Consent Is a Conversation That Keeps Going

She said yes at the door. Not to any one thing; to the evening, to the idea of it, to the man, to the version of herself that had decided, on the train, to be someone who says yes. It was a real yes and she does not regret it.

It is now twenty to ten, and something has changed. Not dramatically. The room is warmer than she expected, the pace is faster, a hand is somewhere she had not pictured, and the feeling in her chest has shifted from anticipation to something closer to watching. She wants to say: slower. Or: not that. Or: could we stop for a minute. And she does not, because she said yes at the door, and it feels, at twenty to ten, that the yes was spent then and she has no more of it to give or to take back. To speak now would be to go back on her word.

This essay is about that feeling, which is one of the most common experiences women describe about intimacy and one of the least often named. Its claim is simple. Consent is not a gate she passes through once. It is a conversation that keeps going, and a yes given at the door is a report on how she felt at the door, nothing more. Everything after it is new information. The question has to keep being asked, and she has to be allowed to keep answering.

The word 同意, consent, collects several different models, and the confusion between them is where the twenty-to-ten feeling comes from.

There is the legal model: consent is an event that either occurred or did not, and the law’s job is to decide which. There is the contractual model: consent is a document, a form, a set of terms agreed in advance, and what was signed is what governs. There is the cultural model that Japan is especially fluent in: consent is inferred, read from the air, from the fact of her having come, from her not having said no. And there is the model that popular sex education has settled on in the last decade, often summarised by the acronym FRIES: consent should be freely given, reversible, informed, enthusiastic and specific.

Classification: FRIES is an educational mnemonic from a reproductive-health organisation, not a research finding; its value is as a checklist of the ways a yes can be hollow. What it supports: that “reversible” and “specific” are part of the definition, not additions to it. A yes that cannot be withdrawn is not a yes, and a yes to the evening is not a yes to the hand.

The first three models are all door models. They locate consent at a moment, a signature, or an arrival, and once the moment has passed they treat the matter as closed. The fourth is a conversation model, and it is the only one that describes what actually happens to a person over the course of two hours.

Michaela Coel’s I May Destroy You, broadcast in 2020, is the most complete recent work of fiction about consent, and it earns that description by refusing to treat consent as one thing.

The orientation a viewer needs: Arabella is a young writer in London, famous for a book that began as tweets, under pressure to deliver a second. One night she takes a break from writing to meet friends at a bar, and the next morning she has a cut on her forehead, a broken phone, and a memory that surfaces in fragments: a man over her in a toilet cubicle. Her drink was spiked. The twelve episodes follow her attempt to reconstruct the night and to keep living while she does, alongside her friends Terry, an actress, and Kwame, a fitness instructor. The tone moves between comedy and something much harder without warning, and the series is structured as a spiral rather than an investigation. The spoiler boundary for this essay is the final episode, which offers several endings and will not be described.

What makes the series useful here is not Arabella’s assault, which is a case where no consent was possible, but the other scenes, which are all cases where consent was given and turned out to be several things at once. Arabella goes home with a man she likes, Zain, and consents to sex; midway, he removes the condom without telling her, and she learns afterwards, from a friend, that this has a name. The series is explicit that the yes she gave did not cover what was done, because the act changed and she was not told. Terry, on a trip, has a threesome with two men she believes are strangers to each other and discovers they arrived together; her consent was informed by a lie. Kwame meets a man through an app, has consensual sex, and is then held down for something he did not agree to, by a person he had just said yes to.

Classification: a drama, fiction, written from the author’s own experience; it establishes nothing statistically. What it supports is a distinction the door models cannot make: in every one of those scenes a yes was real and a violation still occurred, because the yes was to one thing and something else happened. Consent did not fail at the door. It failed at the point where the conversation should have continued and did not.

Dossie Easton and Janet Hardy’s The Ethical Slut, first published in 1997 and revised twice since, is a manual for people in open relationships, and it might seem far from a woman at twenty to ten in a room with one man. Its usefulness is that people who conduct several relationships at once had to work out, decades before the mainstream did, how consent operates when it cannot be assumed.

The book’s architecture: it reclaims the word in its title, argues that jealousy is a feeling and not a verdict, and spends most of its length on the practical technology of agreements. Its central practice is negotiation as a continuous activity rather than a one-time contract: agreements are made, kept, revisited on a schedule, and renegotiated when anyone’s circumstances change. It popularised the yes/no/maybe list as a tool for finding out, in advance and in writing, what each person wants and does not, and it insists that a maybe is a maybe. It is cheerful, Californian, and written from inside a subculture; its evidence is anecdote and long practice, not study.

Classification: a practitioner manual, subcultural, American, anecdotal; it supports nothing about populations and everything about method. What transfers is the method: consent as a standing agreement with check-ins built in, revisable by either party at any time, and a written maybe that protects the person who wrote it. What does not transfer is the tone, and a Japanese reader should not feel obliged to adopt the culture in order to borrow the tool.

The book’s deepest point is one the door models cannot see. If consent is an event, then renegotiating it is a breach. If consent is an agreement, then renegotiating it is the agreement working as designed. The woman at twenty to ten is not going back on her word. She is doing what the word was for.

The mechanism deserves to be spelled out, because the door model is intuitive and its failure is not.

The body updates. Emily Nagoski’s research synthesis on desire describes arousal as a running calculation between an accelerator and a brake, and the brake responds to context in real time: warmth, pace, a hand in an unexpected place, the sense of being watched rather than met. The state she consented in at the door is not the state she is in at twenty to ten, and no amount of good faith at the door can bind a nervous system that has since changed rooms.

New information arrives. A condom is removed. A third person turns out to be known. A location changes. The act she pictured is not the act that is happening. Every one of these is a fact she did not have when she said yes, and Coel’s series is built on the principle that a yes given without the fact does not extend to it.

And there is the sunk yes. Having said yes, she experiences stopping as a cost that she incurs, a withdrawal of something she offered, an embarrassment, a reflection on her reliability. The door model puts the entire cost of changing course on the person who changes it. The conversation model distributes the cost: if the question is being asked, then answering it differently is not a reversal but a reply.

This is why the twenty-to-ten feeling is so common and so silent. She is not weak and she is not confused. She is operating a model of consent in which she has already used up her turn.

The word 性的同意, sexual consent, was almost absent from ordinary Japanese until the last decade. It entered public speech through the #MeToo years and through the long, public case of the journalist Itō Shiori, and it became statutory in 2023, when the Penal Code was reformed to replace the old requirement of proven violence or intimidation with an offence of non-consensual sexual intercourse, 不同意性交等罪.

The reform is unusually instructive here, because instead of defining consent it lists the circumstances in which a person cannot form, express or act on a refusal. The list includes force and threat, mental or physical impairment, alcohol or drugs, sleep or unclear consciousness, not having the time to form or express a refusal, fear or shock at a situation that differs from what was expected, the effects of abuse, and pressure from economic or social position. Classification: statute, public record; the paraphrase is the essay’s and should be checked against the text. What it supports: the legislature itself has described consent as something that can be lost mid-way, through intoxication that accumulates, through a situation that turns out to differ from what was expected, through not having a moment to speak. That is a description of a conversation that stopped, not of a gate that was never opened.

The culture around the law has not caught up. Consent is barely taught in schools; a survey-of-the-air model, in which asking is considered unromantic and the partner is expected to 察する, remains the ordinary practice; and a woman who speaks at twenty to ten risks being heard as difficult, or worse, as having been ambivalent all along. The structural fact underneath is a norm of not troubling others with one’s inner weather, which makes the one sentence she most needs to say the one most expensive to say.

For a woman in Japan whose intimate conversations happen in a second language, the door model has an extra trap: her yes may be the only word she has.

A yes in a second language is often a smaller yes than it sounds. It is the word she has ready, not the sentence she means. “OK” and 大丈夫 both mean something between agreement and endurance, and a partner who does not share her first language will tend to hear the larger version. The check-in, if it comes, may arrive in a register she cannot answer quickly, and the pause while she searches for the words reads, to the other person, as hesitation or as assent depending on what he hopes.

Across cultures there are also different norms about how much pressure is ordinary. What one culture reads as persistence, another reads as coercion; what one reads as coyness, another reads as refusal. The conversation model is the only one that survives this, because it does not rely on reading. It relies on asking, in plain words, and waiting for an answer in plain words, however long that takes.

If consent is a conversation, it has a grammar, and it can be described without turning intimacy into a questionnaire.

It sounds like a check-in that is itself a touch: still good? in a voice that expects either answer. It sounds like a pre-agreed signal, a word or a gesture that means stop or slow and costs nothing to use, so that the sentence she cannot say at twenty to ten has a shorter form. It sounds like the change-of-state rule: any new element, a new act, a new place, a new person, a removed protection, is a new question, not an extension of the old answer. It sounds like an ending that costs nothing: if she stops, the evening does not become a failure or a debt, and nobody sulks.

And it sounds like the conversation after. What was that like. What did you want more of. What would you not do again. The after-conversation is the part that turns one evening’s consent into the next one’s information, and almost nobody has it, because the door model treats the evening as closed when it ends.

None of this is a mood-killer. The belief that asking kills the mood is the door model’s best defence, and it is false in practice: the mood is killed by the woman at twenty to ten who has gone somewhere else and cannot say so. A question brings her back. Silence is what she leaves in.

An honest essay gives the door model its due, and it has three real arguments.

First, constant asking can be its own pressure. Are you sure, are you sure, are you still sure can be a way of not hearing the first answer, and a partner who checks in every ninety seconds is not attending to her; he is managing his own anxiety with her body as the instrument. The conversation model needs restraint as much as it needs questions.

Second, the contract has value. A written yes/no/maybe, agreed in advance, protects the person who wrote it precisely because it exists outside the momentum of the room. Where the conversation model says everything is revisable, the contract says some things were decided when she was thinking clearly, and that decision deserves weight. The two are not enemies. The written map is what the conversation refers back to.

Third, some consent is consent to not being asked. A pre-negotiated scene of surrender, agreed in detail beforehand with a stop word in place, is a case where the person has explicitly chosen an evening in which the questions stop, and the conversation model, applied mechanically, would take that choice away from her. The answer is that the stop word is the conversation, compressed to one word and always available; the questions have not stopped, they have been delegated to her hand.

And there is the legal objection: a consent that is always ongoing is always deniable, and a system that cannot fix consent at a moment cannot adjudicate it. That is true, and it is why law needs events. But a woman at twenty to ten is not adjudicating. She is living, and the model she needs is the one that lets her speak.

Moonlight’s position, offered after the door has had its defence, is that most women have only ever been given the door model, and that the experience of a conversation that actually continues changes what they expect everywhere else.

The structure is what the essay has described. A written Desire and Boundary Map before anything, in which a maybe is never treated as a yes. Spoken re-consent at every transition, not as a formality but as a question that expects either answer. A stop word that belongs to her alone and ends the evening without explanation or cost. And the after-conversation, which is where the next map begins. The provider’s job is to keep the question open, including when she has not thought to ask it of herself.

The objection Moonlight must make against itself is that in a paid context the provider has a reason to prefer that the evening continue, and that his questions could be shaped by that reason. We accept the risk and answer it in two ways: the structure does not depend on his sincerity, because the stop word does not require his agreement; and the provider has his own no, which he uses. Whether the conversation was real is something she judges afterwards, in the after-conversation, and then alone.

She said yes at the door and she meant it. At twenty to ten something has changed, and she has a sentence she is not saying because she believes her turn is over.

Her turn is not over. The yes at the door was the beginning of a conversation, not the end of one, and the sentence she has now is not a reversal. It is the next line. Slower. Not that. Could we stop for a minute. Each of those is consent working exactly as it should, and the only thing wrong in the room is that nobody has asked.

She can say it anyway. She was always allowed to. That is what the word means.