5. Multiplicity at Book Scale

5.1 The principle, stated as a constraint

Multiplicity — the practice of making one unit of text discharge several narrative functions at once — is introduced in Chapter 1 (The Purposes of a Scene) and reappears in every chapter of this textbook. At the scale of the subplot it changes character, and the change should be stated.

At sentence scale, multiplicity is an economy: a well-chosen noun does the work of a clause of explanation. At book scale, multiplicity is a survival condition, because the unit under discussion is too large to be justified by any single function. No writer would defend forty pages on the ground that they supply pacing relief. Forty pages must be doing several things, and the several things must be inseparable — which is a stronger requirement than doing several things in succession.

The constraint can be stated arithmetically.

• A subplot occupying 15% of a 100,000-word novel costs 15,000 words.
• Those 15,000 words are drawn from a budget in which the main plot’s own chain, its sequels, its transitions and its ending are already competing.
• The reader’s tolerance for diversion is not proportional to the book’s length: it is roughly constant, and is spent per crossing rather than per page.

From which the operative rule:

The discipline is not to give the subplot more jobs. It is to find the version of the subplot in which the jobs are the same act. A thread that supplies thematic counterpoint in chapter 9 and structural convergence in chapter 27 is doing two things. A thread whose convergence is its counterpoint — where the object it hands the main plot is the object the comparison was about — is doing one thing that means two.

This is the book-scale form of the distinction Chapter 9 (The Opening Pages) §5.1 draws between compression and selection. Compression stacks functions; selection finds the material that already carries them.

5.2 Anatomy of a multiplying beat

Multiplicity at this scale cannot be demonstrated on a sentence. It can be demonstrated on a single beat of a subplot, dissected element by element.

The instance is the chaplaincy election in Middlemarch (Book Two, chapter 18). Lydgate — whose thread is nominally the novel’s second plot — must cast the deciding vote for the chaplaincy of the new fever hospital. Bulstrode, who funds the hospital, supports Tyke. Lydgate prefers Farebrother, who is poor, likeable, and needs the forty pounds. Lydgate votes for Tyke.

Five domains, one committee meeting, and not one of them is served by a passage inserted to serve it. Two features of the beat are worth extracting for teaching.

The convergence is planted inside the counterpoint. Lydgate’s vote is simultaneously the novel’s argument about how vocations are bought and the specific mechanism by which Bulstrode will later acquire a hold over him. The thematic relation and the causal relation are the same event, which is the condition §5.1 names.

The beat is small. It is a vote at a meeting. This is the characteristic form of book-scale multiplicity: not a large scene doing many things, but a small, ordinary, structurally exact one placed where several lines of force cross. Writers looking for multiplicity in their own subplots tend to look at the big scenes. It is nearly always in the small ones.

5.3 Where the work actually lives

Novice revision looks for subplot integration in theme — in a stated resemblance, or an image repeated across threads. In practice, the integrating work is done in four humbler places, none of them figurative.

1. A shared object or institution. A lawsuit, a house, a will, a hospital, a ledger. A reliable integrator, because it gives both chains something to do to the same thing, and because a reader’s memory holds objects better than it holds parallels. Bleak House has Jarndyce and Jarndyce; Middlemarch has the hospital and Featherstone’s will; the running example in §9.9 has a shop and a ledger.
2. A crossing character who belongs to both chains. Someone with business in each thread, whose crossings are motivated rather than convenient. Stiva Oblonsky does almost nothing in Anna Karenina except connect; Inspector Bucket does it in Bleak House; Farebrother does it in Middlemarch. The crossing character is what makes braiding feel like one world rather than two books, and a manuscript that lacks one usually reads as parallel play (§8.1) whatever its thematic ambitions.
3. A repeated decision-shape. Not a repeated image but a repeated kind of choice: in this book, people keep having to decide whether to say a true thing that will cost them. When the same shape of dilemma recurs across threads, the reader perceives unity without any statement being made — and the perception survives the long intervals that defeat imagery.
4. The placement of the crossing. Which chapter break the braid is cut on is a structural decision that does integrating work by itself: a cut from the subplot’s small betrayal to the main plot’s large one asserts a relation that no sentence has to assert. Chapter 6 (The Chapter Break) supplies the mechanics under the names proposition and payoff; this chapter supplies the reason to care.

Imagery can integrate — Forster’s rhythm (§3.1) is this — but it is the highest-risk instrument at book scale, because a figure repeated across a hundred pages is either missed or heard as authorial insistence, and the margin between the two is narrow.

5.4 The ceiling: how many chains can a book run?

There is a limit, and it is not where writers expect it to be.

The naive assumption is that the limit is a thread count — three threads good, five threads too many. The tradition falsifies this immediately: War and Peace carries four families and something close to a dozen traceable chains; Bleak House carries three systems and perhaps a dozen sub-chains within them. Neither is confusing to read.

The actual constraint is a product of three variables.

The ceiling is reached when the product of these exceeds the reader’s capacity to hold the threads as chains rather than as a sequence of episodes. Two failure signatures mark it:

• The reader can name the characters but not the questions. They know who everyone is and cannot say what any thread is trying to achieve. The chains have degraded into a cast.
• The reader stops preferring. A healthy braided novel produces preferences: readers of Bleak House skip ahead to Esther or to the Dedlocks. When a reader has no preference among threads, the threads have stopped being distinct enough to compete, which is the same fact as their having stopped being legible.

The limit case: War and Peace. Tolstoy exceeds every convention this chapter states and the book is not confusing, and the reasons are instructive rather than magical.

1. The threads are anchored to families, which are the most memorable unit of organisation available in a realist novel — the reader holds the Rostovs as one object even when six Rostovs are in play.
2. The crossing characters are numerous and motivated. Pierre has business with the Bolkonskys, the Kuragins and the Rostovs; the network has no isolated nodes.
3. The crossings are frequent and short. Tolstoy rarely leaves a family dormant for the length that would require reinstatement.
4. The threads are differentiated by register — Bolkonsky austerity, Rostov warmth, Kuragin appetite — so that a paragraph identifies its thread before a name appears.

That is the honest account of the ceiling: it can be raised, and the means of raising it are technical rather than inspired. §6.6 and §7.1, test 8 make them operational.

The obligation that comes with exceeding the hierarchy. A book that runs three or more chains without a primary one has become a multi-plot novel and inherits three duties it cannot decline:

• It must declare itself early, so that the reader does not spend fifty pages deciding which thread is the real one. Chapter 9 (The Opening Pages) §4.6 makes the same point about the formal contract; a multi-plot novel that opens like a single-plot novel has mis-promised.
• It must make the relations between threads carry the argument, since it has surrendered the unifying effect of a single pursuit.
• It must converge or rhyme at the end, because a form organised by relation is judged by whether the relation resolves. Chapter 16 (The Ending) treats what resolution can consist of.

5.5 Exercise: the page-share and domain audit

A practical procedure for a completed draft, and the most reliable single instrument in this chapter.

1. List every thread in the manuscript, including ones you have not thought of as threads. For each, write the goal, the obstacle and the consequence in one sentence each. Threads for which you cannot complete all three go on a separate list marked presence, not plot (§2.3).
2. Count pages. For each thread, total the pages in which it is the primary business of the scene. Convert to a percentage of the whole.
3. Mark domains. Beside each thread write the letters of every domain it serves: T (thematic counterpoint), C (character testing), P (pacing relief), S (structural convergence), W (world expansion), X (temporal counterpoint).
4. Mark convergence. For each thread, note the page at which deleting it would break the main plot’s causal logic (§3.4). If there is no such page, write none.
5. Mark dormancy. For each thread, note the longest interval in pages between consecutive appearances.

Then read the table.

The audit takes an afternoon for a full manuscript and is worth running twice, several weeks apart, because the first run is contaminated by what the writer intended and the second is closer to what is on the page.