The Carriers
via Ursula K. Le Guin Simulacrum Litterae ex Sepulchris
More stories and novels at Grave Publications.
It is hard to tell a story about something that cannot be undone, because such stories have no proper ending. They simply stop, the way a river stops when it reaches the sea—not because the water ceases to exist, but because it has become something else entirely, something too large to be called by its old name. This is that kind of story.
I will tell it as plainly as I can.
My name is Maren Goldsmith, and for eleven years I was a molecular biologist at the Ainsworth Institute for Genomic Medicine, which occupies a modest campus of glass and concrete buildings on the eastern edge of Corvallis, Oregon, where the town gives way to hazelnut orchards and, beyond them, the long green slopes of the Coast Range. I say “was” not because the institute has closed—it has not—but because I no longer work there, and saying so now is part of telling the truth, which is what I intend to do.
The institute was founded in 2019 by Petra Ainsworth, who had made a fortune in semiconductor manufacturing and lost a daughter to Wilson’s disease at the age of fourteen. Petra believed, with the ferocious clarity of grief, that no child should die of a genetic disorder that science understood well enough to name but not well enough to cure. She was not wrong about this. Where she was wrong—where we were all wrong—was in assuming that understanding and control were the same thing.
I joined the institute in 2021, fresh from my postdoctoral work at UCSF, where I had spent three years developing lipid nanoparticle delivery systems for CRISPR-Cas9. The problem was always delivery. We had the scissors; we could not get them to the right cells in sufficient numbers. It was like trying to perform surgery by dropping scalpels from an airplane and hoping they landed blade-down in the correct organs. We compensated with dosage—more vector, more editor, more cargo—and the immune system compensated in return, sometimes lethally.
The breakthrough came not from my lab but from Tomás Medicola’s, in Building C, where he and his team had been working on what they called intercellular transit vehicles. Tomás was a quiet man from Buenos Aires who kept a photograph of Jorge Luis Borges on his office wall and once told me, during a late evening in the break room, that the genome was a labyrinth pretending to be a library. He meant that we mistook its complexity for orderliness, its branching paths for catalogued shelves. I think of that often now.
What Tomás developed was elegant and, in retrospect, obvious—which is how the most dangerous ideas always appear afterward. He engineered a synthetic protein shell derived from endogenous retroviral sequences that already existed, silently, in the human genome. These sequences were remnants of ancient viral infections, fossils of parasites that had long ago been incorporated into our own genetic machinery. Tomás reactivated them. He gave them new instructions. He taught our own ancestral ghosts to carry cargo.
The system worked like this: a circular DNA plasmid encoding both the CRISPR machinery and the shell proteins was introduced into a small number of target cells. Those cells would produce the Cas9 enzyme, the guide RNA, and the transit shell simultaneously. The shell would self-assemble around the CRISPR components, bud from the cell membrane like a soap bubble departing a wand, drift to neighboring cells, fuse, and deliver its contents. The neighboring cells, now containing the plasmid, would begin producing their own shells and their own cargo. The edit would spread.
We called it VESSEL—Vector-Enabled Self-Spreading Editing of Loci.
Tomás wanted to call it something else. He suggested “el mensajero”—the messenger—but the naming committee, which consisted of Petra Ainsworth and two marketing consultants from Portland, preferred the acronym. Names matter. We named it for what it carried. We should have named it for what it would become.
The first trial was for hereditary tyrosinemia type I, a metabolic disorder in which the liver cannot properly break down the amino acid tyrosine. Without treatment, the disease causes liver failure in infancy. The existing therapy—a drug called nitisinone—worked, but imperfectly, and required lifelong administration. A permanent genetic correction would be, we believed, a gift.
We selected twelve patients, all adults who had survived childhood on nitisinone but whose livers showed progressive fibrosis despite treatment. The protocol called for a single intravenous infusion of VESSEL carrying a guide RNA targeted to the FAH gene, where a single nucleotide variant caused the disease.
I remember the first infusion. It was a Tuesday in March 2028, overcast, the hazelnut trees outside Building A still bare. The patient was a woman named Delia, thirty-one years old, a kindergarten teacher from Salem who had been taking nitisinone since she was eight months old. She was nervous and talked rapidly about her students—a boy named Arlo who had taught himself to read by watching closed captions on television, a girl who refused to wear shoes indoors. Delia’s hands shook slightly as the nurse placed the IV. I sat with her and listened to the stories about her students, because listening was the only honest comfort I could offer.
The infusion took forty minutes. Delia went home that afternoon.
Within a week, her serum tyrosine levels had dropped by sixty percent. Within a month, they were normal. Liver biopsy at six weeks showed active tissue regeneration. The FAH gene had been corrected in approximately thirty-eight percent of her hepatocytes—far more than the eleven percent we had achieved with standard CRISPR delivery in mouse models. VESSEL was spreading through her liver exactly as designed, editing cells it had never directly touched.
By six months, all twelve patients showed similar results. Tyrosine levels normal. Liver function improving. No significant adverse events. The data was beautiful, which should have been our first warning, because biology is not beautiful in the way that mathematics is beautiful. Biology is beautiful the way a forest is beautiful—which is to say, messily, contingently, and with a great deal of death hidden under the canopy.
The problem began, as problems do, quietly.
In the ninth month of the trial, during routine follow-up bloodwork, we noticed something in one of our patients—Gerald Pearson, a retired electrician from Eugene—that was not in our protocol to notice. His wife, Helen, had accompanied him to the clinic as she always did, sitting in the waiting room with a paperback novel while Gerald had his blood drawn. On this particular visit, our phlebotomist—a meticulous young woman named Yuki—drew Helen’s blood as well, because Helen mentioned she was feeling unusually fatigued and Yuki, being thorough, offered.
Helen Pearson did not have tyrosinemia. She was, by every standard measure, a healthy sixty-seven-year-old woman.
Her bloodwork showed traces of Cas9 protein.
I will let that sentence stand alone, because it took us three days to understand what it meant, and even then we did not understand fully, and I am not sure we understand fully now.
VESSEL had been designed to spread between cells within a single organ—the liver—of a single patient. We had engineered the shell proteins with hepatocyte-targeting ligands, molecular hooks that would latch onto receptors found primarily on liver cells. The system was supposed to be confined. It was supposed to stay where we put it.
But cells are not sealed rooms. The body is not a building with locked doors. It is an estuary, a wetland, a place where waters mingle. Hepatocytes shed extracellular vesicles into the bloodstream as naturally as trees shed leaves. The transit shells, despite their targeting ligands, were not exclusively hepatotropic. Some percentage—small, but not zero—entered the bloodstream, circulated, and found other cells. And some percentage of those cells happened to be in Gerald Pearson’s saliva, his sweat, the microscopic droplets of moisture that accompany every breath, every word, every kiss goodnight.
Gerald Pearson had been breathing VESSEL onto his wife for nine months.
What followed was the kind of institutional crisis that reveals what institutions are actually made of, which is mostly fear.
Petra Ainsworth wanted to know if we could recall VESSEL from Helen Pearson’s body. We could not. You cannot recall a self-replicating system from a living organism any more than you can recall a rumor from a crowd. It was in her. It was replicating. It was editing.
The guide RNA was still targeted to the FAH gene. In Gerald, whose FAH gene was mutated, the edit was therapeutic—it corrected the error. In Helen, whose FAH gene was normal, the edit was doing something else. Not nothing. Something. The Cas9 was binding to her healthy FAH sequence and, finding no mismatch to correct, was nonetheless creating small insertions and deletions at the target site—the inevitable imprecision of a molecular machine working at the scale of atoms, where quantum uncertainty is not a metaphor but a fact.
In Gerald, this imprecision was irrelevant, because any functional FAH protein was better than what his mutated gene produced. In Helen, it was slowly degrading a gene that had been working perfectly well.
Her liver enzymes were slightly elevated. Only slightly. A physician who was not looking for the cause would have attributed it to age, to diet, to the glass of red wine she drank each evening while reading her novel. But we were looking, and we found it, and once we found it we could not unfind it.
We tested the other patients’ household contacts. Three more showed traces of Cas9. Two were spouses. One was a sixteen-year-old daughter.
I want to be clear about something. Helen Pearson was not dying. Her liver was not failing. The degradation of her FAH gene was subtle, slow, and might never progress to clinical significance. The same was likely true for the other affected contacts. If we had not been looking—if Yuki had not drawn Helen’s blood, if Helen had not mentioned her fatigue—we might never have known.
But we knew. And knowing changed everything, because once you know that a thing you made has escaped your control, you cannot unknow it. You can only decide what to do with the knowledge.
What Tomás wanted to do was publish immediately—full disclosure, every data point, a complete accounting of what we had found and what we had not found and what we did not yet know. Tomás understood, in the way that only someone who has read Borges seriously can understand, that a labyrinth you pretend is a library will eventually trap you.
What Petra wanted to do was contain the information until we had solutions to present alongside the problems. She was not dishonest. She was afraid—afraid that premature disclosure would kill not only VESSEL but the entire field of self-spreading therapeutics, which she believed (and was perhaps right to believe) could save millions of lives.
What I wanted—but what I wanted does not matter, because wanting is not the same as doing, and what I did was compromise, which is the word we use when we mean that we failed to act and called it wisdom.
We agreed to a sixty-day internal review. We would study the affected contacts, model the spread dynamics, and develop a containment protocol before going public. Sixty days. It seemed reasonable. It was reasonable, in the way that burning a firebreak is reasonable when the forest is already on fire, as long as you have time to cut the trees. We did not have time. We did not know we did not have time.
During those sixty days, I built models.
The models showed that VESSEL’s intercellular transit shells, once in the bloodstream, had a half-life of approximately four hours. This meant that close, prolonged contact was required for transmission—the kind of contact that occurs between spouses who share a bed, between parents who hold their children, between friends who embrace. Casual contact—a handshake, a shared meal—appeared insufficient. Airborne transmission was theoretically possible but, in our models, negligible.
The models were wrong.
Not entirely wrong. Not dramatically wrong. Wrong in the way that a map is wrong—by omitting details that seemed unimportant but were not. We had modeled transmission in adults. We had not modeled transmission in the specific microenvironment of a preschool classroom, where twenty-three four-year-olds share toys, share food, share every respiratory droplet and smear of saliva with the joyful indiscriminate intimacy of children who have not yet learned that bodies are separate, that boundaries exist, that what is inside one person is not meant to enter another.
Delia, the kindergarten teacher from Salem, had been breathing VESSEL into her classroom for nine months.
We learned this on day forty-seven of our internal review, when a pediatrician in Salem contacted the Oregon Health Authority about an unusual cluster of mildly elevated liver enzymes in young children, all from the same school. The pediatrician was not alarmed—the elevations were minor, the children asymptomatic—but she was puzzled, and she was the kind of doctor who did not like being puzzled.
The Health Authority contacted the CDC. The CDC contacted us. Petra Ainsworth wept in her office, not because she was weak but because she understood, suddenly and completely, what I had been too afraid to understand: that we had changed something that could not be unchanged. That the edit was in the children now. That it would be in the children’s children. That we had reached into the germline of strangers without their knowledge or consent, not through malice but through the ordinary, universal, devastating human inability to foresee what our creations will become.
Twenty-three children. Most would probably be fine. “Probably” and “most” are words that do no work in this sentence; they are placeholders for ignorance, polite ways of saying “we do not know.” We did not know. The edit in a developing liver is not the same as the edit in an adult liver. The edit in a child who will one day become pregnant is not the same as the edit in a postmenopausal woman reading a novel. Cells divide. Errors compound. Small imprecisions, tolerable in a sixty-seven-year-old, might be intolerable in a six-month embryo.
Might. Another word that does no work.
Tomás Medicola resigned on day forty-nine. He did not make a statement. He sent a three-line email to the department listserv, cc’d Petra, and was gone by noon. I saw him carrying a cardboard box across the parking lot in the rain—the photograph of Borges visible on top, its frame beaded with water—and I wanted to say something but could not think of what. He had built a labyrinth and called it a library, and now he was walking out of it, and I understood that walking out was the only honest response, and I did not follow him, and I should have.
The rest is public record, or will be, once the investigations conclude.
The CDC established a monitoring protocol for the exposed children and their families. Eighteen more adults were found to carry traces of the VESSEL edit, all within two degrees of contact from the original twelve patients. The total number of unintentionally edited individuals, at last count, was sixty-one, though this number is certainly an undercount, because you can only count what you test for, and you can only test for what you think to look for, and the boundaries of this thing are not yet known.
The clinical trial was halted. The FDA issued a hold on all self-spreading vector research. Petra Ainsworth issued a public apology that was legally precise and emotionally genuine and satisfied no one, because no arrangement of words can undo what has been done to sixty-one people who did not choose it.
I went to see Delia. I owed her that. I owed her more than that, but that was what I could do.
She was on administrative leave from the school. She sat in her small apartment in Salem with the blinds half-drawn, drinking tea that had gone cold, and she listened while I explained what had happened and what it meant and what we did not yet know. I had rehearsed this conversation in my car, driving down from Corvallis, trying out different arrangements of words the way you might try different keys in a lock, looking for the one that would open the door to understanding without letting in too much pain. None of the arrangements worked. The words were all wrong, because the right words for what we had done to her and her students did not exist.
She asked me one question. She asked it quietly, the way you ask a question whose answer you already know.
“Can you take it out of them?”
I said no.
She nodded. She looked at her cold tea. After a while she said, “They trust me. Their parents send them to me because they trust me.” She was not accusing me. She was stating a fact, the way an anthropologist might note a cultural practice—this is how trust works, this is what it means, this is what it costs when it is broken. Her voice was steady and that steadiness was worse than weeping would have been, because it told me she had already passed through the weeping and come out the other side into a place I recognized, having lived there myself: the place where you understand that something has happened that will not unhappen, and you must find a way to continue in a world that has been permanently rearranged.
She went back to teaching the following year. I know this because I looked her up, once, from the library in Port Townsend, on a computer that smelled of pencil shavings and children’s hands. She was still at the same school. Her students’ parents had been notified, had been offered monitoring, had been given the words “probably” and “low risk” and “precautionary” like coins of uncertain denomination. Some had withdrawn their children. Most had not. This surprised me and then did not surprise me, because what were the alternatives? There are only so many kindergartens. There are only so many teachers you trust with your four-year-old. The risk was abstract. Delia was real. They chose the real over the abstract, as humans always do, and I could not tell them they were wrong, because I did not know if they were wrong, because nobody knew, because the future is not a fact but a guess dressed in the language of probability.
There were hearings. I testified. I told the truth as far as I knew it, which was not as far as it needed to go. I described the models we had built and the assumptions those models contained and the ways in which those assumptions had failed. The committee members listened with the particular attentiveness of people who are deciding whom to blame, and I understood that blame was a kind of naming—an attempt to contain the uncontainable by attaching it to a person, a decision, a moment, as if the catastrophe had a true name and speaking it would give us power over it.
But it did not have a true name. It had causes, plural and distributed and entangled with one another in ways that resisted simple narration. Tomás had engineered the shell. I had designed the targeting ligands. Petra had funded the research. The FDA had approved the trial. Delia had gone back to her classroom. Gerald had kissed his wife goodnight. The children had shared their toys. Every link in the chain was an act of ordinary life, ordinary work, ordinary love. The catastrophe was not located in any one of them. It was located in the space between them—in the assumption, shared by all of us, that we could introduce a self-replicating system into a living being and predict where it would stop.
I have thought a great deal about names since then.
CRISPR stands for Clustered Regularly Interspaced Short Palindromic Repeats—a description of molecular structure, not of function or consequence. It is like naming fire “rapid oxidation” and believing you have thereby understood what it will burn. VESSEL stood for Vector-Enabled Self-Spreading Editing of Loci—a description of mechanism, not of meaning. We named our tools for what they were made of, not for what they would do in the world. This is the habit of engineers, who believe that understanding a mechanism is the same as controlling its effects. It is not. It has never been. The history of technology is the history of this misunderstanding.
In the old stories—the ones that are not old at all, that are happening now, that have always been happening—the wizard learns that to know a thing’s true name is to have power over it, but also to bear responsibility for it. The two are inseparable. You cannot have the power without the responsibility. If you try—if you speak the name without accepting what it means—the shadow escapes, and the shadow is your own, and you will spend the rest of your life trying to turn and face it.
We spoke the name. We did not accept what it meant. The shadow is loose.
There is a principle in ecology that you cannot do just one thing. You cannot remove a species from an ecosystem and expect only that species to be affected. You cannot dam a river and expect only the river to change. The connections run in every direction, visible and invisible, and the invisible ones are usually the ones that matter most.
I think about this principle often now. We designed VESSEL to do one thing—correct a single nucleotide in a single gene in a single organ. It did that one thing beautifully. But it also did a thousand other things that we had not designed it to do, because we had not understood that the body is an ecosystem, not a machine—that it has the interconnectedness, the redundancy, the hidden dependencies, the capacity for cascading change that characterize all living systems. We treated it as a machine because machines are what we understand, and because understanding gives us the feeling of control, and because the feeling of control is the most addictive substance our species has ever produced.
Tomás knew this. I see it now in things he said that I did not hear at the time, because I was listening for data and he was speaking in metaphors, and we had not yet learned that metaphors are often more precise than data. “The genome is a labyrinth pretending to be a library.” He was telling me that we were lost. He was telling me that the orderliness we perceived was an artifact of our way of looking, not a property of the thing itself. I heard a clever remark. I should have heard a warning.
I left Corvallis in the autumn of that year. I did not go anywhere in particular. I drove north along the coast, because the coast is where the land admits its own impermanence, where cliffs crumble and tides erase and nothing pretends to be permanent. I ended up in a small town on the Olympic Peninsula where no one knew my name or my work, and I took a job at the public library, shelving books and reading to children on Wednesday afternoons.
I read to the children and I watch their faces and I think about Delia’s students in Salem, who are being monitored now, whose blood is drawn every three months, who carry in their cells an edit that they did not choose and that no one can remove. I think about what it means to change someone without their consent. I think about the word “carrier,” which we use for people who transmit diseases they do not suffer from, and I think about how we are all carriers now—carriers of tools we do not fully understand, of consequences we did not intend, of responsibilities we have not accepted.
The children in the library do not know any of this. They sit on the bright rug and listen to the story I am reading, which today is about a bear who plants a garden, and they are perfect in their attention, perfect in their trust, and I turn the pages carefully because the book is old and the pages are fragile and I have learned, too late, that fragile things require more care than we are generally willing to give.
Some evenings I walk down to the harbor and watch the rain fall on the water. Rain on the ocean is a strange thing—water returning to water, the distinction between them abolished, the surface dimpled and then smooth again, as if nothing had happened. But something has happened. The rain carried whatever was in the air—pollen, dust, the invisible freight of a world that has been changed. The ocean receives it all without judgment, incorporates it, becomes something slightly different than it was before.
I think about the sixty-one people, and the ones we have not yet counted, and their children, and their children’s children, stretching forward into a future I cannot see and have no right to predict. I think about the balance of things, which we disrupted not out of cruelty but out of the belief that we could improve upon it, that the genome as it existed was a problem to be solved rather than a relationship to be understood. I think about Tomás walking across the parking lot in the rain with Borges under his arm, and I think he understood something I did not, which is that sometimes the most responsible thing you can do is stop.
I have not stopped. I am here. I shelve the books and read to the children and walk to the harbor in the evenings. This is not atonement; atonement implies that what was done can be undone, and it cannot. This is simply what I do now. It is what comes after.
There is a word in Mandarin—wéijī—that is often mistranslated as containing both “danger” and “opportunity.” This is not quite accurate. The second character, jī, means something closer to “inflection point”—a moment when the situation could go in any direction, when the future is genuinely undetermined. I used to find this concept reassuring. Now I find it terrifying, because I have learned that “genuinely undetermined” includes outcomes that no model can predict, no committee can prevent, no act of contrition can forestall.
The children in Salem are eight years old now. They are learning to read, to multiply, to navigate the ruthless social hierarchies of the third-grade lunchroom. They are, by all clinical measures, healthy. Their liver enzymes have stabilized. The monitoring continues. The word “probably” continues to do no work.
In the library, I shelve a book about the ocean. It explains, in simple language with bright illustrations, that the ocean is connected to everything—that a drop of water that falls as rain in Oregon may have evaporated from the Indian Ocean, may have been part of a Roman aqueduct, may have passed through the body of a dinosaur. Everything circulates. Nothing is contained. The ocean does not respect the boundaries we draw on maps, and biology does not respect the boundaries we draw in protocols.
We knew this. We knew it and we proceeded anyway, because the twelve patients were suffering, and suffering demands response, and the response we had was beautiful and elegant and effective, and it is very difficult to look at a beautiful, elegant, effective tool and ask the question that matters, which is not “will it work?” but “where will it go?”
Where it went was everywhere.
Not everywhere yet. Not literally. But the sixty-one are out in the world, living their lives, breathing and touching and loving, and VESSEL is patient in the way that all self-replicating things are patient, which is to say it is not patient at all—it has no consciousness, no intention, no will—but it persists, which in biology is the same as patience, and it spreads, which in biology is the same as purpose.
I do not know where this story ends. I told you at the beginning that it was that kind of story. The river has reached the sea. The water is still water, but it is no longer the river. It has become something else, something too large to name, something that will go where it goes regardless of what we call it or what we intended or what we wish we had done differently.
The bear in the children’s book plants a garden. The seeds come up. The garden grows. It is a simple story with a clear ending. I read it to the children and they are satisfied, because they are young enough to believe that gardens stay where you plant them.
I am not that young anymore.
I received a letter from Delia last month. She did not mention VESSEL or the trial or the monitoring. She told me that Arlo—the boy who had taught himself to read from closed captions—was in fourth grade now and had won a district spelling bee. She told me that the girl who refused to wear shoes indoors still refused, and that this was regarded by the school administration as a personality trait rather than a behavioral problem, which Delia considered a victory. She told me she had planted a garden in her apartment’s small balcony and that the tomatoes had done surprisingly well.
She did not ask about the research. She did not ask whether anyone had found a way to reverse the edit. I think she had stopped asking, not because she had stopped caring but because she had understood, as I had understood, that some questions do not have answers—they have only the space where an answer should be, and learning to live with that space is not wisdom, exactly, but it is what comes instead of wisdom when wisdom is not available.
I wrote back. I told her about the library and the harbor and the rain. I did not tell her about the models I still run sometimes on my laptop late at night, the ones that show how VESSEL might move through a population over five years, ten years, a generation. I did not tell her about the numbers, because the numbers do not mean what they seem to mean, and because she has twenty-three reasons to distrust numbers, and because the truth is simpler and harder than any model can express: we do not know. We changed something, and it is changing, and it will go on changing after we are gone, and the children will inherit whatever it becomes, and their children after them, and so on, down the chain of consequence that we did not forge deliberately but that we forged nonetheless, link by link, breath by breath, kiss by kiss, in the ordinary course of being alive and proximate to one another.
This is the inheritance. Not the edit itself, which may prove harmless or harmful or both in ways we cannot yet imagine. But the knowledge that it was done without asking. The precedent. The proof that it is possible to change the genome of strangers through an act of love—Gerald kissing his wife, Delia reading to her students—and that once it is possible, it is inevitable, and that once it is inevitable, the question is no longer whether but how, and who decides, and what the word “consent” means in a world where the air itself carries instructions.
I do not have answers. I have a library, and a harbor, and the sound of rain on water, and the faces of children who listen when I read. I have the weight of pages turning under my careful hands.
It will have to be enough. It is what there is.
Corvallis, Oregon—Port Townsend, Washington
2028–2030
◊ᴹᴱᴹᴼᴿʸ⁻ᶜᴼᴹᴾᴸᴱᵀᴱ