Historians study the past, but they work with the historical record. The historical record is whatever we have from the past. Thanks to powerful destructive forces like fire, war, recycling, shredders, and biological degradation, very little of what was produced in the past exists in the present. Good historians know how to work with the historical record and form histories, despite the massive gaps in the record.

In recent years the field of history has benefitted from new ways of reading the historical record, but in my mind there are still several basic types of records. There is always an ongoing process of discovery, but this takes time and money, both which are in short supply.
- The intentionally preserved. Libraries, museums, archives, time capsules, my mother’s photo albums, etc. are created with the intention of preserving records for the future. They are expensive, which means someone must be incentivized to fund them. Think of the cost of supporting authors, building libraries, staffing them, and then maintaining them year after year. Thought is put into what will be preserved and what it will tell future people. As time goes on, additional funding is required to continue the preservation, which requires additional incentives to preserve certain parts of the past and discard other parts. Many of the sources that are preserved are created with the intention of influencing people’s understanding of the past. History books, textbooks, sagas, religious texts, etc. create ideas that are considered worth preserving. Archives tend to choose documents that are important for giving a deeper idea of what happened, although the document were not created for posterity . Museums are highly selective in what they can keep. So these sources are, in a way, passive, but there is a lot of active thought into deciding what will be preserved.
- The randomly preserved. Bodies that fall into a bog happened with no thought of future interpretation. Many archeological sites around important buildings dig up all sorts of ordinary items like tools and kitchenware that were accidentally lost. People who abandon buildings don’t do it with the intent that future historians study their home. Garbage dumps are perhaps the best example of things that were not meant to be preserved. Ships are built with the intention that they float, but when they sink they become treasure troves of historical information.
- The revealed. New technology gives us new ways of seeing the past. Drones, infrared cameras, DNA testing, sonar, have revealed things that could not be seen by a human on the ground. We have had amazing nuggets of historical information in our genes since the dawn of mankind (the one time this phrase was used properly!) but until Crick, Watson, and Franklin came along we had no way to access it. Tree ring interpretation is a field that reveals a lot about prehistoric times using information that is readily available almost everywhere, but it took a lot of environmental science to really figure out how to use it.
- The disturbed and the private. Many graves that have revealed a wealth of historical information had sacred significance to the people who created them, and they were likely meant to be left alone. Private diaries or letters that are kept after someone passed away could be included here if they violate the person’s privacy. Information taken by force can be included here. This area can have ethical considerations for how we obtain information from them.
- The inherited. Family histories and family memorabilia fall into this category. I have much of my grandfather’s history in my head because he gave it to me. much of the inherited historical information is never written down, and thus easy to lose.
Of course this is a simplification, but I think these categories encompass a lot of types of sources. There are always hybrid sources, and sources that change over time. For example, the pyramids of Giza were built with the religious importance of providing an eternal link for Pharaohs between this world and the afterlife. As Egyptian society change they lost their original significance, but they were randomly preserved because nobody had the time or energy to dismantle them.
Students of history are usually quick to understand that the historical record is limited, but they often struggle to understand how to form cohesive historical arguments with the limited record. A good history should align closely with the historical record, and do its best to describe the past by interpreting the record we have to the best of our ability. Source analysis often requires an in-depth knowledge of the source creators and of the forces that preserved the forces. Religious iconography can look like pictures drawn by children if you don’t know how to interpret them, or they can be worth ten thousand words per square inch. Government sources can sound like they are packed full of catch-phrases and government-speak, or they can reveal underlying governmentalities that influence how they reach decisions.
The historical record is also subject to tampering and damage. Qin Shi Huang destroyed all records that predated him so he would be at the start of history. Every time the Pope moved between Avignon and Rome a good chunk of the Papal private library was lost on the way. Every Great Library of the ancient world was eventually destroyed by war or lack of funding. Governments love to classify information to influence the record.
The preservation of the record is beautifully random, which has fundamentally influenced our histories. The humid Amazon rainforest has nearly wiped out all knowledge of peoples who lived their five hundred years ago, but a wealth of information ten times as old in the Fertile Crescent and the surrounding arid regions has been preserved by the dry desert. The discovery of a single source, the Rosetta Stone, opened up ancient Egypt to modern scholars, but decades of work on the Minoan Linear A script has yielded next to nothing.
If we had a perfect historical record, where every action of every human, animal, plant, and other living and non-living entity was on record, historians wouldn’t exist. We would just have data analysts.