There is a particular unease that comes when a story refuses to end. Not the frustration of a cliffhanger engineered to sell the next installment, but something quieter: the sense that a narrative has simply stopped attending to its own closure, leaving a hollow where a conclusion ought to be. It is the difference between a locked door and a door that was never hung. Most of us are trained to read this as a defect. We reach for resolution the way a hand seeks a railing on an unlit stairway—automatically, with a small spike of anxiety when it finds nothing. But what if that anxiety is not a sign of failure, but of friction worth preserving?
Consider two objects. The first is a finished novel: spine creased, last page turned, its world sealed and complete. The second is the Sagrada Família in Barcelona, still under construction after more than 140 years, its towers caught in various states of emergence, cranes rotating above the nativity facade like permanent fixtures of the design. One is a product; the other is a process that has outlasted its original architect, its original patrons, and the historical context of its founding. Yet the Sagrada Família is not a failure of completion. It belongs to a different category of object entirely—one that generates meaning precisely because it remains unfinished. Gaudí understood this. He reportedly remarked that his client was in no hurry. The building’s incompleteness is not a bug to be resolved but the condition of its ongoing life. Each generation of architects and stonemasons must interpret the fractional plans Gaudí left behind, making decisions that are simultaneously faithful and inventive. The result is a structure that demands participatory engagement from everyone who encounters it, because the mind cannot help but complete forms the eye cannot find.
I want to suggest that this participatory friction—the cognitive and emotional labor of inhabiting an incomplete system—is something we are quietly losing, and that the loss matters for more than cathedrals. It matters for how we design products, how we structure arguments, how we plan cities, and how we teach ourselves to tolerate uncertainty. The availability of instant narrative resolution, whether through formulaic screenwriting or the kind of automated text generation that now promises to eliminate writer’s block, threatens to make incompleteness feel like a mistake rather than a feature. When you can press a button and receive a tidy ending, the untidy one starts to look like poverty instead of restraint.
The Unfinished as a Formal Strategy
Not all incompleteness is created equal. There is a distinction between a work that is unfinished because its maker died or lost interest, and a work that incorporates incompleteness as a deliberate formal strategy. The films of Abbas Kiarostami operate in this latter territory. In Taste of Cherry (1997), the narrative follows a man driving through the hills outside Tehran, searching for someone who will bury him after he commits suicide. The film ends not with a resolution to this search but with a sudden cut to grainy video footage of the crew filming the scene, Kiarostami directing his actors, the camera pulling back to expose the artifice. This is not a cheat. It is an ethical gesture. By refusing to grant closure to the protagonist’s dilemma, Kiarostami returns the question to the audience. The film does not tell you what to feel about suicide, mortality, or compassion. It leaves the argument unfinished so that you must finish it yourself, in your own mind, with your own moral materials.
This is the opposite of what most narrative structures are engineered to do. Commercial storytelling—whether in Hollywood blockbusters, prestige television, or the fables we tell in corporate strategy decks—tends toward resolution with the gravitational pull of a well-designed drain. Loose ends are tied. Character arcs complete their promised trajectories. The story’s world is returned to equilibrium, or at least to a new equilibrium that feels legible. There is nothing inherently wrong with this. But it trains us to expect that all systems should behave the same way: that every input should produce a clear output, that every tension should be resolved, that ambiguity is a temporary state on the road to clarity. When we encounter a system that refuses this logic—a film like Kiarostami’s, a building like Gaudí’s, a city plan that embraces organic growth over master-planned finality—we may misread its incompleteness as deficiency rather than design.
Metabolism and the Architecture of the Unfinished
In 1960, a group of Japanese architects released a manifesto titled Metabolism: The Proposals for a New Urbanism. Their central idea was that buildings and cities should be designed not as fixed monuments but as living systems capable of growth, adaptation, and partial decay—like organisms. The metabolists envisioned structures with permanent cores and replaceable modules: apartment units that could be swapped out as needs changed, infrastructure that could expand incrementally rather than in disruptive master-planned phases. The most famous built example, Kisho Kurokawa’s Nakagin Capsule Tower (1972), consisted of a central concrete core onto which individual prefabricated capsules were bolted, each intended to be replaceable over time. The capsules were never replaced. The building fell into disrepair and was eventually demolished in 2022. But the metabolic idea did not fail because it was a bad idea. It failed because the economic and regulatory systems surrounding it demanded completion and stasis rather than ongoing adaptation. A building always in process is hard to finance, hard to insure, hard to sell. It does not photograph well for real estate listings. It resists the closure-obsessed logic of property markets.
Yet the metabolic vision persists in less literal forms. Any city that works—that is, any city that remains livable across centuries—is a metabolic city whether or not its planners intended it. Neighborhoods densify and thin out. Storefronts change uses. Infrastructure layers accumulate: a tram line beneath a bike lane beneath a fiber-optic cable, each installed in a different decade, each operating on a different maintenance cycle. The city is never finished. It is always in a state of fractional completion, and its resilience depends on this fact. A city that could be optimized to a final state would be a city that had stopped responding to its inhabitants. The incompleteness is the operating system.
This is not an argument against planning. It is an argument against the assumption that the endpoint of planning should be a finished object. The best urban plans—think of the 1811 Commissioners’ Plan for Manhattan, which laid down a grid that could accommodate centuries of incremental infill, or the flexible zoning codes of Tokyo that allow residential and commercial uses to mix in ways that keep neighborhoods adaptable—are frameworks for incompleteness. They provide scaffolding rather than a completed structure. They make room for the partial, the provisional, the not-yet-decided.
The Closure Machine
Into this landscape enters the automated story generator. It is a seductive proposition: you supply a premise, a few characters, perhaps a tone or a genre, and the machine returns a complete narrative with beginning, middle, and end. The friction of composition—the staring at a blank page, the false starts, the discovery halfway through that the story you thought you were writing is not the story you are actually writing—drops to near zero. You get resolution on demand. You get an output that looks like a finished product.
The question is not whether these tools can produce something that resembles a story. They can. The question is whether the resemblance matters. A story is not merely a sequence of events that reaches a conclusion. It is a structure of attention, a way of holding certain possibilities open and closing others off, a negotiation between what the writer intends and what the reader brings. When a writer sits with an unfinished draft—when the ending will not arrive, when a character refuses to behave, when the logic of the plot unravels into something more interesting than the outline promised—something is happening that is not visible in the final text. The writer is practicing a kind of negative capability, to borrow Keats’s phrase: the capacity to remain “in uncertainties, mysteries, doubts, without any irritable reaching after fact and reason.” This capacity is not a romantic luxury. It is a cognitive skill with applications far beyond literary fiction. It is the same skill that allows a product designer to sit with an unresolved user need rather than shipping a premature feature, or a strategist to resist the false clarity of a five-year plan that ignores the nonlinear dynamics of the market.
If you never practice sitting with the unfinished—if every narrative you encounter resolves itself on command—you may lose the muscle that makes that sitting possible. The Authors Guild, in its guidance on AI best practices for authors, emphasizes that human authorship involves judgment, intentionality, and a relationship to the work that automated generation cannot replicate (AI Best Practices for Authors — The Authors Guild). The point is not that AI-written stories are aesthetically inferior (though they often are, in ways that are tedious to catalog). The point is that the process of writing—the incomplete, iterative, uncertain process—is itself a form of thinking that the product cannot replace. When we outsource the resolution of narratives to a machine, we are not just getting a story faster. We are bypassing the cognitive work that the unfinished demands.
The Generative Power of the Missing
What is lost when a story ends? The obvious answer is nothing; the story is complete, and you can move on to the next one. But this misses something important about how narrative meaning accumulates. A resolved story is a closed system. Its meanings are bounded by the choices the author made. An unresolved story, by contrast, is a territory with open borders. The reader’s mind moves through it differently, filling gaps, testing interpretations, constructing endings that the text refuses to supply. This is not a passive experience. It is participatory in a way that finished products rarely are.
Consider a phenomenon from a completely different domain: the half-finished bridge. In infrastructure projects, a partially constructed span that is abandoned—whether due to funding collapse, political upheaval, or engineering failure—becomes a different kind of object than either a completed bridge or a bare river crossing. It invites speculation. What was the intended route? Who would have used it? What assumptions about traffic, trade, or population growth were embedded in its design? The incomplete bridge is a narrative artifact as much as an engineering one. It tells a story about ambition, constraint, and the moment when a system stopped making sense. A finished bridge tells a much simpler story: here is a crossing. The half-finished bridge is more interesting, not because it is more useful, but because it requires more from us.
Something similar happens in creative writing pedagogy. The Purdue OWL’s resources on creative writing emphasize the workshop model, in which student writers share unfinished drafts and receive feedback that is deliberately non-prescriptive—suggestions that open possibilities rather than foreclosing them (Creative Writing Introduction — Purdue OWL®). The goal is not to help the writer produce a perfect story on the first attempt. It is to teach the writer to tolerate the discomfort of the partial, to see revision as a form of thinking rather than a form of error correction. This pedagogical approach has structural parallels to the metabolic architecture of the 1960s: it provides scaffolding (the workshop, the instructor’s prompts, the peers’ questions) while leaving the final form deliberately underdetermined. The best workshops do not tell you what your story should be. They help you see what it could be, and then leave you to decide.
Resolution and Its Discontents in Systems Thinking
The preference for resolution is not confined to narrative. It shows up wherever humans design systems. A project plan that maps every milestone from kickoff to delivery feels more professional than one that acknowledges uncertainty. A strategic framework that promises a clear path from current state to desired state is more likely to win buy-in than one that admits the path will shift underfoot. And a tool like an AI story generator that outputs a complete narrative in seconds feels more powerful than a blank document and a half-formed idea. But in each case, the appearance of completeness may be masking a loss of adaptive capacity.
In systems theory, there is a useful distinction between closed and open systems. A closed system exchanges no matter, energy, or information with its environment; it trends toward entropy and equilibrium. An open system stays alive by remaining permeable, by exchanging resources with its context, by never reaching a final state. Most of the systems that matter—ecosystems, economies, organizations, minds—are open systems. They require incompleteness to function. A forest that reached a “final” climax state would be a forest that had stopped responding to fire, drought, and invasion. An organization that optimized every process to a state of perfect efficiency would be an organization that had lost the ability to adapt to surprise. A mind that demanded resolution for every uncertainty would be a mind incapable of creative thought.
This is why the case against resolution is not a case for laziness or vagueness. It is a case for designing systems—narrative systems, architectural systems, organizational systems—that preserve the generative friction of the incomplete. It is a case for leaving some arguments unfinished, not because we cannot finish them, but because finishing them would close off the thinking they make possible. It is a case for valuing scaffolding over monuments, process over product, the half-known over the fully mapped.
Living Inside the Scaffolding
What would it mean to take this seriously in daily practice? Not as a philosophy to admire from a distance, but as a set of habits and defaults that shape how you work, write, and decide? Here are a few entry points, offered not as prescriptions but as invitations to experiment.
First, notice where you reach for resolution prematurely. When a conversation hits an uncomfortable ambiguity, do you rush to fill it with a conclusion? When a project deliverable is 80 percent complete, do you push it to 100 percent because the remaining 20 percent would be the hardest part, or because an unfinished thing feels unprofessional to share? The 80 percent threshold is not a rule; it is a signal. It marks the point at which the marginal cost of additional completeness starts to exceed the marginal value of the learning that comes from shipping the partial version and seeing how it lands.
Second, build practices that normalize the unfinished. Keep a document of open questions rather than a list of resolved answers. Share drafts that are deliberately incomplete, with annotations that explain what is missing and why. In meetings, resist the urge to synthesize every divergent viewpoint into a tidy consensus; sometimes the most useful outcome is a clearer articulation of the disagreement. These practices are small, but they accumulate. Over time, they shift the culture around you from one that treats incompleteness as embarrassment to one that treats it as a working condition.
Third, cultivate a relationship with objects that refuse resolution. Spend time with a Kiarostami film. Visit a building that is still under construction after a century. Read a novel that ends in a way that does not satisfy. These encounters are not merely aesthetic experiences. They are training in a kind of attention that the resolution-obsessed world is constantly eroding. They remind you that not everything needs to end, and that some things are better for not ending.
The Sagrada Família is expected to be completed in 2026, the centenary of Gaudí’s death. When the last tower is finished, something will be gained and something will be lost. The building will become a completed monument, legible in a way it has never been. But it will also lose the particular charge of the unfinished—the sense that the building is still making up its mind about what it wants to be. One hopes that the completion is handled with the same intelligence that has guided the construction so far: that the architects understand they are not resolving a problem but closing a chapter of a story that will continue, in other forms, long after the cranes come down. The scaffolding will be removed, but the thinking that the scaffolding made visible should remain.