The mechanism of this chapter is an economy: a grant is the most expensive instrument a novelist has, so it must be admitted on terms, worked hard once admitted, and timed so that its cost falls where the book can bear it.
The admission requirement
A grant commits the book’s scarcest resource, the reader’s identification and capacity to model minds, for the length of the book rather than of a scene. Its commonest justification, access, is the one for which the cheapest substitutes exist. Before granting a viewpoint for access, a writer has five cheaper instruments: report (someone tells the focaliser what happened), inference (she deduces it, from a smell of fresh ink in a drawer, say), reassignment (the scene is rebuilt so that she can plausibly be there), document (a letter, a message, a ledger) and ellipsis (the event is not narrated, and its consequences carry it). The principle of multiplicity in The Purposes of a Scene (Chapter 1), that a unit should do several jobs at once, becomes at the scale of the grant an admission rule.
A grant should be able to name two domains, and one of them should not be access. In practice the writer completes two sentences before granting: This vantage gives the reader … that no existing vantage could deliver, and The cheaper instruments I rejected are …, and I rejected them because …. There is a good answer to the second, and it is usually the same one: report can state a fact, and only dramatisation can make a decision felt as it is taken. The threshold of two is a working rule, not a law; a grant may serve its domains across the book rather than inside its own pages, as Addie Bundren’s single section in As I Lay Dying serves almost nothing locally and reorganises the novel around it.
The rule runs in both directions. A grant should do more than one job, and the character receiving it should already be doing more than one job in the book. The strongest grants go to characters who were load-bearing before they became viewpoints: Esther Summerson is the hidden centre of Bleak House’s plot, and Levin is the counterweight of Anna Karenina’s argument. A vantage awarded to a character who exists only to hold it reads as an appointment.
Where the distinguishing work lives
Revision usually looks for difference between viewpoints in voice, in dialect, rhythm or verbal tic. Voice is the most expensive place to work and the easiest to overdo, and in close third person among characters of one family it is often unavailable without making some of them sound affected. The work of a grant lives in four cheaper places.
Selection is what a mind registers. Two people enter the shop; one notices that the register drawer sticks, the other that the coat still hangs on its hook upstairs, and they are now distinct at no cost in words, since the sentences were going to describe something. The idiom of judgement is what appraisal leaks into the narration: when a third-person sentence calls a room poky or a brother impossible, the judgement is the focaliser’s, and the reader knows whose head she is in before any name appears. The knowledge boundary is what the vantage cannot know, and it is a resource: a chapter that keeps its boundary produces inference, misreading and dread, and a viewpoint chapter that feels thin has usually been allowed to know too much. Naming is what a vantage calls people, places and acts: a sister who thinks Dov and a brother who thinks of the dead man as his father have been told apart without exposition, and so have a vantage that calls the sale selling the shop and one that calls it what he did.
The first grant and the later grant
A first grant is made while the reader’s investment across the cast is still forming, in practice within the first chapters, and it is cheap because a new vantage simply joins the field. A later grant is made after the book has trained the reader to follow fewer vantages, and its cost is the sum of three charges. Formal surprise is the reader’s discovery that the book is a different shape from the one she thought; it is reduced by declaring early that movement is possible. Attention withdrawal is the pages taken from vantages the reader has invested in; it is reduced by keeping the grant short. Investment cold start is the work of building a new mind from nothing at the point in the book where the reader is least willing; it is reduced by prior presence. The third charge yields the chapter’s most practical rule: a later grant is cheap in proportion to how long the reader has already been curious about its recipient. A character watched from outside for three hundred pages, whose silence has been a standing question, costs almost nothing to enter.
That is why the bounded grant, one chapter or section given to a mind that never holds the reader again, is the standard remedy for a late grant. It pays formal surprise once, withdraws a chapter’s worth of attention, and pays no cold start if the recipient has been present throughout. Its shaping requirement is that the section read as complete in itself, not as the first instalment of a thread the book then drops. As a rough schedule, a grant in the opening chapters needs nothing; in the middle it should be bounded or carry a payoff no report could deliver; in the last quarter it should be bounded, or the book’s subject should be the insufficiency of its own vantages; and in the last pages it must read as the book’s argument rather than its author running out of alternatives.
The ceiling
There is no correct number of viewpoints. The ceiling rises with the book’s length, but only if the length goes into longer stays rather than more vantages; with the reader’s task, sharply if the task is triangulation and not investment; with distinguishability, since fifteen instantly recognisable minds are cheaper than four interchangeable ones; and with the legibility of the pattern. As I Lay Dying gives fifteen focalisers fifty-nine sections in a short novel and is not diluted, because the reader triangulates one journey rather than investing in fifteen people, because each voice is known at once by idiom and obsession, and because the distribution is deliberately unequal: Darl holds nineteen of the fifty-nine sections. A book can carry as many vantages as it can make instantly identifiable and as its reader’s task requires; what it cannot carry is many vantages all asking to be inhabited equally.
Architectures
The four properties of the grant combine into a small number of whole-book shapes, here called viewpoint architectures, and a manuscript nearly always belongs to one whether or not its author chose it. The table gives the commonest, with what each buys and costs.
| Architecture | Shape | Example | Buys | Costs |
|---|---|---|---|---|
| Single fixed | One focaliser throughout | The Remains of the Day | Depth; the boundary as a permanent engine | Everything off-stage reported or inferred |
| Roving third | One mind at a time, moving without a declared pattern | Mrs Dalloway | Coverage at close range | Minds blur without strong differential noticing |
| Alternating dual | Two vantages for the length of the book | Bleak House | Contrast at its clearest; a felt rhythm | The reader prefers one; the other interrupts |
| Rotating ensemble | Three or more in a cycle | The Poisonwood Bible | A social subject | Dilution; the cycle becomes a metronome |
| Hub and spokes | One dominant vantage, others brief and specific | Anna Karenina (two hubs) | Depth and coverage together | An unequal cast, which drafts resist |
| Bounded grant | One section to a mind that never returns | Addie in As I Lay Dying | A late grant at a fraction of the price | The reader may expect a return |
| Documentary | Letters, diaries, transcripts | Dracula | Authenticity; incompatible vantages | Interiority always performed for a reader |
| Mobile omniscient | A narrating intelligence with free access | Middlemarch | Scope, irony, analysis | Needs a narrator strong enough to hold it |
| Refused grant | One character kept outside throughout | The Count in Dracula | Inference; opacity as subject | Readers supply an interior, perhaps the wrong one |
The architectures are adjacent regions rather than boxes, and accomplished novels braid two. Their use in revision is substitution: name the current architecture, name its two neighbours, and rewrite one chapter in the stronger. The substitution that most often helps and is least often tried is the last row applied to one character, because much of what a draft calls developing a difficult character is granting her an interior, and much of what makes such characters powerful is its absence.
See also: the Reference treatment of the grant, with thirteen architectures, the signals that declare a plural book and the historical drift of architectures; the Reference economy of the grant, with the page-share audit’s interpretation table.