The Quote-Back
Keza Paredes ran the clip for the thirty-seventh time at two in the morning, sitting on the cabin floor of the research vessel with her father’s headphones around her neck and the laptop propped against her knees. The clip was nine seconds. The first four were her father’s codas, recorded in 2019 off the coast of Dominica, and the last five were the matriarch’s reply. The reply began with his first coda and ended with something that was not.
She closed the laptop. The cabin smelled of diesel and the wet rope smell of the lines, and somewhere below her the generator ticked. She put the headphones on the bunk and lay down. She did not sleep. At 05:30 the alarm went off and she got up and put on the same clothes she had worn the day before, because the rotation was two shirts and she had not done laundry.
The matriarch was called Marengo by the crew, which Keza hated. Not because the name was wrong but because it made the animal sound like a grandmother in a storybook, and the matriarch was not that. She was roughly forty years old, the largest female in the clan, and she carried a scar across her dorsal fin from a propeller strike in 2011. Keza had her file. She had her father’s file on the same animal, written in pencil in a waterproof notebook, and the pencil marks were still legible because her father pressed hard.
The trial had started on a Tuesday. Fourteen days, alternating transmissions. Day one: the father’s codas, recorded in 1998, broadcast at 06:14:00 on the matriarch’s usual opening slot. Day two: the current EC1 dialect, the living one, broadcast at the same time. Then back. Keza had two hundred and seventeen transmissions in her archive and could have run the analysis in an afternoon. She told herself she needed fresh data. She told herself the archive was contaminated by her own expectations. Both things were true. Neither explained the nights on the cabin floor.
She ran the trials instead.
On day one she transmitted the father’s codas and the clan’s response lasted eleven seconds longer than their response to the EC1 dialect on day two. The rubato match, the timing overlap between her transmission and their reply, was tighter by 0.3 seconds. These were publishable numbers. They would vindicate her father or they would not, and either way the paper would be clean.
She sat in the cabin and looked at the spectrogram. It felt like leaning against a door that opens inward.
On day three, the matriarch produced a novel tail. In sperm whale acoustics, a tail is a coda appended to the end of a standard reply, a kind of postscript. Novel means it does not match any coda in the published catalog for this clan. It is new. It happens rarely. It is the kind of thing that gets a paper into Nature and a researcher into trouble, because a new coda implies learning, and learning implies a teacher, and the question of who taught whom is not a question you can answer with a spectrogram.
Keza transcribed the novel tail against the father’s cassette. She did this at three in the morning, on the cabin floor, with the headphones on. The novel tail was four clicks. Three short, one long. The spacing between the third and fourth click was 0.41 seconds. She scrolled through the father’s 1998 recordings and found the same spacing. Same coda. Same animal, possibly. She could not tell. The father had recorded it and had not transmitted it, because in 1998 he was not transmitting. He was only recording. He was a listener.
The coda appeared twice in the father’s recordings, both times from the same bearing, both times at the end of a long exchange with the matriarch. He had noted it in his field book. Keza had read the field book. The note said: “Matriarch, coda, novel, possible variant of 1+1+3. Worth following up.” He had not followed up. He had drowned in 2019, and the notebook had come to Keza in a box with his sweaters.
She played the clip again. Nine seconds. Four of him, five of her. The reply began with his first coda and ended with something that was not.
The father’s name was Tomás Paredes. He had been a fisherman off Dominica, and in the summers of 1998, 1999, and 2000 he had recorded sperm whales with a hydrophone and a cassette recorder. He had a theory. The theory was that sperm whale codas were not just clan markers, not just “I am from the Eastern Caribbean 1 clan,” but individual signatures, and that a human who learned to produce them could be recognized as a member of the clan. He had called this kinship. He had used the word in a paper in 1999, and the paper had been received with the polite skepticism that the field reserves for claims it does not want to examine.
Keza had spent her career not being him. She had gotten her PhD at Woods Hole. She had published on coda structure in the Pacific. She had been careful. She had not used the word kinship. She had not used the word recognition. She had used the word response, which is measurable and does not imply anything about what the whale thinks.
She had also inherited his tapes. All of them. Three summers on DAT cassettes, three summers on the early digital recorder he had bought in 2001, and two summers on a hard drive that she had found in his desk. The hard drive contained the 2019 recordings, which were the ones she had played on the cabin floor. The 2019 recordings were the ones where he sounded tired. She could hear it in how he counted the clicks under his breath.
She had not planned to come to Dominica. She had planned to run the analysis on the archive and publish the differential-response result and be done with it. But the archive was not enough. The archive was his, and she could not publish from his data without becoming him, and she could not publish from her own data without coming here. So she had come here, and she had set up the trials, and now she was sitting on the cabin floor at three in the morning listening to her father count clicks.
On day five, Keza called her colleague Priya from the satellite phone. Priya was at Woods Hole, and she was the only person Keza had told about the quote-back. Priya had said, “Don’t publish until you know which direction it went.”
“Which direction what went?”
“Whether the matriarch learned it from him or he learned it from her. That’s the whole question, Keza. If she learned it from him, that’s contact. That’s a human teaching a whale a coda. That’s your father’s kinship claim, and it’s extraordinary. If he learned it from her, that’s your father quoting a whale, and it’s a footnote.”
“It’s not a footnote.”
“It’s a footnote if he was just recording. It’s the paper of the decade if he was transmitting.”
“He wasn’t transmitting in 1998. He was only recording.”
“Then it’s a footnote.”
Keza hung up. She did not call Priya again for six days.
The field notes were in a box in the cabin, under the bunk. Keza had brought them because she could not leave them. She had not read them since the first time, when she was twenty-three and had just started her PhD and needed to know what her father had been doing all those summers. She had read them then with the contempt of a young scientist for an old one, and she had not read them since.
She read them now. The 1998 notebook was pencil on waterproof paper, and the pages were stiff with salt. The handwriting was small and careful, the handwriting of a man who had learned to write in two languages and chose the smaller one for field notes. She read about the weather. She read about the boat, a wooden pirogue named Marie, and the captain, a man called Toto who charged by the hour and who Keza had met once when she was twelve. She read about the codas. She read about the matriarch.
On July 14, 1998, her father had written: “Matriarch, coda, novel, possible variant of 1+1+3. Worth following up.”
Keza read this twice. She read it a third time. Her father had not produced the coda. He had recorded it. He had written that he had not produced it.
But the coda was in his recordings. It was in his recordings twice, both from the same bearing, both at the end of long exchanges. If he had not produced it, then the matriarch had produced it, and he had recorded it, and then two decades later his daughter had transmitted his codas and the matriarch had produced it again.
This meant the matriarch had learned it from herself. Or it meant the coda was not novel at all but a clan variant that appeared rarely and that Keza had not cataloged. Or it meant the matriarch had learned it from another whale, a whale that had learned it from the father, which required the father to have been transmitting, which he had not.
Keza sat on the cabin floor with the notebook in her lap and the laptop beside her and the headphones around her neck, and she thought about which of these was true, and she could not decide.
On day eight, the matriarch produced the novel tail again. Same coda. Same spacing. Keza transcribed it and it matched. She sat in the cabin and looked at the two spectrograms side by side, the 1998 recording and the 2026 recording, and the click patterns were identical within the resolution of the hydrophone.
She sat with it. If the matriarch had learned the coda from the father in 1998, she would have had to learn it from a transmission, and the father had not been transmitting. He had been recording. So she had not learned it from him. She had produced it herself, in 1998, and she had produced it again in 2026, twenty-eight years later, in response to hearing the father’s codas. She let it go.
Keza went to the hydrophone. She sat on the deck in the dark and listened. The clan was clicking somewhere to the south, and the clicks came through the speaker like a fire in another room. She listened for an hour. She did not hear the coda. She went to bed.
On day ten, a funder called. The funder was from the National Science Foundation, and his name was Gerald, and he wanted to know when the differential-response paper would be ready. Keza said the data was still coming in. Gerald said the grant ended in December. Keza said she understood. Gerald said, “Keza, this is a good result. The clan responds more strongly to the dead dialect. That’s the paper. Don’t sit on it.”
Keza said she would not sit on it. She hung up. She had not told Gerald about the quote-back. She had not told anyone except Priya.
She sat in the cabin and thought about publishing. The differential-response result was clean. It was publishable. It would vindicate her father’s method, if not his theory. It would say: the clan responds to the father’s codas. It would not say: the matriarch quotes him back. It would not say: the matriarch produced a coda in 1998 that the father recorded and did not transmit, and produced it again in 2026 in response to his codas. It would not say: your father was, in some acoustic sense, part of their world.
Keza could publish the differential result and bury the quote-back. She could do this. It would be easy. The quote-back was a single coda in a single tail, and it could be explained as a clan variant, and no one would ask. She could publish the clean paper and go home and not be her father.
She did not. Not yet. The days passed.
On day twelve, Keza made a choice. She would transmit the quote-back coda. She would send it to the matriarch, the four clicks, three short and one long, with the 0.41-second gap, and she would see what came back. She queued the coda and sat on the deck in the dark and pressed the key.
The coda went out. Three short clicks, one long, into the water. The water was black and flat and the stars were out and the boat rocked gently and the generator ticked.
Nothing came back for forty seconds. Keza counted. At forty-one seconds, the matriarch replied. The reply was not a single coda. It was a sequence. Six codas, one after another, with the rubato of a full exchange, the kind of exchange that happens between clan members and not between a clan and a stranger. The sequence lasted nine seconds. The last coda in the sequence was the quote-back coda. The matriarch had answered with herself.
Keza sat on the deck and listened. The clicks came through the speaker and they were low and close and they had the rhythm of conversation, and she could hear in them something she had not heard before. She did not know what it was. She sat there and listened to it, and she let it play.
After a while the clicking stopped. The water was quiet. The boat rocked. The stars were out.
Keza went to the cabin. She put on her father’s headphones. She did not play the cassette. She sat on the floor with the headphones on and the laptop closed and the notebook in her lap, and she listened to the silence, and in the silence she could hear the matriarch’s sequence, the six codas, the last one the quote-back, and she could hear her father counting clicks under his breath in 2019, and she could hear the generator ticking, and she could hear the water against the hull.
She took off the headphones. She put them on the bunk. She picked up the notebook and opened it to July 14, 1998, and read the line again: “Matriarch, coda, novel, possible variant of 1+1+3. Worth following up.”
She closed the notebook. She turned off the lamp. She lay on the bunk and listened to the water.
In the morning she would pack the equipment. She would not transmit again. She would not publish the differential result, or the quote-back, or the sequence. She would go home and put the tapes in the box with the sweaters and she would not open them again for a long time.
But that was the morning. Now it was night, and the boat rocked, and the water was dark, and somewhere to the south the matriarch was clicking, and Keza listened.