From the Papers: A case study in transcription

The Mary Baker Eddy Papers is working to publish online the correspondence and manuscripts of the Christian Science founder. It is a multi-year project in which we are systematically processing documents from oldest to newest. We began with the year 1872 and are currently working on documents from before 1872, as well as those from 1886.

One of the first steps in this process is transcription verification, in which handwritten documents are digitized. The goal at this step is to create an electronic version that is as exact as possible. Each handwritten word must be identified correctly and faithfully. All spelling, capitalization, and punctuation that the writer used in the original document must be replicated.

Sometimes this process is complicated by various factors—the documents themselves may have deteriorated, ink may have faded, handwriting may be idiosyncratic or difficult to discern because of cross-outs or erasures, and spelling and phraseology may be creative or archaic. All of these factors can make certain words puzzling.

Our transcription verifiers employ various strategies for accurately teasing out these tricky words. These include:

• looking for context cues
• comparing the letters in the word in question with those in easier-to-read words in the document
• studying other correspondence by the same writer
• changing the magnification or resolution of the scanned document

If we still cannot determine a particular word with a high degree of certainty, we use the notation “[ILLEGIBLE]” in its place within the transcription. We know that readers will always be able to see the handwritten word for themselves in the original document, because a scan of it is published side-by-side with its transcription. But using “[ILLEGIBLE]” is nevertheless a last resort; before we do it, the whole Papers team workshops these words in weekly team meetings, attempting to figure them out together. Often we can. We save this “find the word” exercise for the end of each meeting, and the process is collaborative, effective, and fun.

The primary purpose of this thorough and painstaking verification process is to replicate the contents of each document with as high a degree of accuracy as possible. But occasionally there are other benefits. Recently it led to the discovery that one particular document was not what we originally thought it to be.

We were working on a poem assumed to have been written by Mary Baker Eddy.¹ Records indicated she sent it to her friend Augusta Holmes Swasey in 1838. The handwriting appeared consistent with other examples of her handwriting from this period. Yet it was crafted with many embellishing curlicues and flourishes, and over the intervening 182 years the ink had become faint in many places.

As we embarked on the first stanza of the poem, the following exemplifies the thought process of the initial transcription verifier:

What’s the brow, or the eyes lustre, or the…the…what’s that next word? stir? step? of air. Hmmm…stir of air makes more sense than step of air. And there’s what looks like an i dot over the third letter. So stir is a possibility. But what about the last letter? It could be an r with a flourish, like the one at the end of air. But it resembles the p at the end of the word lip several lines down. Hmmm.…

The word was a candidate for the team to workshop at our next meeting. After puzzling over it, the group concluded that the weight of available evidence pointed toward the wording stir of air.

But that wasn’t the end of the story. The initial transcription verifier felt inspired to spend a few moments digging a little deeper. It seemed certain that the word in question was either stir or step. While stir of air did make more objective sense, maybe step of air had some archaic meaning that was simply unfamiliar to our twenty-first century way of thinking.

Googling that phrase alone, however, yielded results having mostly to do with the technicalities of air conditioning, cleaning, or monitoring (for example, “Emission inventory is a basic step of air quality management”). Clearly a dead end. But refining the query to include terms such as literary and poem yielded this search result: “The Heights of Macchu Picchu” (1945), by Pablo Neruda (“… to climb every step of air up to the emptiness …”). This was encouraging. So on a whim, the verifier added two additional terms from Eddy’s letter to Swasey into the search. And up popped the poem “Psyche, Before the Tribunal of Venus,” by Nathaniel Parker Willis (1806–1867). Originally published in 1829, it contained the precise verses under examination.

Willis was a prolific and well-known poet, magazine writer, and editor of his day. He lived in Boston from 1816 to 1831, and during that time became the editor of The Token, an illustrated literary publication known as an “annual.” Annuals of this type included art, poetry, short stories, and essays and were quite fashionable at the time. It was in The Token that Willis published “Psyche, Before the Tribunal of Venus,” which was written about and accompanied by an engraving of the same title by John Cheney (after a drawing by Evariste Fragonard).²

It is not hard to imagine Eddy and her school friends reading, copying, saving, and sharing verses from popular annuals. And it was in this way that we discovered she had not composed the poem our team was working on, as originally thought. Rather, she had excerpted a stanza of “Psyche, Before the Tribunal of Venus” to share with her friend. Although she did not reference Willis, she did employ quotation marks at the beginning and end—which now made sense as well.

Sometimes it’s when the usual methods fail that we’re propelled to make unexpected discoveries. For example, if the words of the poem in Eddy’s letter had been clear from the start, we would have transcribed them correctly, but there would have been no need to look any further. Intuition sometimes provokes a sense that there is more to a story than meets the eye and inspires deeper digging. Also, while archivists and documentary editors have always been motivated to do deep research into the origins of documents, the internet has made it possible to discover in moments what might have previously taken days or longer. In this case, these factors all combined to reveal the true provenance of one particular document.