9. Worked Analyses

Eight close readings and a revision case. The analytical apparatus is the same throughout: name the relation between the chains in Levin’s terms (§3.2); audit the domains served (§4); locate the convergence point by the kernel test (§3.4, §7.1 test 5); and describe the braiding structure and rhythm (§6.4, §6.6).

Where a work is read in translation the translator is named, and the distinction stated in this book’s Introduction holds: structural claims survive translation, texture claims do not. Everything argued below about plot architecture would be argued from any competent translation. Nothing argued below depends on a sentence’s rhythm.

9.1 Eliot, Middlemarch (1871–72) — the web, and the definitive case

The threads. Four, none of them subordinate: Dorothea Brooke and Casaubon, and later Ladislaw; Tertius Lydgate and Rosamond Vincy; Fred Vincy and Mary Garth; Nicholas Bulstrode and his concealed past. A fifth, the Featherstone inheritance, functions as a shared machine rather than as a chain of its own.

A fact of composition that settles the theoretical question. The Lydgate material and the Dorothea material began as separate works. Eliot was writing a novel about a provincial town, with a young doctor at its centre, and separately began a story called “Miss Brooke”; she fused them, and the fusion is the novel. This is the strongest available evidence for the chapter’s central claim about relation: the two chains were not derived from one another, they were set beside one another, and everything that makes Middlemarch one book was constructed after the decision to combine.

Levin’s relation: analogy, at the level of terms. The chains are not causally dependent. Dorothea’s marriage would proceed unchanged if Lydgate’s did not exist, and the reverse. What binds them is a question posed four times with four different answers: what happens to a person who wants their life to be of consequence, in a place that will not fund it?

Domains served by the Lydgate thread. Thematic counterpoint at maximum (it is the novel’s second and harshest answer to its own question); character testing ground for the town rather than for a person — Lydgate is the instrument by which Middlemarch is tested; world expansion (the medical profession, hospital politics, the economics of provincial reform); temporal counterpoint (Lydgate’s ruin is slow, accumulating in small debts and small compromises, against the sharper crises of Dorothea’s marriage); and structural convergence, which arrives late and at a single point.

Convergence, located by the kernel test. Apply test 5: where would deleting the Lydgate thread break the main plot’s causal logic? The answer is Chapter 81, when Dorothea goes to Rosamond to plead Lydgate’s innocence and Rosamond, unexpectedly, tells her the truth about Ladislaw. Dorothea’s marriage to Ladislaw depends on that conversation, and that conversation depends on the Lydgate thread having produced a scandal, a wife and a reason for the visit. Everything before it is analogy; at Chapter 81 the second chain’s kernels become the first’s.

Note what this means arithmetically. The Lydgate thread occupies something close to a third of the novel and converges causally on one afternoon. No modern structural edit would permit that ratio, and the ratio is the novel.

Braiding. The repeated intersection (§6.4, structure 5) executed with unusual density. The threads touch constantly and lightly: Lydgate attends Casaubon and delivers the diagnosis that shapes Dorothea’s marriage; Farebrother carries messages between the Vincy and Garth chains; Bulstrode’s money touches everyone; Featherstone’s will draws four families to one house. None of these touches resolves anything. Their cumulative effect is a world in which everything is connected weakly — which is the argument.

Eliot’s own theory. The narrator’s account of her method in Chapter 15 is the best short statement of multi-plot form in English: she says that she has so much to do in unravelling certain human lots, and seeing how they were woven and interwoven, that all the light at her command must be concentrated on this particular web, and not dispersed over that tempting range of relevancies called the universe.

Three things follow from the figure, and they are the reason this passage is quoted in every serious discussion of the multi-plot novel.

1. A web has no trunk. The image excludes hierarchy. There is no main strand and no subordinate ones; there are crossings, and the crossings are where the material is.
2. The web is a limit, not an expansion. The sentence’s rhetorical work is renunciation: this web and not the universe. Eliot is describing a multi-plot novel as an act of exclusion, which is the opposite of the way expansive nineteenth-century structures are usually characterised.
3. The metaphor is a theory of the attention tax with the sign reversed. Attention concentrated on the relations between lots, rather than distributed across separate lots, is what makes the concentration affordable.

The Finale, and off-page convergence. Eliot resolves several threads by report (§6.4, structure 12), and the compression is deliberate: the Finale’s business is to place each life in a duration longer than the novel, which is temporal counterpoint (§4.6) executed in the last four pages. Chapter 16 (The Ending) treats the Finale in full.

9.2 Tolstoy, Anna Karenina (1875–77) — the convergence that is refused

Read in Constance Garnett’s translation; the structural account holds across Garnett, the Maudes, Pevear and Volokhonsky, Bartlett and Schwartz.

This is the chapter’s central example, because it breaks the rule that a subplot must meet the main plot, and is better for it.

The threads. Two, of comparable weight: Anna, Vronsky and Karenin; Levin and Kitty. Around them a third, weaker chain — Stiva and Dolly — which functions as connective tissue (§8.8).

What actually connects them, stated accurately. Craft accounts of this novel routinely assert that the two plots never touch. That is not true, and the truth is more instructive.

The Part Seven visit is the detail that repays attention, because it is a meeting without convergence — a structure that a model requiring intersection has no name for. The two protagonists occupy the same room for an evening, and the novel’s causal architecture is entirely unaffected. Tolstoy engineers the encounter and then declines to use it, which is a far stronger statement than keeping them apart would have been: he demonstrates that contact is not the same as connection.

The relation: analogy, sustained over eight parts. The chains ask one question — what does it cost to want to live truthfully? — and answer it twice.

Domains. Thematic counterpoint, unambiguously and at the highest level available to the form. Temporal counterpoint (§4.6): Anna’s chain accelerates towards catastrophe while Levin’s proceeds at the rate of harvests, a wedding, a birth, a death, a haymaking — two clocks, and the difference in clock is a large part of the difference in fate. Character testing ground, in the sense that each chain tests what the other’s premises would produce. Structural convergence: none after Part One, and this is the point.

The critical disagreement, presented rather than resolved. Whether the two plots are unified has been argued for a century and a half, and competent critics still disagree.

• The case that it fails. The chains are causally independent after Part One; the novel could be published as two books; the Levin material’s agricultural and philosophical passages digress from any plot at all; the reader’s investment is repeatedly interrupted at the worst possible moments. Contemporary reviewers, including some sympathetic ones, made versions of this complaint.
• The case that it succeeds. The unity is thematic in Levin’s exact sense (§3.2) — analogy carried by an enormous number of small rhymes, marked so persistently that no reader can miss the comparison even while being unable to state it. Tolstoy himself, in a well-known letter responding to this criticism, insisted that the connection lay not in plot or in the relations of the characters but in an inner one, and compared the book’s construction to a vault whose keystone cannot be identified.
• A third position worth putting to a seminar. That the question is malformed, because it applies a unity criterion derived from drama to a form Tolstoy was deliberately constructing against drama. He denied that the book was a novel in the European sense; the refusal of convergence may be a refusal of the whole apparatus that makes convergence seem necessary.

Braiding. Alternating, at part and chapter scale, with the crossings distributed evenly enough that neither thread goes dormant beyond §6.6’s second band. This is what makes the analogy work: Empson’s requirement of proximity (§3.3) is met continuously, so that the comparison is always live in the reader’s memory. A novel with this convergence structure and War and Peace’s block braiding would fail. The thematic-only relation is affordable only because the braid is tight.

The lesson for the chapter. The usual rule requires convergence and permits it to be thematic. Anna Karenina shows that thematic convergence is not a lesser option granted on sufferance but a structure with its own demands — dense braiding, marked analogy, shared world, and a willingness to let a meeting happen and produce nothing. The rule is not converge. The rule is relate, and pay for the relation in the currency it takes.

9.3 Dickens, Bleak House (1852–53) — three systems and one secret

The threads. Three, and they are systems rather than storylines, which is what makes the novel the most instructive convergence structure in English.

1. Chancery. Jarndyce and Jarndyce; Richard, Ada, Miss Flite, Gridley, Krook, Vholes. A suit that consumes lives and finally consumes itself in costs.
2. Chesney Wold. Sir Leicester and Lady Dedlock, Tulkinghorn, Hortense, Rouncewell. A marriage built over a concealment.
3. Esther. Her origin, her illness, her face, her marriage; and Jo, Charley, the brickmakers, Tom-all-Alone’s, which belong to her narrative and are the novel’s third social level.

Levin’s three-level structure, in full (§3.2). Aristocratic, professional-middling, destitute — with the relations of analogy running down the ladder, and one secret running through all three. Levin’s Renaissance dramatists could not have asked for a cleaner instance.

The convergence apparatus. Everything runs on one fact: Esther is the daughter of Lady Dedlock and the law-writer known as Nemo. The mechanism by which it is planted is the finest single piece of construction in Dickens.

In Chapter 2, Tulkinghorn brings an affidavit to Chesney Wold. Lady Dedlock asks who copied it; the question is the merest social nothing; and she nearly faints. A reader has, at that moment, no idea there is anything to recognise. Handwriting is the plant, and it is planted before the existence of the secret is known — which is why the convergence, when it arrives four hundred pages later, feels discovered rather than arranged. Compare §8.9: the coincidences that follow are enormous, and they are underwritten by this one act of preparation.

Convergence by cascade (§6.4, structure 6). The threads do not converge together. They converge in sequence, and the sequence accelerates:

Domains. All six, which is what a novel of this length permits. World expansion is the most conspicuous — Jo and Tom-all-Alone’s exist so that a claim about social connection can be made as narrative. Temporal counterpoint (§4.6) is the least noticed and structurally the most interesting: Jarndyce and Jarndyce runs on a generational clock — it began before the novel and has consumed a lifetime — while Esther’s narrative runs on months. The suit’s slowness is not a background fact; it is the second clock against which every human urgency in the book is measured, and it is why the novel’s argument about institutions can be made without an argument being stated.

Viewpoint and thread together. This is the one close reading in the chapter where the allocation of viewpoint and the allocation of plot cannot be separated. Esther narrates her thread in the first person and in retrospect; the other two are given to a present-tense panoramic narrator who can be in Lincolnshire and in Chancery in the same paragraph. The two narrations are themselves a double plot: an inside and an outside, a person and a system. Chapter 13 (The Viewpoint Shift) treats the dual narration as a viewpoint structure; the point here is that the viewpoint division and the thread division are the same division, and that this is why Bleak House can hold three systems without confusion.

9.4 Tolstoy, War and Peace (1865–69) — braiding at maximum scale

Read in the Maude translation; the structural account is translation-independent.

The threads. Four families — Bolkonsky, Rostov, Bezukhov, Kuragin — carrying perhaps a dozen traceable chains, plus the historical narrative of the campaigns of 1805 and 1812, plus a non-narrative strand of historiographical argument that expands until it occupies the Second Epilogue entirely.

Why it is not confusing. §5.4 gives the general answer; the particulars are worth listing, because they are the transferable craft.

1. Threads are anchored to families, which readers hold as single objects.
2. The crossing characters are numerous. Pierre has business with all four families; Andrei with three; the network has no isolated nodes and therefore no thread that can be forgotten.
3. Crossings are frequent and short, keeping every frame warm (§6.6).
4. Threads are differentiated by register. A paragraph of Bolkonsky austerity, Rostov warmth or Kuragin appetite identifies its thread before a name appears — the memory subsidy of §6.6.

The relation between the two halves of the title. War and peace are not main plot and subplot. They are a temporal counterpoint (§4.6) organised as a moral argument: the war chapters run on the clock of a campaign, in which everything is decided in hours by people who do not know what is happening, and the peace chapters run on the clock of a life, in which everything is decided over years by people who also do not know what is happening. The rhyme between the two is the novel’s thesis about historical causation, and it is made by the alternation rather than by the essays.

Where the essays fail and why they are there. The historiographical passages have no chain, no goal and no convergence, and by §2.1 they are digression (§8.11). Tolstoy’s own justification is that the novel’s narrative demonstrates something the discipline of history cannot state, and that the statement must therefore be made directly. This is the honest position for a chapter like this one to take: the essays fail every structural test and are not thereby shown to be wrong, only to be doing something narrative does not do. Readers have disagreed about them since the first serial instalments, and the disagreement is set as a seminar question at §12.2.

The exception this novel supplies to the chapter’s own rules. Platon Karataev, discussed at §8.5, arrives in the fourth volume and carries the resolution. The general principle extracted there: a late arrival can discharge an early investment when the investment was in a question rather than in a thread.

9.5 Thackeray, Vanity Fair (1847–48) — convergence by contrast

The threads. Two, established in the first chapter as a matched pair and never allowed to separate for long: Becky Sharp, who has nothing and wants everything; Amelia Sedley, who has everything and wants one thing that is not worth having.

Levin’s relation: contrast, executed as a formal system. The two chains are constructed as inversions at every level — origin, temperament, marriage, motherhood, relation to money, relation to truth. This is the most schematic use of a double plot in nineteenth-century English fiction, and the novel’s achievement is that the schema breaks under its own operation.

Domains. Thematic counterpoint at the level of outcome and of terms simultaneously; world expansion (Becky’s chain reaches into places Amelia’s cannot — the Continent, the gaming table, the court, the demi-monde); character testing ground reciprocally, since each woman is the pressure applied to the other; and one decisive structural convergence.

The convergence. Located by test 5: near the end, at Pumpernickel, Becky produces the note George Osborne wrote to her years earlier, proposing that they run away together, and gives it to Amelia. Amelia’s idolatry of her dead husband collapses; she is released to marry Dobbin. Delete the Becky thread and Amelia’s chain cannot resolve — she has no other route out of the sentiment that has held her for six hundred pages. That is the kernel test returning a page number, and it is a beautiful piece of construction: the object that converges the plots is a piece of paper generated by the subplot, kept for twenty years, and deployed as an act of malice that functions as a rescue.

The hijack, converted. §8.3 treats this novel as the great counterexample to the main-plot hijack. The structural note to add here is that the hijack is managed by the narrator. Thackeray’s showman-narrator continually draws attention to the reader’s own preference for Becky, and the novel’s subtitle — A Novel without a Hero — is the formal declaration that neither chain will be permitted to be the moral centre. A double plot with no primary thread is a structure that requires a narrator willing to say so.

9.6 Hugo, Les Misérables (1862) — the subplot’s limit case

Read in the Denny and Donougher translations; the structural account is translation-independent, and the digressions are complete in Donougher and abridged in some earlier English versions — a fact students should check before arguing about proportions.

Why it is here. Because a chapter that could not explain Les Misérables would have described a convention rather than a mechanism. The novel contains at least four kinds of secondary material, and they need to be separated before anything useful can be said.

Waterloo, and the paragraph that changes the argument. The Waterloo book occupies a substantial stretch of Part Two and describes a battle at which no principal character is present. It fails test 1 (no chain), test 2 (no kernels), test 3 (nothing the narrative could not supply), test 5 (deleting it breaks nothing) and test 8 (it is a single uninterrupted block).

And then, in its closing pages, a man moves across the field robbing the dead, lifts an officer clear of a heap of bodies in the course of stealing from him, and is thanked. The officer is Marius’s father; the thief is Thénardier; and the obligation created by that misunderstanding governs Marius’s entire relation to the Thénardiers for the rest of the novel, including his refusal to see what they are.

A hundred pages of history are converted, in a paragraph, into the load-bearing joint of a subplot. Whether that redeems the hundred pages is a real disagreement, and the chapter declines to settle it. Two positions:

• It does not. A single kernel does not justify a proportion of the book; the same kernel could have been planted in two pages, and the digression’s actual function is authorial — Hugo wished to write about Waterloo.
• It does, and the terminal placement is the argument. The point of the digression is that the great historical event and the squalid theft are the same event seen at two magnitudes, and that the private consequence of Waterloo, for this novel, is a lie told by a thief. Compress the history and the disproportion — which is the meaning — disappears.

The sewers work differently and more defensibly. The disquisition on the sewers of Paris precedes Valjean’s carrying of the unconscious Marius through them. It is a descriptive pause functioning as preparation: the reader traverses the described space before traversing it with him, and the passage’s exhaustiveness becomes duration. This is the clearest case in the tradition of a digression paid for by what follows.

The general principle the novel yields. A digression is not a subplot and cannot be assessed as one. It is judged on whether the book has an argument that narrative cannot make, and on whether the writer is one of the small number who can hold a reader through an argument. The tests in §7 will condemn every digression ever written; that is a fact about the tests, and §16 says so.

9.7 Dostoevsky, The Brothers Karamazov (1880) — the thread that takes the ending

Read in the Pevear and Volokhonsky translation; also available in Garnett, McDuff and Avsey.

The thread. Ilyusha Snegiryov, his father the disgraced captain, and the schoolboys — Kolya Krasotkin above all. It occupies several books of a very long novel and appears, on a first reading, to be the least connected material in it.

Its causal origin, which is exact. Dmitri, in a rage, drags Captain Snegiryov out of a tavern by his beard in front of his son. Ilyusha, humiliated on his father’s behalf, bites Alyosha’s finger to the bone — Alyosha being the only Karamazov available. The subplot begins as a consequence of the main plot’s central conflict, which is early convergence (§6.4, structure 3) in its most economical form: the thread is causally generated rather than introduced.

Domains. Character testing ground, primarily and superbly (§4.2): the parricide plot gives Alyosha nothing to do — he mediates among adults whose positions are fixed — while the boys give him a situation in which his elder’s instruction about active love can be attempted and can fail. Thematic counterpoint: the novel’s great question is whether suffering can be justified, asked most famously by Ivan in the form of a tortured child; the Ilyusha thread answers it not with an argument but with a dying child and the people around him. World expansion: the town’s poor, its schoolboys, its doctors and its shame.

The convergence, which is the novel’s last chapter. The Brothers Karamazov ends not with the trial, not with Dmitri’s sentence, not with Ivan’s illness or Alyosha’s vocation, but at a stone outside town after Ilyusha’s funeral, with Alyosha telling the boys that one good memory from childhood may be the thing that saves a person, and the boys shouting for Karamazov.

Consider what that placement claims. The subplot has been given the privileged position — Rabinowitz’s term, treated in Chapter 9 (The Opening Pages) §3.2 and inverted in Chapter 16 (The Ending) — which is the strongest structural statement a novel can make about what it thinks it is about. Everything in the main plot is left partly unresolved (Dmitri’s escape is a plan, Ivan’s recovery is uncertain, Grushenka’s future is open) and the secondary thread receives the closing chord.

The lesson. The domain audit at §5.5 would give this thread three domains and a moderate page-share, and would not predict that it ends the book. Position is a resource the domain model does not measure. A subplot placed last is asserting that the book’s answer lives there, and the assertion is available at almost no cost in pages. It is, in this chapter’s judgement, among the most under-used structural moves in contemporary fiction.

9.8 Shakespeare, King Lear (c. 1606) — where the theory came from

Included, though a play, because Levin’s and Empson’s arguments were built on plays of this kind.

The threads. Lear and his three daughters; Gloucester and his two sons. Shakespeare took the second from a story in Sidney’s Arcadia — the blind Paphlagonian king — and grafted it onto the Lear legend, which had no subplot.

Levin’s relation: analogy in the strong form, with contrast at the level of register. A father misjudges his children; trusts the false; casts out the true; is stripped, exposed, and partly redeemed by the child he wronged. The pattern is the same twice, and the novelty is what is changed:

Why it is the origin of the modern discussion. Two positions, still live:

• The double action is a defect. A. C. Bradley judged the second plot to overload the play and to blur its dramatic clarity, while conceding that it produces an impression of universality unmatched elsewhere in Shakespeare. Earlier neoclassical criticism made the objection more bluntly, and it is the objection the attention tax formalises.
• The double action is the play. Levin’s argument, and Empson’s method. The doubling generalises: what happens to Lear is not a story about a foolish king but a fact about the world, and the only way to assert that in a play is to have it happen twice, to different men, in different registers. This is the Empsonian surplus (§3.3): the meaning belongs to neither action alone.

What the transfer to the novel costs. Two adjustments must be made before this example can be used in a fiction workshop, and both point back to §3.2.

1. A play is held in a single sitting. The analogy between Lear’s plot and Gloucester’s is perceptible across twenty minutes. In a novel the same analogy across eighty pages requires marking. What Shakespeare could leave implicit, a novelist must braid.
2. The stage supplies simultaneity that prose cannot. Edgar leads his blinded father across the same heath on which his king went mad, and the audience sees the space. A novel must construct the connection in language, which is why §5.3 recommends the shared object over the shared image.

9.9 A revision case: Dov’s thread

The following applies the diagnostic protocol to this book’s running example — a contemporary literary novel about three adult siblings, Elena, Nadia and Dov, reconstructing what their father withheld from them in the months before his death. The main plot is the sisters’ reconstruction: Elena wanting to know why their mother changed the locks in March, a question the book leaves open, and Nadia and Elena working out what the ledger in the shop records.

Dov’s thread is the novel’s subplot. Two versions follow.

Version A — the thread as recurring texture

In this version, Dov’s financial situation is present throughout the novel as atmosphere. It has no chain of its own. Its appearances are:

A representative passage, from chapter 12:

The woman from the hardware place two doors down asked her, in the tone people used for asking about illnesses, whether they’d decided anything about the shop. Nadia said they hadn’t decided anything about anything. The woman nodded as though that were an answer, and went back in, and Nadia stood on the pavement a while and counted the shutters that were down along the row. Four, out of eleven. It had been two, in March.

The writing is not the problem. The paragraph does several things well: it locates the shop in a declining row, it gives Nadia something to count, which is her established habit, and it sets a private grief against an economic one.

Diagnosis

Patterns present: presence, not plot (§2.3); the occupational subplot (§8.8), since the material exists to give Dov something to be doing between his appearances; convergence by encounter (§8.10), because chapter 31’s convergence is the three of them happening to be in the same room; and the thematic alibi (§8.13) in embryo, since the thread’s defence — “it’s about what the family is losing” — is true and functionless.

The central observation. Version A contains all the necessary information. Its failure is not of content but of structure: Dov is doing things, and none of the things is a decision. A reader finishing this novel would be able to say what Dov’s situation was and unable to say what Dov did.

Version B — the thread as a chain

This is the version the book uses. The same character, the same shop, the same financial pressure, arranged as a goal–conflict–disaster chain in the sense of Chapter 2 (Scene and Sequel), with a convergence column.

Domain audit. Two domains, stacked, which is §4.8’s working minimum and its most durable pair.

• Thematic counterpoint. Dov’s instinct to protect by silence is his father’s instinct: a man who documented his illness in the back of a ledger and told nobody, and a son who sells a building and tells nobody. The rhyme is not stated. It is produced by the reader holding two withholdings together, which is Empson’s mechanism (§3.3) and Levin’s analogy (§3.2). What the thread asserts is a qualification the main plot cannot make about itself: the sisters are investigating a concealment, and their brother is concealing in the same way, for the same reason, and believes he is being kind.
• Structural convergence. He supplies the evidence. The convergence is literal, late, and — this is the part that matters — the object that converges the plots is the object the counterpoint was about. The ledger is both the proof and the parallel. This is the condition §5.1 names: the two jobs are one act.

A third domain is served more quietly. Character testing ground: the sale applies a pressure the investigation cannot, because the investigation tests what Dov will say and the sale tests what he will do when nobody is watching. The signature is unexcused behaviour (§4.2): nothing forces him to sign in secret.

What the revision did, procedurally

1. Converted a condition into a goal. The shop is losing money became sell the shop before another season of losses. This is the whole of test 1, and it is the most transferable move in the case: a thread acquires a chain the moment its character wants something with an object and a deadline.
2. Supplied a branching point. The buyer’s demand that the building be emptied by month’s end is a kernel: the thread can now go two ways. Version A had no moment at which anything could have gone otherwise.
3. Made the subplot’s clock the main plot’s clock. The deadline is the mechanism by which a parallel thread starts exerting force before it converges. This is the supplied constraint carrier at §4.4, and it is why the middle of the book stops feeling like two activities. The deadline reaches Elena as an empty register drawer and Nadia as Dov’s request that the flat be cleared by the month’s end, made without the reason; the second gives Nadia a motive for putting her hand into the pocket of the coat, since a coat’s pockets are checked before it goes, while the coat itself stays on its hook (Chapter 7).
4. Made the disaster a betrayal in both systems. Signing in secret is Dov’s disaster inside his own chain (he has committed himself and cannot now retreat) and an event inside the main plot (it changes what Elena’s investigation is about). One action, two systems — the multiplicity condition of §5.1.
5. Placed the convergence on an object already planted. The ledger is first seen as an object in Chapter 4 (The Descriptive Detail), named once in a clause about where the father wore his coat, and the looser hand after March the ninth is Elena’s discovery at the chapter break that Chapter 6 (The Chapter Break) works; Dov’s chapter (Chapter 13) has by then shown the reader whose hand it is. Version B does not introduce a new object at the convergence; it explains to the sisters one the reader has been carrying for about half the book. Compare the handwriting plant in Bleak House (§9.3): the strongest convergences are recognitions.
6. Withheld viewpoint until near the climax. Dov is observed from outside for most of the book, and receives a single bounded viewpoint chapter near the climax — the structure Chapter 13 (The Viewpoint Shift) analyses. The plot-side consequence is that his chain is reconstructed by the reader from evidence, which makes the thread’s kernels inferences rather than reports, and does a substantial part of the work of the subplot’s suspense without a single scene of concealment.

Braiding and rhythm

What Version B has not solved

Four problems remain, and a workshop that declared the revision finished would have taught the wrong lesson.

First, the convergence carries a great deal of weight for a single scene. Dov’s reveal must simultaneously resolve his chain, deliver the main plot’s evidence, retrospectively explain the handwriting, and land the thematic rhyme with the father. Chapter 16 (The Ending) treats overloaded final scenes; the risk here is that the reader receives four discharges at once and feels the architecture.

Second, the thread now risks the redundant mirror (§8.4). Dov protects by silence as his father did, and the closeness of the match is the danger. Levin’s analogy requires a change of key, and the current change — a son concealing a sale, a father concealing an illness — may be too small a variation to prevent the rhyme reading as restatement. The available correction is to make the two silences differently motivated. The father kept the record so that someone could check it — that is what the good pen was for; Chapter 7 (Backstory and Flashback) gives the reason in his own words, and the prologue (Chapter 10) plants its first two words. Dov keeps the same ledger in the same book, and enters only the date and the count: nothing about the buyer or the deadline. He continues the form of his father’s honesty and empties it of its content. That is a change of key rather than a repetition, and it costs a paragraph.

Third, the promotion test (§7.1, test 10) returns an uncomfortable answer. Dov has the most at stake of the three siblings, is the only one who acts irreversibly, and is the only one whose chain has a deadline. On three of the test’s four criteria his thread is stronger than the investigation it serves. This book has decided that the novel is Nadia and Elena’s and that Dov’s opacity is a large part of the reason the ending works — but that is a decision, not a fact, and a different novelist would write the other book. The right outcome of a diagnostic protocol is frequently a defensible choice rather than a repair.

Fourth, the empty drawer opens a question the book must manage. Elena now wants to know who took the ledger, and the book must keep her from putting the question to Dov until the convergence, or the convergence arrives early. That she might suspect their mother, who changed the locks, would hold the question open without explaining the mother’s silence; whether it can do so without reading as contrivance is for a later draft to test.