Quiet pages, and patience,
and a cup of tea.

Every novel has its own climate. You enter and find that the air is different.

I have been thinking about this lately, because I picked up a book I last read in my thirties — Penelope Fitzgerald's The Blue Flower — and I felt, immediately, the same cold and dry air I remembered. Not in the prose, not even in the descriptions of weather, but somewhere underneath. The way the sentences sit on the page. The way Fitzgerald withholds.

I think most readers know what I mean. Kawabata has the weather of a January morning: thin sunlight, breath visible. Banana Yoshimoto has the weather of a kitchen at 11pm when someone you love is finally asleep. Tove Jansson, in The Summer Book, has the weather of a small island in August — bright, but with the first cold edge of autumn waiting in the next chapter.

When I was translating, I used to test my drafts by closing my eyes and asking myself: what is the temperature? If the translation was warmer than the original, I had failed. Usually I had failed by being too polite — by adding small softenings that the English did not have, because Japanese is a language that apologises easily and English is a language that does not.

I think this is what people mean when they talk about voice, although the word is too thin. Voice is only part of it. Voice is what the writer says. The weather is what the writer cannot help.

A small exercise, if you are a writer: take down three books you love and open each to a random page. Read one paragraph from each, out loud, and try to describe — only to yourself — what season it is inside that paragraph. Not what season the book is set in. What season the prose itself contains.

I find that my own writing is almost always set in late autumn. I did not choose this. It chose me, the way the cats chose me, by arriving.

I do not read many books a year. I read perhaps twenty-five, and of those, I reread five.

I mention this because I have noticed, in recent years, a quiet competition online about how many books people finish. Fifty-two a year. One hundred. One a week. The list-keeping has a flavour of accomplishment that makes me uneasy. Reading is not exercise. The metric is not pages per hour.

When I was younger I read faster, and I remember less. The books I have kept, the ones that live inside me now, are almost all books I read slowly — sometimes a chapter a day, sometimes a paragraph in the morning and another in the evening, with the rest of the day spent thinking about what I had read without realising I was thinking about it.

Kawabata's Snow Country took me three months the first time. I was twenty-two and I thought I was reading badly. Now I think I was reading correctly.

A book that you finish quickly will leave you quickly. A book you live inside, even briefly, will stay.

I do not say this to scold anyone. If you like to read fast, please read fast — it is your time, your pleasure. But I want to make a small case for the other thing. For the book left open on the table. For going back to page sixty-four because something happened there that you did not notice the first time. For the bookmark that stays in the same place for a week, not because you are stuck, but because that is where the book wants you to wait.

My translator's habit, I think. We read every sentence three times.

I have been trying to write a novel for four years. It has had three titles and two narrators and one room.

The room is the only thing that has stayed. It is a six-mat room on the second floor of a house in Sapporo, where my heroine — her name keeps changing, but let us call her Setsu today — does her ironing on Sunday afternoons. There is a window. There is a low table. There is a kettle on a hot plate, because the kitchen is downstairs and she does not want to go down. That is all.

I cannot move forward in the novel unless I am in this room. When I try to write a scene in a cafe, or at her office, or on the train, the prose goes dead. I have learned to stop fighting this. I now know that the novel lives in the room, and any time I want it to come closer to me, I must go upstairs.

I think every writer has one of these. A place where the work waits. Sometimes it is a real place from your childhood. Sometimes it is invented but feels more real than the room you are sitting in now. The trick is to find it and then to not redecorate. Do not improve the room. Do not add a balcony. Let the room be small. The smaller it is, the more it can hold.

When I read Tove Jansson, I can feel that she had such a room — or rather, an island. When I read Kawabata, I feel he had an inn. Yoshimoto has a kitchen. Fitzgerald, perhaps, a study with a window on a courtyard, in a building she did not own.

Where is your room? You may not know yet. You will know when you next try to write and the words come without effort. Wherever you were in that moment — that is your room. Go back tomorrow.

A friend lent me John Truby's The Anatomy of Story last autumn. I have been arguing with it ever since, which is, I think, the highest compliment I can pay a book about craft.

Truby's book is not famous in the way Save the Cat is famous. You do not see it on every writing-Twitter thread. I am glad of this. I think it is the better book — by which I mean, the book that asks more of you, and gives more in return.

Truby's central claim is that a story is not a structure but an organism. It grows from a single seed (he calls it the "designing principle") and every part of it — character, plot, theme, scene, dialogue — must be expressions of that one seed, or the organism will be sick. This is, I think, true. And it is the opposite of how most plotting advice works, which treats story as a series of beats you arrange like furniture.

His twenty-two steps are more demanding than Snyder's fifteen, and they are stranger. "Moral Need" is one of them — the flaw the hero must overcome not just to win, but to deserve winning. "Opponent's plan and main counterattack" is another. "Battle" and then "Self-revelation" and then "New equilibrium." The vocabulary is not American-cheerful. It sounds, faintly, like a Russian novel.

Truby teaches you to ask what the story is for, before he teaches you what shape it should be. This order matters.

What I argue with, in Truby, is the same thing I argue with in all Western story manuals: the assumption that the protagonist must want something specific, and pursue it, and grow through pursuing it. This is a perfectly good description of a great many stories. It is not, however, a description of The Old Capital or The Sound of the Mountain or The Pillow Book or much of what I love.

Kawabata's Shingo, in The Sound of the Mountain, does not pursue. He listens. He grows old. He notices his daughter-in-law in a new way. He has dreams. The novel is a masterpiece, and Truby's twenty-two steps would not find most of it.

So my advice, if you read Truby — and I do recommend reading him, more than Snyder — is this: take his diagnostic seriously. Ask yourself what your designing principle is. Ask what your character's moral need is. These are deep, useful questions. But do not believe that all twenty-two steps must be present in your book. Some books — quiet books, observed books, the books that move slowly because that is their nature — leave out fifteen of the twenty-two and are better for it.

The organism Truby describes is real. He is just describing one species of it. There are others, growing more slowly, in different soil.

Kawabata is often called a writer of beauty. I think he is a writer of absence.

I reread Thousand Cranes last month, the third time in my life. In my twenties I read it as a love story. In my thirties I read it as a story about grief. Now, in my fifties, I read it as a story about rooms after the person has left them. The tea bowl on the mat. The kimono folded on the chair. The conversation that happens in the same alcove where another conversation, years ago, happened. The novel is full of furniture, and the furniture is doing most of the work.

This is, I think, the great gift of Kawabata to writers. He shows that setting can be a character not because it is described in detail, but because it remembers. His rooms remember. When Kikuji enters the tea cottage, he is not entering a place; he is entering a piece of someone's afterlife.

Western fiction often forgets this. A room is a backdrop, painted in once and then ignored. But in our tradition — and not only ours; you find this in Marilynne Robinson, and in Penelope Fitzgerald, and in the very last pages of Stoner — the room is a witness. It has been there longer than the people in it. It will be there after them.

I think this is why I cannot finish my novel until I have spent enough time in Setsu's six-mat room. I am waiting for the room to remember her. When it does, I will be able to write what she does there, and the writing will feel necessary instead of invented.

This may take another year. I am trying not to mind.

A note on Quilva, and on what it has been like, for the first time in twenty years of working with long sentences, to use software that understands a book is not a long sentence.

I will start with a small experience that surprised me, because everything else I want to say grew out of it.

In my novel there is a character named Hibiki. She is, in chapter one, a girl on a bicycle. In chapter four she is referred to only as the daughter, because we are inside her mother's point of view and her mother does not, in that scene, want to think of her by name. In chapter eleven she has become an adult and is called Hibiki-san by the neighbours. In chapter nineteen she carries a particular blue umbrella that has belonged to her family for three generations, and for the rest of the book the umbrella is, in a sense, also her — the way an object can be a person in Japanese fiction, the way Proust's madeleine is also his aunt.

One afternoon, idly, I opened the part of Quilva called the Atlas and went to Hibiki's profile. There is a small list there called Other Names. I added the daughter. I added Hibiki-san. I added the blue umbrella. I did not know what would happen.

What happened was that the program, quietly, showed me every appearance of Hibiki in the manuscript — including every passage where I had not used her name. The scene where her mother thinks of her without thinking of her. The umbrella in chapter twenty-three, sitting closed by the door. A line of dialogue in chapter thirty where another character says Hibiki-san is late again and I had forgotten I had even written it. The book, suddenly, knew its own shadows.

I sat with this for some time. I have spent my whole working life as a translator, and the central skill of translation is precisely this: tracking the shadows. A character is rarely just their name. They are their title, their nickname, the pronoun another character uses for them, the small object that has stood in for them since chapter three. To translate well, you must hold all of these together in your head, simultaneously, for the length of a book. It is the most quietly exhausting part of the work.

I had assumed, without ever quite framing the assumption, that software could not do this. That a book, to a computer, was a long string of characters, and that the relationships inside it — the way a name and an umbrella could be the same person — were only ever in the writer's head, where they belonged. I had made peace with that. I had built my own systems of notebooks and coloured pens to compensate, the way every long-form writer eventually does.

Quilva, very calmly, does not assume this. It treats a book as what a book actually is, which is a network — a web of relationships between people, places, ideas, objects, and the words that name and rename them — and it maintains that web for you while you write.

Once you notice this, you start seeing it everywhere in the software.

The cards I use to plan a scene are not just notes. They live inside the scene; they carry the same tags as the characters who appear in them; they show me their little icons in the sidebar while I am writing the scene below them. The note I left myself in the planning view three weeks ago is sitting beside me in the writing view today, because the program understands that the note and the scene are the same thing, seen from two angles.

The Timeline is the book again, this time laid out as horizontal strips — main plot, subplot, a character's interior arc — and when I drag a card from one strip to another, the card's tags update themselves to match the move. The book has noticed that I have changed my mind about what this scene belongs to.

The Bubbles view is the book again, this time as a mind map; I can collapse an entire subplot out of sight while I work on the spine, then Alt-click an orphaned scene onto a different chapter and watch the chapter, the timeline, the outliner, and the planner all update at once. There is a friend of mine who taught creative writing for many years using the coloured-card method — cutting out little rectangles of card, pinning them to the wall, moving them around with her hands until the shape of a book emerged. The first time I showed her this view, she went very quiet, and then said: this is the wall. They have made the wall. She is right. It is the coloured-card method as software — except the cards know what is on them, and the wall remembers, and nothing falls behind the sofa.

A book is not a long sentence. It is a web of relationships that happen to come out, in the end, in sentences. Software that understands this is rarer than it should be.

I keep returning to that word — rarer than it should be. The idea that a manuscript is a network is not new. Every working novelist I know already thinks this way; we just have to do the network-keeping privately, in our heads, with notebooks and tabs and an ache behind the eyes by the end of a long writing day. What is new is that someone has finally built the network, and let the writer think inside it. I do not understand why this took so long.

Some smaller things I have grown attached to, briefly:

There is a setting called Typewriter scroll. When it is on, the page moves under your cursor instead of the cursor moving down the page; you type in the same comfortable place on the screen forever, and the words flow up out of view at the top as the page advances beneath you. I did not think I wanted this. I turned it on to humour the developer who had asked me to try it. After fifteen minutes I could not understand how I had written for twenty years without it. My neck used to ache by the end of the second hour of a long manuscript day. Now my eyes stay where they were when I sat down, and the page moves, and I do not.

There is a setting that lets you rename the four levels of your project — Book, Part, Chapter, Scene — to whatever words match your own mental shape of a long work. I renamed mine into Japanese before I started, and the menus then spoke my language back to me, in my structure, and the software disappeared a little, which is what software is supposed to do.

Each scene can hold up to four versions, side by side, and you switch between them with one click. A bold experiment never overwrites the draft you trust. I cannot describe how much that one detail has changed my willingness to try a worse idea, knowing the better one is still there.

And the entire project — outline, cards, timeline, character profiles, research images, manuscript, and the final book design — lives in one file. One file. I can email the file to a friend and she opens it and the whole studio is in front of her, intact. There is a calmness to this that is difficult to convey to anyone who has not lost their work to a cloud outage at exactly the wrong moment.

What it does not do, and where I would not yet recommend it. Quilva is young software. The screenplay industry has standards built around very particular tools, and if you are writing for film at a professional level, you will probably still hand your script off to the established screenwriting programs at the end. The font size in screenplay mode is locked — for good reasons, I now understand, to do with the one-page-one-minute convention — but those reasons take a moment to accept. There is no cloud synchronisation: your project lives on your computer, in a file, and is yours. For me this is a relief. For someone who writes across three devices and a phone, it may be a problem.

The learning curve is also real, though shorter than it looks. The first week I felt I had moved into a house with too many rooms. By the third week I had begun to live in three of them and to know that the others were waiting, in case I needed them. This is, I think, the correct way to approach the software. Do not try to use every tab at once. Pick the room where the work begins for you, and let the others arrive in their own time.

A small note in the spirit of honesty, since I have spent so much of this page in praise. The Quilva team — a small studio in Europe — wrote to me last autumn and offered me a free licence in exchange for testing the software in its late stages and sending back what I noticed. I did, and I sent back many small complaints, several of which they fixed. They are not paying me for this piece. They have not asked me to write it. But I think it matters that you know how the software arrived in my house, and you can weigh what I have said against that knowledge.

I do not often write sentences like the next one, so I will write it carefully. Of every piece of writing software I have used in twenty years of working with words, Quilva is the only one that has treated my book as the thing it actually is. Not a long document. Not a stack of chapters. A living web of relationships that finally has, somewhere outside my own tired head, a place to live coherently. For a slow writer in Kyoto, in her fifties, working on her first novel in her third language, that has turned out to matter very much — and I suspect it would matter, in different ways, for almost any writer who has ever lost track of one of their own shadows.

If you are curious, they have a free demo at quilva.com. I would suggest you try the Atlas first, with a project you already know well. Add a character. Add their other names. Watch the book begin to recognise them.

It is, I think, the moment most people understand what the software is for. It was for me.

The hardest part of translating from English to Japanese is not the words. It is the pauses.

When I translated my first novel, a quiet British book about a widow in Suffolk, I worked very hard on the vocabulary. I made elaborate notes on the names of birds, of garden plants, of the small foods she ate. I asked English friends what a certain kind of biscuit tasted like. I was proud of my draft.

My editor read it and said, gently: Aiko-san, the silence is wrong.

I did not understand. I had translated every sentence. I had not removed any silence. How could it be wrong?

It took me a year to understand what she meant. In English, silence happens between paragraphs. The white space does the work. In Japanese, silence often happens inside the sentence — in the gaps the language naturally leaves, in the particles, in the dropped subjects, in what is not said because it does not need to be. If you translate an English sentence word for word into Japanese, you preserve the words but you destroy the silence, because the silence is in a different place.

A translator is a person who must move a house from one country to another, and rebuild it so that the windows still face the morning.

I think this is why translation taught me to write. Before translation, I thought writing was about choosing the right words. After translation, I knew that writing is about choosing the right silences — and then placing words around them, so that the silences can be heard.

When I edit my own work now, I read it twice. Once for the words. Once for the spaces between them. The second reading is the more important one.

I have read Joseph Campbell. I respect him. But I have always felt that his monomyth was made for a particular kind of person — and that person is not me, and is probably not most of the people I love.

Campbell's hero is called to adventure. He refuses. He is mentored. He crosses a threshold. He faces trials. He is reborn. He returns, changed, to share his gift.

This is a beautiful pattern, and it is true to many lives. But it is also, if we are honest, the shape of a young man's life — usually a young man with the freedom to leave home. It assumes a threshold. It assumes that going is the way change happens.

What about the woman who never leaves her village, but who, over forty years, becomes someone her younger self would not recognise? What about the widower who stays in the same flat for thirty years and slowly, like a stone polished by water, becomes wise? What about my grandmother, who lived in the same house from age nineteen until her death at ninety-three, and was the most transformed person I have ever known?

There is, I think, another kind of journey — one that happens in place. The Japanese have a word for the patience this requires: 我慢 (gaman). It does not mean passivity. It means the active, willed staying-with. The choice not to flee. The hero of this kind of story does not cross a threshold; she lives long enough at the threshold for the threshold itself to dissolve.

If you are writing such a character — a character who does not go — do not feel that you must invent an adventure for her. Trust the room. Trust the slow work. The Western monomyth is one story. There are others. The grandmother in her chair, watching the seasons pass through the same window for sixty years, is also on a journey. Hers is just quieter, and harder to film.

I write in the morning, very early, before the world has its opinions ready.

My routine is unremarkable and I will describe it anyway, because I have found that other writers' routines comfort me, and perhaps mine will comfort someone.

I wake at 5:30. I do not look at my phone. I go downstairs, fill the kettle, and put genmaicha leaves in a small pot. While the water is heating I open the back door for a moment, even in winter, to feel the temperature of the day. The cats follow me. I am not sure if they want food or company; perhaps they do not distinguish.

I take the tea upstairs to my desk. I sit. I do not write yet. I drink the first cup looking at the wall.

The first cup is for nothing. This is important. If the first cup has a purpose, the writing will also have a purpose, and writing with a purpose is usually bad writing. The first cup is for becoming a person who is not yet trying.

Begin in idleness. The work will arrive when it sees that you are not waiting for it.

I write for about ninety minutes. By 8am I am usually done — not because I have run out of words, but because I have learned that ninety minutes is the most my mind can give without starting to perform. After that I would start writing what I think writing should sound like, instead of writing what is actually in me. So I stop. I make a second pot of tea. I go for a walk along the river. The morning is over.

I do not write in the afternoon. The afternoon is for translation work, or for reading, or for the small tasks of a life. The morning is the only room I have, and I do not let other things in.

I came to Tove Jansson late. I was forty-four. A Finnish friend gave me The Summer Book and said, with some firmness, you will love this. She was correct.

What I learned from Jansson, more than anything, is how to end. Not a chapter, but a whole work. Western novels often end with an explosion or a wedding or a death. Jansson ends, almost always, with weather. With someone going inside because it is getting cold. With a boat being pulled up onto the rocks for the winter. With a small gesture that does not resolve the story so much as release it.

This is harder than it looks. To end without resolving requires that the writer trust the reader entirely. You must believe that the reader has been paying attention, and that they will know what to feel without being told. Many writers do not trust their readers this much, and so they explain. They write a last paragraph that tells you what the book was about. I always feel slightly insulted by such paragraphs, even when the book itself was wonderful.

Here is an exercise I have started to do, and which I recommend. Take a novel you love. Cover the last page with your hand. Read up to where your hand begins. Then ask yourself: what would I have done here? Where would I have ended it? Sometimes — not always, but sometimes — you will find that you would have ended the book one page earlier than the author did. And you will be right. Most novels have one paragraph too many. The writer was nervous about leaving.

Jansson is almost never nervous about leaving. She closes the door quietly and goes.

I did not begin writing my own work until I was forty-three. I want to say something about this, in case it is useful.

For twenty years before that, I translated. I worked on other people's sentences. I lived inside other people's rooms. I thought, occasionally, about writing something of my own, and then I would feel a kind of dizziness — a sense that the door to that room was locked, or that I had lost the key a long time ago, or that the room had perhaps never existed and I had imagined it.

When I finally began, I expected to feel liberated. Instead I felt clumsy. The sentences I had spent twenty years admiring in other writers refused to come from my hand. What came was small, and a little stiff, and embarrassingly mine.

It took me three years to make peace with this. To understand that beginning late does not mean catching up. I will never have the body of work that someone who started at twenty-three will have. I have to accept this, and write the smaller, later thing that is actually mine.

The plum tree that flowers in March is not behind the cherry that flowers in April. It is a different tree.

If you are reading this and you are also beginning late — at thirty, at fifty, at seventy — I want to say: it is not too late, but it is also not earlier than it is. You will not write the books you would have written at twenty-five. You will write the books that only the older you can write. Those are real books. They are perhaps the best ones.

My grandmother began painting at sixty-eight. She died at ninety-three with one hundred and twenty paintings. She was not a great painter. But she was an entirely happy one, in a way that earlier-beginning painters often are not.

I think about her often.

There is a Japanese word, ma (間), which is usually translated as "space" or "interval" or "gap." None of these are quite right.

Ma is the active emptiness between two things. The pause between two notes of music — not as silence, but as part of the music. The space in a tea room around a single hanging scroll. The few seconds after a guest has spoken and before you reply, which in Japanese conversation are not awkwardness but courtesy. The gap is the meaning.

Western readers sometimes find Japanese literature "slow" or "uneventful." I think what they are noticing is ma. The narrative is not slow; it is making room. The room is not empty; it is full of what is about to happen, or what has just happened. To read for ma, you have to let the gap do its work. You cannot rush across it. You have to sit in it, the way you sit in a quiet temple courtyard, and wait until the silence becomes something.

I think this is what is missing from a lot of the writing advice I read online. The advice is almost always about filling. Add tension. Raise the stakes. Cut anything that does not advance the plot. These are not bad rules — they are excellent rules for one kind of story. But they have no concept of ma. They treat the gap as a failure. In our tradition the gap is sometimes the whole point.

If you are a writer trying to learn from Japanese fiction, this is the door. Not the cherry blossoms. Not the haiku. The gap. Sit in the gap.

A list, with brief notes. Not the books I am currently reading — those move. These are the books I have not moved in years.

Kawabata, The Sound of the Mountain. I have read this six times. Each time I find another paragraph I did not see before. The novel is about an old man listening to his garden. It is also about everything else.

Tove Jansson, The Summer Book. A six-year-old girl and her grandmother on an island. Almost nothing happens, except that you understand, by the end, what love is. The English translation by Thomas Teal is one of the best translations I have ever read in any direction.

Penelope Fitzgerald, The Beginning of Spring. Set in Moscow in 1913. A printer's life. Fitzgerald began writing novels at sixty. I keep her near for that reason as much as for the prose.

Banana Yoshimoto, Kitchen. I do not always agree with Yoshimoto, but I always trust her. She writes about grief in a way that is very Japanese and very modern at the same time. I read her when I am sad and want to stop being sad without lying to myself.

Marilynne Robinson, Gilead. A Christian American novel, very far from my world, and yet — the voice. An old man writing a letter to a child he will not live to see grow up. The patience of the prose. I keep it for the patience.

Sei Shōnagon, The Pillow Book. A thousand years old and still the funniest thing in my house. She is a courtier in the imperial palace and she is making lists: things that quicken the heart, things that look better at night, things that should be small. I aspire, in my own way, to her lists.

What is on your bedside? I am always curious. Not the new books — the ones you have not moved.