The Days of Abandonment, and the Self That Was Load-Bearing

Consider the sequence rather than the feeling. A man tells his wife, on an ordinary afternoon, that he is leaving. He is calm about it. He takes his things. He goes. And then, over the weeks that follow, a specific list of things stops happening in that flat. The post is not opened. The bills are not paid. The shopping is not done, or is done and then left in the bags. The children are not got out of the door at the right hour. The dog is not walked when the dog expects to be walked. The telephone is not answered, and then, at the worst possible moment, it cannot be.

Nobody decided any of this. Nobody refused. There was no scene in which a woman announced that she was no longer going to run a household. The tasks simply stop having anyone assigned to them, because the person who had been assigned to all of them simultaneously has stopped being able to hold the list.

This is the middle of Elena Ferrante’s second novel, and it is most often described as a descent into madness. That description is available and it is not stupid. This Library thinks it is the wrong shape for what is actually on the page. What is on the page is not, first, a mind coming apart. It is a household discovering, in real time and at the expense of a child and an animal, that it had exactly one functioning member — and that she had not been playing the part of one. She had been the part of the building that carries the weight.

That distinction is the whole essay. A role is something you can stop performing and still be standing. A load-bearing wall is not. And the thing that makes this book frightening rather than merely sad is that it is written from inside the wall.

I giorni dell’abbandono was published in Rome by Edizioni e/o in 2002. It was Ferrante’s second novel, ten years after her first. It appeared in English in 2005 as The Days of Abandonment, translated by Ann Goldstein and published by Europa Editions, at 192 pages. An Italian film adaptation directed by Roberto Faenza, with Margherita Buy as Olga, was released in 2005. In July 2024 the New York Times placed the book at number ninety-two on its list of the hundred best books of the twenty-first century, compiled from ballots submitted by five hundred and three writers, critics and other book people.

The narrator is Olga, thirty-eight, from Naples, living in Turin in a city she has never much taken to. Her husband Mario tells her he is leaving after fifteen years. There is another woman, Carla, considerably younger. There are two children, Gianni and Ilaria, and a dog, Otto. In the flat below lives a man called Aldo Carrano, a musician who plays in an orchestra and whom Olga barely registers as a person for most of the book.

The novel’s central set piece arrives on a day in August. Her son has a high fever. The dog shows the signs of having been poisoned. The lock on the front door will not turn. The telephone will not work. She is inside her own flat with two children and a dying animal and she cannot get out and cannot call for help.

Running underneath all of it is a figure from Olga’s Neapolitan childhood, la poverella, the poor woman: someone left by her husband who lost everything including the name people had known her by, and who drowned herself. Olga does not merely remember her. She keeps checking whether she is becoming her.

Read the August day for what it is procedurally. A key that will not turn is not a metaphor first. It is a door that will not open. A phone that will not work is not a symbol of isolation. It is a phone. A child with a fever needs a thermometer, a decision and possibly a doctor, and those are tasks, and on that day there is nobody in the building to whom they belong.

This is why the novel does its damage. Grief in fiction is usually rendered as weather: something that descends, colours everything, and lifts. Ferrante renders it as an operational failure with a timestamp. She is precise about order. What goes first is not hope and not dignity. It is ordinary competence — the ability to finish a sentence, to remember what you did an hour ago, to work a lock. Judgement goes later. Self-respect goes later still. The small mechanical acts go first, and that ordering is the finding.

And the cost lands on people who did not choose it. Gianni is ill and inadequately attended to. Ilaria is left to manage an adult who is not managing. Otto is dying on the floor. This Library will not excuse that and will not condemn it either. The book does neither, which is one of the few genuinely difficult things about it, and a reader who arrives wanting a verdict on Olga as a mother will find that the novel has quietly declined to hold the trial.

What it offers instead is an accounting. Here is what was being done. Here is who was doing it. Here is what happens when that person stops. The accounting is unbearable precisely because it is an accounting and not an accusation.

This Library has read this novel before, some years back in the reading order, in the piece filed as days-of-abandonment-a-self-larger-than-wife. That essay argued that the small daily acts of the marriage were not expressions of it but scaffolding for a self, and that when the audience for them left, the acts lost their occasion and stopped. It is a good reading and this one does not replace it.

It does push against it in one place. Scaffolding is temporary and external; it holds a building while the building is being made, and it comes down. What Ferrante describes is not that. The functions Olga was performing were not for show and they did not need an audience to be real. Somebody has to buy the food. Somebody has to know when the boy’s shoes stopped fitting. Somebody has to hold the calendar of a family of four in their head, continuously, for fifteen years, and that is not a performance, it is labour, and it was load-bearing.

The difference matters because it changes what the collapse is evidence of. On the scaffolding reading, Olga falls apart because the meaning drained out of her acts. On this reading, the acts had meaning all along, the meaning was that they kept four people and a dog alive, and what collapses is not their significance but their execution — because there was never a second person qualified to execute them.

Mario does not come back for Otto. No system absorbs the shortfall. The building does not redistribute the load, because buildings do not. This is the part the phrase moving on cannot reach.

Here is the part that concerns a reader who is not being left by anybody and has no plans to be.

Olga did not arrive at thirty-eight already unable to be alone. She had wanted things. She had had aspirations of her own and, in the Goldstein text, she says she had put them aside to fall in with his. She is more exact than that elsewhere, and the exactness is arithmetic: she describes having taken away her own time and added it to his to make him more powerful, and having disappeared into his minutes, into his hours. That is not the language of someone who was robbed. It is the language of someone who did a transfer, in increments, over fifteen years, each increment entirely reasonable at the time.

The danger in a long arrangement is not that it will end. Most of them do end, one way or another, and people survive the ending. The danger is subtler and it is about custody. An arrangement can quietly take over functions that a person has stopped performing for herself: being the one who decides what an evening is for, holding her own attention for an hour without a task attached, having an interior that is not routinely narrated to somebody else. These are capacities, and capacities go slack. The arrangement does not steal them. It performs them on her behalf, competently, for years, and she does not notice that she has stopped.

And then, if the arrangement ends, those capacities do not come back in the state they were lent. They come back the way a language comes back after twenty years of not speaking it: recognisable, humiliating, slower than you remember. This house has argued something adjacent in the mission essay time-that-belongs-to-you-is-not-left-over, which makes the case that time which is merely unclaimed by anyone else is not the same as time that is yours, and that forty spare minutes cannot be used by a person who has lost the habit of using them. Olga’s middle section is what that argument looks like at full scale and at the worst possible moment.

The Library keeps returning to the difference between a role and a person, and this novel supplies the mechanism.

A role is efficient. It answers questions before they are asked. It tells you what to want on a Sunday, how to behave at your husband’s work dinner, what your day is for. In the essay holiday-rituals-turn-intimacy-into-roles this house argued that defaults are not defeated by renegotiation, they merely go dormant, and that an occasion reactivates them. What Ferrante adds is the further step: a role does not only prescribe. It substitutes. It will perform, on your behalf, functions you have stopped performing yourself, and it will do so without ever sending you a notice that the substitution has occurred.

This is why the collapse is so fast and so total. Olga is not slowly losing a person she had been maintaining in parallel. There was no parallel. Wife and mother were not garments over a self; for fifteen years they were the operating system, and Ferrante has the nerve to show what happens when the operating system is uninstalled by somebody else without warning.

Which is also why this house has insisted, in intimacy-and-solitude-capacity-to-be-separate, that two people who have dissolved into each other have not achieved maximum closeness but have removed the condition that makes closeness possible. That essay made the case structurally. This novel makes it by demonstration, and the demonstration includes a dead dog.

A reader in English is not reading Ferrante. They are reading Ann Goldstein, and Goldstein’s choices are real choices that a reader is entitled to know about.

Goldstein was for many years head of the copy department at the New Yorker, and came to Italian as an adult, beginning lessons with colleagues around 1987; Europa Editions selected her to translate Ferrante in 2004. She has described her own manner as straightforward and literal-minded, and has said that Ferrante’s sentences tumble out in a way that is hard to make both sensible and grammatical in English, because Italian permits clauses to be piled up in ways English syntax resists. That is a translator telling you, plainly, that the English is calmer on the page than the Italian is.

The title is where the loss is easiest to see. In Italian, abbandono is not one thing. The Treccani dictionary gives the sense of being left or deserted, and it also gives, separately, the bodily sense of limbs relaxed and untensed, and the moral sense of entrusting oneself to another’s will with complete trust, and a mystical sense of surrender. The Italian title therefore holds, in one word, both what Mario did to Olga and what Olga had done to herself over fifteen years. English has abandonment for the first sense and abandon for the second, and they are different words. Goldstein had no move available. The English title is accurate and it is narrower, and every reader of the translation is reading a book whose title has been made to mean one thing.

For a reader working in Japanese the gap is wider still. The National Diet Library catalogue records the Neapolitan quartet in Japanese, translated by Ryosuke Iida for Hayakawa, and records this novel only in Italian, English and German. What that shows is what the catalogue holds; it is not a claim about what any bookshop stocks. But it does mean that most readers here will meet Olga through a second language at best, and that is worth saying out loud rather than letting a review imply otherwise.

Olga becomes foul-mouthed. She thinks and says things about her husband, his lover and her own body that are coarse in a register the earlier Olga did not use. Reviewers noticed it in 2005 and readers still flinch. The standard move is to file this under deterioration: the nice woman is going, look what is coming out of her.

This Library reads it the other way. A woman who has spent fifteen years being the reasonable one has been speaking in a register that was selected for her. Scholarship on the novel has made a version of this point: Esme Hodson, writing in Italian Studies in 2024, describes Olga as trained in canonical reactions and as exercising a surveillance over herself that is not her own, and reads the book as searching for models of female existence that do not have to be conceptualised according to prescriptions about how women behave. Victor Xavier Zarour Zarzar, in MLN in 2020, maps Olga’s disintegration as semantic and linguistic as well as physical, and argues that plotting — the act of putting things into an order — is what does the reconstructive work.

Put those together and the obscenity stops looking like a leak. It looks like the recovery of a register. The vocabulary Olga uses in the worst weeks is vocabulary she has always had and has been declining to use, because using it was incompatible with the function she was performing. Her fury is not a fog. It is analytic. She is, throughout, working out what was done and what she did, and she is doing it in the only words that fit.

It is also, and this matters, not a therapy. Nothing here is being claimed about what saying the unsayable does for anybody. The point is narrower and it is about reading: a reader who classifies Olga’s speech as a symptom has to discard its content, and the content is the argument of the book.

The novel does end, and it ends with Olga alive, the children alive, the household running again, and a tentative relationship beginning with Carrano, the musician downstairs whom she had earlier used badly and misjudged. She sees him play in an orchestra and revises her estimate of him.

What the ending conspicuously does not supply is a lesson. Olga does not emerge wiser in any way she could put into a sentence. She has not learned to be independent. She has not discovered that she never needed him. She is not shown having built, off-stage, the interior life the marriage had been performing on her behalf. She has survived, which is a smaller and more honest thing, and the book declines to upgrade it.

This is the hardest discipline in the novel and the easiest thing for an essay to spoil. Ferrante has been explicit for thirty years that a book does not need its author standing behind it explaining what it meant; the 1991 letter to her publishers at Edizioni e/o, written before her first novel appeared, says that books once written have no need of their authors, and that if they have something to say they will find readers. Whatever else that position is, it is a refusal to supply the moral afterwards. The novel is built the same way.

A reader who wants Olga to be a better person by page 192 is asking the book to say that the collapse was worth it. It was not worth it. Otto is still dead. The value of the last section is not that she grew, but that the functions resumed — and, in the order Ferrante gives them, the acts come back before the feelings do.

This is a publication attached to a business, and the business is bounded, scheduled, paid-for private companionship. So the interest has to be declared here rather than at the bottom, and declared against itself first.

An essay arguing that long arrangements absorb capacities is, straightforwardly, an essay that makes a market for arrangements which cannot absorb anything, because they end at a fixed hour and resume nothing. That is a convenient argument for us to be making, and a reader is entitled to weigh it accordingly.

So here is what this house does not claim. It does not claim that a scheduled evening builds the capacity to be alone; it cannot, and nothing in this essay should be read as suggesting that a purchased hour restores an unshared interior. It does not claim that a marriage is a worse arrangement than any other, or that Olga should have kept something in reserve. It does not offer itself as a hedge against being left. And it does not claim that the capacities described here can be bought back, by us or by anybody, because the argument of the essay is precisely that they are rebuilt by use and not by provision.

What a bounded arrangement can honestly offer is narrower: an occasion that belongs to one person, which somebody else has taken responsibility for organising, and which has no ongoing claim on her afterwards. Whether that is worth anything is her call, and the essay above is not evidence that it is.

The strongest objection is that this reading does the exact thing Ferrante spent the novel refusing. Turn the collapse into a structural finding and you have converted Olga’s worst weeks into a case study with a takeaway, and a takeaway is moralising with better manners. The book is ferociously uninterested in what Olga should have done differently. An essay that extracts a principle from her disintegration has made her useful, and making her useful is a form of not reading her.

The second objection follows from it and is sharper because of where this is published. There is a version of this argument that is simply an old warning to women about depending on men, dressed in structural vocabulary. This Library does not want that reading and has tried to close it off, but wanting is not the same as succeeding, and a reader who comes away thinking do not let yourself need anybody has found a door in this essay that the essay left open. The honest response is not a rebuttal. It is to say that such a conclusion would be a misuse of both the novel and the argument, and that the failure would be partly ours.

A third: a novel is not evidence. Olga is a made thing, arranged by a writer to demonstrate something. Using her household as a finding about households is a category error, and no amount of care in the prose repairs it. The most this essay can claim is that Ferrante has described a mechanism vividly enough that a reader may recognise it, and recognition is not data.

A fourth: this Library is reading a translation, and has said so, but saying so does not undo it. The obscenity argument in particular rests on register, and register is the first thing that moves in translation. Goldstein is an unusually careful translator and the point still stands only as far as the English carries it.

And a fifth, against the book rather than the essay. The novel gives Olga a neighbour, an orchestra and an ending. Many people in her position have none of those, and the tidiness of the last movement is the one place where the book consoles. A reader for whom nothing resumed may find that final section not honest but lucky.

Nothing here is a diagnosis of Olga. No clinical term has been applied to her, and none should be; the essay has deliberately described what stops happening rather than naming a condition, because the novel does the same and because Olga is a character in a book and cannot be assessed.

Nothing here is a diagnosis of a reader either. A person who has organised her life around a shared household has not thereby lost anything, and a person who feels rusty at being alone is not exhibiting a deficit. The argument is about what long arrangements can do, not about what any particular one has done.

Nothing here is a claim about women in general, or about dependence, or about the wisdom of marrying. The essay does not hold that Olga made a mistake, and it does not hold that she did not.

Nothing here rests on the identity of the author. Ferrante has published under that name since 1992 and has declined to be a public person, and this Library takes that at face value and has no view on who she is; the only use made of her pseudonymity is her own stated position that a finished book does not need its author.

The plot particulars above are taken from published records of the work and from reviews of record rather than from a fresh reading of the volume, and no passage of the novel is reproduced here beyond a few short phrases attributed to the Goldstein translation. The scholarly arguments are summarised from what their authors state; the essay does not claim their endorsement of anything argued here, and neither Elena Ferrante, her publishers, her translator nor any cited scholar has any connection to this house, knows of it, or endorses it.