Time Without a Clock
In the early 1930s, an insurance investigator from Hartford, Connecticut — a man whose day job was examining the causes of fires and industrial accidents — began spending his evenings teaching himself the grammar of an obscure language spoken on three mesas in the high desert of northeastern Arizona. His name was Benjamin Lee Whorf, and his conclusions about what that language did and did not contain would haunt linguistics, philosophy, and cognitive science for the next century. Whorf claimed that the Hopi language had no words, grammatical forms, or expressions referring to time as something that flows, passes, or is divided into past, present, and future. The Hopi, he argued, lived in a fundamentally different temporal universe — one in which the Western conception of time as a river with a fixed direction was simply absent from thought.
He was not entirely right. Later scholars — most forcefully the linguist Ekkehart Malotki, who spent years living with Hopi speakers and produced an 677-page monograph in 1983 demonstrating that Hopi does in fact have grammatical means of expressing temporal sequence — corrected the record sharply. The popular version of Whorf's claim — that the Hopi have no concept of time whatsoever — is now broadly considered an overstatement. But the debunking of Whorf's specific claims left intact something more subtle and more interesting: a genuine philosophical difference in how Hopi grammar and cosmology organise events in time. And that philosophical difference, stripped of Whorf's overclaiming, turns out to resonate in unexpected ways with some of the deepest problems in contemporary physics and artificial intelligence.
The Hopi Distinction: Manifesting and Manifested
The genuine philosophical insight in Hopi temporal thought is not that time doesn't exist — it is that the Hopi language draws its most fundamental distinction not between past, present, and future but between two modes of existence: what might be called the manifesting (everything that is in the process of becoming, including what we would call the present and future) and the manifested (everything that has come into concrete, fixed form — what we might call the past, but including also stable present physical objects and facts). This is a distinction not of when something occurs but of what kind of thing it is.
In Western temporal thought, a rainstorm exists in the future as a prediction, in the present as an event, and in the past as a memory or a record. It moves through time; time happens to it. In the Hopi framework, a rainstorm in the manifesting phase is a gathering of intensities — prayers, rituals, atmospheric conditions, social preparations — that are growing toward manifestation. It is not a future event waiting to occur; it is an ongoing process of becoming that will, at some density, cross into the manifested. The past is not lost — it is the realm of the fully concrete, the fixed, the settled. The future is not empty — it is continuously active with the processes that will become the manifested world.
This sounds, to Western ears, initially like mysticism. But it is more precise than it sounds. What Hopi ontology resists is the idea that time is a dimension through which a fixed world moves — the idea that there are discrete moments, each fully real, and the present is just the moment you happen to be standing in. Instead, it treats the present as a zone of active process and becoming, and the past as the accumulated result of what that process has produced. Events are not objects located in time. They are intensities of process that crystallise into fact.
The Block Universe: Physics Catches Up
In 1908, the mathematician Hermann Minkowski delivered a lecture that Einstein later described as making special relativity fully comprehensible. Minkowski showed that Einstein's theory required not space and time to be treated separately but spacetime — a four-dimensional continuum in which every event occupies a fixed position. Past, present, and future are all equally real, equally fixed, equally existing. What we call "now" is simply the three-dimensional slice of spacetime that a particular observer's worldline happens to be crossing at a particular moment.
This is the block universe — or eternalism — and it remains the dominant view among physicists who think carefully about the metaphysics of time. There is no objective "now." There is no privileged present moment that is more real than the past or the future. The past is not gone; it is simply elsewhere in spacetime. The future is not yet to come; it is already fixed, located further along your worldline. You are not moving through time. You are extended through spacetime, and the sensation of temporal flow is a feature of your conscious experience, not a feature of physical reality.
The block universe theory has a strange resonance with Hopi ontology. If all events exist equally across the spacetime block, then the Western intuition of time as a flowing river — with the present uniquely real and the future not yet existing — is an illusion. What exists is a totality of events, some of which are causally connected to others in ways that give rise to the experience of sequence and change. The Hopi resistance to treating time as a dimension through which you pass, and to treating events as discrete located objects, looks, from this vantage, less like a cultural curiosity and more like an intuition that anticipates the block universe by several centuries.
All spacetime events equally real — past, present, future co-exist in four-dimensional manifold.
No objective "now" — present is observer-relative, not ontologically privileged.
Time appears to flow due to thermodynamic asymmetry (entropy gradient), not metaphysical reality.
Events are nodes in a causal network, not beads on a string.
Events exist as degrees of manifestation — no privileged moment, only the manifested and the manifesting.
The present is not a boundary between past and future — it is the zone of active becoming.
Time does not flow — processes intensify and crystallise into settled facts.
Events are processes, not positions; they have intensity, not location.
Quantum Time: Where It Gets Stranger
The block universe is already unsettling to everyday intuition. Quantum mechanics is more so. And when physicists attempt to reconcile quantum mechanics with general relativity — the great unfinished project of theoretical physics — the concept of time becomes almost unrecognisable.
The most significant problem is this: general relativity treats time as a dimension of spacetime, shaped and curved by mass and energy. Quantum mechanics treats time as a fixed external background — a parameter that variables are functions of, not a dynamical entity in its own right. When you attempt to write a quantum equation for the universe as a whole, combining both frameworks, the time variable drops out. The Wheeler-DeWitt equation — the closest thing physics has to a quantum equation of gravity — contains no time variable. It describes a universe in which, at the deepest level, nothing changes. Time, as we experience it, is not fundamental. It is emergent — an appearance that arises from the correlations between subsystems within a timeless quantum state.
The physicist Carlo Rovelli, in his work on loop quantum gravity, has argued that time as we experience it arises not from a fundamental feature of physics but from our particular thermal perspective within the universe — from the fact that we, as observers, are embedded in a region of low entropy that is increasing toward higher entropy, producing the asymmetry we call the arrow of time. At the level of fundamental physics, there is no past, no future, no now. There is only a web of physical processes, each defined by its relations to others. Change is real; time as a container is not.
The Surprising Convergence
The convergence between Hopi ontology and these developments in physics is not causal — no physicist was studying Hopi grammar when they derived the Wheeler-DeWitt equation — but it is intellectually striking. What Hopi grammar resists is precisely what frontier physics now questions: the idea that time is a container, a background, a dimension external to events. What Hopi proposes instead — a world in which processes have degrees of manifestation, in which events are intensities rather than positions, in which the becoming is as real as the become — has a recognisable structural echo in both the block universe (where all events co-exist) and quantum gravity (where time emerges from correlations between processes rather than being an independent backdrop).
Carlo Rovelli's relational quantum mechanics, in particular, proposes that physical properties are never absolute but always relative to a reference system — that the world is constituted by interactions and relations between processes, not by objects with intrinsic properties located in space and time. The world is not a collection of things. It is a network of happenings. Events, not objects. Relations, not properties. Processes, not states.
The AI Problem: Reasoning Without a Clock
Large language models — the most capable artificial reasoning systems currently available — have a peculiar and underexamined relationship with time. They are trained on text produced across years and decades, and they process all of it without any genuine temporal experience. They know that events have dates, that some things came before others, that timelines exist. But their internal architecture gives them no intrinsic temporal dimension: every token in a context window is processed relationally, not sequentially in a way that encodes duration, causation, or becoming. They exist, in a sense, in something like the Hopi manifested realm — a vast repository of settled facts — without the manifesting realm of genuine temporal process.
This produces characteristic failure modes. Current AI systems struggle with:
2. Process vs state conflation. AI systems are far better at reasoning about states (the world as it is at a moment) than processes (things that are in the process of becoming, developing, or intensifying). They tend to tokenise events as discrete facts rather than as points in ongoing processes.
3. Causal vs temporal conflation. Because training data often presents events in narrative sequence, models frequently confuse temporal sequence with causal order — assuming that if A is described before B, A caused B.
4. Static world model. Trained models encode the world as it was at a snapshot in time, with limited ability to reason about how the world changes as a process unfolding over time. Their "knowledge" is the manifested — fixed facts — without access to the manifesting.
What Hopi Ontology Might Teach AI Architecture
The Hopi distinction between manifesting and manifested is not just a linguistic curiosity — it is a powerful ontological framework for reasoning about incomplete, evolving, and probabilistic information. And it maps surprisingly well onto the kinds of reasoning improvements that AI researchers are actively pursuing.
1. Process-Based Knowledge Representation
Current knowledge graphs and retrieval systems typically represent facts as static nodes: entities with properties, linked by relations, each tagged with a timestamp. A more Hopi-inflected representation would treat knowledge as a field of processes with varying degrees of crystallisation. Some facts are fully manifested — settled, verified, stable. Others are manifesting — in process, changing, contested, in different states of becoming in different contexts. This is not merely a technical nicety. It corresponds to how knowledge actually works: most of what an AI system should "know" exists somewhere on a spectrum of certainty, recency, and ongoing revision, not as a binary manifest/non-existent.
2. Relational Event Reasoning
Rovelli's relational physics suggests that properties are relational rather than absolute. An AI reasoning system built on process ontology would represent events not as points with timestamps but as relational processes — defined by their connections to other processes, their dependencies, their rates of change, their causal upstream and downstream. This is closer to how expert human reasoning actually works: a doctor reasoning about a patient's condition is not indexing discrete facts but tracking an ongoing process — systems interacting, trajectories of change, intensifying patterns that may or may not crystallise into a diagnosis.
3. The Temporal Uncertainty Layer
In quantum mechanics, superposition represents the existence of multiple potential outcomes prior to measurement-induced collapse. The Hopi manifesting realm has a structural parallel: things that are in process have not yet crystallised into the single outcome they will eventually produce. An AI architecture that explicitly represents this layer — modelling not just what is true but what is in-process-of-becoming-true, with probability distributions over outcomes weighted by causal and contextual factors — would be significantly more robust to the kinds of temporal reasoning failures listed above. The model would have, in effect, an internal manifesting realm alongside its manifested knowledge base.
4. Eliminating Spurious Temporal Anchoring
A model trained on the Wheeler-DeWitt insight — that time is emergent from correlations between physical processes rather than a background coordinate — would approach temporal reasoning differently. Rather than asking "when did this happen?" as its primary operation, it would ask "what is this event's relational position in the network of processes that constitute this domain?" This shift from absolute to relational temporal reasoning is exactly what physicist-philosophers like Rovelli advocate for physics, and it would directly address the AI failure mode of treating training-time knowledge as if it were universal present-tense truth.
Epistemic Humility as an Architectural Feature
There is a deeper lesson that the Hopi/physics convergence offers AI systems, and it is philosophical rather than purely architectural. Both the Hopi ontology and the Wheeler-DeWitt framework share a resistance to treating any moment — any state of knowledge, any configuration of facts — as privileged and complete. In the block universe, no moment is the "real" present. In relational quantum mechanics, no observation reveals an absolute fact — only a relational one. In Hopi thought, no state of the world is fully fixed while the manifesting continues.
Current AI systems are, at their worst, absolute in a way that none of these frameworks endorse. They present training-time knowledge as present-tense fact. They assert without representing uncertainty. They confuse what is settled with what is true at the moment of inference. A genuinely well-calibrated AI reasoning system would have built into its architecture what might be called temporal humility — an explicit representation of which of its beliefs are fully manifested (verified, stable, reliably sourced) and which are still in the manifesting phase (recent, contested, evolving, likely to crystallise differently as more becomes known).
This is not merely a safety feature. It is an accuracy feature. The world does not stand still while AI systems are trained on snapshots of it. Processes continue to intensify and crystallise. Knowledge accumulates. What was true in one moment may not be manifested in the same form in another. The system that can represent its own epistemic relationship to time — not just what it knows, but when it knew it and how that knowledge stands relative to ongoing processes — is a more honest and more capable reasoner.
The Clock Is a Metaphor
Benjamin Lee Whorf sat in his Hartford apartment in the 1930s, studying Hopi verb forms by lamplight, and saw — through a glass darkly, with conclusions he overstated and a methodology his successors rightly critiqued — something genuinely interesting: that a language could be built on a different temporal ontology than the one his own language embodied, and that this different ontology was not obviously less sophisticated. He was wrong to say the Hopi had no concept of time. He was right to say that how a language organises time encodes assumptions about what time fundamentally is.
Physics spent the twentieth century discovering, through relativity and quantum mechanics, that those Western assumptions — time as a river, the present as a privileged moving point, events as discrete located objects in a temporal container — are not physically fundamental. They are the appearance of something stranger and more relational at the base of reality. In quantum gravity, time dissolves into correlations between processes. In the block universe, all moments are equally real and the present is an observer-relative fiction. In relational quantum mechanics, the world is constituted by events rather than things.
And AI systems, now at the frontier of machine reasoning, carry within their architectures the unexamined assumption that time is a background coordinate, that facts are timestamped objects, that the present is the training distribution — and they fail, reliably and characteristically, at the kinds of temporal reasoning that require more.
What Hopi grammar encoded as an ontological intuition — that the world is constituted by processes of becoming, that events are intensities not positions, that the manifesting and the manifested are different kinds of thing — turns out to be a direction that frontier physics is moving toward independently, and a direction that AI architecture would benefit from incorporating deliberately. The clock is a metaphor. Time, at its root, may be something more like an ongoing ceremony of becoming. And the system that understands this — that can represent not just what is settled but what is still gathering — will be closer to reasoning the way the world actually works.
Comments
Post a Comment