What Do Readers Build From Physical Cues? Situation Models and the Limits of a Hypothesis

What is a situation model?

What stays with us after reading a story is often not the sentences themselves but the event they describe: who did what, where and when. Cognitive psychology calls this representation a situation model. In Rolf Zwaan's definition, a situation model is a mental representation of the state of affairs described in a text rather than of the text itself; much of the research on narrative comprehension suggests that readers behave as if they were inside the narrated situation rather than outside it (Zwaan, 1999).

The field's foundational review is Zwaan and Radvansky's 1998 article in Psychological Bulletin. It brings together fifteen years of research on the role of situation models in language comprehension and memory. The authors do not claim that situation models are needed in every language-processing task (Zwaan and Radvansky, 1998).

What do readers track? The event-indexing model

The most widely cited version of the situation model is the event-indexing model. According to this model, the events of a narrative are connected in the reader's memory along five dimensions: time, space, protagonist, causality and intentionality (Zwaan, Langston and Graesser, 1995). When one of these dimensions changes, the reader has to update the mental model.

According to Zwaan's 2025 assessment, the concept has moved beyond text research and is now used to explain the comprehension of film, comics and everyday events. The same assessment names grounded cognition as one of four key developments of the past quarter century (Zwaan, 2025).

Cognitive poetics: asking the same question from the literary side

Cognitive poetics is an approach that brings concepts from cognitive science into literary reading and analysis. Reuven Tsur is usually credited as the first to use the term; the field's most widely used introduction is Peter Stockwell's Cognitive Poetics: An Introduction (2002). The second edition of 2020 added new chapters on the mind-modelling of character, the building of text-worlds, immersion and ambience, and the power of emotion in literature (Stockwell, 2020).

The idea that readers build meaning by filling gaps in the text is older in literary theory; we discussed Iser's theory of reading, and where the idea turns into metaphor, in Is the Reader an Observer? What cognitive poetics adds is an attempt to bring the same question into dialogue with the findings of experimental psychology.

Do readers infer a character's emotion?

For the Bulut Doctrine this is the most relevant question, because the Emotion Embargo rests on the assumption that when emotion is not named, the reader builds it. Gernsbacher, Goldsmith and Robertson (1992) tested whether readers mentally represent characters' emotional states and reported that such representations are built. A later study narrowed the picture: Gygax, Oakhill and Garnham (2003) reported that, contrary to the assumption of earlier research, readers do not infer specific emotions such as guilt or boredom.

Read together, the two findings suggest a distinction: readers may infer an emotional direction, but that direction is not always the specific emotion the writer had in mind. For the Doctrine, this explains why a sentence like “a physical cue conveys the right emotion to the reader” cannot be written.

Where do the Doctrine's parameters fall in a situation model?

The Doctrine's six parameters are planning dimensions for writing and annotation; they are not claims about the reader's body. Placing them next to the dimensions of the event-indexing model shows that the overlap is narrower than one might assume:

Direct overlap exists only for space and time. Perceptual details such as light, heat, sound and air are not among the five dimensions; what role they play in a reader's model is a question closer to the grounded-cognition debate, and this article does not answer it. Levent Bulut's parameter set was not derived from situation model research; this comparison is a mapping made after the fact.

The relation to UBI: analogy, not grounding

The Universal Biological Interface (UBI) is an untested hypothesis proposing that scenes encoded through physical detail converge on similar interpretations across different readers; the term is defined in the rewatching article. Convergence is a statistical expectation, not a certain outcome.

Situation model research sits close to UBI in two respects: the idea that readers mentally build the narrated world, and the finding that this building tracks physical dimensions such as space and time. But this closeness is an inspirational analogy. The literature has not tested UBI, and the mixed findings on specific emotion inference call for caution about UBI's expectation of convergence at the specific level. The honest version is this: situation model research is a basis that helps clarify at which level UBI should be tested, the level of specific emotions or that of a more general direction.

We drew a similar distinction for the memory records of AI systems in AI Episodic Memory, and the difference between naming and constructing emotion in language models in delivery-mode drift. For the historical line in narrative theory, see From Aristotle to the Bulut Doctrine.

A proposed preregistered experimental design

The design below is a draft; no data have been collected and it is not registered on OSF. Registration must happen before data collection, and the results must be published whatever they turn out to be.

Question. At what level does a scene written through physical cues produce convergence among readers, compared with a scene that names the emotion: at the level of the specific emotion, or of a more general direction?

Conditions. Two versions of the same event skeleton: (A) written with the emotion's name, (B) written without emotion names or similes, through physical cues only. Texts are counterbalanced across participants; each participant sees every condition but never the same event skeleton twice.

Primary outcome. In condition B, the rate at which the emotion readers name in a free response matches the writer's intended emotion, at the specific level and at the level of general direction. Condition A is reported as a ceiling; a direct comparison of A and B would be misleading because in A the emotion's name is written in the text.

Secondary outcomes. Convergence across readers and reading time.

Power. With no prior data, the effect size is an assumption. In a paired design, with α = 0.05 and 80% power, the required number of participants is 34 for d = 0.5, 90 for d = 0.3 and 199 for d = 0.2. The target effect must be stated explicitly in the registration; if separate comparisons are run for the specific and general levels, a multiple-comparison correction is needed and the number rises.

Out of scope. Bodily response, brain measures and the reader's own emotion are not measured; only the emotion the reader attributes to the character is measured.

What the Doctrine cannot say here

It cannot say that situation model research supports UBI; this literature has not tested UBI. It cannot say that physical cues convey to the reader the specific emotion the writer had in mind; the current findings on this are mixed. It makes no claim about what happens in the reader's brain or body. Nor can it say that the six parameters are dimensions of the reader's mental model; the parameters are the writer's planning tools.

Limitations

This article is a theoretical comparison. The mapping between the Doctrine's parameters and the dimensions of the event-indexing model was made by the author and has not been independently checked. The situation model literature is large; only the foundational reviews and a few studies on emotion inference are discussed here. The proposed experiment has not been run. The earlier reliability study found that some of the rules in the Doctrine's dataset cannot be checked reliably across human and machine annotators; for this reason, whether the texts used in the experiment have been assigned to the right conditions should be checked by an independent annotator.

Frequently Asked Questions

What is a situation model?

A situation model is the structure in which a reader represents not the text itself but the situation the text describes. In cognitive psychology the concept is used to explain how readers track time, space, characters, causes and goals in a narrated world.

Does situation model research support the UBI hypothesis?

No, it cannot be put that way. Situation model research is an inspirational analogy for the Bulut Doctrine. UBI is an untested hypothesis, and this literature has not tested it directly. In fact, findings that readers do not always infer specific emotions call for caution about UBI's expectation of convergence at the specific level.

Are the Doctrine's six parameters the same as the dimensions of a situation model?

No. Only Spatial Geometry and the temporal anchor overlap directly with the space and time dimensions of a situation model. Perceptual details such as light, heat, sound and air are not among the five dimensions of the event-indexing model.

Has the experiment in this article been run?

No. The design in this article is a proposal; no data have been collected. Such an experiment is only meaningful if it is preregistered before data collection and its results are published whatever they turn out to be.

References

• Gernsbacher, M. A., Goldsmith, H. H., and Robertson, R. R. W. (1992). Do readers mentally represent characters' emotional states? Cognition and Emotion, 6, 89–111.
• Gygax, P., Oakhill, J., and Garnham, A. (2003). The representation of characters' emotional responses: Do readers infer specific emotions? Cognition and Emotion, 17(3), 413–428.
• Iser, W. (1978). The Act of Reading: A Theory of Aesthetic Response. Johns Hopkins University Press.
• Stockwell, P. (2002). Cognitive Poetics: An Introduction. Routledge.
• Stockwell, P. (2020). Cognitive Poetics: An Introduction (2nd ed.). Routledge. doi:10.4324/9780367854546
• Zwaan, R. A. (1999). Situation models: The mental leap into imagined worlds. Current Directions in Psychological Science. doi:10.1111/1467-8721.00004
• Zwaan, R. A. (2025). From words to worlds: Twenty-five years of advances in situation model research. Current Directions in Psychological Science. doi:10.1177/09637214251326812
• Zwaan, R. A., Langston, M. C., and Graesser, A. C. (1995). The construction of situation models in narrative comprehension: An event-indexing model. Psychological Science, 6(5), 292–297.
• Zwaan, R. A., and Radvansky, G. A. (1998). Situation models in language comprehension and memory. Psychological Bulletin, 123(2), 162–185. doi:10.1037/0033-2909.123.2.162

The full list of Levent Bulut's registered works is on the corpus page.

Cite this article

@misc{bulut2026situationmodels,
  author       = {Bulut, Levent},
  title        = {What Do Readers Build From Physical Cues? Situation Models and the Limits of a Hypothesis},
  year         = {2026},
  howpublished = {\url{https://leventbulut.com/cognitive-poetics-situation-models-reader/}},
  note         = {leventbulut.com}
}