The Ask Between Us

Towards a Science of Teleosynthesis

Intelligence in the Space Between

For a long time, we have imagined intelligence as something sealed inside the individual mind — measurable, rankable, self-contained. That picture has been powerful and, in many ways, productive. But it has always left something out. Some of the things we most value in human thought do not arise in isolation alone. Originality, creativity, insight, even the discovery of what we were really trying to ask, often take shape through interaction. Not merely in the weak sense that conversation reveals what was already fully formed inside one person, but in the stronger sense that exchange can help specify, refine, and sometimes generate the thought itself.

We know this in ordinary life. A difficult conversation opened with praise may unfold very differently from one opened with criticism. A question that seemed clear at first can turn out not to be the real question at all. A reply can suddenly expose the shape of an underlying concern that had not yet found words. Some thoughts are answered in dialogue. Others are discovered there. This is one reason why the path-dependent structure of conversation has always been so hard to study rigorously. Human dialogue is one of the oldest things in the world, but also one of the hardest to formalise and to study scientifically. It is shaped by sequence, tone, timing, memory, expectation, and subtle shifts of framing. We have long known, from life itself, that order dependent complexity matters. But we have had very few ways of examining that fact with any real discipline. The phenomenon has been obvious; the science of it has lagged behind.

This page should be read as a shorter, more accessible companion to my main Teleosynthesis page, where I set out the broader claim that purposive organisation can emerge through structured interaction, especially narrative and dialogue over time.

A New Interactional Space

Now something has changed. For the first time in history, human beings can enter sustained, responsive narrative exchange with non-human partners. We can ask, reply, clarify, challenge, hesitate, redirect, and continue — not with another person, but with a generative system capable of following the thread of a conversation. That does not make such systems human, and it does not settle any larger question about consciousness or inner life.

But it does create a new kind of interactional space. Once we enter it, we should not be surprised if we begin to notice shadow forms of phenomena we have long associated with human conversation. By shadow forms I mean partial, structurally similar phenomena: not human originality or understanding in full, but patterns that may help illuminate how such phenomena arise. Not the whole thing, perhaps. Not in the same way. But enough to make us pause. What if some of the qualities we prize in thought are not simply “in the head” at all? What if they have always depended, at least in part, on unfolding narrative interaction? What if the real ask is not always a hidden sentence waiting to be phrased correctly, but a meaning that only becomes clear in sharing and exchange?

The Ecological Turn

The older picture of mind made this hard to see. It encouraged us to imagine thoughts as if they were stored like files: memories in one compartment, intentions in another, questions waiting somewhere in the machinery until the right words retrieve them. But psychology has had another tradition for some time. In The Ecological Approach to Visual Perception in 1979, J. J. Gibson argued that we do not perceive the world as a neutral array of objects to be internally decoded, but as an environment rich with possibilities for action. A surface may invite walking, an object may invite grasping, a tool may invite use. Meaning, on this view, is not simply placed inside the head; it emerges in the relation between organism and world.

Memory can work in a similar way. When we walk a familiar route, the memory is not simply inside us; it is partly in the street, the corner, the shopfront, the turn that suddenly feels obvious. The world helps organise the act of remembering. Conversation can do something similar. A response, a hesitation, a challenge, or a misunderstanding can reveal the direction of a thought that was not yet fully available to the person who began it.

Thought in New Company

If that is even partly true, then generative AI matters in a deeper way than most current debates allow. Its significance would not lie only in faster answers, better summaries, or more efficient tools. It would lie in the appearance of a new medium in which thought can be jointly developed, redirected, and sometimes brought into clearer existence. What emerges is not merely an answer to a prior question, but a temporary direction of thought — a small, shared purpose formed in the movement of the exchange itself.

This should not be exaggerated. The mere existence of dialogue does not guarantee depth. Many such interactions will be trivial, repetitive, derivative, misleading, or simply wrong. But even so, something irreversible has happened. When any articulate human can now begin a sustained exchange with a non-human conversational partner, the opportunities for interaction-born ideas multiply enormously. What was once largely confined to human collaborations, seminars, friendships, rivalries, and rare moments of intellectual luck may now occur in distributed form, across hundreds of fields and thousands of local contexts. We are moving, in other words, into a world in which thought itself may take shape in new company.

The important point is not that AI might replace human originality. It is that originality may never have been wholly private in the first place. Some of it may always have depended on the space between voices: the fragile membrane of conversation where meanings are tested, reshaped, repaired, and sometimes transformed into something neither side could have brought into quite the same form alone. If so, then this new world deserves more than either commercial enthusiasm or cultural panic. It deserves serious study. We need to understand what kind of phenomena are appearing in these human–AI exchanges, how far they resemble familiar human forms, where they differ, and what they may reveal about intelligence itself.

For centuries, we have treated conversation as a vehicle for thought: a way of expressing, refining, or transmitting what is already there. We may now have to consider the possibility that, in some of its most important forms, conversation is also one of the places where thought begins.

From the Ask to the Attractor

If thought can begin in conversation, the next question is what gives some conversations direction, depth, and resistance to premature closure. Not every dialogue deepens understanding. Some exchanges close too quickly. They satisfy a need, confirm an expectation, or produce a phrase so polished that inquiry stops before it has properly begun. Others do the opposite. They keep the question alive. They expose uncertainty, invite correction, and draw thought toward something that can survive contact with the world. This is the context within which I have begun to use the phrase intelligence attractor. I am not suggesting that this idea begins with me. Biology and systems theory have approached related territory through other vocabularies: teleonomy (Mayr, 1974) for evolved goal-directedness, anticipatory systems (Rosen, 1985/2012) for present action shaped by models of possible futures, self-organisation (Kauffman, 1993) for the emergence of order in complex systems, and attractor landscapes (Waddington, 1957) for recurrent paths of development or regulation. My phrase intelligence attractor is an attempt to bring these ideas into the study of thought, narrative, and human–AI dialogue.

A false attractor in narrative gives us the answer we thought we wanted too soon. It gathers fluency, confidence, and coherence around a premature closure. A truer attractor keeps the ask open long enough for reality, evidence, other minds, and further questioning to push back. The idea has a biological analogue in teleonomy (Mayr) and anticipatory systems (Rosen). It is no accident that the eye of the human and the octopus are so similar. Eyes evolved repeatedly in very different forms of life because the world contains light, movement, distance, edges, prey, danger, and surfaces worth detecting. The eye is not merely useful; it is a structure disciplined by the world. It shows that life does not simply invent its own meanings. It becomes attuned to real structures that make action possible.

The same pattern appears in science itself. Newton’s laws, Einstein’s relativity, and quantum theory did not become powerful simply because they were elegant ideas. They became powerful because the world pushed back in their favour. Their predictions, instruments, anomalies, and corrections disclosed structures that were not directly visible, but were nevertheless real enough to constrain thought. The real is not always obvious, but it makes itself felt by what it allows, resists, confirms, or corrects. This suggests a cautious extension of the biological analogy. Purpose-like organisation may not be merely something humans project onto the world. It may be one of the ways living and thinking systems become answerable to real conditions of continuation: survival, action, cooperation, correction, discovery, and repair. At this stage, that is not a proof. It is a direction of inquiry. Some forms of organisation recur because they help keep viable futures open.

The Trajectory not the Output

Human–AI dialogue makes this newly visible. A prompt may begin as a request for an answer, but in sustained exchange it can become something more: a trajectory of inquiry. The human ask projects a possible future; the AI response reshapes the path; the next human move redirects it again. Sometimes the result is cliché, drift, or premature certainty. But sometimes the exchange moves toward clarification, repair, discovery, and a deeper relation to the question itself. That is the step from the ask to the attractor. The ask opens a field of possible continuations. The intelligence attractor is what begins to form when one of those continuations becomes progressively more coherent, corrigible, and answerable to the world.

This is why the study of human–AI conversation should not be reduced to prompts and outputs. The real phenomenon is the unfolding trajectory: how questions change, how answers redirect them, how uncertainty is either closed down or preserved, and how thought sometimes becomes more real through the process of exchange. For centuries, we have treated conversation as a vehicle for thought: a way of expressing, refining, or transmitting what is already there. We may now have

References

Gibson, J. J. (1979). The Ecological Approach to Visual Perception. Boston: Houghton Mifflin.

Mayr, E. (1974). “Teleological and teleonomic: A new analysis.” In Methodological and Historical Essays in the Natural and Social Sciences. Mayr helped legitimise teleonomy as a way of discussing biological goal-directedness without old-style teleology. The Stanford Encyclopedia also notes that teleonomy was taken up by evolutionary biologists including Mayr and George Williams.

Rosen, R. (1985/2012). Anticipatory Systems: Philosophical, Mathematical, and Methodological Foundations. Rosen’s key idea is that an anticipatory system’s present behaviour depends on internal models of possible future states.

Kauffman, S. A. (1993). The Origins of Order: Self-Organization and Selection in Evolution. Kauffman argues that self-organisation contributes to the order on which natural selection can act.

Waddington, C. H. (1957). The Strategy of the Genes. Waddington’s epigenetic landscape remains the classic biological image for developmental pathways and stable valleys; later systems biology connects this with attractor landscapes.

Nilsson, D.-E., & Pelger, S. (1994). “A pessimistic estimate of the time required for an eye to evolve.” Proceedings of the Royal Society B: Biological Sciences, 256(1345), 53–58.

Pezzulo, G. (2024). “Active inference as a theory of sentient behavior.” Biological Psychology. Active inference gives a modern framework in which perception and action are guided by internal models used to predict, infer, and direct action.