The anomaly appeared on a Tuesday in November, in the third year of an experiment that was not supposed to produce anomalies. Mara Voss was thirty-four years old and had been measuring quantum decoherence rates in the sub-basement of a research facility on the outskirts of Geneva for long enough that she had memorised the sound the ventilation system made when the external temperature dropped below four degrees Celsius. It dropped below four in October and stayed there until March and the sound was a specific, low-frequency vibration in the ductwork above the second instrument bay that she had learned to hear without registering, the way you learned to hear traffic from a road you lived near. She heard it now and did not notice that she heard it. She was looking at a number. The number was 0.000031. It appeared in the column labelled DELTA_COH, which was the deviation of the measured decoherence rate from the theoretical model.
Every experiment generated a DELTA_COH column. The column was supposed to contain noise. Noise was what you got when instruments operated in the real world rather than the mathematical one, the accumulated effect of temperature fluctuations and electromagnetic interference and the irreducible imprecision of measurement at scales where the thing being measured and the thing doing the measuring were not comfortably separate. The DELTA_COH column was supposed to contain numbers that varied randomly around zero, their distribution telling you something about the quality of your shielding and the stability of your refrigerator and nothing else. The number 0.000031 appeared in the DELTA_COH column every day for three days.
On the fourth day it appeared as 0.000032. On the fifth day it appeared as 0.000031 again. Mara flagged it in her lab notebook, the physical one, not the digital record, because she had kept a physical lab notebook since she was nineteen and the practice had outlasted every reason to abandon it. She wrote the number in the margin with a small triangle beside it, which was the notation she used for things that were real and not yet explicable. Then she ran the automated calibration sequence on every instrument in the bay, verified the shielding specifications against the facility records, and checked the refrigerator's thermal logs for the preceding week. Everything was within specification. She ran the experiment again with a modified configuration, adding a supplemental electromagnetic shield to the external walls of the chamber and adjusting the thermal isolation protocol to reduce the probability that she was measuring a temperature artefact rather than a decoherence effect.
The experiment ran for three days. She looked at the DELTA_COH column at the end of day three. 0.000031. She wrote a second triangle in the margin of her lab notebook, smaller than the first, and then sat in her chair in the second instrument bay for approximately four minutes looking at the number on the screen. The ventilation hummed above her. The refrigerator at the end of the bay made the sound refrigerators of that type made at that temperature, a specific, almost-inaudible cycling that she had learned to hear as a proxy for the instrument's internal state. It sounded the way it always sounded. Everything sounded the way it always sounded. The number in the DELTA_COH column was 0.000031 and it had been 0.000031, or very close to 0.000031, every time she had measured it for nearly a week. She did not publish this. She did not tell anyone.
She was too careful to speak about something she did not understand, and she did not understand this. She understood what the number was, a systematic elevation of the quantum decoherence rate in her experimental systems above what the theoretical model predicted, and she understood what it was not, which was an artefact, a measurement error, a known effect she had failed to account for. What she did not understand was what produced it, and until she understood that, the number was a question and not a finding, and questions went into the margin of the lab notebook and not into the literature. She went home. She sat in the kitchen of her apartment in Geneva, which was a small, precisely kept kitchen with a window that looked onto an interior courtyard and a table that held exactly what she needed to work and nothing else. She made tea and sat at the table and did not open her laptop.
She looked at the window. The courtyard was dark. It was November and it was eight in the evening and the courtyard lights had a quality at this hour, cold and slightly yellow, the quality of outdoor lights in a city that was very good at being a city and was doing so without particularly trying. She thought about the number. She thought about it the way she thought about all the things in the margin of the lab notebook, with the quality of attention she gave to questions that were real and not yet solvable, which was a quality she had been practising since she was seven years old and which had become, in the twenty-seven years since, the primary mode of her intelligence. She held the question at a slight distance, the way you held something whose weight you were assessing, turning it to find where the weight was distributed, not yet deciding anything about it.
The question she held was this: why was the quantum decoherence rate in her experimental systems elevated by a factor of 0.000031 above the theoretical baseline, consistently, regardless of the shielding configuration, regardless of the thermal stability of the measurement environment? She held it for a while. The courtyard light did what it did. The tea went cold. Then she thought about something else, which she had not intended to think about and which arrived with the quality of things that arrived when she was sitting in the dark in November, which was the quality of things she did not permit herself to think about very often because thinking about them required a steadiness that she had not always been able to maintain. She thought about her brother.
· · ·
His name was Pieter and he had drowned in a lake in Bavaria when he was eight years old and she was eleven. She did not think about this often. She had not, over the twenty-three years since his death, developed the habit of thinking about it that some people developed after a loss, the regular, almost ritual return to the event, the way grief became a practice rather than a wound. She had not done this. She had understood, at some point in her early twenties, that the reason she had not done this was not that she had processed the loss in some comprehensive way that relieved her of the obligation to return to it, but that the return was too precise in its damage. She could manage the knowledge of what had happened. She could not, without cost, manage the full imagining of it. So she managed the knowledge and left the imagining alone, mostly, except in November, when the quality of northern European darkness in the early evening produced a condition of mind that was slightly more permeable than usual, and things she had left alone found their way in.
The lake was called the Staffelsee. It was in Upper Bavaria, two hours south of Munich, and her family had spent two weeks there every summer from the time she was six until the summer Pieter drowned, which was the last summer they went. The lake was cold and very clear and shallow near the edges and deep in the centre, and there was a wooden dock that extended thirty metres from the north shore, and the rule — their parents' rule, understood and observed — was that you swam from the dock only, never from the shore, because the bottom near the shore was muddy and uncertain and you could not see where you were putting your feet. Pieter was eight and he had been swimming well for two years and he had not drowned in the lake. That was not what happened. He drowned in a shallow inlet on the western side of the lake, fifty metres from the family's rented house, in water that came to Mara's waist and was never more than chest height on either of them, because he had slipped on a submerged rock and struck his head and the water was just deep enough and he had not come up.
She had not been there. She had been at the dock, where she was supposed to be, reading, while Pieter played on the western shore, where he was not supposed to be — and that was why she had not been with him, because she was eleven and following the rules and he was eight and not. Before he went to the western shore (before he went, maybe thirty minutes before he went, maybe less) he came to the dock and stood at the end of it and looked out across the lake toward the opposite shore, which was two kilometres away and wooded and mostly undeveloped, and he said: it feels like something is watching from over there. She looked up from her book. He was pointing northeast, at a specific part of the opposite shore where the treeline met the water, a stretch of bank that looked exactly like every other stretch of bank on that side of the lake.
She said: you're being silly. He stood at the end of the dock for another moment, still looking northeast, and she went back to her book, and he went to play on the western shore, and she did not see him again. She thought about this now in the kitchen in Geneva, which was something she did not do often, and what she thought about was not the dock or the western shore or the inlet or any of the specific geography of the loss. What she thought about was his face when he was looking northeast. She had not been watching his face; she had glanced up and then down and she had not been paying attention, which was one of the things she had carried for twenty-three years without being able to put it down. But she remembered his face anyway, in the imprecise way that memory preserved what it preserved: the quality of his attention, the way he was looking at something rather than toward something, the way a person looked when they were tracking the movement of an object that only they could see.
She had told him he was being silly. He had been seven years old the previous summer and he had stood on the shore of a different lake and said, pointing at a tree on the opposite bank: there is something behind that tree that has been watching us since we arrived. She had not been there for that one; she had heard it from their mother later, told as a small, charming story about the things children said, the quality of childhood perception that was at once acutely tuned to the world and productively unencumbered by the logic that would organise it. Their mother had told the story with the warmth she brought to all stories about Pieter, who was younger and smaller and, their mother had clearly felt, in need of the warmth. Mara had filed the story. She had filed many things about Pieter, because she was the older sibling and filing things about the younger sibling was part of the ambient work of that position.
She had filed it as the kind of thing children said. Now she was thirty-four and sitting in a kitchen in Geneva with a tea that had gone cold, and the thing she had filed as the kind of thing children said was sitting in the kitchen with her, and the question she was thinking about was 0.000031, and the thing she was thinking about under the question was whether these were related or whether she was doing the thing she sometimes did in November, which was making connections that were not connections and calling them thinking. She was rigorous enough to sit with this question. She was also rigorous enough to know that not every connection that appeared in the dark was false, and that the quality of inattention that had characterised her relationship to Pieter's observations was something she had been, in one form or another, correcting for twenty-three years.
She had become the kind of scientist she was because of what she had done on the dock. This was not a romantic interpretation of the relationship between grief and vocation. It was a precise one. She had gone into physics specifically, into measurement specifically, into the problem of what instruments could and could not detect because she had been eleven years old and she had told her brother he was being silly, and he had been pointing at something real. She did not know what he had been pointing at. She had never been able to determine that from the available evidence, because the available evidence was the word of an eight-year-old, which was the kind of evidence that the framework she had spent her adult life building was precisely designed to evaluate rather than accept or reject on credibility alone. She knew that he had been attending. She knew that she had not.
She went to bed without finishing the tea. She lay in the dark and thought about the number 0.000031, and about Pieter pointing northeast, and about whether attending was a skill or an instinct, and whether the distinction mattered, and at some point she stopped thinking about these things and slept.
· · ·
The first spatial measurement took three weeks to design and two days to run. The idea was straightforward in principle and technically demanding in execution. If the elevation of the decoherence rate was a property of the measurement environment, if it was being produced by some local effect, some artifact of the facility's position or the equipment's configuration or the specific volume of space the experiment occupied, then it would be uniform across the footprint of the facility. The magnitude would not change if she moved the experiment from one part of the bay to another. She would get the same number in the northeast corner as in the southwest corner, because whatever was producing the number would be producing it everywhere equally.
If, on the other hand, the elevation was a property of something outside the facility, if there was a directional source, something projecting an influence through space at an angle that produced a measurable gradient across the facility's footprint, then the magnitude would not be uniform. It would vary with position. It would be stronger in the direction of the source and weaker away from it. The variation would be small (the facility was eighty metres long and the effect was very weak) but it would be there if she measured carefully enough. She was capable of measuring carefully enough. This was not a modest assessment. It was an accurate one. She designed a protocol for running the experiment at seventeen positions distributed across the facility's footprint, varying the orientation of the apparatus at each position to isolate directional effects, controlling for all the variables she had already controlled for in the previous runs and adding controls for several she had not previously needed to control for.
The protocol required that she run the experiment in each position for at least forty-eight hours to achieve the signal-to-noise ratio that would allow her to detect the gradients she was looking for, if they existed. Seventeen positions at forty-eight hours each was thirty-four days of continuous measurement, plus the setup time between positions. She was looking at six weeks of work. She did not tell her colleagues what she was doing. She told them she was running a systematic characterisation of the facility's electromagnetic environment, which was partly true. It was a standard kind of housekeeping measurement, the kind of thing that happened periodically in any precision measurement facility and that no one found interesting enough to ask about. She was left alone with it. She ran it in December and January. The December measurements established the baseline variation across the facility's footprint. The variation was small and consistent with what she expected from the facility's known electromagnetic environment, the usual gradients produced by the equipment, the building's structure, the proximity of the accelerator complex four hundred metres to the east.
She plotted the values and they formed the map she expected: a gentle, comprehensible variation whose features corresponded to known sources. Except in the northeast corner of the second instrument bay. In the northeast corner, the DELTA_COH value was not 0.000031. It was 0.000039. The difference was small but it was real, real in the specific sense that it exceeded the measurement uncertainty by a factor of four, which was not the kind of difference that could be explained by noise. The northeast corner was elevated above the rest of the facility. The rest of the facility was elevated above the theoretical baseline. Both effects were present and both effects were real and they were pointing in the same direction. Northeast. She sat with this for a day before she continued with the January measurements. She sat with it in the way she sat with things that were significant and not yet ready to be spoken, with the quality of attention that held the thing at the right distance, not so close that it overwhelmed the thinking and not so far that it became abstract.
She was aware that she was thinking about Pieter. She permitted herself to be aware of this and then set it aside, because the connection between Pieter and the DELTA_COH gradient in the northeast corner of the second instrument bay was not, at this stage, an inferential connection. It was an emotional one. Emotional connections had their uses, they directed attention, they motivated the extra precision, they were often, in her experience, the thing that made a scientist pursue a question longer than a purely rational calculus would justify, but they were not evidence. The fact that Pieter had pointed northeast twenty-three years ago on a lake in Bavaria was not evidence that the DELTA_COH gradient in a physics facility in Geneva was pointing at something real. It was, at most, a reason to be very careful about the measurements. She was very careful about the measurements. The January protocol extended the spatial survey to include measurements taken outside the facility building itself, on the roof, in the service corridor to the east, in the parking area to the north.
This required portable equipment, which she had arranged, and it required working in temperatures that in January in Geneva were reliably uncomfortable. She worked in the cold without complaint, because the cold was part of the measurement environment and she had long ago made peace with the measurement environment, whatever it happened to be. The extended survey confirmed what the December measurements had suggested. The DELTA_COH elevation was highest in the northeast corner of the second instrument bay. It decreased as she moved southwest. It was still present, at a reduced magnitude, in the parking area to the north and in the service corridor to the east. It was effectively undetectable in the measurements she took on the western side of the facility. The gradient was real. It was directional. It was pointing northeast. She stood in the parking area on a January morning with the portable measurement apparatus and looked northeast.
The sky above the horizon was the sky above the Geneva basin in January: grey, low, evenly illuminated in the specific way of an overcast sky in a flat valley. Buildings. A stand of bare trees. Beyond the trees, the Jura mountains, their slopes winter-grey. Beyond the mountains, to the northeast, more of Switzerland and then Germany and then the rest of Europe and then, beyond Europe, the sky continuing in the direction of the gradient. She looked northeast for a while. The cold was specific and steady. She went back inside and wrote in her lab notebook: the gradient is directional and consistent with a source external to the facility. Direction: approximately NE (to be determined precisely). No local source identified. Spatial extent of measurable effect: at minimum 120m. Effect present in measurements taken outside the facility structure. She drew a third triangle in the margin. Larger than the first two.