Bluets, and the Container for What Cannot Be Said Directly

There are two facts in your life that you cannot look at. Not because you have decided not to, and not because you are in denial about either of them, which is a word people reach for too quickly. You know both facts completely. You could recite them. It is simply that when you turn your attention fully towards either one, something happens that is worse than not looking: the thing becomes a scene. You hear yourself narrating it. The grief acquires an audience, and the audience is you, and what was a fact about your life becomes a performance about your life, and the performance is unbearable in a way the fact was not.

So you do what people actually do. You develop an interest. It may be a small and slightly absurd one — a bird, a route, a piece of music played on repeat, the history of a particular kind of tile. You attend to it with a seriousness that is out of all proportion to the thing itself, and anyone watching would say you were distracting yourself. You are not. You are keeping your hands on something while the two facts move past.

In 2009 the publisher Wave Books issued a short book by Maggie Nelson called Bluets: two hundred and forty numbered propositions, each of them a sentence or a short paragraph, all of them concerned with the colour blue. It is usually described as a book about the colour blue, or as a book about heartbreak, or as an example of the lyric essay. It is a cult object now, taught widely, reissued in a tenth anniversary edition, adapted for the stage, and translated into a dozen languages, which is not bad for a hundred and twelve pages published by a poetry press.

This essay proposes something more specific, and it is not a summary. The blue is not the subject of Bluets. The blue is a container. The book has two subjects it cannot approach — a love affair that has ended, and a close friend left paralysed by an accident — and a devotion to a colour is the one thing the writer can keep holding while both of them are approached sideways. The numbered form does exactly the same work as the colour. A proposition can be set down and abandoned. A paragraph would have to arrive somewhere. What follows is an argument about what that arrangement makes possible, why it is not avoidance, and what is wrong with saying so.

Bluets was published by Wave Books, a poetry press, in 2009. It runs to a hundred and twelve pages. Published accounts of its composition place the writing roughly between 2003 and 2006, and the book was assembled rather than drafted straight through: the propositions were reordered a great many times before their sequence was fixed. A tenth anniversary hardcover appeared in September 2019.

Each of the two hundred and forty units is numbered. Nelson calls them propositions, which is a borrowed word and a deliberate one. They are short. They do not build an argument in the way numbered points in a report build an argument. Some of them follow directly from the one before. Some of them change the subject entirely and do not explain why. There are no chapters and there is no plot. The book has been filed as poetry, as memoir, as lyric essay and as autotheory, and the disagreement about which shelf it belongs on is a symptom of the thing this essay is about rather than a problem to be resolved.

The material inside is of three broad kinds, braided rather than sectioned. There is the blue itself: pigments, skies, birds, a brief history of a particular blue, notes on what other people have written and painted and thought about the colour. There is reading, a great deal of it, and the book is candid that most of what it contains is other people’s blue rather than its own. And there is a life, admitted in pieces: a man who has gone, and a woman in a hospital bed.

For the record, because it is often got wrong: the United Kingdom edition was published by Jonathan Cape on the first of June 2017. Fitzcarraldo Editions, which a reader might reasonably assume given its association with this kind of book and with other writers of the same generation, does not publish Bluets. The book has appeared in editions in Turkey, France, Germany, Sweden, Korea, Spain, Norway, Russia, Finland and Italy. It has not appeared in Japanese, a point this essay returns to because it bears on who is reading it and how.

The MacArthur Foundation, which named Nelson a Fellow in 2016, describes Bluets in its own citation as a philosophical meditation on the colour blue, the end of a relationship, and a friend’s grievous injury. That is a fair three-part description and it is worth noticing what it does. It puts the colour first. The two human facts are named second and third, as though they were additional topics rather than the reason the book exists.

The first of them is the ended affair. The book does not narrate it. There is no account of how the relationship began, what it consisted of, what was said at the end, or who decided. What there is instead is aftermath, distributed across the propositions in small charges: the condition of wanting somebody who is not there, the physical stupidity of the body that has not caught up, the particular humiliation of still wanting. A reader finishes the book without being able to say what happened, and that is not an oversight.

The second is harder, and it involves a person who is not the writer. In October 2003, Christina Crosby, a professor of English and of feminist, gender and sexuality studies at Wesleyan University, caught a branch in the spokes of her bicycle, was thrown to the road, and broke her neck. She was fifty. She was left quadriplegic and lived the remaining seventeen years of her life that way, returning to teaching part-time in 2005 and dying in January 2021. She had taught Nelson at Wesleyan in the 1990s; the two were friends. Bluets does not name her. Published criticism does, and Crosby herself published a full account of her injury and her life afterwards, A Body, Undone: Living On After Great Pain, with New York University Press in 2016, written using voice recognition software.

This matters for how the book should be read and for what can honestly be said about it. The injury is not Nelson’s. She was the friend at the bedside, not the person in the bed. Nelson had already written directly about those weeks — the 2007 poetry collection Something Bright, Then Holes contains a long poem written at a friend’s hospital bedside — and she returned to Crosby again in The Argonauts in 2015. Bluets does something different from either. It circles. The friend appears in fragments, at intervals, never as a case and never as a chapter, and the book keeps returning to the colour between the appearances the way a person returns to an ordinary task between visits.

Here is the claim, stated flatly so that it can be argued with.

A book about the colour blue is not what Bluets is. A colour cannot sustain two hundred and forty propositions on its own merits; nothing about blue is that interesting, and the book half admits this, since much of its blue is quoted, borrowed, looked up. What the colour supplies is not content. It supplies a place for attention to rest that will not collapse. The lost man will not hold attention, because attention directed at him produces either longing or self-reproach and both of them close down the page. The friend’s injury will not hold attention either, because attention directed at it produces something worse: the spectacle of a healthy woman contemplating a catastrophe that happened to somebody else. Blue holds. Blue can be looked at for a long time and gives nothing back, and that is precisely its usefulness.

Once the colour is understood as a container rather than a topic, several features of the book stop looking like eccentricities. The heavy reliance on other people’s writing about blue is not padding; it is the construction of a structure large and impersonal enough to stand in a room with two unbearable things. The refusal to narrate either event is not coyness. The tonal flatness in the most painful propositions — the way a sentence about the friend’s body arrives with no more emphasis than a sentence about a pigment — is not coldness. It is the container doing its job, which is to keep everything at one temperature so that nothing in it has to be approached at full strength.

And this explains the book’s real effect, which readers report and reviewers often mislabel as beauty. What Bluets produces is not consolation. It is the experience of standing next to two catastrophes without being asked to face either of them, for a hundred and twelve pages, in the company of somebody doing the same thing. That is an unusual offer. Almost nothing else in the literature of grief makes it.

The numbered form is doing the same work as the colour, and it is worth being precise about the mechanism, because the mechanism is the transferable part.

A paragraph in ordinary prose carries an obligation. It has a first sentence that opens something and a last sentence that has to deliver on it, and the reader is entitled to expect that between them something was resolved, complicated or advanced. A paragraph that opens a difficult subject and then simply stops reads as a failure of nerve. This is why writing about grief in continuous prose exerts such pressure towards conclusion, and why so much of it ends in a lesson that nobody in the situation actually learned.

A numbered proposition carries no such obligation. It can state one thing and stop. The number is a permission slip: it says that this unit is complete at whatever length it happens to be, that the next unit owes it nothing, and that a reader who wants continuity will have to supply it herself. A proposition can be set down and abandoned. This is why the form can hold material that a paragraph would deform — a thought about the friend that goes nowhere, an observation about the man that does not justify itself, a sentence that is only a sentence and is not going to become an insight.

Nelson has been open about where the form came from, and her account is more specific than the received version. Speaking to Susie DeFord in BOMB Magazine in October 2009, she said she was so taken with the form, tone and locutions of Wittgenstein’s Philosophical Investigations that it took her a while to realise that the form for Bluets was in front of her, and that she took roughly half of the book’s locutions from Wittgenstein and half from Goethe: the anecdotal manner of talking about perception from the Theory of Colours, and from the Philosophical Investigations what she called the relentless form of self-questioning, of pulling the rug out from under the speaker each time she has come to perch.

That last phrase is the whole technique. The Philosophical Investigations is a book of numbered remarks that repeatedly dismantles its own positions; it is not the Tractatus, whose numbered propositions build a system, and the difference is the point. Nelson took the form that unsettles rather than the form that concludes. Goethe supplies the other half: a way of writing about a colour as something that happens to a person looking, rather than as a fact about light. Between them the two sources give her a container that can hold observations without ranking them and can abandon a line of thought without that abandonment reading as defeat. The scholar Georgia Walton, writing in English: Journal of the English Association in 2023, argues that the numbered propositions work as sequential lenses through which the book’s narrator sees, and traces the device further back, to Emerson’s image of life as a train of moods like a string of beads. Walton’s reading and this one are compatible. A lens and a container are the same object seen from two sides.

The reason any of this is worth a reader’s time has nothing to do with poetry. It is a claim about grief, and it is this essay’s actual argument.

The instruction our culture gives about difficult feeling is uniform and it is always some version of the same sentence: face it, name it, say it out loud, do not avoid. The instruction is not wrong. For a great many situations it is exactly right, and this Library has argued elsewhere for naming things plainly. But it contains an assumption that does not always hold, which is that the act of addressing a thing directly leaves the thing unchanged.

For certain griefs it does not. There is a class of loss that alters when it is looked at, because looking at it requires an account, an account requires a shape, and a shape requires the person giving it to occupy a position — the bereaved, the abandoned, the one who was there. The moment the position is occupied, the feeling acquires a genre, and the genre brings its own vocabulary, its own cadence, its own expectations about what should be felt and in what order. What comes out is not the grief. It is a competent rendition of grief, and the person producing it knows this while producing it, which adds self-disgust to everything already present.

The friend’s injury in Bluets is the clearest case. Consider what would happen if the book addressed it head-on. A well-woman writing at length about a friend’s paralysis has almost no available registers. She can be devastated, which makes her the story. She can be admiring, which makes the injured woman a lesson. She can be angry at the universe, which is a genre. She can be scrupulous about not centring herself, which centres her more thoroughly than any of the others, since the scrupulousness is now the visible performance. Every direct route converts the friend into material and the writer into a figure. The sideways route does not solve this — a later section takes the objection seriously — but it does something the direct routes cannot: it lets the fact stay in the book as a fact rather than becoming an occasion.

So the second grief in Bluets, the ended affair, is handled the same way for a different reason. There the risk is not indecency but cliché. Nothing is more thoroughly scripted than the account of a woman still wanting a man who has left. Approached directly, the material arrives pre-interpreted; the reader knows the shape before the sentence finishes. Distributed across two hundred and forty propositions about something else, it arrives without its script, in pieces the reader has to assemble, and it lands.

Put the two mechanisms together and the claim can be stated in its general form. An oblique container is not an evasion of a subject. For certain subjects it is the condition under which they become sayable at all.

This is a claim about conditions rather than about virtue, and the distinction matters. It is not that indirection is more tasteful, or braver, or more sophisticated than plain speech. It is that a given thing has a threshold, and below the threshold the thing can be said and above it the thing cannot, and what determines which side of the line you are on is not how honest you are but what structure you have available to say it in. Change the structure and the same person, with the same honesty, can say more.

The structures that work all share a feature: they carry the reader’s attention while the difficult material passes. A colour will do it. So will a route walked every day, a list, a taxonomy, a catalogue of somebody else’s possessions, an account of a practical task performed repeatedly. The container has to be genuinely absorbing or it will not hold, which is why the devotion in Bluets to blue is not ironic and cannot be. If the writer did not actually care about the colour the whole arrangement would collapse into a device, and readers can tell.

There is also a cost, and this essay does not want it hidden inside the argument. A container that permits a thing to be said also limits what can be said about it. The reader of Bluets never learns what happened. She never learns what the friendship was like before, or what the friend thought, or what the affair was. She gets the fact of two catastrophes and the temperature of a mind moving near them, and nothing else. That is a real loss of information, and somebody who needed the account rather than the temperature should read a different book — including, in this case, the injured woman’s own.

Two earlier pieces of work in this Library stand directly next to this one, and the argument here is not usable until it has said where it differs from them.

This Library has read Lisa Taddeo’s Three Women as a book about audience rather than about desire: the decisive variable in a life organised around a want is not the wanting but who was watching, who acquired the right to narrate it afterwards, and what a private want becomes at the moment somebody else is holding the account. That essay located the only available leverage in the choice of who witnesses.

This Library has also read Annie Ernaux’s Simple Passion as a book that removes shame from the register rather than from the content: the material stays maximally shameable, and what changes is the grammar available to the account, which is an inventory and therefore offers a judging reader no surface to put a verdict down on. That essay located the leverage in the form the account takes — a lever that remains available to a person who cannot choose her audience at all.

This essay parts from both, and the difference is the reason it was worth writing. Bluets is not concerned with who is watching; it was published under Nelson’s name to a general readership and has been taught, adapted and translated for fifteen years, which is the least controlled audience imaginable. Nor is its problem the register in which a thing is said. The register of Bluets is not flat in Ernaux’s way; it is warm, funny in places, unembarrassed about longing. Its problem is anterior to both. It is whether there exists any position from which the thing can be addressed without the addressing ruining it. The answer the book gives is that there is no such position, and that what is required instead is a container — a third object, genuinely interesting in itself, that the attention can hold while the material goes past it.

So the three arguments locate the leverage in three different places, and a reader who has only one of them is missing the other two. Audience is a scarce lever, available occasionally. Register is available always, to anyone. A container is available always as well, but it has a precondition the other two do not: you have to actually care about something else. That is not a technique. It is closer to a piece of luck, and the fact that it can be arranged for on purpose is the only practical thing this essay has to offer.

One further note, so that it is not left for a reader to find. This Library has already published a reading of Nelson’s The Argonauts, as a book about relationships that outgrow the labels available to them, with the koseki and the unregistered partnership as its Japanese application. Nothing here repeats it. That essay is about naming; this one is about the impossibility of approaching directly, which is a different problem and, in Nelson’s own chronology, an earlier one.

A reader in Japan meets Bluets from an unusual angle, and two facts about that angle are worth stating plainly.

The first is that the form is not foreign. Roughly halfway through the book Nelson turns to Sei Shōnagon, whose Makura no Sōshi was completed around the year 1002 and consists of lists, notes, complaints, catalogues and observations that connect to one another loosely or not at all. The zuihitsu is not a translated device here; it is a thousand-year-old native one. A Japanese reader encountering two hundred and forty numbered fragments that circle a subject without arriving at it is on ground that Anglophone readers describe as experimental and that the Japanese tradition has regarded as one of the ordinary shapes a book can take. What Nelson adds to the inheritance is the numbering, which comes from philosophy, and the single obsessional object, which comes from Goethe. The discontinuity itself she did not have to invent, and she names her debt.

The second fact is that the book has not arrived. There is no Japanese translation of Bluets, and this essay states what was searched rather than asserting a negative. A search of the National Diet Library catalogue for マギー・ネルソン returns three records, all of them relating to a single 2016 article by Nishiyama Atsuko in the magazine Subaru about The Argonauts; there is no monograph. A search of the same catalogue for ブルーツ and for アルゴノーツ returns no translation of either book. Wave Books lists eleven foreign editions of Bluets, including a Korean edition from 2017, and Japan is not among them. As of the date of this essay, a reader in Japan who wants Bluets reads it in English, or in a language that is not her first, or not at all.

That absence is not a scandal and it is not evidence of anything about Japanese readers. Translation rights are commercial decisions made by particular editors about particular lists, and a short book of numbered propositions from a poetry press is a difficult commercial proposition anywhere. But the absence does have a consequence worth naming. It means that Nelson’s ideas reach Japanese readers mainly secondhand — through essays like this one, through university coursework, through the slow migration of an idea across a language border — and secondhand transmission tends to carry the argument and lose the texture. The argument of Bluets is not separable from its texture. A summary of Bluets, including this one, is a description of a container rather than an experience of being inside it.

Four objections have real force and none of them is disposed of by the argument above. They are given here at their strongest rather than in a weakened form that is easy to answer.

The first is that an oblique container is a way of keeping a reader at a flattering distance. A writer who never narrates the affair never has to show herself behaving badly in it. A writer who approaches a catastrophe sideways is never caught in an unbecoming posture near it. The fragments can be individually beautiful precisely because no single one of them has to carry the weight of an actual account, and the cumulative effect — a mind of great delicacy moving near great pain — is an extremely attractive self-portrait that has been produced without a single claim that could be checked. Obliquity, on this reading, is not a solution to the performance problem. It is a more sophisticated performance, one that has learned that the surest way to be admired is to appear to be withholding.

The second concerns the friend, and it is the serious one. A woman who was not injured has made literature out of an injury that happened to somebody else. Whatever care is taken, the arrangement is asymmetrical: one person is in the bed and the other has the book. The sideways treatment does not dissolve this and arguably sharpens it, since a fragment can take what it needs from the other person’s catastrophe and leave before any obligation attaches. The book knows this; the discomfort is legible in it. Knowing is not resolving. The most that can be said in mitigation is a fact rather than a defence: Christina Crosby was not mute. She was a professor who wrote and published her own full-length account of her injury, on her own terms, with a university press, and a reader who feels the asymmetry should go and read it, because it is the correction that actually exists.

The third is about the form and the fifteen years since. Numbered fragments are now a house style. The mode has been repeated so widely across memoir and the essay that the white space between units has become a signal of seriousness rather than a consequence of anything, and readers have learned to supply profundity to a short paragraph that has been given a number and a gap. The novelist and critic Lauren Oyler has pressed this charge hard — that the fragmented mode is trendy, melodramatic, vague, plagiaristic and inflated — and Louisa Hall, writing in the Los Angeles Review of Books in 2021, answered that these are objections to how a form is used rather than defects in the form itself. Hall is right and the charge still lands on Bluets, because Bluets is one of the books that made the mode available. A form cannot be blamed for its imitators, but a book that was influential is in a different position from a book that was not, and some part of the reverence Bluets now attracts is reverence for a style rather than for this particular use of it. It is worth noticing what happened when the book was adapted for the stage: the version by Margaret Perry, directed by Katie Mitchell at the Royal Court in London between May and June 2024, divided the material among three voices and, in the account of at least one reviewer, became a single oblique narrative rather than a sequence of numbered parts. The numbering, in other words, was the first thing to go.

The fourth objection is aimed at this essay rather than at the book, and it is the hardest. The claim that some things can only be said obliquely is unfalsifiable as stated. There is no experiment. Nobody can produce the direct version of Bluets that Nelson did not write and demonstrate that it fails. Any direct treatment of the same material that a critic might point to can be dismissed as a treatment by a lesser writer, and any successful direct treatment can be dismissed as not really the same material. An argument that cannot be wrong is not a finding; it is a description dressed as one, and a reader is entitled to be suspicious of a publication that offers her an elegant explanation for why a beloved book is right to do what it already does.

That objection is conceded. It is not answered, and the essay would be dishonest to pretend otherwise. What can be offered is a narrowing. The strong claim — that these things could only have been said this way — is indeed unfalsifiable and is withdrawn. The weaker claim is not. It is that direct address to a grief imposes specific and predictable costs, which are namable in advance: the requirement of an account, the occupation of a position, the arrival of a genre, and the self-consciousness that follows. Those costs are observable in the enormous quantity of grief writing that ends in a lesson. An oblique container avoids those particular costs and incurs different ones, which this essay has also named. That is a claim about trade-offs, and trade-offs can be argued with, which is the most that should be claimed for it.

This house sells private companionship, and that interest runs through the argument above closely enough that a reader is entitled to see it stated rather than to work it out.

An argument that certain things become sayable only inside a container is flattering to an arrangement that is itself a container. An evening here has a beginning and an end, a price, a form and a set of terms, and people do in fact say things inside that structure that they have not said elsewhere. It would be very convenient for this house if the reader concluded from an essay about a book that a bounded arrangement is where the unsayable becomes sayable. The convenience is real. It is not cancelled by the argument also being, as far as this Library can tell, defensible.

So the constraints, stated as restrictions on what may be claimed rather than as a disclaimer after the fact. This house diagnoses nobody. Nothing above says that a reader who cannot speak about a loss is avoiding it, is unwell, or has a problem that a paid evening addresses. This house claims no outcome: not relief, not resolution, not the ability to say a thing afterwards that could not be said before. Nothing here is grief counselling, and a person carrying a bereavement or caring for somebody badly injured should be talking to people qualified to help with that, which this house is not.

And this house claims no connection to the book or to anybody in it. Maggie Nelson, Wave Books, Jonathan Cape, the estate and family of Christina Crosby, New York University Press, and every critic and scholar named here have no connection to this house, no knowledge of it, and no part in anything it says. Naming a book is discovery and nothing else. There is no arrangement of any kind behind the recommendation and no link from it to anything that can be bought here.

This essay is a reading. It is not clinical guidance, it describes no treatment, it promises no outcome, and it diagnoses nobody. Nothing above should be taken as advice about how any particular person should handle a bereavement, a caregiving situation, or the end of a relationship.

The central claim — that an oblique container is the condition under which certain griefs become sayable — is this Library’s interpretation of a book, not a finding about how people grieve. Its strong form is unfalsifiable, the essay says so in the body rather than only here, and the strong form is withdrawn in favour of a weaker claim about costs and trade-offs. No proportion, prevalence or frequency is asserted for any of it.

Christina Crosby was a real person who died in 2021. What is stated about her here comes from the public record: her position at Wesleyan University, the date and cause of her injury, her quadriplegia, her return to teaching, her death, and her own book, A Body, Undone, published by New York University Press in 2016. Bluets does not name her; published criticism identifies her, and this essay follows that criticism rather than any private knowledge. Nothing is asserted about her inner life, her view of Nelson’s books, or what passed between the two women. Where this essay describes the effect of a fragment about the friend, it is describing the book’s rendering and not her experience.

Nelson’s account of her own sources is taken from her published remarks, specifically an interview with Susie DeFord in BOMB Magazine in October 2009, in which she names Wittgenstein’s Philosophical Investigations and Goethe’s Theory of Colours. Other published descriptions of the book attribute the numbered form to Wittgenstein’s Tractatus or connect it to his Remarks on Colour; this essay follows the writer’s own statement and notes the disagreement rather than resolving it.

Publication details are taken from publisher and library records rather than from inspection of physical volumes. The United Kingdom edition is Jonathan Cape, 2017; Fitzcarraldo Editions does not publish this book, and any impression to the contrary is corrected here. No passage of Bluets is quoted anywhere in this essay, and no reading is offered of any proposition in isolation.

On the Japanese translation: this essay asserts that none exists on the basis of searches of the National Diet Library catalogue under マギー・ネルソン, ブルーツ and アルゴノーツ, and of the publisher’s own list of foreign editions, all carried out on the date of writing. A catalogue search can miss a record, a forthcoming edition would not appear, and the absence is stated as what those searches returned rather than as a certainty. Nothing here is a statement about Japanese readers, about Japanese publishing, or about national taste of any kind.