Everything in this chapter is an instrument for noticing. None of it is a rule, and the chapter’s central metaphor is the part most in need of scepticism.
The empathy budget is borrowed from economics, and like all such metaphors it naturalises scarcity. A budget presupposes a fixed sum, competing claims and the necessity of trade-offs; it makes the allocation of interiority look like the allocation of money, which is to say like a problem with a correct answer and an unavoidable arithmetic. That framing has real diagnostic power — §5.5 and §7.1 could not exist without it — and it smuggles in an assumption that a great deal of fiction refutes.
Consider what the model must say about the novels this chapter most admires.
War and Peace has upwards of five hundred named characters and grants interiority, briefly or at length, to a startling number of them, including a boy whose interiority we briefly enter before his death. On the budget model this is profligacy: attention divided until nothing is left. On the page it produces the opposite of thinness — a sense, unmatched in the form, that the world is full of people whose insides are as crowded as the reader’s own. Tolstoy does not divide a fixed sum. He enlarges what the reader can hold, by demonstrating repeatedly that another person’s interior is available if one looks, until looking becomes the book’s habit and the reader’s.
As I Lay Dying runs the same argument at the opposite scale. Fifteen consciousnesses in a very short book ought to be impossible. It is not impossible, because Faulkner has changed what a reader is being asked to do — and the budget model, which prices only one transaction, cannot register the change.
Middlemarch ends by observing that the good of the world depends partly on unhistoric acts, and on the number of people who lived faithfully hidden lives and rest in unvisited tombs. That sentence is a claim about the character-system as much as about morality. A novel whose closing gesture is towards the people it did not have room for has understood the scarcity better than the model does — and has also declined to accept it as final.
Three specific limits, then, on everything above.
(1) The scarcity is a property of a form, not of narrative. The realist novel has a protagonist convention, and Woloch’s model describes what that convention does to everyone else. Forms without the convention — the oral epic, the chronicle, the collective narrative, the systems novel — distribute attention on other principles and are not well described as spending. A student who applies the model outside its range will find every one of those forms defective, which is a fact about the model.
(2) Attention is not conserved. The economic metaphor implies that what one character gains another loses. Reading does not work like this. A reader who has been shown that Casaubon has an inside does not thereby care less about Dorothea; they care about her differently, and with finer discrimination. Interiority granted to a minor figure can increase the interest of a major one, because it changes what the major one is standing next to. Any model that cannot represent this is missing something the novels do routinely.
(3) The model measures cost and cannot measure worth. Every empirical source in §3.4 measures comprehension, response and processing load. None measures value. A book may impose costs the model calls prohibitive — fifteen narrators, a corpse’s monologue, a hundred-page excursus, a narration that moves between minds inside a sentence — and be worth every one of them. Some novels spend recklessly and are the better for it, and a pedagogy that can only price the spending has taught economy and not literature.
What the model does reliably is convert a vague dissatisfaction into a specific question. This book feels thin becomes: five vantages hold equal shares and none has a knowledge boundary, or this grant serves access only, or the only character whose motives are never available is the one the plot needs to be wrong. Specific questions can be answered. Vague ones can only be worried at.
And there is a last question, less mechanical than the ten in §7, which belongs at the end because it cannot be answered by an audit. Every allocation of viewpoint makes an argument about whose experience the book treats as fully real. Most writers make that argument without noticing, by following convenience, convention or affection. The chapter’s whole apparatus exists to make the argument visible, and having made it visible, it has nothing further to say about which argument to make.
So the last test is this:
If a reader could see only the shape of your allocation — who is inside, who is outside, and in what proportion — what would they conclude the book believes?
If the answer is the book you meant to write, the architecture is right, whatever the tests say. If it is not, no amount of anchoring will repair it.