuscripts; this paper records it and does not judge.
Zhu Xiangyu (Huo Guoling and others). The claim is that the book has a "double nature," with a history hidden behind the novel, and that the prototype of Lin Daiyu was Zhu Xiangyu; the eighty-chapter text with the Zhiyan Zhai commentary is Cao's complete work, and the core of the commentary is from Cao Xueqin's own hand. This theory denies that the book is unfinished, and so has no need of a layer of continuation.
The Ming royal house and other figures. These theories take the author to be Zhu Cizhao, a son of the Chongzhen emperor ("Honglou yinshi die lian hua"), or Zhu Cilang, the Chongzhen emperor himself, Yinreng, Zheng Keshuang and others (listed by Ouyang Jian in 2015; the proponents of each to be checked one by one). "The Third Prince Zhu" and the like are the persons identified, not proponents. Theories of this kind rest mainly on homophones, as when Jia Wang Xue Shi 贾王薛史, the four great families, is read as jia wang xue shi 家亡血史, "a family ruined, a history of blood."
Qin studies (Liu Xinwu). Liu Xinwu holds that the prototype of Qin Keqing was a daughter of Yinreng, the deposed crown prince of the Kangxi emperor, and that Cao Xueqin "had to delete and alter chapters already written in order to escape the 'literary inquisition'" (reprint). The reconstruction makes Qin Keqing come from the line of the deposed crown prince, in agreement with this view. The project did not rely on Liu's work; Liu's view came first (2003). Similarities and differences are left to paper 8; they are noted here in advance.
General assessments. Yan Zhong said in 2015 that 63 "authors" had been proposed (China News, 2015-11-30; this was his count for that year). Peng Guangqian says that "there is only one true answer to the riddle; the remaining sixty-odd are beyond doubt all false," and argues that Dunmin, Duncheng, Zhang Yiquan, Mingyi, Yurui, Yongzhong and the Zhiyan Zhai commentary should be the standard (The Paper, 2022-03-01). Ouyang Jian argues rather that "divergence is to be valued" and that one should "not rush to conclusions" (lecture of 2015, published by Aisixiang).
5.3 How they bear on the layer this paper depends on
Following §1, "Cao Xueqin" is taken as a role in the text. Seen in this way, the popular theories fall broadly into three kinds.
- Those that take Cao Xueqin as the reviser, such as the Hong Sheng and Wu Meicun theories, are compatible with the layer this paper depends on.
- Those that take "Cao Xueqin" to be a pen name or an invention, such as the Xu Xinggong, Fang Yizhi, Mao Pijiang and Cao Fu theories. They diverge from the common view over "who this person was," a dispute in which this paper takes no part; they do not directly conflict with the layer this paper depends on, and can simply be reported.
- Those that take the Zhiyan Zhai manuscripts to be forgeries, such as Chen Lin's, dispute the authenticity of the materials; as in §4.4, they are recorded and not judged.
6 How the Project Has Used These Conjectures
This section is a disclosure. In its backstage notes (the project's general account, the backstage volume and the volume on political background), the project put together a working chart of layers, used it to think through a good many things, and part of this entered the reconstruction. Following the author's ruling, the section gives the structure of the working chart, the source of each layer and the result of checking; every item that influenced the plot or wording of the reconstruction is set out one by one; inferences that have been withdrawn are reported as withdrawn. Prototype identifications are listed only in a table of sources in an appendix, without argument, and not as evidence.
6.1 The backstage working chart of layers
The backstage working chart has six stages: Xu Xinggong's Story of the Stone; Fang Yizhi's Record of the Passionate Monk; Kong Shangren's A Mirror for the Romantic and the "Wu Yufeng group's" Dream of the Red Chamber; the three-part Mirror for the Romantic of the "late elder brother of the Western Hall"; and finally Cao Xueqin. There is also a theory of a relay of signatures across three generations of Zhiyan Zhai, and more than forty prototype identifications drawn from the literary circle of Fuzhou.
To write this section, the paper traced the chart to its sources, stage by stage. The results are as follows.
First: in the materials of the various writers checked in this round, no one was found to propose this complete chain of hand-offs; the links between the stages were assembled by the project itself. The stages belong to theories that are incompatible with one another. In Liao Yihu's system, for example, there is only one authorial layer, Xu Xinggong; "Kong Meixi, Wu Yufeng" are for him a split of the characters for "Eastern Wu," which leaves no room for intermediate layers such as Fang Yizhi or Kong Shangren. Some individual nodes have sources; that does not give a source to the links between the nodes.
Second: the parts that have sources.
- The theory of Xu Xinggong as author comes from Liao Yihu. Among the prototype identifications used in the backstage, Lin Chong for Daiyu, Ma Xu for Jia Lian and others are found in Liao's readable writings, which contain more than twenty such identifications. Several other identifications are recorded in the backstage as Liao Yihu's, such as Lin Gudu for Xiren, Ni Fan for Ping'er, Hong Shiying for Xiaohong and She Xiang for Jia She, but were not found in Liao's own writings in this round; Appendix 3 lists them as "source not found (once recorded in the backstage as Liao Yihu)," and does not put them under his name. Author's statement: the prototypes contemporary with Xu Xinggong came basically all from Liao Yihu; apart from the Xiangyun item, these prototypes had no influence on the writing of the current draft of the reconstruction. Lines written in earlier drafts on the basis of such prototypes and later withdrawn, such as the Fuzhou thread, in which Lingguan returns to Fuzhou to comment on the book, and the Bao Yong thread, are listed under "withdrawn" in §6.3. This is the author's account of the project's own writing history, not a judgement on Liao's theory.
- The theory of Fang Yizhi as author comes from Chen Xianfu.
- Kong Meixi as Kong Shangren comes from Zhang Zhijian (Sina blog, 2011).
- The phrase "late elder brother of the Western Hall" comes from a note in a disputed manuscript; the problems of that manuscript are left to paper 3.
Third: the parts the project inferred itself: the "Wu Yufeng group"; Fang Yizhi's addition of the Roster of Feeling and of the monk and the Daoist; and the reading of two solar-term dates as Fang Yizhi's "signature" on his final draft (§6.2).
Fourth: the backstage is not consistent with itself, and more than ten points remain unreconciled. For example: whether "mourning the Ming" belongs to Xu Xinggong or to the Wu Yufeng group; whether Jihu Sou was Lin Chong, or the supplementer of some continuation; whether the cyclical dates should be moved back by one sixty-year cycle; and the total number of chapters, first 108 and later changed to 110.
Fifth: wording. In the backstage notes there are about thirty places where a conjecture is worded as a settled conclusion, such as "iron-clad proof," "final ruling," and "solving a riddle unsolved in 200 years of Redology." This paper does not use such wording, and where it quotes it, it gives the source. Where the backstage claims something as "the first" or as "never said by anyone," a literature search is made before publication.
6.2 Withdrawing the solar-term signature
The backstage once read "the thirtieth day of the eleventh month, the winter solstice" in chapter 11, and "the twenty-sixth day of the fourth month, the beginning of Grain in Ear" in chapter 27, as Fang Yizhi's signature on a final draft of 1665–1666, and on that basis, together with the objects in the first eighty chapters and other matters, drew up a timetable for the whole book covering 1655–1674.
The recalculation does not match, and the inference is withdrawn. Recalculated with modern true solar terms and true new moons (sxtwl 2.0.7, UTC+8), the winter solstice of 1665 falls on the fifteenth day of the eleventh month, and that eleventh month is a short month, with no thirtieth day; Grain in Ear of 1666 falls on the third day of the fifth month, seven days later than "the twenty-sixth day of the fourth month." The dates obtained do not support the original inference, and the author has withdrawn it on that basis. With differences of half a month and of seven days, the project does not make the pages of the calendar of the time a precondition of the withdrawal; the pages of the Shixian almanac for those years are wanted only for the record.
Even if a year in which both dates fit were one day found, that would show only that the story could be set in that year. It could not reach "someone finalized the book in that year," still less "this person was the author."
The signature is withdrawn, but the timetable left one thing behind: the sequence of years, the festivals and seasons, and the characters' ages (counted in sui) in each chapter of the reconstruction were arranged by it; the backstage records that Daiyu dies in Rec. ch. 95 on the third day of the third month, the Shangsi festival, with the note "according to the timetable." The text of the reconstruction gives no reign titles, and does not name Fang Yizhi. So the timetable has entered the text in the form of a schedule of festivals and seasons; it now serves only as a working coordinate for the writing, and has no independent reason of its own (Appendix 4, item 2). An attentive reader might work back from the festivals and seasons of the chapters to the years set in the backstage; the project does not make this a claim about the years of the story or of the book's composition. The other thing that entered the text is the Kangxi-era ground of the object layer; its source and its reason are given in §6.3.
6.3 What entered the reconstruction
The backstage materials list twenty-four items, each marked as in the text, withdrawn, or not in the final text; Appendix 4 gives an account of each. The main text selects, grouping the items by the part that the thought of layering played.
Mostly left in the backstage. The great majority of the prototype identifications; Lin Xue commenting on the book, and the theory of a relay across three generations of Zhiyan Zhai; Fang Yizhi's "four avatars"; and the "trace of renaming" in the Prince of Dong'an.
By the author's statement (§6.1), the prototypes of the Liao Yihu line, apart from the Xiangyun item, had no influence on the writing of the current draft of the reconstruction. The backstage once linked the apricot tree outside Ping'er's room in Rec. ch. 109 to Ni Fan's "Old Apricot Studio" (古杏轩), and linked Tanchun's surviving her distant marriage in Rec. ch. 85 to Shao Jiechun's "poetic compensation"; by the author's statement, the writing of these two passages does not rest on the prototypes, and they too are recorded in this group.
Those that reached the plot or the wording.
- In Rec. ch. 109 the Sanhe troupe stages The Peach Blossom Fan. This echoes the theory that Kong Meixi was Kong Shangren, but the final proofreading required that Kong Shangren's name not appear in the text.
- In Rec. ch. 110 Xiangyun's "Rhapsody on My Humble Lodging" (敝舍赋) has the line "I once read the Dongxiyang kao" (尝览《东西洋考》, Zhang Xie's study of the maritime world), which draws on Zhang Xie as a prototype. As far as the prototype identifications of the Liao Yihu line go, this is the only item that influenced the writing of the current draft.
These prototypes suggested material. Their entry into the novel does not, in return, prove that the proposed prototypes are the persons behind the original characters.
Wording changed; the layering account left in the backstage. In Rec. ch. 83 Grandmother Jia admonishes Jia Zheng. An earlier draft of the reconstruction had her speak of "an empty frame" (虚架子); this is now changed to a warning not to care only for appearances in outside affairs, so that Grandmother Jia does not foresee the confiscation of the family's property. The backstage recorded this change as "what is kept is acknowledged": what a former author kept, a later one acknowledges. This paper cannot state which place in the first eighty chapters the former author kept, so the account is left in the backstage, and is not counted as a case in which the thought of layering directly influenced a choice.
Layering conjecture used where it is not load-bearing. The rereading of Qin Keqing. The Roster of Feeling in Rec. ch. 110 says of her: "The world says she ruined the family by her beauty, and names her lewd; yet her sin was not beauty but politics." In Rec. chs. 104 and 105 her background is supplied through Baozhu's account and through Qin Keqing appearing in a dream. This rereading rests on a conjecture: that Cao Xueqin did not himself accept the "lewdness" with which the Mirror for the Romantic layer had cast Qin Keqing (overview §3.5). This is the project's own judgement on the number of layers and on whose each was. The overview (§1.4) has already listed it, together with "the reason for the excision of the Tianxiang Tower episode," as not load-bearing: if it falls, only a part is damaged. The reconstruction makes her come from the line of the deposed crown prince, in agreement with Liu Xinwu's view, as noted in §5.2.
Grounds established independently; the thought of layering only a side aid.
- The Tianxiang Tower is not supplied. The ground is the inviolable: what the first eighty chapters narrate is neither changed nor supplemented (overview §3.5). The layering thought that "a deletion is a blank to be left" serves only as a side aid.
- The object layer takes a Kangxi-era ground, with the dynasty left unnamed. The source is the backstage, where it was originally linked to the Fang Yizhi layer: the backstage used it as "the yardstick for judging whether a given description in the first 80 chapters belongs to Fang Yizhi's original layer." The Fang Yizhi layer and the solar-term signature have both been withdrawn. The present reason is the author's: the glass, the Western medicines, the full-length mirrors and so on described in the first eighty chapters are, within the book, rarities that only the nobility could enjoy, and the project reads this as closer to conditions in the Kangxi reign than in the Qianlong. This is the project's own reading and a basis for its writing; it is not textual research, and this paper does not argue from it to any date of composition.
Withdrawn. The Fuzhou thread, in which Lingguan returns to Fuzhou to comment on the book; the mapping of Yuanchun's and Yuanyang's deaths onto the years 1627 and 1628; the double reading of Lady Wang "blocking the scissors"; the Bao Yong thread; and Zhen Baoyu becoming a monk and delivering the Yushi xinbian (玉石新编).
6.4 What they helped with
Following §1, the popular theories helped the project think; this section sets out how, item by item. The help was of two kinds: of method, and of material.
Of method, five items. Their sources lie mostly in scholarship and in the commentary, not in the popular theories.
| Item | What it helped with | Source | Used as |
|---|---|---|---|
| 1 Two notes at odds, neither judged mistaken | The notes need not rest on the same draft. This idea led paper 1 to register two notes at odds only as possibly two drafts, without judging either mistaken; for an instance see 18.17 and 6.48 in §3.2 | Scholarly layering theories and the commentary's own statements (overview §3.4: "nor did the commentary necessarily all rest on the same draft") | An aid to thinking: paper 1's rule of registration |
| 2 The verdicts may come from a different draft from the later text | Helped the project not to read the verdicts as a design fixed once and for all. But the reasons for the reconstruction's departures from the verdicts finally rest on the criterion and on the characters' choices, not on layering conjecture | Scholarly layering theories | An aid to thinking; the grounds for departure are set out separately (overview §§4.1, 6.2) |
| 3 A deletion is a blank to be left | The Tianxiang Tower episode is not restored. The indirect account in Rec. chs. 104 and 105 is the project's own arrangement, not a restoration of the deleted passage. This is the project's choice, not a general conclusion that "whatever was deleted must not be written" | Commentary 13.89 and others; the ground is the inviolable | The project's choice |
| 4 An unnamed dynasty and the ground of the objects | Anchors only the object layer, not allusions to events; no inference from prototypes to the dynasty or year in which the original was written | The backstage; the present reason is in §6.3 | A constraint on the writing |
| 5 Sifting the materials of the Guiyou manuscript | Looking at the materials of continuations with an eye to layering, and removing them layer by layer, as in "remove mourning-the-Ming from the Guiyou." What is removed, and why it can be, is handed to paper 3 | The backstage's own inference | Passed to paper 3 |
Of material, from popular and outside theories:
- the Dongxiyang kao in Rec. ch. 110 uses Zhang Xie, as found in Liao Yihu's readable writings;
- The Peach Blossom Fan in Rec. ch. 109 uses the theory that Kong Meixi was Kong Shangren, from Zhang Zhijian;
- for the dream after Daiyu's death in Rec. ch. 96, in which envoys from the Palace of the Moon come to meet her, the plot model is the story of the young late-Ming poet Ye Xiaoluan as recorded by her father Ye Shaoyuan in Yaowen (窈闻) (according to the backstage records; this paper has not read the original); Zhang Mi calls Ye Xiaoluan the prototype of Daiyu (in the preface to a book), but the project uses her story only as material, not as evidence of who Daiyu's prototype was.
By the author's statement (§6.1), the other prototypes of the Liao Yihu line had no influence on the writing of the current draft; for the two passages once linked in the backstage, Ping'er's apricot tree and Tanchun's distant marriage, see §6.3.
The popular theories each have their reasons. Their help to the project lay mainly in letting the writer see persons and events that could be drawn on as material, not in proving anything on the project's behalf.
7 The Minimal Working Judgement Adopted in This Paper, and Its Limits
The paper depends only on this. The main body of the text of the surviving first eighty chapters went through layering, and the one who integrated it into a book and fixed its centre was the Cao Xueqin named in the chain of titles and in the commentary (a role in the text).
Known exceptions, each recorded as another layer, and not part of the layer this paper depends on:
- the unfinished, damaged and patched end of chapter 22: 22.104, "from here on damaged and lost; what follows is a later supplement," and 22.106, "this chapter was not finished when Qin [Xueqin] died"; this paper does not rule on who wrote the supplied text or on what it was based;
- the title: the commentator "copied it and commented on it again in the jiaxu year, and went back to the title The Story of the Stone";
- the Tianxiang Tower: the commentator ordered Qinxi to delete it (13.89);
- the missing and supplied chapters 64 and 67;
- the Cheng–Gao edition's alterations to the first eighty chapters.
These are listed in the order that holds if the Zhiyan Zhai manuscripts came first. If the view that the Zhiyan Zhai manuscripts came later were to hold, items 1 to 3 would be recorded as layers of the Zhiyan Zhai lineage, and item 5 would run in the opposite direction; the one layer this paper depends on would not be affected, but the list of exceptions would have to be rearranged.
"Fixed its centre" is the project's working judgement, which this paper does not argue. What the centre is belongs to the overview's interpretation; it is for the overview to argue and to test, and it is not brought back as evidence for this layer.
What stands even if the commentary is not genuine. The Cheng lineage also has, in the chain of titles, the sentence "Cao Xueqin, in the Mourning-the-Red Studio, read it over for ten years and revised it five times" (from the common text; the page of the Chengjia edition to be checked); Cheng Weiyuan's preface says, "only, the book itself records that Mr. Cao Xueqin deleted and altered it several times" (common transcription; no transcription that can be named has been seen); and the inconsistencies are all in the text itself. The Cheng editions keep the book's own account, and Cheng Weiyuan's preface cites it, so the two do not count as independent pieces of supporting evidence. On this basis the paper keeps the working understanding of the segment "the one who integrated it into a book was the Cao Xueqin named in the text," whose historical attribution remains open to testing; the segment "fixed its centre" is not argued in this paper and is not included here. So the paper takes no part in the dispute over authenticity because it has no need to.
Conditions for overturning.
- Reliable evidence that the one who integrated the main body of the text of the surviving first eighty chapters was not the one named in the chain of titles and the commentary, the known exceptions listed above not counting.
- Or evidence that the present text has no layering on the side of composition (revision by the author, merging, or reworking), and was written in one pass.
Since "Cao Xueqin" is taken as a role in the text, the first condition concerns the one named in the text, and has nothing to do with who this person was historically. The second conflicts with the materials of §3 and is unlikely, but it is stated all the same.
Limits. If the judgement on composition changes, what must change are the corresponding attributions, layers and citations; whether the textual interpretation "people have qing" then fails would need a separate argument. The overview (§1.4) has already said that even if the working hypotheses in its not-load-bearing column were overturned, "people have qing" and the method of interpretive reconstruction could still stand.
The paper does not depend on:
- any determination of the number of layers or of whose each was. The main thesis does not; one not-load-bearing place used such a determination once, namely the rereading of Qin Keqing (§6.3);
- any prototype identification;
- any ruling on the authenticity of any material;
- the provenance of the Guiyou manuscript (paper 3).
8 Conclusion
The conclusion of this paper can be stated briefly. The book is layered. The materials are in §3 and the theories in §§4 and 5, all registered by provenance, without judgement on which is right. The paper depends on one layer only, and has stated its exceptions, the conditions under which it would be overturned, and what it does not depend on. The layering conjectures the project used in its backstage are disclosed as they stand in §6: which were only thought about, which entered the reconstruction, and which were withdrawn. Where there is a source, the source is given; what was not found in this round is marked as not found; and only what the backstage explicitly records as set up by the project itself is marked as proposed by the project.
The reader should be able to see three things clearly: what the original materials say, which step is some scholar's interpretation, and which step is the project's own association. The reader should also be able to see, once an association has been withdrawn, which choices in the novel are still kept for other reasons. If these three are kept apart, the paper has no need to win a dispute over authorship.
Appendix 1 Catalogue of Scholarly Theories of Composition
Listed by the groupings of §4. In the reliability column, "reprint" means that the scholar's own words have been seen, but in an online reprint; "second-hand" means known only from someone else's account; "catalogue" means that only a record in a library, publisher's or journal database has been seen; "checked" means that the author's side has checked the record listed. Everything marked "reprint" or "second-hand" is to be checked against print or a full-text database before the next version.
| No. | Scholar | Line | Main claim | Representative works | Opponents | Reliability |
|---|---|---|---|---|---|---|
| 1 | Hu Shi | One book revised (baseline) | Written by Cao Xueqin, the last forty chapters supplied by Gao E; no layer of old drafts | "A Textual Study of Honglou meng" 红楼梦考证, first version published May 1921 as the preface to the punctuated edition of the Yadong Library, Shanghai, revised November 1921; "New Materials for the Study of Honglou meng" 考证《红楼梦》的新材料, Xinyue, 1928 (volume and issue to be checked) | Cai Yuanpei's allegorical school; Pan Chonggui; Dai Bufan; Ouyang Jian | Conclusions: reprint; dates of publication: second-hand |
| 2 | Yu Pingbo | One book revised | The first eighty chapters are by Cao; discusses inconsistencies such as a setting that may be "either south or north" | Honglou meng bian 红楼梦辨, Yadong tushuguan, 1923 | No specific refutation found | Second-hand |
| 3 | Zhou Ruchang, Feng Qiyong | One book revised | "Revised it five times" refers to Cao Xueqin's own revisions; Feng: the Gengchen is the most complete of the early manuscripts made in Cao Xueqin's lifetime | Feng Qiyong, Lun Gengchen ben 论庚辰本, Shanghai wenyi chubanshe, 1978; the sentence quoted in the main text is from "San lun Gengchen ben" 三论庚辰本 (reprint on Sohu); Zhou Ruchang, Honglou meng xinzheng 红楼梦新证, Tangdi chubanshe, 1953 (to be checked) | Dai Bufan; Pan Chonggui; Ouyang Jian | Feng: reprint; Zhou: to be checked |
| 4 | Cai Yijiang | One book revised | A Mirror for the Romantic was an earlier work, the basis on which The Story of the Stone was worked over and revised; against two books combined | "How Was Honglou meng Made?" 《红楼梦》是怎样成书的？ (original source to be checked); two articles against Ouyang Jian, Honglou meng xuekan 1993, no. 3, and 1994, no. 3 | Du Chungeng | Reprint; second-hand |
| 5 | Eileen Chang | One book revised (also touching on two books) | Cao Xueqin was in charge of the rewriting, and the variants are strata; Baoyu's age was lowered from revision to revision; the Land of Illusion appeared only after A Mirror for the Romantic was taken in | Honglou mengyan 红楼梦魇 (Nightmare in the Red Chamber), Taipei: Huangguan, 1977 | No specific refutation found | Reprint |
| 6 | Chen Qinghao | One book revised | Three of the five titles name drafts of different stages; Daiyu took over the place Xiangyun had held as a child in the Jia household | "A First Study of the Composition of the Eighty-Chapter Story of the Stone" 八十回本《石头记》成书初考, Wenxue yichan 1992, no. 2; "A Second Study of the Composition of the Eighty-Chapter Story of the Stone" 八十回石头记成书再考, Honglou meng xuekan 1995, no. 1 | None found | Second-hand |
| 7 | Shen Zhijun | One book revised | Revision in stages, each layer corresponding to a title | Honglou meng chengshu yanjiu 红楼梦成书研究, Beijing: Zhongguo shudian, 2004; "'The New Baoyu' and 'the Old Baoyu': An Essay on How Honglou meng Was Made" “新宝玉”和“旧宝玉”——《红楼梦》成书过程试探 | Opposed to Du Chungeng | Catalogue; reprint |
| 8 | Liu Shide | One book revised | The first eighty chapters consist of two components, the first draft and the revised draft | Honglou meng zhi mi: Liu Shide xueshu jiangyan lu 红楼梦之谜：刘世德学术讲演录 (reissued 2024; first edition to be checked); Honglou meng banben tanwei 《红楼梦》版本探微 | None found | Second-hand |
| 9 | Mei Jie | Outside the three lines (the texts and their relation to Zhiyan) | The book was finished, and its later part was lost through lending; holds the Jing manuscript to be genuine | "Another Look at the Jing Manuscript" 也谈靖本, Honglou meng xuekan 2002, no. 1 | Those who hold the Jing manuscript to be forged, such as Gao Shuwei 2024 | Reprint |
| 10 | Chao Kang, Chen Chung-yi | One book revised | The last forty chapters of the Cheng–Gao edition preserve remnants of a layer of Xueqin's early drafts | Honglou meng xintan 红楼梦新探 (first published in Hong Kong c. 1970; to be checked); Honglou meng yanjiu xinbian 红楼梦研究新编 (Taipei: Linking, c. 1975; to be checked) | None found | Second-hand |
| 11 | Zhou Shaoliang | Two books combined | Two first drafts, the original Story of the Stone and A Mirror for the Romantic, were joined; the romantic episodes come from A Mirror for the Romantic | "'Xueqin Formerly Had a Book Called A Mirror for the Romantic'" “雪芹旧有《风月宝鉴》之书”, Honglou meng xuekan 1979, no. 1 | Cai Yijiang | Reprint |
| 12 | Wu Shih-ch'ang | Two books (two sources) combined | "Tangcun's short prefaces"; two sources, the first draft seen by Mingyi and "what the stone recorded," with Cao Xueqin as the one who combined them and worked them over | On The Red Chamber Dream: A Critical Study of Two Annotated Manuscripts of the XVIIIth Century, Oxford: Clarendon Press, 1961; "On the First Draft of Honglou meng Seen by Mingyi" 论明义所见《红楼梦》初稿, Honglou meng xuekan 1980, no. 1; Honglou meng tanyuan waibian 红楼梦探源外编, Shanghai guji chubanshe, 1980 | Zhou Ruchang | Monograph: catalogue; Hanan's review of 1961 (BSOAS 24.3, pp. 603–605) checked; Hsia's review (JAS 21.1, pp. 78–80) second-hand; article of 1980: reprint; Waibian p. 483: second-hand |
| 13 | Zhu Danwen | One book revised (provisional) | The titles name drafts of one novel at different stages; only two drafts certainly existed; the genealogy of the old draft differs from the present text | Honglou meng lunyuan 红楼梦论源, Jiangsu guji chubanshe, 1992; enlarged edition, Zhejiang guji chubanshe, 2024 | No specific refutation found | The passage on p. 197 as quoted by Xia Wei: quotation checked; the book itself not checked |
| 14 | Du Chungeng | Two books combined | The Story of the Stone, centred on the Rong mansion, and A Mirror for the Romantic, centred on the Ning mansion, were combined | "The Rong and Ning Mansions, Two Books" 荣宁两府两本书, Honglou meng xuekan 1998, no. 3; "On Cao Xueqin's 'Read It Over for Ten Years and Revised It Five Times'" 曹雪芹“披阅十载、增删五次”考, same journal, 2007, no. 3; "The Story of the Stone Could Not Have Evolved out of A Mirror for the Romantic" 从《风月宝鉴》是演变不出一部《石头记》来的 (issue to be checked) | Cai Yijiang | Reprint; catalogue, second-hand |
| 15 | Yurui | Another's old draft (the source document) | "Cao Xueqin obtained it … and deleted and altered it as many as five times" | Zaochuang xianbi 枣窗闲笔, facsimile, Wenxue guji kanxingshe, 1957 | Liu Mengxi: mostly hearsay; its authenticity is the subject of separate study (a chapter in Gao Shuwei 2024) | Second-hand |
| 16 | Dai Bufan | Another's old draft | Brother Stone's old draft was A Mirror for the Romantic, and Cao Xueqin was the reviser who reworked it; four kinds of internal evidence | "Unveiling the Mystery of the Author of Honglou meng" 揭开《红楼梦》作者之谜, Beifang luncong 1979, no. 1; "Brother Stone and Cao Xueqin" 石兄和曹雪芹, same journal, 1979, no. 3; "On Qin Keqing's Late Death" 秦可卿晚死考 (first publication to be checked); all collected in Dai Yun, ed., Hongxue pingyi: waipian 红学评议·外篇, Wenhua yishu chubanshe, 1991 | Nearly thirty articles, collected in Collected Papers of the Debate on the Authorship of Honglou meng 《红楼梦》著作权论争集, Taiyuan: Shanxi renmin chubanshe, 1985; Liu Mengxi, Wenyi yanjiu 1979, no. 4; Chen Xizhong, Hou Zhongyi; Zhang Bibo, Zou Jinxian | Issues: second-hand; the Collected Papers: catalogue record checked |
| 17 | Pan Chonggui | Another's old draft (claim on composition only) | Another, anonymous original author; Cao Xueqin the one who held the brush in revising, supplementing and correcting | Hongxue liushi nian 红学六十年, Taipei: Sanmin shuju, 1991 | General objections from the textual-research school | Second-hand |
| 18 | Pi Shumin | Another's old draft | The Story of the Stone had two authors; Zhiyan was Li Ding | Honglou meng kaolun ji 红楼梦考论集, Taipei: Linking, 1984; Suzhou Li jia yu Honglou meng 苏州李家与红楼梦, Xin wenfeng, 1996; Li Ding yu Shitou ji 李鼎与石头记, Wenjin, 2002 | None found | Catalogue; his own note: reprint; the rest second-hand |
| 19 | Zhao Guodong, Liu Runwei | Another's old draft | Zhao takes Zhiyan Zhai to be the author (1990); Liu holds that Cao Yuan met the "sufficient conditions" for original authorship (1994) | Titles and issues to be checked | Du Zhijun: "plainly both lack the support of solid documentary material" | Second-hand |
| 20 | Ouyang Jian | Priority of the Zhiyan Zhai manuscripts and the Cheng editions | The Cheng editions first, the Zhiyan Zhai manuscripts later | Series of articles from Fudan xuebao 1991, no. 5; Honglou xinbian 红楼新辨, Huacheng chubanshe, 1994; Hongxue bianwei lun 红学辨伪论, Guizhou renmin chubanshe, 1996; Huanyuan Zhiyan Zhai 还原脂砚斋, Heilongjiang jiaoyu chubanshe, 2003 | Ying Bicheng, Fudan xuebao 1992, no. 1; Chen Zhao; Cai Yijiang; Song Mouyang; Yang Guanghan; Feng Qiyong; Xu Naiwei | Second-hand |
| 21 | Xia Wei | One book revised | A Mirror for the Romantic was the first draft, from which the creative thought evolved | "From A Mirror for the Romantic to Dream of the Red Chamber: The Evolution of the Book's Composition and Creative Thought" 从《风月宝鉴》到《红楼梦》——成书与创作思想的嬗变, Wenxue pinglun 2022, no. 6, pp. 201–210 | None found | Reprint; catalogue |
| 22 | Li Qingyu | One book revised | The five titles correspond to four stages in the conception of fiction | "The Development of Cao Xueqin's Conception of Fiction, Seen from the Titles of Honglou meng" 从《红楼梦》题名论曹雪芹小说创作观念的演进, Wenyi lilun yanjiu 2014, no. 6, pp. 138–145 | None found | Reprint (abstract) |
| 23 | Hu Lianhao | One book revised | The location of the Jia mansion was changed from Jinling in the first draft to Chang'an in the finished draft | "From 'Jinling' to 'Chang'an' in the Making of Honglou meng" 《红楼梦》成书从“金陵”到“长安”考辨, Honglou meng banben yanjiu jikan 4, Huaqiao chubanshe, 2024 | None found | Reprint |
| 24 | Tan Dejing | Two books combined | A selective integration of A Mirror for the Romantic | "The Relation between Honglou meng and A Mirror for the Romantic Reconsidered" 《红楼梦》与《风月宝鉴》之关系再探 (original publication to be checked) | None found | Reprint |
Anchor for the history of scholarship: Du Zhijun, "Thirty Years of Redology in the New Period, 1978–2008" 新时期红学三十年：1978—2008, Honglou meng xuekan 2009, no. 2 (reprint).
Appendix 2 Identifications of the Title-Givers and Commentators with Historical Persons
Mostly popular; listed without discussion. "Not checked" means known only from an account, the proponent's own writing not having been read.
| Name | Identified as | Proponent and source |
|---|---|---|
| Kong Meixi | Kong Shangren | Zhang Zhijian, Sina blog, 2011 |
| Kong Meixi | Kong Jihan | Wu Enyu, citing another's view (original not checked); Zhang Guangwen (2022) refutes it: in the jiaxu year Kong Jihan was only fifteen |
| Kong Meixi | Read as kong mei xi 恐霉熙, "Kangxi in dread" | Wu Chengmou |
| Kong Meixi | Combined with Wu Yufeng as a split of the characters 东吴 ("Eastern Wu"), indicating that the author was from Fujian | Liao Yihu |
| Kong Meixi | A fictitious person | Sina blog "Honglou meng xinjie," citing Zhou Chun |
| Wu Yufeng | Wu Meicun | Fu Bo and Zhong Changshan; Wu Chengmou |
| Wu Yufeng | Wu Wen | Ling shuo Honglou 另说红楼 (author not checked) |
| Wu Yufeng | A split of the characters 东吴 | Liao Yihu |
| Wu Yufeng | Homophone of "no jade peak" (无玉峰), an invention | Online self-media account (zimeiti) "Junjian yakan Honglou" 君笺雅侃红楼, 2023–2024 |
| Tangcun | Xu Biao, younger brother of Xu Xinggong | Liao Yihu |
| Tangcun | A younger brother of Wu Meicun | Wu Chengmou |
| Tangcun | Combined with Wu Yufeng and Kong Meixi into "Wu Mei(cun)" | Fu Bo and Zhong Changshan |
| Tangcun | Liang Qingbiao | Wang Zhengkang, 2025 |
| Zhiyan Zhai | Lin Xue | Liao Yihu |
| Zhiyan Zhai | Huang Hui, wife of Hong Sheng | Tumore |
| Zhiyan Zhai | You Tong | Fu Bo and Zhong Changshan |
| Zhiyan Zhai | Mao Pijiang's three concubines | Huang Weimin |
| Zhiyan Zhai | Cao Fu | Sina blog "Cao Fu," 2012 |
| Zhiyan Zhai | The core of the commentary by Cao Xueqin himself | Huo Guoling |
| Zhiyan Zhai | The prototype of Shi Xiangyun (a scholarly view) | Zhou Ruchang, Honglou meng xinzheng (not checked) |
| Jihu Sou | Lin Chong | Liao Yihu |
| Jihu Sou | "Xue'er" of the Deng family, concubine of Hong Sheng | Tumore |
| Jihu Sou | Qin Songling | Fu Bo and Zhong Changshan |
| Jihu Sou | A Ming official who went over to the Qing | The Master of the Zhizhen Studio (Zhizhenzhai zhu 至真斋主) |
| Jihu Sou | Cao Fu | Sina blog "Cao Fu"; also held in scholarship (not checked) |
| Western Hall | Cao Yin's study | The common reading; first proponent not found |
| Western Hall | You Tong (late sobriquet "Old Man of the Western Hall") | Fu Bo and Zhong Changshan take You Tong as Zhiyan Zhai; whether they argue by way of the "Western Hall" not checked |
| Late elder brother of the Western Hall | Found in a note of a disputed manuscript | Left to paper 3 |
Appendix 3 Tracing the Backstage Working Chart, and What Was Taken from It
Table 1 Tracing the six stages
| Stage | What the backstage wrote | What the proponent actually claims | Check |
|---|---|---|---|
| One | Xu Xinggong's Story of the Stone, its characters grounded in the literary circle of Fuzhou | Liao Yihu: Xu Xinggong is the author, "Cao Xueqin" his pen name; no intermediate layers such as Fang or Kong | The authorship claim has a source. The backstage's description of the base draft as "leaning to Laozi and Zhuangzi, against institutions" is not found in Liao's readable writings; "no ending" can only be inferred indirectly |
| Two | Fang Yizhi's Record of the Passionate Monk, begun in 1665 and finalized in 1665–1666; the Roster of Feeling and the monk and the Daoist added by him | Chen Xianfu: Fang Yizhi is the author, and Cao Xueqin his invention | The authorship claim has a source. The addition of the Roster and of the monk and the Daoist, the year of finalization and the solar-term signature have no source that could be found, and are the project's own inference; the solar-term signature failed recalculation and has been withdrawn |
| Three (i) | Kong Shangren's A Mirror for the Romantic | Zhang Zhijian: Kong Meixi is Kong Shangren, implying that Kong Shangren took part in the writing | The identification has a source; no specific argument that he "completed a version of A Mirror for the Romantic" was found |
| Three (ii) | The Wu Yufeng group's Dream of the Red Chamber, mourning the Ming | Wu Yufeng is identified as Wu Meicun, Wu Wen and others, each with its proponent; the theme of "mourning the Ming" is held by several popular writers | "The Wu Yufeng group" not found; the project's own inference |
| Four | The three-part Mirror for the Romantic of the "late elder brother of the Western Hall" | From a note in a disputed manuscript | Left to paper 3 |
| Five | Cao Xueqin, the reviser | Many popular theories take Cao Xueqin as the reviser | This paper takes only the role in the text (§1) |
| Whole chain | A relay across six stages | — | In the materials checked in this round, proposed by no one; the links were assembled by the project itself |
Table 2 What the project took
| Element | Taken from | How it was transformed | Where it is used | Status |
|---|---|---|---|---|
| Kong Meixi as Kong Shangren | Zhang Zhijian | From the identification of a title-giver to a play staged by a troupe | Rec. ch. 109: the Sanhe troupe stages The Peach Blossom Fan; Kong Shangren's name does not appear in the text | In the text |
| Zhang Xie for Xiangyun | Liao Yihu | From prototype to a book Xiangyun has read | Rec. ch. 110: "Rhapsody on My Humble Lodging," "I once read the Dongxiyang kao" | In the text; as far as the prototype identifications of the Liao Yihu line go, the only item in the current draft that influenced the writing |
| Ni Fan's "Old Apricot Studio" | Recorded in the backstage as Liao Yihu's; not verified in this round | Once linked in the backstage | Rec. ch. 109: the apricot tree outside Ping'er's room | The scenery is in the text; by the author's statement, the writing does not rest on the prototype |
| Shao Jiechun | Prototype from Liao Yihu; the "poetic compensation" is the backstage's own inference | Once linked in the backstage | Rec. ch. 85: Tanchun married far away, to Nanming | The plot is in the text; by the author's statement, the writing does not rest on the prototype |
| The story of Ye Xiaoluan | Ye Shaoyuan's Yaowen; Zhang Mi's preface (both according to the backstage records; this paper has not read the originals) | From a record to a dream | Rec. ch. 96: envoys from the Palace of the Moon come to meet her | In the text; used only as material, not as evidence of a prototype's identity |
| Bao Juechu for Zhen Baoyu | Liao Yihu | Once identified in the backstage; by the author's statement, the writing does not rest on this prototype | Rec. ch. 96: after delivering the jade, Zhen Baoyu stays on and sets his heart on the examinations, without bringing the Yushi xinbian and without becoming a monk; the Bao Yong thread is cut entirely | The Bao Yong thread withdrawn; the delivery of the jade rests instead on 18.116 (Appendix 4, item 13) |
| The solar-term signature | The backstage's own inference | A timetable | The festivals, seasons and ages of each chapter arranged by the timetable, with no reign titles | Signature withdrawn; the timetable entered the text as a schedule of festivals and seasons, only as a working coordinate |
| The Kangxi-era ground | The backstage, originally linked to the Fang Yizhi layer | A constraint on the writing | The object layer | Kept, with a reason given separately (§6.3) |
| "The verdicts fixed by an earlier layer, conduct by a later one" | The backstage's own inference | An aid to thinking | The verdicts may come from a different draft | An aid to thinking; the grounds for departure are set out separately |
| "What is kept is acknowledged" | The backstage's own inference | A change of wording | Grandmother Jia's sentence in Rec. ch. 83 | Wording changed; this account left in the backstage |
| "A deletion is a blank to be left" | The backstage's own inference, drawing on the commentary | Not supplied | The Tianxiang Tower | In the text; the ground is the inviolable |
| "Remove mourning-the-Ming from the Guiyou" | The backstage's own inference | Sifting the materials of continuations | — | Handed to paper 3 |
| Lin Xue commenting on the book; a relay across three generations of Zhiyan Zhai | Lin Xue as Zhiyan Zhai from Liao Yihu; the three-generation relay is the backstage's own inference | — | — | Left in the backstage |
| Lingguan returning to Fuzhou to comment on the book | The backstage's own inference | — | — | Withdrawn |
| 1627 and 1628 mapped onto the deaths of Yuanchun and Yuanyang | The backstage's own inference | — | — | Withdrawn |
| The double reading of Lady Wang "blocking the scissors" | The backstage's own inference | — | — | Withdrawn |
Table 3 Sources of prototypes
Following the author's ruling, only the character, the proposed prototype and the source are listed, without argument, and not as evidence.
- "Liao Yihu" means found in Liao's readable writings, namely his Bilibili column and the sample chapters on Douban Read (References).
- "Source not found (once recorded in the backstage as Liao Yihu)" means attributed to Liao in the backstage, but not found in Liao's own writings in this round.
- "Source not found" means that the backstage gives no source, and none was verified item by item in this round.
- "Proposed by the project" is used only for items that the backstage explicitly records as set up by the project itself.
Author's later note: the prototypes contemporary with Xu Xinggong came basically all from Liao Yihu, including the Zhihu pieces that could not be read in this round; apart from the Xiangyun item, they had no influence on the writing of the current draft (§6.1). This note is kept separate: it is not used to reassign each row automatically to Liao Yihu, nor is "not found" taken to mean "proposed by the project."
| Character | Proposed prototype | Source |
|---|---|---|
| Baoyu | Xu Xinggong 徐兴公 | Liao Yihu |
| Baoyu | Cao Yin 曹寅 (middle layer); an early avatar of Fang Yizhi 方以智 | Proposed by the project |
| Daiyu | Lin Chong 林宠 | Liao Yihu |
| Daiyu | Ye Xiaoluan 叶小鸾 (model for her poetic manner, her fate and the scene of her death) | Zhang Mi's preface (according to the backstage records; this paper has not read the original); the project uses her only as material, not as evidence of identity |
| Baochai | Xie Zhaozhe 谢肇淛 | Liao Yihu |
| Yuanchun | Ye Xianggao 叶向高 | Liao Yihu |
| Tanchun | Shao Jiechun 邵捷春 | Liao Yihu |
| Xiangyun | Zhang Xie 张燮 | Liao Yihu |
| Qin Keqing | Xu Tong 徐熥 (in Liao, the prototype of "Qin Ke'er") | Liao Yihu |
| Qin Keqing | The line of the deposed crown prince | Proposed by the project; agrees with Liu Xinwu's view, which came first (2003); similarities and differences in paper 8 |
| Jia Zheng | Xu Ang 徐㭿 | Liao Yihu |
| Jia Lian | Ma Xu 马歘 | Liao Yihu |
| Xue Pan | Xu Zhi 许豸 | Liao Yihu |
| Lingguan | Lin Xue 林雪 | Liao Yihu |
| Jia Qiang | Zheng Bangxiang 郑邦祥 | Liao Yihu |
| Zhen Baoyu | Bao Juechu 包厥初 | Liao Yihu |
| Aiguan | Ai Rulüe 艾儒略 (Giulio Aleni) | Liao Yihu |
| Xiren | Lin Gudu 林古度 | Source not found (once recorded in the backstage as Liao Yihu) |
| Ping'er | Ni Fan 倪范 | Source not found (once recorded in the backstage as Liao Yihu) |
| Xiaohong | Hong Shiying 洪士英 | Source not found (once recorded in the backstage as Liao Yihu) |
| Jia She | She Xiang 佘翔 | Source not found (once recorded in the backstage as Liao Yihu) |
| Xifeng | Chen Yiyuan 陈一元 | Source not found |
| Li Wan | Dong Yingju 董应举 | Source not found |
| Yingchun | Shang Mei 商梅 | Source not found |
| Xiangling | Xu You 许友 | Source not found |
| Baoqin (宝琴) | Xu Bi 许珌 | Source not found |
| Qingwen | Wu Yu 吴雨 | Source not found |
| Yuanyang | Dong Xiaoshuang 董小双 | Source not found |
| Sheyue | Chen Jianfu 陈荐夫 | Source not found (search leads point to a Zhihu page of Liao Yihu's, whose text was not read) |
| Xing Xiuyan | Shen Ye 沈野 | Source not found |
| Feng Ziying | Ruan Zihua 阮自华 | Source not found (search leads point to a Zhihu page of Liao Yihu's, whose text was not read) |
| Jiang Yuhan | Tu Benjun 屠本畯 | Source not found (search leads point to a Zhihu page of Liao Yihu's, whose text was not read) |
| Lady Wang | The wife of Wang Yingshan 王应山 | Source not found |
| Sun Shaozu | Zhong Xing 钟惺 (cultural layer); Sun Changyi 孙昌裔 | Source not found |
| Zhang Dehui | Zhang Limin 张利民 | Source not found |
| Lady You | You Zongzhen 游宗振; or, on another view, no prototype | Source not found; both views stand in the backstage |
| You Erjie | You Jiyuan 游及远 | Source not found |
| Leng Zixing | Jiang Yifang 蒋奕芳; or, on another view, no prototype | Source not found; both views stand in the backstage |
| Ma Daopo | Ma Congfu 马从甫 | Source not found |
| Qianxue | Cui Shizhao 崔世召 | Source not found |
| Wei Ruolan | Yan Jizu 颜继祖 | Source not found |
| Grannie Liu | Zou Shifeng 邹时丰 | Source not found |
| Qiaojie | Wang Gaigong 王盖公 | Source not found |
| Jinchuan | Zheng Huaikui 郑怀魁 | Source not found |
| Liu Xianglian | Wu Yuanhan 吴元翰 | Source not found |
| Jia Yun | Lin Yingpin 林应聘 | Source not found |
| Jia Huan | Wang Yu 王宇 | Source not found |
| Kuiguan | Yu Yingkui 喻应夔 | Source not found |
| Baoqin (抱琴, Yuanchun's maid, not the Baoqin 宝琴 above) | Ye Xianggao's nurse | Source not found |
| Cuimo | An extension of Shao Jiechun | Source not found |
| Ying'er, Yanhong, Cuiyun, Qiutong, Lin Ruhai, Jia Cong, Baochan, Ruhua, Feng'er and others | Recorded in the backstage as having no prototype ("pure Cao layer") | Proposed by the project |
Appendix 4 Table of Twenty-Four Influences
The twenty-four items listed in the backstage materials are accounted for one by one. The chapters of the reconstruction were spot-checked against the project's collections of sealed drafts (Rec. chs. 81–95 from round 54, Rec. chs. 96–110 from round 53). For Rec. chs. 96–99 there are also single-chapter sealed drafts of round 54, which this paper has not obtained; the "Rec. ch. 99" of item 7 rests on the round 53 collection, and a reviewer confirmed it separately against the round 54 single chapter. The reconstruction is revised round by round, and details follow the version used at the time. The column "independent reason for keeping it now" is filled in by the author's ruling: where the backstage or the overview records a reason explicitly, it is entered; where the text of the first eighty chapters or the commentary offers an anchor, the note number or chapter is given; where there is neither, the entry is "no independent reason." A reason need not be evidence from the original; it is enough to say that it is the project's own reason, without raising it into Cao's intention.
Author's statement (§6.1): the prototypes contemporary with Xu Xinggong came basically all from Liao Yihu; apart from the Xiangyun item, they had no influence on the writing of the current draft. Where prototypes of this kind appear in the column "conjecture used and its source," they record only where the backstage got them, and do not mean that the prototype influenced the writing.
The original text listed in the column "independent reason for keeping it now" supports only the part it states. Chapters, plot points, treatments and epithets in the same row that it does not directly determine do not acquire the status of direct evidence because a note number is listed; they are the project's arrangements.
| No. | Item | Conjecture used and its source | Where in the reconstruction | Status | Independent reason for keeping it now |
|---|---|---|---|---|---|
| 1 | Chapter count and the Roster | The 108-chapter scheme comes from continuations modelled on The Water Margin; the underlying original had 110 (backstage) | Written as Rec. chs. 81–110; the Roster of Rec. ch. 110 shows only sixty women, with Baoyu at the head, and the women are not collectively "returned to their places" (归位) | Changed: the earlier 108 overturned | The criterion: the Roster of Feeling is alive (overview §3.5) |
| 2 | Time anchor | Solar-term signature (the backstage's own inference) | The sequence of years, the festivals and seasons, and the ages in each chapter arranged by the timetable, with no reign titles | Signature withdrawn; the timetable entered the text as a schedule of festivals and seasons, with no reign titles | No independent reason; serves only as a working coordinate (§6.2) |
| 3 | The monk and the Daoist, and the frame of the Passionate Monk | Fang Yizhi's design (the backstage's own inference) | In Rec. ch. 106 Baoyu leaves of his own accord and is not taken away; the jade's return to Greensickness Peak is moved to Rec. ch. 110; the chapter title calls him "the lay disciple who feels for the unfeeling" (情不情居士) | Changed | 8.140, Jiaxu upper-margin, "according to Jinghuan's Roster of Feeling, Baoyu is 'feeling for the unfeeling'"; 19.80, Gengchen, citing the Roster, "Baoyu: feeling for the unfeeling." These support only the use of "feeling for the unfeeling" in the chapter title; the epithet "lay disciple," his leaving of his own accord in Rec. ch. 106, and the moving of the jade's return to Rec. ch. 110 are the project's arrangements |
| 4 | The object at "letting go" | The cloak given by Xichun (backstage) | At the end of Rec. ch. 105 the scarlet felt cloak is fixed as something against the cold, not a monk's robe | In the text | No independent reason |
| 5 | The ground of "grief is in living on" | The backstage's own inference | Baoyu neither dies nor becomes a monk; the book closes in Rec. ch. 110 with him teaching his son by lamplight | In the text | The criterion (overview §6.5, "Grief is in living on") |
| 6 | Xiren's ending | Lin Gudu as prototype (source not found); materials from continuations | In Rec. ch. 81 she marries Jiang Yuhan at the Flower Festival; the title of Rec. ch. 106 reads "Hua Xiren is indeed constant from beginning to end" | In the text | The heading "Hua Xiren, constant from beginning to end" recorded in commentary 20.20, which supports only the direction of her ending; the Flower Festival, the marriage and other details have no independent reason |
| 7 | Jia Qiang and Lingguan | Lin Xue and Zheng Bangxiang as prototypes (Liao Yihu); "from feeling in vain to feeling fulfilled" (情虚改情融; the backstage's own inference) | Rec. ch. 99 recalls Lingguan striking Jia Qiang; in Rec. ch. 109 he marries her with full ceremony, in the Sanhe troupe | In the text; the Fuzhou thread withdrawn | The episode in the first eighty chapters in which Lingguan traces the character qiang 蔷, which supports only the continuation of their relationship; the striking and the marriage in the Sanhe troupe are the project's extension |
| 8 | Yuanyang | Dong Xiaoshuang as prototype; the mapping onto 1627 and 1628 | In Rec. ch. 84 Yuanyang dies following her mistress | In the text; the mapping of years not kept | Yuanyang's vow in chapter 46 ("I'll cut off my hair and become a nun; and if not that, there's still death"), which supports only the direction; taking the path of death, and setting it in Rec. ch. 84, are the project's choices |
| 9 | Yuanchun | Ye Xianggao as prototype (Liao Yihu); the backstage's retitling of "Lament of Cangxia" (苍霞悲歌; Cangxia as in Ye Xianggao's Cangxia cao 苍霞草) as "Lament of the Torn Net" (裂网悲歌); the drafts of the reconstruction have only ever used the latter | In Rec. ch. 91 Yuanchun dies, in the same chapter as the confiscation | In the text | 18.115, Gengchen double-line interlinear, "The Palace of Eternal Youth foreshadows the Imperial Consort's death," which supports only Yuanchun's death; its falling in the same chapter as the confiscation, and the rest, are the project's arrangements |
| 10 | Carrying out the confiscation | Qing law and legal anachronism (the backstage's own inference) | In Rec. ch. 91 the Embroidered Guard draws up the inventory, and so on | In the text | A constraint on the writing, like the Kangxi-era ground |
| 11a | The Tianxiang Tower not supplied | "A deletion is a blank to be left" (the backstage's own inference) | The deleted passage is neither supplied nor written | In the text | The inviolable (overview §3.5) |
| 11b | Rereading Qin Keqing | Cao Xueqin did not himself accept the "lewdness" given her by the Mirror for the Romantic layer (the project's conjecture, overview §3.5) | Baozhu's account in Rec. ch. 104, Qin Keqing's dream visit in Rec. ch. 105; in the Roster of Rec. ch. 110, "yet her sin was not beauty but politics"; she comes from the line of the deposed crown prince | In the text; not load-bearing (overview §1.4) | The verdict is a pronouncement made at the moment of chapter 5; the project rereads it by the criterion, recording it as "the world says" and then turning it over (overview §3.5). If the conjecture about whose layer it was falls, this rereading is left with only this one ground. Agrees with Liu Xinwu's view; similarities and differences in paper 8 |
| 12 | "What is kept is acknowledged" | The backstage's own inference | Grandmother Jia's sentence in Rec. ch. 83; the old draft's "empty frame" changed | Wording changed; this account left in the backstage | So that Grandmother Jia does not foresee the confiscation; this paper cannot state which place the former author kept (§6.3) |
| 13 | Zhen Baoyu delivering the jade | Bao Juechu as prototype (Liao Yihu); the story that Zhen Baoyu becomes a monk | In Rec. ch. 96 Zhen Baoyu, after delivering the jade, stays on in the Jia household and sets his heart on the examinations; he brings no Yushi xinbian, and is not a monk | Partly in the text; the monk story not in the final text | 18.116, Gengchen double-line interlinear, "The Handan Dream foreshadows Zhen Baoyu delivering the jade," which supports the delivery itself; his staying on and setting his heart on the examinations are the project's arrangements |
| 14 | Bao Yong | A second projection of Bao Juechu | Cut from the whole book | Withdrawn | — |
| 15 | Lady Wang's death | The double reading of "blocking the scissors" | In Rec. ch. 98 Jia Zheng dies of anger, and Lady Wang follows him in death | The plot in the text; the double reading withdrawn | No independent reason |
| 16 | Baochai | Xie Zhaozhe as prototype (Liao Yihu); "extension of personal names" (the backstage's own inference) | In Rec. ch. 107 she marries Yucun; in Rec. ch. 108, confiscation, exile and death | The ending in the text; "extension of personal names" left in the backstage | Jia Yucun's line in chapter 1, "the hairpin in the casket waits its time to fly" (钗于奁内待时飞), and 1.141, Jiaxu side note, "Daiyu having been presented, Baochai follows at once," are materials the project chose to read in this way, and on them it arranges her marriage to Yucun; the notes do not say so. This reading cannot be satisfied together with the chapter-head note to chapter 28, "Baochai … together to the end" (宝卿得同终始; the note speaks of Baoyu and Baochai); the overview (§6.4) records this as it stands, as soft evidence not accommodated. The confiscation, exile and death in Rec. ch. 108 are a separate plot arrangement of the project's |
| 17 | Xiangling | Xu You as prototype (source not found) | In Rec. chs. 107 and 109 she recovers the memory of her past and returns to her family; her Roster verdict in Rec. ch. 110 reads "though on two plots a lone tree grows, while this body lives it goes home" | In the text | Driven mainly by the criterion; the prototype only a loose association (as the backstage records explicitly) |
| 18 | Ping'er, Tanchun, Baoqin, Feng Ziying | Once linked in the backstage to the prototypes Ni Fan, Shao Jiechun, Xu Bi and Ruan Zihua | Ping'er's apricot tree in Rec. ch. 109; Tanchun married far away, to Nanming, in Rec. ch. 85; Baoqin married far away, and Feng Ziying going overseas to the South Seas | In the text | Author's statement (§6.1): the prototypes had no influence on the writing of the current draft; the plot choices do not rest on the prototypes, and no other reason is recorded explicitly |
| 19 | Xiangyun and Zhang Xie | Zhang Xie as prototype (Liao Yihu) | Rec. ch. 110: "Rhapsody on My Humble Lodging," "I once read the Dongxiyang kao" | In the text | No independent reason; the prototype is the source. As far as the prototype identifications of the Liao Yihu line go, the only item in the current draft that influenced the writing |
| 20 | Kong Shangren | Kong Meixi as Kong Shangren (Zhang Zhijian) | Rec. ch. 109: the Sanhe troupe stages The Peach Blossom Fan; Kong Shangren's name does not appear in the text | In the text | No independent reason |
| 21 | Daiyu's death | Lin Chong as prototype (Liao Yihu); Ye Shaoyuan's Yaowen | She dies in Rec. ch. 95, without burning her poems; in Rec. ch. 96 envoys from the Palace of the Moon come to meet her | In the text; Lin Chong only in the backstage | 18.117, Gengchen double-line interlinear, "The Peony Pavilion foreshadows Daiyu's death," which supports only Daiyu's death; the date of her death, the poems not burned, and the dream of Rec. ch. 96 (§6.4) are the project's arrangements |
| 22 | The Prison Temple | Commentary and materials from continuations | The Prison Temple line of Rec. chs. 92–94 | In the text | The "Prison Temple" recorded in commentary 20.20 and 26.24 (paper 1), which supports the Prison Temple episode; distributing it over Rec. chs. 92–94, and the way the rescue is written, are the project's realization |
| 23 | Shishu and Daishu | A theory of taboo avoidance (the backstage's own inference) | The final text uses "Shishu" (侍书) | In the text (choice of character) | No independent reason |
| 24 | Against mourning the Ming, against amnesty | A symmetry of selection between the two texts (the backstage's own inference) | The book writes no roving bandits and no collapse of the realm, and no "the orchid and the cassia flourish together"; Baoyu neither becomes a monk nor takes his own life | In the text | The discipline of acceptance in overview §4.6 (political passages removed; the frame of mourning the Ming not adopted); for Baoyu neither becoming a monk nor taking his own life, see item 5 |
Appendix 5 Items to Be Checked
To be completed before the next version. What cannot be found will be marked as such, and nothing will be supplied merely to complete the list.
- Scholarship:
- Dai Bufan: the first publication of "On Qin Keqing's Late Death"; the page ranges of the three articles in Hongxue pingyi: waipian; an article-by-article list of the rebuttals of 1979–1980, following the table of contents of the Collected Papers; the author of the article "The Commentary Is Iron-Clad Proof" (脂批就是铁证). - Wu Shih-ch'ang: p. 483 of Honglou meng tanyuan waibian; chapter and page of the English monograph on "Tangcun's short prefaces." - Zhu Danwen: the original text of pp. 197 and 200 of Honglou meng lunyuan. - Chao Kang and Chen Chung-yi: the editions of their two books. - Du Chungeng: the issue in which "The Story of the Stone Could Not Have Evolved out of A Mirror for the Romantic" appeared. - Cai Yijiang: the original source of the phrase "suspecting the neighbour of stealing the axe." - Hu Shi: the volume and issue of "New Materials for the Study of Honglou meng" in Xinyue. - Hawkes: page numbers of the introduction to the first Penguin volume. - Whether scholars have made a similar argument from the Western objects in the book to the period in which the story is set or was written (for example Fang Hao). The project's reading does not depend on it; if such work is found, it will be reported.
- Inconsistencies within the text: the remaining examples, to be checked case by case against the base text; the reasons given by proponents of the chapter 13 example; the earliest proponents of the examples of duplicated persons, of genealogy, and of Dong'an and Dongping.
- Outside records: Mingyi's poems, to be checked again in Honglou meng juan, compiled by Yi Su; the two quotations from Zaochuang xianbi, "Zhiyan" and "Zhiyan Zhai"; a nameable transcription of the sentence "accounts of the author differ" in the Chengjia preface, and the page of the Chengjia edition itself; the page of the Chengyi foreword; the original collections of the poems of Duncheng and the others, including the source of the phrase "writings left of ox-demons."
- Variants between manuscripts in the chain of titles: whether the Wu Yufeng sentence and the sentence on the second commentary of the jiaxu year are peculiar to the Jiaxu manuscript; first find the collation notes themselves, then check manuscript by manuscript.
- Popular theories: Liao Yihu's Zhihu pieces; the four prototypes recorded in the backstage as Liao Yihu's; the first proponents and sources of the theories on Kong Meixi, Wu Yufeng and the "Western Hall"; the proponents of the theories of the Ming royal house; the source and date of Yan Zhong's count.
- Versions of the reconstruction: Rec. chs. 96–103 to be based on the round 54 sealed drafts, and Appendix 4 spot-checked again.
Appendix 6 Differences from the Overview v0.8
In the following places this paper is authoritative, and the next version of the overview will be revised accordingly (Chinese and English together):
- §3.4, "Its main thesis depends on one point only: that the book is layered, and that the one who finally fixed its centre was Cao Xueqin": to be revised following §1 and §7 of this paper, to read "the main body of the text of the surviving first eighty chapters went through layering, and the one who integrated it into a book and fixed its centre was the Cao Xueqin named in the chain of titles and in the commentary"; with a note on the known exceptions (the end of chapter 22 described as "unfinished, damaged and patched," not as "a supplement after Qin [Xueqin]'s death"), and a note that "fixed its centre" is a working judgement, not argued in this paper.
- §3.5, "an upper-margin note signed 'Old Tablet' (Jihu Sou): 'That the whole chapter hides how Keqing died is my own great act of compassion'": the text quoted is the Jing upper-margin note of 13.21, and the manuscript should be named, as "Jing upper-margin note." Add a note that the Gengchen note 13.90 records a shorter text to the same effect ("a great act of the compassionate heart," dated "spring, renwu," not signed Hu Sou, "Old Tablet"), so that it is not said to be found only in the Jing.
- §3.5, "Chapter 13 carries three notes": with reference to §3.2 of this paper, add 13.87, "this chapter has only ten pages, because with the Tianxiang Tower episode deleted, four or five pages are gone," 13.51, "this supplies undeleted text of the Tianxiang Tower," and the Gengchen note 13.90, so that the page count of the deleted draft no longer rests on the Jing manuscript alone.
Appendix 7 Acknowledgements and Use of AI
I thank Zesi Chen and Zhen Zhang for a great deal of analysis and help in the course of the project.
The project used AI assistance from several providers, including Claude (Anthropic), ChatGPT (OpenAI), Gemini (Google), DeepSeek, GLM (Zhipu), Meta AI and Grok (xAI). The research for this paper, the checking against the base text and the drafting were carried out by AI according to criteria the author set. A further round of searching for scholarly and popular theories was made with ChatGPT Deep Research; its report served only as a record of where to look for sources, and every citation in this paper leads to a book, article, web page or image that was actually read. The outline and the main text were both reviewed. The English version is a rewriting of the Chinese version for English-speaking readers, drafted with AI assistance and checked against the Chinese; where the two differ in substance, the Chinese is authoritative. Every ruling was made by the author, who bears responsibility for the text.
References
Only works actually read, or whose records have been checked, are listed; for reliability see Appendix 1. For online reprints and reports, the place where they were read is given.
Base text and this series
- Han Qin 秦汉. People Have Qing: Origins, Method and Conclusions of an Interpretive Reconstruction of the Last Thirty Chapters of Dream of the Red Chamber. Pre-release v0.8. Zenodo. Concept DOI: 10.5281/zenodo.22985375.
- Han Qin 秦汉. An Audit of Forward References in the Zhiyan Zhai Commentary on the First Eighty Chapters of Dream of the Red Chamber: What Is Adopted and What Is Not. Pre-release v0.4. Zenodo. Concept DOI: 10.5281/zenodo.23007235.
- Jiangshange Yin 江山阁隐, maker. Honglou meng zhipi ben wanmei ban 红楼梦脂批本完美版 (Honglou meng with the Zhiyan Zhai Commentary: The Perfect Edition). Electronic edition circulating online (PDF; according to its maker, based on the Gengchen manuscript, incorporating commentary from the Jiaxu, Wangfu, Qixu and other manuscripts). The "project base text" of this series.
- Cao Xueqin 曹雪芹. Zhiyan Zhai chongping Shitou ji 脂砚斋重评石头记. Manuscripts: Jiaxu, Jimao, Gengchen, Mengfu (Mongol Prince's Mansion), Qixu (Qi Liaosheng preface); commentary of the Jing manuscript (original lost; known from transcriptions).
Studies
- Beifang luncong editorial office 《北方论丛》编辑部, ed. Honglou meng zhuzuoquan lunzheng ji 《红楼梦》著作权论争集. Taiyuan: Shanxi renmin chubanshe, 1985 (catalogue).
- Cai Yijiang 蔡义江. "Honglou meng shi zenyang chengshu de?" 《红楼梦》是怎样成书的？ Reprint, Aisixiang.
- Chang, Eileen (Zhang Ailing 张爱玲). Honglou mengyan 红楼梦魇. Taipei: Huangguan, 1977. Online text; the preface to Mingyi's poems quoted in §3.5 is from the essay "Er xiang Honglou meng" 二详红楼梦 (Nunu Shufang).
- Du Chungeng 杜春耕. "Rong Ning liang fu liang ben shu" 荣宁两府两本书. Honglou meng xuekan 1998, no. 3. Reprint, Aisixiang.
- Du Zhijun 杜志军. "Xin shiqi hongxue sanshi nian: 1978–2008" 新时期红学三十年：1978—2008. Honglou meng xuekan 2009, no. 2. Reprint, Aisixiang.
- Feng Qiyong 冯其庸. "San lun Gengchen ben" 三论庚辰本. Reprint, Sohu.
- Gao Shuwei 高树伟. Honglou meng Jingcang ben bianwei 红楼梦靖藏本辨伪. Beijing: Zhonghua shuju, 2024. ISBN 9787101165395 (catalogue).
- Hanan, P. D. Review of Wu Shih-ch'ang, On The Red Chamber Dream. Bulletin of the School of Oriental and African Studies 24.3 (1961): 603–605.
- Hu Lianhao 胡联浩. "Honglou meng chengshu cong 'Jinling' dao 'Chang'an' kaobian" 《红楼梦》成书从“金陵”到“长安”考辨. Honglou meng banben yanjiu jikan 红楼梦版本研究辑刊 4. Huaqiao chubanshe, 2024. Reprint, Tencent News.
- Hu Shi 胡适. "Honglou meng kaozheng" 红楼梦考证. Reprint, Aisixiang (the source of the "pedlar's load" sentence of Cheng Weiyuan's preface quoted in §3.5).
- Li Qingyu 李清宇. "Cong Honglou meng timing lun Cao Xueqin xiaoshuo chuangzuo guannian de yanjin" 从《红楼梦》题名论曹雪芹小说创作观念的演进. Wenyi lilun yanjiu 2014, no. 6.
- Liang Yusheng 梁羽生, on Wu Shih-ch'ang. Online text at 99csw: https://www.99csw.com/book/5551/195513.htm (title to be supplied).
- Lin Yi 林沂 and Cai Zhengxing 蔡政星. "Honglou meng banben lunzheng zongshu" 《红楼梦》版本论争综述. Originally in Ming Qing xiaoshuo yanjiu. Reprint, Sohu.
- Liu Mengxi 刘梦溪. "Cao Xueqin he ta de wenxue shijie" 曹雪芹和他的文学世界. Reprint, Aisixiang.
- Mei Jie 梅节. "Ye tan Jingben" 也谈靖本. Honglou meng xuekan 2002, no. 1. Reprint, Aisixiang.
- Shen Zhijun 沈治钧. "'Xin Baoyu' he 'jiu Baoyu': Honglou meng chengshu guocheng shitan" “新宝玉”和“旧宝玉”——《红楼梦》成书过程试探. Reprint, Aisixiang.
- Tan Dejing 谭德晶. "Honglou meng yu Fengyue baojian zhi guanxi zai tan" 《红楼梦》与《风月宝鉴》之关系再探. Reprint, Gudai xiaoshuo wang.
- Wu Shih-ch'ang 吴世昌. "Lun Mingyi suo jian Honglou meng chugao" 论明义所见《红楼梦》初稿. Honglou meng xuekan 1980, no. 1. Reprint, Aisixiang.
- Wu Shih-ch'ang. On The Red Chamber Dream: A Critical Study of Two Annotated Manuscripts of the XVIIIth Century. Oxford: Clarendon Press, 1961 (catalogue).
- Xia Wei 夏薇. "Cong Fengyue baojian dao Honglou meng: chengshu yu chuangzuo sixiang de shanbian" 从《风月宝鉴》到《红楼梦》——成书与创作思想的嬗变. Wenxue pinglun 2022, no. 6. Reprint, Gudai xiaoshuo wang.
- Zhan Dan 詹丹, on dialect and the dispute between Dai Bufan and Chen Xizhong and Hou Zhongyi. Guangming Online, 2019-08-17.
- Zhou Jianyu 周建渝. "Shitou ji de xushu cengci ji qi gongneng yu yiyi" 《石头记》的叙述层次及其功能与意义. Zhongguo wenhua yanjiusuo xuebao 中国文化研究所学报 (Journal of Chinese Studies) 58 (2014).
- Zhou Shaoliang 周绍良. "'Xueqin jiu you Fengyue baojian zhi shu'" “雪芹旧有《风月宝鉴》之书”. Honglou meng xuekan 1979, no. 1. Reprint, Aisixiang.
Popular writings and reports
- Liao Yihu 廖一壶. "Xu Xinggong yu Honglou meng" 徐兴公与《红楼梦》, column on Douban Read (sample chapters); Bilibili column "Shitou ji de zuozhe Xu Xinggong, pingdianzhe Zhiyan Zhai Jihu Sou, xuxiezhe" 《石头记》的作者徐兴公、评点者脂砚斋畸笏叟、续写者, 2022-03-17.
- "Xuezhe: Honglou meng zuozhe shi Fang Yizhi, shu yongyi 'diao Ming fan Qing'" 学者：红楼梦作者是方以智 书用意“吊明反清”. China News (chinanews.com), 2011-10-31.
- Zhang Zhijian 张志坚. "Guanyu Honglou meng (Zhiping ben) zhong de 'Meixi'" 关于红楼梦（脂评本）中的“梅溪”. Sina blog, 2011-03-17.
- Tumore 土默热. Tumore hongxue: Honglou meng chuangzuo zhenxiang yu zuopin zhendi xintan 土默热红学：红楼梦创作真相与作品真谛新探. Jilin renmin chubanshe, 2005. Pang Guanqing 逄冠卿. "Tumore hongxue de 'shi lun'" 土默热红学的“十论”. Hangzhou ribao 杭州日报, 2016-08-12.
- Fu Bo 傅波 and Zhong Changshan 钟长山. "Honglou meng zuozhe: Wu Meicun" 《红楼梦》作者：吴梅村. Reprint, Sina blog "Lao Kang," 2012-04-30 (original publication to be checked).
- Mao Lianquan 冒廉泉. Xuan'an bu xuan: Mao Pijiang zhu Honglou meng qishisan zheng 悬案不悬：冒辟疆著《红楼梦》七十三证. Huaxia chubanshe, 2016 (catalogue).
- Chen Lin 陈林. Poyi Honglou shijian mima 破译红楼时间密码. Jiangsu meishu chubanshe, 2006 (catalogue).
- Huo Guoling 霍国玲 et al. Honglou jiemeng 红楼解梦. Dongfang chubanshe (catalogue).
- Liu Xinwu 刘心武: book excerpts and interviews. Guangming Online, 2003-08-20 and 2006-01-01.
- Yan Zhong 严中 on the "authors" proposed in Redology, China News, 2015-11-30; Peng Guangqian 彭光谦, The Paper, 2022-03-01; Ouyang Jian 欧阳健, lecture at Hanzhong, 2015, published by Aisixiang.