3. Why Viewpoint Allocation Is a Structural Decision

This chapter asserts that granting viewpoint is expensive. The assertion is correct, and it can be argued from four independent directions — narratological, socio-formal, cognitive and historical. The four do not merely corroborate one another; they disagree productively, and the disagreements are where the teaching is.

3.1 Focalisation: separating who speaks from who perceives

The indispensable move was made by Gérard Genette in Narrative Discourse (1972; Eng. 1980), and it is the most useful distinction in this chapter, because every later term depends on it. Earlier criticism had run two questions together under the heading of “point of view”:

• Who speaks? — the question of voice, of the narrating agent.
• Who perceives? — the question Genette named focalisation, of the position from which narrative information is regulated.

The two vary independently, and the failure to separate them is why so much craft advice about viewpoint is unusable. A narrator may speak in the first person while perceiving from a position not their own (Nick Carraway recounting Gatsby’s boyhood, or Michaelis’s testimony); a narrator may speak in an impersonal third person while perceiving strictly through a single character’s senses and knowledge (Kafka’s The Trial, which almost never departs from what Josef K. can register). Person is a fact about the speaker; focalisation is a fact about the distribution of knowledge.

Genette’s tripartite scheme:

Genette then subdivides internal focalisation by how it behaves across a text, and this is where the scheme becomes a tool for the present chapter rather than a taxonomy of sentences:

• Fixed — one focaliser throughout. The Remains of the Day; What Maisie Knew.
• Variable — the focaliser changes, one at a time, across the text. Madame Bovary; Mrs Dalloway; most contemporary multi-viewpoint fiction.
• Multiple — the same events are focalised successively through different characters. The Sound and the Fury; Rashōmon-type structures; the epistolary novel of contradictory testimony.

The distinction between variable and multiple internal focalisation is the most under-used item in the whole narratological toolkit, and it maps onto a craft distinction this chapter needs. Variable focalisation buys coverage: each vantage sees different material, and the reader assembles a wider world. Multiple focalisation buys incommensurability: each vantage sees the same material, and the reader assembles a doubt. The first is the ordinary multi-viewpoint novel. The second is the epistemological structure proposed as a sixth domain at §4.6.

Two cautions about the scheme, which students should carry.

First, Genette himself noted that focalisation is rarely constant for long, and that the categories describe dominants rather than absolutes. A novel in fixed internal focalisation will contain sentences no character could have produced. This is not a defect in the novels; it is a limit on the descriptive power of the categories.

Second, the scheme describes restriction of information, not intimacy. A narrative can be internally focalised and cold (The Trial); externally focalised and unbearably intimate (Hemingway). Craft advice that equates “close third” with “internal focalisation” has collapsed a knowledge relation into a tonal one.

3.2 Four refinements: Stanzel, Bal, Cohn, Booth

Genette’s scheme is a skeleton. Four further bodies of work put flesh on it, and each supplies something this chapter needs.

Franz Stanzel’s typological circle (A Theory of Narrative, 1984; Eng. trans. Charlotte Goedsche) organises narrative situations not as discrete boxes but as positions on a continuum that returns on itself. Stanzel proposes three narrative situations:

The value of the circle for this chapter is that it is a circle. Stanzel’s point is that the situations shade into one another by degrees, so that a novel can drift from figural towards authorial without any identifiable sentence where the change occurs — which is the phenomenology of the failure pattern at §8.6, the omniscient slip. A model with discrete boxes cannot describe a gradual leak; a model with a continuum can.

Mieke Bal’s refinement (Narratology, 1985; 4th ed. 2017) is the most consequential for practical analysis. Bal argues that focalisation is not a property of a narrator but a relation between a subject and an object: there is always a focaliser and a focalised, and the analytic question is always double — who sees and what is seen. This produces two gains.

1. It makes non-character focalisation describable. A crowd, an institution, a camera position or an unspecified observer can occupy the focalising position. §6.3’s architectures 9, 11 and 13 need this.
2. It makes the object of vision analysable. To say that a chapter is focalised through Nadia is only half a description; the other half is what Nadia is permitted to look at. A character granted viewpoint but pointed only at plot information is the information mule of §8.2, and Bal’s formulation is why the fault is visible as a fault.

Bal is also the source of the useful observation that focalisation is embedded: a character may focalise another character who is themselves focalising something. Wuthering Heights runs three deep — Lockwood focalising Nelly focalising Heathcliff — and the reader’s access to the innermost layer is filtered twice.

Dorrit Cohn’s Transparent Minds (1978) supplies the vocabulary for what a grant of viewpoint actually delivers at sentence level. Cohn’s three modes for third-person contexts:

The distinction matters here because the three modes cost different amounts of empathy budget. Psycho-narration is cheap: it reports a mind without requiring the reader to inhabit it, and a book can afford it for a dozen characters. Narrated monologue is expensive: it requires the reader to adopt the character’s idiom and judgement, and it is the mode in which a viewpoint character is truly inhabited. Quoted monologue is more expensive still. A draft that “has five viewpoint characters” may in fact have one inhabited consciousness and four psycho-narrated ones, and the diagnosis is invisible until the modes are named.

Wayne Booth’s The Rhetoric of Fiction (1961; 2nd ed. 1983) contributes the axis of distance and the reminder that it is plural. Booth insisted that distance operates between several pairs of positions at once — narrator and character, narrator and reader, implied author and narrator, reader and character — and that a change in any of them alters the reader’s judgement. His essay “Distance and Point-of-View” (1961) is the compact statement.

Booth’s contribution to the present argument is a warning. He observed that inside views produce sympathy more or less mechanically: sustained access to a consciousness tends to align a reader with it, more or less regardless of what the consciousness contains. This is the mechanism this chapter calls sympathy engineering (§4.4), and Booth’s point is that it is not optional. A writer who grants viewpoint to a character has already, by that act, argued for them. A novel that gives an abuser a hundred pages of interiority and expects the reader’s verdict to be unaffected has misunderstood the instrument.

Two further sources should be named. Boris Uspensky’s A Poetics of Composition (Eng. 1973) analyses point of view on four separable planes — ideological, phraseological, spatio-temporal and psychological — the sharpest available account of how a narration can be inside a character on one plane and outside on another. Shlomith Rimmon-Kenan’s Narrative Fiction (1983) offers a compact synthesis with a similar three-facet division, and remains the most usable single-chapter introduction. Norman Friedman’s pre-Genette taxonomy is treated at §6.2.

3.3 Woloch: character-space, the character-system, and the distribution of attention

This chapter’s “empathy budget” is a serviceable metaphor. Alex Woloch’s The One vs. the Many: Minor Characters and the Space of the Protagonist in the Realist Novel (Princeton, 2003) is the rigorous instrument the metaphor gestures at, and it is this chapter’s principal source.

Woloch begins from an observation that is obvious once stated and rarely pursued. A novel contains many implied human beings; it has room to attend to very few of them. Every character in a realist novel therefore stands in a relation of friction between the fullness of the person they are asked to imply and the limited textual room they are given. Woloch names that room character-space:

Character-space — the particular and delimited position that a character occupies within the narrative as a whole; the relation between an implied human being and the space the narrative affords them.

And he names the arrangement of many such spaces into a single structure the character-system:

Character-system — the arrangement of multiple character-spaces into a unified narrative structure, in which the attention given to any one character is a function of the attention withheld from the others.

Three consequences follow, and each converts a craft intuition into an analytic claim.

First: attention is scarce, and scarcity is structural rather than incidental. A novel’s attention is not merely limited by length in the way a suitcase is limited by volume. It is distributed, and the distribution has a shape. Woloch’s insistence is that this shape is a formal feature of the work — as available to analysis as its plot or its imagery — and that criticism has systematically ignored it because it appears to be a mere consequence of the story rather than a construction in its own right.

Second: minor characters are not small; they are compressed. This is Woloch’s most useful single claim for a writer. A minor character is an implied person subjected to a squeeze, and the squeeze registers in the character’s presentation as distortion. The flatness of the flat character is not an absence of depth but the visible mark of depth being excluded. Hence the characteristic forms minor characters take in the realist novel: the eccentric with one repeated gesture, the grotesque, the type, the walk-on defined by an occupation. Woloch reads such figures as bearing the pressure of their own exclusion, and the reading changes what a writer sees on their own page. A tic is not characterisation; it is what characterisation is compressed into. This book calls the effect character compression, to keep it apart from Chapter 2’s compression of page-space against story-weight.

Third: the distribution is a political fact. Woloch’s title is not decorative. The realist novel’s formal problem — one protagonist, many others, finite pages — reproduces at the level of form a social problem about whose experience counts. He develops this most fully on Le Père Goriot and on the English realist tradition, and the argument does not require the reader to accept any particular politics to be usable: it requires only the recognition that deciding whose interiority the reader will inhabit is a decision about whose experience the book treats as fully human. In a novel about a strike, granting viewpoint to the mill owner and not to the workers is not a neutral craft decision. Neither is the reverse. Both are arguments, and the reader receives them as arguments whether or not the writer intended one.

Woloch offers, memorably, the formulation that the minor character functions as something like the novel’s proletariat — labouring to produce a protagonist’s significance while being denied the space in which significance can be represented. The phrase is polemical, and it is the fastest route into the argument for a student who has been taught to think of viewpoint as a technical choice.

What this gives the chapter that the empathy-budget metaphor does not.

The practical upshot is the test at §7.1, no. 10 — the character-compression test — which asks not what the viewpoint characters gained but what happened to everyone else. It is the only test in the protocol that looks away from the page under revision, and it is, in this author’s experience of teaching the material, the one that most often changes a manuscript.

A caution. Woloch’s book is about the nineteenth-century realist novel, and its claims are calibrated to a form with large casts, a strong protagonist convention and an omniscient narrative habit. Applying it to a two-hander, a modernist consciousness-novel or a genre thriller requires adjustment, and the adjustment is not always small. The chapter uses Woloch as an instrument for noticing distribution, not as a theory that predicts value.

3.4 The cognitive strand: minds, models, and the limits of the empathy claim

Three empirical and cognitive literatures bear on viewpoint allocation. The first two support the chapter’s claims; the third substantially qualifies them, and reporting the qualification is more useful than suppressing it.

Theory of mind and levels of intentionality. Lisa Zunshine’s Why We Read Fiction (2006) proposes that narrative fiction exercises the ordinary human capacity for theory of mind — attributing mental states to others — and that the pleasure of much fiction lies in the difficulty of the attributions it demands. Zunshine borrows from Robin Dunbar the notion of levels of intentionality: I believe is first order; I believe that she wants is second; I believe that she wants him to think that she does not know is fourth. Zunshine reports that ordinary readers manage roughly four levels comfortably and degrade beyond that, and she reads passages of Woolf and Richardson as deliberate exercises at the ceiling.

The application to viewpoint is direct and is not usually drawn. A viewpoint grant changes the level of intentionality the reader must track for that character, and for everyone the character thinks about. Before a grant, a character is an object of attribution: the reader guesses at them from behaviour. After a grant, they become a subject, and the reader must now hold both what they think and what others think about them. Adding a viewpoint character therefore does not add one item to the reader’s load; it adds a new set of relations between that character and every other. The load grows faster than the cast does, which is a precise mechanism for what “dilution” feels like from inside.

Fictional minds and the continuing-consciousness frame. Alan Palmer’s Fictional Minds (2004) argues that narratology’s traditional focus on techniques for presenting speech and thought has produced too narrow an account of how readers construct characters. Readers do not assemble a character out of quoted thoughts; they maintain what Palmer calls a continuing-consciousness frame — a running model of a whole mind that persists between appearances and is inferred from action, dialogue, physical description and report as much as from interiority. Palmer’s later work on social minds (Social Minds in the Novel, 2010) extends this to intermental, group-level thought — the “we” of a village or a family that thinks collectively.

Two consequences follow for this chapter, and they pull in useful directions.

• A character does not need viewpoint to have a mind. This is the strongest available argument against granting viewpoint reflexively. Readers construct rich fictional minds from external evidence; the question a writer must answer is not how else will the reader know he has an inner life? but what does the reader need that only direct access supplies? That is test 1, the necessity test.
• Collective focalisation is not a gimmick. Palmer’s social minds give theoretical standing to architecture 9 in §6.3, the collective we, and to the intercalary chapters of The Grapes of Wrath.

Situation models and the cost of a shift. Zwaan and Radvansky’s synthesis (1998) holds that readers build a situation model of the state of affairs described, tracking five dimensions — protagonist, space, time, causality, intentionality — and paying a measurable processing cost whenever one changes. Chapter 5 §3.3 uses this to price the transition. The point specific to this chapter, and this chapter’s inference rather than Zwaan and Radvansky’s finding, is that the protagonist dimension is plausibly the most expensive of the five, because it is the dimension to which the other four are indexed: a change of protagonist typically drags space, time and intentionality with it, whereas a change of space often leaves the others intact. This is the cognitive basis for the claim that a viewpoint change is among the most costly moves available, and for the finding that a legible pattern reduces the cost — Emmott’s contextual frames (1997) permit frame recall rather than frame construction, so that the twelfth alternation in a book with a declared architecture is cheaper than the first.

The Deictic Shift Theory literature (Duchan, Bruder and Hewitt, eds., Deixis in Narrative, 1995) formalises the same phenomenon from a linguistic direction: readers adopt a deictic centre — a here, a now, a who — and a viewpoint change requires them to abandon it and construct another.

And now the qualification, which matters.

The empathy-budget metaphor invites a romantic corollary: that novels build empathy, that more viewpoints mean more empathy, and that the moral value of fiction lies in this. Suzanne Keen’s Empathy and the Novel (2007) is the necessary corrective, and the chapter reports it rather than suppressing it.

Keen surveys the empirical literature and finds the popular claim — that novel-reading reliably produces real-world empathy and pro-social behaviour — substantially unsupported. Her findings, stated carefully:

• Narrative empathy is real as an in-reading phenomenon. Readers do feel with characters, and the feeling is measurable.
• It does not reliably transfer. The inference from felt empathy during reading to altered behaviour afterwards is not supported by the evidence; Keen notes that empathy may even be easier with fictional persons because they make no demands.
• Authors’ empathetic intentions frequently misfire. Keen distinguishes bounded strategic empathy (directed at an in-group who already share the author’s assumptions), ambassadorial (directed at a specific out-group audience for a specific end), and broadcast (aimed at all readers) — and observes that the third, the most morally ambitious, is the least reliable.
• Empathy and judgement are separable. A reader can feel with a character and still condemn them, which is what the strongest novels of interiority arrange.

The point is reinforced by the fate of the best-known experimental finding in the area. Kidd and Castano’s 2013 report in Science that reading literary fiction improved performance on a theory-of-mind task was widely publicised; a large registered replication by Panero and colleagues (2016) failed to reproduce the effect. The honest summary for a textbook is that the relation between reading fiction and social cognition is real but small, unstable across studies, and much weaker than the cultural rhetoric around it.

What survives, and what this chapter needs, is the narrower claim: inside views produce alignment within the world of the book. Booth’s mechanism holds. Whether it improves anyone is a separate question with a discouraging evidence base, and a chapter that sold viewpoint as a machine for producing moral improvement would be selling something the research does not support.

3.5 The historical strand: the prohibition is younger than the novel

This is the chapter’s principal corrective, because it changes the status of every caution that follows, and it should be stated plainly.

The dominance of the single limited third person is recent and local. For most of the novel’s history in English, French and Russian, the default narrative situation was a mobile narrator with free access to any consciousness in the book, exercising that access whenever the material called for it — including within a single continuous scene.

A brief historical survey makes the point better than argument.

The turn is attributable to a small number of identifiable interventions. Henry James, in the prefaces later collected as The Art of the Novel (1934), developed the idea of the reflector or centre of consciousness — a character through whose registering intelligence the events of the novel are refracted — and argued that a novel gains in intensity what it loses in scope by such restriction. James’s famous image of the house of fiction with a million windows, at each of which a watcher with a unique field of vision stands, is a statement of the same principle: the aperture and the watcher together make the subject.

Percy Lubbock’s The Craft of Fiction (1921) converted James’s practice into precept. Lubbock’s preference for the dramatic showing of the scene over the narrator’s picture, and his treatment of shifting vantage as a lapse of control, are the proximate origin of what a modern workshop teaches. It is worth noticing that Lubbock’s book is a work of Jamesian advocacy, not of neutral description, and that E. M. Forster’s Aspects of the Novel (1927) dissented almost immediately — Forster argued that the novelist’s right to shift vantage is a resource, and that what matters is not consistency but the reader’s confidence in the writer’s control. Forster’s position is essentially the one this chapter adopts at §8.1.

The conventional force of the prohibition then hardened through mid-century American creative-writing pedagogy, which needed teachable rules, and through the commercial editing practices of the second half of the century. By the time a contemporary student encounters “don’t head-hop”, it is presented as a fact about reading rather than as the residue of a specific aesthetic programme with a date and an address.

Stated plainly: “head-hopping” names the violation of a convention that Fielding, Eliot and Tolstoy did not hold. They moved between minds within a single scene continually and masterfully, and any account of narrative that classifies their practice as an error has produced a false description of the art in order to preserve a rule of thumb.

Two qualifications keep this from being an argument for anarchy.

The convention is real, and readers have been trained in it. A contemporary reader of contemporary fiction does experience an unsignalled mid-scene change as a wobble, because their expectations were formed by a hundred years of books that did not do it. Conventions are not laws, but they are also not nothing: violating one has a cost that must be paid for.

The nineteenth-century practice is not the same manoeuvre performed carelessly. What Eliot and Tolstoy do is governed. The narrator’s mobility is established from the first page, exercised consistently, and accompanied by a narrating voice strong enough to hold the transitions together. The reader is never uncertain whose narration this is, only which mind it has stepped into. The workshop error is not omniscience but an unstable narration that has not established the right it is exercising. §8.1 develops this at length; the discrimination it arrives at is that the fault is in the establishment, not in the movement.

3.6 The first principle, stated — and qualified

We can now state the chapter’s first principle with more precision than the metaphor allows.

The allocation principle. A novel’s viewpoint structure distributes a finite quantity of reader attention, identification and mind-modelling capacity across its cast. The distribution is a formal feature of the work; it determines what the book can witness and how well it can know anyone; and it deforms, in patterned and legible ways, every character it excludes.

And now the qualifications, which are where the more interesting teaching lies.

(1) The pool is not fixed by nature; it is fixed by construction. The budget metaphor implies a quantity given in advance, like a sum of money. It is nearer to a muscle: a book that trains its reader well can carry more than a book that does not. As I Lay Dying asks a reader to hold fifteen focalisers in about two hundred and fifty pages, and readers manage, because Faulkner has made each voice instantly identifiable by idiom, because the sections are short, and because the reader’s task is not to invest in fifteen people but to triangulate one event from fifteen positions. Change what the reader is asked to do with a viewpoint, and the arithmetic changes.

(2) Not every grant asks for identification. The budget metaphor assumes a single currency. In fact a grant may be spent for irony (the reader is placed inside a mind in order to see past it), for dread (the reader is placed inside a mind about to be destroyed), for comedy, or for testimony. Several of these are cheaper than identification because they require less: a reader can occupy Jason Compson’s section without any wish to be him.

(3) Depth is not proportional to page share. A single well-chosen chapter can produce more felt knowledge of a person than two hundred diffuse pages. The revision case at §9.9 turns on this, and it is the argument for the bounded grant (§6.3, no. 6).

(4) The scarcity is culturally and formally specific. Woloch’s model is calibrated to the realist novel with its protagonist convention. Forms that do not have that convention — the collective narrative, the systems novel, the polyphonic epic, the oral epic — distribute attention differently and are not well described as spending a budget. §16 presses this hardest.

(5) The principle describes tendencies in reading, not laws of literary value. Every empirical source in §3.4 measures comprehension and response, not merit. A book may impose costs the model calls prohibitive and be worth every one of them.