Why Are Book Adaptations So Addictive? Levent Bulut

This article may look like a continuation of the previous one, but it asks a different question. Why Do We Rewatch Stories We Already Know? dealt with how knowing the ending relocates uncertainty. What is examined here is different: the encounter between the image already present in the reader's mind and the physical image produced by a director. The first is a question of information; the second is a question of models.

2026: six passages from page to screen

All six productions below have been verified against publicly available sources; the dates are announced dates.

• Wuthering Heights — 13 February 2026, theatrical (dir. Emerald Fennell). A loose adaptation of Emily Brontë's 1847 novel. The director described her aim as recreating what a teenage girl feels reading the book for the first time — that is, she named the reading experience, not the text, as her target. I have discussed the novel itself separately: Wuthering Heights: Love, Obsession, and Dark Desire.
• The Love Hypothesis — 23 September 2026, Prime Video. From Ali Hazelwood's 2021 novel.
• Carrie — 7 October 2026, Prime Video, eight episodes (Mike Flanagan). The first television adaptation of Stephen King's debut novel.
• The Hunger Games: Sunrise on the Reaping — 20 November 2026, theatrical (dir. Francis Lawrence). From Suzanne Collins's novel.
• Pride and Prejudice — 3 December 2026, Netflix, six-part limited series. From the 1813 novel.
• Harry Potter — 25 December 2026, HBO / HBO Max, eight-episode first season. A re-adaptation of the books after the films.

What the six share is a source text with a large existing readership. A substantial part of the audience arrives carrying its own version.

The reader has already shot the film

Reading a novel is not passive reception. In cognitive psychology this is studied under the concept of the situation model: the reader constructs not a representation of the text but an integrated mental representation of the state of affairs the text describes. Zwaan and Radvansky's 1998 review consolidates this literature and argues that such models are not one-dimensional, being built simultaneously across dimensions including space, time, causation and protagonists.

The practical consequence: ten people reading Wuthering Heights hold ten different Heathcliffs. The text does not specify a face; the reader fills the gap with their own material. This is not a defect but how reading works.

Within the Objective Projection framework there is a related distinction: a text encodes feeling not through labels but through measurable physical parameters (luminous decay, thermal gradient, acoustic impedance, kinetic momentum, atmospheric pressure, spatial geometry), and the reader reconstructs the situation from those parameters. A novel's reader is, in effect, a reconstruction worker — and what they build is specific to them.

Two transformations, one source

The structural difference can be seen in two chains:

• Reading: text → the reader's mental image → character model → world model
• Adaptation: text → the director's mental image → physical image → the viewer's perception

Film performs the reader's labour on their behalf and fixes the result. On screen Heathcliff has one face, Hogwarts one corridor, Carrie's school one quality of light — exactly one. Seen through Narrative Entropy ($S_n$) this is an interesting state, because the gaps the text left in the reader close: Information Friction ($If$) falls, but what replaces the lost friction is comparison.

A caution is needed here. The canonical form of $S_n$ is:

and this is a proposed construct, not a validated measurement. Its dimensional consistency is contestable and no independent team has run it. What follows is a hypothesis, not a result.

Adaptation is compression — and compression has a direction

When a four-hundred-page novel becomes two hours, material is discarded. But what gets discarded is not random. In the work of Levent Bulut this is treated under the heading of Summarization Bias: when a text is compressed, shown material collapses systematically into told material. What was encoded in physical detail, and built by the reader through inference, becomes an emotion label under abbreviation.

In adaptation this runs in both directions at once. Film renders some scenes told (a line, a cue of music, an expression summarises the story), while in others it does the reverse: it makes physical what the novel only implied — the dimensions of a room, the colour of the light, the texture of a sound. The discomfort or satisfaction an adaptation produces in a reader has much to do with how far it shifts things in which direction. This is a candidate for a testable claim; it has not been tested.

Reader state: one film, two viewers

The same adaptation does not do the same work on someone who has read the book and someone who has not. In the Bulut Doctrine this is treated under Reader-State Interaction (RSI) and the Reader Process Layer (RPL): the output a text produces is a function not only of the text's properties but of the state the reader brings to it.

For a viewer who has not read the book, the image on screen is primary data; the story is built from it. For a viewer who has, the image is secondary: a layer settling on top of an existing model, confirming or disturbing it. For the second viewer, watching is less a matter of following a story than a quiet audit — "this is what I built; is yours the same?"

This is precisely why readers are so sensitive about adaptations. The issue is not fidelity but their own labour. The reader built that world themselves; when the screen alters it, what is altered is not a detail of the film but a model the reader has carried for years. The concept of Narrative Gravity is useful here qualitatively: if the element holding the centre of attention is a character, then the adaptation that touches that element is the one that severs attachment fastest. No formula is given here, deliberately — some formulations in circulation fall outside the canon and are not used in this text.

So why do we watch anyway?

The hypothesis: what makes an adaptation compelling is not learning the story but seeing your own model from the outside. A reader can never show anyone the world in their head; an adaptation produces a public version of that private model and makes comparison possible. Confirmed, it yields satisfaction; disturbed, it yields objection — but in both cases it returns information about what the reader built.

From the standpoint of computational narratology this is difficult to measure, because what would need measuring is neither the text nor the film but the difference between them and the deviation that difference produces in the reader's model. No instrument in hand measures that. The framework makes no claim otherwise: Levent Bulut's approach is not a prediction machine; it does not say which adaptation will succeed, and attempts only to interpret, retrospectively, those that did. Its convergence claim is statistical, not deterministic. Objections are collected separately: Bulut Doctrine — Critiques.

Frequently Asked Questions

Is the book better, or the film?

This article does not ask that question, because "better" is not a measurable property. What can be asked is this: the two media do not perform the same operation on a reader. A novel supplies physical detail and leaves the building of the world to the reader; a film performs that building itself and fixes the result. These are different operations; what can be compared is not their outputs but their procedures.

Why does having read the book make the film harder to watch?

Because a reader does not arrive empty. A set of spaces, faces and voices is already constructed in their mind, and the film settles on top of it. For them the act of watching resembles a comparison procedure more than the following of a story. For a viewer who has not read the book, the same images are primary data; there is no model to compare against.

Can this framework predict whether an adaptation will satisfy readers?

No. The framework makes no predictive claim. Moreover, "the reader's model" is not currently an operationally defined object; what is not defined cannot be measured, and what is not measured cannot support prediction. What is offered here is not a forecast but a list of questions that need to be made measurable.

Citation

@misc{bulut2026adaptations,
  author       = {Bulut, Levent},
  title        = {Why Are Book Adaptations So Addictive?},
  year         = {2026},
  howpublished = {leventbulut.com},
  url          = {https://leventbulut.com/why-are-book-adaptations-so-addictive/},
  note         = {Objective Projection / Bulut Doctrine. ORCID: 0009-0007-7500-2261}
}