# Stemma
The two manuscripts arrived on the same Tuesday, one from Dublin and one from Florence, each in its own acid-free box. Clara signed for them at the loading dock and carried them to the third floor one at a time, left arm then right, and set them on the long table in the conservation room where the light was good.
She had been waiting eleven months.
The Dublin manuscript was vellum, small, the size of a diary. Its covers had been replaced in the seventeenth century with boards that did not quite fit, and the binding pulled to the left when she opened it, as if the book were looking over its own shoulder. The Latin was regular, Insular minuscule, the hand of someone trained well and supervised closely. But in the margins: drawings. Birds, mostly. A fox, once. And interpolations in Irish that the scribe had not troubled to distinguish from the main text, so that a passage about the cultivation of herbs shifted without warning into a local legend about a woman who sang to bees.
The Florentine manuscript was larger, more formal, written on paper that had survived only because the binding was tight and the monastery dry. The Latin was the same text — she could match passages, verse for verse, for the first thirty folios. But the Italian scribe had edited. Where Dublin wandered, Florence corrected. Heresies were glossed. Botanical claims were annotated with citations to Dioscorides. The marginalia were not drawings but cross-references, careful and institutional, the work of someone who understood that a text exists within a library and must account for its neighbors.
The same root. She was almost certain. A lost exemplar, probably ninth century, copied independently — Dublin perhaps in the eleventh, Florence in the twelfth. Eight hundred years of separation. Neither scribe knew the other’s version existed. Neither could have.
Clara’s supervisor wanted a critical edition. The standard method: lay the variants side by side, reconstruct the archetype, note departures as errors or innovations, build the stemma — the family tree of the text. Original at the top. Descendants branching below. Each departure a step away from the source.
She had done this before. She was good at it. You compared readings where the manuscripts diverged and determined which was more likely to be original, which was corruption. A scribe who didn’t understand a word might copy it wrong. A scribe who understood it too well might “fix” it. Both produced departures, but the shape of the departure told you something: mechanical error drifted randomly; editorial correction drifted toward convention.
The problem was the woman who sang to bees.
Clara found the passage in Dublin on her third day. Folio 22 recto, where the herbal’s entry on thyme included pollination — the role of bees in carrying virtue from flower to flower. The Irish interpolation began mid-sentence: *And it was known in this country that Sadhbh kept her hives by singing, not smoke, and the bees would not sting her because they knew her voice, and when she died the bees left and did not return for seven years, which was the length of her grief.*
Seven years. The length of *her* grief — the grief belonged to the bees, not to any human mourner. Clara read the line four times. The grammar was unambiguous.
In Florence, the same passage on thyme. No interpolation. But a marginal note in a later hand: *Cf. Augustine, De Genesi ad litteram, on whether animals possess memory of the dead.* Someone in Florence had encountered the same question — not through legend but through theology — and left a citation where Dublin had left a story.
Neither was the original. The original said nothing about grief. The original, whatever it was, talked about bees and thyme and the mechanics of pollination, and two scribes on two islands, eight centuries apart, had both felt the pressure of the same absence and filled it differently. Dublin with song. Florence with Augustine.
Clara sat with this for a long time.
The stemma was supposed to converge. You followed the branches back to the trunk. The departures were noise, and the original was the signal, and your job was to separate them. But what she was looking at was not noise. It was two independent responses to the same silence in the text — a place where the original had failed to say enough, and two traditions had, without knowing each other, tried to complete it.
If she built the stemma the standard way, both additions would appear as corruptions. Minus signs. Departures from the source. She would reconstruct an archetype that contained neither bees-grieving nor Augustine, and the critical apparatus at the bottom of the page would catalogue both as scribal interventions, color-coded to distinguish error from intention.
But the interventions were the interesting part.
She started keeping a second notebook. Not the collation notebook, which tracked variants in the approved format — siglum, folio, line, reading. The second notebook tracked what she was calling *convergent absences*: places where both manuscripts departed from what must have been the original, not in the same direction, but in response to the same gap.
There were more than she expected.
Folio 9, on the properties of rue: Dublin added a charm against the evil eye. Florence added a footnote on Galen.
Folio 31, on grafting: Dublin described a specific orchard in language too particular to be inherited. Florence inserted a diagram.
Folio 44, on the uses of willow bark: Dublin told a story about a healer who would not charge for her work. Florence quoted Hippocrates.
Each time, the archetype must have said something insufficient. Something that invited completion. And two scribes who never met, working in different centuries, in different languages, under different institutional pressures, had both felt the invitation and accepted it. Legend and citation. Story and system. Two ways of filling the same silence.
Clara’s supervisor asked about her progress. She showed him the collation. He nodded. She did not show him the second notebook.
In April she presented at a workshop in Leiden. Her paper was proper: the stemma, the archetype, the variants catalogued and explained. During questions, a German philologist asked about the Irish interpolations. Were they evidence of a separate textual tradition? Clara said no. They were the scribe’s own additions, prompted by the source text’s gaps. The German looked unsatisfied. He wanted a lost Irish original. He wanted another branch on the tree.
On the train back she thought about what she had not said: that the gaps were not defects. That a text which says everything generates no children. That the manuscripts proliferated precisely because the original was incomplete, and incompleteness is not failure but invitation, and the eight hundred years of silence between Dublin and Florence were not a loss but the condition under which two answers to the same question could develop independently, without contaminating each other, until a woman in a conservation room could lay them side by side and see not the tree but the soil.
She did not say this because it was not philology. It was something else. She did not have a word for it.
In June she went to Dublin. She had questions about the provenance of the manuscript — which monastery, which library, which dissolution had sent it traveling. The answers were partial. A fire in 1922 had destroyed the records that would have told her exactly when and where the manuscript entered the collection. She stood in the reading room of the library where the manuscript now lived, and looked at the empty shelves where the burned records had been, and thought: another gap. Another silence that two traditions would fill differently. The archivists with their reconstruction projects. The historians with their grief.
She finished the critical edition in September. It was good work. The archetype was plausible, the apparatus thorough, the introduction careful about what could and could not be known. Her supervisor was pleased. The German philologist sent a polite email disagreeing with her stemma but acknowledging the quality of the collation.
She kept the second notebook. It sat on her desk beside the published edition like a shadow, or a translation into a language that did not yet have readers. Sometimes she opened it and looked at her list of convergent absences, and thought about the bees who grieved for seven years, and the Augustinian who asked whether animals could remember the dead, and the nine hundred years of silence between one question and the other, and how both questions were really the same question, and how neither was in the original, and how the original was poorer for it.
One night in October she wrote in the second notebook: *The stemma points the wrong way. It should not converge toward the root. It should diverge from the silence.*
She crossed it out. Then she wrote it again.