Late Spring, and the Marriage She Is Talked Into

Consider how an agreement is actually arrived at. Not the dramatic version, in which somebody is threatened or deceived by an enemy and gives way, but the ordinary version, which is far more common and has almost no vocabulary attached to it. A person says she does not want a thing. The people around her do not argue. They are warm about it. Over some weeks, one of them removes the reason she gave, another mentions what other people are saying, a third arranges an occasion that would be rude to decline. At no point is she overruled. At the end of it she says yes, and means it.

What happened there is difficult to name because every individual move in it was decent. Nobody lied for advantage. Nobody raised a voice. If you interrogate any single participant they will tell you, accurately, that they only wanted the best for her and that the decision was hers. And it was hers. She made it. She can give reasons for it, and the reasons are good ones, and she will hold them for the rest of her life.

There is a film about this, and it is not generally described as being about this. Late Spring, made at Shochiku in 1949 by Ozu Yasujiro, is usually introduced as a tender study of a daughter reluctant to leave her widowed father. That description is not false. It is simply pitched at the level of feeling, and the film is doing something at the level of procedure that the feeling conceals.

This Library has cited Ozu before, in an essay on Wong Kar-wai, where his restraint was set against Wong’s as a different kind of stillness — his an order absorbing change, Wong’s a labour of continuous refusal. He has never had a piece of his own here. This is the first, and it is not an appreciation. It is an argument that the famous serenity of this particular film is not its atmosphere. It is the thing under examination.

One disclosure before anything else, because it governs every sentence that follows. This Library did not screen the film for this essay. What is described below — the plot, the shots, the substance of what characters say — comes from published synopses, from the scholarship, and from reviews of record, and is described rather than reproduced. Where the published accounts disagree, that is said. Where they are silent, nothing is supplied.

The film was released in Japan on 19 September 1949, with a preview screening on the thirteenth. It runs 108 minutes. Ozu directed and wrote it with Noda Kogo, from a novel by Hirotsu Kazuo called Chichi to musume, Father and Daughter. Ryu Chishu plays Professor Somiya Shukichi, a widower. Hara Setsuko plays his daughter Noriko, who is twenty-seven and keeps his house. Sugimura Haruko plays Aunt Masa, Shukichi’s sister. Miyake Kuniko plays Miwa Akiko, a young widow. Tsukioka Yumeji plays Aya, Noriko’s divorced friend. In 1950 the film was voted best Japanese film of 1949 by the critics of Kinema Junpo, and took four Mainichi Film Awards, including one to Hara for her acting.

The situation is this. Aunt Masa decides it is time Noriko married and sets about it. Noriko declines, and gives a reason: she does not wish to marry anyone, because marrying would leave her father alone and unable to manage. This is not an evasion. It is true, and everyone can see that it is true, and it is the load-bearing objection in the entire film.

So Masa does not attack the reason. She removes it. She tells Noriko that she is also trying to arrange a match between Shukichi and Mrs Miwa. If that comes off, Noriko has nothing left to say. Shortly afterwards, at a Noh performance, Noriko watches her father greet Mrs Miwa warmly, and something in her closes. Later, pressing her to meet a prospective husband, Shukichi tells her he intends to marry Mrs Miwa. She agrees, reluctantly, to the meeting. She finds she likes the man. Under pressure from every direction, she consents.

Two things about that consent are worth fixing before anything is built on them. The man, Satake, is never shown on screen; the audience meets him only as a description, and one of the descriptions is that he resembles Gary Cooper. And the wedding ceremony itself is never shown either. Ozu omits the scene that any other filmmaker would have treated as obligatory. What we are given instead is the household before it and the father after it.

And then the fact that makes the film something other than a domestic drama. After the wedding, drinking with Aya, Shukichi admits that he was never going to marry Mrs Miwa. He invented it. He fabricated a future for himself in order to remove his daughter’s reason for staying. He goes home to the empty house, sits down, peels an apple until the peel drops, and lowers his head.

Set out as a sequence, the machinery is unusually clean, and its cleanness is the point. Noriko states a reason. The household does not dispute the reason; disputing it would have been an argument, and an argument is a thing you can lose or win or walk out of. Instead the household abolishes the condition the reason depends on. When the condition is gone, the refusal has nothing to stand on, and she is not defeated so much as left without a position.

That is a categorically different operation from persuasion, and it is worth being exact about the difference. Persuasion addresses the person: it offers her considerations and invites her to weigh them. Excuse-removal addresses the world: it alters the facts so that the conclusion she was resisting becomes the only reasonable one available to somebody with her own values. She is never asked to abandon anything she believes. She is simply placed in a world where what she believes now points the other way.

The second feature is that the pressure is distributed. The aunt raises it. The friend has a view. The colleague’s engagement is mentioned. The father, who does not want her to go at all, becomes its most effective agent. No single one of these people is doing enough to be called responsible, and each of them, asked, would say they had merely mentioned something. This is why nobody in the film can be identified as the one who did it, and why a viewer looking for the villain finds only relatives.

The third feature is the one that makes the serenity legible. There is no scene of confrontation because confrontation would have given her something to refuse. A raised voice announces itself as pressure and licenses resistance; concern does not. Warmth, in this arrangement, is not a decoration on the pressure. It is the delivery mechanism, and it works better than force would, because force produces a grievance and kindness produces a debt.

And a debt is unusually difficult to argue with. This Library made the point once about a condiment — a television drama in which somebody squeezes lemon over a shared plate without asking, and the offence turns out to be not greed but having decided on another person’s behalf what they would like, and having left them with an altered plate and no way to object without seeming petty. Late Spring is that structure at the scale of a life. The alteration is a marriage. The objection would require first disputing that it was for her benefit, which is a much larger and more hostile conversation than any of the individual moves seem to warrant.

The father’s invented remarriage is the most admired thing in the film and it deserves to be looked at coldly, because the admiration is doing work.

Shukichi does not want Noriko to leave. The published accounts are consistent that his loneliness after the wedding is the film’s closing image and that it is not presented as relief. He lies anyway, and he lies at his own expense: the fabrication costs him the thing he wants, and buys her a future he believes she needs. As an act of self-denial it is genuine, and it is the reason audiences leave the film moved rather than angry.

Consider what it is, mechanically. It is a false fact about the world, supplied to a person who is using facts about the world to decide what to do with her life. She revises her position in the light of it, correctly. Her reasoning is sound. Her inputs are not, and the person who corrupted them did so out of love and never told her.

This Library has a definition of betrayal that it arrived at in an essay about the inside of a married person’s head: a betrayal is the unilateral alteration of terms the other person is still relying on. That definition was built to be lenient about fantasy and strict about the things fantasy gets confused with, and it catches this squarely. Shukichi altered, by himself, without her, the single term her whole decision rested on, and let her go on planning around the altered version.

It is also the exact inverse of a problem this Library examined in a piece on a Japanese drama about four musicians who agree not to ask each other about their pasts. The line drawn there was that your own content is yours to withhold, but another person’s ability to decide is not, because that is not about you at all. Shukichi does not withhold an input. He manufactures one. If withholding the thing somebody needs in order to act is the failure that essay described, supplying a counterfeit is the same failure with the polarity reversed, and it is harder to see because it looks like generosity rather than reticence.

None of which makes him a villain, and the film is right not to treat him as one. It makes him the clearest available illustration of the essay’s actual claim: that the most effective interference with somebody’s choosing is performed by people who love them, at cost to themselves, and is experienced by everybody involved as care.

The Occupation context is relevant here for one precise reason, and only that reason. It is not offered as background colour.

Films made in Japan in 1949 passed through the censorship apparatus of the American Occupation before production could be approved. The film scholar Kyoko Hirano, in her study of Japanese cinema under the Occupation, reports that its officials considered the custom of arranged meetings for prospective marriage partners, miai, to be feudalistic, on the ground that the custom seemed to them to downgrade the importance of the individual. Hirano notes that had that policy been rigidly enforced, this film could not have been made at all, since the marriage of its heroine to a man she has met once at a single arranged meeting is its central event.

The detail that matters is recorded by Lars-Martin Sorensen, who worked through the censorship files on the film. In the synopsis the filmmakers were obliged to submit before approval, Noriko’s decision to marry was presented as a collective family decision rather than an individual choice, and the censors apparently rejected this. Other changes from the same process are on the record: the trip to Kyoto was originally motivated by a visit to the dead mother’s grave, which is absent from the finished film, possibly because it would have read as ancestor worship; a line about the ruins of Tokyo was replaced with a word meaning dusty; a reference to conscription by the Japanese Navy was softened to forced work.

Put those two things together and you get a fact so strange that it is usually mentioned and then dropped. An occupying authority which disapproved of arranged marriage on the ground that it diminished the individual required the script to convert a collective family decision into an individual choice. The individualised consent visible on screen is therefore not simply what Ozu and Noda observed. It is at least in part a required rendering, produced under review, by a foreign administration, in the name of the autonomy of the person consenting.

This Library does not claim to know Ozu’s intention, and Sorensen’s own reading — that Ozu was subverting the censorship to exalt tradition against imposed Western trends — is a contested one that other scholars dispute. The point here needs no view on intention at all. A film in which a woman’s agreement is manufactured by her family had its own account of that agreement rewritten, by a censor, into the form in which agreement was supposed to look free. Whatever else Late Spring is, it is an object that underwent the process it depicts.

The legal frame is what makes all of this visible as machinery rather than as custom, and it had changed within the previous eighteen months.

Article 24 of the Constitution that came into force in 1947 states that marriage shall be based only on the mutual consent of both sexes, and shall be maintained through mutual cooperation with the equal rights of husband and wife as a basis. The word doing the work is only. The revising Civil Code statute, Law No. 222, promulgated on 22 December 1947, carried the consequences through: the household system and the authority of the head of the household were abolished, and an adult no longer required a family elder’s permission in order to marry.

So in September 1949, when this film opened, the entire legal weight of a marriage rested on one thing: whether the woman had agreed. There was no longer anybody whose permission counted, no register of family authority that could substitute for her word, nothing else to point at. Her yes was the whole of it.

And the film’s argument, if it has one, is that this is exactly the condition under which the production of a yes becomes the most important activity in the room. When consent was one requirement among several, the machinery of the household was one influence among several. When consent is the only requirement, the household does not stop working; it redirects. Everything that used to be exercised as authority is now exercised as concern, because concern is what still has purchase on the one input that now matters.

It is worth adding what the form actually was, in numbers, so that nobody reads this as a story about a peculiar family. The National Institute of Population and Social Security Research collects, in its national survey of marriage and childbearing, a retrospective series on how couples met. For marriages contracted in the years 1945 to 1949, the series records arranged marriages at 59.8 per cent and love matches at 21.4 per cent. By the late nineteen-sixties the two lines had crossed; in the most recent figures arranged marriages are under a tenth. Noriko is not in an unusual arrangement. She is in the majority one, at the moment when the law had just stopped underwriting it and left it standing on her agreement alone.

Here is where the essay has to be careful, because the obvious conclusion is the wrong one. The obvious conclusion is that Noriko’s consent is fake, that she was engineered, and that what looks like her decision is really her aunt’s. That reading is tidy and it is false, and it insults her.

What the psychological literature on motivation offers here is a description rather than a verdict. Self-determination theory, developed by Edward Deci and Richard Ryan, distinguishes degrees to which a value that originated outside a person becomes part of how she regulates herself: an external demand can be swallowed whole and enforced by shame or approval-seeking, or it can be identified with, taken on as genuinely one’s own, and integrated with everything else she wants. The model’s useful implication for this essay is that an externally originating value is not thereby a foreign body. Internalisation is a real process with real degrees, and a want that arrived from outside can end up being, in every sense that matters to the person having it, hers.

There is also a small, honest piece of experimental evidence about the specific mechanism this film runs on, and it should be reported with its weaknesses rather than its headline. Social psychologists have studied a technique in which a request is accompanied by an explicit statement that the person is free to accept or refuse. A pre-registered meta-analysis published in 2023 examined fifty-two experiments and found an overall medium effect on compliance. It also found that when the analysis was restricted to the seven studies judged to be at low risk of bias, the effect disappeared, and that the reproducibility indicators for this literature are very poor. So: an interesting hypothesis about why declaring somebody’s freedom to refuse can increase the rate at which they comply, and not a demonstrated fact. Nothing in this essay rests on it, and it is cited here with its failure attached.

What the essay does rest on is a plainer observation that requires no laboratory. Two things are true of Noriko’s decision at once. It is hers: she made it, nobody physically prevented anything, she gave reasons, and by the end she is defending the marriage to her own father and apologising for what she calls her selfishness. And it is not hers: the reason she gave for refusing was removed by other people without her knowledge, and the central fact she reasoned from was false.

The instinct is to ask which of those is the real one. The essay’s claim is that the question has no answer, and that the absence of an answer is the finding rather than a failure to reach one. Agreement is produced in conditions arranged by others, and it is still agreement. A person can be genuinely the author of a choice she would not have made if the people who love her had behaved differently, and there is no operation you can perform on her afterwards that separates the two.

There is an image in this film which is probably the most interpreted shot in world cinema, and it would be cowardly to write about the film and avoid it, and worse to use it as decoration. So here is the only thing this essay wants from it.

At an inn in Kyoto, the night before the trip home, Noriko and her father lie on separate futons. She talks; she admits she had found the thought of his remarriage distasteful; he reassures her; she looks over and he is asleep, or appears to be. She looks at the ceiling and appears to smile. There follows a shot, about six seconds, of a vase on the floor in front of a paper screen with the shadows of branches on it. Cut back: she is close to tears. Then the same vase again, for about ten seconds, and the film moves on.

The scholar Abe Mark Nornes, writing about this crux, observes that nothing in all of Ozu’s films has provoked such conflicting explanations, and that everybody seems compelled to weigh in on it. He is right, and the published readings are genuinely incompatible with one another: the vase as stasis and transcendence; the vase as a container for the viewer’s feelings; the vase as a non-narrative element wedged into the action; the vase as a false point of view, since she is never shown looking at it; the vase as a feminine symbol; the vase as traditional Japan. Gilles Deleuze took it as an image of time itself, the unchanging form within which a change from near-joy to near-weeping occurs.

This essay declines to add a seventh. What it takes from the shot is the disagreement, not the object. At the precise moment when the film would have to tell us what Noriko feels about the thing being arranged for her, it cuts away to an object and holds. The interior is not shown. It is left as a slot, and six decades of criticism have filled the slot with six incompatible contents, each of them plausible, none of them in the film.

That is a useful discipline to accept, and it is aimed at this essay as much as at anybody. A piece arguing that Noriko’s consent was produced is itself a filling-in of the slot. The film does not say she was produced. It shows a sequence and withholds the verdict, and the withholding is not an oversight — it is the same director who declined to show the wedding.

This publication is attached to a service in Japan that sells private companionship: an evening, arranged in advance, paid for, with a stated beginning and a stated end. That is the commercial position from which the foregoing argument was written, and it is stated here rather than in a footnote because it is an interest, not a disclosure.

The interest is direct and it runs in an unflattering direction. A house that sells stated terms profits from an argument that unstated terms are how people get talked into things. If a reader finishes this essay with a heightened suspicion of ordinary kindness in her own family, that suspicion is, commercially speaking, useful to us. We would rather say that now than have it noticed later.

So four things are refused here, and each of them costs something. This essay does not claim that a paid, bounded arrangement is a purer form of consent than an unpaid one; this Library has argued at length elsewhere that paid or unpaid is the wrong axis, and nothing in Late Spring changes that. It does not claim that anything on offer here is a remedy for the condition the film describes, because an evening is not a remedy for a life. It makes no claim about any reader’s family, her marriage, or how she came to be in either. And it offers no advice: nobody reading this is being told to reconsider a decision, raise a subject, or say anything at all to anybody.

What stated terms actually buy is narrower than the essay’s own logic would like. They buy a reduction in one specific failure mode: the case where a person is reasoning from facts somebody else has quietly arranged. They do nothing about the deeper one, which is that wants are formed in company and always have been. An arrangement written down in advance is still an arrangement somebody proposed, in a market somebody built, to a person whose sense of what she might want was assembled by the world she lives in. That is not a problem stated terms can solve, and pretending otherwise would be the same trick in a cleaner suit.

Five objections, and the essay does not survive all of them intact.

The first is priority, and it is straightforward. This reading is not new. Roger Ebert, writing about the film, said that it is about two people who do not believe that a woman of a certain age must want a husband, and about how they are undone by their tact, their concern for each other, and their need to make others comfortable by seeming to agree with them. Robin Wood argued that the film is about the sacrifice of Noriko’s happiness in the interest of continuing tradition, and that everyone in it, including the father and finally Noriko herself, is complicit. Between them they have most of the mechanism. What is added here is the legal and censorship frame and the refusal to resolve the two truths about her consent, which is a narrower contribution than the essay’s confidence might suggest.

The second is Ozu’s own method, and it cuts deeper. Ozu said of himself, in a remark repeated everywhere, that he only made tofu, because that is what a tofu maker makes. He said he wanted to make people feel without resorting to drama. He called his most celebrated film too melodramatic. A body of work built on the refusal to underline, made by a man who declined to tell audiences what to conclude, does not obviously invite a reading that identifies a mechanism and names its operators. It is entirely possible that the serenity is the mood, that Ozu was depicting something he found sad and inevitable rather than diagnosing a procedure, and that an essay which converts his reticence into evidence of a hidden system is doing to the films what the critics did to the vase.

The third follows from the second and is the sharpest. Reading calm as machinery may say more about the reader than the film. A critic trained to look for hidden coercion will find it in any room where nobody is shouting, and the finding will be unfalsifiable, because the absence of visible pressure is being treated as evidence of pressure. That is a bad epistemic position to occupy and this essay occupies it. The honest response is not a defence but a limit: the claim is offered as a reading that makes a particular pattern visible, not as a discovery about what was really going on, and a reader who finds the film simply sad has not missed anything.

The fourth is about Hara Setsuko, and it is the one this essay is least able to answer. Her performance carries an interiority the script does not state, and every account of the film agrees that whatever is happening in Noriko is happening in her face rather than in her lines. An argument that treats Noriko as produced by her circumstances risks denying her precisely the thing the essay says she is owed, which is a self that is doing the wanting. This Library made an adjacent point in a recent piece on how restraint gets read: restraint performed by a man arrives as an act of withholding, a decision visibly made and held, while restraint performed by a woman tends to arrive as composure, as a quality of her rather than as something she is doing. An essay that describes Noriko’s agreement as produced is at risk of the same move at a higher level of sophistication — granting her a mechanism instead of an interior. The objection is registered and not rebutted.

The fifth is the commercial one, already conceded above and worth repeating in its harshest form. A house that sells explicitly stated arrangements has written an essay establishing that unstated ones are how people are moved without knowing it. The scrupulousness of the essay does not neutralise that. If anything it is what makes it work.

It does not claim that arranged introduction is a bad way to meet somebody, or that the marriage at the end of this film is unhappy. The film supplies no information about the marriage at all — the husband is never seen and the ceremony is never shown — and any account of what became of Noriko is invented by whoever is telling it.

It does not claim anything about Japanese families. The mechanism described here is not national. It is what happens wherever a decision matters to more than one person and the people involved are fond of each other, which is everywhere, and a reader in any country who recognises the sequence is not recognising something foreign.

It does not claim to know Ozu’s intention, and where the scholarship disagrees — on whether he complied with the censorship, subverted it, or simply worked around it — this essay takes no side. Nor does it claim that the censorship record settles the meaning of the film. It establishes something about the document, not about the mind that made it.

It asserts nothing whatever about Hara Setsuko’s life. She retired from acting in 1963 and lived privately until her death in 2015, which was not reported for some months at her family’s request. A great deal has been written about why she stopped and about her relations with Ozu. Almost all of it is speculation, it is identified as speculation by the sources that repeat it, and none of it appears in this essay.

It contains no clinical claim and no diagnosis. Nothing here says that any reader has been manipulated, that any agreement she has given was invalid, or that any relationship she is in operates like the household in this film. The self-determination material is cited as a model of how motivation is described, not as an assessment of anybody, and the compliance experiment is cited together with the finding that its effect vanished in the least biased studies.

And it does not claim that the problem has a solution. The essay’s own position is that a want can be produced in a person by people acting in good faith, and that the resulting choice is genuinely hers and genuinely not. That is not a diagnosis to be cured or a trap to be escaped. It is a description of the conditions under which anybody has ever agreed to anything, offered so that the next time a reader hears herself say she has decided, she has somewhere to stand while she looks at the sentence.