text {isle}

one small islet in the sea of digital text
April 23rd, 2013 by Heather Asbeck

Encoding Epistles: New Contexts, Old Artifacts

The Flagg Letters

So now that I have (very, very briefly) explained the technological underpinnings of digitizing old documents, I want to talk about an encoding project I have been a part of.  Our university library has a collection of letters shared between a father and son in 1853.  The father, Gershom Flagg, was a farmer in rural Illinois, not too far from St. Louis.  His son, Willard Flagg, was a student at Yale.  He later became a prominent Republican IL state senator, as well as one of the founders of Illinois Industrial University, which later became the University of Illinois.  These letters chronicle the aspects of daily life of a farmer vs. daily life of a student; Gershom tells of the apple harvest and how many barrels of cider have been made, while Willard tells of the people he has met, things he has seen, and the books he is reading, accompanied by occasional requests for more money.  These letters, aside from providing information about the writers, also showcase their shared referents – people, places, other mutual correspondents, and publications such as books and periodicals.

In order to begin the project, as a group we had to decide which letters to choose, and provide a rationale for the choice.
Individual students read the letters and shared the contents with the class, and 18 letters were chosen for the project.  Then high-resolution images of the chosen letters were created, which were subsequently converted to smaller web-ready versions and uploaded into a shared file.  Summaries of each letter were compiled, and each of us transcribed our assigned letters.  We also learned about the TEI guidelines and the XML encoding process.  After being sucked into a whirlwind of information and being swirled around in a cloud of TEI guidlines, XML, elements and attributes, and a constellation of letters which dealt with a wide variety of topics, we had to make some decisions.
What gets encoded?

That is determined by our editorial rationale, and the decisions we make as a group.  (See an example editorial rationale at the Wide, Wide, World Digital Edition website.)  We can decide to code anything (and everything) – names, places, events, punctuation (or lack thereof), linguistic anomalies, and pertinent physical details about the documents such as holes, wax seals, illegible writing, etc.  Since that is a monumental task, we have to decide what matters most for our project’s purposes.  So what is important?  Do line breaks matter?  In poetry, maintaining line breaks is essential; but in a letter?  Probably not.  Punctuation?  Well, in these letters, there was a lack of punctuation – Gershom’s sentences begin with capital letters, but never have periods.  Rather than artificially inserting them, which would require editorial decisions that could affect the meaning of the letters, we decided to leave the (lack of) punctuation as written.  So what else is important?  Because of the nature of the letters, we chose to tag people, places, events, organizations, modes of travel, and media references (other letters, books, and periodicals).  This analysis shifted our focus from looking at the documents as data, rather than solely as literature or historic record.

XML markup

The markup explains how and why something has been coded, defines different attributes (name, date, place, etc.), denotes textual features (line and page breaks, paragraphs, etc.), and allows for cross-references and search functions.

The process of coding is a long and arduous one.  Many lines of code precede and follow this example, but this gives an idea of how the opening information – place of writing, date, and salutation – would be encoded.





<placeName>Paddock’s Grove Ills.</placeName>

<date>27 November 1853</date>


<salute>Dear Willard</salute>



As you can see, there quite a few lines of code for a very simple first two lines of the letter.  As you can see, every item has been labeled with opening and closing tags, which define not only what type of thing it is (placeName, date, etc.), but also what role it plays in the letter (opener, dateline, salutation).

Another example shows how different items may be tagged within an epistolary context.  If you follow the link, the example shows many different types of items being tagged, from people and places to periodicals and poems.

As I worked on the Flagg Project, I discovered that these letters are documents that provide a physical link between father and son, bridging the geographic divide between them.  Books, periodicals, and other media mentioned in them provide a wider context and background information for the letters, as well as a shared frame of reference between father and son.  They read the same newspapers, periodicals, and shared letters from loved ones.  These these texts provided the common context for their writing, and framed their communication.  But as the letters are passed along from father and son, saved by generations of their descendents, and were eventually given to the university library, the chain of custody widens the contextual sphere with each transfer, and each receiver has a new frame of reference and context for viewing the letters.  As the letters are passed along, they serve as a common referent for a successively broader audience.  For an online community of readers, the letters provide the same benefit that the shared media did for the Flaggs: context.  By creating digital media versions of the letters, the digital screen becomes a shared contemporary context for perusal.  These embedded shared contexts are like nested Russian dolls, with ever widening contextual spheres, but also with some gain and loss with each changing context.  Each new audience contributes a different contextual understanding, and a redrawing or widening of the boundaries of the letters’ community or public.

nesting dolls


Leave a Reply

Your email address will not be published. Required fields are marked *