The Self · Saturday 12 September 2026
You are kneeling on a cold veranda in Heian-kyō, a little before sunrise, beside a woman who is about to be handed a stack of blank paper — and who will do something with it that nobody had thought to do before.
心ときめきするもの · kokoro tokimeki suru mono
Act one · The bundle
The shutters are still up. The lamp has been burning all night and is down to a thumb of oil. Outside, in the grey before colour, there is a noise in the eaves: sparrows, feeding their young. You will want to remember that noise in a minute.
The woman beside you is about thirty. We do not know her name. History calls her Sei Shōnagon — "Sei" from her father's family name, Kiyohara, and "Shōnagon" from a minor court post — which is a little like being remembered as Smith, Junior Clerk. She has served the Empress Teishi since 993. She reads Chinese, which women were not supposed to do well, and she is very, very quick.
Into this room, one day, comes the Empress's brother, Fujiwara no Korechika, Minister of the Centre. He is carrying paper. In the year 1000, paper is not stationery; it is closer to silver. A bundle of blank notebooks is a serious gift, and the question of what to write on it is a real question.
"One day Lord Korechika, the Minister of the Centre, brought the Empress a bundle of notebooks. 'What shall we do with them?' Her Majesty asked me. 'The Emperor has already made arrangements for copying the Records of the Historian.' 'Let me make them into a pillow,' I said. 'Very well,' said Her Majesty. 'You may have them.'"
The Pillow Book, closing section · Ivan Morris's translation, 1967
The Emperor is having the great Chinese history copied out — the Shiki, Sima Qian's Records of the Grand Historian — the most serious thing a person can do with paper. Sei Shōnagon's answer is a joke whose punchline scholars are still arguing about. Shiki, the history. Shiki, to spread out as bedding. Make it a history? No — make it a pillow.
And then she does the thing. Handed the most expensive blank space in Japan, she does not write a history, or a poem cycle, or a moral treatise. She writes down what she notices. Her carriage. Her cat. The way a certain kind of man gets out of an ox-cart. What annoys her. What, for no defensible reason, makes her chest tighten with pleasure at five in the morning.
A thousand years later it is still in print, and the histories the Emperor was so carefully copying are read mostly by specialists. This is the first thing the day wants to tell you: the record of a small noticing outlasts almost everything built to last.
She opens the book by refusing to praise a season. She praises an hour of each season — the exact moment when that season is most itself. Before you see her answer, give yours. Drag each slider to the hour of the day when spring, summer, autumn and winter are at their best. The window will follow you.
Every prompt here is skippable. Nothing you write is scored, and there are no right answers on this side of the page.
Morris, 1967
"In spring it is the dawn that is most beautiful."
McKinney, 2006
"In spring, the dawn — when the slowly paling mountain rim is tinged with red…"
Her four: spring — dawn. Summer — night. Autumn — dusk. Winter — the early morning. Notice that two of her four are the edges of the day, the hours when almost nobody is watching. Notice also that those two sentences above are the same sentence. Translation is the first honest wobble of the day: the woman you are about to read is always partly the person who carried her into your language.
Act two · The list
Scattered through the book — 164 of them, by the usual count — are the monozukushi, the "list-things": a heading, then a handful of items and nothing else. No argument. No moral. Depressing Things. Hateful Things. Rare Things. Things That Arouse a Fond Memory of the Past. Things That Should Be Large. Things That Should Be Short.
And one heading that, once you have read it, you will never quite stop running in your head. In the Japanese it is 心ときめきするもの — kokoro tokimeki suru mono. Kokoro, heart-and-mind, the two not separated. Tokimeki: the flutter, the throb, the quickening. Here is the whole of it.
Things that make one's heart beat faster
Morris's translation, section 16 in his numbering, section 29 in the standard Japanese one. The original opens: 雀の子飼ひ。ちご遊ばする所の前わたる。— suzume no ko kai. chigo asobasuru tokoro no mae wataru.
Seven entries. Not one of them is an achievement.
Look at what is actually in there. A bird feeding. Walking past children. A room that smells of last night's incense. A mirror going slightly cloudy — she is delighted by the slight cloudiness of a mirror. Washing your hair and dressing beautifully when nobody is coming. And rain thrown against the shutters on a night you are waiting for someone, so that for half a second the weather sounds like footsteps.
Two things make this list a piece of equipment rather than a pretty page. First, every entry is a moment, not a category. She does not write "nature" or "romance" or "self-care." She writes the second the sparrow's beak goes in. Second, three of the seven happen with nobody else present. She is not listing things that look good said out loud. That is the difference between a list of what you love and a list of what you would tell people you love, and it is the whole difference.
Here is your stack. Before you write anything: guess. How many things do you think you can find that genuinely make your heart go faster — not things you approve of, not things that sound good, things that actually do it?
Now the rule that makes this work, and it is her rule, not ours: name a moment, not a kind of thing. Not "music" — the half-second before the drums come back in. Not "my dog" — the way she sighs when she finally lies down. If you can't picture it happening at a particular time of day, it isn't specific enough yet.
Things that make my heart beat faster
The page is blank. That is exactly what she was handed.
Skippable, like everything on this side. The gauge counts words and word-types — nothing more. It cannot tell whether an entry is true, and it is not marking you.
Act three · The threads
Read her list again as a scientist would. Four of seven are sounds or smells — the sparrows, the incense, the rain on the shutters, and arguably the announcement at the gate. Three of seven happen alone. Five of seven happen at the edges of the day, early or late. Nobody told her to build it that way. She simply wrote what quickened her, and a shape showed up.
That is the trick worth stealing. You do not find out what you care about by deciding what you care about. You collect enough small true instances and then look for the thread. So: tag your entries. Then the page will draw the threads — and tell you, with numbers, what it can and cannot see.
Add a few entries above and they will appear here as islands to tag.
Which senses are involved? (any number)
Who else is there?
Roughly what hour does it happen?
Act four · The other list
If the book were only about beauty it would be unbearable. It is not. Two sections before the heart-quickening list sits Hateful Things (nikuki mono), and it is one of the great sustained complaints in literature. Someone who breaks into your story to supply a detail you were coming to. A dog that barks at a visitor who was meant to be secret. Fleas. And, best of all, this — which is a thousand years old and describes last Tuesday:
"You've just settled sleepily into bed when a mosquito announces itself with that thin little wail, and starts flying round your face."
Hateful Things · Meredith McKinney's translation, 2006
Here is why that matters for the instrument in your hands. It is the same eye. The attention that catches a mirror going faintly cloudy is the attention that catches a guest overstaying by four minutes. You do not get one without the other. People who are moved easily are irritated easily, and being told to be more positive is, quite often, being told to see less.
Permission, for one panel only, to be as petty as a Heian court lady. Small irritations, please — the specific kind. Not injustice; the sound of someone else's cutlery.
Hateful things
Empty. Suspiciously virtuous of you.
Optional, and deliberately light. If something genuinely heavy is sitting behind one of these, that is not a list item — that is a conversation, and it is worth having with somebody you trust.
Act five · Far yet near
近うて遠きもの。宮のまへの祭り。思はぬはらから、親族の仲。鞍馬のつづらをりといふ道。十二月のつごもりの日、正月のついたちの日のほど。
遠くて近きもの。極楽。舟の道。人の仲。
Act five · The instrument
Sei Shōnagon had no way to measure a quickened heart. She had a word for it and a piece of paper. Modern laboratories have rather more: heart-rate monitors, breathing belts, electrodes on the skin, and — since 2011 — a small camera called the GooseCam, developed by Mathias Benedek and Christian Kaernbach, which films a patch of forearm and runs a two-dimensional Fourier transform over the image to detect the exact instant the hairs stand up.
Point all of that at a person listening to music they love and you get what the literature calls chills, or frisson. In 2001 Anne Blood and Robert Zatorre put musically trained volunteers in a PET scanner with music the volunteers had chosen themselves, precisely because it gave them, in their words, "shivers-down-the-spine." As the chills got stronger, blood flow shifted in regions tied to reward and emotion — ventral striatum, midbrain, amygdala, orbitofrontal cortex. Heart rate, breathing and muscle activity moved too.
"Subjective reports of chills were accompanied by changes in heart rate, electromyogram, and respiration."
Blood & Zatorre, PNAS 98, 11818–11823, 2001
So: the tokimeki is real, and it is partly public. It leaves marks on instruments. And then comes the finding that turns this whole page around.
In 2021 Rémi de Fleurian and Marcus Pearce read 167 research reports on music-evoked chills and put the numbers side by side. Across four studies that measured actual goosebumps in people who had just pressed a button to say I feel it now, the hairs on the arm actually rose in only 57%, 43.1%, 40.7% and 40% of them.
When people say they felt chills, how often did the instrument agree?
Share of participants reporting music-evoked chills in whom measurable piloerection — hair standing up — was actually observed. Four studies, as tabulated by de Fleurian & Pearce, Psychological Bulletin 147(9), 2021.
Read that again, because it is the most useful sentence of the day. The people were not lying. Something happened to them. It is the instrument that only catches some of it. Your inner life is not fully visible to a camera pointed at your arm — and the review says so out loud: it is "likely that MECs, as reported by participants, do not always involve piloerection."
Which makes Sei Shōnagon's method — a word, a page, and paying attention — not a charming antique. It is the other half of the measurement.
This is the closest a web page can honestly get to a chills study. The page will build a short piece of music out of pure tones, live — no recording, just oscillators. It has the features the literature says are the usual suspects: a slow crescendo, a rising line, and a new voice entering near the end. Press and hold NOW whenever you feel anything: a lift, a shiver, a tightening, the thing you can't name. Afterwards you will see your own presses drawn against the actual loudness of what you heard.
Sound never starts on its own. If you would rather not listen, press Play anyway with sound muted and watch the loudness curve draw itself — then mark where you expected to be moved. That is a real measurement too, of a different thing.
How often does something — music, a film, a landscape, a voice, a memory — give you an actual physical shiver?
And what does it, when it does? (any number, or none)
If you answered "never," you are in ordinary company and nothing is missing from you: in one large survey between 8% and 23% of people said they never or rarely get chills, and the review's own estimate is that around 90% is the upper limit of people who can have them at all. Shōnagon's heading does not require goosebumps. It requires noticing.
Act six · The time capsule
Here is the obvious objection to the whole exercise. Sparrows? A cloudy mirror? Why would anyone record that? It feels like the definition of not worth keeping. Which is precisely the bias that five studies in 2014 went looking for.
Ting Zhang, Tami Kim, Alison Wood Brooks, Francesca Gino and Michael Norton had people build time capsules: write down a recent conversation, a song, an inside joke, an ordinary exam question. Then predict how interesting they would find it months later. Then they came back and read it.
People underestimated, every time. And the crucial result is in Study 3, which had people document both an extraordinary day — Valentine's Day — and an ordinary one, then randomly assigned which they reread three months later.
What people predicted vs what they actually felt, three months later
Mean "how interested are you / would you be in reading this" ratings, 1–7. Zhang, Kim, Brooks, Gino & Norton, Psychological Science 25(10), 2014, Study 3 (n = 130 of 152 at follow-up).
The special day was predicted about right: 4.36 expected, 4.65 felt. The ordinary day was badly misjudged: 4.04 expected, 4.73 felt — and the gap was large by the standards of this kind of research. Across the studies the pattern held: the more ordinary the thing recorded, the more its future value was underestimated.
One more number, because it is humbling. In a later study people predicted they would remember 73% of a conversation they had written down. A month later they actually remembered 42%. We are confident librarians of an archive that is quietly deleting itself.
So make the prediction, then make the capsule. Two dials first — and they are dials about you, not a test.
If you read today's page again three months from now, how interesting do you think it will be? (1 = not at all, 6 = very)
And what percentage of what you wrote today do you think you will still be able to recall without looking?
Skippable. Nothing here is sent anywhere by this page, and nothing here is graded.
Act seven · Your own heading
The lists are famous. The real invention is smaller and stranger: she made up the headings. Nobody gave her "Things That Make One's Heart Beat Faster." She noticed that a certain set of unrelated moments belonged together, and then built a category with exactly the shape of her own attention, and then went looking for more.
That is a tool you can own by tonight. A heading no one else would write. Things that are better in the rain. Things I only like because of who showed them to me. Rooms that sound like a particular year. Things that are funny only at the wrong moment. Write the heading first. Then hunt.
Game · Guess her heading
Honest wobbles
This subject attracts more confident nonsense than any other, so here is every soft place in today's story, including in its science.
Not a single page in her hand survives. The book reaches us through four manuscript families — Sankan-bon, Nōin-bon, Sakai-bon, Maeda-bon — which disagree about how many sections there are (roughly 190 to 330, depending on which you count and how) and even about whether the lists are scattered through the book or gathered together. The order you read her in is an editor's decision.
"Sei Shōnagon" is a court label. Her personal name is unrecorded; "Kiyohara no Nagiko" is a guess that scholars still argue over. She was born around 966, entered Empress Teishi's service in 993, and after about 1017 the record simply stops. Two later traditions have her dying poor as a nun, or married to a provincial official. Both may be invented.
The pun on Shiki (the Chinese history) and shiki (bedding) is one theory. Others: a notebook kept by the bed; a writer's stock of poetic place-names, utamakura; simply a precious thing. There is no scholarly consensus, and anyone who tells you flatly what the title means is choosing a side.
Morris's 1967 Shōnagon is an elegant lady diarist; McKinney's 2006 Shōnagon talks. McKinney's own reason is worth having: "Without the vividness of her personality, the words turn to dust." Both are faithful. Neither is her. Where this page quotes, it names the translator.
You will read that people who get chills from music score higher on Openness to Experience. Probably true — but the standard NEO personality questionnaire contains an item that asks, more or less, whether you get chills. Measuring openness with a chills question and then discovering that open people get chills is not nothing, but it is not a clean finding either. The 2021 review flags it explicitly.
Whether familiar music produces more chills: studies disagree. Whether heart rate rises: studies disagree. Several researchers now think "chills" is at least two or three different experiences — a warm one, a cold one, a moved-to-tears one — wearing the same word, and that treating them as one thing is exactly why the results keep quarrelling.
A handful of typed entries and four sliders are not a personality, a diagnosis or a score. This page counts words and tags. It cannot see you. The whole reason for the mirror at the end is that a mirror is the only instrument here that is honest about being one.
The long thread
The mirror
This is not a result. It is your own words and tags counted and handed back, the way a mirror hands back a face — with no opinion about it. Read it, then tell the page where it is wrong.
The graded part
Do this at home
You need: one sheet of paper folded in half, a pen, and one week. Nothing else, and no app.
1. On the front, write a heading nobody else would write. Not "good things." A heading with a shape — Things that are better cold. Things I notice only when I'm early.
2. Carry it. For seven days, add an entry the moment it happens, and obey the one rule: a moment, not a category. Aim for a sentence you could film.
3. On the inside-back, keep a second list: Hateful Things. Small ones. Enjoy yourself.
4. On day seven, read the front list aloud to someone and ask them one question: what do you think these have in common? They will see a thread you can't, because they are not standing inside your attention. Then put the sheet somewhere you will find it by accident in three months. That is the part the 2014 studies say you will underrate — and the part you will be gladdest of.
For a younger explorer: make it a game for two. Each of you writes a secret heading, then you swap and have to guess the other's heading from the items alone. It is surprisingly hard, and it is exactly what a biographer does.
Six things worth knowing
Sei Shōnagon wrote down a mirror going faintly cloudy, and a thousand years later strangers in languages she never heard of are still nodding at it. Nothing about her life was preserved because it was important. It was preserved because it was noticed. So: what did you notice today that nobody will ever record unless you do it?
Sources