The Lover and the Way Memory Rewrites Desire

A woman approaching seventy writes about a girl of fifteen and a half, crossing a river in colonial Indochina in a man’s hat and gold shoes, who is approached by a wealthy Chinese man considerably older than her and enters a relationship with him that lasts about eighteen months.

That has to be said at the front, without the book’s own lyricism attached. The narrator is fifteen. The man is an adult. This Library does not treat that as a romance, does not read it as material about desire without naming what it is, and has nothing to offer in its defence. The fact that the account is written by the person it happened to, decades later, and in extraordinary prose, does not convert the arrangement into something else.

The month’s theme is the courage to ask, and the sentence underneath it is: I would rather say what I want and let someone choose freely.

The reason the book is in this reading list is that its actual subject is separable from its material and is the thing the atlas title points at. This is not a memoir of an affair. It is a demonstration of how a remembered want gets built — and the building, not the affair, is what the book is doing.

What is unusual about this book is not its frankness. It is that it refuses to behave like a memory and insists on behaving like a construction.

It opens with a face — the narrator being told, in a lobby, that she is ruined, and then a long passage about her own face at eighteen and at seventy, as a thing that happened rather than a thing she had. Before any event is narrated, the book has established that the subject under examination is an image.

Then it keeps breaking its own surface. The narrator shifts between I and she, sometimes in the same paragraph, so that the girl is alternately the writer and a figure the writer is looking at. Scenes are introduced with statements that she does not know: not sure of the year, not sure whether it was before or after, not sure what she felt.

And the central image of the whole book — the girl on the ferry in the hat — is announced as a photograph that was never taken. Nobody recorded it. It does not exist and could not be produced. The most fixed and most repeated picture in the book is explicitly an invention of the woman writing it.

None of that is carelessness and none of it is modesty. It is the method. Duras is not recovering a past; she is showing you a seventy-year-old assembling one, with the joins left visible, and the visibility is the book’s whole claim.

Follow that method to its consequence and the atlas title becomes precise.

The girl in this book does not know what she wants. She is fifteen, her family is in a state of financial and emotional disorder, her mother is unreachable, one brother terrifies her and another is dying, and she moves toward the man for reasons that are not separable from money, escape, spite, curiosity and the desire to be removed from where she is. The account gives no evidence that she had a formed desire she could have stated.

The woman writing is the one with the desire. Over fifty years the event has been handled, turned, rewritten and composed into something that has a shape — and that shape is what the book presents. Not a record of a want. A want that has been made, afterwards, out of an event that did not contain one.

Which is the claim worth taking away, and it is not a claim about unreliability. It is a claim about what desire is. A remembered want is authored. It is not stored and retrieved; it is produced, in the act of telling, by a person who has since acquired the vocabulary, the distance and the motive to produce it.

And the reason this is not merely literary is that almost nobody experiences it that way. A person’s own account of what she has wanted feels like retrieval. It feels like looking at a record. The book’s service is to make the authoring visible in a case where it cannot be denied — because the central image is admitted to be a photograph that does not exist.

Now the month, and this is the essay’s actual contribution.

The courage to ask presumes something that is rarely examined: that there is a want available to be stated. The whole discourse around asking treats the want as given and the saying as the difficulty — find the nerve, use the words, accept the risk.

This book suggests the prior problem. A want that has not been composed cannot be said, not because of fear but because there is no sentence. What exists instead is a condition: a pressure, a restlessness, several incompatible impulses, an attraction with no clear object. The girl on the ferry is in exactly that state and it is not a character flaw. It is what wanting looks like before anybody has authored it.

So the courage to ask has a precondition nobody mentions. You need a version of the want that holds still long enough to be put into a sentence — and that version is made rather than found. It is produced by telling, repeatedly, to somebody or to a page, until the thing acquires edges.

Which reverses the usual sequence. The standard advice is: work out what you want, then find the courage to say it. What this book suggests is that the saying is how the working-out happens, and that waiting to be sure before speaking is waiting for a state that speaking is the method of reaching.

And it explains a failure this chamber has described from several directions — the woman who cannot say what she wants and concludes she must not want anything. She is not empty. She is unauthored, which is a completely different condition and has a completely different remedy.

That argument has an obvious and dangerous other half, and Duras is honest about it, which obliges this essay to be.

If a remembered want is authored, then the faculty that composes a sayable want is the same faculty that can manufacture one that was never there. The book does not resolve which it has done. It offers competing versions of scenes, lets the reader see the narrator choosing, and in places suggests that the adult has supplied the girl with feelings the girl did not have.

Which matters practically, because the retrospective construction of desire is not always benign. A person can compose an account in which she wanted something that was done to her, and the composition can be sincere, and it can be a way of making an event survivable by converting it into a choice. Duras’s material is exactly the kind where that is a live possibility, and the book’s refusal to settle it is either its greatest honesty or its greatest evasion depending on the reader.

This Library’s position is the narrow one and it is stated so the previous section cannot be misread. Authoring a want is a legitimate and necessary operation when the thing being given edges is yours. It is something else when the function of the authoring is to retroactively convert something you did not choose into something you did.

The test is not introspective comfort, because the manufactured version is frequently the more comfortable one. A more useful check is whether the account is correctable — whether you can say it to somebody and let them find it implausible without the whole structure needing to be defended. A composed want holds up under contradiction. A manufactured one usually cannot be examined at all.

It is worth warning a reader about the style, because the book is short and is frequently recommended as though it were easy.

The sentences are fragmentary, the tenses move without signalling, the chronology loops, and paragraphs stop where an explanation would ordinarily go. Events are narrated twice in different forms. The word that recurs most is a negation — not knowing, not remembering, no longer able to say.

And the style is the argument rather than a manner. A book claiming that memory composes rather than records cannot be written in orderly retrospect without contradicting itself, because orderly retrospect is precisely the illusion it is refusing. So the prose withholds the coherence a memoir normally supplies, and the reader experiences construction instead of being told about it.

Which also means the book resists the use most commonly made of it. It is difficult to extract a love story from it, because the parts that would make a love story — motive, sequence, interiority, a declaration — are the parts the method removes.

The translation matters here more than usual. This is prose whose effects live in its rhythm and its refusals, and which English edition a reader holds will change the experience considerably.

The book is read as a French story that happens to take place in Indochina, and that reading removes the thing that makes the arrangement legible.

The narrator’s family are poor whites in a colony — holding the status of the colonising population and none of its money. That combination is the engine of the entire situation. Her whiteness is her only asset and it is an asset precisely because of the colonial order; the man’s money is substantial and his position is, in that order, beneath hers despite it. The relationship is shaped at every point by a hierarchy in which each party has something the other’s position denies them.

Which means this is not a story about two individuals whose circumstances happen to differ. The attraction, the secrecy, the impossibility of the relationship being acknowledged, and the family’s simultaneous profit from it and contempt for it are all functions of the colonial arrangement rather than of anybody’s psychology.

And the book is substantially less interested in interrogating that arrangement than a contemporary reader will want it to be. The Chinese man is written as an object of the narrator’s composition rather than as a person with an account of his own; he is beautiful, wealthy, weak, and almost entirely silent on his own behalf. Whatever he wanted is not in this book.

That is a real limitation and it should be named rather than absorbed. A book about how memory authors a want has authored one party completely and left the other as material.

I would rather say what I want and let someone choose freely.

Three things, and the first is permission rather than instruction.

Stop waiting to be sure. If you cannot currently state what you want, the most likely explanation is not that you want nothing and not that you are repressed. It is that the want has not been authored yet — and authoring happens by telling, out loud, in provisional and contradictory versions, until something acquires edges. Nobody arrives at a clear want by introspecting harder.

Second, tell it somewhere it can be revised. A version stated once to somebody who will remember it exactly is worse than useless at this stage, because you will then be held to a draft. What the process needs is a place where the account can be said badly, corrected, and said differently next time without anybody keeping score.

Third, apply the honesty test from this book. A want you have composed should survive being doubted. If you can say it and let somebody find it strange or implausible, and the thing holds, it is yours. If the account has to be defended completely or not examined at all, that is worth knowing before acting on it.

And the thing the month’s sentence gets exactly right, which this book shows by failing to have it: the second clause. A want that cannot be refused has not been stated. The girl on the ferry was never in a position to be refused, which is the reason nothing she did there was asking.

The Lover was published in French in 1984 by Marguerite Duras, in English translation by Barbara Bray. Edition, translation and current availability should be verified at the publication gate, per the atlas entry’s note on external works, and which translation a reader holds materially affects the prose.

The narrator of the book is fifteen and a half and the man is an adult. This Library states that plainly, does not treat the relationship as a romance, does not aestheticise it, and offers nothing in its defence. The essay engages the book for its account of retrospective composition, which is separable from its material, and a reader who would prefer not to engage with the work at all has a good reason.

The book is described in outline and no passage is quoted, including its opening and the much-discussed image of the river crossing, which are characterised rather than reproduced.

The relationship between Duras and her narrator is not settled here. The book is autobiographical in origin and is not a document; this essay treats the narrator as a constructed voice rather than as the author reporting, which is the reading the book’s own method invites.

The claim that a remembered want is authored rather than retrieved is this Library’s formulation of what the book demonstrates. It is an argument about the book, not a finding about memory, and no claim is made about how memory functions physiologically.

The warning that the same faculty can manufacture a want that was not there is this Library’s addition, included so the preceding argument cannot be used to validate any retrospective account whatever. It is not a clinical claim and it is not a statement about anybody’s history; this house treats nothing, and a reader for whom this is not abstract should be asking a professional rather than a Library.

And the account of the colonial arrangement is given in outline. The observation that the book authors one party and leaves the other as material is this essay’s criticism rather than a consensus reading.

What stays from this book is the image at its centre, and what stays about the image is that it does not exist.

A girl on a ferry, in a man’s hat, in worn gold shoes, on a river in the heat. It is the most reproduced picture in late twentieth-century French literature and nobody ever photographed it. Duras says so. There was no camera, no one present who would have thought to record it, and the reason she returns to it so often is that nothing fixes it — it has no original, so it can be made again each time.

That is the book’s entire proposition in one object. The most durable thing in a life may be something that was never documented, composed afterwards by somebody who needed it to exist, and none the less true for having been made rather than found.

And it says something useful about the month, which is otherwise a month about nerve. What you want is not a fact waiting to be reported. It is a photograph that was never taken, and it comes into existence by being described — badly at first, differently each time, to somebody, out loud.

Which is what an evening at Moonlight is arranged to be, stated as narrowly as this book requires. Not a place to arrive with a finished account of yourself: a stated number of hours with a beginning and an end, terms set in advance and set by you, where a want can be said in a provisional version, corrected, and said differently — where nobody is keeping a record to hold you to, nothing is inferred from your silence, and being wrong about yourself out loud costs nothing. The sentence does not have to be right. It has to be said for the first time somewhere.