Is time a reality that flows on its own, or something constructed, experienced, and narrated? One puzzle — and physics, the brain, story and literature each hand you a different face of it.
2026 · Book Recommendations · No. 29
We assume the one thing that needs no explanation is time — it flows evenly, the same for everyone. This issue's four books each pry that intuition open from a different field: a physicist argues that the single "now" and time's very direction are not fundamental features of the world; a science writer shows the brain holds no single clock; a historian shows "time travel" is a modern idea, invented only in 1895; and Proust shows that lost time hides in the senses, and only art can buy it back. The point isn't to memorize four opinions — it's to interrogate the most familiar and most alien thing there is from four angles that don't talk to each other.
| Book | Author | Year | The one thing it makes clear |
|---|---|---|---|
| The Order of Time | Carlo Rovelli | 2017 | "Time flows evenly and there is a shared now" is, in physics, an illusion — time's direction comes solely from rising entropy, which in turn comes from our blurred view |
| Time Travel: A History | James Gleick | 2016 | "Time travel" is not an ancient myth but an idea invented in 1895 — a projection of our anxiety about the future and our regret about the past |
| Why Time Flies | Alan Burdick | 2017 | The brain has no single clock; subjective time is constructed, and the density of novel experience sets how long a stretch feels in memory |
| In Search of Lost Time | Marcel Proust | 1913–1927 | Deliberate recollection is pale; the truly vivid past hides in involuntary, sense-triggered memory — and only art can buy it back from time |
Rovelli is a theoretical physicist, and the whole book methodically skins away our most deeply rooted intuitions about time. First cut: there is no shared "now." Relativity long ago proved that time's rate depends on speed and gravity — a clock on a mountaintop runs faster than one at sea level. This is no measurement error; it's real. So the question "what is happening on Mars right now" has no objective answer: there is no single instant valid for every place in the universe.
The second cut goes deeper: in the fundamental equations that describe the world, there is no distinction between "past" and "future." Newtonian mechanics, Maxwell's equations, quantum mechanics — reverse the time variable and they hold just the same. So where does our overpowering sense that "time has a direction and cannot be run backwards" come from? Rovelli's answer is entropy — the only physical quantity that distinguishes past from future is disorder, which only ever increases.
Here lies the book's most counterintuitive step: entropy is low in the past and high in the future because our view of the world is "blurred." As macroscopic creatures we cannot see each individual molecule, only the averaged state — and it is precisely this forced blurring that gives birth to the distinction between past and future. Adopt a view that resolves every microstate, and the direction of time vanishes. The arrow of time is not an objective property of the world; it is a by-product of how creatures like us deal with it.
And when everything is peeled away, what remains? Not a clock, Rovelli says, but a web of relations between events. The world is not a collection of things but of events; even a stone is just an event that changes very slowly. And the reason we still genuinely feel time flowing is that we ourselves are precisely beings made of memory and anticipation. Time is not a property of the world; it is the form of our existence, the source of our identity.
Rovelli is a standard-bearer for loop quantum gravity, and "time does not exist at the fundamental level" rests heavily on the frontier theory he himself champions — as yet unconfirmed by experiment. The poetic prose sometimes outruns the rigor of the argument, and readers can mistake a bold hypothesis for settled physics.
Rovelli's "there is no shared now" holds almost literally for BigCat's distributed-systems background. Distributed systems likewise have no global clock and no absolute "simultaneity" — which is exactly why Lamport logical clocks and vector clocks exist: what a system can reliably establish is causal order (happens-before), not a true total order. To try this week: pick a place in your design that "orders by global timestamp" and ask a Rovelli-style question — does this really need absolute simultaneity, or is causal order enough? Most obsessions with "globally consistent time" are the same illusion as the human craving for a "universal cosmic now," and the price is availability and latency. Treat the relations between events as primary, rather than some external clock, and the architecture often clears up at once.
Gleick's thesis is surprisingly specific: the idea of "driving a machine through time" was invented by H. G. Wells in The Time Machine in 1895. Before that, no culture imagined time this way. The ancients had prophecy, cycles, nostalgia for golden ages — but not "build a device, sail into the future, and steer back." Why precisely the end of the 19th century?
Because that was the first time humanity was struck by the thought that the future would be radically different from the present. The Industrial Revolution, Darwin, the accelerating pace of technology made "the future" for the first time a strange place, worth both dreading and imagining. Wells treated time as a fourth dimension — no different from the dimensions of space except that our consciousness moves along it. The time-travel story is a by-product of this brand-new conception of time.
Gleick's real insight is that time-travel stories were never about physics; they were about emotion. They are tools for handling two knots: anxiety about the future, and regret about the past ("if only I had..."). The grandfather paradox, going back to rewrite history, writing a letter to your younger self — all of these are ways of wrestling with "the past that cannot be changed" and "the future that cannot be known." What they really press on is determinism versus free will: if the future is already written, does the effort of this moment count for anything?
He also points to an irony: the more we can instantly record everything — photos, data, a cloud that never deletes — the more we are trapped in a flat, eternal present, and the more we lose the feeling of time actually passing. To be able to go to any moment is, in a sense, to have truly lived none of them.
Gleick is learned and wide-ranging, but the book is loosely structured — more a literary and scientific ramble around time travel than a tightly building argument. Readers after a hard answer to "is time travel physically possible" will be disappointed; its ambition is always cultural history, not physics.
The illusion Gleick exposes — treating "the future" as a place that already exists, waiting to be predicted — is exactly what BigCat should watch for in the AI wave. Most judgments of the form "what AI will be in three years" smuggle in a Wellsian assumption: the future is a single, fixed destination. To try this week: take one confident prediction about AI's future and rewrite it as three diverging stories, each paired with "the earliest observable signal." This isn't indecision; it's swapping "predict a point" for "monitor a distribution." Gleick's reminder earns its keep: the future is not a place but a story we tell ourselves and bet on — and since it's a story, you should keep more than one version ready.
Burdick is a New Yorker writer who spent years in the labs that study time, and reaches a plain but subversive conclusion: there is no single clock in the brain. Timekeeping is distributed — different regions handle milliseconds, seconds, the circadian rhythm (the last governed by the suprachiasmatic nucleus), and they do not add up to one unified "sense of time." Your experience of time is a construction stitched together on the fly by several systems, not the objective reading of some clock.
The puzzle closest to daily life: why does time speed up as we age? One explanation is proportion — to a 5-year-old a year is a fifth of a life; to a 50-year-old it is a fiftieth, and naturally feels shorter. But Burdick weights a second mechanism more heavily: the density of novel experience.
The key distinction: time felt in the moment and time felt in retrospect are two different things, and often the opposite. A day crammed with novel stimuli feels frantic and fast in the moment, yet long in memory (dense with markers); a month of routine doesn't feel slow as it passes, yet compresses to an instant in recollection (no anchors). Childhood is long because almost everything is a first time; midlife is fast because the brain automates the repetition and barely records it.
Burdick also debunks the legend that "time slows down in a crisis" — the neuroscientist Eagleman had people free-fall and measured their in-the-moment time resolution, and found time did not actually slow: fear simply encoded those few seconds into memory with unusual density, so they only "seemed" long in recollection. Once again: the sense of time is constructed by the brain, not passively received.
Burdick's strength is taking the reader into the lab; his weakness is the lack of a unifying theory — the book reads like a platter of fine reportage, and you finish knowing many intriguing fragments without assembling a full map of "how the sense of time works."
Burdick's law — "the density of novel experience sets the length of time in memory" — is a tonic for BigCat as a mother. The feeling that a child "grew up in the blink of an eye," that one's own years run "faster and faster," is one and the same mechanism: repetition smooths the days flat. To try this week: schedule one small thing you and your child have never done before — a street you've never walked, a dish you've never cooked, a little place you've never been. This isn't a diversion; it's deliberately planting memory anchors by Burdick's mechanism. A counterintuitive corollary: to make a stretch of time "grow longer in memory," don't live it more leisurely — live it more freshly. Childhood is long precisely because nearly everything back then was a first.
The seven-volume In Search of Lost Time is often treated as a daunting monument, but its core is one clear discovery. Proust distinguishes two kinds of memory: voluntary memory — what you deliberately recall, which yields a dry inventory of facts, pale and distorted; and involuntary memory — triggered abruptly by a smell, a taste, a texture, in which a whole stretch of the past floods back intact, with its original feeling attached. The former is what you can "summon"; the latter is what "ambushes" you — and the latter is the true one.
Then the famous madeleine: the narrator dips a small cake into lime-blossom tea, and at the touch on his tongue the whole Combray of his childhood "rises out of a cup of tea." The point is not nostalgia but a discovery — a past that deliberate searching cannot recover is preserved whole in the senses, awaiting a chance key. This is exactly why an old song or a long-gone perfume can strike a person defenseless.
From this comes the book's ultimate wager (cashed out in the final volume, Time Regained): time destroys everything — people age and die, love cools, memory fades. The only thing that can buy back "what is lost" from time is art. Capture those involuntary memories, write them into a work, and lost time gains a kind of permanence in another form. The destination of In Search of Lost Time is the narrator finally resolving to write the very book we hold in our hands.
This echoes the first three in a striking way: the physicist strips time of its objectivity, the neuroscientist says the sense of time is constructed, the historian says time travel is a projection of modern anxiety — and Proust offers an exit: since the living past hides in the senses and in memory, awakening it and fixing it through art is the only truly effective way a person can resist the passing of time.
Proust's sentences are famously long — one can wind on for half a page — and easily put off a first-time reader; the seven volumes are a real barrier for the busy, so pair it with a guide and read selected volumes. It captures a sense of time belonging to a leisured, aristocratic life, at some distance from today's pace — but for that very reason, it forces you to slow down.
Proust's "voluntary memory is pale, involuntary memory is vivid" points straight at the blind spot of heavy note-takers and second-brain users like BigCat. Our notes are almost entirely products of voluntary memory — structured, searchable, yet bone-dry: they keep the conclusion and lose the context. And the smell, the sound, the bodily state of the moment are the very keys that later let you "live a whole stretch again." To try this week: when you record an important idea, add one line of the sensory anchor of the moment — where you were, what you heard, what triggered it. This isn't artistic indulgence; it's burying a key to involuntary memory for your future self. A step further: have AI periodically re-surface these anchors and stage a Proustian "recovery" — what comes back will be far more than the single conclusion you wrote down.
List every point in the design that "must order by absolute time" and ask each one: would switching to causal order (happens-before) break it? If most really need only causal order, you are paying for a nonexistent "shared now" with availability and latency. Cases that genuinely need a total order (financial reconciliation, say) are few, and usually secured by an explicit consensus protocol rather than an assumption of synchronized clocks — separate the two classes and the architecture clears up.
Run the numbers: thinking back to the last "fast" month, how many things can you name that were unique to it and wouldn't have happened in any other month? Three or more = it has anchors in memory, you were just busy in the moment; none = it was smoothed flat by repetition. By Burdick's mechanism, "adding time" to life comes not from adding leisure but from adding novelty — deliberately breaking the routine, even with a street you've never walked.
First separate the two kinds of memory: what you can actively "summon" (mostly faded to a fact-list) and what suddenly "ambushes" you (vivid but uncontrolled). Proust's claim is that the latter is truer. The one fixing action available: write it down — not the conclusion, but the sensory detail that triggered it (the light that day, the melody, the smell). A recorded sensory anchor is the only preparation you can make for involuntary memory; the rest is left to chance.