Baxnet Ideas · Founder note
Napoleon Had Letters. We Have an Export Problem.
TL;DR: Future historians may inherit an enormous quantity of private communication that survives without enough structure or context to be understood.
Napoleon’s letters to Josephine are now historical material. They are used to study a marriage, a public figure, a period of war, and the private voice that existed beside the official one.
It is tempting to imagine that our messages will do the same job for the future, only in far greater detail. We leave behind ordinary conversations almost continuously: family logistics, political arguments, courtship, gossip, illness, work, jokes that make no sense outside one friendship.
The future should be spoilt for material.
I suspect the opposite problem is coming. Future historians may inherit an enormous quantity of communication that survives badly: detached from its software, stripped of context, difficult to authenticate, and shared by people who never agreed on what should happen to it.
A letter can outlive the system that carried it
A paper letter remains a legible object after the postal system that delivered it has changed. It can move from a drawer to a family collection, then to an archive. The envelope, address, paper, handwriting, date, and physical sequence all help explain what it is.
That does not make a historical correspondence complete. We see what happened to survive and what later collectors considered worth keeping.
The scale of that work is visible in the Fondation Napoléon’s General Correspondence project. Its print edition ran to 15 volumes. When the first volume was put online in 2022, the original 2,283 annotated letters had grown to 2,647 as more material was discovered and corrections were made.
The archive is still changing two centuries later. What we can read now is the result of repeated human decisions to keep, find, authenticate, organise, and explain.
Our chats start with far more material. They also depend on far more machinery.
A chat depends on its machinery
The text of a message is only part of the record. A useful chat archive may also need its sender, timestamp, reply relationship, reactions, attachments, edits, deleted-message markers, participant changes, and system events.
Without that structure, an export can preserve every word and still lose the conversation.
Exactly means very little if the message it answered has disappeared. A photograph may have prompted thirty replies but sit in a separate media folder with a generated filename. Sam left could explain the abrupt silence that follows, unless the export drops membership events as interface clutter.
This is already familiar to archivists working with email. US National Archives guidance says that, like paper letters, email can document events and relationships. It also warns that email programs are not built for long-term preservation, and recommends saving messages separately with attachments and metadata such as the sender, recipient, and time. Its advice is practical because keeping the account is not the same as keeping an intelligible record.
Chats make the problem messier. The National Archives’ own electronic-message guidance recognises that some messages lose their value when separated from the wider platform or records that supplied their context.
A historian in 2126 might have 80,000 lines of a political conversation and still be unable to tell who was in the room for the important week.
The group chat has no single donor
There is another difference between a set of love letters and a modern group chat.
A 17-person family conversation has 17 contributors. One member may be able to export it, but that does not make every story, photograph, voice note, and vulnerable disclosure theirs to place in an archive. The same problem appears in school groups, campaign rooms, workplaces, and old friendship chats.
Who gets to decide that a shared conversation has historical value? Would the archive remain closed for fifty years? Could people remove their own media while preserving the surrounding sequence? What happens when one participant wants the record kept and another wants it gone?
There will not be one clean answer. Preservation and publication are different decisions, and a responsible archive needs room between them. Material can be kept without being searchable by the public. Access can be delayed, restricted, or refused.
This matters for products that help people understand their message history, including Mimoto. A useful export should remain under the person’s control while carrying enough structure to be intelligible elsewhere. Helping somebody preserve their own history does not quietly grant permission to expose everybody inside it.
Napoleon’s letters survived because people kept making decisions about them. Our messages will need decisions too. Leaving billions of words behind is the easy part.
Further reading: The Correspondence of Napoleon, Preserving Your Digital Memories, and National Archives guidance on electronic-message context.