Problems of provenance

Yes, I also write about History (capitalized – meaning the discipline, not the hobby).

And as a historian, I teach my students the difference between primary sources and secondary sources.

A primary source is one created at the time one is studying. A secondary source is created later, about the subject one is studying.

This isn’t as clear-cut as it sounds.


Take, for example, the Scopes trial of 1925, in which John Scopes, a high school biology teacher, was tried for teaching evolution, which was against the law. The transcripts of the trial are a primary source. The Wikipedia article about the trial is a secondary source. Ray Ginger’s book “Six Days or Forever? Tennessee v. John Thomas Scopes”  is a secondary source.

Or is it? It was written in 1958.  It is a secondary source if you’re studying the Scopes trial. But if you’re studying the arguments about evolution during the 1950s, then Ginger’s book is a primary source.

This is the problem with using primary sources to teach history. What era or subject are they primary for?

So I’m looking at Net Texts, since I’m always seeking good resources, and went to their section on Revolution and the New Nation, which mentioned primary source documents, and I noticed that some of them weren’t, at least not for the American Revolution – several were images from the 1820s supposedly portraying events from the Revolution. They seemed to get around the issue by putting “1820s” as the end date of the unit, but I felt that was deceptive. I went to the source, which they claim was the National Archives, assuming that the site would be more reliable in this regard.

It wasn’t. Throughout the National Archives teaching resources, primary sources were emphasized but many weren’t primary to the issue being studied. In many places they were fine, but in some they were later interpretations or were presented with no date so that students couldn’t tell whether they were primary or not.

(From the Archives, a portrayal of Washington General George Washington and a Committee of Congress at Valley Forge. Winter 1777-78. Copy of engraving after W. H. Powell, published 1866 — Powell was born 24 years after Washington died.)

The year after Washington’s death, Parson Weems published the first hagiographic account of the general and President, not only introducing the cherry tree story but glorifying Washington’s character in general. In the early years of the U.S., such moral tales were important to establishing an American identity and ideology. William Henry Powell’s mid-19th century work is more part of this traditional than of the actual events of the Revolutionary War.

In other words, Powell’s painting is a primary source for the 19th century, not 1777-78.


This is not an idle point. Consider, for a moment, movies. Movies are frequently used to teach history. I used to show Hollywood film clips to illustrate historical events if I felt they were portrayed accurately. But over time I realized that these were being seen as primary sources for the era being portrayed, rather than the era the film was made, and that I was encouraging bad history by showing them in that context.

In other words, the American movie Reds (1981) says more about our revision of communism in the 1980s than it does the Russian Revolution. Going back further, the film Battleship Potemkin (1925) says more about the 1920s than the 1905 mutiny it portrays, and Birth of a Nation is about attitudes in 1915 rather than during Reconstruction. The Battle of the Bulge (1965) is about our ideas of WWII as seen from the 1960s. Lawrence of Arabia (1962 – set during WWI), Charge of the Light Brigade (1968, set during the Crimean War), Zulu Dawn (1979 – set in 1879 South Africa) are about the British dealing with the history of their Empire in the 60s and 70s, once it was lost. Breaker Morant (1980, set during Boer War) and Gallipoli (1981, set during the Great War) show Australian culture coming of age during the 80s. Same thing with films from every nation.

That’s why there are so many remakes (such as Mutiny on the Bounty in 1935, 1962, 1984).  Each generation takes the stories and interprets them in light of current issues and concerns. It’s a signifiant contextual distinction.

The Web

And it’s very tricky now with the web. NPR recently reported on the first web page (Tim Berners-Lee’s demo page from 1990) having been lost. Very few web pages have what we could call provenance (in fact, the term for web purposes seems to refer to accuracy rather than history).

Original web creations will be difficult to date, and thus difficult to use as primary sources. A meme like Selleck Waterfall Sandwich will be difficult to use. One has to rely on secondary sources to date its origin at 2010. But when I search on Google and set dates further back, it pops up in web pages referring to the meme as far back as 2003.

How will we reliably date less popular items? Anyone using Google to check a student paper for plagiarism knows that tons of text has been copied from website to website – it is impossible to tell the origin of many passages of writing. There is often no “date of publication”.

Not that I’m putting Selleck Waterfall Sandwich on a par with a painting of George Washington or a note written by Thomas Jefferson.

But you get the idea.

1 comment to Problems of provenance

1 comment to Problems of provenance