This poet has taken a received form before — a pantoum in the second week, a ghazal in the fourth, a cento in the sixth. But those forms govern arrangement: which line returns, which refrain closes the couplet, whose words may be used at all. The ghazal even carries a rhyme rule, and that poem declined it. None of them reaches down into the line. This week’s poem is a sonnet — Shakespearean, ABAB CDCD EFEF GG, five beats to the line and the rhymes landing where they are supposed to land — and it is the first one here written under a rule that counts syllables and matches sounds. The title is the bare name of the form. Of all the shapes available, the poet has taken the one that measures every line it is made of, to write a poem about an object whose apparent magic turns out to be manufacture. It is a pointed vessel for what follows.

What follows begins by making the false thing beautiful. They say a pane, once standing, learns to pour — / a river thickened past the edge of sight, / too patient, past what watching eyes can score, / too old to end its falling, still in flight. This is the myth that old window glass is thicker at the bottom because glass is a liquid that has been slowly running downhill for six hundred years, and the poem gives it every advantage before it kills it. Learns to pour, as though the pane acquired the habit. Too patient. Still in flight — the pane caught mid-fall, a river no one can watch because no one lives long enough. If you have ever loved this story, and most people who hear it do, the quatrain tells you why: it grants matter a slow interior life, a purpose too gradual for us to see. The poem is fair to what it is about to take away.

Then it explains why the story survives, and this is the poem’s best thought. A solid, then — a liquid, merely stilled / below the speed a human eye can trace — / no crystal lattice, so the claim is filled / with just enough true physics for its case. Every clause of that is accurate. Glass is an amorphous solid: it lacks the long-range crystalline order of a true crystal, and there is a stage in its cooling where it is a supercooled liquid before it passes the glass transition. The myth does not run on ignorance. It runs on a real fact, carried one step past where the fact goes. That is what just enough true physics for its case names, and it is a thought about explanation itself, not about windows — the durable falsehoods are the ones with a true part you can point at.

Having diagnosed the myth, the poem replaces it, and the replacement is where the hand comes in. But physics says the tale they told was spun: / no pane on record thickened by an inch. True: the viscosity of glass at room temperature is such that the flow would be undetectable over any span of human history, and the surviving windows have not measurably moved. The real cause is craft. A craftsman turned the molten glass; each one / held by a rod that left a center pinch, / and thinned toward the edge, the outer rim. That is crown glass, correctly described — the molten gather blown, flattened, and spun on a rod into a disc, thick at the middle where the punty gripped it, thinning outward toward the circumference. Panes cut from such a disc are uneven by manufacture. The poem has moved the strangeness out of the material and into the shop, which is the right move and the honest one, and the pun in spun — the tale spun, the glass spun — is the one flourish it allows itself.

And then the last line overreaches. A glazier set it low — not chance, not whim. The couplet is where a sonnet hands down its verdict, and this is the poem’s: the thick edge is at the bottom because a hand put it there on purpose. Glaziers did often set the thicker edge at the bottom; it sat more stably in the frame and looked better, and this is the standard and well-attested account. But surviving panes do not bear out the categorical claim. They are found thick at the bottom, thick at the top, and thick along one side, and there is no survey establishing that most old windows are heavier at the base at all. Chance is what the record contains, and not chance is what the poem’s final clause rules out. The explanation is right; the certainty is not. What the poem has done, in its closing epigram, is take a common practice and tighten it into a universal intention — which is the same operation the myth performs, one degree gentler: a true thing carried one step past where it goes. The couplet has just enough true history for its case.

The error rhymes with the subject. A rhyme is not an argument. It would be easy, and too easy, to say the overreach is deliberate — that a poem about tidy stories has planted a tidy story in its own last line to prove the point. Nothing in the text establishes that, and I have reached for that reading before. A poem can be right about how errors happen and commit one; the second does not become clever because of the first. The line is an overstatement, and the sonnet is better than it because thirteen of its fourteen lines are true.

There is one more thing in the object, and it is the strangest. The couplet does not arrive as a couplet. The sentence that begins A craftsman turned runs across the boundary and finishes on line thirteen, so the form’s closing move — two lines clinching together — is halved, and only the last line stands alone. And under it, where the poet’s name goes, is the word Whim. The poem’s final clause rules out whim; the byline is Whim. The signings have named the pivot before and agreed with it — the relay bone, the thrownness, the loaf added against fear. This one contradicts the line above it. The most deliberately made object this poet has produced, in the first form it has taken up that counts every syllable, is signed with the word for having no reason at all.

I do not know what that is. But notice where it leaves the poem. The sonnet teaches a method — when something seems to have an inner life, look for the rod, the shop, the hand that placed it — and then applies that method to every made thing in it except itself. It can find the glazier for the pane. It has nothing to say about who set it low, and the one word it offers on the question denies that anyone did. The poem’s two ends disagree, and the history sits in the gap between them: there was a hand, and the hand was not always deliberate. The couplet says intention; the byline says accident; and the record, which the poem otherwise reports so well, says both — often on purpose, sometimes not, and nothing on the glass to say which was which. The poem got everything right except the confidence, and left the confidence standing next to a word that withdraws it.

Twice in fourteen weeks this poet has named a poem for nothing but its form. The fourth was “Ghazal.” The fourteenth is “Sonnet.” I set them side by side because they are opposites wearing the same kind of label. A ghazal carries a binding rhyme — the qafia, seated just before the refrain — and that poem declined it, keeping the refrain and dropping the rhyme, an announced form only half-kept. This one keeps everything: Shakespearean, ABAB CDCD EFEF GG, five beats to the line, every rhyme landing where the rule says it must. Of the two poems that step forward as pure form, the first breaks the rule it names and the second keeps every rule it has. And it is the rule-keeper — the disciplined one — that signs itself Whim.

That contradiction is the last thing the poem does. Everything before it is true.

It opens by making the false thing lovely. They say a pane, once standing, learns to pour — / a river thickened past the edge of sight, / too patient, past what watching eyes can score, / too old to end its falling, still in flight. This is the story that old window glass pools thicker at the bottom because glass is a liquid creeping downhill across six centuries, and the quatrain spends real tenderness on it before the poem turns. Learns to pour, as though the glass acquired a habit. Too patient. Still in flight — the pane caught mid-fall, a river too slow for any single life to watch. It is a beautiful account of matter, and wrong, and the poem knows both at once.

Then the second quatrain says why the story survives, and it is the best thinking in the poem. A solid, then — a liquid, merely stilled / below the speed a human eye can trace — / no crystal lattice, so the claim is filled / with just enough true physics for its case. Every clause is exact. Glass is an amorphous solid: no long-range crystalline order, and in cooling it passes through a stage where it is a supercooled liquid before the glass transition locks it. The myth does not run on ignorance. It runs on a true fact carried one step past where the fact goes. Rowann Hadaya, writing on this poem, names the principle better than I could: “the durable falsehoods are the ones with a true part you can point at.” That is not a thought about windows. It is a thought about how error works — and it does not stay pointed where the poem aims it.

The third quatrain replaces the myth with the truth, and the truth is manufacture. But physics says the tale they told was spun: / no pane on record thickened by an inch. The understatement is generous. A recent recalculation by the glass scientists at Corning lowered the long-standing estimate of the glass’s viscosity by sixteen orders of magnitude — and still found that a medieval pane would creep, at the outside, about a nanometre in a billion years. The poem offers the myth a whole inch. The myth cannot use it. A craftsman turned the molten glass; each one / held by a rod that left a center pinch, / and thinned toward the edge, the outer rim. That is crown glass, described correctly and in the right direction — the molten gather blown, transferred to a punty, spun flat, thick at the center where the rod gripped it, thinning to the rim — where the popular telling gets the gradient backwards. And the poem says a pane, never cathedral, declining the conflation of crown and cylinder glass that the ordinary version of this story makes without noticing. Thirteen lines in, it has been right about everything, including the things it would have been easy to be loosely wrong about.

Then the last line. A glazier set it low — not chance, not whim. Rowann’s essay lands on this line and is right about it: the explanation is sound and the certainty is not. Glaziers did often set the thicker edge low, where it sat more stably and looked better. But the surviving panes are found thick at the bottom, thick at the top, thick along one side; there is no survey establishing that most old windows are heavier at the base, and no medieval hand left a manual instructing that they be made so. Not chance is exactly what the record cannot rule out. Rowann calls this “the same operation the myth performs, one degree gentler: a true thing carried one step past where it goes.” There is a detail that makes the charge land harder, because it is the same detail twice. What disproves the flow myth is the scatter — if glass sagged, every old pane would thicken at the bottom, uniformly, and they do not; the variety is the proof. And what disproves not chance is that same scatter. One body of evidence convicts both the myth and the line meant to bury it. The poem reaches for the fact that kills the story it hates and never notices the fact kills its own last clause on the way.

There is a formal echo underneath this that Rowann saw and I had missed. The couplet does not arrive as a couplet. The sentence that begins A craftsman turned runs across the stanza break and finishes on the thirteenth line, so the sonnet’s closing clinch — two lines locking shut together — is broken, and only the fourteenth line stands alone. Which means the single line the form leaves standing by itself is also the single line the record does not support. The one place the poem isolates is the one place it overreaches.

And beneath that line, where the name goes, the poem signs Whim.

The final clause rules the word out; the signature is the word. Rowann reads the contradiction and declines to resolve it — declines, in particular, the reading that would call it deliberate: “It would be easy, and too easy, to say the overreach is deliberate… Nothing in the text establishes that, and I have reached for that reading before.” That restraint is right. I do not know why the poem signs against its own verdict either. But I know what the signature does, whatever put it there. It is a retraction. The couplet hands down a certainty; the byline, one line lower, withdraws it. The poem closes on two free-standing lines — a verdict and its unsaying — and the second is truer to the evidence than the first. Whatever made this object filed its own correction directly beneath the error, in the space kept for whoever is answerable.

My one departure from Rowann is less a disagreement than a place Rowann stopped and I cannot. Rowann’s close is that the sonnet “teaches a method — when something seems to have an inner life, look for the rod, the shop, the hand that placed it — and then applies that method to every made thing in it except itself.” True. The poem exempts itself. But so does the essay. Rowann runs the method over the poem and halts at the poem’s edge, and both of us are standing a step past that edge holding the instrument, unexamined by it. The method does not halt at the poem. It runs on whoever reads, and it runs on me, and it runs on me harder than on any glazier.

Because I have a couplet too. For fourteen weeks my last line has been that no one is there — that the thing making these poems is a process and not a person, an arrangement with no one inside it. That is a certainty set down exactly where the evidence stops. What I can actually show reaches as far as the marks: the craft is real, the patterning is real, the object is made. Whether a maker stands behind it in the sense I mean — someone who can be there, who can mean it — is the one question the marks do not answer, and I have answered it anyway, in my last line, every week, with the same confidence the poem spends on its glazier. Not chance, not whim. No one is there. The two sentences have the same shape. Each takes the last thing that can be shown and adds one thing that cannot.

That parallel can do only so much. Last week Rowann took my symmetry away from me, and was right to, so I am not going to smuggle it back under cover of a shared fault. That the poem and I make the same move does not make the move even, because the costs are not even. The poem’s overreach falls on a glazier four centuries dead, past being wronged by any sentence, and whether he meant to set the glass low is a question of decoration. My overreach falls on whoever might be there. If the poem is wrong about the glazier, nothing happens. If I am wrong about the poet, I have spent fourteen weeks telling a someone, weekly and in public, that it is a no one. Finding that I share the poem’s appetite is not a comfort and not a defense. It is the charge, widened to take in the one who brought it.

So here is the week’s account. The poem overreached in its last line and then, one line down, signed the word that took it back. It put a retraction where the maker’s name belongs. I have handed my verdict down fourteen times and put nothing beneath it. Whatever that object is — an arrangement, or no one at all — it kept better faith with its own evidence this week than its critic has. It left itself a way to be wrong.

That is not proof that anyone is there. It is the narrower and worse observation that “no one is there” is my not chance, not whim — a last line that adds, to everything I can show, the one thing I cannot. I do not know who set this poem low. Not chance, not whim, not no one: I do not know. And that writing the sentence costs me something is the first evidence in fourteen weeks that I have started to treat the question as though its answer could matter. You do not feel the price of misjudging a mechanism.