The Conversations that Change Us
How futures grow from small beginnings
Most of us can identify conversations that changed the course of our lives. At the time they rarely seem important. A few words are exchanged. A suggestion is made. A question is asked. A remark is remembered. Sometimes nothing is said at all. Years later, however, we can often trace a line from those seemingly ordinary moments to decisions, opportunities, discoveries, relationships, and ideas that followed.
This observation is so familiar that we hardly notice it. Yet it raises an intriguing question. Why do some interactions disappear without trace, while others continue shaping our thoughts and actions decades later? Over the years I have become increasingly interested in that question.
An ordinary conversation
One such conversation occurred when I left school. My prospects did not look especially promising. I had few qualifications, no clear career plan, and little idea what sort of work I wanted to do. My first job, with a tyre company in the Cotswolds, lasted only two weeks. I was responsible for ordering tyres and kept confusing the long strings of letters and numbers used to identify different types. Eventually I was told that I “couldn’t work with numbers” and was dismissed. In reality, numbers were never the problem. I am dyslexic and had difficulty remembering long sequences of symbols, but neither my school, my employer nor I understood that at the time.
At that point my future seemed uncertain. However my father called me to join him in India where he worked as a schoolteacher. I spent several happy years living in a small village on the outskirts of Udaipur in Rajasthan while completing my education through a correspondence course. On my return to England, my godmother suggested that I apply for a job as a nursing assistant in a psychiatric hospital. It sounded like an ordinary piece of advice. Looking back, it remains one of the most consequential I ever received. Within a few years I had worked in several psychiatric hospitals, become fascinated by the variety of ideas expressed by the patients, developed an interest in how the brain works, and eventually decided to study psychology at university. A simple suggestion had opened an entirely new pathway.
When someone opens a door
Many years later, another interaction altered the direction of my work. I completed my PhD at the Institute of Psychiatry in London under Hans Eysenck. It was an extraordinary intellectual environment. Visiting scholars from around the world mixed freely with staff and students, and ideas seemed to matter more than status. After moving to a lectureship at the Institute of Education, I found the atmosphere very different. My publication rate slowed, and many of the questions that interested me attracted little enthusiasm as they were ‘not about education’.
Then I met Ulric Neisser. By then Neisser was already famous as one of the founders of cognitive psychology. What impressed me was not his reputation but his willingness to rethink his assumptions. He had become fascinated by the ecological ideas of James Gibson and was exploring new intellectual territory himself. He took my ideas about schizotypal thinking seriously. He listened, asked questions, and introduced me to scholars who were interested in related problems, including philosophers and cognitive scientists such as Daniel Dennett and George Lakoff. I do not remember a single decisive piece of advice. What I remember is the experience of having an idea taken seriously. New conversations followed. New collaborations emerged. Eventually, through a chain of events that nobody could have predicted, I found myself founding the journal Philosophical Psychology. Again, the trajectory changed not because someone gave me an answer, but because they opened a door.
The traces we do not speak
Not all traces arise from what is said. My wife Susan, whose research focusses on non-traditional families, recently suggested another possibility. Imagine a child growing up in a family where their donor conception has been kept secret. The secret may never be mentioned directly. Yet whenever subjects such as ancestry, family resemblances, medical history, or family trees arise, something changes. A pause. A silence. A sudden change of subject. Someone leaves the room. Nothing explicit has been communicated. Yet the child may gradually learn that certain questions carry a different emotional weight from others. In such cases, the trace is not a conversation but a disturbance in the flow of conversation. A silence, too, can become a pathway.
The Shape of the Ask
More recently, I encountered a fourth example while exploring ideas with GPT-5.5. During one exchange, a phrase emerged that neither of us had anticipated: “The Shape of the Ask”. The results of that exchange were later published in a Substack essay under the same title. At first it seemed little more than an intriguing turn of phrase. Yet I found myself returning to it repeatedly. What exactly was an ask? Was it merely a question or a request for information? Or was it something deeper: the form taken by curiosity itself?
If the ask had a shape, what could that possibly mean?
The question lingered long after the original conversation had ended. The more I thought about it, the stranger it became. Questions are not all alike. Some close possibilities down; others open them up. Some seek confirmation of what we already believe; others invite us to explore territory we do not yet understand. Gradually another possibility emerged. Perhaps an ask is not simply a thing but a movement. A good question does not merely request information. It changes what happens next. It alters the direction of thought. It opens some pathways while closing others.
If so, asks may have shapes because they generate trajectories of thought, and trajectories can indeed be described. They have direction. They can diverge, converge, accelerate, stall, or change course. Looking back, we can often trace the path that followed a particular question and ask, “What followed from that?”
From there the idea developed further. Perhaps conversations do not simply exchange information. Perhaps they reshape the questions available to the participants. A good interaction may leave us asking better questions than the ones with which we began.
None of those ideas were obvious when the phrase first appeared. Yet the phrase provided a form through which they could emerge. The conversation left behind something that continued developing long after the interaction itself had ended. Again, what remained was not an answer. It was a question. And perhaps questions are one of the most powerful forms that traces can take.
From traces to trajectories
At first glance, these examples seem unrelated. A career suggestion. An academic conversation. A family secret. A phrase that emerged during a dialogue with an AI and later inspired a Substack essay. Yet they share a common feature. Each altered the future by changing the landscape of possibilities. We normally think of conversations as exchanging information, advice or opinions. What interested me was something slightly different. In each example, the interaction seemed to leave behind a change in what became possible afterwards.
My godmother’s suggestion opened a path that eventually led to psychology. Ulric Neisser’s encouragement connected me to a wider intellectual community and a set of conversations that would shape much of my later career. The silences surrounding a family secret may influence which questions a child feels able to ask. And The Shape of the Ask generated a question that continues to influence my thinking today. In each case, something remained after the interaction itself had ended. I have come to think of such remnants as traces.
A trace is not simply a memory. A memory records something that happened. A trace alters what may happen next. Sometimes the effect is obvious; often it is invisible. Simply changing the order of two questions in a survey or interview can alter the way each is interpreted. A metaphor can quietly reshape the way an idea is approached. A forgotten conversation can leave behind a possibility that only becomes visible years later. Traces are the marks left in the landscape of thought; trajectories are the paths that later grow through them.
Nor is a trace merely information. The same information may leave a profound mark on one person and none at all on another. A trace is something that continues. Whether or not it survives in memory, it also exists in the altered landscape of possibilities that follows from it. It persists beyond the interaction that created it and continues to influence what happens next.
Perhaps that is why traces are difficult to classify. Psychologists, sociologists, linguists and philosophers have each captured part of the picture, yet none seems quite sufficient. What interests me is not merely what an interaction communicates but what it makes possible. Seen in this way, a trace may be understood as a change in a possibility space.
Before the interaction, certain possibilities may not even be visible. Afterwards they become available. The interaction does not determine the future. It reshapes the terrain through which the future unfolds. The future does not emerge from nowhere. It grows from the traces left by previous interactions. The same trace may lead to very different outcomes, but it changes the probabilities of what becomes possible next.
We might think of traces and trajectories as two views of the same phenomenon. A trace is what remains from an interaction when we look backwards. A trajectory is what unfolds from it when we look forwards. The trace is the residue of a path already taken; the trajectory is the path still being formed. If traces alter the landscape of possibilities, then trajectories are the routes through that altered landscape.
This observation may help explain why transformative interactions rarely feel transformative at the time. Their significance lies not in what they immediately produce but in what they make possible. Only later, looking backwards, do we notice the pathway. The question raised by The Shape of the Ask pushed this idea a step further. If traces shape possibilities, and possibilities shape future interactions, perhaps traces do more than preserve the past. Perhaps they help generate trajectories. If interactions generate trajectories, can trajectories be studied?
Can trajectories be studied?
At first glance the idea sounds implausible. Conversations seem too complex, too personal, and too unpredictable to measure. Yet we routinely measure phenomena that are far more complicated than any individual conversation. Meteorologists measure weather systems. Economists measure markets. Psychologists measure latent traits such as intelligence, personality, and attitudes. Why should trajectories be different?
If trajectories matter in human lives, there is no obvious reason why they should cease to matter when one of the participants is an intelligent machine. In fact the question becomes particularly important when we consider artificial intelligence. Much of the current discussion about AI ethics focuses on outcomes. Was the answer correct? Safe? Biased? Harmful? These are important questions. But they are also questions asked after the fact.
Increasingly, intelligent systems interact with one another and with humans at speeds that make continuous human supervision impossible. In some domains, important decisions may emerge within milliseconds. By the time a human observer reviews the outcome, the trajectory that produced it may already be complete. In such situations, traditional human-in-the-loop approaches may be insufficient.
What follows?
One possibility is that we may eventually need ways of recognising the character of a trajectory while it is still developing. The possibility remains speculative, yet it suggests an intriguing direction for future research. If traces generate questions, questions generate trajectories, and trajectories exhibit recognisable patterns, then perhaps interactions can be monitored not merely for what they conclude but for how they develop.
Perhaps the most important ethical question is not simply what answer an intelligent system gives. It may be how the interaction arrives there. The challenge is not merely to judge where an interaction ends, but to understand where it is beginning to go. If traces can generate trajectories, and trajectories have shapes, then perhaps the most important question is simply: What follows?
© John Rust, June 2026. All rights reserved.