One of my favorite courses to teach is my department’s class on historical methods. For a final assignment in the course, students have to create a bibliography of sources that could support the writing of an upper-level seminar paper, and annotate several of the entries. My students often ask how many primary sources is enough, and they’re never very satisfied by my honest answer of “it depends.”

I started thinking about this from their perspective. Surely there’s a threshold historians should be able to articulate that represents “enough”? And why do we make such a big deal about primary-source research anyway?  Why isn’t it enough to build a composite picture of the past from secondary sources alone? Or from a lot of secondary sources with a smattering of primary sources here and there?

I went looking for an article that would answer these questions, and I couldn’t find one. That’s probably because to professional historians the answer seems so obvious as to not need any clarification. But to students starting out as practicing historians, the answers are not particularly clear.

(We should be sure we’re all on the same page about what primary, secondary, and tertiary sources are. This resource, from the Library of Congress’ Teaching with Primary Source program, is a good place to start.)

Primary sources are left to us from the era we’re examining. They were created in (or close to) the same moment as the events we study – they’re oral traditions, letters, diaries, account books, laws, judicial records, birth certificates, death records, journals, commonplace books, maps, paintings, sketches, etchings, prints, photographs, chairs, milk churns, carriages, planes, fossils, textiles, and debris (to offer an incomplete list.) Because these sources offer us tiny glimpses of the past as it happened—fragments of thought, feeling, opinion, process, and effect—we gather as many of them up as we can to try and piece together a holistic picture of not only what happened and to whom, but how and why it happened at all.

Primary sources might once have been meant for public audiences—like propaganda posters, or newspapers—or they might have been things that were private—love letters, confessions, stories meant only for one community, but not another.  They are a fraction of everything ever created about a place and time.  They tend to record the voices of those who had power, time, and/or money. They favor durable goods over flimsy objects. In Western society they favor the literate over those who couldn’t read and write.

It’s not enough for someone to have created something. Someone else had to see value in that something and want to keep it safe because of that value. Someone had to pass it along to someone who could place it in their memory, or in a book, a library, an archive, or sometimes someone’s attic. People had to file the something, or conserve it and display it, or repeat it to others so that it wouldn’t be lost. Value judgements were made over and over, and some things were thrown away, torn up, sunk, forgotten, and burned.

We try to write the story of the past based on what is left.

It isn’t as simple as looking or listening to what’s left to us and believing it. We have to put the thing into the context of its time, and ask questions about why it was created, and why it survived. Often, we may vehemently disagree with the politics, values, ethics, and ideas of the people who wrote or created or stamped or painted the things with which we, as historians, interact. We come to this work with our own ideas about what’s valuable in life, and whose voices matter, and we have, as a profession, proved willing and able to ignore large swathes of humanity because we didn’t examine our own biases when we came to do the work.

We can’t sit down with our forebears and ask them questions directly, but we can sift through the fragments of evidence they left behind and try and reconstruct what they meant, why they did things, who they treasured, what mattered to them, what didn’t matter at all. We have to look high and low for every fragment we can find. We have to think seriously about where to look. We have to think carefully about how to look. And it’s only when we’ve done that that we can begin to weave together a story about the past. That story becomes the secondary source, written or crafted or told or shared at some distance from the original event. We need both in order to do our job, but the principle is the same each time: learn from the people who went before you. Listen to as much as you can.

That’s why the answer to “how many primary sources do I need?” is “it depends.” And at its heart that answer cannot be about a number. At its heart the answer must be about justice—doing justice to the subject at hand, but also practicing justice in the world we’re building now. How do we recover as many voices as we can? How do we get the stories right? By diligently searching for all the fragments; by listening to the quiet places; by thinking about who we are, and what we’re doing this for.