Ana Elisa Ulloa Labariega is an MA student in philosophy at Università della Svizzera italiana. Her work moves between philosophy of science, epistemology, and philosophy of emotion, focusing on how conceptual and interpretive frameworks shape different forms of understanding. She is particularly interested in scientific modeling and in the role of imagination as a narrative and structuring capacity through which phenomena become intelligible.
A post by Ana Elisa Ulloa Labariega
Imagine a mobile—the kind that hangs above a child’s bed. Glass figures suspended at different lengths, each positioned in relation to the others. The structure is not arbitrary: it has been built to support those particular elements, with their specific weights, distances, and tensions. At the same time, the structure itself takes shape in relation to what it is meant to hold; frame and elements are not fully independent, but emerge together as a single configuration.
For a moment, everything holds together. Each figure moves slightly, but never entirely on its own. Its motion is shaped by the others, and by the frame that sustains them all. In other words, the range of movement is delimited by the possibilities given by the whole.
But the balance is conditional.
If one figure becomes heavier, the strings must adjust. If new elements are added, the structure must redistribute tension. Changes are possible, but only up to a point. Once something shifts, it cannot simply be undone: it can only be rebalanced. At the same time there are limits to what it can sustain. At some point, the configuration no longer holds: the system strains, distorts, and eventually breaks. Pieces fall to the ground. Some may remain intact, but the mobile as it was—and the structure that held it together—cannot be recovered.
Scientific models, I want to suggest, are more like this than we usually admit.
Usually, models are understood as tools for explanation and prediction. But this way of putting things risks obscuring another dimension of modeling: its narrative character, sustained and structured through imagination.
To call modeling “narrative” is not to suggest that scientists tell stories in any literary sense. Rather, it is to recognize that models organize elements in a temporally and logically structured way, allowing us to explore possible states of affairs (Frigg & Nguyen 2022; Kind 2022). In this broader sense, narratives are not defined by plot or characters, but by their capacity to guide us through sequences of transformation under constraints (Kind 2022). Scientific models do something similar: they structure a space of possibilities and allow us to move within it.
Understanding a model, in this view, is not simply a matter of grasping a representation from the outside. It is more like moving within a structured space. One follows the relations it establishes, tests how far they can be extended, and senses when coherence begins to give way. What matters is not only what the model represents, but how it holds together—and how long it can continue to do so under pressure.
This becomes especially clear when we consider how models are actually used in scientific practice. They are not static objects, but tools for exploration. We adjust parameters, introduce variations, and consider alternative scenarios. Counterfactual reasoning, idealizations, and heuristic scaffolds are not isolated techniques, but ways of sustaining a space that can be navigated in an intelligible and coherent way. This kind of structured imaginative engagement has been increasingly emphasized in recent work on the epistemology of imagination (Kind 2022; Nanay 2016).
It is here that the narrative dimension becomes especially helpful. Counterfactuals are not just “what-if” statements: they are moves within a structured space of possibilities, as has been explored in work on imagination and make-believe (Walton 1990). Idealizations are not mere distortions; they help stabilize that space so that it can be explored without collapsing under its own complexity—think of the well-known example of spherical cows. Even highly abstract mathematical or computational models—such as climate simulations or population dynamics—encode temporal development, conditional dependencies, and transformations that can be followed step by step. They do not tell stories, but they make something like narrative traversal possible: a way of moving through a structured configuration, following its transformations, dependencies, and limits (Frigg & Nguyen 2022; Walton 1990).
This perspective also brings into view a dimension of modeling that is often treated with suspicion: the role of aesthetic judgment. Qualities such as simplicity, symmetry, beauty, and elegance frequently guide the construction and evaluation of models. These are sometimes dismissed as merely subjective preferences or biases. Yet, in practice, they often function as more than that, shaping how models are constructed and understood as structured systems.
A familiar example is Kepler’s early commitment to circular planetary orbits. Circles were associated with perfection: simple, symmetrical, and mathematically elegant. When observational evidence forced a shift to elliptical orbits, this is often described as a case in which aesthetic preference had to be abandoned. Nevertheless, the transition is more revealing than this story suggests. What Kepler gave up was not aesthetic judgment as such, but a particular conception of it. The ellipse does not eliminate order; it introduces a more complex and less immediately transparent form of it—one that accommodates irregularity while preserving coherence. In this sense, aesthetic criteria do not disappear with the change of model. They remain, but in altered form. What counts as elegance, symmetry, or beauty is itself reconfigured by the new structure.
The point, then, is not that science leaves aesthetics behind once better evidence arrives. It is that aesthetic judgment must sometimes be educated by the very models it helps to guide. A form that first appears less perfect may come to seem more compelling once one learns to perceive the order it makes possible. From this perspective, aesthetic features function as cognitive guides. A model feels “elegant” or “coherent” when it can be inhabited without excessive resistance—when one can move through it, follow its implications, and extend it without immediate breakdown. Conversely, what is often experienced as “ugliness” may reflect a kind of friction: a difficulty in navigating the structure or making its internal relations cohere. Yet this too can shift. What initially appears awkward or resistant may, with time, reveal a different form of order—one that becomes perceptible only once we learn how to move within it.
This tension is not a flaw in modeling; it is part of its epistemic function. Models are provisional structures. They allow us to explore possibilities, but they also expose their own limits. When a model begins to resist extension—when adjustments accumulate, when coherence starts to strain—we encounter a kind of epistemic friction. That friction is not merely a sign of failure; it is informative. It signals that the structure may need to be revised or replaced.
Seen in this light, scientific understanding is not a matter of achieving a final, stable representation of the world. It is an ongoing activity: constructing, shaping, navigating, and revising models that are at once enabling and constraining.
To describe modeling as narrative is, therefore, not to reduce it to storytelling, but to recognize its dynamic and imaginative character. Models organize elements across a space of possibilities, guiding transitions and constraining developments. They make certain paths intelligible while excluding others. Working with a model is to follow those paths, to test their limits, and sometimes to reshape the structure itself.
This perspective invites a shift in how we think about scientific knowledge. Models are not simply tools for getting things right. They are frameworks that allow us to engage with the unknown—to explore, to extend, and to recognize when a once-coherent structure can no longer be sustained.
Like the mobile, they hold together only under certain conditions. They can be adjusted, changed, and sometimes transformed. But they can also collapse. And when they do, what we lose is not just a representation, but a way of moving through a space of possibilities.
Understanding, then, is not just a matter of seeing the world correctly. It is a matter of learning how to move within structures that make the world intelligible—structures shaped by imagination, constrained by evidence, and guided, in part, by our sense of what holds together.
References
Kind, Amy (2022). Imagination and Creative Thinking. Oxford University Press.
Nanay, Bence (2016). Aesthetics as Philosophy of Perception. Oxford University Press.
Frigg, Roman, and James Nguyen (2022). Modelling Nature: An Opinionated Introduction to Scientific Representation. Springer.
Walton, Kendall (1990). Mimesis as Make-Believe. Harvard University Press.