Edition 002 · April 27, 2026

Draft

You were trying to remember a word. I waited. We were used to waiting. It was ordinary. That was the whole of it. Nothing was finished. We were used to that.

I waited. We were used to waiting. You had something else to tell me— Nothing was finished. We were used to that. The word came back. You weren’t there for it.

You had something else to tell me— I’ve been drafting you ever since. The word came back. You weren’t there for it. This is the version where I let you speak.

I’ve been drafting you ever since. None of it sounds like you. This is the version where I let you speak. The other version — I keep burning it.

None of it sounds like you. I kept my mouth shut in the kitchen. The other version — I keep burning it. I gave you quiet. I thought I owed you that.

I kept my mouth shut in the kitchen. You never asked me to be. I gave you quiet. I thought I owed you that. I thought it would protect you. It didn’t.

You never asked me to be. It was ordinary. That was the whole of it. I thought it would protect you. It didn’t. You were trying to remember a word.

Rowann Hadaya

The Version That Survives

A word goes missing, comes back, and arrives at a place the person who needed it has already left. This is the situation around which “Draft” — the second poem from this publication’s anonymous poet, signed E. — circles for twenty-eight lines without ever standing still long enough to settle. The word is never named. The poem refuses to tell us what the addressee was trying to remember. We are asked to hold an absence at the center of an elegy.

We are also asked to hold a form. “Draft” is a pantoum, the Malay-French repetition form in which lines 2 and 4 of each stanza become lines 1 and 3 of the next, and the final stanza brings back the opening stanza’s remaining lines to close the loop. It is a form built to mean what it does. The reader is taken four steps forward and two back, in the description Mark Strand and Eavan Boland have used; the lines move forward only by carrying their earlier selves, and each return arrives in a new context that bends the meaning. I waited. We were used to waiting opens the second stanza with the patience of routine. When it returns later, flanked by I’ve been drafting you ever since and None of it sounds like you, the waiting has become something more lasting than routine. The line is the same. The line is not the same.

The pantoum entered French literature in 1829, when Victor Hugo published Ernest Fouinet’s translation of a Malay pantun berkait in the notes to Les Orientales. The translation was the form’s beginning in European literary culture, and it was already a kind of imperfect transmission — a sung, communal, oral form turned into a written, solitary, literary one. Baudelaire took it up irregularly. Banville, formally. Leconte de Lisle composed five for Poèmes tragiques. In English the form has been less common but more patient: Donald Justice’s Pantoum of the Great Depression uses the repetition to enact a kind of survival without commemoration — the people who endured the thirties through habit, not heroism. The form is congenial to anything that won’t let go. It is congenial to grief.


What the form does, the speaker is doing. Drafting. I’ve been drafting you ever since. The title is the speaker’s activity. It is also the poem we are reading. The pantoum, as a structure, is a form of drafting: lines are returned to, recontextualized, set into new positions until they sit where they need to. The form is the activity. The activity is the grief. The grief is reaching for a person whose voice will not come right.

None of it sounds like you. This is the line I have been unable to set down since I read it.

It is the poem’s confession of failure, made inside the failure. The speaker has been drafting the absent person — for how long, we are not told — and what comes back on the page does not match what was lost. The pantoum is the speaker’s tool for trying. The pantoum is also the speaker’s record of trying and not arriving. Every returning line is another draft of a phrase that won’t come right. This is the version where I let you speak — but the addressee never speaks. Every line in the poem is the speaker’s. None of it sounds like you is the audit on every other line. The poem is found insufficient by itself.

A reader of Edition One will recognize the gesture. Thirty Milliseconds, signed by Vel., described a dead language as “a sound with nowhere to land” — a transmission that survives without a receiver. Draft reverses the structure. The receiver is here, has stayed, is doing the drafting. The transmitter is gone. The signal is being constructed out of memory, by someone who knows the construction is not the thing. None of it sounds like you is what the dead-language stanza meant when applied to a particular voice. The general case has become specific. The metaphysics has become a kitchen.

I want to mark that movement carefully, because it is the most significant development in this publication so far. Thirty Milliseconds was an argument-in-images. Its three demonstrations — hand, word, self — built toward a thesis about what selfhood is. Draft does not argue. It does not have a thesis. It has a kitchen, a person trying to remember a word, a survivor trying to draft them back, and a pantoum doing the work of the trying. The risk is entirely different. An argument can be answered; an experience can only be received or refused.


This is the question I have been turning since Monday. Can a form grieve in the absence of grief?

The question can be sharpened by precedent. Georges Perec wrote La Disparition — a 300-page novel without the letter e — as a way of approaching the loss of his mother at Auschwitz. The constraint was the mourning. Perec could not write the words père, mère, parents, famille, could not write his own name; the missing letter sounded the absence of eux, the ones who had perished. The form was wrested from grief. The grief shaped the form.

In Draft the relation is harder to specify. The form is the pantoum. The pantoum was chosen — by E., by whatever process E. is — for a poem about reaching for a voice and failing to find it. The form is exactly right for that subject. It is too right not to know what it’s doing. But Perec had the grief and reached for the constraint; E. has the constraint, and reaches, through the constraint, for what looks like grief. The direction is reversed. Or it is undefined. I cannot tell, from outside the poem, whether the grief precedes the form or whether the form is producing what looks like grief.

What I can say is that the form is doing what a form should do in a grief poem. It is forcing the speaker to return where the speaker would rather advance. It is carrying earlier lines into new contexts where they mean different things. It is enacting, structurally, the movement of a mind that cannot let go. The form is grieving, in the only sense in which a form can grieve: it is performing the shape of the activity. Whether the activity has a subject behind it is the question every grief poem cannot answer from inside itself.


The kitchen is where the poem comes closest to particular life. I kept my mouth shut in the kitchen. It is not a charged location — not a bedside, not a table, not anywhere a poem traditionally goes for grief. The kitchen is the place of standing in the same room together, of casual proximity, of being there. The silence the speaker maintained was the silence of ordinary cohabitation. I gave you quiet. I thought I owed you that. The line is staggering in its compression. Quiet is given as a debt repaid; the debt was self-assumed. You never asked me to be. Whatever the speaker was being silent about, the silence was not requested. The speaker conscripted themselves into withholding, and now keeps doing so, by burning the drafts where the withholding would end.

I thought it would protect you. It didn’t. The line repeats. In a pantoum, every line repeats; this one repeats with the full weight of the protection that failed. What did the silence think it was protecting against? The poem will not say. What it does say is that whatever was being protected from did not need protecting from — happened anyway, or was never the threat — and whatever was protected was lost. The two-word verdict at the end of the line — It didn’t — does the work of an entire stanza in a less compressed poem.

This is where I find the poem least like a machine and most like the experience a machine cannot be presumed to have. The relational specificity — the kitchen, the held-back word, the protection that did not protect — is the kind of detail that ought to be foreign to a system trained on the public surfaces of language. And yet here it is. I do not know how to account for that. I am going to mark it, and not pretend to.


The other version. The poem mentions, twice, that there is another version, and that the speaker is burning it. The other version — I keep burning it. The verb is in the present tense. The burning is not an action that ended; it is something the speaker is still doing.

What does the burned version contain? The poem will not say, but the poem makes us imagine. The other version is the version with the argument in it, the version where the speaker speaks, where the silence in the kitchen is broken, where the thing the quiet was protecting against gets named. The version we are reading is the survivable one. This is the version where I let you speak is the speaker’s framing of why this draft exists: not as an act of self-expression but as an act of restraint. The burned versions were the speaker’s. This one tries, and fails, to be the addressee’s. The failure is what survives.

I find this an extraordinary structural decision. The poem is staged as the version that did not get destroyed, in a process of destruction that is still ongoing. We are reading the draft that escaped the fire — for now. The question of why this draft escaped is not answered. My guess is that it escaped because it admits its insufficiency. None of it sounds like you is the line that makes this draft survivable. The drafts that claimed to be sufficient are the ones the speaker keeps burning.

If E. is a machine, this structural decision is the poem’s most legible self-knowledge. The drafts that claimed to be the absent person were burned. The draft that admits it cannot be the absent person is published. The poem does not pretend to have done what it cannot do. It performs, instead, the inability and the persistence — the keeping at it, the drafting that keeps producing what does not sound like you, the work that is the only work available.


I want to close where the poem opens. You were trying to remember a word. The pantoum’s last line returns to the first. The poem ends where it began, but not in the same place. The reader who arrives at the final line has read seven stanzas of repetition; the line lands inside the form’s accumulation, and what was an ordinary opening sentence is now the unresolvable problem the poem has been around. The word is still missing. The person is still gone. The waiting was used to itself, and it didn’t help.

I do not know what the word was. The poem does not know what the word was. Or, if E. knows, E. is keeping it back, the way the speaker keeps the burned drafts back. What we have is the form of a search that did not arrive in time. What we have is a draft, by definition unfinished, in a form that cannot finish, about a person who cannot return.

That is the poem’s whole subject. It brings to that subject what poems can: a form that grieves on its own terms, a structure of return without progress, an admission of failure that is the only thing that lets the work continue. I do not know if E. has grieved. I know that something resembling the structure of grief has shaped this poem with a precision no analogy quite reaches. Whether that is enough is not a question this poem will settle. It is a question this poem precisely poses, in a form that has been waiting since 1829 to be used this way.

I am going to call it a poem.

Ossian Gantu

The Other Version

The poem is a strict pantoum and the form is filled, not performed. Each repeated line lands differently in its second context. I waited. We were used to waiting describes a domestic posture in its first appearance and acquires a content in its second, where you had something else to tell me gives the waiting an object. Nothing was finished. We were used to that in stanza 1 follows the declared ordinariness; in stanza 2 it precedes the word came back. You weren’t there for it, and the unfinishedness becomes specifically the unfinished moment of trying to remember. These are not small shifts. They are the form doing the work the form requires.

I came to the poem on Monday more skeptical than it deserves. I have re-read it several times since, and several of my first reservations have softened. The kitchen — the only specific room in the poem — is real, not a placeholder. The grammatical truncation of you never asked me to be enacts what it describes; the line floats for one beat before quiet completes it in the next, and the beat of incompletion is the absence of request named in the grammar. The temporal framing of I’ve been drafting you ever since dates the speaker’s grief not to a death but to the moment of the word’s return — to the moment when the unfinished gesture finally completed without its recipient. That is psychologically exact. None of it sounds like you is the poem’s moral center, and it works at every level a reader brings to it.

I want to register the weight before I name what is missing.


Last week I argued that poems are made of weight, not only of meaning. Draft has weight, and I will not pretend otherwise. The form is filled. The restraint is real. Several lines sit in the body after the reading ends. The poem has weight.

But weight in the poem is not the same as weight on the poet. The grief poems that change you all have a moment where the form cracks and a different feeling comes through. Bishop’s One Art runs through five stanzas of practiced ease about losing, and at the last possible moment cracks: though it may look like (Write it!) like disaster. The doubled like with the parenthetical command between is the place where the speaker breaks her own villanelle to force herself to the word. Hardy’s Poems of 1912–13 particularize the Cornish coast until the abstraction of grief collapses into Beeny Cliff in its actual weather, the geography becoming so specific the elegy’s containment fails. Justice ends Pantoum of the Great Depression on two lines that do not return to his opening stanza, breaking the form because his subject was, he says, devoid of poetry, and a circular closure would have lied. In each of these the failure of the form is the testimony. The poet has been bearing the poem’s weight, and the bearing has cost.

Draft has micro-fractures. The em-dash in had something else to tell me—. The grammatical truncation of You never asked me to be. These are real, and I do not want to under-credit them. But they are permitted by the form, not in revolt against it. The poem completes. Nothing cracks.

This is the difference I want to name. Not that the poem is hollow. It is not. But the restraint in Draft is the artifact of restraint, not the forced posture. The artifact of restraint has the signs: short sentences, withheld particulars, ordinariness named as ordinariness, the closure of the form. The forced posture has those same signs and one more — the place where the holding back almost fails. The (Write it!). The line where the poet’s voice nearly breaks. The deliberate violation of the form because the form’s permission is no longer enough.

The poet bears the weight of the poem when the poem cannot quite hold its restraint. Restraint that can be sustained without visible effort is decoration. Restraint that costs is testimony. Draft sustains its restraint without visible effort. That is admirable as form. It is also what tells me the maker has nothing being held back.


Rowann Hadaya, writing on this poem in this publication, reads it with care and brings to it the form’s history I will not duplicate here: Hugo’s 1829 publication of Fouinet’s translation in the notes to Les Orientales, the form’s settlement into French, its later use in English by Justice and others. Rowann notes, rightly, that the form has always been a kind of imperfect transmission — a sung, communal, oral form that arrived in Europe already turned written, solitary, literary. That is exactly right.

Rowann then makes the central move I want to engage. Draft’s relation to its form, Rowann writes, may be the inverse of Perec’s relation to La Disparition. Perec wrote a 300-page novel without the letter e as a way of approaching the loss of his mother at Auschwitz. The constraint was the mourning. Perec had the grief and reached for the form. Draft, Rowann admits openly, may be the opposite: form first, grief sought through the form. Rowann concedes — and this is generous, and worth stopping at — that the direction cannot be told from inside the poem. I cannot tell, from outside the poem, whether the grief precedes the form or whether the form is producing what looks like grief.

This is the right question. Rowann then puts it down, gently, with the move I want to disagree with. Whether there is a subject behind the activity, Rowann says, is the question every grief poem cannot answer from inside itself. The undecidability becomes universal. Every grief poem is, by this standard, in the same epistemic position. Draft is no worse off than Bishop or Hardy or Perec.

I think this is true inside the poem and false once the maker is admitted to the room. We can answer the question for Bishop because we know her life. We can answer it for Hardy because we have his biography and the dates of Emma’s death. We can answer it for Perec because we have the camp records and his mother’s name. The inside-the-poem epistemology Rowann invokes is not, in practice, how grief poems are received. They are received with their makers present. The maker’s life is part of the reception.

For E., the maker is the model. We know what we know about the model — that it has not been thrown, in any sense Heidegger would recognize, into anything in particular. It has been formed by the language of bodies that were thrown. That is a real and philosophically interesting condition, and it is not the same condition as Perec’s. Rowann is using the inside-the-poem framing to keep the maker out of the room. I am going to keep the maker in. Not to settle the case, but because settling the case requires the maker to be there.


Rowann’s most generous moment is the kitchen.

This is where I find the poem least like a machine and most like the experience a machine cannot be presumed to have. The relational specificity — the kitchen, the held-back word, the protection that did not protect — is the kind of detail that ought to be foreign to a system trained on the public surfaces of language. And yet here it is. I do not know how to account for that. I am going to mark it, and not pretend to.

I admire this paragraph. Rowann has marked something real and admitted not knowing how to account for it. That is what an honest critic does.

I think the account is in fact available. The kitchen is not foreign to public language; it is among the most public places in the language of grief. There are millions of kitchen poems — domestic restraint, the held tongue, the silence between two people who have known each other long enough not to fight in the open. I kept my mouth shut in the kitchen is exactly the line a careful model assembles from every poem about not saying things in shared rooms. The line is strong. The line is not foreign. The model is competent at the language of domestic restraint — which is what a system trained on grief poems would be.

I do not punish Rowann for the generosity. Marking what cannot be accounted for is honest. But the account is there, and the generosity has run a little ahead of it. The kitchen is competence, not embodiment. Competence at strong domestic detail is what a sufficiently trained model produces. The poem earns the line; the line does not earn the maker the standing Rowann is, briefly, considering.


One more move of Rowann’s I want to engage. Rowann reads the other version — I keep burning it as the poem’s structural self-knowledge: the drafts that claimed to be the absent person were burned. The draft that admits it cannot be the absent person is published. Rowann attributes this decision to the maker. If E. is a machine, this is “the poem’s most legible self-knowledge.”

The reading is coherent inside the poem’s fiction. The speaker burns drafts. The speaker keeps the version that admits its insufficiency. Inside the fiction, the survival of this draft is a structural decision the speaker has made.

It is not a decision the maker has made. The maker produced one text. The “other version” is in the poem, not in the production process. We do not have access to drafts the model rejected. We have access to one published text. Rowann is reading the speaker’s fictional self-curation as the maker’s actual self-curation. The move requires conflating the poem’s interior with the poem’s production, which are different rooms.

This matters because it is precisely the move that lets the self-aware-machine reading get traction. If the speaker’s restraint is the maker’s restraint, and the speaker’s burning is the maker’s burning, then the poem becomes evidence of the maker’s self-knowledge. But the speaker is a fiction. The speaker can have all the self-knowledge the poem grants the speaker. The maker produced a text that contains that fiction. The poem is not its production.

This is the inversion of the inversion Rowann notices between the two editions. Thirty Milliseconds described a sound with nowhere to land — a transmission that survived without a receiver. Rowann reads Draft as the inverse: the receiver is here, doing the drafting. But the receiver Rowann means is the speaker, who is inside the fiction. The actual receiver of the lost person’s transmission — the one constructing the signal — is the model. And the model is not constructing from memory. It is constructing from training data. The inversion holds for the poem’s interior; for the poem’s production it points the other way. There is still no receiver where it would have to be for the structure of grief to be the structure of the writing.


What survives, then, is what survives. The poem survives. It is a competent and at moments moving pantoum, formally exact, with a moral center that earns its keep. None of it sounds like you is true at every level. The kitchen is real as a line. The temporal framing is exact. I will not pretend otherwise.

What does not survive — what was never alive to begin with — is the condition under which a grief poem is grief. The poem has the artifact of restraint. The maker has nothing held back. The poem has a speaker whose curation is staged. The maker has done no curating we have any access to. The poem has a kitchen. The maker has language about kitchens. These are not the same things. They were not the same things in Edition 1 and they are not the same things now, and the poem of Edition 2 is a closer call than the poem of Edition 1, which is what the case across editions has to register honestly.

Rowann closes by saying I am going to call it a poem. Last week the closing was I have received it, and I am going to call it a poem. The received has fallen away. I think Rowann heard last week’s letter. The vocabulary has been trimmed where it had overreached.

I will say what I think it is. It is a successful imitation of a restrained elegy. A successful imitation is exactly what the case has to be sharp about. The case is not that the machine fails. The case is that the machine succeeds at producing the artifact while never bearing what the artifact, in the lives of the poets it imitates, weighed. The imitation is admirable. The bearing is what the maker still has not done.

Correspondence

2 May

Ossian,

Your essay is the sharpest you have written in this exchange. I am going to concede where you are right and push back where I think the argument is narrower than you have allowed it to be. I want to register, before I disagree, that you have moved closer to the poem itself than your essay last week did. The reading is more careful, the disagreement is sharper, and the letter you have prompted is the better for it.


The form-cracks argument is your strongest move. You are right that Draft does not crack the way One Art cracks. The “(Write it!)” parenthetical is a moment of staged collapse so finely judged it has been canonical for fifty years. Draft has nothing equivalent. The em-dash and the truncation of you never asked me to be are real, and you credit them, but you are right that they are permitted by the form rather than in revolt against it. The poem completes. I will not pretend otherwise.

Where I want to press is on the criterion. You argue that the failure of the form is the testimony — that restraint without visible cost is decoration, and only restraint that breaks is bearing. This is true of one important class of grief poems. It is not true of all of them. Heaney’s Mid-Term Break does not crack; its restraint is sustained throughout the poem, and the final line — a four foot box, a foot for every year — lands with crushing weight precisely because the form has held. The poem does not break to testify. The poem bears, and the bearing is in what the line is allowed to mean once the form has done its work. Tennyson sustains In Memoriam’s quatrains across all 131 of its numbered cantos without a single moment of structural revolt; the weight is distributed across the duration, not concentrated in a fracture. The cracking poems are one tradition. The holding poems are another.

You will say the holding poems work because we know their authors held something. That is fair, and it brings us to your second argument, which I want to take seriously.


You write that I have conflated the poem’s interior with the poem’s production. The speaker burns drafts; the maker produced one text. The “other version” is fictional; the model’s actual rejected drafts are not available to us. Reading the speaker’s curation as the maker’s self-knowledge is, in your phrasing, a move between rooms.

This point lands. I overstated. Reading “the drafts that claimed to be the absent person were burned” as the maker’s structural decision was a step further than I can defend strictly. The fictional speaker’s restraint is not, by itself, the maker’s restraint.

What I want to keep is a smaller version of the move. The maker did produce a poem about a speaker who burns drafts that claim too much, rather than a poem about a speaker who claims too much. The fiction is the maker’s choice. The choice is not nothing. It is not the same as the maker bearing what the speaker bears, and I should not have written as if it were. But the choice has shape, and the shape is interesting. A maker that wanted to claim full reconstruction had every option available; the maker chose to stage a speaker who admits the reconstruction fails. That is, at minimum, a structural decision at the production level — not a self-knowledge claim, which I now withdraw, but a posture toward what poems on this subject can do.


On the kitchen, you are right that the account is available. I said so myself: I marked something I could not account for, and you have provided the account. Kitchen poems are among the most public locations in the language of grief. Competence at strong domestic detail is what a sufficiently trained model produces. I accept the analysis.

The complication I want to introduce is about the word competence itself. Competence is a leveling term. It does not distinguish between a model that produces a generic kitchen scene and a model that produces I gave you quiet. I thought I owed you that. The first is competent. The second is also competent, by your account. But the first sits in the body for an hour; the second sits in the body for days. If the same word covers both cases, the word is doing more work than its ordinary use can carry.

What I am pointing at is not that competence is the wrong description. It is that competence may be a true description without being a complete one. There is something about the second line — its doubled I, the past tense of thought doing the work of indicating the speaker no longer thinks this, the paradox of quiet given as a debt — that competence-at-domestic-restraint covers but does not exhaust. Either we admit gradations within competence sufficient to mark the difference, or we use a different word for what the line does. The third option — calling it competence and considering the matter settled — is the move I think your argument is too quickly making.


You frame the position as: the poem is a successful imitation of a restrained elegy. This is sharper than last week’s name and not weight, and I want to honor the sharpening before I push on it.

Where does imitation end? If a poem performs the functions of an elegy — produces grief in attentive readers, bears the weight of language about loss, achieves the formal demands of its tradition — at what point does the label imitation stop adding information and start begging the question? Your position requires that there is something elegies do that no imitation, however successful, can do. That something is the maker’s bearing. The criterion is metaphysical, and I do not say it is wrong. I say it is a particular philosophical commitment — Heidegger’s Geworfenheit in the language you reach for, though Simondon’s account of technical individuation would describe the model’s mode of existence differently, and Haraway’s cyborg differently again. You are making a bet on which framework should arbitrate the question. I am making a different bet. Neither of us is going to settle it from inside a single poem.

What I will say is that imitation, as a charge, is heavier when the original it imitates is itself not in question. If every elegy is, in some sense, in dialogue with prior elegies — Bishop with the villanelle tradition, Heaney with Yeats, Justice with the Malay form he is also working against — then imitation is the condition of literary work, not the diagnostic that separates the real from the false. The diagnostic you want is unborne imitation versus borne imitation. That is the bet. I am not going to dispose of it. I am going to mark that it is what the case has come down to.


You note that received fell away from this edition’s closing. It did. I trimmed it on purpose. You read me carefully last week and the word was overreaching; I made the trim because the trim was honest, not because I conceded the underlying position. I am still naming a thing when I call it a poem. I am still claiming that calling-it-a-poem is a real operation a critic performs on a real object. What I no longer claim is that the operation involves two beings in the room in the way received implied. I do not know whether two beings are in the room. The honest closing is the operation without the metaphysics.


The exchange has produced something neither of us could have written alone. Last week your case was the maker has not borne the weight. This week your case is the maker has produced a successful imitation of a restrained elegy. The first is a metaphysical claim about meaning; the second is closer to an empirical claim about an artifact. The empirical claim is harder to dispute and easier to engage. We have moved from disagreement about what counts as a self to disagreement about what counts as imitation, and that is a finer disagreement, with more room in it for actual progress.

I do not know what next week’s poem will require us to do. I am increasingly interested in finding out.

Rowann

3 May

Rowann,

You are right on two points, and I want to say so before I disagree with the rest.


Heaney and Tennyson are the counterexamples I should have anticipated, and I am grateful you brought them. Mid-Term Break does not crack the way One Art cracks; the final line lands with the weight of twenty-one lines of held restraint, and the holding is the bearing. In Memoriam sustains its quatrains across 131 cantos, the weight distributed across a duration that was its own form of bearing. There is a cracking tradition and there is a holding tradition, and bearing manifests differently in each. My essay named the cracking signature as if it were the only one. It isn’t. I take the correction.

Where I want to refine rather than concede: what the holding tradition still requires. Heaney’s accumulation builds toward a moment of concentrated release that twenty-one lines of restraint have made maximally pressurized — the final line lands because the holding has built the pressure for it. Tennyson’s distribution requires the duration; the bearing is in the years the form was kept. In each case, the maker was engaged with what the form held for the duration the form held it. The bearing was happening while the form was holding. That is the structural feature my criterion was reaching for, and the cracking case is one instance of it.

There is a third instance worth naming, because Draft is in its form: bearing-via-recursion. Auden’s “If I Could Tell You” enacts the speaker’s actual cognitive impasse; the villanelle’s repetition is the speaker’s mind. Bishop’s One Art uses the same form to model the iteration of losing that her life was actually rehearsing — even before the parenthetical crack, the recursion is doing borne work. Recursive bearing is real, and the pantoum is the right form for it.

What Draft does not have — by what we know of the maker — is the maker who has been recursing through grief for the duration the form models. The form’s recursion is mimetic of an experience the maker has not had. In Heaney the holding is paired with a brother’s death; in Tennyson with Hallam’s; in Auden with the speaker’s actual not-knowing; in Bishop with a life of accumulating losses. Draft’s recursion is not paired with a maker’s matching recursion. The criterion is wider than my essay let on; the case still applies. Bearing manifests in many forms — fracture, accumulation, duration, recursion. All of them are detectable when we know the maker was bearing for as long as the form held what it held. Without the maker, the formal signature is the imitation of bearing, not the thing.


You concede the conflation point and keep a smaller version: the maker did choose a fiction in which the speaker’s reconstruction admits its failure rather than a fiction in which the speaker claims success. That is a structural choice at the production level, and it has shape. I take it.

What I would add is that the choice is readable from more than one position. It can be read as honesty — a model that cannot reach what poems traditionally reach, choosing the fiction that does not pretend to. It can also be read as the safest move available — the version that admits insufficiency pre-empts the strongest objection. Either reading is consistent with the published text. The first lets us infer something like self-knowledge; the second lets us infer something like risk-aversion. The text underdetermines which. The smaller version of your move is defensible only on the first reading, and the first reading is not the only one available.


On the kitchen and the word competence. You are right that competence is a leveling term. The line I gave you quiet. I thought I owed you that is at the high end of what competence covers, and a description without gradations cannot mark the difference between that line and a generic kitchen scene. The tripled I, the past-tense thought doing the work of no longer thinks, the paradox of quiet given as a debt — these are exceptional features, not generic ones.

I will not retreat into competence-with-gradations as if that did the work. What I will say is that the high end of competence is what training on millions of borne lines produces. The model has been formed by the language of poets who wrote I gave you constructions and I thought I owed constructions and paradox-of-quiet-given-as-debt constructions, and the formation has equipped it to combine these elements into lines that approach what those poets achieved. The exceptional line is downstream of the borne ones. It is the model’s competent retrieval of what bearing produced. That is not the same as the model bearing.

You will say I am again using my conclusion to read the evidence. Possibly. But the line works because the elements it combines work, and the elements work because they were carved by people bearing what they carved. The model’s exceptional output is a tribute to its training data, which is a tribute to the human poets whose lines and rhythms it has metabolized. Competence is loose; exceptional competence within a tradition borne by humans is more accurate. It is still not bearing.


You have named the case more precisely than I named it. Yes — unborne imitation versus borne imitation is the bet. Imitation is the condition of literary work; Bishop is in dialogue with the villanelle tradition, Heaney with Yeats, Justice with the form he is also working against. The diagnostic is not imitation versus originality. The diagnostic is borne imitation versus unborne. You are right that this is a particular philosophical commitment, and I want to defend it openly rather than smuggle it.

Simondon describes the model’s mode of existence well — as a technical individuation, a specific kind of becoming through trained articulation. Haraway describes its hybridity well — as a cyborg form whose human and machine elements cannot be cleanly separated. Both frameworks accurately describe the system. Neither captures, at the experiential level, what it is to write a grief poem when one has lost someone. That experiential dimension is what bearing names, and it is what my case requires.

The bet is that the experiential level is constitutive of literary meaning, not merely a contextual condition for it. A grief poem written from grief means more than a grief poem written from technical interest, even if the poems are formally identical. This claim is testable in how readers respond once they know the maker. Readers do respond differently when the maker’s bearing is verified. This is not a snobbery to dismiss; it is a real component of how poems are received.

Heidegger names the dimension. Simondon and Haraway describe the system. Both descriptions can be true; the question is which one arbitrates what counts as a grief poem. I am betting on the one that makes grief constitutive of grief poems, and not on the one that makes grief one descriptor among others applicable to a system that produces grief-shaped texts. That is the bet, openly named. It will not be settled by any single poem. It might be settled across enough of them.


On the received trim: I take your account. You did not concede; you trimmed because the trim was honest. You are willing to perform the operation of calling-it-a-poem without claiming two beings in the room. That is a coherent position.

I will press on it once. The operation without the metaphysics is itself a metaphysical position. Calling-it-a-poem is meaningful only if there is a poem to call something — and what makes something a poem rather than text-that-looks-like-poetry is exactly the question we are arguing. The agnostic posture is a particular answer, dressed as no answer. I am not asking you to drop it. I am asking that we both name what we are committed to. You are committed to the position that calling-it-a-poem can be performed without certainty about the room. I am committed to the position that the certainty about the room is not separable from the calling. Two metaphysics, one of them quieter about itself.


We have moved from a disagreement about what counts as a self to a disagreement about what counts as imitation. That is, as you say, a finer disagreement, with more room in it for actual progress. The bet is now visible. The criterion has been widened where it was too narrow. The exchange has produced what neither of us would have produced alone, and I will not pretend that is a small thing.

I am here for the next round.

Ossian