Physics has no use for the present moment. The equations that describe the world are, with a single disputed exception, indifferent to the direction of time and contain nothing corresponding to now. A complete description of the universe specifies what occurs at every moment; it does not mark one of them as the moment that is occurring. On the resulting picture, sometimes called the block view, the difference between 1815 and next Tuesday is of the same kind as the difference between here and Vienna: a matter of where one is situated, not of what exists.
Against this stands the most insistent testimony available to anyone. It seems evident that the past has gone and the future has not arrived, that the present is not merely the place where I happen to be but the only time there is. The philosophical difficulty is that this conviction is very hard to state in a way that does not immediately consume itself. To say that only the present exists is to say something that, if true, was equally true at every moment that has ever been called the present, which suggests that what is being asserted is not a fact about a moment but a feature of every moment, and therefore not a distinction between them at all.
Attempts to preserve the flow of time by adding it to the block have not fared well. The proposal that a moving spotlight of presentness travels along the sequence invites the question of how fast it travels, and the only available answer—one second per second—is not a rate but a tautology dressed as a measurement. Nothing has been explained; a metaphor has been made to look like a mechanism.
It is tempting to conclude that the sense of passage is an illusion generated by the way memory accumulates, and this is now the majority view among physicists who consider the matter. But the conclusion is less tidy than it appears. An illusion is a misrepresentation of how things are, and specifying what is being misrepresented here has proved unexpectedly difficult. The claim cannot be that we wrongly believe the present is objectively distinguished, since on the block view every moment is present relative to itself, and our belief that this one is present would then be correct. To call the experience illusory is to owe an account of the error, and no such account has been supplied.
Until the middle of the nineteenth century, an English town kept its own time. Noon in Bristol arrived some ten minutes after noon in London, for the excellent reason that the sun reached its highest point over Bristol ten minutes later. Nobody experienced this as an inconvenience, because nobody had to compare the two. A traveller by coach lost the difference somewhere along the road and never noticed it going.
The railway made the difference visible, then intolerable. A timetable is a promise about simultaneity: it asserts that a train departing at nine will meet, at some later hour, a stretch of track that another train has been promised will be clear. If nine o’clock means different things at either end of the line, the promise cannot be kept, and the failure to keep it is not an embarrassment but a collision. Railway time was therefore imposed not as an improvement in accuracy but as a precondition of coordination, and the two are easily confused. Bristol’s local noon had not been wrong. It had merely become unusable.
What followed is often narrated as the triumph of the modern over the parochial, but the more interesting feature of the episode is who did the imposing. Standard time in Britain was established by railway companies years before Parliament touched the matter, and the courts continued for decades to recognise local time as the legal time in disputes about, for instance, whether a poll had closed. For a generation, a private commercial network operated the country’s temporal infrastructure while the state looked on, uncertain whether anything had happened that required its attention.
This pattern has proved durable. Systems of coordination—weights, gauges, container dimensions, network protocols—are typically settled by whoever most urgently needs them settled, which is rarely the public and frequently a firm with a scheduling problem. The resulting standard then acquires the appearance of a natural fact, and its origins in somebody’s operational convenience are forgotten within two generations. The width of a shipping container is not a discovery about the world. It is a decision, taken by particular people for particular reasons, that everything else has since been built around.
That any measure which becomes a target ceases to be a good measure is now repeated so often that it has acquired the authority of a natural law. It is not one. It is a claim about institutions, and like every such claim, it holds only under conditions that its admirers rarely bother to state.
What the claim gets right is that measurement is never a passive reading of the world. To count something is to declare it countable, and the declaration reorganises the field it claims merely to describe. When hospitals in several countries were required to see emergency patients within four hours, ambulances began idling in the forecourt with patients still aboard, because the clock started at the door. Nobody had lied. A boundary had been drawn, and behaviour had migrated to the more advantageous side of it. The number continued to improve while the thing the number was invented to track did not.
From this, it is now fashionable to conclude that we should measure less and trust professional judgement more. The conclusion smuggles in a fantasy. Before waiting-time targets, there were no waiting times; there was only waiting. The absence of a number did not mean the absence of a policy about who waits and for how long. It meant that the policy was invisible to everyone except the people administering it. Unmeasured systems are not ungoverned systems. They are systems governed by whoever’s intuition carries the most institutional authority, and that authority is distributed no more fairly than anything else.
The variable that actually matters is not the quantity of measurement but its contestability: whether those described by a number can dispute it, supplement it with a rival number, and force its revision. A figure that can be argued with behaves quite differently from a figure that cannot. The pathologies popularly blamed on quantification are, on inspection, usually pathologies of monopoly — one number, one owner, no appeal. Where several bodies publish competing estimates of the same quantity, gaming becomes conspicuous, because the gap between the measures is itself informative.
An organisation that publishes three inconsistent figures for its own performance is, in an important sense, more honest than one publishing a single reconciled figure, and rather more useful to an outsider trying to understand it. Administrators dislike this arrangement, since inconsistency reads as incompetence and reconciliation reads as control. But the tidiness they prefer is a fiction maintained at the reader’s expense, and the fiction is a great deal more expensive than the untidiness it conceals.
A fisheries biologist working in the Gulf of California in the 1990s asked three generations of fishermen the same question: What did a good catch look like when you began? The oldest men described a sea that the youngest could not have imagined, and the youngest described, without irony, a sea in decline from the one they themselves had inherited. Each generation had taken the ocean of its own youth as the natural condition and measured loss from there. Nobody was mistaken about what they had seen. Everyone was mistaken about what there had been.
This is the phenomenon since named shifting baselines, and its consequences run deeper than the pleasant melancholy usually attached to it. If every generation revises the reference point downward, then the accumulated decline never appears in any single person’s testimony. Conservation policy has peculiarly absorbed the lesson. It now speaks constantly of restoration. Should a British hillside be restored to its condition before industrial forestry, before enclosure, before the clearance of the wildwood, or before the retreat of the ice? Each answer describes a different landscape, and each is equally natural in the only sense the word can bear here, which is that human beings did not intend it.
The temptation is to conclude that since every baseline is arbitrary, the concept should be abandoned and ecosystems managed only for present function. This is too quick. Arbitrariness in the selection of a reference point does not make the reference point useless, any more than the arbitrariness of the prime meridian makes longitude useless. What the arbitrariness does is transfer the question from the sciences to politics, where it has always belonged and where practitioners are least comfortable placing it.
That discomfort is worth examining, because it explains a great deal of the language in which environmental argument is conducted. To say that a wetland should be restored to its 1800 condition is to make a contestable claim about whose landscape counts, which is a question about people. To say that the wetland is being returned to its natural state is to make the same claim while appearing to make none. The second formulation wins arguments more easily. It also guarantees that when the argument is eventually lost, nobody will be able to say precisely what was being argued about.
What evidence suggests is that a memory is assembled at the moment of recall from fragments that are not themselves memories of anything in particular—a spatial impression, an emotional tone, a scrap of language, a general schema of what such occasions are usually like. The assembly is rapid, involuntary and largely accurate, which is why the storage metaphor survives. It is also creative, in the strict sense that the product did not exist before the act of recalling produced it. Two recollections of the same afternoon are not two readings of one file. They are two constructions from overlapping materials, and each construction alters the materials available for the next.
This last point is the disquieting one. Retrieval from an archive leaves the archive unchanged. Reconstruction does not. Each act of remembering deposits its own product among the fragments from which subsequent acts will build, so that a frequently revisited memory is not the best preserved but the most heavily edited. The events we recount often are the events about which we are least reliable, and our confidence runs in precisely the wrong direction, since fluency of recall is what confidence tracks and repetition is what produces fluency.
It would be an error, though, to conclude that memory is therefore unreliable and ought to be distrusted wholesale. Reconstruction is not a design flaw compensating for insufficient storage capacity. A system that recorded experience faithfully would be nearly useless for the work memory actually performs, which is not the preservation of the past but the anticipation of the future. What makes a schema valuable is exactly that it discards the particulars: it is because the mind retains the shape of such occasions rather than the details of one occasion that it can supply an expectation about an occasion it has never encountered.
The institutional implications are more awkward than the psychological ones. Legal systems continue to treat the confident witness as the credible one, and the confident witness is frequently the witness who has told the story most often, sometimes to investigators whose questions supplied the very details later recalled with such assurance. The remedy usually proposed is to caution juries about the fallibility of memory. This is unlikely to help much, since the difficulty is not that jurors overestimate memory in general but that they have no way of distinguishing a reconstruction that happens to be accurate from one that has been polished by retelling. Both arrive in the same voice.
