The Immutable Past · What Is
What Is Pace
You were taught that a period of time has a speed — that the fifties were slow, that the world has accelerated, that you can lose your pace and get it back. The foundation I work from says frequency is undefined at the boundary the ordinary sentence points at, so a period has no rate to have. A pace is an arithmetic done here: marks counted against a counter, both read in the present. That is why the same complaint has been filed in the same words for a century and a half, why two people in one hour honestly report opposite speeds, and why the decades people file as slower are, on the measured ratio, the faster ones.
Life was slower then. It is one of the most-said sentences in English and nobody asks the speaker to defend it, because it does not sound like a claim. It sounds like a memory. Behind it sits a picture so ordinary that stating it feels pedantic: time comes in stretches, each stretch runs at a speed, and you live at whatever speed your stretch is running. The fifties were slow. The nineties were quicker. Ours is a blur. On this picture a decade has a tempo the way a river has a current, and being alive in it is being carried at that rate — which is why you can be swept along, why you can fall behind, and why some people move somewhere quiet to get back to a slower one. I held this picture for most of my life. It is, I will argue, wrong about where the rate lives, and three ordinary situations show it before any foundation is consulted.
One clarification first, because two nearby subjects get merged constantly and this piece is about only one of them. How long a stretch of time was and how fast it went are different quantities. Remembered length is a duration; pace is a rate, a count against a count. I have written about duration elsewhere and am not re-running it here. This piece is about the rate.
What the ordinary picture commits you to
State it strongly, because the weak form is not what anybody believes.
The picture makes three commitments. First, that a rate belongs to a period — that the 1950s had a speed as they had a weather, and that the speed was a feature of the decade rather than of anyone in it. Second, that the rate is there whether or not anyone reads it, so acceleration is something that happened to the world and not something that happened to the readers. Third, that the rates of different periods are comparable — that ours is faster than theirs — which requires both to be quantities of the same kind, available at once, to be compared.
Three ordinary situations do not fit inside those commitments, and each one breaks a different commitment.
Case one · a complaint older than the machines it blames
In 1881 the New York physician George Beard published American Nervousness and named the cause of the exhaustion he saw in his patients: modern civilization, distinguished from the ancient by five things — steam power, the periodical press, the telegraph, the sciences, and the mental activity of women. In 1903 Georg Simmel opened his lecture on the metropolis by locating the metropolitan character in the intensification of nervous life produced by the rapid and uninterrupted change of external and internal impressions. Both are marked here as report, not grounds. What they report is a complaint that is word-for-word the one being filed now, about a world with no aircraft, no broadcast, and no network.
The second commitment says acceleration is a fact about the world. Then the reading of it should move when the world does. It does not. The complaint is the same in 1881, 1903 and this morning, across periods whose measured rates of change are not remotely comparable. A reading that stays fixed while its supposed object varies enormously is not a reading of that object. Something is being reported. It is not the period.
Case two · one hour, two speeds, no disagreement about the facts
Two people sit in one meeting for one hour. For the first it goes fast; for the second it drags. Ask them during — at the forty-minute mark, on a note passed under the table — so that nothing here depends on memory or on how long the hour will later seem. They agree completely about the contents. Same agenda, same slides, same interruptions, same number of minutes on the clock on the wall. They disagree about the speed.
The first commitment says the rate belongs to the interval. There is one interval and it has one set of contents, and if the rate were among those contents there would be one reading of it. There are two, both honest, neither correctable by pointing at the room. Whatever each is reading, it is not a property of the hour.
Case three · the rate with nothing underneath it
Try to state the speed of 1955. Not a change that happened in 1955 — the speed of the year itself, the thing the sentence claims it had. You cannot finish the sentence without a second term, because a rate is a ratio and a ratio needs two counts: miles per hour, beats per minute, marks per something. Say fast and the immediate question is faster than what, counted per what, and the answers that can actually be completed name something present. Faster than now. Faster than I expected. Faster than there is evidence of.
The third commitment says the two rates are comparable quantities. But only one of them has ever been computed, and it was computed here. Life was slower then has the grammar of a measurement and not the arithmetic of one. It names a period, asserts a rate, and supplies no denominator — and a rate with no denominator is not a small quantity or a disputed one. It is the shape of a quantity with the quantity missing.
So the rate is not in the period, is not the same for two people in one interval, and has not, so far as I can find, ever been computed the way the picture says it exists. Something else is doing the work.
What the foundation says instead
I work from a small set of numbered primitives and results derived from them. The result doing the work today has been run twice in this series and both times under a different clause of itself: under warmth and colour, where longing aimed a thermal vocabulary at a domain that has no temperature, and under length and distance, where an age turned out to be a count taken here. Today it runs under the clause that has been sitting in it unused since it was written, which is the one that names frequency.
P3: to exist is to vary. Nothing that exists is ever still, and that which does not vary does not exist, in the only sense of exist available to varying things. P0, the axiom: the past is immutable — a consistency constraint on the record set, which admits exactly one consistent assignment that no present operation can alter. From those, R5: temperature, frequency, colour and all time-evolution descriptors do not apply to the Past. Not zero. Undefined, the way a slope is undefined at a point that is not on any curve.
Now be literal about what a rate is, because the derivation is three lines and each depends on the literal reading. A counter is a thing that varies at a rate someone trusts — the piece on age worked this out and I carry it over rather than re-deriving it. A rate is then two counts and a division: how much of one thing varied, per how much of the counter varied. Both terms are variation. A rate is not merely described by time-evolution vocabulary; it is made of it.
So: the Past does not change, by the axiom. What does not change does not vary. A rate requires a varying numerator, and there is none — not a numerator that came out at zero, which would be a measurement, but no term to put there at all. Frequency is undefined at the boundary, and the speed of 1955 is undefined in exactly that sense. Case three is what that looks like from inside the language: a sentence people say fluently, with a hole where the arithmetic goes, which nobody notices because nobody has ever tried to finish it.
It is worth pausing on how strong that is. The foundation carries its own list of ways it could be shown false, and one entry on that list reads: any reliable measure that assigns a well-defined temperature, frequency or colour to the Past itself, rather than to a record in the present, violates R5. That is the falsifier. And life was slower then is that falsifier, said casually, thousands of times a day, by people who have never heard of the theory and are not trying to break anything. The framework is not saved from it by cleverness. It is saved by the fact that nobody cashes the sentence — that in a century and a half of complaining about the speed of the age, the measurement has not been produced, because there is no instrument aimed in that direction.
Where the rate actually is
The vocabulary locks history as the artifacts: present, surrounding, and revisable. Not the Past under another name — the marks, here, on this side, which is where any count is taken. Put that beside the counter and the arithmetic assembles itself.
First marked construction
A pace is a present ratio: marks counted against a counter, both read now, over the record that is on the shelf now. There is no operation that reads a rate off a period. There is only an arithmetic performed on a record, and the answer belongs to the record and to the moment of counting.
Consequence of R5 with P3 and P0, using the locked definition of history and the counter carried over from the age piece. Mine: the foundation says frequency does not apply at the boundary and does not say what the word pace is doing instead.
Notice what this does not forbid. Patents per year, miles of track laid per decade, transactions per second, deaths per thousand — these are real quantities, honestly computed, and the construction says they are exactly what they look like: present ratios over surviving records. The illegitimate move is not the arithmetic. It is the transfer afterward, when the answer is lifted off the shelf and hung on the decade as something the decade had. The nineteenth century industrialised at such a rate is a sentence about a record. The nineteenth century was slower is the same arithmetic wearing a claim it did not earn.
And once the count is located on the shelf, a second thing follows immediately. R9: the Past is inaccessible, and the abundance of present records makes it appear otherwise. The piece on history found that at full strength in an archive room. Turn it over and it says something about thin shelves too. Records do not survive at a constant rate. The further back a period sits, the fewer of its marks are still here to count — and on the construction above, fewer marks per unit of counter is not merely less evidence of a pace. It is, arithmetically, a slower one.
Second marked construction
The past reads slow to a reader here, because its record is thin. Apparent tempo tracks the survival of marks, not anything a period did. Every generation stands in the densest archive it will ever occupy, which is why every generation reports its own time as the fast one and the time before it as slower — and why the report does not move when the world does.
Mine, from the first construction with R9. The foundation gives inaccessibility and gives the abundance of present records as the illusion of access; the step from record density to apparent rate is my addition, and case one is its evidence rather than its proof.
Case one falls out of that. Beard in 1881 and Simmel in 1903 were standing in their own dense archive with a thinning one behind them, and so are we, and so will anyone be. The complaint is stable because the geometry that produces it is stable, and it will be filed in 2103 about a world we would find unrecognisably quick.
There is a harder version of the same point available, and it is checkable. In The Rise and Fall of American Growth, Robert Gordon argues that the transformative century ending around 1970 outran everything since, and reports United States labour productivity growing at 2.82 percent a year over 1920–1970 against 1.50 percent over 1890–1920 and 1.62 percent over 1970–2014, with total factor productivity in the middle window running at close to triple the rate of the periods on either side. I mark that as report, not grounds, and I will mark its edges too: it is a United States series, not a global one; the post-1970 slowdown is close to consensus while the size of the 1930s and 1940s spike is disputed — William Nordhaus, building an independent series, does not find it — and Gordon’s forecast for the decades ahead is an open bet with named opponents, not a finding. With all that said, the direction is the one that matters here. The decades a great many people file as slower are, on a measured ratio computed over surviving records, the faster ones. The felt ordering and the counted ordering come apart, and whatever the felt one is reading, the count is not it.
A rate with no denominator is not a small quantity or a disputed one. It is the shape of a quantity with the quantity missing.
Why the same hour runs at two speeds
The shelf ratio is not what case two was about. Nobody in that meeting was counting artifacts. Something else was being read, and the foundation is precise about what a person reads.
P4: R = A / E, Reality as the Actual over the Expectation, where the Expectation comprises subconscious prediction and ideas. R8: the human experiences only the quotient — not the Actual, not the Expectation, only the ratio between them. So a felt pace is not a count of marks. It is a reading of arrivals against a denominator that is already predicting arrivals, and the denominator is host-side and belongs to one person.
Case two is then not a puzzle. One hour, one set of contents, two denominators, two quotients. The person for whom it dragged brought a denominator expecting more than the hour delivered; the person for whom it flew brought one expecting less. Neither is misreading the room. There is no host-free reading of the room to be had.
And R2 gives the part that stings. Records update expectation forward and never re-open. So a stretch of dense arrivals raises the denominator, and the denominator stays raised when the stretch ends. The week after the launch is not a slack week; it is an ordinary week read against a denominator the launch built. The return from a fast job, a fast city, a fast year, to an ordinary one is felt as a fault in the ordinary one, and there is no fault in it. The piece on streaks worked the same forward reach under a run of wins, and I carry the shape over rather than re-deriving it.
Third marked construction
Felt pace is a quotient — arrivals read against a denominator that predicts arrivals — so it is not a property of the interval, and two hosts in one hour can honestly report opposite speeds. It is also asymmetric in time: the dense stretch raises the denominator and the denominator is still raised afterward, which is why the ordinary week that follows is read as having gone slack.
Consequence of P4 and R8 with R2. Mine: the primitives define the quotient and do not carry an expression for a rate of arrivals into it, and I say below what that costs.
Figure 01
Three objects, one word
- The period No rate Does not vary (P0, P3), so there is no numerator and frequency does not apply (R5). Undefined, not zero. This is the object the ordinary sentence points at and the one that has nothing to give it.
- The record A counted ratio Marks per counter-unit, both read now, over what survives. A real quantity, computed here. Thin where the record is thin (R9), which is most of what slower reports.
- The host A quotient Arrivals against a denominator that predicts arrivals (P4, R8). One per person. Raised by a dense stretch and still raised afterward (R2).
Figure 02
The slower century, counted
- The window Gordon calls the special century
- The windows on either side of it
What this costs
It costs the comparison. Everything is faster now is built to be a claim about now against then, and there is no then available to be the second term. Both terms in the sentence are here: a dense shelf and a raised denominator. The comparison people believe they are making across a century is being made, each time, between two things in the room.
It costs the return. You cannot go back to a slower time. Not because it is gone — that would leave the wish coherent and merely unsatisfiable — but because it was not slow. The longing has the form of a longing for a place and there is no place. The piece on nostalgia found the same result under warmth and colour; this is the harder case, because a warmth is a feeling with no arithmetic behind it and a rate looks like a number, so the wish for the old pace feels like a wish for something specific and measurable. It is not. What was measurable about it is on your shelf.
It costs the era as a defendant. If you are worn out, the exhaustion is real and it has causes, and not one of the causes is a decade. That will read as a cheat to anyone who has correctly noticed that more is asked of them than was asked of their parents — so let me say the rest of it. More may well be asked. That is a claim about counts on a present shelf: hours worked, jobs held, people cared for, notifications delivered. It is checkable, and this argument does not touch it. What the argument says is that the counting happens here, that the comparison is between two present records, and that the answer, whatever it comes out to, belongs to the record and not to an age.
R14, once more, on this idea as on every other: an idea is oriented and prejudiced in a single act, and so is this one. It is kind to the person who suspects their nostalgia is running a trick on them, and unkind to the person whose complaint about the speed of modern life is a complaint about a real and documentable load. Both readings are available and only the first one is what the piece argues. The second is a matter for the count.
The three constructions, and where each fails
- 01 · A pace is a present ratio over a present record
- Fails if a rate can be read off a period by an operation that does not go through present records. That is the falsifier the foundation already lists, in its ordinary clothes: an instrument aimed at elapsed variation itself rather than at a present state and a trusted rate. None exists, and case three is why. Produce one and this construction goes with it.
- 02 · The past reads slow because its record is thin
- Fails if apparent tempo did not track record density. It predicts something specific and checkable: unusually well-documented remote stretches — a decade with a famous diarist, a heavily filmed war, a city with an intact newspaper run — should read faster than their neighbours to a reader working from those records, and thinly documented recent ones should read slower. If density and apparent tempo came apart, this would be doing no work.
- 03 · Felt pace is a quotient with a lagging denominator
- Fails if two hosts in one interval reliably reported the same speed, or if the return from a dense stretch to an ordinary one reliably read as ordinary. The second is the sharper test, because it has a direction: the slack week should follow the fast month and not precede it. A symmetric effect would count against.
Where this is unfinished
Four edges, stated so someone can hold me to them.
The first is the rate at which a denominator moves. The third construction needs the Expectation to rise during a dense stretch and to come down afterward at some speed, and the foundation supplies no rate, no floor and no scale for that update. The piece on streaks stated the same gap and I inherit it rather than filling it. Everything I have said about the slack week is a claim about direction and not about how long it lasts, and I have no business saying how long it lasts.
The second is whether the record carries a magnitude or only an order. The axiom gives one consistent assignment and is silent on distance between marks; the piece on age read that as order-only and I inherit the reading and mark it as a reading. It matters more here than there, because a rate needs a magnitude in both terms. My answer is that the magnitude comes entirely from the counter, which is a present thing that varies, so the arithmetic survives the order-only reading intact. That is my answer and not the foundation’s, and if the record turned out to carry a metric of its own, the first construction would need rewriting rather than merely re-labelling.
The third is the machine, which arrives in this subject the moment anyone says the machines have sped everything up. A machine computes events per second in full and better than any of us. On this piece’s account it therefore has the middle tile of Figure 01 completely and the right-hand tile not at all, because the quotient is defined only where a host stands and a host is defined as a living system. Whether that is a difference in kind or an artefact of the present definition is a question the foundation records as open and instructs a piece to state rather than settle in passing. Stated.
The fourth is the boundary I drew at the start. Rate and remembered length are different quantities and I have argued about the first. They interact — a stretch that left many marks is both counted fast and remembered long, and there is a piece to be written about why those come apart as often as they agree. This is not that piece, and I have not assumed an answer to it.
A limit, stated plainly
If your days are too full, nothing above is a treatment for that. The load is a count on a present shelf — hours, obligations, people who need you, an illness, a second job — and relocating the arithmetic does not remove a single item from the list. A metaphysics can take a bad picture off a feeling. It cannot take a shift off a rota, and it must not be offered as though it could.
And if the feeling that life is running away from you has become a loop — a thing you are reading many times a day, with dread in it — that is not what this argument is for either. The reasoning is not a reason to be calmer and there is no schedule on which understanding a mechanism stops it. If it will not stop, that is work for a person, and the person should be a clinician. I would rather say that here than have it inferred from silence.
On a Tuesday
- When you say the world has sped up, supply the missing term. Faster than what, counted per what? The honest answer names two things you can point at today. If you cannot name them, you have not made a claim yet, and the feeling is still worth taking seriously as a feeling.
- If a week reads slack, check the denominator before you blame the week. Ask what it was set by and when. A week following a launch is not a slower week; it is an ordinary week being read by an instrument the launch adjusted.
- Slowing down is an act, not a return. There are two moves and they are different: reduce the arrivals, or change what you expect of them. Both are performed now, in the present, on your own account. Neither is a journey backward, and the one that people actually make when they move somewhere quiet is the first.
- When you want the old pace, name the marks. What specifically was on the shelf then that you want on it now — a long evening, a letter, a Sunday with nothing in it? Once there is a list, it is a plan. Left unlisted, it stays a longing for a rate that no period ever had.
None of which is fatalism, and the foundation is careful about the difference. P5: the host can occupy a position from which its response to a readable condition is revisable. R13: agency is relocated, not abolished; the host is the site of conscious responsibility. What is closed is the past, which was never running at a speed anyone could go back to. What is open is the count you take from here and the denominator you take it with, and both of those are yours today.
No period ever had a speed. A pace is an arithmetic done here — marks against a counter, both read now — and a feeling of pace is a quotient carried by one person. The past reads slow because its shelf is thin, and the age you are standing in will read slow to whoever is standing in the next one.