What We Do

Collection and Preparation

To prepare the correspondence for publication, we start with copies of the original manuscript pages. Deciphering these manuscripts is often our first challenge, as was the case with the following pages written to Madison by Tench Coxe, ca. 22 November 1801. (Click on any of the following graphics to see an image of the full page.)

Once we've gathered the manuscripts for a volume, we type up transcriptions and then proofread the typed copy against the manuscript copy. During this process of collection and proofreading, we decide how each document will appear in the volume. If a document is unavailable in other printed sources and/or addresses topics of some importance, we are likely to print it in its entirety. We almost always print in full Madison's correspondence with other major political figures and his rare family letters. We often print as abstracts lengthy missives, previously published material, and letters which deal with routine subject matter. An abstract quotes the entire body of the document, but ignores salutations, complimentary closings, and signatures. We don't usually print documents such as passports, ships' certificates, dinner invitations, and other routine paperwork, unless the document provides key information regarding Madison's whereabouts or important events. Occasionally we discover references to a document that we are unable to find; in this case we prepare a "letter not found" entry describing everything we know about the missing document.


Research and Annotation

After collecting, identifying, proofreading, and sorting the documents we research and annotate them. We try to determine the dates of undated letters, identify the correspondents in unsigned and unaddressed letters, and describe persons and subjects with which we think our readers might be unfamiliar. This step becomes especially important in the Secretary of State and Presidential series, where consular, ministerial, and even domestic correspondents often refer to European, South American, and Mediterranean affairs, to seizures of American ships abroad, and to a broad array of foreign characters and events.

Checking and Editing

Next, all the letters are tandem proofread (research editors reading aloud to the series editor). Then all 1,800 to 2,200 typed pages of the volume are passed to our research editors who check the letters and footnotes for style and accuracy. At this stage, we make trips to the National Archives, the Library of Congress, and other repositories to check questionable readings, dockets, and marginal notations not clearly visible on our copies. The manuscript is then given to the editor-in-chief for review, and finally to the copy editor to prepare the manuscript for the press. After the entire volume has been checked, a clean copy is printed out, proofread again, then delivered to the University of Virginia Press.

To Press!

After an initial reading by our editor at the press, the manuscript goes to the printer to be typeset. Within a few months we receive page proofs, which we proofread against our in-house copy to catch any errors that might have slipped in during typesetting. The corrected pages go back to the press, along with the index which we have prepared, checked, and proofread. Approximately two months after that, we receive page proofs of the typeset index to proofread against our in-house copy, correct, and send back to the press. During these months any further corrected manuscript pages are also sent back and forth between us and the press until they are approved and finalized.

About nine months after the press process began, the finished volume is released and review copies are sent out. (The printed pages above exhibit the final, printed version of Tench Coxe's handwritten letter shown at the top of this page.) In the meantime the process has started all over again for the next volumes; typically we have three volumes in production at any given time.

Special Cases, including Code

This is a lot of work! And in some cases the process is even more complicated. Some documents—those that are badly defaced, torn, undated, misdated, or unsigned, for instance—require special attention. Included in this category of "special cases" are letters written in code.

James Madison, Thomas Jefferson, and others used codes for a variety of reasons. Mail was more insecure in the eighteenth century than it is today. In the late 1790s especially, private correspondence that went through the U.S. mail was subject to inspection by postmasters, many of whom were Federalist in their politics. A common complaint at the time was that letters were tampered with and their contents sometimes published in the newspapers. To avoid this fate, Madison and Jefferson often sent their letters unsigned or partially in code.

Sensitive diplomatic correspondence was also encoded, as well as being sent in multiple copies by different routes. Private letters sent overseas were also subject to seizure in those uncertain times and were often written in code.

Here's an example: the letter on the left (below) was written in code by James Monroe to Madison on September 8, 1795, when Monroe was U.S. minister to France. In this case we are lucky enough to have a copy of the code key (below on the right) shared by Madison, Monroe, and Jefferson. This is a simple number code, in which letters, words, or combinations of letters are represented by numbers. Madison's clerk used this code key to supply the words that appear written above the numbers Monroe wrote in his letter.

Coded letters of that time, all laboriously encoded and decoded by hand, are littered with errors. As editors, we want to know the writer's intention as well as the message received (that is, what the writer encoded as well as what the receiver decoded), so we decode the letter again ourselves, using the key, and print that version (see below left). We compare our version with the clerk's decoding and record any discrepancies in the annotations that follow the letter in the published volume (see below right).


In some cases, we do not have a code key. For instance, to interpret the diplomatic code used by Charles Pinckney when he was U.S. minister to Spain in the early 1800s, we must rely on the decoding done at the State Department. In such cases we have to reconstruct a code key as best we can from the existing decoded letters.