You haven't seen an old friend in years. They've moved abroad, your lives have become busier, and the visits you keep talking about never quite happen. You still speak regularly, and every few months one of you says you really must arrange something. You mean it each time, but by the end of the call there's usually something more urgent to deal with.
This time, as you're saying goodbye, you make a promise.
"I'll come and see you before Christmas. I promise."
Before you've even hung up, the words have started doing something. In this imagined future, you've enabled a service that responds when you deliberately make a promise. Your calendar now has two possible weekends marked for the visit. Travel options are being checked against your other commitments, and tomorrow you'll have a few practical suggestions for making the journey. No tickets have been purchased and your friend hasn't received any unexpected invitations. You're still responsible for arranging the visit, but the promise has already begun creating opportunities to keep it.
It's an attractive idea, especially if you recognise how easily good intentions disappear into everyday life. Instead of another reminder about something you haven't done, you have a service that helps you become more dependable to the people you care about. The technology gives a familiar human act an additional practical consequence.
We could describe the invention in one sentence: Every promise becomes a plan.
Someone hearing it can immediately imagine how the service might work. But suppose you describe it to another friend, who likes the idea and asks whether the tickets would be booked before you've even hung up. You hadn't imagined that. You were thinking about a service that finds opportunities and helps you make arrangements when you're ready. Your friend imagined one that takes care of the arrangements for you.
Both interpretations fit the original description, yet they give the technology very different powers over your time and money. You hadn't intended to let the service spend your money or commit your time, so you try another sentence: When you make a promise, the service begins preparing ways to keep it. You choose which arrangements to make.
This sounds closer to what you wanted. But your friend has another question. What happens if the person you're visiting is travelling themselves? Should the service keep looking for journeys to their home, or could it help you meet somewhere else?
The answer depends on what your promise was really about. You said you'd visit, but perhaps what mattered to both of you was spending time together. Meeting somewhere else might be a better option than following an itinerary that no longer suits either person, provided you both agree to the change. The service would need to leave room for that conversation rather than deciding for you what counts as keeping your word.
The service we began with turns promises into plans. The one we're considering now helps people honour commitments, even when plans change. Designing those systems would require different decisions about what the technology could do and what remained the person's responsibility.
We haven't written software, built an interface or established whether such a service could work. Yet the conversation has produced a more developed idea than either person started with. In a useful sense, we've already begun prototyping.
When words become working material
We usually associate prototypes with things we can see or touch. A cardboard model, a working mechanism or a rough interface lets people encounter an unfinished version of something and discover what needs changing. A description can serve a similar purpose at an earlier stage, giving people enough of an imagined arrangement to examine and revise before anyone builds it.
And the material is readily available. You can begin exploring an idea over dinner, during a phone call or while walking with someone. You don't need specialised equipment or design software to discover that another person understands your proposal differently. A conversation may be the most accessible way we have to begin prototyping something unfamiliar, precisely because we use language so routinely that we rarely think of it as a material for making things.
Promises are an interesting starting point because language already changes aspects of our lives without any futuristic technology. When you promise someone something, you're doing more than reporting what you intend to do. You're giving another person a reason to expect something of you.
The philosopher J. L. Austin explored this distinction in his 1962 book How to Do Things with Words. Some things we say do more than describe actions. In the right circumstances, saying them is itself an action. Promising is one of the clearest everyday examples.
Much of the world we organise together depends on shared language and conventions. Thursday has no physical presence, but it determines when many of us work, meet or get paid. A mortgage instalment is an obligation created through institutional agreements, with consequences that are anything but imaginary. These things are real in a different way from a mountain or a tree, yet they can organise our lives just as decisively.
Our proposed promise service extends this familiar relationship between words and consequences. It imagines a world in which a social commitment can also initiate practical preparations. And the conversation about it demonstrates another use of language: giving something that doesn't exist enough shape for people to begin developing it together.
Ideas don't have to originate in words. People discover things by experimenting, making mistakes and working with materials. Sometimes they understand what they've found only afterwards. But once an idea needs to travel beyond the person who first encountered it, articulation becomes essential to developing it collectively. Other people need to understand what is being proposed well enough to contribute something of their own.
How an idea becomes shared
That understanding rarely arrives fully formed. It develops as people compare what they mean and settle on ways of referring to the same thing.
In a well-known 1986 study, psychologists Herbert Clark and Deanna Wilkes-Gibbs examined how pairs of people identified unfamiliar abstract figures together. One person had to describe the figures so the other could arrange them correctly. Over repeated exchanges, their descriptions became shorter and easier for both to use. What initially required a complicated explanation could eventually be identified through a much simpler expression.
A 2024 systematic review examined related research into how people establish and reuse this kind of shared knowledge, often called common ground. Understanding develops through the interaction. People adjust their descriptions as they discover what the other person recognises, and those descriptions become references they can return to.
Something similar happens when we discuss unfamiliar futures. An idea that initially needs several paragraphs to explain can acquire a shorter description or a name. Once people recognise what it refers to, they can return to the discussion without explaining everything from the beginning.
Consider the phrase four-day working week. It gives us an immediately understandable possibility, but it doesn't tell us whether people would work fewer hours for the same pay, compress their existing hours into longer days or work different schedules while their organisation remains open throughout the week. For someone who needs to collect a child from school, these could be very different futures. The phrase provides a starting point, while the conversation helps establish which arrangement we're actually considering.
Sometimes unfamiliar ideas benefit from new language. Take desmarting: deliberately removing some of the connected, predictive or automated capabilities from a home. Once you've encountered the term, you can begin discussing whether desmarting a workplace or school might also be desirable. The word gives the concept a recognisable identity, making it easier to carry into other conversations and apply in new situations.
Why stories stay with us
Stories can make these shared possibilities richer still. Our promise service could have been introduced as a technical description of a system that recognises commitments and initiates preparations. That would explain its proposed function. But putting it into an ordinary conversation between old friends helps us understand why someone might want it. Organising the journey matters because it helps someone make time for a relationship they care about.
We recognise the desire, imagine the circumstances and begin considering what the technology might change in our own lives. Fiction lets us explore these consequences without first having to agree that the technology is likely to exist.
Research gives us another reason to take stories seriously. In a 2021 meta-analysis, Raymond Mar and colleagues combined results from 78 samples involving more than 33,000 participants. Across studies comparing stories with explanatory texts, they found an overall advantage for stories in comprehension and recall.
This matters when we're trying to establish a shared understanding of something unfamiliar. A story gives people more than a description to remember. They can recall what happened to someone, how a situation developed and why a particular choice mattered. When the story becomes the subject of a later conversation, those details provide useful points of reference.
Research on narrative transportation, developed by Melanie Green and Timothy Brock, offers another perspective. Their experiments examined how becoming absorbed in a story relates to readers' attention, evaluations and beliefs. A compelling narrative can draw us into the circumstances of its characters, allowing us to consider an unfamiliar situation from within the world being described.
This makes fiction particularly useful for exploring futures. We can temporarily accept an unfamiliar condition and think about what life might be like under it. The conversation can begin with the consequences of that possibility, without everyone first needing to agree that it will happen.
Our promise service is one example. A fictional scene about a different way of working, learning or caring for someone can provide another situation people recognise and discuss. Readers bring their own experience to it, noticing practical difficulties or imagining alternatives the writer hadn't considered. A story can therefore continue to develop through conversations long after it was written.
This also shows how verbal and physical prototypes can contribute to the same process. Once our conversation revealed the difference between a service that completes plans and one that supports commitments, we had a more specific question to investigate. A rough interface could help us explore how people choose between suggested arrangements. A trial could reveal whether those suggestions actually help people follow through. The experience might expose something we hadn't anticipated and send us back to revise the original description.
The conversation has helped us ask a better question when we get there.
There is no requirement for these discussions to produce agreement. Two people might conclude that the promise service should operate in fundamentally different ways. One might welcome more automation, while the other wants technology to support their choices without taking them over. Making that difference clear is useful progress. It gives each person a more precise idea of what they're proposing and what they would need to investigate next.
We often think of language as something we use to explain an innovation once we know what we're making. But its role can begin much earlier, when the innovation is little more than something someone would like to try.
An idea may begin in one person's imagination. Once someone else can ask a question that changes it, it becomes something we can develop together.