Two Ways to Keep a Song

This month, the Library of Congress is offering two accounts of preservation.

The first fits inside a piece of metal about as tall as a quarter. It contains synthetic DNA encoded with digital copies of collection materials: Thomas Jefferson's draft of the Declaration of Independence, Francis Scott Key's handwritten lyrics for “The Star-Spangled Banner,” the L'Enfant plan for Washington, congressional records, a family tree containing more than 1,500 names, and 1890 Passamaquoddy audio recordings. The device has been placed in a time capsule in Philadelphia. If the plan holds, someone will open it in 2276.

A separate chip carries instructions for decoding the DNA.

The second account fills a gallery. Alive in Many Hands, the American Folklife Center's 50th-anniversary exhibition, brings together recordings, photographs, instruments, recipes, games, tools, and objects made through communal life. There is a banjo signed by more than 260 people, the sewing machine used to join panels of the AIDS Memorial Quilt, and a palmetto basket from a Houma tradition that Janie Verret Luster helped revive.

The exhibition's premise is expressed in motion: traditions endure when people continue to practice, adapt, and pass them on.

Both projects are trying to keep something available to the future. They obey very different rules.

For molecular storage, waiting is a feature. Digital information is translated into sequences made from DNA's four bases, written into many short strands, and sealed for long-term keeping. Addresses help a future reader locate the strands belonging to a file. Redundancy and error-correcting codes make it possible to reconstruct data when some strands are damaged or missing.

The most promising version is deliberately quiet. A technical review of DNA stability in data-storage systems describes archival DNA as something close to “write once, read never”: a preservation copy protected from humidity, temperature changes, and repeated handling, while more ordinary working copies absorb the wear of use. Accessing the material can introduce loss. Leaving it alone can help it last.

A living tradition has the opposite problem. Leave it alone for 250 years and the record may survive while the practice disappears.

A recording can preserve the sound of a song. It cannot, by itself, preserve the moment when somebody decides to teach it, the relationships that make the teaching welcome, the language that holds its meanings, or the authority to say when and how it should be heard. Those things continue through use. Repetition does not degrade the tradition in the way handling can degrade a storage medium. Repetition is one of the ways the tradition stays alive, even though every repetition may also change it.

The distinction is not an argument against archives. Human chains break. People die. Communities are displaced. Languages are suppressed. Material records can hold open a route that history tried to close.

The Passamaquoddy recordings inside the DNA capsule make that relationship especially clear. The wax cylinders lasted long enough to carry voices from 1890, but their recent life did not come from storage alone. In a collaboration involving Passamaquoddy elders, the recordings were sonically restored, reviewed, transcribed in the community's language, and given richer metadata. Traditional Knowledge labels added cultural protocols and attribution to the catalog. The work included digital repatriation: making the recordings available again to the people whose history they contain.

The archive had not failed because it needed that work. The archive became useful because it could meet that work.

This is where the tiny vial's futuristic confidence becomes productively incomplete. A decoder can explain how to turn molecular sequences back into files. It cannot decide what the files mean, which descriptions have become obsolete, whose permission governs their use, or what a future institution owes the people represented inside them. Technical readability and cultural legibility are different preservation problems.

Institutions often speak about durable storage through capacity, format, migration, and lifespan. Those are real obligations. A file that cannot be recovered cannot rejoin anyone's life. But long-term preservation also needs a living account of stewardship: who can revise a description, how communities remain connected to the record, where consent can change, and what forms of access support renewal rather than extraction.

That work does not fit neatly inside the medium. In fact, some of it should not be frozen. The American Folklife Center's documentation guidance treats archives, public programs, fieldwork, and community participation as parts of the same mission. The collection is not only a container. It is also a set of relationships maintained around what the container holds.

There is a temptation to give the two preservation projects separate moral roles: the machine keeps the data, the people keep the meaning. Reality is less tidy. People designed the encoding, chose the materials, wrote the decoder, built the capsule, and will have to find it. Instruments and archives help people remember across distances no single chain of teachers could cross. Preservation is technical and relational all the way down.

Still, the contrast changes what counts as success. The DNA device can be judged by whether its files return intact after 250 years. A living tradition cannot wait until 2276 for its first test. Its continuity is being decided in every present act of teaching, practicing, describing, correcting, and making room for another person to carry it.

When the capsule is opened, a future librarian may be able to turn the molecules back into an audio file. The harder question will not be answered by the chip: who can tell them what they are hearing?

That question continues in Seven Degrees of Commons, an experiment that turns one Wikipedia article into a visual path through Wikimedia Commons. Its collage only advances when the archive can name a relationship—and leaves the break visible when it cannot.